TAGA Journal

TAGA Journal Technical Association of the Graphic Arts

Volume 4, 2008

Copyright 2008 SWANSEA PRINTING TECHNOLOGY LIMITED All Rights Reserved ISSN 1748-0337 (Print) ISSN 1748-0345 (Online) The TAGA JOURNAL is published by SWANSEA PRINTING TECHNOLOGY LTD on behalf of the TECHNICAL ASSOCIATION OF THE GRAPHIC ARTS Reproduction in any form by any means without specific written permission is prohibited. The technical content of the TAGA JOURNAL is the sole responsibility of the authors and the articles do not necessarily reflect the views of the editor or the publisher. Any product names mentioned herein appear as a matter of information only and do not imply endorsement by the TECHNICAL ASSOCIATION OF THE GRAPHIC ARTS, the editor or the publisher. For complete instructions on how to submit articles for review and possible inclusion in this publication, please see http://www.tagajournal .com SWANSEA PRINTING TECHNOLOGY LTD 7 Glanmor Park Road Sketty Swansea SA2 0QG United Kingdom Email: [email protected]

TAGA JOURNAL VOL. 4

© 2008 SWANSEA PRINTING TECHNOLOGY LTD

Editor’s Note

Tim Claypole Welcome to the fourth volume of the TAGA Journal, the premier peer reviewed Journal for all professionals who use printing and related processes to create products. This represents a milestone for the TAGA Journal. Swansea Printing Technology Ltd took over the publication in 2005. This was a dark time for TAGA with worries about its future and much restructuring. One of the possible victims of this need to focus on core activities was the TAGA Journal. Hence, Swansea Printing Technology stepped in to ensure the survival of the TAGA Journal, as the only peer reviewed publication for printing science and technology. Since then TAGA has recovered its strength, delivering one of the premier annual International conferences for those in graphic applications. Thus, TAGA, now successful under GATF stewardship, is in a position to run the Journal. Volume 5 will be published by GATF on behalf of TAGA. The editorial office, which handles the peer review process, will move from WCPC (Welsh Centre for Printing and Coating, Swansea University) to GATF. Mark Bohan will take over as the editor. Mark as well as being VP Research at GATF has had a long involvement with the Journal. He was one of the group of TAGA members who first formulated the plan for a peer reviewed Journal in San Diego hotel in the early 00’s. There is something of a Swansea continuity as Mark at that time was a member of the academic staff in WCPC. The successful publication of the Journal to date would not have been possible without the work of others. In this regard, I would like to particularly thank my colleagues, Eifion Jewell and David Bould. Many thanks goes to Christine Hammett, the secretary/administrator of WCPC, for the hard work administering the peer review


TAGA JOURNAL VOL. 4

process, particularly chasing reviewers and authors, monitoring progress and formatted the papers. Finally, my thanks is to the authors who have submitted their work and the reviewers who have ensure the quality of the Journal. We at the WCPC pass the TAGA Journal back to TAGA and wish it every success in the future.

Tim C. Claypole Editor TAGA Journal

TAGA JOURNAL VOL. 4


TABLE OF CONTENTS 1 - 10

The use of non-destructive methods in the study of IJ ink penetration Tadeja Muck, Branka Lozo, Lenka Otáhalová, Markéta Držková, Marie Kaplanová

12 - 24

Water-Borne and Solvent-Based Flexographic Inks – Influence on Uncovered Area and Ink Leveling on PEcoated Paperboard Behudin Mesic, Gunnar Engström, Magnus Lestelius

25 - 34

Characterisation of the Lithographic Printing of Conducting Films B. Ramsey, D. Harrison, D. Southee

35 - 58

Grey Balance Control with a Re-Purposed GATF Colour Circle Manfred H. Breede

59 - 71

Diffractive glossmeter for measurement of dynamic gloss of prints Raimo Silvennoinen, Mikko Juuti1, Hanna Koivula, Martti Toivakka and KaiErik Peiponen

72 - 83

The Effect of Paper Properties on the Color Reproduction for Digital Proofing of Gravure Publication Printing Yu Ju Wu, Alexandra Pekarovicova and Paul D. Fleming

84 - 100

Factors Affecting the Printing Strength of Kaolin-Based Paper Coatings J.C.Husband, J.S.Preston, L.F.Gate, D.Blair and P.Creaton

101 - 115

A Comparison of Densitometric and Planimetric Measurement Techniques for Newspaper Printing Maria S. Wroldsen, Peter Nussbaum, Jon Y. Hardeberg


TAGA JOURNAL VOL. 4

117 - 125

The Effects of Paper Coating on Gravure Ink Mileage Curve Renmei Xu, Yu Ju Wu, Alexandra Pekarovicova, Paul D. Fleming, and Michelle X. Wang

126 - 148

From Multi Channel Publishing towards a Ubiquitous Media Environment Maria Åkesson and Carina Ihlström Eriksson, PhD

149 - 164

The evaluation of colour difference equations and optimization of DE2000 Martin Habekost, Dr. rer. nat., Dr. Katrin Rohlf

165 - 177

Effect of Toner Fixing Temperature on Print Properties in the Electrophotographic Process Timo Hartus

179 - 191

Significance of Print Quality in Variable Data Printing Åsa Gidlund, Thomas Mejtoft & Sofia Demnert

192 - 205

Effect of drying temperature profile and paper on mechanical print quality in heatset offset printing Timo Hartus

206 - 218

Factors Impacting the Evaluation of Printer Profile Accuracy Robert Chung

219 - 231

Characterization of Conductive Polymer Inks based on PEDOT:PSS Erika Hrehorova, Marian Rebros, Alexandra Pekarovicova, Paul D. Fleming, and Valery N. Bliznyuk

TAGA JOURNAL VOL. 4


T. MUCK, B. LOZO, L. OTAHALOVA, M. DRZKOVA, M. KAPLANOVA

1

The use of non-destructive methods in the study of IJ ink penetration. Tadeja Muck, Branka Lozo*, Lenka Otáhalová**, Markéta Držková**, Marie Kaplanová** University of Ljubljana Faculty of Natural Sciences and Engineering Chair of Information and Graphic Technology Snežniška ul. 5 SI-1000 Ljubljana Slovenia [email protected]

*University of Zagreb Faculty of Graphic Arts Getaldićeva 2 10 000 Zagreb Hrvatska [email protected]

**University of Pardubice Faculty of Chemical Technology Department of Graphic Arts and Photophysics Studentská 95 532 10 Pardubice Czech Republic [email protected] [email protected] [email protected]

Abstract The final phase in printing process is drying of printing inks on the printing substrate. The mechanism of drying and ink distribution on the substrate surface depends on the ink characteristics as well as on many other factors, like surface energy, roughness, sizing and porosity. The scope of the study is to evaluate the applicability of several nondestructive methods in the analysis of vertical and radial ink distribution on diverse InkJet prints. The dye-based Ink-Jet inks printed on different types of coated and uncoated papers were analyzed. The non-destructive measurement methods were used as follows: Slit-scanning densitometry, photoacoustics and confocal laser scanning microscopy – CLSM. The achieved results were verified by the use of the destructive measurement method consisting of microtome cross-sections in z-direction of the prints and optical microscopy. The destructive method gives a real insight into the depth of penetrated ink and enables the correlation with the studied non-destructive measurement techniques. Keywords: Ink-Jet ink, Confocal laser scanning microscopy, Photoacoustics, Crosssection analysis, Slit-scanning densitometry

1

Introduction

The quality of ink-jet prints depends to a large degree on the interaction between the printing substrate and Ink-Jet ink. The penetration of printing ink should be low in order to retain high optical colour density [1,2]. The surface of the paper also plays an important role: the absorption capacity for printing inks should be controlled. It should not dust and should have a suitable level of smoothness as well as other properties necessary for good operability and printability [3,4]. The efforts have been done in explanation of the ink-substrate interaction by non-destructive methods. A few studies have included the application of different spectroscopic and microscopic methods (Raman Spectroscopy, Fourier Transform infrared photoacoustic


TAGA JOURNAL VOL.4

2

THE USE OF NON-DESTRUCTIVE METHODS IN THE STUDY OF IJ INK PENETRATION

spectroscopy and UV resonance Raman spectroscopy) in the study of ink distribution inside the print [5]. This study extends recent results in investigation of radial and vertical distribution of Ink-Jet printing ink, both on and beneath the paper surface, on different types of papers, and determines the impact of the paper surface on print quality. Consideration of advantages and drawbacks of each method is discussed with the aim to find method that can provide satisfactory insight into Ink-Jet ink penetration in the substrate structure.

2

Materials and Methods

2.1 Types of paper In the study five completely different types of printing substrates were used (Table 1). The reason for the use of such a wide range of samples was to check precisely the limits of the applied non-destructive and destructive measuring techniques. Table 1: The base characteristics of printing samples. Sample mark

Sample description

Grammage Thickness

Z

Zweckform Photo Paper, recommended for Ink-Jet 130 g/m2, 170 μm prints, one-side coated, high glossy

K

Carton board, one-side coated multilayered white 300 g/m2, 430 μm board

I

ICP Permanent Paper, quality corresponds to standard 70 g/m2, 110 μm ISO 9706, no surface treatment

T

Tissue paper, Three-ply with a low percentage of 60 g/m2, 185 μm secondary fibers

A

Scheufelen’s offset standard paper, coated, wood-free, 150 g/m2, 115 μm without fluorescent whitening agents

/

For better understanding and gain of more complex insight in the results of photoacoustic measurements, standard coated paper (mark A) was included in the study. The prints were done on the Canon BJC 8500 ink jet printer with magenta water-soluble dyebased ink. Beside magenta ink, only for the photoacoustic measurements also black ink was used. Printing quality was 1200 dpi with colour intensity 100%. 2.2

Non-destructive methods

2.2.1 Slit-scanning densitometer is usually used in analytical chemistry for evaluation of thin layer chromatograms. The separated compounds could be qualitatively and quantitatively evaluated by measuring the optical density in reflection or transmission mode. Therefore the samples with printed magenta dots were prepared. The diameter of spots and spacing between spots was 3 mm. They were arranged as different lines and tracks. By slit-scanning densitometer both the radial and vertical distribution of printing ink were investigated. Also, the method was used for evaluating the influences of paper on the non-homogeneity of the signals from different tracks and different lines of the models. The testing conditions for measurements were remission mode and a wavelength 580 nm. 2.2.2 Confocal laser scanning microscopy (CLSM) was at first used in medicine, especially in analyzing of pathophysiological samples. In the field of graphic arts the CLSM represents a

TAGA JOURNAL VOL.4



3

non-conventional method for the non-destructive study of the interaction between Ink-Jet inks and the printing substrate. It was used in the investigation of three-dimensional paper structures, the determination of pore volume [6,7] and the evaluation of ink penetration in the z-direction [8]. The type of the analyzed test form was the same as for the Slit-scanning microscopy. The spots of the printing ink were observed at 460 µm × 460 µm scan area. The air objective with numerical aperture 0.6 was used. The laser beam wavelength was 458 nm and resolution in the z-direction 2.7 µm. The sample preparation, i.e. cutting the samples in z-direction was done virtually, by computer software. This software calculated the number of slices necessary to pass from the most fluorescent to the non-visible part of the ink. The measurements were done on printed samples covered by immersion oil, which enabled the proper transparency of the samples. 2.2.3 Photoacoustics is the non-destructive method for the study of optical and thermal properties of a printed area at different depths beneath the sample surface at given wavelength of excitation radiation. This implies the examination of ink penetration into the paper structure. The physical principles of indirect photoacoustic effect and its possible applications in the field of graphic arts are already well established [9,10,11,12]. Magenta and black Ink-Jet prints were measured together with the appropriate blank paper. Samples were cut 11 mm in diameter and mounted on plexi-glass backing. As a reference, the “carbon black” was used. A diode pumping solid-state 20 mW laser (Suwtech, 532 nm) was used as the light source. The generated acoustic signal was detected by a capacitor microphone Brüel and Kjaer type 4166 and preamplifier. The output signal was fed to a DSP lock-in amplifier Stanford Research Systems model SR830. The laser light was modulated using the internal oscillator of the lock-in amplifier. The measurements were controlled and processed by computer. To acquire basic parameters of studied samples, the following methods were used. Optical spectra for optical density and whiteness determination were measured according ISO standard 13655 by a reflection spectrophotometer Eye-One. Paper roughness was evaluated by Parker Print Surface Roughness tester (model 58-04-00-0001). The mass-per-area of all substrates was determined through analytical weighing of 100 square centimetres paper samples (Karl Schröder KG sample cutter) and thickness of the paper sheets was measured with a thickness gauge. The mean density of individual substrates was determined from thickness and mass-perarea values. 2.3

Destructive method

2.3.1 Cross-section analysis is the destructive method which enables a detailed access to the undersurface migration of printing ink through the inside layers of the printing substrate. The samples were immersed into the polyethylene glycol, frozen at −30 °C and covered with Keyser glycerol Jelly. The cross cutting was done by microtome Cryostat Leica CM 1850. The final microtome slice thickness was 30 µm. The slices were then examined by optical microscope Zeiss Option Axioscope and captured by a Canon EOS Digital Rebel camera.

3

Results and discussion

3.1 Non-destructive methods 3.1.1 Slit-Scanning Densitometer The magenta prints of all printing samples were measured line by line with the Slit-scanning densitometer. The measurements were done in remission mode. The peak area value indirectly © 2008 SWANSEA PRINTING TECHNOLOGY LTD

TAGA JOURNAL VOL.4

4


presents the optical density of printed samples. The evaluated standard deviation values show the homogeneity of ink distribution on the printed surface. Table 2: The average peak area and RSD measured by densitometer in remission mode, for magenta ink printed on all four samples. Z1

Z2

K1

K2

I1

I2

T1

T2

average

47802.4

48468.8

41904.5

41602.4

40114.8

38897.3

25010.0

24899.8

st. dev

411.6

546.5

1637.7

1501.5

1133.0

577.1

2730.0

2242.4

RSD

0.9

1.1

3.9

3.6

2.8

1.5

10.9

9.0

Figure 1: The peak area of the magenta prints measured by densitometer in remission mode on all four samples. From the Table 2 and Figure 1 it can be seen that the highest intensity (optical density) of magenta ink is achieved on the Z sample and the lowest value on the T sample. On the Z sample the ink remains on the paper surface while on the T sample it penetrates into the paper. The relative standard deviation (RSD) values show that on the Z sample the printed Ink Jet inks are homogeneously distributed (Table 2). Such characteristics were expected from the paper producer’s specification which guarantees a brilliant colour reproduction. The T sample has the highest RSD value – around 10, because of its rough and open surface structure. The RSD value for the K sample is higher than for the uncoated I sample. This can be explained by the type of coating of the K sample, which is excessively hydrophobic. After the drying process, dye-based ink remains on such a surface in a non-homogeneous layer. Therefore, the K sample is not suitable for Ink Jet printing.

TAGA JOURNAL VOL.4



5

Figure 2: Surfaces of Z (left) and K (right) magenta prints observed by CLSM. To confirm last statement, an image was taken of the Z and K surfaces of magenta prints by Confocal Laser Scanning Microscopy. Figure 2 shows the surfaces of both prints observed by CLSM. The size of the captured slice of the Z sample was 146.2 × 146.2 micrometers in x and y directions. The z direction of the sample presents the thickness, i.e. virtual thickness expressed as the CLSM resolution in z direction, which was 24.7 micrometers for the Z sample. The size of the captured slice of the K sample was 206.8 × 206.8 in x and y directions and the CLSM resolution in z direction of the sample was 12.4 micrometers. The picture proved the conclusions from Slit-scanning densitometry. On the surface of K sample the non-uniformity of ink layer as well as the holes in it are clearly visible. The ink on the Z sample is homogeneous distributed. The micro cracks visible on the surface are used only for ink’s faster fixation. 3.1.2 Confocal Laser Scanning Microscopy (CLSM) The CLSM measurements of vertical ink distribution were done on the magenta printed area for each of the samples. By the computer software the slices in z-direction were virtually cut. The maximum depth of penetrated ink was calculated using the number of slices from the top of the printed area (the most fluorescent part of the magenta dot) to the lowest visible ink trace in the sample (Table 3). Table 3: Results of CLSM vertical distribution of magenta ink. Sample

Aver. Nu. of slices

Slice thickness

Max. ink thickness (µm)

Z

32.23

0.73

23.53

K

32.64

0.74

24.15

I

32.12

0.80

25.96

T

89.51

0.81

72.50


TAGA JOURNAL VOL.4

6


Figure 3: The printed surface of I sample evaluated by CLSM software (virtual slice). Figure 3 shows the example of printed surface evaluated by CLSM software (virtual slice) for the I sample. The stack size is 206.8, 206.8 and 31.3 µm for x, y and z-directions, respectively. The deepest ink penetration is present on the T sample and the shallowest on the Z sample (Table 3). These results correspond to those of the Slit-scanning densitometer. On the Z sample, ink forms a layer on the paper surface; the optical density of such a layer is high. The measured ink thickness for both T and I samples is high, because ink penetrates deeper into the uncoated substrates (Table 3). 3.1.3 Photoacoustics The preliminary measurements of spectra density were done before the photoacoustic measurements. Figure 4 shows the influence of the base sample characteristics on the spectra density changes. 1.4

Optical density of magenta prints

Z 1.2

A K

1.0

I T

0.8 0.6 0.4 0.2 0.0 350

400

450

500

550

600

650

700

750

λ (nm)

Figure 4: Optical density spectra changes for different substrate types.

TAGA JOURNAL VOL.4



7

The optical density spectrum differs for each type of sample, meaning that different colours will be perceived on diverse media. The change in the appearance of printing colour is mainly due to the different ink distribution in z-direction; in addition, in the case of the Z sample the dissimilar low wavelength spectrum is probably due to presence of FWA (fluorescent whitening agents). The highest optical density is achieved on the Z sample and the lowest on the T sample. These results are completely in accordance with both previous applied non-destructive methods. Photoacoustic measurements were done in the modulation frequency range from 10 Hz to 1 kHz and modulation frequency dependencies for both intensity and phase of photoacoustic signal were evaluated for the individual samples. In modulation frequency range, where the absorbing layer is thermally thin, i.e. the signal arises from the whole thickness of the layer, the photoacoustic signal is influenced by structural and thermal properties of the substrate, the so-called sample backing. The theory predicts the decrease of the photoacoustic signal intensity with increasing density of the backing. This is supported by the measured dependence of the normalized intensity of photoacoustic signal on mean density of substrates (Figure 5). Only for the K sample is the value of normalized photoacoustic signal lower than the values for substrates with similar mean density. The value corresponds to the value for the A sample, whose mean density is considerably higher. This reflects the real structure of the K sample, as the main effect on its upper side density comes from coating layer, as is also the case with the A sample. With increasing modulation frequency the influence of the substrate should decrease and the photoacoustic signal should be driven by the absorbing layer (i.e. the impact of ink density becomes more significant). In accordance with this assumption, the dependence on substrate mean density observed for low frequencies diminished when the highest frequencies were used.

Normalized photoacoustic intensity at 50 Hz

0.10 0.09 0.08 0.07

Z

0.06

A K

0.05

I

0.04

T

0.03 0.02 0.01 0.00 0

200

400

600

800

1000

1200

1400

Mean density (kg/m³)

Figure 5: Dependence of normalized photoacoustic intensity at 50 Hz on mean density of substrates. For depth-profiling, the phase information of photoacoustic signal is also important. Differences between individual samples are more pronounced for the phase lag frequency dependence than in the case of intensity. As the reference carbon black sample used for the experiment has shown some irregularities particularly for lower modulation frequencies, it was not possible to © 2008 SWANSEA PRINTING TECHNOLOGY LTD

TAGA JOURNAL VOL.4

8


gain absolute values of the phase lag; therefore, only phase lag differences between each sample and carbon black reference were calculated (see Figure 6). However, ink penetration depth still can be roughly estimated. According to the measured photoacoustic phase profiles, the ink layer thickness of the studied samples is greater than several microns; penetrating into the substrate structure to depth on the order of tens of microns. From the relative differences of phase values for individual samples and the overall character of phase curves further conclusions can be drawn. Generally, the phase lag of the photoacoustic signal is minimal for strongly absorbing samples with high thermal diffusivity and increases with decreasing values of sample optical absorption coefficient and thermal diffusivity. Comparing the phase lag of photoacoustic signals originating at the same depth beneath the sample surface, the weaker photoacoustic signal, the greater phase lag. The phase lag also increases with increasing depth into the sample, as the photoacoustic signal from deeper layers is delayed due to thermal transport to the sample surface. Assuming thermal properties in the range typical for respective kinds of materials, at the highest modulation frequency used, the penetration depth of photoacoustic signal is approximately 10 microns and at the lowest it is one order higher. This means that at high frequencies, the photoacoustic signal is influenced only by the uppermost layer of the samples, while at low frequencies the maximum penetration depth also plays a role. For samples studied, the phase lag increased in the higher modulation frequency region in following order: A, K, Z, I and T (Figure 6); toward the lower frequencies, the phase curves get near each other except for the T sample, whose phase lag remained apparently greater than for other samples. In the studied range of modulation frequencies it is also possible to distinguish uncoated and coated papers based on the shape of curves.

Photoacoustic phase (sample − ref.) (degrees)

15

0 Z A

-15

K I T

-30

-45

-60 0

5

10

15

20

25

30

35

Square root of frequency (Hz)

Figure 6: Dependence of photoacoustic phase reduced by carbon black values on square root of frequency. Further experiments with the T sample have shown significant influence of compactness of its three layers on modulation frequency dependence of both intensity and phase of photoacoustic signal. In low modulation frequency region, the maximum difference measured for compact and split sample was 0.042 for normalized photoacoustic intensity and −20.8 degrees for photoacoustic phase. On the contrary, for high frequencies it was −0.027 for normalized photoacoustic intensity and 10.3 degrees for photoacoustic phase. Thus, the generally known

TAGA JOURNAL VOL.4



9

ability of photoacoustic method for investigation of the subsurface structure was verified to be suitable for the type of samples studied. 3.2 Destructive method 3.2.1 Cross section analysis and Optical Microscopy (OM) The cross section analysis was used as a destructive method. The samples were cut by microtome, observed by optical microscopy and captured by camera in order to measure the real depth of penetrated ink. All the analyzed samples caused various difficulties in cross section analysis (as the magenta dye ink is water soluble, the microscopy evaluation of the prints should be done as quickly as possible after the sample preparation). The K sample (carton board) was found to be the most problematic. Due to its thickness neither microtome cross-section nor optical microscopy was done. Table 4: Results of vertical distribution of magenta ink Sample

Sample thickness (µm)

Ink thickness (µm)

Z

170

20–34

K

430

—

I

110

39–65

T

185

66–79

Table 4 shows that the results obtained by the microtome cross section measurements are in correlation with all presented non-destructive measurements: slit-scanning densitometry, photoacoustics and CLSM (Table 3). Results of the cross section analysis and CSLM are in very good agreement particularly for Z and T sample. In case of the I sample, a possible reason of deviation between ink thickness values determined by these methods is inhomogeneity common for uncoated papers, influencing both the ink thickness in X-Y plane and uneven deformation of sample thickness during preparation for cross section analysis.

Figure 7: The microtome cross section analysis of Z sample with added micrometer measuring tool. Figure 7 presents the measurement of the ink distribution in the z-direction on the microtome cross-section. The thickness of the penetrated ink is measured by inserting the micrometer measuring tool after the OM image has been taken. 3.3 Comparison The results obtained by the microtome cross-sections present the real insight into the vertical ink distribution in z-direction of the print as well as into the surface thickness of the ink layer.


TAGA JOURNAL VOL.4

10


Those facts make this destructive method a reliable control measurement for all the applied nondestructive methods. 3.3.1 Cross section analysis vs. Slit-Scanning Densitometry The real thickness of the Ink-Jet ink printed on the Z paper is in the range of 20–34 micrometers. This result corresponds to that obtained by slit-scanning densitometry showing the highest optical density for the Z sample. This was explained by the remaining of Ink-Jet ink in a compact layer on the paper surface. The Z paper is the high quality photo paper. Its coating layer is thin and has a very good absorption characteristics. The correlation is also obtained for other measured samples: The real ink layer thickness in the z-direction of the I sample is in the range of 39–65 micrometers and of the T sample it is 66–79 micrometers. The results obtained by slit-scanning densitometry show that the T sample has the lowest optical density and that the value for the I sample is in between the extremes (as well as the K sample for which the cross section was not possible). 3.3.2 Cross section analysis vs. Confocal Laser Scanning Microscopy (CLSM) The results obtained by CLSM are in the close correlation with those of the microtome crosssection analysis. The real thickness of the Ink-Jet ink layers are precisely in the same range with those measured by CLSM for the Z (20–34 vs. 23.53 micrometers) and the T (66–79 vs. 72.50 micrometers) samples. Both methods agree for the I sample in between of the Z and T extremes. 3.3.3 Photoacoustics Phase of the photoacoustic signal in higher modulation frequency region decreased in order A, K, Z, I and T (see Figure 6). As the photoacoustic phase depends both on optical and thermal properties and the optical density measured by conventional spectrophotometer (see Figure 4) decreased in the same sequence except for the Z sample, it reflects dissimilarity in Z sample structure in comparison with other samples. The Z sample is printed on photo paper, whose coating is expected to have special properties intended to ensure optimal distribution of the ink, so the ink penetration profile is probable to be specific. At the other end of the modulation frequency range, only the phase lag of T sample is apparently higher in comparison with remaining samples. This is in accordance with maximum penetration depths determined namely by CSLM (see Table 3). Unambiguous determination of the exact ink-penetration depth could not be so far addressed by photoacoustic measurements due to lack of parameters needed to be known for assessment of such a complex multilayered system. At this stage, by the other methods used, the penetration depths for Ink-Jet ink printed on respective substrates were determined, but neither destructive cross section analysis nor other non-destructive methods provided detailed depth-profiles of the samples. Satisfactory discrimination of the phase curves measured for the individual samples proves sufficient sensitivity of photoacoustics, to observe differences in both optical and thermal properties of subsurface structure. In further research, correlation with results of all discussed methods applied to extended and better defined sets of samples should enable the actual depth-profile of Ink-Jet prints to be established.

4

Conclusions

The relevance of information obtained by different non-destructive and destructive measurement techniques is achieved. Each of the studied methods, despite their limitations, gives an explanation of the ink – paper interaction as well as the impact of the substrate on the print quality. A detailed study of the obtained results shows a good correlation among the applied nondestructive methods: slit-scanning densitometry, photoacoustics and CLSM. Not only the correlation among the non-destructive methods themselves, but the correlation with the results TAGA JOURNAL VOL.4



11

of the destructive method is achieved. The cross section analysis by microtome and optical microscopy give the real insight into the ink distribution in the z-direction. Slit-scanning densitometry, photoacoustics and CLSM are introduced as the methods complementary to each other and relevant for future applications.

Acknowledgements This work, was partly supported by the Ministry of Education, Youth and Sports of the Czech Republic, project No. MSM 0021627501.

References Glittenberg D., Voigt A. and Donigian D., Novel pigment starch combination for the online and offline coating of high quality inkjet papers, Pap. Technol, Vol. 44, No. 7, pp.36-42, 2003 2 Muck T. and Hladnik A., Evaluation of radial and vertical distribution of ink jet inks in paper, Professional papermaking, Vol. 2, No. 2, pp.62-64, 66-68, 2004 3 Lyne M. B. and Aspier J. S., Paper for Ink-Jet printing, Tappi Journal, pp.106-110, 1985 4 Baudin G. and Rousett E., Effect of paper properties on print quality, Imaging Science and Technology, pp.120-124, 2001 5 Lozo B., Vyörykkä J., Vuirinen T. and Muck T., Nondestructive microscopic and spectroscopic methods for depth profiling of ink jet prints, J. imaging sci. technol., Vol. 50, No. 4, pp.333-340, 2006 6 Goel A., Tzanakakis E.S., Huang S., Ramaswamy S., Hu W-S., Choi D. and Ramarao B.V., Confocal laser scanning microscopy to visualize and characterize the structure of paper, AICHE Symposium Series, Vol. 96, No. 324, pp.75-79, 1999 7 Auran P.G. and Bjorkoy A., Measuring the pore volume distribution of papers by CLMS for printability, STFI, Proceedings, p.220, 1999 8 Hoang V., Huy H.L., Wei S. and Parker L.H., The interactions of ink-jet inks and uncoated papers, 55th Appita annual conference, pp.285-292, 2001 9 Kaplanova M. and Katuscakova G., Photoacustic Study of the Thermal Effusivity of Cellulose and Paper, Springer series in Optical Science Vol. 69., p.180, 1992 10 Kaplanova M. and Cerny J., Photoacoustic and Photocalorimetric Study of UV Curable Inks and Varnishes, Advances in Printing Science and Technology, p.107, 1994 11 Kaplanova M. and Cerny J., Photoacoustic Study of the Ink and Paper Interactions, Advances in Printing Science and Technology, p.301, 1997 12 Drzkova M., Otahalova L. and Kaplanova M., Photoacoustic Study of Printed Samples, The First European Conference on Color in Graphics, Imaging and Vision (CGIV), p.307, 2004

1


TAGA JOURNAL VOL.4

12

WATER-BORNE AND SOLVENT-BASED FLEXOGRAPHIC INKS – INFLUENCE ON …

Water-Borne and Solvent-Based Flexographic Inks – Influence on Uncovered Area and Ink Leveling on PE-coated Paperboard Behudin Mesic*, Gunnar Engström** and Magnus Lestelius** *

Affiliated at Samrt Print BrobyGrafiska Education 37 SE-686 80 Sunne Sweden [email protected]

** Affiliated at Karlstad University Dept. of Chemical Engineering Faculty of Technology and Science SE-65188 Karlstad Sweden [email protected]; [email protected] .

Abstract A smooth and a rough PE-extrusion-coated paperboard, with and without corona treatment, were printed flexographically using a solvent-based and a water-borne ink with the objective to examine the influence of surface roughness and wetting of the paperboard on ink transfer, uncovered area (UCA), ink-film thickness distribution, ink leveling, and print unevenness. The results showed that ink transfer was significantly lower for the solvent-based ink, but because this ink has a stronger pigment, less ink was needed for a given ink density. Both inks were equally prone to yield UCAs, and both inks also formed dry ink layers with the same thickness distribution in the 2–4 mm range. Both inks also leveled out to same extent. The leveling was controlled by the corona treatment and reduced the UCAs. The smooth substrate exhibited the least UCAs. On a submillimeter level the print unevenness for the water-borne ink exhibited patterns that resemble crawling worms or curled thread ends aligned in the direction of the printing. The pattern for the solvent-based ink was grainy and without orientation. These patterns are suggested to reflect the splitting pattern of the ink during the transfer to the paperboard. Keywords: printability, solvent-based ink, water-borne ink, surface roughness, PEextrusion-coated paperboard, ink transfer, uncovered area, ink leveling, print unevenness.

1

Introduction

Flexography is a progressive and widely applicable printing method for a wide range of substrates and finished products, such as foil, plastic film, corrugated board, paper, paperboard, or even fabric [1,2]. It is a printing method that has greatly expanded and is often preferred for package printing [3,4]. Flexography generally uses low-viscosity inks, either solvent-based or water-borne, which dry very quickly between the printing units in the printing press [5]. Solvent-based inks are formulated to contain a blend of alcohol that dissolves the resin and produces a liquid ink, while water-borne inks are generally dispersions, using water and amines for the liquefaction [6]. Both solvent-based and water-borne inks solidify by loss of solvent/water (drainage and evaporation). Solvent-based inks give good adhesion, gloss, and flexibility on polymer-coated substrates and aluminum foil, and they provide good resistance to water, acid, and alkali [7].

TAGA JOURNAL VOL. 4


BEHUDIN MESIC, GUNNAR ENGSTRÖM AND MAGNUS LESTELIUS

13

Some disadvantages with using solvent-based inks are that they require venting, because odor is given off, and that environmental hazards increase, because the inks contain highly inflammable solvents. Water-borne inks share many common features with solvent-based inks and are said to be cost effective, when material and compliance costs are factored in [7]. Water-borne inks have advantages compared to solvent-based inks: they wash up with water, they do not produce vapors hazardous to health, they reduce fire hazards and insurances costs, and their cost is less affected by fluctuations in the price of crude oil. However, water-borne inks also have disadvantages: they require more energy for drying, and they do not properly wet printing substrates with low surface energy, such as polymeric films (for example, polyethylene (PE) film and PE-coated board). Another disadvantage with water-borne inks is the handling of the waste and recycling of packaging material [7,8]. A common and serious problem in flexographic printing, as well as with solvent-based and water-borne inks, is uncovered area (UCAs). The UCAs consists of a number of small dots, typically 0.04-0.9 mm2 in size, that are supposed to be covered with ink but are not. When printing with water-borne inks, both local areas of low surface energy and depressions [2,9] in the surface of the printing substrate have been reported to be able to cause UCAs. Experience within the printing industry is that solvent-based inks cause less UCAs than water-borne inks. Because solvent-based inks have significantly lower surface tension than water-borne inks, this suggests that solvent-based inks wet the printing substrate better and also level out better. Therefore, solvent-based inks are believed to be forgiving for areas of low surface energy of the printing substrate. This paper is focused on UCAs and the issues of how much of the UCAs stem from local areas with low surface energy and how much stems from depressions in the surface of the printing substrate. To shed light on this issue, a printing trial was performed in which a smooth and a rough PE-extrusion-coated paperboard, with (44 mJ/m2) and without (40) the DE2000 equation shows a more coherent relation to the perceived differences then the DECMC-equation. Overall the DE2000 and the DECMC-equation show a very similar behavior with a slight advantage for the DECMC-equation, since the r2-values are in a more coherent band compared to the DE2000-equation. The DEab equation shows the largest spread of data point throughout the entire chroma range of the samples. With the increasing complexity of the colour difference equations the data points are in a more coherent band. 3.3 Performance of the four difference equations against Hue of the evaluated samples The third evaluation of the colour difference equations is against the hue angles of the colour samples. The evaluation against the hue angles of the tested colour samples is important to determine if which equation performs better at certain hue angles, especially between DE94 and DE2000, since an extra term was introduced to the equation to weight the colour differences according to the position of the tested colour in colour space. The graphs for this are shown in figures 8 and 9.

1 0.9

Correlation

0.8

)r(ab

0.7

)r(94

0.6

r(c mc)

0.5

)r(2000 )r(cmc

0.4 0.3

r(2000)

355.64

357.58

276.15

317.86

234.48

r(ab) 243.38

181.1

195.92

Hue Angle

173.29

130.5

149.43

98.5

105.29

77.99

31.41

71.72

r(94) 11.83

1.77

0.2 0.1 0

Figure 8: Threedimensional plot of the r2-values in regards to Hue Angle

Figure 9 contains the same data points as figure 8 but only the fourth order polynomial curve is shown.


TAGA JOURNAL VOL. 4

158

THE EVALUATION OF COLOUR DIFFERENCE EQUATIONS AND OPTIMIZATION….

1

0.9

DE ab DE94

0.8

DE2000 0.7

DE cmc

0.6

R

0.5

Y

G

B

0.4 0

20

40

60

80

100

120

140

160

180

200

220

240

260

280

300

320

340

360

Hue Angle

Figure 9: Performance of all four colour differencing equations in relation to the hue angle values of tested colour swatches

From Figure 8 and 9 it can be seen that DEab, DE94 and the DECMC-equation show a negative trend towards the blue-violet region of the hue angle chart. It needs to be noted that the performance of the DE2000 equation does not show a large variation in relation to the hue angle. It is clearly visible that the DE2000 curve (blue curve) shows a much better correlation between the rating done by the human observers and the numerical expression of colour differences using this equation. The improved performance of DE2000 vs. DE94 and DEab can be attributed to the correctional term that was introduced in the DE2000 equation to improve its performance in especially this region of the visible spectrum. 3.4 Overall performance indicators of the colour difference equations In addition to the figures 4 – 9, which show how each of the colour difference equations correlate to Lightness, Hue and Chroma of the evaluated samples, other means are necessary to evaluate which of the four equations correlate the most with the perceived visual difference of the 24 human observers used in this test. A measure for this is the average r2-value of the visual rating scores and the standard deviation these r2-values for each one of the colour differencing equations. The following table depicts this in more detail. Table 2: Average r2-value, standard deviation and coefficient of variation for all four colour difference equations Average r2-value Standard Deviation Coefficient of variation (Standard deviation/mean)

DEab 0.818 0.161 0.197

DE94 0.851 0.124 0.146

DECMC 0.877 0.108 0.123

DE2000 0.885 0.104 0.118

The table 2 shows that the DE2000 equation has the lowest coefficient of variation of all four colour difference equations and highest average r2-value. The coefficient of variation is defined as the division of standard deviation by the mean value. It allows for a more valid comparison between data sets with different mean values. The lower the number the lower is the dispersion

TAGA JOURNAL VOL. 4


M. HABEKOST, K. ROHLF

159

of the data. Since DE2000 has the lowest coefficient of variation it means that the numerical colour difference values correlate quite well with the observed colour differences. From table 2 is can be seen that the data for the DE2000 equation has the lowest dispersion. In figure 10 these coefficient of variation values are shown in a bar graph. 0.25

0.2

0.15

0.1

0.05

0 DEab

DE94

DEcmc

DE2000

Colour difference equation

Figure 10: Colour difference equations vs. coefficient of variation values 3.5 Optimization of the DE2000 equation The DE2000 equation contains three weighting factors kL for lightness, kC for chroma and kH for hue angles. These values are set by default to a value of 1. This means that there is no emphasis on the lightness (kL), chroma (kC) or hue (kH) of the measured colours. In the past attempts have been made to see if the DE2000 equation can be optimized [15], [16]. Gilbert et al [15] found that the DE2000 equation was only marginally better than the DECMC equation, while Melgosa et al. [16] introduced another weighting function. This weighting function improved the performance of the DE2000 equation, but the improvement was statistically not significant enough. A recent study by Johnson & Green [17] done in this area found that an adjustment of the parametric factors to the values of kL = 1.5, kC = 1 and kH = 0.5 in the DE2000 as well as the DE94 equation improves the performance of these equations versus the DEab equation. Luo et al [11] did a verification of DE2000 using industrial data and came to contradicting results in regards to the weighting factors kL, kC and kH. In one case a ratio of 2:1:1 got better results and in the other case a ratio of 1:1:1 corresponded better on how the human observers perceived the colour differences of the samples. It needs to be said, that for the 2:1:1 ratio untrained observers were used and in the case of the 1:1:1 ratio professional observers were used. These two studies were the starting point for the attempt to improve the r2-value between the ranking done by the human observers and the numerical difference obtained using the DE2000 equation with the default values for weighting factors. Based on the work done by Johnson & Green [17] and Luo [11] it seems that an emphasis on the kL factor seemed to improve the correlation between observed colour differences and numerically calculated differences. The optimization of the weighting factors for DE2000 was done with the help of MAPLE software. The maximum of the average r2-values was obtained using the Optimization Package


TAGA JOURNAL VOL. 4

160


within MAPLE. The package uses built-in NAG (Numerical Algorithms Group) C library routines that are well documented and widely used in the scientific community. In particular, the SQP (Sequential quadratic programming) method was invoked using NLPSolve, which is a function call within MAPLE to solve nonlinear programs. The numerical method, as well as the NAG library itself is well documented [20]. The DE2000 equation was entered together with the rating scores obtained from the 24 observers. The criteria for the optimized weighting factors were to get an improved average r2value and a lower standard deviation. As a result the following weighting factors were found: kL = 0.98 kC = 1.05 kH = 0.97 The average r2-value improved slightly from 0.8845 to 0.8861 and the standard deviation improved slightly from 0.1045 to 0.1038. Therefore no real improvement was achieved. The weighting factors are also quite similar to each other with no real emphasis on one or the other. The MAPLE worksheet is listed in the appendix. This attempt to improve the performance of DE200 equation resulted only in a very minor improvement in regards to a reduced standard deviation of the r2-values. The average of the r2values could not be improved. Melgosa et al [16] introduced a chroma-tolerance weighting function, which they have borrowed from the LCD equation, to improve the performance of the DE2000 function [18]. Although performance was improved it was statistically not significant. Gilbert et al [15] did an evaluation of colour difference formulae and found the correlation efficients for DE2000 were only marginally higher than that for DECMC. Based on the results and results found in literature only one instant gives concrete new weighting factors for DE2000. Other results show only minor improvements that indicate the use of the DE2000 equation with the default weighting factors of 1. Therefore the DE2000 formula in its current form is the best colour difference equation that is currently available.

4

Conclusions

As this study shows, there are vast differences in regards to the correlation between visually observed colour differences and numerically calculated difference, depending on which colour difference equation is used. Consequently it is important to name the equation used to determine colour difference. The evolution of colour difference equations demonstrates that with increasing equation complexity the correlation between numerical and observed difference improves. The DE2000 equation correlates slightly better with the observed colour differences than the DECMC equation. The r2value for the correlation between numerical and observed colour difference is 0.877 for the DECMC equation and 0.883 for the DE2000 equation. Both previously named equations correlate numerically better with the perceived colour differences than the DEab-equation, which has an average r2-value of 0.818. The attempt to improve the performance of the DE2000 equation by modifying the weighting factors kL, kC and kH resulted in similar new weighting factors that gave only in a minor improvement in regards to the average r2-value. This agrees with the majority of the work done by other researchers in this area. Unless an internationally agreed set of new weighting factors for DE2000 is available one should continue to use the default weighting factors of 1.

TAGA JOURNAL VOL. 4



161

Acknowledgements: This project was made possible through the generous support of the Faculty of Communication and Design through a Project Grant. We would like to acknowledge the support and help of Peter Roehrig, Print Technician at the School of Graphic Communications Management, in preparing the printed colour samples. We would also like to thank Assistant Professor Chris Kular for the Ishihara Colour blindness test plates. We would like to thank Dr. Abhay Sharma, Chair of the School of Graphic Communications Management at Ryerson University for his guidance and support. We would also like to thank all the volunteers who helped us with this project.

References: 1 2 3 4 5

6 7 8 9 10 11

12 13 14

15

16

Luo R., 75th CIE Jubilee conference presentation, Ottawa, May 2006 CIE., Colourimetry, 2nd ed., CIE Publications No 15.2. Vienna: Central Bureau of the CIE, 1986 Heidelberg Publication, Heidelberger Druckmaschinen AG, Colour and Quality, p.81, 1995 Clarke F.J.J., McDonald R. and Rigg B., Modification to the JPC79 colour differencing formula, J. Soc. Dyers and Col., 100: 128 – 132, 1984 Berns R.S., Altman D.H., Reniff L., Snyder G.D. and Balonen-Rosen M.R., Visual determination of suprathreshold colour-differences using probit analysis, Col. Res. App., Vol .21, 459 – 472, 1991 CIE., Technical report: Industrial colour-difference evaluation. CIE Publication 116, Vienna: CIE Central Bureau, 1995 Billmeyer Saltzman., Principles of Colour Technology, Berns, R.(Ed.), 3rd edition, Wiley & Sons, New York, p. 114, 2000 CIE, Technical Report: Improvement to industrial colour-difference evaluation, CIE Publication 142, Vienna: CIE Central Bureau, 2001 Luo M.R., Cui G. and Rigg B., The development of the CIE 2000 colour difference formula, Col. Res. Appl. 26: 340 – 350, 2001 Sharma G., Wu W. and Dalal E.N., Col. Res. Appl., 2005, 30, 2005 Luo M.R, Minchew C., Kenyon P. and Cui G., Verification of CIEDE2000 using industrial data, AIC Colour and Paints, Interim Meeting of the International Colour Association, Proceedings, 97 – 102, 2004 Johnson T., Green P., Issues of Colour Measurement and Assessment, TAGA Proceedings, 518 – 534, 1991 Buering., Deutsche farbwissenschaftliche Gesellschaft e.V. , Presentation 2001, http://www.dfwg.de/doc/dfwg-homepage-419.htm , retrieved Feb 4, 2006 Basimir R., Costello G., Dibernardo A., DiPiazza J., Kuna D., Paulius K., Rybny C. and Zawackni W., A Comparison of Visual and Spectrophotometric Evaluations of Paired Colour Prints, NPIRI Task Force on Colour Measurement, Rochester, NY, TAGA Proceedings, pp. 558-578, 1995 Gilbert J.M., Daga J.M., Gilabert E.J., Valldeperas J. and the Colourimetry Group of the Spanish Fastness, Evaluation of colour difference formulae, Colour. Technol. 121, 147 – 152, 2005 Melgosa M., Huertas R., Yebra A. and Perez M.M., Are Chroma Tolerance dependent on Hue-angle?, Col. Res. Appl., 29: 420 – 426, 2004


TAGA JOURNAL VOL. 4

162


17 Johnson A. and Green P., The Colour Difference Formula CIEDE2000 and its Performance with a Graphic Arts Data Set, TAGA Journal, vol. 2, pp. 59 – 71, 2006 18 Kim D.H., The influence of parametric effects on the appearance of small colour differences, PhD thesis, University of Leeds, UK, 1997, Chapter 6 19 Kim D.H., Cho E.K. and Kim J.P., Evaluation of CIELAB-based colour difference formualae using a new data set, Col. Res. Appl. 26, 369 – 375 2001 20 Numerical Algorithm Group, http://www.nag.com/numeric/cl/cldocumentation.asp retrieved January 28, 2008

Appendix The following table list the L*a*b*-values of the colour tested. Table 4 L*a*b*-values of the tested samples Sample # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

TAGA JOURNAL VOL. 4

L* 49.66 75.94 50.43 61.68 50.29 40.27 76.81 39.65 67.66 50.03 68.99 81.85 79.74 79.93 80.61 71.12 89.37 68.28 79.79 80.27 79.08 76.83 57.54 57.58 60.03 48.50 49.65 49.34 48.98 41.47 33.10 32.38 31.65

a* 36.17 -4.40 20.20 -9.85 -17.20 19.31 8.24 6.51 -21.34 12.36 -5.50 7.62 2.58 0.96 -7.32 9.08 -3.69 -15.29 -3.98 5.96 -14.09 13.42 -23.03 -18.11 46.99 -19.95 38.68 -27.65 26.01 -19.86 -2.60 14.51 -10.32

b* -1.53 29.44 23.56 20.37 -0.33 -17.47 28.76 -30.60 17.77 2.59 28.73 -7.99 13.19 -8.91 8.57 0.28 13.46 -4.36 -11.17 1.97 2.64 40.63 -5.99 2.13 -8.19 -39.01 23.62 16.33 -1.53 0.46 -1.41 2.43 -19.92



34 35

56.99 61.57

56.15 -25.67

163

-4.28 -35.96

Worksheet used in MAPLE software for the optimization of the weighting factors for DE2000 > restart > # INPUT TEST COLOR DATA AND COMPUTE DE2000 value > # M[i,j] will be an array where i is the number of data pts for a given test color (1 to 4), and j is the number of the test colors (1 to 35) > # DeltaLprime35 etc. are data files with the 35 values for the corresponding terms in the DE2000 formula > > for j from 1 by 1 to 35 do for i from 1 by 1 to 4 do L[i, j] := op(1, fscanf("DeltaLprime35", "%f")); C[i, j] := op(1, fscanf("DeltaCprime35", "%f")); H[i, j] := op(1, fscanf("DeltaHprime35", "%f")); RT[i, j] := op(1, fscanf("RT35", "%f")); SL[i, j] := op(1, fscanf("SL35", "%f")); SC[i, j] := op(1, fscanf("SC35", "%f")); SH[i, j] := op(1, fscanf("SH35", "%f")) end do end do > > for j from 1 by 1 to 35 do for i from 1 by 1 to 4 do DE2000[i, j] := sqrt(L[i, j]^2/(kL^2*SL[i, j]^2)+C[i, j]^2/(kC^2*SC[i, j]^2)+H[i, j]^2/(kH^2*SH[i, j]^2)+RT[i, j]*C[i, j]*H[i, j]/(kC*SC[i, j]*kH*SH[i, j])) end do end do > > # INPUT RATER INTEREST SCORES FOR THE COLOR SAMPLES > for j from 1 by 1 to 35 do for i from 1 by 1 to 4 do RI[i, j] := op(1, fscanf("RaterI35", "%f")) end do end do > > # CALCULATE CORRELATION COEFFICIENT FOR EACH TEST COLOR RSQ[j], j=1..35 > # n=number of samples (n= 4) > n := 4 > > for j from 1 by 1 to 35 do sdri[j] := (n*(sum(DE2000[k, j]*RI[k, j], k = 1 .. n))-(sum(DE2000[k, j], k = 1 .. n))*(sum(RI[k, j], k = 1 .. n)))/(n*(n-1)); sdsquared[j] := (n*(sum(DE2000[k, j]^2, k = 1 .. n))-(sum(DE2000[k, j], k = 1 .. n))^2)/(n*(n-1)); sRIsquared[j] := (n*(sum(RI[k, j]^2, k = 1 .. n))-(sum(RI[k, j], k = 1 .. n))^2)/(n*(n-1)); RSQ[j] := sdri[j]*sdri[j]/(sdsquared[j]*sRIsquared[j]) end do > > # Compute average R^(2 )value, and variance © 2008 SWANSEA PRINTING TECHNOLOGY LTD

TAGA JOURNAL VOL. 4

164


> RSQAve := 1/35*(sum(RSQ[k], k = 1 .. 35)) > varRSQ := 1/34*(sum(RSQ[k]^2, k = 1 .. 35))-1/1190*(sum(RSQ[k], k = 1 .. 35))^2 > > with(Optimization) > soln := NLPSolve(RSQAve, {.1 kHmax := op(2, op(3, op(2, soln))) kHmax := 0.970995457163454034 > RSQAvemax := op(1, soln) RSQAvemax := 0.886121262375657914 > stddevRSQmax := sqrt(subs(kC = kCmax, subs(kL = kLmax, subs(kH = kHmax, varRSQ)))) stddevRSQmax := 0.1037577101 > subs(kL = 1, subs(kH = 1, subs(kC = 1, RSQAve))); sqrt(subs(kL = 1, subs(kH = 1, subs(kC = 1, varRSQ)))); 0.8845555344 0.1044760282

TAGA JOURNAL VOL. 4


T. HARTUS

165

Effect of Toner Fixing Temperature on Print Properties in the Electrophotographic Process Timo Hartus Department of Forest Products Technology Paper and Printing Technology Helsinki University of Technology P.O. Box 6400 FI-02015 TKK Finland Email:[email protected]

Abstract Fixing by fusing is the final stage in the electrophotographic printing process. It is an important part of the printing process and greatly determines the final print quality. Melting begins at the surfaces of the toner particles, which at first start to cohere and then as melting progresses the filming or fusing toner starts to wet and adhere to the paper. The print evenness and gloss are seen to develop simultaneously with the adhesion phase as a function of spreading on the paper surface. The aim of the work is to define the effect of fusing temperature on the development of toner adhesion, print evenness and print gloss. To achieve this aim, a new test method has been developed for the estimation of toner interaction in electrophotographic applications. The results of this study show that the degree of toner adhesion greatly determines the filming properties and hence the final evenness and gloss level of a printed surface. Keywords: Adhesion, Electrophotography, Paper, Testing, Toner, Fusing

1

Introduction

The electrophotographic printing process generally includes six main steps, which are charging of the paper surface, exposing, developing, transferring toner to paper, fusing on the paper and cleaning of the photoreceptor [1]. Fusing is the final process in an electrophotographic printer or copier, and it determines print quality properties, such as fixing strength, glossiness and density of the image [1]. In the fusing step, toner is permanently fixed to paper by effectively melting the toner onto the paper surface. The toner particles become fixed on paper by heating them above their glass transition temperature. In addition to heating, pressure can be used to enhance the fixing result [2]. During heating, toner particles partially melt, predominantly on their surfaces, and thereby begin to bind to each other and onto the paper. Thermal changes in the toner during the fusing process can be divided into three stages [3]: 1. Warming (temperature increase of toner particles and paper) 2. Softening (melting of the toner begins from the particle surfaces and they start to cohere and adhere to each other) 3. Melting (partly melted toner adheres to the paper) Immediately after heating is removed, the melted toner starts to cool and the toner particles become bonded to the paper. If the toner has not fused sufficiently, the adhesion to paper is poor.


TAGA JOURNAL VOL. 4

166

EFFECT OF TONER FIXING TEMPERATURE ON PRINT PROPERTIES …

Energy consumption during the fusing process can be split into several parts: the first is consumed by the toner particles in initial surface melting leading to their cohering and adhering to each other. In this stage, the toner changes from a powder to a highly viscous particulate mass, reaching a maximum viscosity at a temperature ranging from 120 to 130 °C [3]. Subsequently, the remaining energy required is consumed to complete the melting, associated with a lowering of the viscosity in turn allowing the toner to penetrate controllably into the paper. The process, however, is subject to limitations. For example, a too high fusing temperature has a negative effect on paper and print quality, even if the toner adhesion may reach its maximum. The most significant negative effects are yellowing and curling of the paper [3]. Following the fusing stage, the print temperature should decrease sufficiently so that the toner hardens. If the fixing temperature is too high, however, the still warm and soft toner can cause detrimental smearing. Furthermore, too high a drying temperature may cause too low toner viscosity, leading to excessive spreading of toner in images, and an increased tendency for slump, or flow, into the paper. Clearly, if the temperature applied is too low, i.e. below the glass transition temperature of the toner, no fixing occurs [3]. Though the glass transition temperatures of toners are typically more or less the same in practice, quite significant differences in the melting temperature ranges of toners can be found [3]. Electrophotographic toners consist mainly of polymer resin and pigment [1]. In addition charge control and flow agents and waxes may be used. Resin types used in toners include styreneacrylates, polyesters and epoxides. Toner pigments themselves do not have significant affinity to a paper surface, thus the main factor determining the toner-paper interface is the toner polymer phase, and indirectly, depending on polymer state and pigment concentration, the viscosity of toner. The secondary factors are thermal conductivity, surface energy and toner pile height [4,5]. Toner rheological properties must be optimized to allow for the correct balance of coalescence, spreading and penetration into paper [1]. Toner thermal properties are a compromise between their being soft enough for the fuser, but not too soft in order to prevent the toner particles from being fused onto the photoreceptor at the development stage, or in the storage before use. In addition, the paper must be optimally designed to have the necessary thermal response characteristics to support the fusing and cooling cycle. Its structure must also be porous and permeable enough for the toner to be able to partly penetrate into the paper. Thick paper grades need a longer time to warm than thin papers with low surface porosity. This may cause incomplete fixing of toner in the printing and copying process. Both toners and papers combine to give a wide variety of melting rates according to the system used and their individual properties. Although a toner with a small particle size and a more spherical shape undoubtedly produces a higher quality print, it needs some additional heating energy compared to the conventional toner in order to produce the same fixing level and print gloss [6,7]. The real fusing temperature reached on the print can vary greatly in the printing process. A large range of paper grades are commonly accommodated in a single process under equal conditions. The printing quality, however, varies if the thermal properties of papers differ too widely from each other. The gloss of the printed image and the evenness of the gloss have a subjective effect on the print quality. Although the legibility of extremely glossy prints decreases, high gloss is often associated with high print quality [8]. The roughness of the paper greatly determines the roughness of the print [9]. The gloss of a print is strongly depending on the underlying paper roughness [8]. Printing normally increases the gloss level but it often increases the gloss variation [10]. Typically, the raising of fusing temperature in electrophotography increases the gloss level of the final print [11].

TAGA JOURNAL VOL. 4


T. HARTUS

167

The purpose of this study is to explain the effect of different paper grades on print quality and fixing degree in the electrophotographic printing process. To achieve this, the effect of fixing temperature on the quality properties of electrophotographic prints with various types of digital printing papers is studied, and the results used to clarify the possibility of controlling the electrophotographic print gloss by adjusting the fixing temperature. In this latter context, the development of print gloss and roughness as a function of fixing temperature is also studied.

2

Experiments

All the prints, thermal fusing processes and measurements were carried out under standard laboratory conditions (23 °C and 50 % RH). 2.1 Materials Studied papers were: 80 g/m2 uncoated copy paper (1), 130 g/m2 coated art paper (2) and 200 g/m2 glossy coated art paper for digital printing (3). Papers were cut to a size of 40 mm x 200 mm for the experiments. Two commercial black toners (A and B), which are commonly used in copy machines, were examined in this study. The toners differed in shape and particle size. Toner A was a conventional toner with the somewhat greater particle size than toner B, which in turn was described as a chemical toner with a more spherical shape. According to the thermogravimetric analyses, differential scanning calorimetric analyses and surface energy determination, however, the thermal properties and surface chemical properties of both toner grades were virtually identical, showing only some differences in the melting energy consumptions of the toners [3,12,7]. They were thus assumed to be constituted of the same material, differing only in particle size and shape, and hence having the same thermal and chemical properties. The impact of particle size and shape could therefore be studied. The formulation and physical properties of the studied toners are presented in Table 1.

Toner A Toner B

Table 1: Formulation and property facts of the toners Particulate Lubricational Glass Specific melt size wax transition energy / µm (on /% temperature / J/g average) / °C 8.5 5 65.6 14.0 (platy) 7.0 15 66.6 19.8 (spherical)

Surface energy (at 150 °C), mJ/m2 19.5 17.7

2.2 Printing Toners were spread on the papers by a hand coater using a bar applicator, the application procedure being the following: A stencil board, thickness 0.5 mm, was set on the paper to be used as substrate. Toner was spread over the paper by pulling a handcoater bar over the stencil board a number of times so that an even toner layer was created. The print areas accepted for the analysis were defined as having optically even full coated black surfaces. The target ink (toner) amount applied after metering was approximately 2 g/m2, on a fully coated paper surface. Each accepted black area was roughly 3 cm wide and at least 10 cm long. The even thickness of the “printed” toner layer was visually controlled before fixing toner on the substrate in the experiment. Only even, full colour, black area was accepted for testing. These chosen samples were then further evaluated by measuring the optical density with a Vipdens 2000 densitometer (Vipdens 2000, Viptronic GmbH) without a polarization filter. The


TAGA JOURNAL VOL. 4

168


density of the prints was taken as a control of the amount of toner on the paper. Print density was evaluated only on full printed areas. 2.3 Fixing The toner was thermally fixed using IR-heaters. Two prints, one of each toner, were heated together so that the different toners could be evaluated on the same paper substrate simultaneously. The sheets were placed on a plywood carriage for transporting under the heating unit. The heating unit consisted of two sections of IR heating elements (Elstein Hochtemperaturstrahler HTS/1 250). The distance between the thermal elements, mounted above the papers, and the print samples was 50 mm. Both the IR heating elements emitted 250 W each and generated a maximum surface temperature of 700 °C. The samples were kept for various controlled times under the IR-heaters until the chosen target temperatures were reached. Surface temperatures of prints were measured by a portable Dickson IR-temperature meter. The measuring time of the meter was less than one second. The temperatures reached at the end of the fixing stage were chosen to be 75, 100, 125, 150 and 200 °C, respectively. 2.4 Toner adhesion measurements The toner fixing and film forming properties was characterized by a specially designed adhesion measurement, using a print tack-meter, which is a special accessory mounted on a laboratory printing device [13]. The instrument is illustrated in Figure 1. The thermally fixed prints were taped on the printing sector. The measurement disc is a 20 mm wide rubber coated aluminium roller. In this study, however, the measurement disc was pre-inked with pick test oil (IGT, normal tack). The distribution of pick oil was made by a high speed inking device (speed 100 m/s, time 10 s). The disc was weighed before and after the tack oil distribution, so that the total amount of pick oil on the measurement disc was determined, and found to be about 8 mg, which corresponds to an amount of 2 g/m2 after application on the paper. The distance between the disc (the upper position, when the disc is totally released from the print) and print on the printing sector was constant; approximately 5 mm. The difference in the distance between the disc and the printing sector has naturally a great effect on measured releasing time, because the moving time of the measurement disc is included into the release time and that is why it has to be constant to get comparable results. The meter is commonly used to survey the change of print tack or setting and drying of offset ink as a function of time after printing, and it records the release time of a measurement disc as it is removed from a printed surface under constant force, the print substrate being taped onto the printing sector. The release time of the measuring disc is termed the printack-value. More commonly, tack measuring instruments record the release force needed to remove such a disc. Comparison between time and force may allow greater resolution between the samples with high evenness and high toner adhesion, or good surface strength, in general (the area of long release times in printack) in further studies using the tack approach. In general, several measurements are made from a print as a function of time to monitor drying of the print. The number of printack-measurements from every print varied between 10 and 20. In this study, the average printack-value of each measured test sheet was calculated from the multiple individually measured values.

TAGA JOURNAL VOL. 4


T. HARTUS

169

Figure 1: A Universal Testprinter laboratory printing device with Printack-meter. 2.5 Print gloss measurements Print gloss was measured by a gloss meter (Vipgloss-1, Viptronic GmbH) after the thermal fixing and cooling of the toner on paper or on microscopy glass slides. The device measures the gloss value according to the standard DIN 16537, in which a 45 ° angle of incidence is in use. Possibly this gloss measurement standard may have set samples some different relative order than the other commonly used gloss standards would have done. The measured gloss values are relative values between 0 (no gloss at all) and 10 (very high gloss). At first, the gloss meter was calibrated onto the low gloss and onto the high gloss standard surfaces, after this the glosses of the prints were measured. Five measurements were made from every sample, and from these values the average values were calculated. The measurement area diameter is 3 mm, which is the minor diameter of ellipsoidal illuminated area in the device. A relative small measurement area allowed the gloss measurements from the samples on microscopy glass slides also. 2.6 Roughness of the print Roughness (rms) values of toner films on microscopic glass slides were measured with an atomic force microscope (AFM) using a tapping mode [14]. The basic mode of an AFM records


TAGA JOURNAL VOL. 4

170


the forces between the sample surface and a fine tip mounted on a cantilever. The tapping mode atomic force microscope (TM-AFM) is a development of the contact mode AFM. The oscillation is driven by a constant driving force and the amplitude of its oscillation is monitored. The tip touches the surface at the bottom of each oscillation, and this reduces the oscillation amplitude of the cantilever. The feedback control loop of the system then maintains this new amplitude constant as the oscillating, or tapping, tip scans the surface. This is accomplished by the z component of the scanner, which changes the tip height to adjust exactly for the surface topographic variations as the tip scans the sample surface. Measured areas were necessarily fairly small, 5 μm x 5 μm. For the practical AFM measurements, the toners were spread on microscopic glass slides by a hand coater in a manner identical to that used for the paper sample preparation, and they were likewise heated using the IR-heater. The measurements gave information only about the part of roughness which is due to toner itself. The measurements were carried out in air atmosphere, and only on those samples which were fixed well enough to avoid dusting in the sensitive environment of the AFM (only samples, which were heated during fixing in the oven to over 100 °C were analyzed using this method). The sample surface tested had also to be fairly even with no excessive variations in the surface height.

3

Results and discussion

Printack-values measured from unfixed or partly fixed prints (low fusing temperatures) were analogous with the friction meter generated values found in a previous study [3]. In the latter approach an adhesive tape is applied to the fused image by controlled and repeated application of pressure. The tape is then pulled from the paper under controlled force conditions, at a given speed and angle according to ASTM D 1894 (Standard Test Method for Static and Kinetic Coefficients of Friction of Plastic Film and Sheeting). The printack-values decreased once the toner particles had started to adhere together. At the second step, the printack-values started to increase once toner particles had begun to stick also onto the paper surface. The adhesion to paper was seen to begin at a temperature range of 100 120 °C. This result correlates also with previous toner adhesion measurements [7]. Printackvalues increased until the surface has reached the final state of evenness (film roughness). The final evenness of a surface was seen to depend mainly on the original roughness of the paper. Once the fixing was completed, only minor changes occurred in the printack-value and these only if the evenness of the print increased. The development of printack-values as a function of fusing temperature of the prints is illustrated in Figure 2.

TAGA JOURNAL VOL. 4


T. HARTUS

171

Printack-value, ms

75 70

A1 A2

65

A3 B1

60

B2 B3

55 50 60

80

100

150

200

Fusing temperature, °C Figure 2: Printack value of prints using toners A (conventional toner) and B (chemical toner) as a function of temperature on papers (1) – 80 g/m2 uncoated, (2) – 130 g/m2 coated and (3) – 200g/m2 glossy coated. In the case of the non-fused toner samples (the samples which were heated below 60 °C); the pick oil penetrated into the toner layer. The splitting occurred in the dry toner layer. The penetration depth determined the splitting depth of the toner layer. A part of toner is removed with the oil onto the disc and a part remains on the paper. The measured Printack-value is determined mainly according the friction forces between the toner particles. At somewhat higher temperatures, i.e. those samples which were heated to above the toner glass transition temperature in the range from 80 to 100 °C, and hence some film forming had occurred, the toner has been partly melted on its surface and begun also to fix onto the paper surface. In this case, the pick oil can only partially penetrate into the toner layer. The splitting occurs in the middle of the unmelted toner layer which is between the formed surface film of the toner and the toner which is beginning to fix on the paper. This toner layer on the tack wheel was thinner than in the case of totally non-fused toner surface, and this, together with the smoother surface creating stronger oil interface forces due to the more even oil layer distribution, may account for the slightly lower observed adhesion forces. The samples which were heated over 100 °C were more completely melted, and a hard surface film is seen to have formed, such that the pick oil cannot penetrate into the toner layer. The splitting then occurs between oil and the toner surface. The toner film surface evenness determines the measured printack-value in this case. No failures were observed in the toner layer or paper surface itself. The more even surface means a closer, more uniform oil film contact and hence the greater Printack-value. Print gloss is, as already described, quite strongly dependent on the paper gloss [15]. The roughness of paper greatly determined the maximum gloss level which could be reached. The gloss of electrophotographic prints increased as the fixing temperature increased, especially in the cases of thick ink layers, depending on the melting degree of the toner. The greater the melting degree of toner, the higher the print gloss obtained. The print gloss was seen to develop similarly to the printack-values as a function of temperature (Figure 3). First, we see that the gloss-values are almost stable when the toner particles start to melt and adhere together. At the second step, the surface of the toner layer being largely melted, the gloss-values start to increase strongly and the print becomes more even, the toner particles also begin to stick onto the paper


TAGA JOURNAL VOL. 4

172


Gloss

surface. In this temperature range, approximately from 100 °C to 200 °C, the printack-values show quite a clear correlation with the measured gloss values. The gloss increases with higher fusing temperatures.

10 9 8 7 6 5 4 3 2 1 0

A1 A2 A3 B1 B2 B3

60

80

100

150

200

Fusing Temperature, °C Figure 3: Gloss (relative scale 0-10) of prints using toners. A (conventional toner) and B (chemical toner) as a function of temperature on papers (1) – 80 g/m2 uncoated, ( 2) – 130 g/m2 coated and (3) – 200g/m2 glossy coated. The difference in the gloss values between toner A and B is significant, as is their mechanism of development. Toner B has greater roughness (Fig. 4), lower final gloss potential (Fig. 5), though the gloss of toner B does rise earlier in the heating series, and lower printack-value (Fig. 2) at every studied fusing temperature. Toner B, therefore, though having reduced gloss potential compared to Toner A, does have the ability to generate equal gloss to Toner A over a range of greater microroughness levels. These data support the hypothesis that Toner A, being of larger particle size and a more flat shape, may be able to mask some of the underlying larger scale paper roughness and so give a coverage effect. The limit in gloss, however, is apparently determined by the fine-scale fusing structure and the degree of filming, which eventually is inversely dependent on toner particle size. The connection between print roughness and print gloss for the toner samples on the microscope glass slides heated to the different fusing temperatures is seen in Figure 5. It can be clearly observed that the toner shape and particle size have a significant effect on the development of print roughness (film forming) and gloss under the various fusing conditions. When toner B is used, higher fusing temperatures are needed to reach the same print evenness and gloss than when using toner A, though, as observed in Figure 4, the gloss-roughness relationship continually favours toner B.

TAGA JOURNAL VOL. 4


T. HARTUS

173

45 40

Roughness rms, µm

35 30 25 20 15 10 5 0 100

150

200

Temperature, °C Figure 4: Connection between roughness (filming) and fusing temperature of toner (◦ toner A and ■ toner B) measured on microscopy glass slides. Combining Figure 4 and Figure 5 we can see that print smoothness and print gloss develop with increasing temperature for both the toners. Interestingly, toner A shows an initial range of gloss following a trend with rms smoothness, however, the highest gloss is reached by an apparent secondary effect, i.e. increasing gloss at equal roughness. It is once again suspected that the gloss development for toner A is strongly related to the distribution of particles and polymer on the paper surface, and that initial gloss may be related to physical coverage of the paper derived from the larger toner particle size. Only if it is heated further, the gloss becomes secondarily developed at the higher temperature rather than only being a function of rms roughness. On the other hand, toner B shows a continuous relationship between smoothness and gloss and we can interpret this function of heating as a direct melting, fusing and flowing of the toner.


TAGA JOURNAL VOL. 4

174


10 9

Gloss value

8 7 6 5 4 3 2 1 0 0

10

20

30

40

Roughness rms, μm Figure 5: Connection between roughness (filming) and gloss (◦ toner A, - conventional toner and ■ toner B – chemical toner) of toner fused at different temperatures on microscopy glass slides. There may be further likely reasons for the reduced final gloss potential of Toner B, for example it may be related to the greater amount of lubrication wax in the formulation in order to produce good quality with fast performance in black and white printing, e.g. FAX application of multifunction machines, and purposefully to remove the gloss from text to aid readability. Toner B having smaller particle size and a more spherical shape and also some greater specific melting energy consumption may have different thermal conductivity properties than Toner A. The possible difference in the thermal conductivity of the unmelted toner powders may also be a reason why toners A and B reached different gloss levels at the defined applied fusing temperature. It is possible that Toner B did not reach as uniform a temperature distribution as toner A during the limited heating time, and this could manifest itself as lower gloss level. The connection between print gloss and printack-value is indicated in Figures 6 and 7 for the different paper surfaces using each toner, respectively. The uncoated paper (1) resulted in the lowest printack-value (the lowest adhesion) at given gloss for both the toners, reflecting the impact of larger scale roughness on the pick oil layer distribution. The coated papers (2) and (3) did not display any differences in the printack-value or print gloss.

TAGA JOURNAL VOL. 4


T. HARTUS

175

10 9 8

Gloss

7 6

A1

5

A2

4

A3

3 2 1 0 50

60

70

80

Printack-value Figure 6: Connection between print gloss and toner fixing degree at different fusing temperatures with toner A ( conventional toner) on paper (1 – 80 g/m2 uncoated, 2 – 130 g/m2 coated and 3 – 200g/m2 glossy coated.

10 9 8

Gloss

7 6

B1 B2 B3

5 4 3 2 1 0 50

60

70

80

Printack-value Figure 7: Connection between print gloss and toner fixing degree over a range of fusing temperatures with toner B (chemical toner) on paper (1 – 80 g/m2 uncoated, 2 – 130 g/m2 coated and 3 – 200g/m2 glossy coated. .


TAGA JOURNAL VOL. 4

176

4


Conclusions

Adhesion measurement employing pick oil using a tack testing device provides a new way to produce fairly easily and reproducibly a useful estimation of toner fixing degree and print surface evenness. The Printack-device used in this work measures the release time of a disc from the printed surface, instead of the release force needed to remove the disc as is measured in most other such devices. Further study would be needed to determine if there are any relative merits between the methods. Incomplete fixing of toner on paper is generally caused by a too low fusing temperature, provided the conditions for surface energy and adhesion are met. The final print evenness and the print gloss clearly depend on the melting degree of toner at the fixing stage, in addition to the original paper roughness. The gloss and evenness of the print are increased with increased fusing temperature. High gloss of the print is obtained at high temperature after complete melting of toner. Though the glass transition softening temperatures of the studied toners were approximately equal, there was quite a significant difference in the effective melting temperature ranges of toners. It is obvious that the difference in the toner melting energy consumption determines partially the final gloss and evenness of printed surface, and although the fusing temperatures as measured were the same, the thermal energy distribution in the toners was likely to be different. The differences in toner particle size and shape reflected these effective thermal fusing properties, in that the coarser toner particle size acted to reduce print gloss though it achieved greater smoothness by potentially improved paper coverage and hence greater gloss at lower initial temperature. Higher temperatures were needed to bring the transition at constant smoothness to the then higher gloss potential. Higher temperatures were needed for the finer toner to achieve fixing and to manifest its glossing potential. Both the toner and the paper properties, therefore, have an important effect on the apparent fixing temperature of the toner on paper. In an electrophotographic printer there is usually no possibility to adjust fusing temperature or time in the printer fusing unit, and so the right paper grades for predefined toner types have to be selected to obtain optimized printing quality.

Acknowledgements The author would like to thank Mr. Mikko Jokinen for AFM analyses of prints. The author is also grateful to Professor P.A.C. Gane for his valuable comments during writing this article.

References 1 2 3 4 5

6

Schein L.B., Electrophotography and Development Physics, 2nd Edition, SpringerVerlag. New York, 1992 Oittinen P. and Saarelma H., Printing, Papermaking Science and Technology, book 13, Fapet Oy, Helsinki, 1998 Hartus T., Adhesion of electrophotographic toner on paper, Graphic Art in Finland 31(1), 14-19, 2002 Britto I.L., An Evaluation of Factors that Control the Fixing of Toner to Paper in Laserprinters, NIP 7 Vol 1. pp. 386-400, 2001 Sipi K., Formation of Quality and the Effect of Toner in Combined IR-fusing and Hot Roller Fusing, M.Sc. thesis. Helsinki University of Technology, Espoo Finland 121 p, 1998 Heilmann J., Heikkilä I., Majava M. and Oittinen P., Print Quality in Hot Air Fusing of Toners, IT&T´s Eleventh International Congress on Advances in NonImpact Printing Technologies. October 29 – November 3, Hilton Head, South Carolina. 214-218, 1995

TAGA JOURNAL VOL. 4


T. HARTUS

7 8 9

10 11

12

13 14

15

177

AL-Rubaiey H., Hartus T. and Oittinen P., The influence of Flash Fusing Variables on Image Fixing Quality, Graphic Art in Finland 312, 7-10, 2002 Handbook of Print Media: Technologies and production methods Kipphan, H. (ed.) Springer Berlin 2001 pp. 223-226, 2001 MacGregor M.A., Johansson P.-Å. and Béland M.-C., Measurement of smallscale gloss variation in printed paper, Topology explains much of the variation for one paper. Proceedings of the 1994 International Printing and Graphic Arts Conference, 33-43, 1994 MacGregor M.A. and Johansson P.-Å., Gloss uniformity in coated paper; Measurements of commercial papers, TAPPI Coating Conference Proceedings, 1991 Briggs J.C., Tse M.-K., Cavanaugh J. and Telep D.A., The Effect of Fusing on Gloss in Electrophotography, IS&T NIP14 International Conference on Digital Printing Technologies. October 18-23, Toronto, Canada, 1998 Hartus T. and AL-Rubaiey H., The Influence of Flash Fusing Variables on Image Fixing Quality. Experimental Report, Helsinki University of Technology, Laboratory of Media Technology. Otaniemi Finland 58 p, 2000 Testprint, Inc. Product Information, from http://www.testprint.com/ Retrieved 19.11.2005 Helsinki University of Technology, Laboratory of Paper and Printing Technology, Research Equipments and services, from http://www.tkk.fi/Yksikot/Paperi/laitteet.html/ Retrieved 15.1. 2006 Oittinen P., The limits of gloss in prints, Paperi ja Puu 65(11), 718–724, 1983


TAGA JOURNAL VOL. 4

TAGA JOURNAL VOL. 4


A. GIDLUND, T. MEJTOFT, S. DEMNERT.

179

Significance of Print Quality in Variable Data Printing Åsa Gidlund*, Thomas Mejtoft** & Sofia Demnert*** *STFI-Packforsk Linköping University ***

**

STFI-Packforsk Royal Institute of Technology (KTH) [email protected]

Royal Institute of Technology (KTH)

Abstract Print quality and variable data printing are two important topics when discussing digital printing and its ability to be a commercially successful printing technology. The aim of this paper is to determine the significance of print quality when using variable data printing to customize printed matters. In this study three levels of colour print quality and three levels of personalization; entirely static, personalized with name, and personalized with name, text and image, were established. Sample material was obtained by creating fictitious advertisements in the personalization levels, which were then printed at the print quality levels. A panel of respondents was asked to arrange the fictitious advertisements in order of preference. At all print quality levels, personalizing with only name led to a rather small increase in respondent preference compared to the static advertisements. However, personalizing with name, text and image led to a significant increase in preference compared to a static or name personalized advertisements. Hence, to attract the respondents a higher level of personalization should be used. The respondents were more forgiving towards low print quality when the advertising information was personalized. Furthermore, personalization contributes more at lower print quality levels, indicating that print quality is more significant in static printed matters or printed matters with a low level of personalization. Keywords: Digital printing, Print quality, Variable data printing

1

Introduction

Before the 20th century, printing was the only way of spreading marketing messages and information to customers when referring to non-personal communication channels [1]. During the later half of the 20th century many new, non-printed, channels for reaching a mass audience have been introduced. Communication media like radio, television and the Internet have changed the way that information and marketing messages are pushed and pulled to and from customers. The new channels, with the global spreading of the Internet leading the way, have however made print just an output channel among others. Until professional digital printing was introduced in the beginning of the 1990’s, printing was only possible to use when trying to reach a mass audience with static messages. Digital printing, in contrast to conventional printing technologies does not require a static printing plate (e.g. [2,3]), and is therefore suitable for producing mass customized [1] printed advertising material in e.g. one-to-one marketing [4]. Customized printing is based on the “combination of variable information with output devices that do not require intermediate films or plates” [3] Using digital printing for customization to make documents more effective as a business tool is a major economic driving force identified by the non-profit digital printing industry consortium PODi (e.g. [5,6]). Variable data printing can be used to gain business success since customized advertising material can make a significant increase in response rate compared to static advertisements [7]. However, earlier studies indicate that the use of variable data printing is in general low among


TAGA JOURNAL VOL. 4

180

SIGNIFICANCE OF PRINT QUALITY IN VARIABLE DATA PRINTING

printing houses [8] even though the technology has been available for over ten years. During this period the print quality of digitally printed matters has steadily improved and is today classified as good enough for many applications [2]. Print quality is a general measure of the success of a colour printing system and is an important customer requirement, along with other requirements such as cost, productivity, connectivity and reliability [9]. Furthermore, print quality is one of customer’s top considerations when purchasing printed matters [10]. The digital printing equipment available today with the capacity to produce variable data prints range from inexpensive office copy-printers to high-speed production equipment, which also means large variations in print quality. Furthermore, the importance of colour has been stressed in earlier research [7]. The response rate from a mailing campaign increased by 46% when colour was added to a static advertisement. For a customized advertisement the increase for colour was 167%. Print quality and variable data printing are two important topics when discussing digital printing and its ability to be a commercially successful printing technology. The aim of this paper is to determine the significance of print quality when using variable data printing to customize printed matters. The following research questions have been formulated to investigate the problem: • •

2

What effect do different types of personalization have on the respondents’ preference at different print quality levels? Is colour of any significance for the experience of personalized advertisements?

Material and Method

2.1 Research Approach In spring 2006, 28 respondents participated in a significance evaluation of print quality and variable data printing. The group of respondents was selected to be equally distributed over age and gender. In this study, three levels of colour print quality and three levels of personalization were defined (Figure 1). The print quality levels were established by technical measurements and visual assessments. Fictitious advertisements were created at the three personalization levels. To adapt the advertisements to the panel of respondents, each respondent completed a questionnaire with information forming the basis for the personalization. The advertisements were then printed in the three colour print quality levels. In all, nine colour samples were obtained. In addition, three black and white samples were printed in the medium print quality level, one for each personalization level. An evaluation was conducted to determine the significance of print quality in personalized printed matter. The respondents were asked to arrange the fictitious advertisements with different levels of print quality and personalization levels in order of preference.

TAGA JOURNAL VOL. 4



181

Quality level verification Technical measurements Visual assessment

Quality levels

Sample production Printer 1

QBW QL QM QH

Printer 2 Printer 3

Personalization levels

Panel of respondents

P1 P2 P3

Twelve samples for each respondent

Significance evaluation

Order of preference 1.

Questionnaire Visual assessment

Analysis

Figure 1: Schematic overview of the research process. 2.2 Personalization and Evaluation Material Design The questionnaire, which formed the basis of the personalization of the test material, concerned travel and specifically the respondents’ opinion of travel. The respondents were asked whether he/she preferred to travel in Sweden or abroad, if he/she preferred metropolises, seaside resorts or skiing resorts, with whom the respondent preferred to travel and how important high hotel standard, reasonable prices and the destination itself were when choosing destination. The respondents were also asked to grade how much they enjoyed different activities (e.g. golf, art, tourist attractions etc.) using a five-point Likert Scale, spanning from very uninterested to very interested [11]. The questionnaire also contained a section with background data e.g. age and gender.


TAGA JOURNAL VOL. 4

182


DREAMTRAVEL

08-652 52 00

Bästa Åsa Gidlund!

Bästa Åsa Nordin!

DREAMTRAVEL Vi vet att människors drömmar ser olika ut. Därför har vi på DreamTravel 08-652 52Sverige 00 och till andra spännande länderDREAMTRAVEL arrangerat resor, både inom över

hela världen, i snart 50 år. I vårt program finns många intressanta resor, 08-652 52 00 både för Dig om en Thomas semester Mejtoft! utomlands och för Dig som Bästa Vi vet att människors drömmar ser olika ut. Därför harsom vi pådrömmer DreamTravel hellre upptäcker vad Sverige har att erbjuda. arrangerat resor, både inom Sverige och till andra spännande länder över Vi vet att människors drömmar ser olika ut. Därför har vi på DreamTravel hela världen, i snart 50 år. I vårt program finns många intressanta resor, arrangerat resor, både inom Sverige och till andra spännande länder över både för Dig som drömmer om en semester utomlands för Dig som Just nu har vi ettoch oemotståndligt till Dig! helaerbjudande världen, i snart 50 år. I vårt program finns många intressanta resor, hellre upptäcker vad Sverige har att erbjuda. både för Dig som drömmer om en semester utomlands och för Dig som I Nice kombinerar Du lata sol- och baddagar medvad storstadssemesterns shopping hellre upptäcker Sverige har att erbjuda. och aktiviteter. Här i staden, som är franska rivierans största turistort, är man Just nu har vi ett oemotståndligt erbjudande till Dig! mycket noga med att bevara sitt historiska arv. Gamla byggnader hålls efter, de ockrafärgade tvättas Härerbjudande lyckas mantill Dig! Just har vioch ettkonserveras. oemotståndligt Kultur, äventyr och några av världens allraoch främsta sevärdheter väggarna väntar Dig i nu rena förenaavsvunnen skönhet med en blommande turistindustri. Hotell Vendante är Peking, huvudstaden som ligger i norra utkanten den nordkinesiska slätten. ett trevligt i eni Indiska vacker byggnad Beläget oceanenfrån ligger1800-talet. vackra Mauritius, som lockar med fantastiska Mest imponerande är förstås Kinesiska Muren, sommellanklasshotell nästan 700 milinrymt lång sträcker sig från Gula havet i öster till sin sista utpost i Gobiöknen. I Peking orörda stränder, korallrev och färggranna fiskar. På trestjärniga Hotell Bretin, 3.510 :- vackert beläget mellan de smaragdgröna bergen och det turkosa havet, har 3 nätter finns så mycket att upptäcka. Hotellet Gloria är avfrån turistklass. Priset gäller del i dubbelrum med kontinental frukost. Du nära till både stranden och den vackra kolonialstaden. För Dig som gillar 7 dagar från 8.680 :vandring har även de mäktiga bergen mycket att erbjuda. Priset gäller del i dubbelrum med frukostbuffé.

Uppfyll Din dröm - ring oss redan idag!

1 vecka från 11.995 :-

RESEDRÖMMAR? RESEDRÖMMAR? RESEDRÖMMAR? RESEDRÖMMAR? RESEDRÖMMAR? RESEDRÖMMAR? Uppfyll Din dröm - ring oss redan idag!

Priset gäller del i dubbelrum med halvpension.

Uppfyll Din dröm - ring oss redan idag!

Figure 2: Example of fictitious advertisements used as test material in the significance evaluation. All advertisements were designed as an offer for a trip to a specific destination and hotel. The advertisements were created in 210 x 148 mm (A5) and consisted of a logotype, a catch phrase, an image and a text area (Figure 2). All advertisements were in Swedish since all respondents were native speaking Swedes. As in an ordinary advertisement, the variable images in this study were mainly selected to give the respondent a feeling of the destination. Three levels of personalization were determined and denoted level 1, level 2 and level 3. The level 1 advertisements were entirely static and hence all respondents were given the same advertisement. In these advertisements the respondents were addressed as “Dear Traveller”. The level 2 advertisements were the same as level 1 and hence had the same offer and the same image. However, in these advertisements the respondents were addressed with their own name (e.g. “Dear Thomas Mejtoft”). In the level 3 advertisements, the offer was adapted to each respondent in terms of destination, hotel standard and hotel location. The choice of image was based on the destination or the preferred activity while on vacation. The hotel standard and the vicinity of the hotel were based on the respondent’s answer on the importance of the hotel standard and the preferred activity while on vacation respectively. This resulted in the use of nine different images. Further, the respondent was addressed by name in the same manner as the level 2 personalization. 2.3 Defining and Verifying Print Quality Levels To create and define different levels of print quality a test form was printed on a number of different paper grades in a number of different printers. Based on a visual evaluation performed by the authors, three paper/printer combinations were chosen to represent the three levels of colour print quality. The levels were denoted low, medium and high. To verify the accuracy and distribution of these three colour print quality levels a more comprehensive evaluation was performed in terms of technical measurements and a visual assessment with a panel. A test form consisting of a test chart for technical measurements and images for visual evaluation was printed in all three printers. The test chart contained colour patches for measurements of print gloss, print mottle, sharpness, and colour gamut. The test form was printed on the chosen paper grade in the three printers. The colour gamut volume was calculated on the basis of CIELAB-values from printed areas of full tone cyan, magenta, yellow, red, green, blue, black and paper white. The CIELAB-values were measured with a GretagMacbeth Spectroscan spectrophotometer. The settings used were D50, two degrees and the No filter. The test areas for print mottle and sharpness were scanned at 300 ppi using an Epson Expression 10000 XL desktop scanner. The measures of these quality

TAGA JOURNAL VOL. 4



183

factors were calculated using Matlab® image analysis routines, which are developed to correlate with the visual appraisal. Print mottle was measured from areas of full tone cyan and full tone green. Sharpness was measured from two regions; a black vertical and horizontal line on white background and a black vertical and horizontal line on a yellow background. Hence, the sharpness metrics black raggedness vertical, black raggedness horizontal, colour raggedness vertical and colour raggedness horizontal could be calculated. Print gloss measurements were performed with a Zehntner glossmeter ZLR 1050M. Measurements were done on a full-tone black area and on a full-tone green area at an angle of 75º, according to the standard [12]. As mentioned earlier, a visual assessment was conducted to establish the accuracy of the print quality levels used in the study. The outcome of a visual assessment depends on the motif used [13]. Further, an image evaluation will be more useful if several types of images are used [10]. To avoid any influence from the different images included in the study, this assessment involved all nine images used in the advertisements at each of the three colour print quality levels, in total 27 samples. The samples were presented to a panel of 14 experienced print quality observers. The images were presented three at the time; one image printed in each print quality level. The method used was category scaling [10]. Each observer was asked to express their opinion of the general print quality of each sample on a 10-point numeric rating, where a larger number implied better print quality. The assessment was performed in a standard daylight viewing illumination, D50. The result was presented as the mean value for each print quality level. Earlier research [14] has shown that experienced print quality observers and inexperienced respondents might judge print quality differently. Therefore, another visual assessment was performed, this time using the respondents participating in the significance evaluation of the advertisements. Hence, it could be investigated if this difference could affect the results of this study. The samples used were the level 1 advertisements printed in the three colour print quality levels, i.e. the most frequent advertisement in the significance evaluation. The assessment was performed using the same method and under the same conditions as the assessment with the experienced observers. The data regarding each respondents view on the print quality levels also gave another dimension in the analysis of each respondent’s answer in the significance evaluation. This data made it possible to adjust the results and base the three colour print quality levels on how each individual respondent perceive the print quality at the different levels. That is, if a respondent perceived the established medium level quality to be the best, the ranking of this sample was adjusted to be used in calculating the mean value of the high print quality level. Furthermore, this assessment enabled a comparison between the experienced print quality observers and the inexperienced respondents regarding appraisal of print quality. 2.4 Sample Production The three colour print quality levels were all produced by using the same paper grade in different printers. To produce the samples, the three levels of personalization were then printed at the three colours print quality levels. In addition, the three levels of personalization were printed in black and white on the paper/printer combination producing the print quality level denoted medium. Hence, a total of twelve types of samples were produced, nine colour and three black and white. The encoding of the samples is presented in Table 1.


TAGA JOURNAL VOL. 4

184


Table 1: Encodings of levels of personalization and print quality.

Black & White

Level 1

Personalization Level 2

(Static)

(Name)

(Name, Text & Image)

Level 3

P1-QBW

P2-QBW

P3-QBW

Low Color

P1-QL

P2-QL

P3-QL

Medium Color

P1-QM

P2-QM

P3-QM

High Color

P1-QH

P2-QH

P3-QH

For each type of sample a unique set was produced for each respondent and all samples of personalization level 2 and level 3 were conformed to the specific respondent based on the answers obtained from the questionnaire. Even though the static sample could be produced using conventional printing technology, all samples were printed in digital printing to ensure consistency of print quality between the three levels of personalization. In general, conventional printing is regarded to give slightly higher print quality than digital printing. This implies that the static advertisements could potentially get a higher print quality if printed using conventional printing technology. However, this quality difference is constantly changing as the quality of digital printing is improving. As the purpose of this study is not to compare digital and conventional printing the potential quality differences between the two printing technologies will be disregarded. Since all respondents’ sets of samples were unique, there was a risk of copy to copy variation in the production of the samples. Nevertheless, to ensure the accuracy between the different copies and ensure a minimal variation between the different sets, all samples were visually investigated after production. The samples were found to be visually indistinguishable and hence any possible copy to copy variation should not affect the results of this study. 2.5 Significance Evaluation Rank order [10] was the method used in the significance evaluation of the advertisements. This method is suitable when having few samples that are easy to mix up and when the effort from the respondents should not be too demanding. The evaluation was carried out by the 28 respondents. Each respondent was presented to a unique set of all 12 samples. The respondents were asked to view the advertisements one at the time – look at the image and read the text – and decide how appealing each sample was to them, i.e. how well each sample caught their attention. The respondents were then asked to sort the samples in order of preference. The order of precedence was translated to ranking points. The best ranked sample got twelve points; the second best got eleven points and so on down to the lowest ranking sample that got one point. The evaluation was performed in a standard daylight viewing illumination, D50. Each respondent performed a colour vision deficiency test before the evaluation. This test contained two samples from the Ishihara test plates, designed to separate the persons with colour defects from those with normal colour appreciation [15]. No respondents showed any tendencies towards colour defects.

3

Results

3.1 Print Quality Levels The accuracy of the defined print quality levels was confirmed both by the technical measurements (Table and Table ) and the visual assessment by the experienced observers. The high quality samples had larger colour gamut volume and higher print gloss than the medium quality samples. No significant differences between these two print quality levels were found in print mottle or sharpness. The medium quality sample had larger gamut volume and lower colour raggedness, i.e. better sharpness, than the low quality sample.

TAGA JOURNAL VOL. 4



185

Table 2: Technical measurements of colour gamut volume, print mottle and print gloss. Colour Gamut Print Mottle Print Gloss Volume Cyan Green Black Green QL

219050

0.06

0.25

67.08 68.85

QM

308330

0.03

0.14

32.57 42.73

QH

417910

0.02

0.25

97.80 99.10

Table 3: Technical measurements of sharpness. Black raggedness Colour raggedness Horizontal Vertical Horizontal

Vertical

QL

5.51

18.75

44.83

38.07

QM

5.78

5.26

5.63

4.96

QH

7.66

6.18

7.66

5.73

The visual assessment of all the nine images in the three colour print quality levels, performed by the experienced observers, showed that the print quality levels were accurately defined (Figure 3), regarding both order and distribution. Print Quality Levels 10 9 8 7 6 5 4 3 2 1 0

5.5 4.3 2.9

QL

QM

QH

Figure 3: The results from the visual assessment of all the nine images in the three colour print quality levels (with 95% confidence intervals), performed by the experienced observers. The visual assessment performed by the inexperienced respondents, indicated that they also found the print quality levels as accurately separated. Hence, in general both the experienced and the inexperienced respondents assessed the print quality levels similar. However, as expected, the individual observers had different views on the range, order and distribution of the print quality levels. 3.2 Outcome of Significance Evaluation The results from the significance evaluation were cross-referenced to the background data collected from each respondent. It could be concluded that there were no considerable differences in opinion based on gender or age. In the significance evaluation, the respondents put the samples in order of precedence. The significance evaluation showed that the samples with level 3 personalization were, at all three colour print quality levels, preferred over the level 1 and level 2 personalization samples (Figure 4 and Table 2).


TAGA JOURNAL VOL. 4

186


12 10 8 6 4

P3

2 P2

0 QBW

QL

QM

QH

P1

Figure 4: Mean value according to print quality level and personalization level. Table 2: Mean value according to print quality level and personalization level. Level 1


(Static)

(Name)


Black & White

2.82

3.04

6.18

Low Color

3.46

3.68

8.79

Medium Color

7.39

7.86

10.11

High Color

6.86

7.14

10.68

Level 3

Basically print quality is something that is perceived by the customer rather than objectively determined by a printing house, printing press manufacturer or an expert panel. As mentioned earlier, the visual assessment performed by the respondents indicated that there were differences in how the different respondents appraised the print quality. Adjustments were made to the data from the significance evaluation using the outcome of the visual assessment performed by the respondents. This made small but noticeable corrections to the results (Figure 5).

12 10 8 6 4

P3

2 P2

0 QBW

QL

QM

QH

P1

Figure 5: Adjusted mean value according to print quality level and personalization level.

TAGA JOURNAL VOL. 4



187

Table 3: Adjusted mean value according to print quality level and personalization level. Level 1


(Static)

(Name)


Black & White

2.82

3.04

6.18

Level 3

Low Color

3.43

3.64

8.93

Medium Color

7.00

7.54

10.00

High Color

7.29

7.50

10.64

It could be noted in Figure 5 that the dip in Figure 4 in grading between medium and high quality for level 1 and level 2 personalization evened out when the results were adjusted to conform with how each respondent perceived the print quality.

4

Discussion

Defining accurate and relevant print quality levels was a difficult task since print quality is a subjective matter. It was also hard to analyze the exact print quality experience of each respondent. However, the results from the general visual assessment with experienced print quality observers indicated that the print quality levels were relevant. Adjusting the results from the significance evaluation and basing the three colour print quality levels on how each individual respondent perceived the print quality of each different level further limited the influence of any differences in perceived print quality among the respondents. The influence of the choice of personalization levels, type of advertisement, layout etc. can not fully be determined from this investigation. However, the use of a general subject like travel, a basic layout and three distinct personalization levels would give a good indication of the general outcome of an analysis of the significance of print quality in variable data printing. As mentioned earlier, digital printing and conventional printing differ in their ability to produce customized printed matters. While digital printing can be used for both static and variable data printing, conventional printing technologies, such as offset, can only be used for printing static matters. If cost and speed is taken into account conventional printing technologies are often used for longer run lengths and digital printing for shorter run lengths or variable data printing since the production cost for digital printing is higher than conventional printing for longer runs (e.g.[2,16]). Personalization Level 1

Level 2

Level 3

(Static)

(Name)

(Name, Text & Image) Conventional Printing Technologies Digital Printing Technologies

Figure 6: Possible printing technologies to use at different levels of personalization. Translating this study into a commercial perspective, means that when moving from a level 1 personalization (static) to a level 2 personalization (name), a change in printing technology has to take place. Accordingly, a level 3 personalization (name, text & image) also has to be printed using digital printing, since some information is personalized (Figure 6). However, it should be


TAGA JOURNAL VOL. 4

188


noted that in this study all samples have been printed in digital printing to ensure consistency of print quality between the three levels of personalization. The mean value over all print quality levels of a level 2 personalization does not differ substantially from using a level 1 personalization, leading to an increase of 5.8% in appraisal by the respondents. Further, adding personalized text and images (level 3) improves the appraisal by another 65%. As mentioned before, customizing some part of a print requires digital printing. Taking into account the higher production cost of digitally printed advertisements (if a long total run length is presumed), the motivation to personalize by using name only is in fact low. Regarding print quality, it is possible to distinguish a mean increase of 33% in appraisal when using colour at the low print quality level instead of black and white printing. However, a further increase of 53% is noticed when increasing the print quality from the low level to the medium level. When increasing the print quality even more there was only a 3.6% increase in appraisal. Looking at personalization level 3, the low quality samples differed only slightly from the medium and high quality samples (8.9 vs. 10.0 and 10.6). However, at level 1 and level 2 personalization, the low quality colour samples clearly differed from the other two levels (level 1: 3.4 vs. 7.0 and 7.3; level 2: 3.6 vs. 7.5 and 7.5). This indicated that the respondents were more forgiving towards low print quality when the advertising information was personalized. 12 10.6

10.0

10

8.9

8

7.0

7.5

7.3 7.5 P1

6.2 6

P2 P3

4

2.8 3.0

3.4 3.6

2 0 QBW

QL

QM

QH

Figure 7: Adjusted mean value, sorted according to print quality level.

TAGA JOURNAL VOL. 4



189

12 10.6 10.0

10

8.9

8

7.0

7.5 7.5

7.3

QBW

6.2

QL

6

QM

4 2.8

3.4

QH

3.6 3.0

2

0 P1

P2

P3

Figure 8: Adjusted mean value, sorted according to personalization level The black and white samples showed the same tendency as the colour samples, i.e. a slight increase in preference between the level 1 and level 2 personalization, but a considerable increase between level 2 and level 3 (Figure 7). At each personalization level the colour samples were more appealing to the respondents than the black and white samples (Figure 8). Further, a level 3 personalization black & white advertisement is less appealing than both level 1 and level 2 personalization colour samples at high and medium print quality in colour. However, as noted in Figure 7 and Figure 8 the respondents believed that the level 3 personalized black and white sample were more appealing than the level 1 and level 2 personalization of the low and medium print quality in colour. This means that if the cost of colour printing, in comparison with black and white, is much higher than the cost of personalization, a black and white printed matter with high personalization could be preferred. This research study is just a brief introduction to investigating the interaction between variable data printing and print quality. Nevertheless, in this investigation the importance of specific images, subject and layout of the advertisements have become noticeable. Future research must take in consideration these factors by extending the investigation even further. For example, the choice of images could affect the results not only by the motif but also by the balance and composition of the colours in the images. That is, an image with highlights would most likely print differently than a highly saturated image. Furthermore, the selection of the image on the static sample is important, since this image is the one that most samples had printed. In this study this was the image that the correction was based on, which raises the validity of the investigation. The levels of personalization could be divided into more unique levels to separate the effect of personalized text and images and investigate these two individually. This is especially important since the results from this study show that personalization with name only (level 1) does not make any significant change in appraisal from a static sample. Could appraisal be raised even further by for example removing the name from level 3 personalization, and making the personalization more “invisible” to the respondent? This could also be studied with for example eye-tracking equipment to investigate how the respondents observe each sample before making a decision of order or preference.


TAGA JOURNAL VOL. 4

190

5


Conclusions

With the type of printed advertisements, print quality levels and personalization levels used here, personalizing with only name led to a rather small increase in respondent preference compared to the static advertisement. However, personalizing with name, text and image led to a significant increase in preference compared to a static or name personalized advertisements. Hence, to attract the respondents a higher level of personalization should be used. The respondents were more forgiving towards low print quality when the advertising information was personalized. Furthermore, personalization contributes more at lower print quality levels, indicating that print quality is more significant in static printed matters or printed matters with a low level of personalization. In general, colour prints seemed to be preferred over black and white prints. No differences were found between the three sets of colour advertisements and the black and white advertisements regarding the significance of print quality and personalization. The highest personalized black and white sample was more appealing than the lower personalization levels of the low and medium print quality in colour. Consequently, a black and white printed matter with high personalization could be preferred if the cost of black and white printing is much lower than colour printing. The conclusions in this study were consistent with previous research regarding both the effectiveness of variable data printing (e.g. [5,7]) and the importance of print quality [10]. However, this study was inconclusive with earlier research that suggests a considerable increase in the response rate when using name personalization instead of static advertisements [7]. Instead, the results from this study indicated an insignificant difference between static and only name personalized advertisements. Future research should consequently include a fourth personalisation level in which the text is personalised but without the name personalisation to clarify the significance of text personalization. In contrast to other studies (e.g. [7]) this study has focused on the respondents experience in a qualitative study rather than actual response rates in a quantitative study. This illustrates that it is possible to get knowledge on the importance of print quality in relation of variable data printing in a time and cost effective study. This study further advances the scientific knowledge by introducing the print quality concept in the discussion of variable data printing. The two concepts are individually important in both research and the industry and this study gives knowledge about where resources should be focused to make a printed matter more effective in catching the attention of the customer.

Acknowledgements The authors would like to dedicate a special thanks to all the respondents in the significance evaluation and the observers in the visual assessment for their participation. We would also like to thank our supervisors Prof. Nils Enlund, KTH, Prof. Björn Kruse, Linköping University, Dr Marianne Klaman, STFI-Packforsk, and Dr. Per-Åke Johansson, STFI-Packforsk for their valuable comments. The Kempe Foundations, the EU Structural Fund, the Swedish printing research program T2F, and the industry participants in the DigiPrint project are gratefully acknowledged for their financial support. The authors are grateful for the scholarship from Tryckeriföreningens stiftelse för utbildning och forskning (TUF), that made possible to present this research at the 2007 Taga Conference in Pittsburgh, USA.

TAGA JOURNAL VOL. 4



191

References 1 2 3 4 5 6 7

8 9

10 11 12 13

14

15 16

Kotler P., Wong V., Saunders J. and Armstrong G., Principles of Marketing, Fourth European Edition, Pearson Education Limited, 2005 Kipphan H., Handbook of Print Media, Technologies and Production Methods, Springer, 2001 Romano F.J., Lee B., Rodrigues A. and Sankarshanan., Professional prepress, printing, and publishing, Prentice-Hall, 1999 Peppers D. and Rogers M., The One to One Future, Currency Doubleday,1993 PODi., Best Practices in Digital Printing, Third Edition, Caslon & Company, January 2003 PODi., Best Practices in Digital Print, Sixth Edition, Caslon & Company, 2006 Broudy D. and Romano F., An Investigation: Direct mail responses, Based of color, personalization, database, and other factors. Digital Printing Council, White Paper, Retrieved July 5, 2005, from http://www.gain.net/PIA_GATF/PDF/romano1.pdf, 1999 Mejtoft T., Strategies for Successful Digital Printing, Journal of Media Business Studies, Vol. 3, No. 1, pp. 53-74, 2006 Dalal E.N., Rasmussen D.R., Nakaya F., Crean P.A. and Sato M., Evaluating Overall Image Quality of Hardcopy Output, Proceedings from IS&T's 1998 Image Processing, Image Quality, Image Capture, Systems Conference, Portland, Oregon, pp. 169-173, 1998 Engeldrum P., Psychometric Scaling: A Toolkit for Imaging Systems Development, Imcotec Press, Winchester, 2000 Saunders M., Lewis P. and Thornhill A., Research Methods for Business Students, Third Edition. Prentice Hall, 2003 ISO 8254-1:1999, Paper and board -- Measurement of specular gloss -- Part 1: 75 degree gloss with a converging beam, TAPPI method, 1999 Field G.G., Test Image Design Guidelines for Color Quality Evaluations, Proceedings from IS&T/SID’s Seventh Color Imaging Conference: Color Science, Systems and Applications, Scottsdale, Arizona, pp. 194-196, 1999 Cui L.C., Do experts and naive observers judge printing quality differently? Proceedings from IS&T/SPIE’s 2004 Image Quality and System Performance Conference, San Jose, California, pp. 132-145, 2004 Ishihara S., The Series of Plates Designed as a Test for Colour-Deficiency, 24 Plates Edition. Kanehara Trading Inc. Tokyo, 2004 Mejtoft T., The Cost of Digital Printing in Newspaper Production, STFI-Packforsk Report 111, STFI-Packforsk, 2005


TAGA JOURNAL VOL. 4

192

EFFECT OF DRYING TEMPERATURE PROFILE AND PAPER ON MECHANICAL ….

Effect of drying temperature profile and paper on mechanical print quality in heatset offset printing Timo Hartus Department of Forest Products Technology Paper and Printing Technology Helsinki University of Technology P.O. Box 6400 FI-02015 TKK Finland [email protected]

Abstract In this paper, the mechanical surface strength properties of offset prints were studied from six different papers and three different heatset drying temperature profiles. The most common types of uncoated and coated papers used in the heatset offset printing process were included in this research. Print gloss, roughness and abrasion resistance were investigated. Roughness values were measured both using the Bendtsen method, and the possibility investigated to characterize paper and print surface evenness by a modified print tack measurement application. The gloss of the dried prints compared to the gloss of the corresponding paper increased predictably. However, as shown by Marttila [1] high drying temperature diminished the gloss level of prints on uncoated papers, related to an observed roughening of the surface made up of the fibre mat In the case of coated papers, the gloss level of prints increased slightly as a function of drying temperature, and exhibited diminished post-print roughness, indicating either the insulating nature of the coating, or its base paper isolating property in respect to ink diluents. Abrasion resistance of the prints increased on uncoated paper grades as the heatset drying temperature rose. The greater surface roughness value predictably also slightly increased the smearing tendency compared to coated print surface in the rub tests, despite the lower overall values of ink transfer. The prints on coated papers had some poorer abrasion resistance when they were dried at high drying temperature than when they had been dried under lower drying temperature. Keywords: Heatset printing, abrasion resistance, gloss, roughness, paper

1

Introduction

During offset printing, viscous ink is transferred onto the printing plate through a train of rollers to form of a thin ink layer. Heatset inks set different than sheet fed inks and have additional parameter on drying. Heatset offset printing provides flexibility and added value to print on coated papers, as well as improved quality on newsprint [2]. In the heatset process, the setting and drying of the ink is promoted by applying surface heat to the print. Both coated and uncoated paper grades are used, and the paper plays a significant role in defining print quality. However, less work on ink gloss and print rub has been done for heatset prints compared to coldset. Due to the high speeds increasingly being used in heatset offset printing, the degree of drying of the print has become a most important question. To avoid problems such as smearing of prints, TAGA JOURNAL VOL. 4


T. HARTUS

193

evaporation has to be as fast as possible. Heatset inks contain 30-40 % mineral oil-based solvent (boiling temperature range about 240-290°C), of which 70-90 % evaporates in the dryer [1]. The other main components are pigment and resin. Resins bind pigment particles together and to the paper. In inks, resin can be in either liquid or solid state, termed soft and hard resins respectively. Pigments are insoluble and crystalline. Simultaneously, therefore, with the evaporation of oil, binders harden, and the ink film develops a solid and glossy appearance. The amount of residual oil in the print is depending mostly upon the drying temperature encountered in the heatset drying oven. Residual oil proportion is also dependent on interactions of the oils and binders in ink and paper compounds [3]. The roughness of the paper greatly determines the roughness of the print [4]. Print roughness largely determines the final print gloss level [5]. The gloss of a print is strongly depending on the underlying paper roughness [6]. In principle, the gloss of paper determines the limits for print gloss [7]. Roughness and porosity affect ink gloss after printing and it shows that ink film splitting event is influenced by these parameters [8]. Recent work has shown that there is also an expected contribution from the refractive index of the ink layer, the higher the refractive index, the greater the Fresnel reflectance [9, 10, 11]. Additives like waxes in ink can contribute to the forming of high print gloss level. In cases where increased drying temperature acts to diminish print gloss, it is accepted as being due to surface fibre roughening, especially on uncoated and lightweight coated papers [12, 13]. With conventional printing methods, evenness or homogeneity of paper greatly determines the limits for micro scale gloss uniformity of the print [7, 14]. Although applying ink normally increases the gloss level as part of the image, it often increases the gloss variation also [15]. Typically, the raising of drying temperature in heatset offset increases the gloss level of the final print with coated paper grades, whereas with uncoated paper grades, gloss value can decrease because of fibre roughening. Fiber roughening is the commonly suspected cause of gloss decrease in heatset printing in the drying stage. The problem is significant when printing paper is uncoated or only lightly coated [16]. Lowering of drying temperature diminishes fibre roughening, but in that case the residual solvent content in the print increases and that may produce setoff smearing problems or weak abrasion fastness. Ink setting and ink drying on paper play important roles in determining the tendency for smearing of the print, either in terms of setoff or in print rub. That is the reason why there have been developed numerous methods for following ink setting and drying after the printing process, for instance, ink tack force development (splitting force) [17, 18, 19]. Also, a novel method and device have been proposed to quantify the point-to-point variation in the rate of ink tack development at millimeter scale [20]. Similarly, abrasion resistance tells how resistant a print surface is against rubbing. There are numerous print abrasion resistance (rub off) testers available in the market. They nearly all adopt a similar rubbing principle and they all cause failure within the ink or between the ink and the paper surface by either rotation. More recently, a new test was proposed by [21], following earlier work by [22], in which a strain was applied to an ink layer without inducing actual rub. This is thought to be more easily related to practice and example correlations were given. In this study, however, the aim is to determine the intrinsic cohesion and adhesion of the ink on paper and so traditional abrasion testing is considered relevant to the job in hand. The purpose of this study is to explain the impact of different paper grades and dryer temperature profile in the print drying stage on print quality and fixing degree (adhesion and hardening) in the heatset printing process. To achieve this, the effect of drying temperature on the quality properties of heatset offset prints with various types of printing papers was studied, and the results used to clarify the possibility of controlling the heatset offset print gloss by adjusting the drying temperature. In this latter context, the development of print gloss and roughness as a function of print drying temperature in a heatset oven is also investigated. Also, it is also observed if the printed surface parameters can predict the resistance against mechanical failures in prints in the cases of different papers and heatset drying oven temperatures.


TAGA JOURNAL VOL. 4

194

2


Experimental

The studied papers were commercial products: 45 g/m2 news (1), 45 g/m2 heatset news (2), 45 g/m2 LWC (3), 56 g/m2 SC paper (4), 75 g/m2 LWC (5) and 80 g/m2 uncoated copy paper (6). Papers were cut to a size of 55 mm x 250 mm for the experiments and printed area was 50 mm times 200 mm in every print. These definitions are summarized also in Table 1.

Sample

Table 1: Summary of commercial paper grades. Type Coated Uncoated

1 2 3 4 5 6

Mechanical Mechanical Mechanical Mechanical Mechanical Woodfree

Newsprint Heatset news LWC Supercalendered LWC Copy paper

Basis weight / g/m2 45 45 45 56 75 80

All the printing experiments were done using one commercial, commonly used, black ink for heatset printing process1. Papers were printing with a Universal Testprinter2 laboratory printing machine using a printing speed of 1 m/s, at a nip pressure of 630 N and forming a printing length of 200 mm. The width of printing disc was 50 mm. In the beginning, 0.3 cm3 of ink were distributed onto the inking device (Universal Testprinter High Speed Inking Device). The distributed ink was spread until ink forms an even ink film. The spreading time was about ten seconds at the rotation rate of 100 m/min. Then a rubber-coated offset disk was set in contact with the ink roller on the inking device, duration five seconds. After each inking of the offset disk, an ink addition of 0.02 cm3 was made on the inking roller to keep ink level constant on the offset disk. Although the ink amounts on the offset disk were nearly constant, the ink uptake in the prints varied greatly among the paper grades. No fountain solution was used – this was omitted primarily for ease of laboratory printing but also to eliminate the effects of moisture variation during the experiment, although in practice this is expected to have a significant effect, especially on fibre roughening. Ink setting onto the papers was estimated by splitting force measurements. The measurements were carried out using a print tack meter, Printack (Figure 1), which is a special accessory mounted on the Universal Testprinter -laboratory printing device [23], and records the time taken to remove a print contact disk, a 20 mm wide rubber coated aluminium roller, via the action of a spring loaded cantilever. This time parameter is related to the viscosity properties, mass of ink available for splitting (roughness dependent) and disc-adhesion properties of the ink, especially being sensitive to the elongation or shortness of the ink as a function of time and extension. The spring force was adjusted to about 2 N before the tests. The print tack meter makes one pull from one position and the print sector turns to the next measurement point. Up to 30 measurements can be done from one print. Whilst it is difficult to represent a given tack force by this cantilever method, as might be offered by other techniques, such as Deltack [24] or ISIT [25], it is useful to compare the release time parameter for the properties of the same ink used on different papers.

1 2

Black heatset production ink for LWC and SC papers, manufactured by Sun Chemical. Universal Testprinter is a product of Testprint B.V. The Netherlands.

TAGA JOURNAL VOL. 4


T. HARTUS

195

Printack

Measurement disc

Print sector

Figure 1: The print tack meter (Printack) connected to the laboratory printing machine The time delay between the end of printing and the first splitting measurement was one second. The ink levels used for the splitting force tests were kept the same as for the drier samples. The release time of the measuring disc is termed the print tack-value. The distance between the disc (the upper removed position, when the disc is totally released from the print) and print on the printing sector was constant; approximately 5 mm. The difference in the distance between the disc and the printing sector has naturally a great effect on measured releasing time, and that is why it has to be constant to get comparable results. Immediately after printing the papers by the Universal Testprinter, the prints were extracted from the printing sector and manually transferred and taped at both ends onto a special, selfmade, print carriage. The print carriage has an even area for 50 mm wide print instead of the more usual 40 mm wide even area commonly used in Prüfbau-type printability testers. All the other parameters and properties of the print carriage were maintained the same as in the carriages designed for Prüfbau-type printability testers. The carriage-mounted prints were subsequently dried in a FograHot oven. The speed of the carriage passage through the oven was 0.10 m/s. The oven consists of three hot air blowing sections and an on-line temperature meter, which automatically measures surface temperature of the print just after the drying session. No air forming filter was in use. Drying temperature profiles used in this study are presented in Table 2. The purpose of the selected drying temperature profiles was to create a regime of fast heating, and to maximize the web temperature differences, whilst still being able to get an acceptable drying result after the drying units in order to study the changes in print due to temperature rising. Table 2: Temperature profiles of hot air dryer. Temperature of dryer units / °C Heating profile code 1st unit 2nd unit A 300 300 B 300 200 C 150 150

3rd unit 300 150 100

Print gloss was measured by a gloss meter3 (L&W Gloss tester SE224) after the drying of prints. The devise produces gloss values according the standard T480. There was a 75° angle of incidence is in use. The gloss values are relative values from 0 to 100. Five measurements were taken from every sample, and, from these, the average values were calculated. Printed samples were then further evaluated by measuring the optical density with a Vipdens 2000 densitometer (Vipdens 2000, Viptronic GmbH) without a polarization filter. The density 3

L&W SE224 gloss meter is a product of Lorentzen & Wettre


TAGA JOURNAL VOL. 4

196


of the prints was taken as a control of the amount of ink on the paper, and also for calculation of relative abrasion fastness values of prints. Print density was evaluated only on full printed areas, which in this case meant that the inking disc was loaded to a constant amount but the papers accepted different weights. Bendtsen roughness of studied paper and the dried prints was measured according to the standard SCAN P21-67 with an L&W SE114 Bendtsen Tester4. The evenness of the papers and dried prints after pick testing was estimated by a specially designed measurement, using the print tack meter described above. The fixing degree of digital toner was characterized with this equipment in a previous study [26]. In this study also, the tack measurement disc was “pre-inked” with pick test oil (IGT, normal tack). The distribution of pick oil was made by a high speed inking device (speed 100 m/s, time 10 s). The disc was weighed before and after the tack oil distribution, so that the total amount of pick oil on the measurement disc could be determined and found to be about 14 mg, which corresponds to an amount of 2.8 g/m2 after application on the paper. In general, when used as a tack device, several measurements are made from a print as a function of time to monitor drying of the print: the number of print tack measurements from every print is 20. In this study, the average print tack value, in the presence of the pick oil, of each measured test sheet was calculated from the multiple individually measured values, as the test is applied to set ink and the strength of the ink and its adhesion values are being studied which are constant over the time of testing after drying. A PATRA Rubproofness Tester was used for measuring the degree of ink rub off from prints. The test area was round with a diameter of 3 cm. Prints are rotated against a non printed paper surface (uncoated copy paper 80 g/m2, paper 6, was used throughout as the contact paper) under a controlled pressure, being applied by two standard loads having weights of 1 144 g and 134 g, respectively. The number of rotations was 50 in each test. The loosened/transferred ink amount was estimated by measuring the ISO brightness value of the contact besmeared rubbing papers. The ISO brightness values were measured by an L&W Elrepho 2000 spectrophotometer. The prints after the abrasion resistance tests were photographed by a Highwood HW-10L digital microscope connected to a computer. The magnification factor was 200. The images were viewed by a Video viewer program. The images were grey level inverted to optimize the resolution power in the printed image.

3

Results and Discussion

3.1 Printing The various different types of papers took significantly different ink amounts from the offset printing disc. The papers which were produced especially for the heatset printing process, however, gave almost equal print density levels, and reflect the uniformity of papers devised for the market sector. The uncoated news and copy paper took the greatest ink amounts and produced the lowest print density levels. Printed ink amounts and respective measured density values are collected in Table 3.

4

L&W SE114 Bendtsen Tester is a product of Lorentzen & Wettre

TAGA JOURNAL VOL. 4


T. HARTUS

197

Table 3: Ink amounts on the prints and metered density values. Density Ink amount / g/m2 1.86 1.54 2.47 1.18 0.94 1.53 1.40 1.66 1.17 1.71 2.07 1.17

Paper Code 1 2 3 4 5 6

3.2 Setting of the prints The setting of ink on the papers was studied as described above, by metering splitting time with the print tack meter immediately after printing, before the drying stage. The disc release times from the prints are visualized in Figure 2. The coated papers have the longest overall setting times. The slowness of tack rise of the coated papers illustrates that the LWC grades studied are probably clay-rich and, therefore, not fast setting, because ink can not be absorbed directly into the substrate, but rather ink vehicle needs to be separated from the viscosifying ink layer. Furthermore, the surface is more even and ink/paper contact area is smaller than it is with the uncoated papers, with their porous and rough surface. The uncoated papers (newspaper and uncoated copy paper), which are not commonly used in the heatset printing process, gave the shortest setting times, because of the relative ease of ink absorption into the substrate with a quite rough surface structure and also relatively greater surface area. The uncoated papers which are developed especially for the heatset printing process (paper code 1 and 4) have a more tightly packed surface structure and so behaved more like the coated papers. 8 7

Release time, ms

6 5 3

4 4

3 2 1

1

5

2 6

0 0

20

40

60

80

100

Time, s

Figure 2: Setting of ink on studied papers / Release times in the print tack meter (- paper 1, - paper 2, - paper 3, - paper 4, - paper 5 and - paper 6) The step-wise behaviour of some of the print tack curves reflects the measurement method, in which separation time is recorded and not actual tack force. The nature of the extensional properties of the ink to filament breakage can remain constant over certain periods, despite the expected progression of concentration.


TAGA JOURNAL VOL. 4

198


3.3 Drying of the prints A part of the prints, which was not used for the splitting measurement, was dried in the FograHot-oven using selected heating temperatures previously described in Table 2. The web temperatures just after the hot air blowing unit were measured. The observed web temperatures are collected in Table 4. The effect of paper and temperature profile produced significant variations in web temperature between the prints. Predictably, the highest web temperatures were measured from the prints which were dried using the hardest drying, profile A, and with uncoated paper grades. The lowest web temperatures were measured from the prints which were dried using the softest drying, profile C, and with coated paper grades, respectively. All the prints were dry enough after the drying sessions to avoid any setoff smearing. Table 4: Print surface temperature (°C) in the dryer after the 3rd heating unit Paper Code 1 2 3 4 5 6 Heating profile A 210 205 184 216 186 220 B 140 135 125 113 130 108 C 122 126 117 119 124 88 3.4 Estimation and development of print gloss and print roughness The results of Bendtsen roughness measurements are visualized in Figure 3. Evenness of pick response from the papers and prints estimated also by a print tack devise, Printack, and by the developed application, are shown in Figure 4. The samples, which produce greatest roughness values, had lowest pick oil print tack-values, as might be expected from lower contact area, and, therefore, lower thin film forces, due to surface roughness. The uncoated samples, paper and print, 2 and 6, had the greatest roughness and also the poorest pick contact evenness values. Respectively, the coated papers and the prints on coated paper had the lowest roughness and also the highest print pick strength evenness values.

Roughtness (Bendtsen), ml/min

The roughness of the prints was, under certain conditions and print-paper combinations, greater than the roughness of the corresponding unprinted papers. The uncoated papers had the greatest difference in the roughness level before and after heatset printing. Print tack pick measurements gave significantly higher values from the papers than from the prints.

300 250 P

200

A B C

150 100 50 0 1

2

3

4

5

6

Samples

Figure 3: Roughness–values (Bendtsen) measured from the papers and the dried prints with different temperature profiles. Samples 1-6 mean the codes of studied papers

TAGA JOURNAL VOL. 4


T. HARTUS

199

300

Release time, ms

250 P

200

A B

150

C

100

50 1

2

3

4

5

6

Samples

Figure 4: Print tack pick-oil release times measured from the papers and the dried prints with different temperature profiles. Samples 1-6 mean the codes of studied papers The connection between the roughness values and pick-oil release times in the print tack meter values is evident (Figure 5). Print tack pick-oil measurements separate better the samples which have low roughness than does the Bendtsen roughness measurement method. It can be observed that at a range of high roughness, the Bendtsen-method separates samples better than the print tack pick-oil measurement in both the cases of papers and prints. 300

250

Release time. ms

Coated papers 200 Uncoated heatset papers 150

100 Uncoated papers

50

0 0

0,001

0,002

0,003

0,004

0,005

0,006

0,007

1/(Roughness)2

Figure 5: Print tack pick-oil release times of papers and prints as a function of Bendtsen roughness (■ means on print and Ο means on paper) It can be observed that Figure 5 can be fitted to an exponential decay function where both unprinted and printed papers tend to fall on the same line. The air leak method of the Bendtsen roughness [27] depends on the square of the cross sectional lateral roughness area, i.e. obeys Poisueille flow, whereas the print tack pick-oil value may be expected to follow the two dimensional contact area as also surface gloss is proposed to correlate to the surface roughness [28], which, for Gaussian surface profile statistics, will depend on the reciprocal of the roughness area such that the graph should follow a 1/x2 behaviour. Surface roughness and coating pore structure control the pick-oil or ink tack dynamics [29], i.e. smooth paper surface with small scale pores is proposed in this study to have faster pick-oil or ink setting rate, and


TAGA JOURNAL VOL. 4

200


also longer print tack release times, than paper surfaces having higher roughness or a more rough-grained pore structure.

3.5

Abrasion resistance of prints

ISO Brightness of contract rub paper

In Figure 6, the ISO brightness values of the PATRA rubbed papers are presented. High brightness value means that only a small amount of rubbed ink has been transferred to the rubbing paper, i.e. good abrasion resistance. In Figure 6, the rubbed prints, i.e. the remaining printed surfaces after rubbing, are presented as inverted microscopic images, in which the white areas correspond to the retained black printed ink and the dark areas correspond to the paper surface where ink has been rubbed away. 94 92 90 88

A B

86 84 82 80

C

1

2

3

4

5

6

Samples

Figure 6: Abrasion test results estimated from ISO brightness values of the contact rubbing papers. (Studied papers are marked using the codes 1-6 and dryer heating profiles using the codes A, B and C as in the previous Figures) The application of drying temperature is making a glossy and even print surface and supports the fixing of ink onto the paper surface by causing an intimate contact between paper and the ink layer. The better fixing/adhesion degree is reached mainly by the greater contact and greater adhesion forces affecting between the ink and paper layers. Also, it is assumed that few chemical bonds can be formed between fibres and ink binders [30]. Too little residual solvent content may cause a brittle ink layer which may reduce the mechanical strength between ink and paper, and in this case, mechanical failures between the layers can occur easily under quite small stress, especially in the case of coated papers, where high drying temperatures are required to fix ink to the paper coating [3]. On the other hand, too much residual oil can lead to smearing. From Figure 7, it can be seen that the hard drying temperature, profile A, caused most fibre roughening. From the raised fibre surfaces more ink was rubbed away than from the more even or low lying print surfaces. The fibres, which have been raised above the print surface plane, are easily seen in prints 1A, 2A, 4A and 6A.

TAGA JOURNAL VOL. 4


T. HARTUS

201

Figure 7: Prints after the abrasion test (gray scale inverted images; more black colour in image corresponds to more white in the print). Paper codes 1-6 and drying temperature profiles A, B and C

Gloss-value, %

3.6 Gloss of the prints The measured gloss values (Figure 8) showed, as expected, print gloss to be higher than the gloss of the unprinted papers. No significant gloss differences were found between the drier heating profiles. Naturally, high drying temperature caused minor lowering of gloss especially in uncoated paper grades, which is probably due to the observed fibre roughening. In coated paper grades high heating temperature produced even higher gloss than lower drying temperatures.

90 80 70 60 50 40 30 20 10 0

P A B C

1

2

3

4

5

6

Samples

Figure 8: Gloss of studied papers and prints (P means unprinted paper before printing, A, B, C refer to the printer dryer heating profile) The measured print tack pick-oil release times, i.e. strength evenness of the printed surface, correlated well with the measured gloss values. We see that there are different correlations with gloss depending on whether we are considering unprinted or printed samples, which are clearly observed in Figure 9. This may be due to surface chemical differences between papers and prints, though more likely it is the extra dependence of gloss on ink refractive index. This is also


TAGA JOURNAL VOL. 4

202


supported by the analysis of print tack pick oil release times as a reciprocal function of roughness.

90 80 70

Gloss, %

60 50 40 30 20 10 0 0

50

100

150

200

250

300

Release time, ms

Figure 9: Gloss of papers and prints as a function of print tack pick-oil release time (■ means a print and Ο means a paper) 3.7 Effect of surface evenness on abrasion resistance of the prints Only minor variations were found in the abrasion resistance between the heatset drying temperature profiles of prints on any one paper (Figure 10). Roughness or print tack pick oil measurement results alone do not explain abrasion resistance of the prints in this study. Selected paper type, however, did produce major differences in the abrasion resistance. Uncoated, news and copy paper (1, 2 and 6), had the best abrasion resistance, though they had high roughness and low print tack pick-oil values. The coated papers (3 and 5) had the highest surface evenness but they gave poorer abrasion resistance than news and copy paper. The SC paper (4), which was produced specially for heatset, had evenness between coated and uncoated papers, but its abrasion resistance was the poorest of all the prints.

TAGA JOURNAL VOL. 4


T. HARTUS

ISO brightness of the contact rubbing papers

100

203

6 1

95 2

90

3

85

5

80 75 4

70 65 60 55 50 0

50

100

150

200

Release time, ms

Figure 10: ISO brightness of the contact rub papers as a function of print tack pick-oil release time of the prints The lack of correlation between purely surface-related measurements and print rub resistance has been discussed in depth by [31] in which it was shown that print rub resistance is a combination of substrate compressibility, as well as contact smoothness and surface pore structure. The differences between single coated and multicoated grades were demonstrated, and the increased rub resistance of single coated papers compared with multicoated papers was accounted for by the manifestation of the compressibility of the substrate through the single coating layer. A similar argument can be used here now for the uncoated grades, i.e. compressibility of the uncoated grades is greater than for the coated papers, and the protection of the exposed ink layer within the uncoated paper surface roughness provides for improved resistance to abrasion.

4

Conclusions

The roughness of the paper determines greatly the final print roughness or evenness level. The surface roughness also determines greatly the limits for the surface gloss. High drying temperature causes fibre roughening in the heatset printing process, especially in the case of uncoated paper grades, and this is associated with diminishing gloss. On the other hand, higher drying temperature produces a more even ink surface strength and a glossy ink film on paper, manifest in raised print gloss-sheet gloss differential. The method for predicting surface evenness by print tack measurements using tack-oil as an inter-medium proved to be useful for defining the surface strength parameters in terms of contact forces. The connection between contact evenness and print or paper gloss has been demonstrated. Abrasion resistance is depending on developed interactions between ink and paper or paper coating. Paper surface structure is one important factor in adhesion of ink on paper. Rough surfaces have greater surface area for good mechanical fixing. The bonding of paper coating onto base paper has to be stronger than the adhesion of ink to the coating to prevent delamination at the coating base paper interface. Neither paper nor print roughness/evenness provided an unequivocal estimate of print abrasion resistance.


TAGA JOURNAL VOL. 4

204


If heatset drying temperature is too high, it can cause too low residual solvent content, and ink film becomes too hard and rigid. Too rigid an ink layer seems to weaken the mechanical strength between ink and paper, or the paper coating layer, and, in this case, mechanical failures between the layers can occur easily under small stress. In the case of coated papers this failure can be between ink and paper coating. Also fibre roughening became a reason for poorer abrasion resistance in the cases of some hard dried prints.

Acknowledgements The author is grateful to Professor P.A.C. Gane for his valuable advice and comments during writing this article.

References 1

2 3 4

5 6 7 8 9

10

11 12

13 14

15 16 17

Marttila J., Oittinen P., Grönlund A., Hartus T., Korhonen J. and Marchart P.,Heat-set offsetpainatuksen kuivatusolosuhteiden hallinta, Teknillinen korkeakoulu. Graafisen tekniikan laboratorio. Julkaisu 19. 40 p. (in Finnish), 1995 Wells N., Hybrid heatset and coldset printing, Newspapers & Technology, 1-5, 2001 Hartus T. and Oittinen P., Characterisation of the Drying Properties of Heatset Inks by thermal Methods, Graphic Arts in Finland, Vol. 25, no 1, pp. 9-15, 1996 MacGregor M.A., Johansson P.-Å. and Béland M.-C., Measurement of smallscale gloss variation in printed paper. Topology explains much of the variation for one paper, Proceedings of the 1994 International Printing and Graphic Arts Conference, pp. 33-43, 1994 Bousfield D.W., A model to predict the leveling of coating defects, Tappi J., Vol 74, no 5, pp. 163–70, 1991 Handbook of Print Media: Technologies and production methods, Kipphan, H. (ed.) Springer Berlin 2001 pp. 223-226, 2001 Oittinen P., The limits of gloss in prints, Paperi ja Puu Vol. 65, no 11, pp. 718–724, 1983 Jeon S.J. and Bousfield D.W., Print gloss development with controlled coating structures, Journal of Pulp and Paper Science, Vol. 30, no 4, pp. 99-104, 2004 Preston J.S., Hiorms A.G., Elton, N.J. and Ström J., Application of imaging reflectometry to studies of print mottle on commercially coated papers, Tappi Coating & Graphic Art Conference, Atlanta, Proceedings, 2006 Sorjonen M., Jääskeläinen A., Peiponen K-E. and Silvennoinen R.,On the assessment of the surface quality of black print paper by use of a diffractive opticalelement-based sensor, Measurement Science and Technology, Vol. 11 (2000), pp. N85-N88, 2000 Granberg H., Spectral Light Absorption and Gloss of Cyan & Magenta Inks, Acreo, T2F Report f59, 2002 Chinga G., Stoen T. and Gregersen Ø., On the roughening effect of laboratory offset printing on SC and LWC surfaces, Journal of Pulp and Paper Science, Vol. 30, no 11, pp. 307-311, 2004 Sederholm R., The influence of Coating layer on heatset roughening, MSc. Thesis, Helsinki University of Technology, 59 p. (in Finnish), 1998 Béland M.-C. and Bennett J.M., Effect of local microroughness on the gloss uniformity of printed paper surfaces, Applied Optics Vol. 39, no 16, pp. 2719-2726, 2000 MacGregor M.A. and Johansson P.-Å., Gloss uniformity in coated paper; Measurements of commercial papers, TAPPI Coating Conference Proceedings,1991 Korhonen J., The Effect of Paper and Ink on Heatset Drying, Graphic Arts in Finland, Vol. 22, no 2, pp. 16-22, 1993 Concannon P. W. and Wilson L.A., A Method for Measuring Tack Built of Offset Printing Ink on Coated Paper, TAGA Proceedings 44 Vancouver, 1992

TAGA JOURNAL VOL. 4


T. HARTUS

205

18 Plowman N., Ink tack – Its effects and results, Graphic Arts Monthly Vol. 61, pp. 68, 1989 19 Gane P.A.C., Seyler E.N. and Swan A., Some novel aspects of ink/paper interactions in offset printing, International Printing and Graphic Arts Conference, Halifax, Nova Scotia, Tappi Press, Atlanta, pp. 209-228, 1994 20 Xiang Y., Bousfield D.W., Hassler J., Coleman P. and Osgood A., Measurement of local variation of ink tack dynamics, Journal of Pulp and Paper Science, Vol. 25, no 9, pp. 326-330, 1999 21 Gane P.A.C., Kozlik T, and Schoelkopf J., Print Rub Determination: A Novel Laboratory Method to Simulate Practice, Pulp and paper Canada Vol. 106, no 10, pp. T21 -T219, 2005 22 Gumbel R., Converting prints on matt coated paper, International Symposium on Paper Coating Coverage, Helsinki. Training Centre of Finnish Forest Industries AEL-Metsko, 32pp, 1989 23 Testprint, Inc. Product Information, from http://www.testprint.com/ Retrieved 19.11.2005 24 Prüfbau. Technical Information, Prüfbau Multipurpose Tack Measuring System Deltack, Prüfbau, Peißenberg/München 33 p, 2003 25 Gane P.A.C. and Seyler E.N., Tack development: an analysis of ink/paper interaction in offset printing, Tappi Coating Conference Proceedings. San Diego, Tappi Press, Atlanta, pp. 243-260, 1994 26 Hartus T., Effect of Toner Fixing Temperature on Print Properties in the Electrophotographic Process, TAGA Journal, 3.3, pp 165-177, 2008 27 Heinemann S., Estimation of Surface Roughness and Air Permeabilility of Paper According to Bendtsen, Papier (Darmstadt), Vol. 50, no 5, pp. 233-241, 1996 28 Gate L., Windle W. and Hine M., The relationship between gloss and surface microtexture of coatings, Tappi Journal, Vol. 56, pp. 61-65, 1973 29 Xiang Y. and Bousfield D.W., Influence of Coating Structure on Ink Tack Dynamics, Journal of Pulp and Paper Science, Vol. 26, no 6, pp. 221-227, 2000 30 Hartus T. and Oittinen P., Päällysteen ja painovärin kuivumistulokseen vaikuttavat vuorovaikutusilmiöt (Drying Interactions of Paper Coating and Heatset Ink), Uudistuva Paperi -seminaari, Espoo, Poster presentation (in Finnish), 1995 31 Gane P.A.C., Ridgway C.J. and Gliese T., A re-evaluation of factors controlling print rub on matt and silk coated papers, Tappi Coating and Graphic Arts Conference, New Orleans, Tappi Press, Atlanta, 2006


TAGA JOURNAL VOL. 4

206

FACTORS IMPACTING THE EVALUATION OF PRINTER PROFILE ACCURACY

Factors Impacting the Evaluation of Printer Profile Accuracy Robert Chung Rochester Institute of Technology School of Print Media 69 Lomb Memorial Drive Rochester, NY 14623, USA [email protected]

Abstract The evaluation of printer profile accuracy is affected by many factors and the printing device represents only one such factor. This paper examines the effect of profiling target layout and patch size on printer profile accuracy. Two digital color printing devices were evaluated using a single color measurement system. The evaluation avoids temporal variation associated with the color printing device on purpose. It was found out that (1) the spatial color uniformity of the device bears a larger impact on its colorimetric accuracy than factors such as target layout and patch size; (2) the use of the random target over the visual target helps minimize spatial non-uniformity in the shadow region of the color gamut; and (3) there is no significant difference in colorimetric accuracy when reducing the patch size of the random target from 6mm to 4mm using the color measurement instrument in this experiment. Keywords: ICC, profile, accuracy, colorimetry

1

Introduction

Many people in North America live in the world of eight-and-half-by-eleven. The notepad is 8.5 by 11; the magazine is 8.5 by 11; and the most popular paper size for desktop printers is “Letter Size,” and that is 8.5 by 11 inches. Having the same dimension helps to keep documents in a neat pile. There is finite area in an 8.5 by 11 inch space. If one takes one inch out as margins, the printable area of an 8.5 by 11 page is 6.5 x 9 inches or 58.5 in2. This translates into 37,741 mm2 (one inch is equal to 25.4mm). You may question what’s all the fuss! When the first CGATS-endorsed profiling target, IT8.7/3, came out in mid-1990s, it contained 928 6mm patches and the entire target would fit into an 8.5 by 11 page with 6mm patch size nicely. As a matter of fact, the “Letter Size” page with one-inch margin can accommodate 1,048 6mm patches. When CGATS introduced the IT8.7/4 profiling target with 1,617 6mm patches [1], the magic world of 8.5 by 11 no longer could cope with the addition of patches that were deemed necessary by the color management communities. The only recourse without changing the paper size while keeping the target intact is to reduce the patch size. For example, if one reduces the patch size from 6mm to 4mm, it can accommodate up to 2,358 patches. Incidentally, neither IT8.7/3 nor IT8.7/4 specifies the patch size as the normative part of the specification. The default patch size of 6mm came from the days of IT8.7/3 and constraints in color measurement instrumentation. In other words, the patch size of 6 mm is not a part of the standard, but the user will confront the issue of target size when implementing color management.

TAGA JOURNAL VOL. 4


R. CHUNG

207

In addition to the increase in number of patches, CGATS also introduced a random layout of the target along with the visual layout in the informative part of the standard. Figure 1a is a reduced size of the IT8.7/4 random target and Figure 1b is a reduced size of the IT8.7/4 visual target.

Figure 1: Two default layouts of the IT8.7/4 target 1.1 Problem Statements Printer profile accuracy depends on the following immediate factors: the profiling target, the color printing device, the color measurement instrument, the profile-making software, the CMM, etc. Depending how printer profile accuracy is tested, spatial uniformity of the printing device becomes a factor if it involves comparison of colorimetric measurements from more than one location within the sheet. If it involves comparison of colorimetric measurements between the press run that generates the press profile and the press run that applies the press profile, color repeatability of the printing device then becomes a factor. Thus, it is important to ask, “Which testing condition is suitable for testing printer profile accuracy?” There are two default target layouts, i.e., random vs. visual, mentioned in the IT8.7/4 specifications. The concept of randomizing patches is believed to minimize the effect of inherent device noises associated with inking evenness, ink starvation, etc. The question of interest becomes, “To what extent is the random target more effective in minimizing printer noise than the visual target?” The dimension of IT8.7/4 with a patch size of 6mm no longer fits within the letter size. A question of equal interest becomes, “Is there an adverse effect in color measurement accuracy when patch sizes are reduced?” These questions regarding target orientation, patch size, and measurement noise were initially explored by Chung [2] and are now more fully examined. The researcher recognizes that the experimental findings will not be absolute, but relative to the variables compared, e.g., visual layout vs. random layout, 6mm patch size vs. 4mm patch size. Furthermore, specific findings depend on the testing conditions, e.g., printing devices tested and color measurement instrument used.

2

Literature Review and Pilot Study

A printer profile contains Look-up Tables (LUTs) between CMYK and CIELAB. It assumes that there is no variability within a device so that each combination of CMYK is mapped to a unique CIELAB values, and vice versa. In fact, a printer can vary both spatially and temporally


TAGA JOURNAL VOL. 4

208


[3]. By understanding the magnitude of the variations, one can begin to have a handle on the estimation of printer profile accuracy. Variability is the degree of repeatability or precision of the measured outcome. When a person measures anything only once, he knows nothing about its variability. If one measures a printed color chart more than once, he can assess the variability associated with a color measurement system or a printing device. 2.1 Instrument Repeatability Researchers at Rochester Institute of Technology devised a method to estimate color measurement instrument repeatability [4]. The Committee for Graphic Arts Technology Standards also recommends the method whereby cumulative probability plot of ∆E distribution from a multi-patch color chart is used for evaluation [5]. The procedures are to (1) measure a printed IT8.7/3 (basic) target twice using one color measurement instrument; (2) calculate colorimetric difference between any one of the two colorimetric measurements and its average; and (3) arrange the ∆E distribution of all 182 patches in the form of relative frequency (better known as a pie chart) or as a cumulative relative frequency (CRF). The CRF of ∆E becomes a graphic depiction of the instrument repeatability. A Spectrolino/Spectroscan was used in this study. Measuring a printed color chart twice is the minimum number of times to assess repeatability of a measurement system. The repeatability of this instrument, as shown in Figure 2, shows that one-half of the time, the uncertainty of a color measurement is 0.2 ∆E or less; and 90% of the time, the uncertainty of a color measurement is 0.4 ∆E or less. For purpose of comparing variations induced by different factors, the scaling of the CRF curve is normalized from 0 to 10 ∆E in the x-axis. 1.0 0.9

CRF Curve -- Instrument Repeatability

0.8 0.7 0.6 0.5 0.4 0.3 0.2

Spectrolino/Spectroscan

0.1 0.0 0

1

2

3

4

5 ²E

6

7

8

9

10

Figure 2: Instrument repeatability of Spectrolino/Spectroscan 2.2 Spatial Uniformity of the Printing Device Spatial uniformity of a CMYK printing device can be estimated similarly with the use of a minimum of two identical IT8.7/3 (basic) color blocks printed on the same sheet (Figure 3). Colorimetric difference between any one of the two colorimetric measurements and its average are calculated. The ∆E distribution of all 182 patches in the form of cumulative relative frequency (CRF) becomes a graphic depiction of the spatial uniformity of the output device [6].

TAGA JOURNAL VOL. 4


R. CHUNG

209

Figure 3: Two IT8.7/3 (basic) color blocks with one rotated Figure 4 shows an example of spatial uniformity of a digital press using the analysis method outlined above. By comparing with the instrument repeatability, one can see that one-half of the time, the uncertainty of spatial uniformity is 0.6 ∆E or less; and 90% of the time, the uncertainty of a spatial uniformity is 1.5 ∆E or less. 1.0 CRF Curves -- Spatial Uniformity

0.9 0.8 0.7 0.6 0.5 0.4 0.3

iGen3

0.2


0.1 0.0 0

1

2

3

4

5 ²E

6

7

8

9

10

Figure 4: Spatial uniformity of a digital press 2.3 Temporal Consistency of the Printing Device Temporal consistency of a CMYK printing device can be estimated similarly with the use of two IT8.7/3 (basic) color blocks printed at different time. Here, spatial uniformity (the average of the sample measurements) is differentiated from temporal consistency (the difference between the sample value and its target value). Figure 5 shows the CRF curve between two digital press runs. One can see that one-half of the time, the uncertainty of temporal consistency of the digital press is 2.6 ∆E or less; and 90% of the time, the uncertainty of temporal consistency is 5.7 ∆E or less. Multiple CRF curves may be generated between sampled press sheets and the target to reflect temporal consistency of the press run.


TAGA JOURNAL VOL. 4

210


CRF Curves -- Temporal Consistency 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 NexPress Run 1 vs. NexPress Run 2

0.2


0.1 0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 5: Temporal consistency between two press runs Instrument repeatability, spatial uniformity, and temporal consistency are independent factors that impact the repeatability of color measurement data. When evaluating printer profile accuracy, one has to live with the uncertainty of a color measurement instrument and the spatial uniformity of the printing device. 2.4 Profiling Target Layout Different target layout and the profile-making software react to spatial non-uniformity differently. There was a curiosity to learn if target layout influences the color measurement accuracy of the profiling target that, in turn, impacts the printer profile accuracy. If the spatial uniformity of an output device is good, patch layout of the profiling target should have little impact on the colorimetric accuracy of the resulting profile. On the other hand, if the spatial uniformity of an output device is poor, patch layout of the profiling target and the profilemaking software may have a significant impact on the colorimetric accuracy of the resulting profile. Thus, to determine the effect of target layout on printer profile accuracy remains as an objective of the study. 2.5 Measurement Error due to Patch Size Spooner discusses the effect of color measurement error due to patch size [7]. When the light of the measuring instrument that illuminates in all directions of a translucent substrate such as paper, some of the light that diffuses laterally out of the lighted area diffuses back to the lighted area. If the measured area is equal to the patch size, then some laterally diffused light will exit through the sample edges and back, and thus the measurement is influenced by adjacent colors. He called such an effect, lateral diffusion error (LDE). To avoid this type of measurement error, ISO 5/4 [8] specifies that the patch size should be 2mm larger on all sides from the measured area. The default width of a color patch in the IT8.7/4 target is 6mm wide. The diameter of the Spectrolino/Spectroscan’s hold-down aperture, used in the experiment, is 5mm. The measured area is about 4mm (Figure 6). In this case, there was a concern regarding the measurement error when reducing the patch size from 6mm to 4mm.

TAGA JOURNAL VOL. 4


R. CHUNG

211

Patch size (6mm)

Measured area (4mm) Hold-down aperture (5mm)

Figure 6: Schematics of patch dimension and measured area

3

Methodology

In this research, the IT8.7/4 (full or 1,617 patches) targets were used to build printer ICC profiles. The IT8.7/3 (basic block with 182 patches) target was used as input to the evaluation process. These targets were printed to a KPG Approval color proofer and a Xerox DocuColor 6060 digital printer. The GMB Spectrolino/Spectroscan was used to measure all color patches (CIELAB, D50, 2 degree). In other words, one set of color measurement was made from one KPG Approval print and one Xerox Docucolor 6060 print containing the targets as described. In addition, the GMB ProfileMaker 5.0 was used for ICC profile construction and CHROMIX ColorThink 3.0 Pro for data extraction from ICC profiles. ∆Eab was used to express the color difference as opposed to using other ∆E formulas because it is recommended by CGATS and ISO 12647. 3.1 Testing the Effect of Target Layout To test if there is a significant difference in colorimetric accuracy between the visual target and the random target, two devices, KPG Approval and Xerox 6060 and the standard 6mm target were chosen for the experiment. It is hypothesized that the 2,540 spots/in (spi) with dye diffusion thermo transfer based KPG Approval is a spatially uniform output device. Thus, the difference in target layout will have less impact on colorimetric accuracy than that of the 600 spi dry toner based Xerox 6060. Below is the experimental procedure for testing the effect of target layout. a)

b) c)

Determine spatial uniformity of the devices by printing two IT8.7/3 (basic) color blocks within an A3 sheet. The ∆E distribution (CRF curve) between individual measurements and their averages is an indication of the spatial uniformity of the output device. More importantly, the average CIELAB values between the two corresponding patches of the two IT8.7/3 (basic) targets represent the reference values when assessing colorimetric accuracy of ICC profiles made from different layouts and from different patch sizes. Print the IT8.7/4 visual target and the IT8.7/4 random target to KPG Approval and Xerox 6060. Construct ICC profiles, with the same CMYK constraints using GMB ProfileMaker 5.0. Test colorimetric accuracy of these ICC profiles by means of output simulation. This is done using the Worksheet feature of ColorThink 3.0 Pro to perform A-to-B or device-to-PCS color conversion. Briefly, a CIELAB list can be generated from a CMYK list via a specific ICC profile and the absolute colorimetric rendering intent. The CIELAB list is the simulated outcome of printing the CMYK target. Because no physical printing device is used, there is no process variation involved.


TAGA JOURNAL VOL. 4

212


d)

Compute colorimetric difference (CRF curve) between the simulated output and the reference value established in Step (3.1a). The CRF curve is a measure of the printer profile accuracy.

3.2 Testing the Effect of Patch Size To test if there is a significant difference in colorimetric accuracy due to patch sizes, KPG Approval with both the 6mm (visual and random) and the 4mm (visual and random) targets were chosen as the testing conditions. It is hypothesized that if patch size is a significant factor, larger colorimetric errors will be detected in the ICC profiles built from reduced patch sizes. The testing procedure is similar to Testing the Effect of Target Layout.

4

Results and Analysis

If there are colorimetric differences due to target layout or patch size, the difference has to be relative to the inherent spatial variation of the device. The researcher will use the results of the spatial uniformity that includes color measurement system error as the starting point to discuss the effect of target layout and patch size on printer profile accuracy. 4.1 Spatial Uniformity of Output Device The spatial uniformity of KPG Approval is shown in Figure 7. Colorimetric differences were between individual measurements and their averages. The ∆E statistics shows that the median ∆E is 0.3 and the 90-percentile ∆E is 0.5. The maximum ∆E of 4 was from the color patch ID 92 with %dot area value of 70C, 100M, 20Y, and 0K. This purplish color patch has the largest spatial color difference. Color patch ID 92 with a coordinate of H1 is located at the bottom center of the IT8.7/3 (basic) target. There was no physical flaw associated the patch and the cause of the color difference was unknown. 1.0 0.9

CRF Curves

0.8

CRF

0.7 0.6

²Eab

Median

0.3

90% Max.

0.5 4.0

0.5 0.4 0.3 0.2

Approval_Average vs. Approval_90

0.1

Approval_Average vs. Approval_0

0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 7: Spatial uniformity of KPG Approval The spatial uniformity of Xerox 6060 is shown in Figure 8. The ∆E statistics shows that the median ∆E is 0.5 and the 90-percentile ∆E is 1.2 with a maximum ∆E of 2.4. Figure 7 and 8 help verify that KPG Approval is more spatially uniform than Xerox 6060 digital printer.

TAGA JOURNAL VOL. 4


R. CHUNG

213

1.0 0.9

CRF Median 90% Max.

CRF Curves

0.8 0.7

²Eab 0.5 1.2 2.4

0.6 0.5 0.4 0.3 0.2

X6060_Average vs. X6060_90

0.1

X6060_Average vs. X6060_0

0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 8: Spatial uniformity of Xerox 6060 4.2 Colorimetric Differences Due to Target Layout To test the colorimetric accuracy of ICC profiles by means of output simulation, a CMYK list of the IT8.7/3 (basic) target was set up in ColorThink 3.0 Pro as a Worksheet. By specifying an ICC profile and absolute colorimetric rendering intent, the software transforms the CMYK list into a CIELAB list via the A-to-B look-up table (LUT). The researcher realizes that the accuracy of the color conversion depends on the CMM and algorithms used in the transform. This is how one would simulate the output device without temporal variability of the printing device. The CIELAB list derived from the above simulation is known as the sample. The sample CIELAB list and the reference CIELAB list, derived from Step (3.1a) of the methodology, are used to calculate ∆E between them. The comparison of colorimetric accuracy between target layouts of KPG Approval is shown in Figure 9. Curve A is the spatial uniformity of KPG Approval, as described in Figure 7. Curve B is the colorimetric difference associated with the profile built from the IT8.7/4 6mm version of the visual target. Curve C is the colorimetric difference associated with the profile built from the IT8.7/4 6mm version of the random target. A

1.0 0.9

CRF Curves

0.8

B

0.7 0.6 0.5 0.4

C

0.3

Approval_Average vs. Approval_V6 Approval_Average vs. Approval_R6 Approval Non-uniformity

0.2 0.1 0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 9: Printer profile performances between two target layouts of KPG Approval Figure 9 shows that the colorimetric error, via output simulation, is around 0.7 ∆E at the 50 percentile. This value is almost three times greater than the spatial uniformity of the Approval or 0.25 ∆E. This means that KPG Approval can match a colorimetric specified color with any possible combination of CMYK values with an average ∆E error of 0.7. In this instance, the error comes from the color measurement device, profile-making software, and CMM used in the ColorThink 3.0 Pro software. Figure 9 also reveals printer profile accuracy between the two target layouts. While the median ∆E from the visual target is slightly smaller than that of the random target, it was uncertain if such a difference is significant. As stated earlier, KPG Approval is a uniform color output


TAGA JOURNAL VOL. 4

214


device. Profiles made with different target layouts may have little effect on their colorimetric accuracy. More importantly, Figure 9 shows that there are more occurrences of large ∆E (greater than 4) values in the visual target than in the random target. Upon further analysis with the use of ∆E sorting feature in the ColorThink 3.0 Pro (Figure 10), these color patches were found to be all 4color black tints with patch ID from 129-134 in the IT8.7/3 (basic) target. The significance of larger errors in 4-color black tints is that the use of gray component replacement (GCR) may help reduce printing variation.

Figure 10: Color patches with ∆E values greater than 4 found in KPG Approval visual target There was only one color patch with ID 92, as shown in Figure 11, having a ∆E larger than 4 in the random target and the patch is made up of 70C, 100M, 20Y, and 0K, the same color patch that was found to have the largest color difference due to spatial non-uniformity. Given that there are choices, the use of the random target over the visual target, couple with the profilemaking software, can help minimize color measurement errors in the shadow region of the color gamut.

Figure 11: Color patches with ∆E values greater than 4 found in KPG Approval random target The colorimetric accuracy comparison between target layouts of Xerox 6060 is shown in Figure 12. Curve A is the spatial uniformity of Xerox 6060, as described earlier. Curve B is the colorimetric difference associated with the profile built from the IT8.7/4 6mm version of the visual target. Curve C is the colorimetric difference associated with the profile built from the IT8.7/4 6mm version of the random target. A

1.0 0.9 0.8

CRF Curves

C

0.7 0.6 0.5

B

0.4 0.3

X6060_Average vs. X6060_V6 X6060_Average vs. X6060_R6 X6060 Non-uniformity

0.2 0.1 0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 12: Printer profile performances between two target layouts of Xerox 6060

TAGA JOURNAL VOL. 4


R. CHUNG

215

Figure 12 shows that colorimetric errors, via the Xerox 6060 output simulation, are greater than 1.5 ∆E at the 50 percentile. This value is, again, three times greater than the spatial uniformity of Xerox 6060 or 0.5 ∆E. Without further testing if the difference between spatial uniformity and temporal consistency are significant, Table 1 reconfirms that colorimetric accuracy of a printer profile begins with spatial uniformity that is inherent in the output device. Table 1: Comparison of colorimetric errors by output simulation ²E at 50 percentile Spatial nonuniformity

Visual target

KPG Approval

0.25

0.7

0.9

Xerox 6060

0.5

1.7

1.5

Randomized target

In term of the effect of target layout on colorimetric accuracy of Xerox 6060 printer profiles, Figure 12 suggests that the median ∆E from the visual target is slightly larger than that of the random target. It is uncertain if such a difference is significant. Figure 12 also shows that there are more occurrences of large ∆E (greater than 4) values in the visual target than in the random target. Using the ∆E sorting feature in the ColorThink 3.0 Pro, it was found out that the same group of color patches with patch ID from 129-134 in the IT8.7/3 target yielded larger ∆E values from the visual target as discussed in the Approval case. Both the magnitude and the occurrence of larger ∆E (greater than 4) were reduced in the random target. 4.3 Colorimetric Differences Due to Patch Size Colorimetric accuracy of two KPG Approval profiles, i.e., Visual_6 and Visual_4, is shown in Figure 13. Curve A is the spatial uniformity of the Approval, as described in Figure 7. Curve B is the colorimetric difference associated with the profile built from the IT8.7/4 6mm version of the visual target and the curve C is that of the IT8.7/4 4mm version of the same target. A

1.0

B

0.9 0.8

CRF Curves C

0.7 0.6 0.5 0.4 0.3

Approval_Average vs. Approval_V6 Approval_Average vs. Approval_V4 Approval Non-uniformity

0.2 0.1 0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 13: Device simulation of KPG Approval using visual target in 6mm and 4mm Figure 13 shows that (1) the median ∆E between the visual 6mm and the visual 4mm KPG Approval targets is the same; (2) there are larger ∆E values associated with the reduced patch size. For example, ∆E at the 90 percentile increased from 1.9 ∆E to 2.6 ∆E as the patch size reduced from 6mm to 4mm. Spooner’s lateral diffusion error (LDE) may very well explain why this had occurred. But the effect of lateral diffusion error did not support the two random targets (Figure 14). As shown in Figure 14, colorimetric differences are relatively small between the random 6mm and the random 4mm targets.


TAGA JOURNAL VOL. 4

216


A 1.0

B

0.9 0.8

CRF Curves C

0.7 0.6 0.5 0.4 0.3

Approval_Average vs. Approval_R6 Approval_Average vs. Approval_R4 Approval Non-uniformity

0.2 0.1 0.0 0

1

2

3

4

5

6

7

8

9

10

²Eab

Figure 14: Device simulation of KPG Approval using random target in 6mm and 4mm By stepping back and examining both Figure 13 and 14 together, it became evident that (1) the random target tends to yield better profile accuracy than the visual target by reducing the occurrence of larger ∆E, and (2) the random target is less affected by the effect of reduced patch size.

5

Discussion

This research work was initially motivated by curiosity of the causal relation between the profiling target variations and the printer profile accuracy. Although measurements can also be affected by ambient illumination entering the instrument, it was realized that printer profile accuracy is the response of a system, and not a single element alone. Figure 15 summarizes the factors that may influence printer profile accuracy. Major factors that contribute to the evaluation of printer profile accuracy include printer repeatability, color management, color measuring instrument, and the evaluation method used. In this research, the researcher paid special attention to the effect of profiling target and the type of printer on printer profile accuracy. He held a number of factors, e.g., color measuring instrument, profile-making software, as constants in the experiment.

Figure 15: Factors contributing to the evaluation of printer profile accuracy Based on our own research work, the magnitude of variation due to color measuring instrument is the smallest of all variations in the system. The smaller patch size of the profiling target can cause additional measurement error due to geometric precision of the positioning mechanism and the halftone structure of the print. It is recommended that more instruments with different measured areas be included as a follow-up study. Particularly, the alignment between the TAGA JOURNAL VOL. 4


R. CHUNG

217

instrument and the patch becomes critical as the measured area approaching to the size of the color patch. Lateral diffusion errors will be more severe for profiling target with small patch sizes in randomized layout than for profiling target in visual layout. This is because the magnitude of these errors varied with the size of the area being measured and the location of the adjacent colors [7]. Printing repeatability is a major factor in the evaluation of printer profile accuracy. Spatial uniformity of printing is adopted as a baseline when evaluating colorimetric accuracy of color printing devices. Temporal consistency and run-to-run variability associated with color printing devices were purposely avoided by means of output simulation with the Worksheet feature of the ColorThink 3.0 Pro. Other than variations in profiling target layout and patch size, color management parameters, e,g., profile-making software, CMYK constraints, CMM, were held as constants in this experiment. As such, the researcher had to accept errors in profile predicted values via device simulation that was three times in ∆E magnitude in comparison to the spatial uniformity of the printing device. If the researcher were to experiment with the profile-making software settings, including the computational ability of the CMM, he would have reduced colorimetric errors in the simulation process. Printer profile accuracy may be assessed by different evaluation methods. When a synthetic color target is used, it involves the A-to-B color conversion where the ‘A’ space is a CMYK color space. The use of a colorant-based synthetic target, such as the IT8.7/3 (basic), proved to be effective in the evaluation of color repeatability of printing devices in color-managed workflows. Other target, e.g., the ECI 2002 profiling target, with ECI published characterization data set as the reference was also used [9]. In this case, the device uniformity is excluded in the evaluation process. There are two limitations with the use of a colorant-based synthetic target as experienced in this research: (1) only the A-to-B LUT of the printer profile is evaluated and the B-to-A LUT is not; (2) the findings do not necessarily predict visual responses of pictorial color image match. As suggested in Figure 15, it is useful to include pictorial color images with synthetic colors that are sampled from various pictorial scenes as the input target in the evaluation process. It is recommended that pictorial color images and an image-based synthetic target are included in the evaluation of printer profile accuracy in the context of digital proofing workflow as a follow-up study.

6

Conclusion

There was little research on conducting printer uniformity and temporal consistency when filmbased proofs were the norm and pressmen were asked to print to visually match the supplied color proofs. ICC-based color management changed all that. In particular, color proofs are to match standard printing conditions. Color matching between two printing devices begins with accurate ICC profiles. This paper is an attempt to better understand factors that impact printer profile accuracy. Colorimetric accuracy of a printer ICC profile can be affected by a number of factors. From the variables tested, device uniformity is the most important factor, i.e., the more uniform the device is, the more repeatable the color will be. Factors such as target layout and patch size are of secondary importance. The use of the random layout over the visual layout of the profile target helps minimize spatial non-uniformity in the shadow region of the color gamut. This requires that the profile-making software have the data fitting and color space modeling capabilities.


TAGA JOURNAL VOL. 4

218


Measurement noise is proportional to the spatial non-uniformity of the printing device. Measurement noise from the randomized target layout may be smoothed out by profile-making software. So, printer profile accuracy is a combination of target layout and data smoothing ability of the profile-making software. Back to the world of 8.5 by 11. When reducing the IT8.7/4 profiling target from 6mm to 4mm, there are colorimetric differences between the two visual targets, i.e., there are more larger ∆E values associated with the reduced patch size. But this effect did not repeat in the random targets with different patch sizes. In this case, there were no significant colorimetric errors caused by the reduced patch size. For all practical purposes, there is no penalty to use the random target at the 4mm reduced patch size.

Acknowledgments The author wishes to thank Mr. Steve Upton of CHROMIX who shared many insights and new features in the ColorThink 3.0 Pro software. He wants to recognize his graduate student assistant, Dimitrios Ploumidis, for providing the laboratory support. He is also grateful to his RIT colleagues, Franz Sigg and Edline Chun; and his industry colleagues, David McDowell, Gary Field, and William Birkett, for reviewing the paper.

References 1 2 3 4

5 6 7 8 9

CGATS IT8.7/4, Graphic technology – Input Data for Characterization of 4-Color Process Printing – Expanded Data Set, 2005 Chung R., The Effect of Profiling Target Variations on Colorimetric Accuracy of Printer Profiles, Test Targets 6.0, pp. 19-24, 2006 Sigg F., Spatial Uniformity of Offset Printing, TAGA Proceedings, pp 649 – 658, 2007 Chung R. and Shimamura Y., Conducting a Press Run Analysis, Proceedings of the 28th IARIGAI Research Conference, Advances in Color Reproduction, GATF, pp. 333-345, 2001 CGATS, Recommended Industry Practice, Color characterization data set development — Analysis and reporting, 2006 Chung R., Gravure Research Agenda: Achieving Repeatable Color in Packaging Printing, Gravure, pp. 44-49, 2006 Spooner D.L., Evaluation of a Method for Correcting for Measurement Errors Caused by Adjacent Colors, TAGA Proceedings, pp. 416-427, 2002 ISO 5/4, Photography - Density Measurements - Part 4: Geometric Conditions for Reflectance Density, International Organization for Standardization, 1995 Sharma A., WMU Profiling Review, Western Michigan University, 2005

TAGA JOURNAL VOL. 4


E. HREHOROVA, M. REBROS, A. PEKAROVICOVA, P.D. FLEMING, V.N. BLIZNYUK

219

Characterization of Conductive Polymer Inks based on PEDOT:PSS Erika Hrehorova, Marian Rebros, Alexandra Pekarovicova, Paul D. Fleming, and Valery N. Bliznyuk Department of Paper Engineering Chemical Engineering and Imaging Western Michigan University 4601 Campus Dr. A-217 Parkview Campus Kalamazoo, MI 49008-5462 Email: [email protected] [email protected] [email protected] [email protected] [email protected]

Abstract The main driving force for implementation of printing in manufacture of flexible electronics is the possibility to reduce the cost by high speed R2R processing at ambient conditions. This work focuses on characterization of inks that can be used to print simple components or layers for various applications in electronics. More specifically, inks containing conductive polymer, poly(3,4-ethylenedioxy-thiophene)-poly(styrene sulfonate), known as PEDOT: PSS, were tested. In order to deposit smooth and uniform functional layers, it is important to optimize ink spreading and leveling on the substrate. PEDOT:PSS is commercially available as an aqueous dispersion with high surface tension (71 mN/m), which leads to poor ink spreading and substrate wetting. In this work, addition of alcohols and surfactants was used to lower the surface tension of polymer ink and the effect of concentration on dynamic and static surface tension was studied. Dynamics of ink spreading were also tested using dynamic contact angle measurements. It is shown that ink containing primary alcohol wets the surface more readily than surfactant containing systems. Addition of alcohols and surfactant had also positive effect leveling and bulk conductivity of final inks. More uniform layers were produced when using inks with lower surface tension. Increased bulk conductivity by addition of ethylene glycol is believed to be a result of conformational change and increased interaction between polymer chains. Stronger interactions between polymer chains were also confirmed by rheological measurements. Keywords: conductive polymers, surface tension, wetting behavior, topography, rheology

1

Introduction

Nowadays, more and more electronics manufacturers are embracing printing technologies as high-potential manufacturing methods for mainstream electronic components. However, to fully utilize the benefits of printed electronics, manufactures need advanced materials that are well suited for specific electronic applications and also printing systems, and are available in commercial quantities. Functional materials needed for printing of electronic components include i) conductors, ii) semiconductors and iii) dielectrics.


TAGA JOURNAL VOL.4

220

CHARACTERIZATION OF CONDUCTIVE POLYMER INKS ….

This work considers the properties of conductive inks that can be used in various printed electronics applications. Ink conductivity can be achieved by different mechanisms, such as incorporating metallic or other conductive particles into a non-conducting polymer matrix, or by using polymers that exhibit electrical conductivity in a suitable solvent. Among conductive polymers, solvent based (xylene, toluene) polyaniline inks [1, 2] and water-based poly (3,4ethylenedioxy-thiophene)-poly(styrene sulfonate) (PEDOT:PSS) inks [3, 4] are widely studied for applications in organic electronics. Several different printing methods are employed in these studies, such as flexography, offset lithography and gravure printing, as well as ink-jet printing. The present work focuses on characterization of PEDOT:PSS based polymer inks. PEDOT: PSS is commercially available as a water-soluble polyelectrolyte system with good film-forming properties, high visible light transmittance, and excellent stability [5]. Some applications of PEDOT:PSS include antistatic coatings, conductive layers in organic light emitting diodes (OLEDs), capacitors and thin film transistors [6]. A PEDOT:PSS complex is prepared by oxidative polymerization of ethylenedioxythiophene (EDOT) in aqueous dispersion using sodium peroxodisufate as the oxidant. A template polymer (usually polystyrene sulfonic acid – PSS) is present during the polymerization [5]. The PSS in the resulting complex acts as a source for the charge balancing counter ion. Moreover, it keeps the PEDOT chains dispersed in water, forming stable, easy to process, deep blue microdispersions [6]. The chemical structure of PEDOT: PSS is shown in Figure 1. As already mentioned, PEDOT:PSS is commercially available as aqueous dispersions. The water-based nature of such polymer systems gives rise to the issues of substrate wetting and ink spreading. Water has a high surface tension and thus water based inks are very often formulated with alcoholic co-solvents and/or surfactants in order to lower surface tension for printing. The addition of alcohols lowers the surface tension monotonically with increasing concentration, due to a preferential adsorption of the organic molecule at the liquid-air interface. Surfactants, however, quickly reduce the surface tension at very low concentrations up to the critical micelle concentration (CMC), due to a strong adsorption of the surfactant at the liquid-air surface. At concentrations higher than the CMC, the surface tension is practically constant, because any additional amount of surfactant will form micelles in bulk [7]. O

O

O

O

O S

S

S

S

S O

O

O

SO3H

SO3H

SO3H

SO3H

SO3

O

O

_

_ SO3H

+

S

+

O

O

SO3H SO3

Figure 1: Chemical structure of poly(3,4-ethylenedioxy-thiophene)-poly(styrene sulfonate) complex (PEDOT:PSS) Although the static surface tension is widely used in the printing industry to predict wetting behavior of printing inks, especially water based printing inks, it is also important to characterize interfacial surface tension under dynamic press conditions, where the ink is under constant compositional change. A new liquid-air interface is created characteristically in the order of milliseconds. Dynamic surface tension of inks containing surfactants and other polymers is determined by diffusion, adsorption and desorption processes and it is typically higher than the equilibrium (static) surface tension.

TAGA JOURNAL VOL.4



221

Analysis of wetting behavior of different functional fluids and their interaction with polymeric substrates is an important subject in order to develop and optimize various materials for electronics manufacture. The quality of the interface between functional layers in an electronic device is crucial for its performance [8]. In general, the wetting process is reflected by the contact angle, defined as the angle that a liquid makes with a solid surface. The equilibrium relation of a three phase system can be described by the Young – Dupre equation [9, 10]:

γ lv cos Θ = γ sv − γ sl ,

(1)

where Θ is the contact angle and γ is the surface/interfacial tension at the liquid-vapor interface (lv), solid-vapor interface (sv) and solid-liquid interface (sl). As already discussed, there are several different ways to reduce surface tension of water based systems. With PEDOT:PSS dispersions, addition of secondary alcohols has yet another positive effect on the resulting films. It has been reported that electrical conductivity can be enhanced by addition of different organic compounds. The conductivity improvement is strongly dependent on the chemical structure of the compound. Among the alcohols, ethylene glycol and glycerol were found to be the most efficient [11]. Enhancement of conductivity is believed to be a result of an increased interchain interaction caused by conformational change of the PEDOT chains from the coil structure into expanded-coil or linear structures [12]. When formulating conductive polymer inks, it is necessary to consider the method of printing and desired application, because requirements for ink properties can greatly vary for different printing processes. In addition, it is important to understand the effect of individual components employed in ink formulation on electrical properties of printed layers. Additives typically used for PEDOT-PSS inks or coatings include co-solvents and surfactants, bonding and cross linking agents, adhesion promoters and additives for conductivity improvement [13]. This work studies the effects of ink additives such as co-solvents and surfactants on ink properties in both fluid and solid form. More specifically, we studied the effect of addition of ethylene glycol, ethanol and surfactant on dynamic and static surface tension of PEDOT: PSS dispersions. The dynamic contact angle of the resulting ink was measured in order to examine the wetting behavior on poly(ethylene terephthalate) (PET) substrate before and after corona treatment. Surface topography and conductivity of resulting films on glass substrates as well as on PET were also studied. The effect of solvents, and additives on rheological behavior of PEDOT:PSS was examined.

2

Materials and Experimental Procedures

Materials The conductive polymer (Baytron® P) dispersion was obtained from H.C.Starck GmbH & Co, which contains 1.2-1.4% of PEDOT:PSS in water. Three different types of PEDOT: PSS based inks were prepared. Other materials, such as ethylene glycol, ethanol and TWEEN80 (nonionic surfactant) were purchased from Sigma Aldrich. Ethylene glycol was used in the formulation of PEDOT:PSS based inks to enhance conductivity. Ethyl alcohol and TWEEN80 were used to decrease surface tension of the PEDOT:PSS dispersion. Table 1 shows the tested ink compositions and ID’s that will be used throughout this article. 2.1


TAGA JOURNAL VOL.4

222


Table 1: Composition of tested PEDOT: PSS based inks Ink ID

Ink Composition

PEDOT:PSS

Pure PEDOT:PSS (Baytron® P)

EG-PEDOT:PSS

Ethylene Glycol in PEDOT:PSS (50% v/v)

EtOH-EG-PEDOT:PSS

Ethanol in EG-PEDOT:PSS (25% v/v)

TWEEN80-EG-PEDOT:PSS

Surfactant Tween80 in EG-PEDOT:PSS (0.31 wt %)

2.2 Procedures A SensaDyne Tensiometer was used to measure dynamic surface tension of the inks during addition of ethylene glycol, ethanol and surfactant to the PEDOT:PSS dispersion. This test is using the maximum differential bubble pressure method [14] based on creation of air bubbles in the fluid, at the end of two orifices with different diameters. The differential pressure of the formed bubbles is measured and the surface tension of the liquid is directly proportional to the pressure difference. Measurement of the static surface tension of inks was done using the contact angle analyzer FTA200 from First Ten Angstroms. The values of ink surface tension were calculated from the pendant drop[15] shape of the ink. Wetting behavior of inks was tested by measuring dynamic contact angle using the FTA200. The values of contact angle were calculated from the sessile drop [16]. PET (DuPont) was used as a substrate for contact angle measurements. A Corona treatment (SOA, Inc.) was applied in order to increase surface energy of the PET substrate [17]. The rheological behavior of PEDOT:PSS based inks was studied using a TA AR 2000 Dynamic Stress Rheometer together with Rheology Advantage software. Concentric cylinder geometry was employed to measure the ink samples. In order to eliminate possible shear history effects from loading, each sample was allowed to equilibrate for 2 minutes. The geometry was maintained at a constant temperature using a circulating water bath (25 °C). A steady state flow test was performed in the range of shear rates from 0.0001 to 2000 s-1. During this test, a shear rate is applied and viscosity measured when the material reaches steady state flow. After the viscosity is measured, the shear rate is again increased and the process repeated yielding a viscosity flow curve. Measured flow curves were then fitted with appropriate flow models, with the help of Rheology Advantage Data Analysis software. In order to observe the changes in surface topography of PEDOT: PSS based films; the tested inks were solution casted onto glass slides and dried for 30 minutes at 120 °C. Solution casting is a simple drop casting, where a certain volume of liquid is dropped onto the substrate and let dry to produce solid polymer film. White Light Interferometry in the Vertical Scanning Mode (WYKO RST-Plus microscope) was used to study topography of the resulting films. White light interferometry is a non-contact method for optical surface profilometry of various surfaces [18]. In vertical scanning mode (VSI), the interferometric lens scans the surface at varying heights by vertical movement through the focus and captures interference data at fixed intervals. The interference signal for each point of the sample surface is recorded, providing information about the fringe modulation, which is consequently used to calculate the surface height profile [19]. Another method for surface topography characterization employed in this study was Atomic Force Microscopy (AFM). An Autoprobe CP machine (Thermomicroscopes, USA) operated in a tapping mode and with typical scan sizes from 30 x 30 μm2 down to 2 x 2 μm2 was used to study the morphological features at different levels of structural organization and depending on the

TAGA JOURNAL VOL.4



223

preparation conditions of the samples. Root mean square (RMS) roughness was then measured and used as a quanity characterizing perfectness of the polymer film on glass. The conductivity of PEDOT:PSS based inks casted on glass was measured using a Keithley 2400 multimeter in four-probe mode. Measured resistance, cross-sectional area, and length of tested film sample can be used to calculate resistivity, ρ (Ω-cm). The inverse of resistivity yields conductivity (S/cm).

3

Results and Discussion

3.1 Static and Dynamic Surface Tension Firstly, ethylene glycol was added to the PEDOT:PSS dispersion. This addition caused decrease of static surface tension of PEDOT:PSS ink from 70.7±0.7 to 59.3±0.2 mN/m as measured by the pendant drop method. However, the surface tension of EG-PEDOT:PSS measured under dynamic conditions is higher than the static (equilibrium) surface tension and depends strongly on time of interface existence - surface age. For shorter surface age, the alcohol molecules have less time to migrate onto the newly created interface and thus the surface tension is higher than that measured under equilibrium conditions (Figure 2). 90

80

75.5 70.6

70.0

Surface Tension [mN/m]

70

60 53.3 50

40

30

20

10

0

1.12 sec

1.52

1

1.86 sec

Infinite Time = static ST

Surface Age [s]

Figure 2: Dynamic surface tension of EG-PEDOT:PSS ink Figures 3 and 4 show the change in dynamic surface tension and bubble frequency of EGPEDOT: PSS ink during addition of ethanol and nonionic surfactant TWEEN80, respectively. It can be seen, that addition of alcohol into the system caused a gradual decrease in surface tension within the measured range of ethanol addition. It has been reported [20] that the surface tension decreases relatively slowly or is almost constant when the ethanol content in ethanol/water mixture exceeds 20 vol%. Therefore, and also due to low concentration of polymer, we have used only up to 20 vol% of ethanol in this work. In the case of the surfactant TWEEN80 (Figure 4), the initial drop in surface tension is more dramatic and further addition of surfactant caused a rather slow decrease in surface tension, indicating that the system is above the CMC of the tested surfactant at the measured bubble frequency. This conclusion is also confirmed by a steady bubble frequency observed.


TAGA JOURNAL VOL.4

80.00

1.2

75.00

1.0

70.00

0.8

65.00

0.6

60.00

0.4

55.00

0.2

50.00

BF [bubble/sec]



224

0.0 0

5

10

15

20

25

EtOH Addition [% v/v]

80.0

1.2

75.0

1.0

70.0

0.8

65.0

0.6

60.0

0.4

55.0

0.2

50.0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

BF [bubble/sec]


Figure 3: Surface tension and bubble frequency changes during addition of ethanol to EGPEDOT: PSS ink

0.0 0.35

Surfactant Conc. [wt %]

Figure 4: Surface tension and bubble frequency changes during addition of TWEEN80 to EG-PEDOT:PSS ink Static surface tensions of the tested inks are shown in Table 2. The “rule of thumb” in the printing industry is to have the surface tension of ink at least 10 mN/m lower than the surface energy of the substrate to be printed on. Typical values of static surface tension for water-based inks used in gravure or flexo printing are in the range of 28 - 45 mN/m [21]. The lowest surface tension was found for EtOH-EG-PEDOT:PSS ink. However, printing of such ink might be still problematic for some polymeric substrates with lower surface energy. Table 2: Static surface tension of tested PEDOT:PSS based inks Ink ID PEDOT:PSS EG-PEDOT:PSS EtOH-EG-PEDOT:PSS TWEEN80-EG-PEDOT:PSS

TAGA JOURNAL VOL.4

Static Surface Tension [mN/m] 70.7 ± 0.9 59.3 ± 0.2 37.4 ± 0.2 41.8 ± 0.1



225

3.2 Dynamic Contact Angle The dynamic contact angle was measured for all prepared inks on corona treated PET substrate. As expected, the lowest contact angle was found for the ink with the lowest surface tension (EtOH-EG-PEDOT:PSS). It can be seen from Figure 5, that the contact angle of pure PEDOT:PSS ink and inks containing only alcohols stabilize after a short time (around 1.5 sec), corresponding to initial spreading of the ink drops on the substrate. In the case of surfactantcontaining system, the contact angle did not reach a stable value, even after 30 seconds and continued to decrease. This behavior has been seen previously for dynamic contact angle of liquids containing surfactants [22]. Figure 6 shows the contact angle of TWEEN80-EGPEDOT:PSS ink on PET substrate after 2, 5 and 10 minutes of observation. 85 PEDOT:PSS

Contact Angle [deg]

75

EG-PEDOT:PSS

65 TWEEN80- EG-PEDOT:PSS

55 EtOH- EG-PEDOT:PSS

45 35 0

5

10

15 Time [s]

20

25

30

Figure 5: Dynamic contact angle of PEDOT:PSS inks on PET substrate

Figure 6: Contact angle of TWEEN80-EG-PEDOT:PSS ink on PET substrate During printing, however, there is only a very shot time available for ink to spread on the substrate before it goes into the drying station. Thus, the dynamic contact angle is more important than contact angle at equilibrium conditions. Therefore, it is reasonable to look at contact angle measurements only for short time periods (Figure 7). It can be seen, ethanol is more efficient than the surfactant (TWEEN80) at short time scale. This is valid for both static and dynamic conditions for the tested system.


TAGA JOURNAL VOL.4

226


85

Contact Angle [deg]

75 PEDOT:PSS

65

EG-PEDOT:PSS EtOH-EG_PEDOT-PSS

55

TWEEN80-EGHPEDOT-PSS

45

35 0

0.2

0.4 0.6 Time [s]

0.8

1

Figure 7: Dynamic contact angle of PEDOT:PSS based ink at short time scale 3.3 Surface Topography In order to avoid effects of the substrate and study only the effects of ink formulation of surface topography, conductive polymer films were drop casted on glass slides and dried at 120 °C for 30 minutes to assure complete removal of ethylene glycol from polymer films. It can be seen from Figure 8, that addition of alcohols into the PEDOT:PSS system significantly improves uniformity of the film surface at the millimeter scale (2.5x1.9 mm2) as measured by vertical scanning interferometry (VSI). This is due to lower surface tension of the alcohol-containing ink and thus improved wetting of the glass substrate. However, AFM scans made at micrometer scale (10x10 μm2) show smoother surface of PEDOT:PSS films. EtOH-EG-PEDOT:PSS film show appearance of some larger domains, which can be a result of conformational change of polymer chains and swelling of the PEDOT:PSS complex, indicating stronger interchain interactions caused by alcohol addition [12]. It was found that the RMS roughness of PEDOT:PSS film measured by VSI was reduced from 902 nm down to 67 nm by addition of ethylene glycol and ethanol. Simultaneously, the RMS roughness measured by AFM shows only 2.3 nm for PEDOT:PSS and 12.9 nm for EtOH-EG-PEDOT:PSS. A similar effect on RMS roughness measured by AFM was found for addition of glycerol into the PEDOT:PSS system [23].

TAGA JOURNAL VOL.4



a) PEDOT: PSS

b) PEDOT: PSS 10x10 μm2

c) EtOH-EG-PEDOT:PSS

d) EtOH-EG-PEDOT:PSS 10x10 μm2

227

Figure 8: Surface topography of PEDOT:PSS and EtOH-EG-PEDOT:PSS ink film on glass substrate as studied by a), c) VSI and b), d) AFM 3.4 Rheological Behavior Viscosity curves (viscosity vs. shear rate) or flow curves (shear stress vs. shear rate) can provide information about ink processing and performance and thus, they are very important to monitor. Low shear rates can be related to storage conditions of materials, such as sedimentation, phase separation and structure retention. High shear rates give information about performance on the press. Factors influencing ink flow characteristics typically include pigment content, particle shape and size distribution, binder system and other additives used in ink formulation. In the case of polymer solutions, flow properties depend on polymer concentration, temperature, molecular weight and polydispersity index. Polymer solutions of higher concentrations deviate from Newtonian behavior more than diluted solutions, due to higher number of chain entanglements per unit volume [24]. Figure 9 shows the viscosity curves of four different PEDOT:PSS based inks. The initial polymer solution (PEDOT:PSS) first shows an increase in viscosity with increasing shear rate. This can be due to orientation of polymer chains with applied shear and thus, increasing the polymer-polymer interaction up to the point of maximum viscosity, after which it slowly shearthins. Addition of EG to PEDOT:PSS solution lowers the concentration of the polymer, however, as already mentioned, it causes the polymer chains to expand, resulting in stronger interchain interactions, which led to increased viscosity at lower shear rates. Addition of surfactant caused even further viscosity increase at lower shear rates. The possible explanation for such behavior is that because the concentration of surfactant in the ink formulation was above the CMC (3 g/l), there is a possibility of formation of rod-like micelles of surfactant molecules [24] and generation of additional entanglements within the polymer system.


TAGA JOURNAL VOL.4

228

CHARACTERIZATION OF CONDUCTIVE POLYMER INKS …. 10.00

PEDOT:PSS EG-PEDOT:PSS EtOH-EG-PEDOT:PSS TWEEN80-EG-PEDOT:PSS

viscosity (Pa.s)

1.000

0.1000

0.01000 1.000E-3

0.01000

0.1000

1.000 10.00 shear rate (1/s)

100.0

1000

10000

Figure 9: Viscosity curves for different PEDOT:PSS based inks In order to quantitatively compare measured curves, mathematical model function fittings can be used to describe flow curves (shear stress, σ, vs. shear rate, γ) using only a few parameters. There are a lot of fitting models available. When selecting a flow model, it is important to consider whether the samples is (i) idealviscous (Newtonian), (ii) shear-thinning, (iii) shearthickening, (iv) without a yield point or (v) showing a yield point. For inks tested in this work, the Herschel-Bulkley model [25] was found to be the best fit (Equation 2). This model incorporates the elements of Newtonian, Power Law and Bingham models, such as yield stress (σy), consistency coefficient (K) and power law index (n). σ = σ y + K ×γ n (2) Figure 10 shows the flow curves of tested inks fitted with the Herschel-Bulkley model for low shear rates (0.001 - 1 Pa). At shear rates higher than 1 s-1, results show linear dependence of shear stress on shear rate with similar slope of around 0.8 and y intercept -0.9 for all inks. Resulting model parameters are presented in the Table 3. 100.0

shear stress (Pa)

10.00

PEDOT-PSS EG-PEDOT:PSS EtOH-EG-PEDOT:PSS TWEEN80-EG-PEDOT:PSS

1.000

0.1000

0.01000

1.000E-3

1.000E-4 1.000E-4

1.000E-3

0.01000

0.1000

1.000 10.00 shear rate (1/s)

100.0

1000

10000

Figure 10: Flow curves for different PEDOT:PSS based inks fitted with mathematical model according to Herschel-Bulkley

TAGA JOURNAL VOL.4



229

Table 3: Solid content and parameters of Herschel-Bulkley model for tested PEDOT:PSS based inks Sample ID PEDOT:PSS EG-PEDOT:PSS EtOH-EG-PEDOT:PSS TWEEN80-EG-PEDOT:PSS

Solid Content [%] 1.29 0.92 0.82 1.21

σy [Pa]

K [Pa.sn]

n

Standard Error

1.39x10-3 1.17x10-3 3.09x10-4 2.22x10-3

0.122 0.115 0.112 0.118

0.982 0.961 0.832 0.936

5.30 2.89 3.80 4.68

When using the Herschel-Bulkley model, the following applies [26]: • Newtonian fluids >> σy = 0 and n = 1 • Power law fluid >> σy = 0 and n ≠ 1 (n > 1 for shear-thickening and n < 1 for shearthinning fluids) • Bingham fluid >> σy ≠ 0 and n = 1 It can be seen from the Table 3 that all of the tested inks exhibit yield stress. However, low values indicate that only a small stress is required to induce the flow. Addition of EG caused only slight decrease in yield stress and consistency coefficient (also known as viscosity coefficient), even though the solid content decreased by almost 30%. On the other side, further decrease in % of solids by ethanol addition led to a decrease in yield stress of nearly one order of magnitude. Considering the power law index, the lowest value was calculated for EtOH-EGPEDOT:PSS, while the rest of tested inks show an n value close to 1. The model fitting software used in this work also calculates standard error of the fit as shown in the Table 3. A reasonable fit gives a value of less than about 20 [27]. 3.5 Conductivity The electrical conductivity of the PEDOT:PSS based inks was calculated from resistance measurements on films casted on a glass substrate. As it was previously reported, addition of ethylene glycol to PEDOT:PSS dispersion enhances conductivity of the resulting films up to 200 S/cm [12]. In our case, addition of 25 vol% of ethylene glycol resulted in conductivity increase from 5.3 to 92.1 S/cm. Addition of ethanol to EG-PEDOT:PSS caused a decrease in conductivity, which can be attributed to lower concentration of polymer in dispersion. On the other hand, the presence of surfactant TWEEN80 in EG-PEDOT:PSS has led to increased conductivity (Table 4). Conductivity results also correlate well with roughness of casted films. The highest roughness measured at millimeter scale was found for PEDOT:PSS and the lowest for TWEEN80-EG-PEDOT:PSS inks indicating that the surface uniformity is very important to the charge transport of conductive polymer layer. Table 4: Conductivity of tested PEDOT:PSS based inks Sample ID PEDOT:PSS EG-PEDOT:PSS EtOH-EG-PEDOT:PSS TWEEN80-EG-PEDOT:PSS

Conductivity [S/cm] 5.3 ± 0.1 92.1 ± 0.1 62.4 ± 0.2 115.6 ± 2.1

Conductive polymers continue to gain importance in many electronics applications, due to the possibility of processing from solution, which enables the use of printing technologies for their deposition and patterning. Properties of polymer inks need to be often optimized for printing in order to achieve desired properties of final prints. Additives used traditionally in graphic ink formulation can work in conductive ink formulation as well, although their effect on electrical performance is very often not known.


TAGA JOURNAL VOL.4

230


This work studied the effect of ink additives on processing and final properties of PEDOT:PSS based polymer inks. PEDOT:PSS is a conductive polymer widely studied for printed electronics applications. The surface tension of PEDOT:PSS aqueous solutions was reduced by addition of alcohols (ethylene glycol and ethanol) and non-ionic surfactant. It was found that addition of ethylene glycol to PEDOT:PSS solution lowers the static surface tension by 11 mN/m. Further decrease can be achieved by addition of ethanol or surfactants, whereas ethanol showed higher effectiveness in lowering both, dynamic and static ink surface tension. This was also confirmed by dynamic contact angle measurements, where the surfactant-containing ink showed higher contact angle and slower rate on ink spreading than the ink containing ethanol. Efficient ink spreading and leveling is crucial for smooth and uniform deposition of PEDOT:PSS layers. Uniformity and smoothness of functional layers for printed electronic devices is even more important than for graphic inks. In printed electronics, non-uniform film can lead to lower device performance or even short circuit. Surface topography studies showed that addition of alcohols into PEDOT:PSS solution helped in improving the resulting film roughness on a millimeter scale, however the opposite effect was detected at the micrometer range due to swelling of PEDOT:PSS molecules. Conductivity of PEDOT:PSS films was improved by addition of ethylene glycol. Rheological tests were performed by using rotational viscometry to obtain viscosity curves for tested inks, which were further analyzed by using Herschel-Bulkley flow model. Higher values of yield stress for ethylene glycol containing inks confirmed increased interaction between PEDOT:PSS chains.

References Makela T. et al., Utilizing Roll-to-Roll Techniques for Manufacturing Source-Drain Electrodes for All-polymer Transistors, Synthetic Metals, vol. 153, pp. 285. – 288, 2005 2 Hrehorova E.L., Wood K., Pekarovic J., Pekarovicova A., Fleming P.D. and Bliznyuk V., The Properties of Conducting Polymers and Substrates for Printed Electronics, Proceedings of IS&T Digital Fabrication DF01, pp. 197 – 202, 2005 3 Fenoll M., Catusse R. and Rousse E., Gravure Printing: Material Characterization for All-Organic Capacitor, IARIGAI’s 32nd International Research Conference on Digitalization and Print Media, pp. 83 – 92, 2005 4 Zielke D, Hübler A.C., Hahn U., Brandt N., Bartzsch M., Fügmann U. and Fischer T., Polymer –based organic field-effect transistor using offset printed source and drain structures, Appl. Phys. Lett., vol. 87, 123508, 2005 5 Groenendaal L., Jonas F., Freitag D., Pielartzik H. and Reynolds J.R., Poly (3,4ethylenedioxythiophene) and Its Derivatives: Past, Present, and Future, Adv. Mater.,vol. 12, no. 7., pp. 481-494, 2000 6 Kirchmeyer S. et.al., Scientific importance, properties and growing applications of poly(3, 4-ethylenedioxythiophene), J. Mater. Chem., vol. 15, pp. 2077 – 2088, 2005 7 Huibers P.D.T., Lobanov V.S., Katritzky A.R., Shah D.O. and Karelson M., Prediction of Critical Micelle Concentration Using a Quantitative Structure-Property Relationship Approach. 1. Nonionic Surfactants, Langmuir, vol. 12, pp. 1462 – 1470, 1996 8 Fahlman M. and Salaneck W.R., Surfaces and interfaces in polymer-based electronics, Surface Science, vol. 500, pp. 904 – 922, 2002 9 Young T., Miscellaneous Works, G. Peacock, ed., (Murray, London), Vol. I, 418 pp, 1855 10 Dupre A., Theoris Mecanique de la Chaleur, (Paris), 368 pp, 1869 11 Ashizawa S., Horikawa R. and Okuzaki H., Effect of Solvent on Carrier Transport in Poly(3,4-ethylenedioxithiophene)/poly(4-styrenesulfonate), Synthetic Metals, vol. 153, pp.5-8, 2005 12 Ouyang J. et al., On the mechanism of conductivity enhancement in poly (3,4ethylene-dioxythiophene):poly(styrene sulfonate) film through solvent treatment, Polymer, vol. 45, pp. 8443 – 8450, 2004

1

TAGA JOURNAL VOL.4



231

13 Kirchmeyer S., et al., Intrinsically Conductive Polymers, Kunststoffe, 10, pp. 202– 208, 2005 14 Klaus J.P. et al., U.S. Patent 4,416,148, 1983 15 Andreas J.M., Hauser E.A. and Tucker W.B., Boundary Tension By Pendant Drops, J. Phys. Chem., vol. 42, pp. 1001, 1938 16 Nutting G.C. and Long F.A., The Change with Time of the Surface Tension of Sodium Laurate Solutions, Journal of the American Chemical Society, vol. 63, pp. 84 – 88, 1941 17 Thomas P.E. and Raley G.E., U.S. Patent 4,351,784, 1982 18 Caber P.J., Interferometric profiler for rough surfaces, Applied Optics, vol. 32, no. 19, pp. 3438 – 3441, 1993 19 Veeco Metrology Group, WYKO Profilers Technical Reference Manual, Version 2.2.1, 1999 20 Sharma M.K., Surface Phenomena in Coatings and Printing Technology, in Surface Phenomena and Fine Particles in Water-Based Coatings and Printing Technology, Plenum Press, NY and London, 1 pp, 1989 21 Podhajny R.M., Surface Phenomena in Coatings and Printing Technology, in Surface Phenomena and Fine Particles in Water-Based Coatings and Printing Technology, Plenum Press, NY and London, 41 pp, 1989 22 Rebros M., Fleming P.D. and Joyce M.K., UV-Inks, Substrates and Wetting, Proceedings of the 2006 TAPPI Coating & Graphic Arts Conference, Atlanta, 2006 23 Snaith H.J., Kenrick H., Chiesa M. and Friend R.H., Morphological and electronic consequences of modifications to the polymer anode ‘PEDOT:PSS’, Polymer, vol. 46, pp. 2573 – 2578, 2005 24 Gupta R.K. Polymer and Composite Rheology, Marcel Dekker, Inc., New York, 390 pp, 2000 25 Mezger T.G., The Rheology Handbook, 2nd Ed., Vincentz Network, Hannover, pp. 299, 2006 26 Hemphill T, Campos W. and Pilehvari A., Yield-Power Law Model More Accurately Predicts Mud Rheology, Oil & Gas Journal, vol. 91, no. 34, pp. 45-50, 1993 27 TA Instruments, Rheology Advantage Data Analysis: Rheometrics Series Getting Started Guide, Revision E, pp. 60, 2004


TAGA JOURNAL VOL.4

TAGA JOURNAL VOL. 4
