Proceedings

8th CONFERENCE ON INFORMATION

AND GRAPHIC ARTS TECHNOLOGY

Ljubljana, Slovenia, 7−8 June 2018

Hosted by:
Proceedings
8th CONFERENCE ON INFORMATION
AND GRAPHIC ARTS TECHNOLOGY

Ljubljana, Slovenia, 7−8 June 2018

Editors: Raša Urbas, Nace Pušnik

Technical editor: Nace Pušnik

Rewiever: Raša Urbas

Authors accept responsibility for the whole paper.

Design, text layout and cover design: Nace Pušnik

Publisher: University of Ljubljana

Faculty of Natural Sciences and Engineering
Department of Textiles, Graphic Arts and Design
Chair of Information and Graphic Arts Technology

In the book of abstracts and proceedings the typeface family Mantika is used,
which was designed by Jürgen Weltin.

Electronic edition: 1st e-ed.

URL:
https://www.ntf.uni-lj.si/igt/wp-content/uploads/sites/8/2018/06/8_CIGT_Proceedings.pdf

Kataložni zapis o publikaciji (CIP) pripravili v Narodni in univerzitetni knjižnici v Ljubljani

COBISS.SI-ID=295289344
ISBN 978-961-6900-24-9 (pdf)
Organizer
University of Ljubljana,
Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Chair of Information and Graphic Arts Technology

Under the Auspices

International Circle of Educational Institutes
of Graphic-Media Technologies and Management

Sponsorship
University of Ljubljana,
Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Chair of Information and Graphic Arts Technology

Organizing committee
Conference director
• Raša Urbas − University of Ljubljana, Slovenia
Members
• Jure Ahtik − University of Ljubljana, Slovenia
• Aleš Hladnik − University of Ljubljana, Slovenia
• Barbara Luštek Preskar − University of Ljubljana, Slovenia
• Klemen Možina − University of Ljubljana, Slovenia
• Deja Muck − University of Ljubljana, Sloveni
• Nace Pušnik − University of Ljubljana, Slovenia

Programme and review committee
Chairman
• Aleš Hladnik − University of Ljubljana, Slovenia
Members
• Bojana Boh Podgornik − University of Ljubljana, Slovenia
• Luk Bouters − Ghent University, Belgium
• Sabina Bračko − University of Ljubljana, Slovenia
• Helena Grabijelčič − University of Ljubljana, Slovenia
• Diana Gregor Svetec − University of Ljubljana, Slovenia
• Nemanja Kašiković − University of Novi Sad, Serbia
• Branka Lozo − University of Zagreb, Croatia
• Sanja Mahovič Poljaček − University of Zagreb, Croatia
• Klementina Možina − University of Ljubljana, Slovenia
• Deja Muck − University of Ljubljana, Slovenia
• Živko Pavlović − University of Novi Sad, Serbia
• Nace Pušnik − University of Ljubljana, Slovenia
• Dimitrina Todorova − University of Chemical Technology and Metallurgy, Bolgaria
• Raša Urbas − University of Ljubljana, Slovenia
• Urška Vrabič Brodnjak − University of Ljubljana, Slovenia
Sponsors
Table of contents
INVITED LECTURES

14 BIOMASS UTILIZATION IN POLYMER SYNTHESIS AND NANOCELLULOSE PRODUCTION

Matjaž KUNAVER 1, Ema ŽAGAR1 & Raša URBAS2
1
National Institute of Chemistry, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

21 COMPUTATION APPROACH FOR REALISATION OF CONTEXT-AWARE ROBOTS

Tomislav STIPANČIĆ1, Duška ROSENBERG2 & Bojan JEREBIĆ1
1
University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Croatia
2
University of London

27 3D PRINTING FOR DESIGN AND ART

Barbara WILKESMANN HOČEVAR 1, Deja MUCK 2
1
Chemets d.o.o., Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

ORAL PRESENTATIONS

32 BIOCOMPOSITE PLA FILAMENT FOR 3D PRINTING

Diana GREGOR-SVETEC, Urška STANKOVIČ ELESINI,
Mirjam LESKOVŠEK, Urška VRABIČ BRODNJAK & Raša URBAS
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

37 COLORIMETRIC CHANGES CAUSED BY UV VARNISHING

Ivana CAR, Igor MAJNARIĆ & Branka LOZO
University of Zagreb, Faculty of Graphic Arts, Croatia

42 COMPUTER GENERATED HANDWRITING SIMULATION

BASED ON ALTERNATION OF CHARACTERS
Mihael Aleksander MAHKOVIC, Klementina MOŽINA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

48 DEFINING THE OPTIMAL TRAINING SET SIZE FOR CAMERA-

BASED COLOUR MEASUREMENT OF GONIOCHROMATIC PRINTS:
A PRELIMINARY STUDY
Ivana TOMIĆ, Sandra DEDIJER, Dragoljub NOVAKOVIĆ & Ivan PINĆJER
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design, Serbia

53 DIFFERENCES BETWEEN DESIGNERS’ AND USERS’

PERCEPTION OF INSTRUCTIONS DESIGN
Dorotea KOVAČEVIĆ, Maja BROZOVIĆ
University of Zagreb, Faculty of Graphic Arts, Croatia

58 ELECTRON BEAM CURING OF PRINTING INKS IN PACKAGING INDUSTRY

Jakub GRIGER 1, Dávid ŠARISKÝ 1, Zuzana ŠTROMAJER 1,
Michal ORAVEC 2, Michal ČEPPAN 2 & Viera JANČOVIČOVÁ 2
1
Chemosvit Fólie, a.s., Slovakia
2
Slovak University of Technology in Bratislava, Faculty of Chemical and
Food Technology, Department of Graphic Arts Technology and Applied
Photochemistry, Slovakia

8th Conference on Information and Graphic Arts Technology 7

63 EQUIPPING THE ARS VIVA INSTITUTE WITH WEB COMMUNICATION CHANNELS
Anja ŠKERJANC 1, Jure VIDMAR 2, Primož PREVC 2 & Matija MAROLT 3
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
University of Ljubljana, Faculty of Electrical Engineering, Slovenia
3
University of Ljubljana, Faculty of Computer and Information Science,
Slovenia

69 EXTAND OF PAPER COATING CRACKING

Klemen MOŽINA 1, Jure AHTIK 1, Špela BEČAJ 1, Urška KRAJNC 1,
Matej PIVAR 1, David RAVNJAK 2 & Diana GREGOR SVETEC 1
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
B & B, Papirnica Vevče, Labels and Flexpack, Slovenia

74 FABRICATION OF CAPACITIVE SLIDERS

Tanja PLEŠA 1, Urška KAVČIČ 1, Gregor LAVRIČ 1 & Deja MUCK 2
1
Pulp and Paper Institute Ljubljana, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

80 FORMATION AND CHARACTERISATION OF SHELLAC FILMS

Urška VRABIČ BRODNJAK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

84 MONTE CARLO DENOISING OF COMPUTER GENERATED IMAGES

BY NON-LOCAL MEANS FILTERING
Helena GABRIJELČIČ TOMC, Aleš HLADNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

90 RECONGNITION OF WORDS WITH JUMBLED LETTERS

Nace PUŠNIK 1, Primož WEINGERL 1, Vladimir DIMOVSKI2 & Irma PUŠKAREVIĆ 2
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design, Serbia

95 THE DEVELOPMENT OF THE GENERATIVE BRANDS’ CATEGORIZATION

Anže KAVČIČ, Helena GABRIJELČIČ TOMC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

101 THE EFFECTS OF TYPEFACE AND IMAGE COMPLEXITY ON CONSUMER' VISUAL

ATTENTION AND ATTITUDINAL RESPONSES IN ADVERTISING
Irma PUŠKAREVIĆ 1, Uroš NEDELJKOVIĆ 1, Ivan PINĆJER 1,
Gregor FRANKEN2 & Nace PUŠNIK2
1
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design, Serbia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

107 THE INFLUENCE OF BACTERIAL NANOCELLULOSE ADDITION

ON MECHANICAL PROPERTIES OF PAPER
Gregor LAVRIČ, Igor KARLOVITS, Tanja PLEŠA & Mija SEŽUN
Pulp and Paper Institute, Ljubljana, Slovenia

8th Conference on Information and Graphic Arts Technology 8

112 THE POSSIBILITY OF USING ARDUINO BASED WATER LEVEL SENSOR
FOR TISSUE PAPER ABSORPTION RATE AND CAPACITY MEASUREMENT
Igor KARLOVITS, Urška KAVČIČ, Tanja PLEŠA & Mija SEŽUN
Pulp and Paper Institute, Ljubljana, Slovenia

117 USE OF IMAGE SEGMENTATION FOR WOVEN FABRICS’ POROSITY

AND ALPHA MAP DETERMINATION
Tanja Nuša KOČEVAR, Živa ZUPIN
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

123 USER EXPERIENCE OF WAIT-ANIMATION PROGRESS INDICATORS

Jesenka PIBERNIK, Jurica DOLIĆ, Roberta STANIĆ & Lidija MANDIĆ
University of Zagreb, Faculty of Graphic Arts, Croatia

POSTER PRESENTATIONS

129 AVAILABILITY OF E-TEXTBOOKS ON THE SLOVENIAN MARKET

Grega TOMAŽIN 1, Urška STANKOVIČ ELESINI 2
1
University of Ljubljana, Academy of Fine Arts and Design, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

135 BLUEPRINTING - AN ANCIENT BUT STILL LIVING WAY OF PRINTING TEXTILES

Veronika KRAVCOVÁ 1, Ondrej PANAK1, Petra FORTE TAVČER 2
1
University of Pardubice, Faculty of Chemical Technology,
Department of Graphic Arts and Photophysics, Czech Republic
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

139 COLOUR AND ODOUR PERCEPTION AT COSMETIC PACKAGING

Kaja JOKIĆ, Barbara BLAZNIK & Sabina BRAČKO
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

145 COLOUR OF TASTE

Gregor FRANKEN
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

151 COMPARISON OF DIGITAL AND TRANSFER PRINTING OF COTTON FABRIC

Michaela ČESÁKOVÁ 1, Ondrej PANAK1, Petra Forte TAVČER 2
1
University of Pardubice, Faculty of Chemical Technology,
Department of Graphic Arts and Photophysics, Czech Republic
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

158 COMPARISON OF SCREEN PRINTED AND ARDUINO-BASED SENSORS

FOR ALCOHOL DETECTION
Urška KAVČIČ 1, Igor KARLOVITS 1 & Matej PIVAR 2
1
Pulp and Paper Institute, Ljubljana, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

163 CREATION OF INTERACTIVE BOARD FOR LEARNING ABOUT TYPOGRAPHY

Tevž ZRINSKI, Klementina MOŽINA & Dejana JAVORŠEK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

8th Conference on Information and Graphic Arts Technology 9

168 DESIGN OF A GIFT CERTIFICATE USING PROTECTIVE ELEMENTS IN A STANDARD TRIAD
Nadezhda SAVCHUK, Olga NOVOSELSKAYA
Byelarussian State Technological University,
Informatics and Web-Design Department, Belarus

174 DEVELOPMENT OF DISPERSION BARRIER COATING FOR GREASE,

MINERAL OIL AND WATER VAPOUR
Jelena VASILJEVIĆ 1, Barbara GOLJA 1, Ivan JERMAN 2, Ervin ŠEST 2, Matic ŠOBAK2,
David RAVNJAK 4, Mateja KERT 1, Marija GORJANC 1, Brigita TOMŠIČ 1, Tea KAPUN 3,
Diana GREGOR-SVETEC 1 & Barbara SIMONČIČ 1
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
National Institute of Chemistry, Slovenia
3
Pulp and Paper Institute, Slovenia
4
Paper Mill Vevče d.o.o., Slovenia

180 ENZYMATIC DEINKING OF FLEXOGRAPHIC PRINTED PAPER

Mija SEŽUN 1, Gregor LAVRIČ 1, Gašper DOKL 2 & Matej PIVAR 3
1
Pulp and Paper Institute, Ljubljana, Slovenia
2
University of Ljubljana, Biotechnical Faculty, Slovenia
3
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

184 EVOLUTION OF NATURAL PATTERNS FROM RANDOM FIELDS

Lovrenc ŠVEGL 1, Igor GRABEC 2
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
University of Ljubljana, Faculty of Mechanical Engineering, Slovenia

189 EYE TRACKING ANALYSIS OF INTERNAL FACIAL FEATURES

IN OBSERVATION AND RECOGNITION PROCESS
Andrej ISKRA, Helena GABRIJELČIČ TOMC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

195 HOW COLOR SHAPES BRAND IDENTITY

Urška STANKOVIČ ELESINI 1, Karla MRKOBRADA 1 & Jani TOROŠ 2
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
Innovative Management Institute Ljubljana, Slovenia

201 IMPACT OF THE AGE OF THE PIGMENT PRINTING PASTE ON THE PRINT QUALITY
Petra FORTE TAVČER
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

207 INFLUENCE OF OPTICAL PAPER PROPERTIES OF RECYCLED AND NON-RECYCLED

PAPERS FOR DIGITAL PRINTING ON COLOUR REPRODUCTION
Ivana JURIČ, Dragoljub NOVAKOVIĆ, Nemanja KAŠIKOVIĆ,
Sandra DEDIJER & Neda MILIĆ
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design, Serbia

214 POLYLACTIC ACID FIBERS CONTAINING FLUORESCENT PIGMENTS

Viera JANČOVIČOVÁ, Barbora ĎURICOVÁ, Veronika HRABOVSKÁ & Katarína HABEROVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Department of Graphic Arts Technology and Applied Photochemistry, Slovakia

8th Conference on Information and Graphic Arts Technology 10

219 PREVENTION OF STARCH DEGRADATION IN PAPER MAKING
PROCESS FOR RECYCLED PACKAGING PAPER
Dimitrina TODOROVA, Ventsislav BLYAHOVSKI & Spas LADZHOV
University of Chemical Technology and Metallurgy, Bulgaria

225 PRINTING BRAILLE WITH SPECIAL EXPANDABLE SCREEN PRINTING INKS

Karin KUKEC, Raša URBAS
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

230 SCIENTIFIC ILLUSTRATIVE DESIGNS IN GRAPHIC PRODUCTS

Tadej PREZELJ, Urška VRABIČ BRODNJAK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

233 SELECTING THE TRADEMARK COLOUR DEPENDING

ON THE COMPANY’S BUSINESS AREA
Ivan PINĆJER, Ivana TOMIĆ, Irma PUŠKAREVIĆ & Gojko VLADIĆ
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design, Serbia

238 SELF-FOLDABLE KNITTED BAGS FOR HOUSEHOLD STORAGE

OF BAKERY PRODUCTS BASED ON BIOMIMETICS
Darja RANT, Alenka PAVKO-ČUDEN
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

243 STUDY OF ACCELERATED AGING OF BLACK INKJET PRINTS

Michal ORAVEC, Lukáš GÁL, Michal ČEPPAN & Patrícia PARAČKOVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Department of Graphic Arts Technology and Applied Photochemistry, Slovakia

248 STUDY OF CHANGES OF WRITING INKS ON THE PAPER DOCUMENTS

Silva STAŠKOVÁ, Milena REHÁKOVÁ & Aleš HÁZ
Slovak University of Technology in Bratislava, Faculty of Chemical
and Food Technology, Institute of Natural and Synthetic Polymers,
Slovak Republic

252 SYNTHESIS, COATING AND EVALUATION OF ANTIMICROBIAL

MICROCAPSULES ON PAPER
Bojana BOH PODGORNIK 1, Boštjan ŠUMIGA 2, Barbara GOLJA 1,
Matej ŠUŠTARŠIČ 2, Barbara ŠUMIGA 2 & David RAVNJAK 3
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
Pulp and Paper Institute, Ljubljana, Slovenia
3
Papirnica Vevče d. o. o., Slovenia

258 TLC AND GC/MS ANALYSIS OF INK JET PRINTING INKS

Barbara BLAZNIK 1, Franci KOVAČ 2 & Sabina BRAČKO 1
1
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia
2
University of Ljubljana, Faculty of Chemistry and Chemical Technology,
Slovenia

261 TYPEFACE AND ECOLOGICAL DESIGN

Klementina MOŽINA, Alenka TOMAŠIČ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

8th Conference on Information and Graphic Arts Technology 11

267 USE OF BIOCOMPOSITE FILAMENTS FOR 3D PRINTING
Maruša ROPRET, Helena GABRIJELČIČ TOMC & Deja MUCK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

USING UAV RGB CAMERA AS A COLOUR MEASURING DEVICE

273 Andrej UČAKAR, Aleš HLADNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Slovenia

8th Conference on Information and Graphic Arts Technology 12

 Invited
lectures
 Invited Lectures

BIOMASS UTILIZATION IN POLYMER SYNTHESIS AND NANOCELLULOSE PRODUCTION

Matjaž KUNAVER 1, Ema ŽAGAR 1, Raša URBAS 2

1
National Institute of Chemistry, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia

ABSTRACT: Biobased platform chemicals can be provided through lignocellulosic conversion in biorefineries.
Acid catalyzed glycolysis of biomass wastes is such example. The liquefaction reaction was the basic platform
for production of polymers, adhesives, energy and nanocellulose. We used the liquefied wood as a polyol in
the polyester synthesis, which was applied in polyurethane foam production. Liquefied biomass reacts with
thermosetting systems containing melamine–urea–formaldehyde resin and it was found that a 50% addition
of liquefied wood met the European standard quality demands for particle boards. Two fuels, namely liquefied
cotton fibres and liquefied biomass, were used in gas turbine. Stable combustion was achieved and experimen-
tal results indicate successful utilization of these biofuels in professional gas turbines. The same liquefaction
process was used for isolation of nanocrystalline cellulose (NCC) from biomass. This method represents a
novelty and is a model procedure for NCC isolation from different natural cellulosic sources with high yields
and with high crystallinity index.

Keywords: biomass, lignocellulosic sources, liquefaction, nanocrystalline cellulose, gas turbine.

1 INTRODUCTION

Biobased platform chemicals can be provided through lignocellulosic conversion in biorefineries. However,
to make such a process economical, the combination of high value products generation with bioenergy pro-
duction is essential. An overview of some applications and methods are presented in this contribution. Since
the petrochemicals may soon be limited, considerable attention has been given to the selective conversion of
biomass into chemicals such as 5-hydroxymethylfurfural and dimethylfuran by selective chemical catalysts.
The conversion of biomass by depolymerization and fermentation is presently the most potential approach.
Recent studies of acid catalyzed depolymerization with organic solvents have allowed to increase yields of
basic chemicals such as furfural, hydroxymethylfurfural and levulinic acids (Jasiukaitytė, 2009; Dutta, 2014).
Although biomass based lignocellulosic materials are complex, three major transformation pathways are more
commonly used:
• fast pyrolysis at high temperatures and subsequent conversion into the liquid products and the gaseous
products that are used for energy production;
• thermochemical conversion, at elevated temperatures and with the use of different liquefying agents and
catalysts. Here, solvolysis and depolymerization take place simultaneously with the liquid final product
being rich in hydroxyl groups;
• biochemical modifications (enzymatic conversion into ethanol and other depolymerization products).

1.1 Chemicals from biomass

Biomass is primarily composed of cellulose (40–50%), hemicellulose (25–35%) and lignin (15–20%). The com-
position varies considerably depending on types of biomass, part of the plant, etc. Cellulose is the major
structural component of the plant cell walls and is organized into fibrils, which are aligned parallel to each
other, surrounded by a matrix of hemicellulose and lignin.
A multitude of different types of biomass has been tested, for example: algae, cultivated crops, agricultural
waste, forest resources, etc. The term that is usually associated with such conversion is biorefinery. A variety
of technologies have been used, including fermentation, pyrolysis, thermochemical conversion, hydrothermal
liquefaction, catalytic conversion, etc. (Zhang, 2010).

1.1.1 Hemicelluloses
Hemicelluloses, hetero polymers that are present along with cellulose in plant cell walls contain xylose,
mannose, galactose rhamnose and arabinose. They can easily be isolated from the plants by different purifi-
cation methods using acids, organic solvents and alkaline reagents. Other methods include steam explosion,
ultrasonication, extrusion and microwave irradiation. Further conversion into hydroxymethyl furfural, furfural

8th Conference on Information and Graphic Arts Technology 14

 Invited Lectures

and even propionic acid can be achieved by reaction in a catalyzed ionic liquid, fermentation processes or
hydrothermal processes (Peng, 2012).

1.1.2 Lignin
Lignocellulosic biomass is mainly cell wall material and lignin is one of three principal components beside
cellulose and hemicelluloses. Lignin serves as a binding agent adding strength to the cell walls. It is a cross-
linked amorphous polymer synthesized from aromatic alcohol precursors, p-coumaryl, coniferyl and sinapyl
alcohols and has a very complex structure.
Lignin without chemical modification can be used as a filler, anti-oxidant, UV-stabilizer, surfactant, etc. The
chemical modification of lignin (Azadi, 2013; Khalil, 2012) has much better potential for multiple applications
and can be classified into three main categories:
• fragmentation;
• modification by creating new chemical active sites;
• chemical modification of hydroxyl groups.

1.1.3 Nanocellulose
Nanocellulose, the carbon neutral and sustainable material with outstanding mechanical properties, is one of
the most promising materials and has been a subject of recent interest. Many new nanocomposite materials
with attractive properties have been prepared by physical incorporation of NCC into a natural or synthetic
polymeric matrix and also holds promise in many different applications such as in nanopapers, coatings, ad-
hesives, optical sensors, biomedical scaffolds, filtration membranes, electronic devices, foams, aero gels, etc.,
which make NCC of significant interest from a scientific and industrial perspective. Major studies over the last
decades have been dedicated to the use of nanocellulose as a filler in nanocomposites to improve mechanical
and barrier properties. It can be easily chemically modified due to the abundance of -OH groups on the surface
of the nanocellulose through the acetylation reaction, TEMPO oxidation, silylation, polymer grafting, all with
the purpose of modifying the polarity of the nanocellulose surface and of improving the compatibility with
different polymer matrices (Habibi, 2014). Due to the nanocellulose high functionality and so many possible
chemical modifications the material applications for nanocellulose are almost limitless.

1.2 The liquefaction process

The liquefaction of wood and other lignocellulosic materials in the presence of phenol or polyhydroxy alco-
hols has been intensively studied by several authors and has been described in details (Yamada, 2001). Such
a liquefaction process is usually carried out at elevated temperature and in the presence of an acid catalyst.
The liquefied wood contains depolymerized products from the β 1-4 glucosidic bond cleavage of the cellulose
and hemicelluloses molecules as well as from low molar mass oligomers. The complex structure of the lignin
molecule is broken to smaller fragments. According to Lin et al., the dominant products were guaiacylglycer-
ol-α-phenyl-β-guaiacyl ethers, followed by guaiacol, triphenylethanes, diphenylmethanes, benzocyclobutanes
and phenylcoumaranes (Lin, 1995).
While a large number of papers have been published describing the liquefaction process, the characteriza-
tion of the product and the reaction pathways, little information is available on the application of the more
specialized techniques in biomass liquefaction. One of these is microwave driven wood liquefaction, where
microwave heating has been used in a very efficient and fast liquefaction of wood. The efficiency of liquefaction
and the time needed for the complete liquefaction was reduced significantly (Kržan, 2006).
A novel approach to very efficient energy input during the thermochemical conversion of lignocellulosic
biomass into liquefied depolymerized products is the use of ultrasound technologies (Kunaver, 2012). High
energy ultrasound has been used in liquefaction reactions with different wood waste materials. The most
common wood waste materials were chosen in order to prove the efficiency of the ultrasound process and to
establish a different way for recycling wood wastes, found in the municipal waste deposits in large quantities.
The reaction times were shortened up to nine times when using the ultrasound process with smaller residual
particles and with no influence on the hydroxyl number of the final products.

2 EXPERIMENTAL

2.1 Biomass liquefaction

The liquefaction of wood was carried out in a 1000 mL three-neck glass reactor, equipped with the mechanical
stirrer and condenser. The reactor was charged with 100 g of wood and 300 g of glycerol – diethylene glycol
mixture (4:1). 9 g of p-toluenesulfonic acid was added. The proportion of the constituents in the reaction

8th Conference on Information and Graphic Arts Technology 15

 Invited Lectures

mixture was chosen after the initial trials, with the goal to maximize the glycerol content as a by-product
of the biodiesel production. A certain proportion of glycerol was replaced with diethylene glycol to provide
the required degree of viscosity reduction for the final product. The mixture was heated for 3 hours at 180°C
while being constantly stirred. A sample was withdrawn from the reaction system periodically and immersed
in cold water to quench the reaction.
The extent of liquefaction was evaluated by determining the residue after the washing out the sample with
dioxane and water (4:1 v/v). The residue was dried in an oven at 105°C to constant weight. The conversion
yield was calculated as the weight percentage based on the starting wood material.

3 RESULTS & DISCUSSION

3.1 Polyesters

The hydroxyl value of the liquefied wood has been determined by several authors and is generally determined
to be between the equivalent of 200 and 500 mg KOH/g. The actual value depends on the liquefaction time.
This large number of hydroxyl group sites means that the products have applications in those areas in which
the hydroxyl component can be used in complex polyester synthesis. Kunaver et al. used the liquefied wood
as a polyol in the polyester synthesis (Kunaver, 2009). Saturated polyesters are defined as polyesters whose
components contain no polymerizable double bonds, the opposite of what occurs with alkyd resins and un-
saturated polyesters. Saturated polyesters are synthesized by the condensation of bifunctional or polyfunc-
tional monomers containing hydroxyl or carboxylic groups. The hydroxyl groups in the current investigation
originate from liquefied wood. The polycondensation process takes place at temperatures of 150 to 260°C.
Water as the main reaction byproduct, is removed from the system by the creation of a slight vacuum, gas
stream condensation or an azeotropic process.
The products were characterized using FTIR, GPC/SEC and viscosity measurements. The polyesters have hy-
droxyl values that were reduced due to esterification, from 1043 mg KOH/g of the liquefied wood to 400 to
800 mg KOH/g. 22 to 23% of the polyhydroxyl alcohols in the polyester formulations were replaced by wood
derivatives. A rise of the average molar mass was achieved together with a reduction of the hydroxyl group
content. Such modification of the polyester reactivity and complexity is favorable for further utilization in
polyurethane synthesis.

3.2 Adhesives

Hydroxyl groups in liquefied lignocellulosics can also react with different reactive sites in thermosetting sys-
tems as well as in two component systems.
Kunaver et al. applied a mixture of liquefied wood and melamine – urea – formaldehyde resin as an adhesive
in wood particle board production (Kunaver, 2010). The properties of laboratory prepared particle boards are
presented in Table 1, together with the values required by European standards.

Table 1. Properties of laboratory prepared particle boards (simulation of the industrial process).
Property (corresponding EN standard) Required values Achieved values
Board thickness (EN 324) [mm] 16.00 ± 0.30 16.10
Density EN 323 [g/cm3] ∕ 0.72
Internal bond strength (EN 319) [N/mm2] above 0.35 0.94
Surface soundness (EN 311) [N/mm2] above 0.80 1.96
Bending strength (EN 310) [N/mm2] above 13.00 21.10
Swelling in Thickness (EN 317) [%] below 15.00 8.20
Formaldehyde release Perforator method (EN 120) [mg/100 g board] below 8.00 3.20

Formaldehyde release was lower than 8 mg/100 g in all experiments due to the positive influence of the liq-
uefied wood components. It can be concluded that the products of the liquefied lignin with their aromatic
character behaved as a formaldehyde scavenger. Lower formaldehyde emissions from particle boards due to
the use of the liquefied wood, are extremely important in the provision of better quality of life. On the basis
of the presented values one can conclude that liquefied woods can be used as a substitute for synthetic resin
precursors in adhesives that are used for particle board production.

8th Conference on Information and Graphic Arts Technology 16

 Invited Lectures

3.3 Polyurethanes

Liquefied lignocellulosic materials are considered to be an alternative feedstock for polymer synthesis. Recent-
ly, considerable attention has been given to the preparation of environmentally friendly polymeric products
from liquefied biomass materials and their derivatives.
In the liquefaction process, the hydroxyl group-containing species in the wood components can be used as
polyols for several different purposes. The polyhydric units that are present in the resulting liquefied/deri-
vatised wood can be used in the creation of polyurethane foams, polyurethane resin precursors and in the
recently developed wood-polyalcohol based urethane adhesives (Čuk, 2015).
We combined the recycled PET polyester with liquefied wood polyester for the production of the polyure-
thane/polyisocyanurate foams. Glass transition temperature, density, and water absorption of the foam in-
creased with increasing amount of glycerol in liquefied wood, compressive stress increased too, while thermal
conductivity was not affected.

3.4 Liquefied biomass as a fuel in gas turbine

Second generation biofuels are increasing their share in the energy source portfolio, which is mainly driven
by the availability of low quality residual biomass and other lignocellulosic materials as well as by their in-
offensive production in comparison with food competitive feedstocks. Two fuels with low carbon footprint,
produced by liquefaction in acidified polyhydroxy alcohols were tested with one being made from cotton
fiber (90% cellulose) and the other one from lignocellulosic biomass (Seljak, 2012). Lower calorific value of
the product is 20.2 MJ∙kg-1 and thus only 20 kJ∙MJ-1 of energy value is consumed for fuel production when
ultrasonic assistance is applied. For comparison, production of bio-oil, which has lower calorific value, uses
72 kJ∙MJ-1. Additionally, both fuels were also neutralized to analyze the impact of product reactivity on com-
bustion performance.
To analyze impact of fuel properties on the combustion performance and exhaust emissions, multiple param-
eters were varied in an experimental study conducted on a laboratory scale gas turbine. To examine effects of
viscosity, fuel was preheated to different temperatures, ranging from 80 to 110°C and airflow varied between
0.15 and 0.23 kg/s. Turbine inlet temperature varied accordingly. Combustion chamber inlet temperature was
kept between 400 and 500°C by means of regeneration. In this way, evaporation of droplets was promoted and conditions
closely resembled those in commercially available systems. Results were compared to those obtained with diesel fuel,
which provided an insight into phenomena influencing performance and exhaust emissions in small co-gen-
eration plants.
Stable combustion of all fuels with CO emission measurements below 500 ppm was achieved even in a flow field
un-optimized combustion chamber. THC emissions remained reasonably low, while NOx emissions increased
with increased turbine inlet temperature and to a lesser extend with increased fuel temperature. Although
pollutant emissions while utilizing the analyzed biofuels, especially those of CO, are higher than correspond-
ing emissions with the diesel fuel, experimental results indicate successful utilization of the analyzed biofuels
in professional gas turbines.

3.5 Nanocellulose

The same liquefaction process was used for the isolation of the nanocrystalline cellulose from biomass (Kunav-
er, 2016). The method is a novelty and a model procedure for NCC isolation from different natural cellulosic
sources with high yields and with high crystallinity index. The process of preparing NCC from different natural
sources uses glycols as the main reactant and an acid catalyst in low concentration (only 3%). Here, during
the one step reaction, lignin, hemicelluloses and the more disordered components of the cellulosic fibers
are liquefied, only the crystalline cellulose remaining as a solid residue. The yields, crystallinity index and
morphology of thus produced NCC were comparable to those NCC products that were described in recently
published papers (Fan, 2012; Texeira, 2010).
The liquefaction reaction, using glycols and mild acid catalysis, was optimized and applied to four model
materials, namely cotton linters, Chinese silver grass, spruce wood and eucalyptus wood. The liquefaction
reaction took place in a glass reactor with constant mixing at 150°C for 180 minutes. Afterwards, the reac-
tion mixture was diluted 1:1 with 1,4-dioxane and centrifuged at 8000 rpm for 20 minutes, re-dispersed and
centrifugation repeated until a clear supernatant liquid was obtained. The % recovery of the nanocrystalline
cellulose, the crystallinity index of the nanocrystalline cellulose and the average crystal dimensions are pre-
sented in the Table 2.

8th Conference on Information and Graphic Arts Technology 17

 Invited Lectures

The SEM micrographs of the isolated NCC are presented in Figure 1. The method can be applied to different
cellulose-containing biomasses. The main benefit of the process arises from the ability to prepare stable NCC
suspensions in an organic medium at 10 times greater loadings than can be achieved in aqueous suspensions.
The liquid residues contain significant quantities of levulinic acid and different sugars that were derived from
cellulose and hemicelluloses.

Table 2. The NCC recovery percentage, the crystallinity index (CrI) and the average NCC crystal length and
crystal width
NCC recovery CrI Avg. NCC crystal Avg. NCC crystal
Biomass [%]
[%] length [nm] width [nm]
Cotton linters 74.5 ± 6.0 80.0 242.0 ± 8.0 12.7 ± 0.4
Chinese silver grass 55.6 ± 4.0 62.8 250.0 ± 17.0 8.9 ± 0.2
Spruce wood 61.5 ± 3.2 63.0 235.0 ± 23.0 8.9 ± 0.1
Eucalyptus wood 63.0 ± 8.5 66.0 306.0 ± 13.0 9.0 ± 0.1

Figure 1. SEM micrographs of NCC: cotton linters (a), Chinese silver grass (b),
spruce wood (c) and eucalyptus wood (d).

When dealing with wood or similar lignocellulosic materials, some sugars derived from hemicelluloses were
identified as was some glucose. It is possible to isolate the levulinic acid selectively from liquid residues and
sugars that were identified in the liquid residue from the wood liquefaction process could be used as a feed-
stock for several fermentation processes in ethanol production.

3.6 Nanocellulose applications in paper and wood coatings

This NCC was mixed in various concentrations into the starch and polyvinyl alcohol coatings (PVA), as an
additive for improving the printability and mechanical properties of selected printing materials (Medvešček,
2017). The main purpose of the study was to optimize the surface properties and thereby to improve the printa-
bility of paper as well as to improve the quality of digital personalized data prints on the security documents.
Uncoated as well as coated paper samples were coated with the prepared starch and PVA coatings and with
the addition of different NCC concentrations. Mechanical measurements and printing analyses were made. It
was established that the addition of 3% of NCC improved the printability of the paper with ink jet printers
since the NCC gives more interlinked structure of the paper and due to its high hydrophobic character, a better
absorption of water-based inks.
NCC was mixed in different concentrations in three different wood coatings and applied to three different sur-
faces: wood, glass and metal (Skok, 2017). By increasing the proportion of the NCC, the viscosity and hardness
increased. There were no major changes in the contact angle. Gloss and roughness of the surface were inversely
proportional. The gloss increased to some extent, and the roughness decreased. According to some published
papers a slightly positive effect of nanocellulose addition on the hardness and modulus of the coatings was
observed while scratch resistance improved consistently (Veigel, 2014).

8th Conference on Information and Graphic Arts Technology 18

 Invited Lectures

4 CONCLUSIONS

The possibilities of converting lignocellulosic biomass and biomass waste in particular into valuable chemi-
cals and raw materials for further use in polymer chemistry are numerous. Intensive studies of this challenge
have been governed for the last two decades. The biorefinery concept has been firmly established and the
only danger that has to be guarded carefully is that the production does not compete with food production.
It provides new prospects of utilizing renewable resources for production of energy, organic chemicals and
polymers through complex processing technologies. Lignocellulosic biomass, the most abundant sustainable
feedstock on earth contains cellulose, hemicellulose and lignin in various proportions. Each of these three
biomass components can then be converted into different platform chemicals. One of the newest challenges
is the isolation and application of nanocellulose, an outstanding material which is carbon neutral, sustain-
able, recyclable and non-toxic. It has the potential to become a true green nanomaterial, which can be used
in high performance applications. The final cost of bioproducts depends upon the processing costs therefore;
high tonnage, high yields and the reduction of the synthesis steps are the most important development trends
nowadays and in the future.
All the products can be regarded as higher added value materials, decreasing the dependency to the crude oil
production and at least not increasing the greenhouse gas emissions.

ACKNOWLEDGMENT
The authors gratefully acknowledge the Slovenian research Agency for financial support (program P2-0145).

5 REFERENCES

• Azadi, P., Inderwildi, O.R., Farnood, R. and King, D.A. 2013. “Liquid fuels, hydrogen and chemicals from lignin:
a critical revew.) Renewable and sustainable energy reviews, 21, 506–523.
• Čuk, N., Fabjan, E., Grželj, P. and Kunaver, M. 2015. “Water-based polyurethane/polyisocyanurate foams made
from recycled polyethylene terephthalate and liquefied wood-based polyester polyol.” Journal of Applied
Polymer Science, 132(8).
• Dutta, S. and Pal, S. 2014. “Promises in direct conversion of cellulose and lignocellulosic biomass to chemicals
and fuels: Combined solvent nanocatalysis approach for biorefinery.” Biomass and Bioenergy, 62, 182–197.
• Jasiukaitytė, E., Kunaver, M. and Strlič, M. 2009. “Cellulose liquefaction in acidified ethylene glycol.” Cellu-
lose, 16, 393–405.
• Fan, J., Li, Y. 2012. “Maximizing the yield of nanocrystalline cellulose from cotton pulp fiber.” Carbohydrate
Polymer, 88, 1184–1188.
• Habibi, Y. 2014. “Key advances in the chemical modification of nanocelluloses.” Chemical Society Reviews,
43, 1519–1542.
• Khalil, H.P.S.A., Bhat, A.H. and Yusra, A.F.I. 2012. “Green composites from sustainable cellulose nanofibrils:
a review.” Carbohydrate Polymers, 87, 963–979.
• Kržan, A. and Kunaver, M. 2006. “Microwave heating in wood liquefaction.” Journal of Applied Polymer
Science, 101(2), 1051–1056.
• Kunaver, M., Jasiukaitytė, E. and Čuk, N. 2012. “Ultrasonically assisted liquefaction of lignocellulosic mate-
rials.” Bioresource Technology, 103(1), 360–366.
• Kunaver, M., Jasiukaitytė, E. Čuk, N. and Guthrie, J.T. 2009. “Liquefaction of wood, synthesis and character-
ization of liquefied wood polyester derivatives.” Journal of Applied Polymer Science, 115(3), 1265–1271.
• Kunaver, M., Medved, S., Čuk, N., Jasiukaitytė, E., Poljanšek, I. and Strnad, T. 2010. “Application of liquefied
wood as a new particle poard adhesive system.” Bioresource Technology, 101, 1361–1368.
• Kunaver, M., Anžlovar, A. and Žagar, E. 2016. “The fast and effective isolation of nanocellulose from selected
cellulosic feedstocks.” Carbohydrate Polymers, 148, 251–258
• Lin, L., Yoshioka, M., Yao, Y. and Sharaishi, N. 1995. “Liquefaction of wood in the presence of phenol using
phosphoric acid as a catalyst and the flow properties of the liquefied wood.” Journal of Applied Polymer
Science, 52(11), 1629–1636.
• Medvešček, S. 2017. “Influence of nanocrystallized cellulose on paper printability.” Master’s thesis, University
of Ljubljana.
• Peng, F., Peng, P. and Sun, R.C. 2012. “Fractional purification and bioconversion of hemicelluloses.” Biotech-
nology Advances, 30, 879–903.
• Seljak, T., Oprešnik, S.R., Kunaver, M. and Katrašnik, T. 2012. “Wood, liquefied polyhydroxy alcohols as a fuel
for gas turbines.” Applied Energy, 99, 40–49.
• Skok, A. 2017. “The influence of the nanocrystalline cellulose on the coatings properties.” Diploma thesis,
Faculty of polymer technology.

8th Conference on Information and Graphic Arts Technology 19

 Invited Lectures

• Teixeira, E.M., Correâ, A.C., Manzoli, A., Leite, F.L., Oliveira, C.R., Mattoso L.H.C. 2010. “Cellulose nanofibers
from white and naturally colored cotton fibers.” Cellulose, 17, 595–606.
• Veigel, S., Grüll, G., Pinkl, S., Obersriebnig, M, Müller, U. and Gondl-Altmutter, W. 2014. “Improving the
mechanical resistance of waterborne wood coatings by adding cellulose nanofibres.” Reactive & Functional
Polymers, 85, 214–220
• Yamada, T. and Ono, H. 2001. “Characterization of the products resulting from ethylene glycol liquefaction
of cellulose.” Journal of Wood Science, 47, 458–464.
• Zhang, L., Xu, C. and Champagne, P. 2010. “Overview of recent advances in thermos-chemical conversion of
biomass.” Energy Conversion and Management, 51, 969–982

Corresponding author:
Matjaž KUNAVER
National Institute of chemistry, Department of polymer chemistry and technology
Hajdrihova 19, SI-1000 Ljubljana, Slovenia
phone: +386 1 476 0363, fax: +386 1 476 0300, e-mail: [email protected]

Co-author(s):
Ema ŽAGAR
National Institute of chemistry, Department of polymer chemistry and technology
Hajdrihova 19, SI-1000 Ljubljana, Slovenia
phone: +386 1 476 0203, fax: +386 1 476 0300, e-mail: [email protected]

Raša URBAS
University of Ljubljana, Faculty of Natural Sciences and Engineering
Snežniška 5, SI-1000 Ljubljana, Slovenia
Phone: +386 1 200 32 83, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 20

 Invited Lectures

COMPUTATION APPROACH FOR REALISATION OF CONTEXT-AWARE ROBOTS

Tomislav STIPANČIĆ1, Duska ROSENBERG2 & Bojan JERBIĆ1

1
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb
2
University of London

ABSTRACT: Humans are using memories, guesses and other implicit information stored or collected to reason
about most appropriate solutions. Unlike humans, robots do not understand context by default. Compared to
conventional approaches where robots are preprogramed to react to a finite number of environmental occur-
rences, contextual awareness can enable modeling of humanlike adaptation skills. Computational models pre-
sented in this work could be understood as context-to-data interpreters that transform contextual information
into data, allowing machines to make context-driven decisions. The basic model contains three main parts.
The first part is used to track and collect significant environmental information. The second part represents
formal knowledge about the domain of interest. The model also contains a probabilistic component realized
by a Bayesian Network. The overall methodology is presented through three separate examples illustrating
reasoning based on: (i) phenomenon of social capital, (ii) human bodily awareness and (iii) human emotions.

Keywords: Affective robotics; Context-awareness; probabilistic reasoning; knowledge representation.

1 INTRODUCTION

Contemporary systems are usually programmed for a limited range of activities foreseen in advance by a
system developer. Such systems cannot act in any unpredicted situation by default. Their reactions are based
only on expected environmental stimuli. Such a reactive system can be very fragile if something unexpected
occurs. This is why robots are so impressive in factories, but so incompetent in any human environment. In
contrast, the system that is able to partially realize context can potentially do both: it can act reactively and
it can comprehend the present and predict future results or actions. It seems that humans and animals are
adapting to their natural environment in a similar way (Barrett, 2017). In most cases, contemporary machines
are using explicit knowledge. In contrast, contextual perception presumes much more implicit understanding.
Research into new methodologies and paradigms is therefore directed toward the development of adaptive,
anthropomatic and cognitive agent capabilities. To achieve this kind of technology it is good to bear in mind
a couple of things. It is not possible to predict all occurrences or changes that arbitrary environment could
derive. Deterministic chaos as a phenomenon of the real world that inevitably obstructs absolute expectations,
always producing slightly changed situations (Stipancic, 2008). Chaos is present in both, temporal and space
continuum, resulting in inconsistencies and uncertainties in all dimensions. Every environment is naturally un-
structured, which can be revealed if observed by using an appropriate scale. In other words, if a sub-molecular
level is neglected from this analysis, it is not possible to completely determine any environment, no matter
how tight the applied tolerance ranges may be. This is connected with issues of sensitivity and instability and
may result in malfunctioning, even if small environmental changes occur. How to deal with such challenges?
One way is to accept deterministic chaos as a natural phenomenon just as it is accepted by nature.
This paper emphasises the direction in context modelling where insights taken from three separate use case
scenarios are discussed. In Section 1context modelling is outlined. In Section 2 theoretical explanation of the
model is given together with the structure of the proposed model. In this way, detailed insights into “rational”
and “probabilistic” parts of the computation mechanism are provided. Three use cases where the methodol-
ogy is tested are presented and discussed in Section 3 together with directions for future work together with
conclusions.

2 THEORETICAL EXPLANATION OF THE MODEL

To mathematically describe the model, a multi-agent approach is used. In the model of interaction (Fig. 1),
all agents (artificial agents or humans form a part of the same environment) and are able to communicate
mutually, interact and share information within the same time domain.

8th Conference on Information and Graphic Arts Technology 21

 Invited Lectures

Figure 1. The model of interaction.

In this way, agent-to-agent and agent-to-environment interaction can be both mathematically described as:

(1)

where the information collected by sensors and the model are mathematically defined as:

(2)

(3)

In (2), vectors representing sensors Snm are used to detect a targeted phenomenon. Information
acquisition represents the first step in contextual perception of the environment. Therefore, these vectors
contain information acquired by sensors that are placed ubiquitously into the environment in a meaningful
way (4).

(4)

Based on (1), there is at least one function G to describe context of an environment at a given moment ti, us-
ing information from (2) altogether with a set of criteria defined in (3) that generates a desired (optimal)
robot behaviour, Gopt. As a part of the vector , the marks CO and BN defined in (3) are abbreviations for Case
Ontology and Bayesian Network, respectively.
By following the presented mathematical formulation, a hypothesis of this paper is:
By finding the function Gopt defined in (1) and respecting the information stored in (3) along with other in-
formation collected by sensors (2), it is possible to alter a behaviour of an artificial agent based on targeted
implicit or contextual information.

2.1 The model as a computation mechanism

In essence, all human cognitive processes are seen here as context-driven. In (Dey, 2010), context is defined as
any information that can be used to characterize the situation of an entity. A system is context-aware if it uses
context to provide relevant information and/or services to the user, where relevancy depends on user’s task.
Context-aware applications look at who’s, where’s, when’s, and what’s of entities and use this information to
determine why a situation is occurring.
To provide implicit or context driven decision-making capabilities to the artificial systems it is proposed a new
computation mechanism that contains the following components: (i) data acquisition and transformation,
(ii) semantically defined knowledge, and (iii) Bayesian Network (BN), as shown at (Fig. 2). The overall meth-
odology is presented in this paper through three separated use case scenarios, which are explained in detail
in (Stipancic, 2016, Jerbic, 2015, Stipancic 2017).

8th Conference on Information and Graphic Arts Technology 22

 Invited Lectures

Figure 2. Computation mechanism for context to data transformation.

An environment in this vision becomes a space constantly analysed by smart sensors to detect significant
changes. In relation to the real world, humans perceive only the information currently classified as significant
and the majority of other occurrences remain hidden because the nature of such events is not relevant at the
moment. This work adopted the conceptual framework from situation theory, a mathematical theory of infor-
mation, where “…recognition is made of the partiality of information due to the finite, situated nature of the
agent (human, animal or machine) with limited cognitive resources. Any agent must employ necessarily limited
information extracted from the environment in order to reason and communicate effectively...” (Devlin, 2008).
The second part (ii) holds an expert’s knowledge about the domain of activity. This part is used for logical or
rational reasoning. It is called Case Ontology because it represents just a small part of the world in relation to
the model application. The computation mechanism highly relies on predefined knowledge about the environ-
ment. In this case, knowledge is a subjective view of the system designer about targeted context or situation.
Some authors consider certain types of context as important while characterizing a situation of a particular
entity (Dey, 2010). Such contextual information can answer questions like: where, who, when and what. These
represent the core of the knowledge implemented in Case Ontology. Ontology Web Language (OWL), used to
define Case Ontology, follows the principles of Open World Assumptions (OWA) (Loyer, 2005). Such ontology
can respond to a query by providing more than one right answer, thus allowing ambiguities in solutions. By
combining inputs from sensors, ontology defines possible solutions in the form of robot responses.
The third part (iii) of the mechanism enables reasoning under uncertainties implemented in Bayesian Network
and is used to ensure a single solution in relation to perceived context. Bayesian (Believes) Networks (BNs)
reflect beliefs about the most appropriate solution in relation to perceived phenomenon. They allow the use of
prior knowledge needed for capturing domain concepts, variables and probability values as well as building
a graphical representation. BNs are convenient if evidence is not provided. While building BN in this work, a
handcrafted approach is used (Daniel, 2003). This approach is usually time consuming and can be used to build
small BNs. At the same time, this approach is very convenient when subjective experiences of a real human
expert need to be coded in a computation model. The overall procedure of BN development is depicted in (Fig.
3). The first step in a BN development procedure is to define variables of interest, which are network nodes,
and place them into a network topology. Arcs in BN connect the nodes with the direction indicating causal
relationships. Condition Probability Tables (CPTs) quantify relationships between connected nodes. In the
methodology used in this work, information about conditional probabilities has to be calculated in advance.
By altering such information, the system designer gets the opportunity to define system priorities and/or to
achieve certain goals. Each node in accompanied CPT contains probabilities emerged from influences of parent
nodes. Given the specification of BN, it is possible to compute posterior probability distributions for each of
the nodes, so-called “beliefs”.
Determination of probability values within CPTs is the most important task in a design of BNs because those
probabilities directly alter the network behaviour. Such values are often determined by using data mining
techniques applied on some larger amount of data that describe a targeted phenomenon. In the approach
used in this work, conditional probabilities are determined by analysing qualitative descriptions of relations
between network nodes. A more comprehensive description of this procedure can be found in (Jerbic, 2015,
Stipancic, 2017, Stipancic 2016).
The next phase in the development of BN is testing the network performance through three scenarios (or more
than three, if needed). In that phase, the model designer fine-tunes the network behaviour. The first scenario
represents an extremely positive situation. The second scenario represents a neutral situation and the third one
represent an extremely bad situation. The main goal of this procedure is to ensure general network behaviour
in all situations.

8th Conference on Information and Graphic Arts Technology 23

 Invited Lectures

Figure 3. BN development procedure.

To find out secret or inner influences that parent nodes have on child nodes and to test a quantitative part
of BN the method called Sensitivity Analysis (Oakley, 2004) is used. The procedure can provide more insights
into inner reasoning mechanisms of the network based on different node inputs and reduction in the system
entropy. Shannon’s Entropy is a measure for uncertainties of a particular event associated with a probability
distribution of a possible event (5).

H (P) = − ∑ P(s)* log 2 P(s)

(5)
s∈S

This study employs the entropy reduction method to determine a decrease in query node’s entropy before –
H(Q), and after – H(Q/F) the evidence is provided to some particular node in the network. The method helped
in determining those nodes to which query nodes (robot response variables) are significantly sensitive (6).

I = H (before) − H (after) = H (Q) − H (Q / F) (6)

The aim is to provide proofs to all BN nodes one by one and to validate and measure how that affects query
nodes. These nodes that cause the most significant reduction in entropy are the most influential ones for mak-
ing decisions or changing the network reasoning output. Such insights can be used in the following two steps
of the BN development procedure (Fig. 3) where the network reasoning could be additionally tested or refined.
The last step in this procedure represents the integration of the model in accordance to the model application.

2.2 First Use Case Scenatio – reasoning based on social capital phenomenon

In sociology, the concept of SC indicates the expected mutual benefit emerged from cooperation between indi-
viduals within a group. A value realized through social contacts can be measured by determining the increase
in group productivity. In (Daniel, 2005), SC is defined as a common social resource that facilitates sharing of
information and building knowledge through continuous interaction.
By implementing this concept into a group of industrial robots on assembly assignments some interesting
system capabilities emerge, such as: system scalability, auto-recovery and partial contextual awareness. The
system scalability resulted with increased overall group productivity because all the system components (ro-
bots and other system equipment) are classified and defined within the core ontology. By adding new Working
Places, which are defined as a class along with all accompanying subclasses within the Case Ontology, it is
easy to increase the overall production capacity. The second principle of auto-recovery can be recognized in
such cases where some Working Place fails in performing its primary function, due to a defect or something
similar. By using the Bayesian Reasoning part of the model, other working places can rearrange priorities and
continue production. The third principle of partial context-awareness can be found in the way sensors are
used while collecting information from the environment. Sensors are placed seamlessly to provide continuous
flow of information. A final BN for this use case contains fourteen nodes in total where five of them represent
query nodes used to control robot reactions. Detailed explanation of this work is available at (Stipancic, 2016).

8th Conference on Information and Graphic Arts Technology 24

 Invited Lectures

2.3 Second Use Case Scenario – reasoning based on bodily awareness

Some authors emphasize the process of perception as the very first step in qualia development (Haikonen,
2012). Among various definitions, qualia is defined as: the ways things look, sound, and smell, the way it feels
to have a pain, and more generally, what it is like to have experiential mental states (...) qualia are experiential
properties of sensations, feelings, perceptions, and, more controversially, thoughts and desires as well (Gut-
tenplan, 1994). It seems that qualia appear in the human mind as a mental picture (subjective interpretation)
of perceived environmental occurrences. How perceived information appears in the form of subjective experi-
ences today still remains a question. Bearing this in mind, qualia in this work are used with extreme caution.
To simulate bodily-awareness qualia, a new cognitive model is proposed. The developed model additionally
combines the visual perception of the robot itself, enabling it to build a kind of ”mental” representation of
its own body/existence within the environment. In the robot’s workspace, the human operator as well as any
other dynamic obstacle can appear as an object that can arbitrarily change its course and speed. By using
the cognitive model, the robot is able to avoid, approach or escape from any kind of object while performing
its spatial movements. If an obstacle is too near, the robot will decrease its speed to further ensure the safe
operation and to plan its next movements while heading to the final movement point. Detailed explanation
in (Jerbic, 2015).

2.4 Third Use Case Scenario – reasoning based on bodily awareness

The main hypothesis of this study is the idea that emotion may perform an adaptive function that requires a
certain degree of processing complexity. Several studies reveal that cognitive processes in humans are highly
intertwined with emotions. Emotions are considered to signal a person and motivate appropriate responses
in relation to situations (Kim, 2005).
Emotions are necessary information for our wellbeing, our everyday experiences and even cognitive processes
(Balduzzi, 2009). Ortony, Clore, and Collins defined emotions as valenced reactions (e.g., affective reactions
based on the perceived goodness or badness of things) and asserted that emotions are determined by how the
eliciting situation is understood by a person (Ortony, 1988).
A final BN for this use case contains 22 nodes in total, where six of them represent query nodes used to control
the robot reactions. Detailed explanation of this work is available at (Stipancic, 2017).

3 DISCUSSION & CONCLUSION

The approach presented in this work builds on the notion that human cognition has the ability to handle un-
certain information (Doya, 2007). It does not, however, attempt to explain how the brain interprets perceived
phenomena. This work is more focused on human representations, meanings and manipulation of uncertain
information in order to examine the effect of uncertainty on the design of technical systems. In this way, the
aim is to reflect subjective experiences of real human experts as they pick up information from the environ-
ment. This methodology is highly convenient when big data used to describe some phenomenon and build a
model is not available.

Figure 4. The simultaneous – contrast illusion.

Desired robot reasoning can be explained by examining (Fig. 5) where two squares are having exactly the
same gray color value. By adding different backgrounds to both squares, the square on the left is perceived as
different from the square on the right. It seems that a change in context where objects are placed can change
the way how people see them Adelson, 2000, Stipancic, 2010). This change in perception is triggered by mech-
anisms that are much more abstract than a simple true – false logic.
To validate a methodology, the third part of the computation mechanism is assessed in all use cases. BN is
validated from the aspect of information entropy reduction. Some hidden and relative influences between
the network variables are revealed. In this context, the method leads to better understanding of the overall
system behaviour in relation to a particular application. Query nodes are sensitive to more then one variable
whereby those nodes that are the closest to query nodes and those with the strong positive connections are

8th Conference on Information and Graphic Arts Technology 25

 Invited Lectures

the most influential. In this way the methodology presented in this paper shows potential contribution to
the design of context-aware robots.

ACKNOWLEDGMENT
Authors would like to acknowledge the support of Croatian Scientific Foundation through the scientific project
ACRON – A new concept of Applied Cognitive RObotics in clinical Neuroscience.

4 REFERENCES

• (Adelson, 2000) Adelson E. H. Lightness Perception and Lightness Illusions. The New Cognitive Neurosciences,
2nd (ed.), MIT Press, Cambridge, pp. 339-351, 2000.
• (Balduzzi, 2009) Balduzzi D., Tononi G. Qualia: The Geometry of Integrated Information. PLoS Computational
Biology, 2009.
• (Barrett, 2017) Feldman Barrett L. How Emotions are Made: The Secret Life of the Brain. Houghton Mifflin
Harcourt. ISBN 9780544133310, 2017.
• (Daniel, 2005) Daniel B.K., McCalla G., Schwier R.A. Data Mining and modeling social capital in virtual learn-
ing communities. In Proceedings of the 12th International Conference on Artificial Intelligence in Education
2005, Amsterdam; pp.2000-2008.
• (Devlin, 2008) Devlin K., Rosenberg D. Information in the Study of Human Interaction, in Adriaans P & Van
Benthem J (Eds): Philosophy of Information volume 8, Handbook of the Philosophy of Science, pp. 692-693.
Elsevier, ISBN: 978-0-444-51726-5, 2008.
• (Dey, 2010) Dey A.K. Context – Aware Computing,” in: Krumm, J. (Ed.), Ubiquitous Computing Fundamentals,
pp. 321-352, 2010.
• (Doya, 2007) Doya K., Ishii S., Pouget A., Rao R.P.N. (Eds.), “Bayesian Brain – Probabil. Appr. to Neural Cod-
ing,” MIT Press 4, 36, 2007.
• (Guttenplan, 1994) Guttenplan S. D. Qualia: The Geometry of Integrated Information. Oxford, OX, UK; Cam-
bridge, Mass., USA: Blackwell Reference, 1994.
• (Haikonen, 2012) Haikonen P. O. “Consciousness and Robot Sentience,” World Scientific Publishing Co. Pte.
Ltd. Singapure, 2012.
• (Jerbic, 2015) Jerbic, B., Stipancic, T., Tomasic, T. Robotic Bodily Aware Interaction within Human Environ-
ments. Proceedings of 2015 SAI Intelligent Systems Conference, London, 305-314, 2015.
• (Loyer, 2005) Loyer Y., Straccia U. Any-world assumptions in logic programming. Theoretical Computer Sci-
ence, 342 (2), pp. 351-381, 2005.
• (Oakley, 2004) Oakley J.E., O’Hagan A. Probabilistic sensitivity analysis of complex models: a Bayesian ap-
proach. J. of the Royal Statistical Society: Series B 2004; 66(3), 751-769.
• (Ortony, 1988) Ortony A., Clore G. L., Collind A. The cognitive structure of emotions. Cambridge University
Press, UK, 1988.
• (Stipancic, 2010) Stipancic T., Jerbic B. Self-adaptive Vision System. Emerging Trends in Technological In-
novation. Camarinha-Matos LM, Pereira P, Ribeiro L (ur.). Heidelberg : Springer Verlag, pp. 195-202, 2010.
• (Stipancic, 2016) Stipancic T., Jerbic B., Curkovic P. A context-aware approach in realization of socially intel-
ligent industrial robots. Rob. and comp. integrated manuf. Elsevier, 37, pp. 79-89, 2016.
• (Stipancic, 2017) Stipancic T., Ohmoto Y., Badssi S. A., Nishida T. Computation Mechanism for Situated Sen-
tient Robot. Proc. of the 2017 SAI Computing Conf., London: IEEE, pp. 64-73, 2017.

Corresponding author:
Tomislav STIPANČIĆ
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb
Ivana Lucica 5, 10000 Zagreb, Zagreb, Croatia
e-mail: [email protected]

Co-author(s):
Duska ROSENBERG
Emeritus Professor (iCOM) University of London, FNCUP, FRSA
Bedford Square London WC1

Bojan JERBIĆ
Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb
Ivana Lucica 5, 10000 Zagreb, Zagreb, Croatia
[email protected]

8th Conference on Information and Graphic Arts Technology 26

 Invited Lectures

3d PRINTING FOR DESIGN AND ART

Barbara WILKESMANN HOČEVAR1, Deja MUCK2

Chemets d.o.o., Velesovska cesta 20, 4208 Šenčur
1

2
University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia

Abstract: 3D printing has been known around the world for at least three decades. Different technologies exist,
and consequently, different materials with which they can print. Originally, 3D printers were used for
the production of prototypes, above all in the fields of industrial design development, construction and the
production of technical products. With the development of new materials, the use of 3D printers is spread-
ing day-by-day to new areas, from printing food, bio-printing, and 3D printed electronics, to the building
industry and art.
A German company Voxeljet which produces larger format industrial printers, with a consequently greater
work volume (maximum platform 4m x 2m x1m) in the last four years, has stepped up its search for solutions
to different applications in the fields of art and building construction. In the article, different unique, specific
and demanding applications are presented, which were produced with the assistance of various 3D technolo-
gies, from scanning, the preparation of suitable data for 3D printing and manual post-production processing
of printed parts to their placement in the environment.

Keywords: 3D printing, building construction, art, Voxeljet.

1 INTRODUCTION

Voxelyet 3D printing technology is based on the principle of depositing droplets, or rather spraying a bind-
er (furan resin) onto a base material (foundry sand of different granulations) or depositing a binder onto a
powder-based material PMMA, which is further strengthened with an epoxy resin for better durability and
solidity (1, 2). Printed items, produced on the basis of foundry sand or the polymer based material PMMA can
be later kitted, which allows for additional sanding, polishing and painting. The advantage of such production
methods is a shorter production time and a short time to market, lower production costs in comparison to
conventional processes, but above all the WOW FACTOR, as everyone asks: “How is that possible?”

2 EXPERIMENTAL

In the following projects from the fields of archaeology, art, architecture and building construction:are pre-
sented.

A facade and a uniquely designed ceiling

The design and production of a unique ceiling is one of the most expensive elements in building a home. The
price can exceed USD 30,000 even without the costs of planning and construction (3, 4).
Aztec Scenic Designs, a company specialised in the creation of unique facades and ceilings, found a solution
for their design in 2014, by using 3D printing. Their designer, Benjamin Cremer, drew up the desired form,
while an outside contractor, Marc Leonard, transformed it via the use of a computer programme into a data
file, which enabled it to be 3D printed. All parts were additionally worked on: golden decorative flakes were
added, and it was given it an ‘aged look’. They were then fixed to wooden panels, which enabled them to be
attached to the ceiling (Figure 1).
The advantage of this method of production was in the speed of it, as 8 printers were used. They printed the
demanding and precise patterns simultaneously, and in the end, the material used in producing them was
rather negligible, costing cca. USD 200.

8th Conference on Information and Graphic Arts Technology 27

 Invited Lectures

Figure 1. Uniquely designed celling.

The final product and attachment to the ceiling took only 2 days. In this case, the designer designed the ceil-
ing, but it is also possible that archaeological finds be 3D scanned, and in the same manner, reproduced and
‘brought to life’.

A replica of an existing portal in Berlin and a source material for sculptors

This huge project relates to the reconstruction of the Liebkencht historical portal in Berlin. Upon deciding that
a city palace be built in Berlin, there was also a wish to have a full-sized replica of the Liebkencht portal (Fig-
ure 2) in it, with the goal being to maintain the originality and historical meaning of the structure. TrigonArt,
3D scanning specialists from Berlin, scanned the structure three-dimensionally and created a high-resolution
model. In addition, they prepared a 3D model of it for presentation purposes and also for its reproduction,
which was taken care of by the Voxeljet company from Augsburg. The entire construction encompassed approx-
imately 100 individual printed parts, their dimensions being 1.5 × 1.0 × 1.0 metres. It would have been possible
to print them in even bigger dimensions, 4 × 2 × 1m, but this would have created handling and transportation
problems, due to the weight and size of the pieces.
The historical portal replica completely matches the original, both geometrically and scale-wise, which was a
huge advantage for sculptors, as the 3D printed structure served as a source material for their work, enabling
them to produce a perfect copy of the Liebknecht portal in sandstone. The advantage of this method of pro-
duction was in the speed of it, but more importantly in maintaining the original, as there would have been a
risk of it being damaged if it were replicated in the classical method of doing so (5, 6).

Figure 2. full-sized replica of the Liebkencht portal.

Film and entertainment

“In the film, Zero Dark Thirty, the filmmakers couldn’t get the right night-vision goggles, because they were
$60,000 per unit. So, I modelled them from some photos that had leaked on the internet.” PROP SHOP 3D
SUPERVISOR, JET COOPER. On the basis of a computer- drawn image, a plastic model was printed using a
VX1000 3D printer . The primary material was PMMA plastic, which was impregnated with epoxy for greater

8th Conference on Information and Graphic Arts Technology 28

 Invited Lectures

strength. The produced goggles and mask were additionally hand worked, kitted, sanded and painted, so that
the achieved final look was that as shown in the accompanying pictures (Figure 3), and were used by the actors
during filming (7, 8).

Figure 3. Original and printed night-vision goggles.

Building construction and interiors

The Voxeljet company, in first testing the application of their sand prints, chose to produce a ‘mini- concrete
wall’ into which their company logo can be seen, which now stands in front of their office building. The com-
pany decided to use the classical method of concreting (pouring concrete into a wooden panelled mould)
for the main part, while use a 3D printed wax insert in combination with a sand mould for the text and the
company’s logo, in order to make them ‘stick out’, which was later on impregnated with epoxy. This gave the
mould better strength and thus enabled easier concreting, as shown in the figure 4. The 3D printed ‘negative’
of the logo, that allowed concreting through an opening in the text.
Cast concrete and the removal of the printed mould. The advantage of such a solution is, that the concrete can
be poured in a more desirable and sheltered place, where weather and light do not impact on the production
process. In the same way, the process is ideal for demanding, and non-linear lines. 3D printed sculptures can
enhance interior spaces, for example, in homes, halls, exhibition places, shops, hallways .... Imagination knows
no boundaries.

Figure 4. Sand print of company logo.

The basis for the production of any kind of sculpture, non-bearing columns or shapes is a 3D computer draw-
ing, which enables the 3D printing of those desired elements. Size is also not a problem, as it is possible to
assemble and glue, post production, just like blocks. Knowledge of postproduction surface finishing gives us
the advantage that seams are no longer seen, and the finished products take on an even more pleasant appear-
ance. Into such finished products, we can fit many different devices, such as room scent fresheners, speakers,
temperature gauges ....

The company, Chemets, has taken advantage of this by printing a torso (Figure 5), utilizing Voxeljet’s tech-
nology. With post-production impregnation, the required hardness was achieved, allowing the possibility
of post-production finishing. Half of the raw printed torso was painted in black, while the other half was
professionally finished and painted.
The lower part has an opening for a room freshening device. The torso is intended to be used at trade fair
presentations, the purpose being that the company ‘stands out from the crowd’, and that potential business

8th Conference on Information and Graphic Arts Technology 29

 Invited Lectures

clients are attracted to take a closer look at the company’s showcased products and services. The inner part
of the sculpture contains a device for ‘aroma marketing’, which releases a pleasant scent into the air at pre-
determined intervals. The intertwinement of scent and visual image enhances the company’s recognition and
consequently assists in increasing sales.
In such cases, we can see that companies which deal in 3D printing and 3D printers, together with manufac-
turers, tradesmen and craftsmen are changing the way we think and use 3D technologies and are looking for
the advantages, which these new technologies bring with them.

Figure 5. Sand printed natural sized torso.

3 CONCLUSIONS

The trend is moving faster in the direction that design is, and will be, all the more important, above all because
there is a trend towards personalization, uniqueness and quick implementation.

In this way, companies active in 3D printing and with 3D printers, together with manufacturers, craftsmen and
artisans are changing the way in which 3D technologies are understood and the manner in which they can be
used. The trend is moving all the more quickly in the direction that design is more and more important, above
all because trends are moving towards personalization, uniqueness and short times to market.
For the general recognisability of a company and its brand name, it is necessary to connect with other com-
panies, and with their products, favourably influence our senses, eyesight, smell, hearing and taste. Therefore,
it is crucial, that we turn to technology, and use it to our advantage.

4 REFERENCES

1. Chemets. http://www.chemets.si (last accessed on 03. 03. 2018)

2. Voxeljet. https://www.voxeljet.com/ (last accessed on 13. 03. 2018)
3. http://3dprint.com/1547/custom-3d-printed-ceiling-amazing (last accessed on 21. 02. 2018)
4. http://3dprint.com/1547/custom-3d-printed-ceiling-amazing
5. www.voxeljet.de/en/case-studies/case-studies/printed-memorial (last accessed on 10. 01. 2018)
6. www.voxeljet.de/en/case-studies/case-studies/entirely-3d-printed-room/ (last accessed on 10. 01. 2018)
7. www.propshop.co.uk/in-the-name-of-a-new-art/ (last accessed on 19. 01. 2018)
8. www.empireonline.com/features/3d-movie-prop-printing (last accessed on 10. 01. 2018)

Corresponding author:
Barbara WILKESMANN KOČEVAR
Chemets d.o.o.
Velesovska cesta 20, 4208 Šenčur
phone: +386 4 235 44 70, fax: +386 4 235 44 77, e-mail: [email protected]

Coauthor(s):
Deja MUCK
University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 84, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 30

 Oral
presentations
 Oral presentations

BIOCOMPOSITE PLA FILAMENT FOR 3D PRINTING

Diana GREGOR-SVETEC, Urška STANKOVIČ ELESINI, Mirjam LESKOVŠEK,

Urška VRABIČ BRODNJAK & Raša URBAS
University of Ljubljana, Faculty of Natural Sciences and Engineering

ABSTRACT: In Fused Deposition Modelling (FDM) usually ABS, PLA, and PA are used as thermoplastic materi-
als. Nowadays, the focus is in developing biocomposite thermoplastic materials, among others to produce 3D
filaments. To improve some properties and broaden the applicability, PLA polymer is blended with bio- and
non-bio-degradable resins and mixed with fillers, such us micro-and nanoparticles or fibres. In our research
the biocomposite filament was produced from PLA, with added 50% of corrugated cardboard dust, which was
composed of broken fibres, fibrils, fines, starch glue and other particulates resulting from corrugated cardboard
production. Structural, morphological and mechanical properties of the biocomposite filament PLA/cardboard
dust were compared with the properties of the pure PLA filament. Compared to pure PLA, biocomposite PLA
filament had lower density, inferior tensile properties, relaxation transitions occur at lower temperature,
thermal stability is lower, while the bending resistance is the same and the compression resistance is higher.

Keywords: 3D filaments, PLA, biocomposite, cellulose, structure.

1 INTRODUCTION

The most spread 3D printing technology nowadays is material extrusion of thermoplastics filaments, commer-
cially named as Fused Deposition Modelling (FDM). Usually acrylonitrile butadiene styrene (ABS), polylactide
(PLA), and polyamide (PA) are used as thermoplastic materials. PLA is biodegradable aliphatic polyester, ob-
tained from corn, sugar cane, potato or other plants. PLA has some positive as well as some negative proper-
ties, such as low density resulting in lightweight products, low flammability, high moister absorption ability,
high ductility and low toughness (Gupta et al., 2007). To improve some properties and broaden applicability,
PLA polymer is blended with bio- and non-bio-degradable resins and mixed with fillers, such us micro-and
nanoparticles or fibres (Poh et al., 2016; Honglin et al., 2014; Atsuhiro et al., 2008).

Nowadays, the focus of different researches is in developing biocomposite thermoplastic materials, therefore
our goal was to determine the properties of the biocomposite 3D filament. The biocomposite filament was
produced from PLA, with added 50% of corrugated cardboard dust, which was composed of broken fibres,
fibrils, fines, starch glue and other particulates resulting from corrugated cardboard production.

2 EXPERIMENTAL

The measurement of the thermal stability was determined by the sample inspection with the Mettler FP84HT
Hot stage thermal measuring cell and microscope. With this method the changes during heating of the fila-
ments, were analyzed and detected. The analyzed starting point began at 30ºC; with the heating speed 5ºC/
min, while the final temperature of the heating ended at 200ºC.

Dynamic Mechanical Analysis (DMA) was performed using Q800 DMA analyser (TA Instruments). The meas-
urement was performed in dual cantilever bending mode on the samples with the length of 35 mm. The meas-
urement was performed at the frequency 1 to 10 Hz and the temperature range 0–160ºC, with temperature step
3ºC/min, rising up to 160ºC. Dependent on the temperature and the frequency, the transition temperatures
in the polymer, the storage modulus (E’), the loss modulus (E’’) and the tan delta (tan δ) were determined.

Analyses of the filament surfaces and the cross-sections was performed by the Scanning electron microscope
JSM-6060 LV (Jeol) (SEM), at different magnifications (50× and 1.000×).

Time of the sound wave propagation through the filament at a frequency of 160 Hz was measured using the
Puls Propagation Meter PPM-5R (H. Morgan, Co). The longitudinal speed of the sound waves traveling through
the filament was calculated as a ratio of distance and time of puls propagation.

8th Conference on Information and Graphic Arts Technology 32

 Oral presentations

The filaments were qualitatively analysed with the FT-IR spectrometer Spectrum GX I (Perkin Elmer). All the
spectra were recorded over the range 4000–800 cm–1, with the resolution of 4 cm–1 and averaged from 64 scans.

The mechanical properties were determined using the tensile testing machine Instron 5567 (Instron). The
tensile properties of the filaments were determined according to the ASTM D2256. The bending resistance
was determined with the 2-point bending test and the compression load was measured by compressing the
filament to half its thickness.

3 RESULTS & DISCUSSION

Results showed that after corrugated cardboard dust was added to the PLA polymer the density of the filament
decreased from 1.26 g/cm3 to 0.961 g/cm3 while longitudinal speed of the sound waves traveling through the
filament increased for 0.5 km/s. The later applies that the structure of the biocomposite filament was more
oriented in direction of filament axes and more uniform compared to pure PLA filament.

ATR-FTIR spectra in the 4000 to 2000 cm-1 range showed typical cellulose peaks, which confirmed the presence
of corrugated cardboard dust in the biocomposite PLA sample (Figure 1). The peak at 3312 cm-1 represents the
vibration of the -OH group in the cellulose. With pure PLA filament, the peak of 3509 cm-1 was detected in this
region, representing the asymmetric and symmetric expansion of C-H bond of the PLA. The same peak was
detected for the biocomposite filament PLA/cardboard dust, but at a much lower intensity.

Figure 1. FTIR analysis of 100% PLA and PLA+50% of corrugated cardboard dust sample,
in the range of 3600 and 2000 cm-1.

In the range between 1800 and 800 cm-1 higher peak intensity was detected for the pure PLA filament, compared
to the biocomposite filament PLA/cardboard dust (Figure 2). There were no other significant differences detect-
ed between the filaments in this range. For the biocomposite filament PLA/cardboard dust, at the wavelength
of 1648 cm-1, the peak represents the vibration of CH2 group and the vibration of intermolecular hydrogen
bonds in the cellulose. This again confirms the presence of the cellulose, though, despite the addition of 50%
corrugated cardboard dust, the content of the PLA component stands out.

Figure 2. FTIR analysis of 100% PLA and PLA+50% of corrugated cardboard dust sample,
in the range of 1800 and 800 cm-1.

8th Conference on Information and Graphic Arts Technology 33

 Oral presentations

Thermal analysis has shown, that the first changes in the structure of the pure PLA filament occurred at 147ºC
and the melting was completed at 172ºC. The results showed that the pure PLA filament was more thermally
stable sample. At the biocomposite filament PLA/cardboard dust the first changes were noticed at 120ºC. The
part of the biocomposite, where the cardboard dust was present, firstly caused the color change (it yellowed)
and then melted at 140ºC, whereas the PLA component completely melted at 150ºC.

In the Figure 3 to 5, the dynamic mechanical properties of the pure PLA and the biocomposite filament PLA/
cardboard dust are presented. As it can be seen from the diagram in Figure 3, both samples have rather low
storage modulus E’, which applies to the high willingness and low elasticity of both materials. The temperature
of the relaxation transition of the biocomposite filament PLA/cardboard dust was slightly lower (58.04ºC) than
for the pure PLA filament (59.39ºC), while the belonging interval was wider (from 53.69 to 63.91ºC) as in the
case of pure PLA (from 56.53 to 62.26ºC). According to the results of the longitudinal speed of the sound waves
traveling, the biocomposite filament PLA/cardboard dust has shown slightly better uniformity of the structure,
which could also be proven by the wider relaxation transition interval. Later also applies to the higher share
of better organised segments. The peaks of the loss modulus E’’ (Figure 4) for both samples were prominent
and were reached at 56.56ºC and 52.72ºC for the PLA and the biocomposite PLA/cardboard dust, respectively.
From the tan δ curves in Figure 5 can be seen that the values of the damping factors were higher in the case
of the pure PLA, probably because of the more intensive movement of molecular segments.

Figure 3. Storage modulus (E`) vs. temperature for the PLA and PLA/cardboard dust filament samples.

Figure 4. Loss modulus (E``) vs. temperature for the PLA and PLA/cardboard dust filament samples.

8th Conference on Information and Graphic Arts Technology 34

 Oral presentations

Figure 5. Loss factor (tan δ) vs. temperature for the PLA and PLA/cardboard dust filament samples.

Images of the filament’s surface obtained by the SEM revealed no differences. The cross-section of the pure PLA
filament revealed a porous structure. On the other hand, the cardboard dust in biocomposite filament acts as a
filler in the polymer (PLA) matrix thus structure was less porous, more uniform and surface smoother (Figure 6).

PLA PLA/carboard dust PLA/cardboard dust

50-× 50-× 1000-×

Figure 6. Cross-section of the filaments (SEM; 50× and 1000× magnification).

Mechanical properties of the two analysed filaments under tensile, bending and compression load were also
evaluated (Table 1).

Table 1. Mechanical properties of the pure PLA and bicomposite PLA/cardboard dust filaments.
Tensile Strain at Strain at Elastic Load at Load at
strength max. load break modulus bending compression
(MPa) (%) (%) (GPa) (N) (kN)
PLA 56.46 5.91 12.46 1.446 2.71 8.85
PLA/
11.57 3.27 4.12 0.844 2.73 10.19
card. dust

The addition of the corrugated cardboard dust to the PLA polymer influenced the mechanical properties in
different ways. As the tensile strength and elastic modulus of the biocomposite filament were substantially
lower, the load at bending was equal, whereas the load needed to compress the filament to half its thickness
was 15% higher. The biocomposite filament PLA/cardboard dust was weaker only at loads acting in direction
of filament axes, as the tensile strength was 5-times lower, strain at break 3-times lower and elastic modulus
almost 50% lower compared to the pure PLA filament.

8th Conference on Information and Graphic Arts Technology 35

 Oral presentations

4 CONCLUSIONS

Structural, morphological and mechanical properties of the biocomposite PLA filament were compared with
properties of the pure PLA filament. With the addition of the corrugated cardboard dust to the PLA polymer the
changes in density, uniformity, chemical structure and transition temperatures of the samples were detected.
The structure of both filaments was mainly amorphous, with low order of structural elements, nonhomo-
geneous, porous. The biocomposite filament had lower density, the relaxation transitions occurred at lower
temperature, and lower thermal stability. The biocomposite filament also showed inferior tensile properties,
much lower tensile strength, breaking length, elastic modulus and storage modulus in comparison with the
pure PLA filament. Both filaments were tough thus bending filament till 15º didn’t reveal noticeable difference
in the bending load. By compressing the filaments to half their thickness higher load was determined for the
biocomposite filament, which was in agreement with the morphological study of the filament cross-section.
SEM image analysis has revealed less porous structure for biocomposite filament, which also showed a bit
higher sound wave velocity propagation meaning more condensed structure in the longitudinal direction of
filament.

ACKNOWLEDGEMENT
Research is a part of the project CEL.KROG Discarded potentials of biomass. Programme »Potential of bio-
mass for development of advanced materials and bio-based products« is co-financed by EU Structural Funds
in Slovenia.

5 REFERENCES

• Atsuhiro, I., Masaya, N. and Hiroyuki, Y. “Cellulose Nanofiber-Reinforced Polylactic Acid.” Composites Science
and Technology, 68 (2008): 2103–2106.
• Gupta, B., Revagade, N. and Hilborn, J. “Poly(Lactic Acid) Fiber: An Overview.” Progress in Polymer Science
32 (2007): 455–482.
• Honglin, L., Guangyao, X., Chunying, M. Peng, C., Fanglian, Y., Yong, Z., Chuanyin, Z. and Yizao W. “Me-
chanical and Thermo-Mechanical Behaviors of Sizing Treated Corn Fiber/Polylactide Composites.” Polymer
Testing 39 (2014): 45–52.
• Poh, P.S.P., Chhaya, M.P., Wunner, F.M., De-Juan-Pardo, E.M., Schiling, A.F., Schantz,J.T., van Griensven, M. and
Huntmacher, D.W. “Polylactides in Additive Biomanufacturing.”

Corresponding author:
Diana GREGOR-SVETEC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 3272, fax: +386 1 200 3270, e-mail: [email protected]

Co-author(s):
Urška STANKOVIČ ELESINI
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 3229, fax: +386 1 200 3270, e-mail: [email protected]

Mirjam LESKOVŠEK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 3266, fax: +386 1 200 3270, e-mail: [email protected]

Urška VRABIČ BRODNJAK

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 3282, fax: +386 1 200 3270, e-mail: [email protected]

Raša URBAS
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 3283, fax: +386 1 200 3270, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 36

 Oral presentations

COLORIMETRIC CHANGES CAUSED BY UV VARNISHING

Ivana CAR, Igor MAJNARIĆ & Branka LOZO

University of Zagreb, Faculty of Graphic Arts

ABSTRACT: This paper examines the influence of UV LED varnish on color change of varnished substrates. In-
dustrially manufactured dyed papers in 8 different colors (green, red, blue, cyan, yellow, brown, black and light
yellow) were UV varnished on Roland VersaUV LEC-300 printing machine in UV LED Inkjet printing technique.
Each sample of dyed papers was varnished in four different surface coverage in half-ton values ​​​​of 20%, 40%,
60% and 80%. Varnished papers were submitted to colorimetric analyses using X-Rite eXact spectrophotome-
ter. From the obtained L*a*b* values, the difference in coloration (ΔE2000) was calculated by CIE Lab ΔE2000
equation. The results of the colorimetric analyses were presented in the form of three-dimensional CIE Lab
diagrams showing the color changes in brightness and chromaticity. Color change ΔE2000< 1 is considered
visually unnoticeable in contrast to color change ΔE2000> 1 that is visually noticeable. Above the color change,
some samples showed significant change in chromaticity.

Keywords: color change, UV varnish, half-ton value, dyed papers.

1 INTRODUCTION

Paper, as the most important carrier of information in the past, has expanded its application. A large variety
of paper grades are produced to suit the special requirements. Paper can be impregnated, coated, laminated,
creped, molded etc (Holik, 2006.). One of the methods used for applying the impregnated layers on the surface
of the paper is UV Inkjet printing technology. In this technology, the transparent UV LED varnish in liquid form
is directly applied on the printing substrate and it is cured by LED UV light (Majnarić, 2012.). This paper analyze
how different surface coverage of the transparent UV LED varnish applied by UV Inkjet printing technology
influence on color change of varnished substrates. Varnished substrates used in the experiment are industrially
manufactured dyed papers in 8 different colors (green, red, blue, cyan, yellow, brown, black and light yellow).
Generally, materials are colored by the use of pigments or dyes that differ in their application. When a pig-
ment is used to color substrate, the finely divided, insoluble solid remains throughout the coloration process
(Lewis, 2006.). Dyes, in the other hand, penetrates the substrate in soluble form, after which it may or may not
become insoluble (Vernardakis, 2006).
By the Herbert Holik’s book »Handbook of Paper and Board«, dyes are the oldest class of synthetic specialty
chemicals employed in the production of paper. Anionic direct dyes are the dominant class of dyes which ac-
count for 52 % of the total worldwide turnover. Anionic direct dyes are the sodium salts of azo dyes containing
sulfo groups (for water solubility) or copper phthalocyanines which also contain sulfo groups. They have a
high affinity to bleached chemical pulps and often do not need additional fixing agents. The colors attainable
may not be as brilliant as with acid and basic dyes but the lightfastness is usually adequate. Basic dyes are the
salts (chlorides, hydrochlorides, sulfates, and oxalates) of color bases. Although they are tinctorially strong,
basic dyes have poor affinity to bleached furnishes, besides theirs mottling tendency and poor lightfastness.
Acid dyes are all water-soluble salts of colored organic acids which dissociate in water to form colored ani-
ons. They have little affinity to paper fibers. Cationic direct dyes retain the planar molecular structure of acid
direct dyes. Cationic groups, incorporated in the structure, accelerate their affinity to paper fibers. They are
moderately adsorbed on bleached lignin-containing stocks and produce fairly good bleed fastness. Fixative is
usually not required for these dyes.
For the most efficient usage, the dyes are mostly added continuously and fully automatically into the stock
flow. Stock or internal dyeing is the most widely used paper dyeing process. The choice of dyeing conditions
largely depend on the raw materials used in papermaking (recycled fibers, stone groundwood, TMP, CTMP,
unbleached or bleached chemical pulp, type and portion of filler) and on its preparation process e.g. a higher
degree of beating of the pulp results in a deeper coloring (Holik, 2006.).

The pH conditions are very important in stock dyeing and have tendentious toward paper production in the
neutral or alkaline pH range. These conditions need dyes with a very good affinity to the paper stock in a
neutral medium and/or very effective fixatives and retention aids.

8th Conference on Information and Graphic Arts Technology 37

 Oral presentations

Batch addition, as one of the stock dyeing processes, involve thorough mixing of the additives with the paper
stock which increase optimal fixation due to longer contact time between the fiber and dye. In the other hand,
the time required for color correction and color change is relatively long.
Continuous addition can attain the desired shade more quickly but have a lower color yield for intensely
colored papers (Holik, 2006.).

2 EXPERIMENTAL

Four types of digitalized printing forms with surface coverage in 20%, 40%, 60% and 80% of half-ton values ​​
were constructed for the needs of experiment.
They were used for LED UV varnishing of industrially manufactured dyed papers in 8 different colors (green,
red, blue, brown, cyan, black, yellow and light yellow) in four different surface coverage. Each standard printing
substrate used in the experiment had a slight difference in weight (gmin= 112,45 g/m2 for blue dyed substrate,
gmax= 121,05 g/m2 for green dyed substrate).
LED UV varnish was applied on the surface of industrially manufactured papers by LED UV printer Roland
VersaUV LEC 300 in UV LED Inkjet printing technique.
By UV printing of felt side (FS) four characteristic samples were made: prints with the surface coverage of
20%, prints with the surface coverage of 40%, prints with the surface coverage of 60% and the prints with the
surface coverage of 80%.
Varnished and not varnished papers were submitted to colorimetric analyses using spectrophotometer model
X-Rite eXact Advanced (mode M1, light condition D50, standard observer 2°). Measurements were made on
10 randomly chosen samples.
Program x-Rite Datachatcher were used for calculating output results such as CIE L*a*b*, c*, H*, x, y, Y, and
spectral reflection in the visible part of the spectrum. All values were measured three times and average val-
ues were used for further analysis. From the obtained L*a*b* values, the difference in coloration (ΔE2000) was
calculated by CIE Lab ΔE2000 equation. The results of the colorimetric analyses were presented in the form of
three-dimensional CIE Lab diagrams showing the color changes in lightness and chromaticity. Color change
ΔE2000< 1 is considered visually unnoticeable in contrast to color change ΔE2000> 1 that is visually noticeable.
Above the color change, some samples showed significant change in chromaticity.

Figure 1. Schematic presentation of the performed experiment.

8th Conference on Information and Graphic Arts Technology 38

 Oral presentations

3 RESULTS & DISCUSSION

In this work the analysis is directed to LED UV Inkjet printing technique which enables the coating of transpar-
ent LED UV varnish which can result in color change on the surface of printing substrates, besides the visual
effect of gloss. The color change is not the same for all dyed substrates neither is equal for the same amount
of applied varnish (Dolić, 2014.).
Based on the measurements, the results showed that for some dyed substrates (i.e. for green) varnished by
20% half-ton coverage the color change is remarkably smaller (ΔE0-20%= 1,42) than for 40% varnish coverage
(ΔE0-40%= 21,06) (Figure 2).
Red and blue dyed substrates showed the linear color change for subsequent increase of varnish coverage. For
a minimum varnish coverage of 20%, the color change was barely visible (ΔE0-20%= 1.07-1.09) for both samples.
For the maximum varnish coverage of 80% color change was visible (ΔE0-80%= 4.69). On 3D and 2D graphs the
color change of three differenty dyed substrates (green, red and blue) is presented. Each color change is the
result of 4 different ways of LED UV varnishing (varnish coverage of 20 %, 40 %, 60% and 80 % half-ton value)
and all the differences in color change are compared with the unvarnished samples.

Figure 2. Color changes of red, blue and green dyed substrates caused by UV LED varnishing with the differ-
ent surface coverage in half-ton values of 20%, 40%, 60% and 80%: a) 3D diagram of dyed substrates, b) 2D
diagram of dyed substrates.

The brightest colored substrate (light yellow) showed low but slightly visible color change with the maxi-
mum varnish coverage of ΔEmax= 3.96 (Figure 3). The black sample was the only one to keep the color change
at low level (ΔE0-80%= 1,39) for the maximum varnish surface coverage of 80% half-ton value. The common
characteristic of both substrates (black and light yellow) is that minimum varnish coverage of 20% half-ton
value is almost undetecteble.

Figure 3. Color changes of brown, light yellow and black dyed substrates caused by UV LED varnishing with
the different surface coverage in half-ton values of 20%, 40%, 60% and 80%: a) 3D diagram of dyed sub-
strates, b) 2D diagram of dyed substrates.

8th Conference on Information and Graphic Arts Technology 39

 Oral presentations

Same samples (brown and cyan) showed a change in chromaticity for different surface coverage of UV var-
nish (Figure 4). Besides the visible color change at 20% of the surface coverage (ΔE0-20%= 1.55), brown colored
substrate also showed the color change of the tone (chromaticity) with tendency in direction of -b coordi-
nate (blue). The largest change is noticeable between 40% and 60% of the surface coverage.
Cyan colored substrate showed the linear color change for subsequent increase of varnish coverage (ΔEmax= 6.65),
but its color change is already visible with minimum varnish coverage of 20% half-ton value (ΔE0-20%= 1.55).
Cyan also showed the color change of the tone with tendency in direction of +a coordinate (red).

Figure 4. Color changes of yellow and cyan dyed substrates caused by UV LED varnishing with the differ-
ent surface coverage in half-ton values of 20%, 40%, 60% and 80%: a) 3D diagram of dyed substrates, b) 2D
diagram of dyed substrates.

4 CONCLUSIONS

By application of UV varnish on industrially manufactured dyed papers most samples showed visible color
change (ΔE> 1) for varnish coverage of 80% half-ton value. Such effects are undesirable and could be adjusted
for each sample by application of lower surface coverage (20%, 40% or 60%) of UV varnish.
The green sample showed the highest color change visible in 40%, 60%, and 80% of surface varnish coverage
and is thus considered unsuitable for LED UV varnishing.
On the other side, all three samples (red, blue and cyan) with the linear color change showed barely visible
color change for a minimum varnish coverage of 20% (ΔE0-20%= 1,07-1,55). For the maximum varnish coverage
of 80%, color change was noticeable (ΔE0-80%> 4,5) and also unsuitable for LED UV varnishing in all three cases.
The black sample showed the slightest color change, barely visible even with application of maximum surface
coverage of 80% of UV varnish. It is thus considered to be the most suitable for LED UV varnishing applications.
Visible color change of some UV varnished samples can be also used to provide additional aesthetic effect.

5 REFERENCES

• Holik, H. (Ed.), 2006. Handbook of Paper and Board. Weinheim: WILEY-VCH Verlag GmbH & Co. KGaA
• Majnarić, I., Bolanča Mirković, I., Golubović, K., 2012. „Influence of UV curing varnish coating on surface
properties of paper“ Paper presented at the Tehnički Vjesnik-Technical Gazette (1330-3651) 19 (2012), 1; 51-56
• Lewis, A. P., 2006. „Colored Inorganic Pigments“ In Book Coatings Technology Handbook, Third Edition,
Tracton A. A., 28 Jul 2005, Taylor & Francis Inc
• Vernardakis, G. T., 2006. „Pigment Dispersion“ in Coatings Technology Handbook, Third Edition, Tracton A.
A., 28 Jul 2005, Taylor & Francis Inc
• Dolić, J., Pibernik, J., Majnarić, I., 2014. „Influence of UV Varnish Pattern Effect on Print Quality“ Paper pre-
sented at The Journal of imaging science and technology (1062-3701) 58 (2014), 6; 60501-1

Corresponding author:
Ivana CAR
University of Zagreb, Faculty of Graphic Arts,
Department of Materials in Graphic Technology
Getaldićeva 2, 10 000, Zagreb, Croatia
phone: +385 95 544 47 83, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 40

 Oral presentations

Co-author(s):
Igor MAJNARIĆ
University of Zagreb, Faculty of Graphic Arts,
Department of Printing
Getaldićeva 2, 10 000, Zagreb, Croatia
phone:+385 91 288 20 29, e-mail: [email protected]

Branka LOZO
University of Zagreb, Faculty of Graphic Arts, Department of Printing
Getaldićeva 2, 10 000, Zagreb, Croatia
phone: +385 98 480 864, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 41

 Oral presentations

COMPUTER GENERATED HANDWRITING SIMULATION BASED

ON ALTERNATION OF CHARACTERS

Mihael Aleksander MAHKOVIC, Klementina MOŽINA

University of Ljubljana, Faculty of Natural Sciences and Engineering, Department of Textiles, Graphic Arts
and Design, Chair of Information and Graphic Arts Technology, Snežniška 5, SI-1000 Ljubljana, Slovenia

ABSTRACT: The aim of this research was the digitalisation of a handwriting with the emphasis on designing
several variations of a single individual letter. The development of a typeface that would successfully simu-
late a human handwriting was not possible in the past. OpenType enabled the creation of the so-called smart
typeface. The most important features are the alternate substitution and the ligature substitution. We aspired
to create a typeface supported both by professional and less advanced text editing programs. Therefore, our
algorithm needed to be formatted into two functions only: the function for contextual alternates and the lig-
ature function. Eva typeface consists of 1243 characters: 87 capital letters, 784 lowercase letters, 237 ligatures,
30 digits, 59 punctuation marks and 46 symbols. No intervention of the user needed, the font alternates among
formative variations of individual characters in professional and less advanced text editing programs alike.

Keywords: character, handwriting, OpenType, typeface.

1 INTRODUCTION

A double alphabet, consisting of capital and lowercase letters, represents the basis for a unique human hand-
writing that an individual begins to learn in their childhood. Over the years, the handwriting changes and
evolves, finally forming a unique design that (to a certain extent) reflects the writer’s personality. A literate
human being develops a unique collection of characters sharing certain features. Due to the “hand-written”
nature of a human script, its characteristics are never used in an identical design. The key characteristic of a
handwriting is therefore minor inconsistency, which is also an important tool to identify an actual handwriting
from its digitalised version.
The aim of this research was the digitalisation of a handwriting with the emphasis on designing several var-
iations of a single individual letter. For a better understanding of letter-design, the development of Roman
lettering (Možina, 2003) and of individual handwriting was researched (Trstenjak, 1986).
In typography, the development of digitalised typefaces induced the tendency to produce digitalised hand-
writings. Several such versions of handwritings exist; however, very few of them employ a random alternation
among several versions of an individual character and thus succeed to imitate the actual human handwriting.
The majority of digitalised handwritings include a single version of an individual character, which reveals its
digitalised nature (Bear, 2017).
The development of a typeface that would successfully simulate a human handwriting was not possible until
OpenType, as earlier file-formatting did not support a large quantity of characters inside a single file. Open-
Type and its different functions have enabled the creation of the so-called smart typeface (Adobe, 2017; Leming,
2017a). The most important OpenType features that contributed to the development of a quality handwriting
digitalisation are two subspecies of OpenType substitution function, i.e. the alternate substitution and the
ligature substitution (Leming, 2017a). There are four categories in the placement of the substitution function,
namely an original point of the letter and the point of progression. Each point bears the value X and value Y,
which can be controlled within the function. They can be written as a move of a certain letter, as the adaptation
of space between two characters and a meaningful move of a letter, i.e. a move that happens under certain
conditions (a preliminary character or a combination of characters) (Leming, 2017b). Visually, the substitution
is the most perceptible change the function can make in a text. Two categories represent the foundation of
each substitution, i.e. target and substitute. The same as there are several types of functions to change letter
placement, there are also many types of substitution function. These are basically the substitution of one target
with one substitute, the substitution of several targets with one substitute and the substitution of one target
with several substitutes (Leming 2017b). In our research, a contextual substitution is of great importance. It
is a simple substitution of one target (or several) with one substitute (or several) that occurs when specific
conditions are met (Leming, 2017b). In OpenType, randomness is a long-desired function that has not been
achieved yet (Leming 2017c). The “randomness” simply means a complex order that can be hidden by includ-
ing several classes of letters, alternating under certain conditions. Another prominent OpenType deficiency
is the order of code writing. OpenType code structure is extremely inflexible and enables only sequential rule
writing – as opposed to branching in commands (Seinfert 2017).

8th Conference on Information and Graphic Arts Technology 42

 Oral presentations

2 EXPERIMENTAL

2.1 Methods and materials

Our first step was to define the requirements for the handwriting. It had to be at least partially connected
(between letters in words), visually interesting, not too common and it had to reflect the author’s style. Fur-
thermore, the author of the handwriting had to be available to provide us with a suitable quantity of sample
text. The next step was to provide sample texts in English and Slovenian. For the English sample, we used a
text generator (Bibakis, 2017). The chosen lettering was the script of a dramaturg Eva Mahkovic. To translate
the sample text into a digital format, we used Nikon D5000. For further processing, we used Adobe Photoshop
CC (to improve the contrast and achieve better visibility) and Adobe Illustrator CC (to design single weight
letter shapes). Moreover, we used Adobe Illustrator CC to design curve outlines. The letter curve outlines were
then imported into Glyphs (2.2.2. version), where the typeface was completed. Our equipment was a MacBook
Pro (Retina, mid 2015) with an additional screen (21.5 inch, mid 2011).

2.2 Designing smart typeface

The first step in designing a smart typeface was to analyse the handwriting. The results of our analysis were
as follows:
• The handwriting is mostly connected between letters within a word (lowercase letters), yet it includes ran-
dom interruptions that follow no perceptible rule. We grouped the letters according to similar connective
links. The linkage appears mostly in the x-height and can be divided into two groups, i.e. upper (top of
x-height) and lower linkage (base of x-height). The third option is for the letter not to have any link with
previous or following letter at all.
• The handwriting does not link the capital and lowercase letters. The capital letters are independent and
mostly resemble the well-known Latin alphabet capital letter design with little or no deviation.
• The third specificity of the handwriting is the diversity of a certain letter design. The main characteristic of
each handwriting is that two letters are never the same; however, they still share common characteristics.
Nevertheless, a particularity of our chosen handwriting is that at certain letters, several forms exist for the
same character.
The findings of our visual analysis were important to structure the architecture of OpenType functions. To
achieve the best possible approximation to an actual handwriting, we decided to form three versions of each
character.
To design individual characters, we used a sample text. Using Adobe Illustrator CC, we drew single weight
versions of letters and characters, and set the stroke weight to 1 pt. The typeface weight remained the same
throughout the typeface (sample text was written using a single weight pen). The character outlines were then
imported into Glyphs. We used Glyphs to further design the characters and achieve their final form – through
correction of letter connective links and curve orientation.
The capital letters being independent with no link to the letters that follow, we classified them into three
groups, i.e. one for each version of the letter. With lowercase letters, we had to take into account the option
of linkage (upper, lower or no linkage). We classified lowercase letters into 27 groups. If each lowercase let-
ter anticipates three linkage options, the final number of combinations is nine (attention should be paid to
the link with the previous and the following letter). As we wanted the number of letters to be as diverse as
possible, we included further three versions of each letter for all nine combinations. In classifying numbers
and other characters, we followed the same rule as in classifying capital letters. The numbers and other char-
acters (symbols, punctuation marks etc.) stand alone, i.e. independently, and thus need to be classified only
into three groups.
The majority of an OpenType code designed typeface is in the function of contextual substitution (GlyphsApp
2017; Microsoft Corporation, 2017; Typotheque, 2017). Figures 1–3 present only certain codes, namely Figure
1 presents the vowel and consonant rotation code (capital letters). Figure 2 presents the lowercase letter
substitution function which occurs between letters in the middle of words. Figure 3 presents the capital and
lowercase letter combination substitution function.

8th Conference on Information and Graphic Arts Technology 43

 Oral presentations

Figure 1. OpenType vowel and consonant rotation code (capital letters).

Figure 2. Lowercase letter substitution function between letters which occur in the middle of words.

Figure 3. Capital and lowercase letter combination substitution function.

3 RESULTS & DISCUSSION

Eva smart typeface consists of 1243 characters: 87 capital letters, 784 lowercase letters, 237 ligatures, 30 digits,
59 punctuation marks and 46 symbols. The choice of characters enables the use in the Slovenian and English
language. The structure of OpenType formatting is designed for the typeface to be useful in professional and
less advanced text editing programs alike. Compared to (the majority of) other digitalised handwritings on the
market, our typeface includes a greater number of diverse characters and several versions of a single character.
The result is better quality simulation of a human handwriting. The advantage is best observed in lowercase
letters where each letter appears in 27 different versions.
Capital letters: There were no significant specifics in design. As capital letters have no connections to the letter
before or after, the digitalised version closely follows the original handwriting design. A change was required
only in the case of diacritic letters (Č, Š, Ž) and different versions of letter Y. At former, the placement of the
caron had to be changed due to the restriction of ascender height.
Lowercase letters: The design of lowercase letters differed from the original handwriting, which was due to
the need for a link between two letters at the same height for all lowercase letters. The unification resulted in
the deviation in the letter design, yet the change was (at the majority of letters) so small that it did not affect
the typeface character. Figure 4 presents all 27 versions of the letter “h”.

8th Conference on Information and Graphic Arts Technology 44

 Oral presentations

Figure 4. Design variations of lowercase “h”.

Ligatures: Our typeface includes a number of ligatures that solve the problem of the upper linkage point
between certain letter combinations. The ligature design closely resembles the letter combination form of
the original handwriting (Figure 5). As we had no restrictions in design, we were able to closely follow the
interesting linkage between letters we noticed in the visual analysis of the original handwriting. An interesting
linkage (designed as perceived in the original handwriting) is for example a ligature of lowercase “o” and “z”
(Figure 6). The formatted ligatures enable the digitalised handwriting to feel even more genuine.

Figure 5. Different ligatures for “b” and “i” combination.

Figure 6. Linkage in ligature of “o” and “z”.

Digits: Similar to capital letters, digits also stand alone. Formatting digits, we had to be careful with the ones
that went above the ascender height or below the descender height. Such an example was the digits 3 and 7. To
avoid cutting away a part of those characters’ outlines, we changed their form so that they fit into the type grid.
Punctuation marks and other characters: Punctuation marks were designed without changing the original
form of the handwriting. The only exception was the brackets. Their form was altered to fit into the type grid.
To make the typeface as useful as possible, we also designed various other characters. Their design follows
the characters of the original handwriting with a minor adaptation in their size.
Comparing the original handwriting with the digitalised Eva typeface reveals relative success of the digital-
isation (Figure 7). Despite the fact that there is still room for some improvements, the Eva typeface imitates
the original handwriting well.

8th Conference on Information and Graphic Arts Technology 45

 Oral presentations

Figure 7. Comparing original (up) handwriting and digitalised version (down).

4 CONCLUSIONS

In typography, OpenType functions have opened up many new possibilities that are now widely used by ty-
pographers all around the world. Our research has brought a digitalised version of a human handwriting that
strives towards the closest possible approximation of the original handwriting. As two characters of a human
handwriting are never exactly the same, such a task demands designing a great number of characters that
would enable a quality simulation of a handwritten typeface. To achieve the typeface functionality, we had
to define a function for an automatic letter variation alteration without the user’s intervention. Therefore, we
used one of the most important OpenType formatting functions, i.e. the substitution function.
As we wanted to create a typeface that could be used by both professional and amateur users, a correct Open-
Type functions code structure had to be provided – the one that can be used by professional and simpler text
generators alike. In the designing of the OpenType rules, we used two substitution function subspecies, namely
the contextual alteration function and ligature function.
In our opinion, an interesting upgrade of our research would be software that enables an automatic design of
an individual human handwriting. Current handwriting generators (Calligraphr, 2017; Your Fonts, 2017) are
severely limited and enable the user to produce a non-connective typeface with a single variation per character.
The development of software that would enable an automatic production of connective and non-connective
human handwriting with several variations per character would be progress in typography.
Our typeface Eva consists of 1243 characters, i.e. 87 capital letters, 784 lowercase letters, 237 ligatures, 30 dig-
its, 59 punctuation marks and 46 symbols. No intervention of the user needed, the typeface alternates among
formative variations of individual characters in professional and less advanced text editing programs alike,
thus creating a quality digitalised simulation of a handwriting.

5 REFERENCES

• Adobe. OpenType. URL: http://www.adobe.com/products/type/opentype.html (last accessed on 6. 3. 2017).

• Bear, J. H. Script fonts mimic handwriting styles. URL: https://www.thoughtco.com/script-font-informa-
tion-1073829 (last accessed on 17. 6. 2017).
• Bibakis, V. Random Text Generator. URL: http://randomtextgenerator.com/ (last accessed on 11. 3. 2017).
• Calligraphr. Transform your handwriting or calligraphy into a font. URL: https://www.calligraphr.com/?r-
tom=muscriptfont (last accessed on 18. 6. 2017).

8th Conference on Information and Graphic Arts Technology 46

 Oral presentations

• GlyphsApp. Features, Part 3: Advanced Contextual Alternates. URL: https://glyphsapp.com/tutorials/fea-

tures-part-3-advanced-contextual-alternates (last accessed on 15. 3. 2017).
• Leming, T. Introduction. URL: http://opentypecookbook.com/index.html (last accessed on 6. 3. 2017a).
• Leming, T. Rules. URL: http://opentypecookbook.com/rules.html (last accessed on 8. 3. 2017b).
• Leming, T. Common Techniques. URL: http://opentypecookbook.com/common-techniques.html (last ac-
cessed on 15. 3. 2017c).
• Microsoft Corporation. Advanced Typographic Extensions – OpenType Layout. URL: https://www.microsoft.
com/typography/otspec/TTOCHAP1.htm (last accessed on 15. 3. 2017).
• Možina, K. 2003. Knjižna tipografija. Ljubljana: University of Ljubljana.
• Seinfert, G. 3–17 January 2017. “Handwriting font & OpenType”, talk on web site. Available on URL: https://
forum.glyphsapp.com/t/handwriting-font/5348/18.
• Trstenjak, A. 1986. Človek in njegova pisava. Ljubljana: CZNG.
• Typotheque. OpenType feature support in applications. URL: https://www.typotheque.com/fonts/open-
type_feature_support (last accessed on 15. 3. 2017). Your Fonts. Make Your Own Handwriting Fonts. URL:
http://www.yourfonts.com/ (last accessed on 18. 6. 2017).

Corresponding author:
Klementina MOŽINA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 42, fax: +386 1 200 32 70, e-mail: [email protected]

Co-author:
Mihael Aleksander MAHKOVIC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 42, fax: +386 1 200 32 70, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 47

 Oral presentations

DEFINING THE OPTIMAL TRAINING SET SIZE FOR CAMERA-BASED COLOUR

MEASUREMENT OF GONIOCHROMATIC PRINTS: A PRELIMINARY STUDY

Ivana TOMIĆ, Sandra DEDIJER, Dragoljub NOVAKOVIĆ & Ivan PINĆJER

Department of Graphic Engineering and Design, Faculty of Technical Sciences,
University of Novi Sad, Novi Sad, 21000, Serbia

ABSTRACT: In order to be used as a colour measuring device digital camera needs to be characterized, i.e.
a function that maps image RGB data to the colourimetric values had to be established. This function is
commonly formed by using the set of samples with the known colourimetric values (training set). The aim
of this study was to define the optimal size of a training set (minimum number of samples that enables the
satisfactory accuracy of colour measurement) for camera-based colour measurement of prints that exhibit
goniochromism. Camera characterization was performed by Artificial Neural Networks, where the accuracy of
colour measurement was assessed for the networks trained with a different number of samples and for three
measuring geometries. The results showed that the optimal size of a training set depended on the measuring
geometry. In general, the best results were obtained for training sets with no more than 200 samples.

Keywords: pearlescent pigments, camera, colour measurement, goniochromism.

1 INTRODUCTION

Colour measurement based on the use of digital cameras (often called colour estimation) relies on camera
characterisation – the process of forming a function for converting camera RGB data to colourimetric values
[1, 2]. In practical application the most frequently used is empirical characterisation, where the link between
RGB and colourimetric data is derived from the samples with the known colour values (so-called “training
set”) [1]. The applicability of created function is afterwards assessed on a different set of samples, which are,
ideally, not used for training (the “test set”).

The accuracy and complexity of empirical characterisation are highly dependent on the method used to form
characterization function, as well as the number and choice of samples within the training set. It was shown
[2-4] that the most accurate estimation of colorimetric values is achieved with polynomial regression and
Artificial Neural Networks (ANN). The abovementioned two methods led to almost identical results when
applied to the same set of samples [3]. The choice of the samples in the training set depends on the specific
application. In general use, it is desireable to choose the minimum number of different colours that evently
covers the gamut of a camera [5].

Regarding the number of samples, more accurate results are evidently achieved if more samples are used for
training. However, a higher number of samples increases the complexity of characterization. Hence, the op-
timal size of a training set is normally determined as a minimum number of samples that produce the satis-
factory accuracy of colour measurement. The influence of the number of samples on the estimation accuracy
was investigated extensively in case of the products printed with absorption pigments [2, 3, 6]. In this work,
we were interested in the camera-based measurement of prints enhanced with pearlescent pigments. Due
to their ability to provide goniochromatic effect (change of perceived colour with the change of a viewing
or illumination angle) [7] pearlescent pigments are often used in the printing (for luxurious packaging and
special effect prints).

Since pearlescent pigments alter the optical properties of the printed samples, we assumed that the accuracy
of the camera-based colour measurement of goniochromatic prints would be lower in comparison to the meas-
urement of samples printed with conventional pigments. To prove this assumption, we conducted a preliminary
study to evaluate to which extent the optimal size of a training set changes when the printed samples whose
colours are estimated by the camera are enhanced with pearlescent pigments. Considering that the appearance
of the goniochromatic samples changes with the alteration of a viewing angle, camera-based measurement
was performed in a way to mimic required measuring geometries [8], as explained in the following section.

8th Conference on Information and Graphic Arts Technology 48

 Oral presentations

2 EXPERIMENTAL

Colours of AGFA IT8 7/2 test chart with the modified gray scale that contained 26 instead of 24 colours (290
colours in total), were printed on 300 g/m2 matte art paper on Xerox DocuColor 252 digital printer. After dry-
ing, they were overprinted with the pearlescent pigments (Iriodin® Rutile Feinbleu) dispersed in aqueous ink
base (percentage of the pigment in the mixture was 10). For overprinting, we used screen printing technique
(43 l/cm), where the mixture of pearlescent pigment and an ink base was applied to the prints three times.

Printed and enhanced patches were measured with Gretag Macbeth Auto-Eye 640 gonio-spectrophotometer
(D65/10°). Following the recommendations of [8], CIELAB values were obtained for three measuring geom-
etries – 45°/asp15°, 45°/asp45°, 45°/asp75° (“asp” denotes the distance of the viewing angle from specular
reflection). The values served as ground truth for camera characterization and for evaluating its accuracy.

The digital camera used in the experiment was Canon EOS 550D with 18-55 kit lens. To mimic the gonio-spec-
trophotometer setup, fluorescent light source (D65 simulator) was positioned in 45° from the normal to the
sample (Figure 1), while the position of the camera was changed to the aspecular angles of 15°, 45° and 75°
by using of a holder designed to enable the rotation of a camera (sample to camera distance was fixed to 40
cm). The adequate exposition was determined by light meter Seconik L-758DR DigitalMaster (Figure 1) and
was not changed during the capturing. Capturing the samples was performed in a dark room. Each printed
sample was captured separately and recorded as a RAW file.

Figure 1. Using the light meter to determine the adequate camera setup.

RAW files were further processed in MATLAB® (R2011a) with dcraw open source software. Only demosaicking
and dark-current noise removal were performed, and average RGB values of 400x400 pixels from the center of
an image were taken into account as inputs to the characterization algorithm.

Camera characterisation model was based on the Artificial Neural Networks as in [3, 9]. Separate ANNs were
formed for each of the measuring geometries. Each net had 3 nods in the input layer (RGB values obtained
from the images), 18 nods in the hidden layer and 3 nods in the output layer (corresponding CIELAB values,
obtained by gonio-spectrophotometer). The total number of samples was divided randomly into three parts
– training, validation and test set in the following ratios: 3:6:1, 5:4:1, 5:4:1, 6:3:1, 7:2:1 and 8:1:1, similar to [3].
Such an approach enabled changing the number of the samples in training set (87, 116, 145, 174, 203 and 232,
respectively) while keeping the test set size unchanged (29 samples). The validation set was used not only
to obtain the desired number of the samples in training and test set but also to prevent the overfitting and
to increase the generalization ability of a net. Therefore, the training was always stopped at the moment the
validation error (MSE) started to increase. The total performance of a model was assessed by evaluating the
errors over the test set.

In order to reduce the influence of random samples’ selection on the results of estimation, training and test-
ing procedure was repeated 100 times (for each of the defined ratios). In each repetition, we obtained ΔE*ab
colour differences between estimated (values obtained by ANN) and measured (obtained by gonio-spectro-
photometer) CIELAB values of the chosen samples. The average ΔE*ab error over 100 repetitions, as well as the
average of the maximum and minimum errors [9], were used as measures of accuracy of camera-based colour
measurement. To determine the optimal size of a training set we observed the change of average errors with
the increase of the number of samples in the set. If there were no significant change in the estimation errors
after a certain number of samples, that particular number was regarded as optimal size of a training set.

8th Conference on Information and Graphic Arts Technology 49

 Oral presentations

3 RESULTS

The average values of the mean estimation errors over 100 repetitions for ANNs trained with a different number
of samples are presented in Figure 1, while the corresponding standard deviations are shown in Table 1. Figure
2 contains the average of the minimum and maximum errors. The results are presented for all three measuring
geometries (viewing angles).

Figure 2. Average of the mean estimation errors over 100 repetitions

with respect to the training set size and viewing angles.

Table 1. Standard deviation of the average ΔE*ab values over the 100 repetitions.
Training set size (number of samples)
Viewing angle 87 116 145 174 203 232
15° asp 1.53 0.79 0.33 0.45 0.33 0.37
45° asp 1.27 0.71 0.59 0.59 1.06 1.03
75° asp 2.11 2.07 1.24 2.48 1.26 1.28

Figure 3. Average of (a) minimum and (b) maximum estimation errors over 100 repetitions
with respect to the training set size and viewing angles.

From the presented results it is clear that the estimation errors decreased when the number of samples in the
training set increased (as expected). The results obtained for three viewing geometries were slightly different,
where the highest errors in almost all the cases were noticed for measuring geometry 45°/asp75°.

By observing the average estimation errors and the corresponding SD values, it can be seen that errors for the
samples measured in the viewing angles of 15° and 45° asp were quite similar and pretty much stable when
the number of samples in the training set exceeded 150. The average of minimum estimation errors followed
the same trend (Figure 3a), while for the maximum errors there were no significant changes when training set
contained more than 200 samples (Figure 3b). Mean estimation errors corresponding to the viewing angle of
75° exhibit slightly higher variations in comparison to the rest of the measuring geometries. This is especially
the case for the average of the maximum errors as seen in Figure 3b. We believe that this changes can be related
to the random selection of the samples within the training set.

8th Conference on Information and Graphic Arts Technology 50

 Oral presentations

4 DISCUSSION

In their work [3], Cheung et al. concluded that the accuracy of camera-based colours measurement where
ANNs were used for camera characterization does not improve significantly if the number of the samples in
the training set exceeds 100. In the aforementioned work, the position of the camera was not changed (meas-
uring geometry 0°/45°) since the camera was used to determine the colour of the samples that do not exhibit
goniochromism.

By analyzing the results presented in this work, it can be concluded that for assessing colours of samples
printed with the chosen pearlascent pigment there was no significant change in estimation erros if the number
of samples in the training set excedeed 150 (for measuring geometries 45°/asp15° and 45°/asp45°) or 200 (for
measuring geometry 45°/asp75°). If all the errors were taken into account (average of the mean, maximum
and minimum for all the three measuring geometries) differences in the estimation accuracy were insignificant
when a number of samples in the training set exceeds 200.

In this work, we used the same methodology for chosing the samples in the training set, as well as the test
chart of almost identical colour gamut as in [3]. Hence, it can be concluded that in order to achive the same
degree of accuracy of camera-based colour measurement of goniochromatic prints in comparisson to the con-
ventional, trainig set had to contain double the number of samples. This result confirms the assumption that
the goniochromatic pigments lower the accuracy of camera-based colour measurement of the printed samples.

5 CONCLUSION

The development of more accurate sensors and the new techniques for camera characterisation made digital
cameras viable solution for colour measurement of printed products. In case of empirical camera characteri-
zation, the accuracy of the colour measurement can directly be related to the number of the samples (training
set) used to form the function that transforms RGB to CIELAB values. The goal of our work was to define the
optimal size of the training set for camera-based colour measurement of prints that exhibit goniochromism.
Due to the complexity of printing and capturing process, we relied on assessing the colourimetric values for
printed samples enhanced with only one type of goniochromatic (pearlescent) pigment. Camera characteri-
zation model was based on Artificial Neural Networks, and the accuracy of colour measurement was assessed
for the networks trained with a different number of samples. Since the goniochromatic samples require mul-
ti-angular measurement, colourimetric values were estimated for three measuring geometries.

The results showed that the pearlescent pigments’ coating lowers the accuracy of camera-based colour meas-
urement. The optimal size of a training set was shown to be 200 samples – which is double in comparisson to
the optimal set for conventional prints. It was also shown that the optimal set size depended on the measuring
geometry. In particular, more samples were needed to accurately estimate colourimetric values correspond-
ing to the larger aspecular viewing angles (geometries where the position of the detector is farther from the
specular reflection). Since in this work we used only one type of pearlescent pigments, further experiments
are needed in order to generalize presented findings.

ACKNOWLEDGEMENTS
The research is supported by the Ministry of Education, Science and Technology Development of the Republic
of Serbia, project number: 35027 “Development of software model for scientific and production improvement
in graphic industry”.

6 REFERENCES

1. BALASUBRAMANIAN, R.: Device Characterization, In Digital Color Imaging Handbook, CLC Press, ISBN
9780849309007, Boca Raton, (2003), pp. 281-395.
2. HONG, G., LUO, M. R. & RHODES, P. A.: A study of digital camera colorimetric characterization based on
polynomial modeling, Color Research &Application, Vol. 26 (2001) No. 1, pp. 76-84, ISSN 1520-6378.
3. CHEUNG, V. et al.: A comparative study of the characterization of color cameras by means of neural networks
and polynomial transforms, Coloration Technology, Vol. 120 (2004) No. 1, pp. 19-25, ISSN 1478-4408.
4. LIU, Y. et al.: Camera characterization using back-propagation artificial neutral network based on Munsell
system, Proceedings of SPIE 6621, International Symposium on Photoelectronic Detection and Imaging 2007,
ZHOU, L. (Ed.), pp. 6621 - 6621 - 7, ISBN: 9780819467638, Beijing, China, SPIE, Bellingham (2008).

8th Conference on Information and Graphic Arts Technology 51

 Oral presentations

5. CHEUNG, T. L. V., WESTLAND, S.: Methods for Optimal Color Selection, Journal of Imaging Science and
Technology, Vol. 50 (2006) No. 5, pp. 481-488, ISSN 1943-3522.
6. DE LASARTE, M. et al.: Influence of the Number of Samples of the Training Set on Accuracy of Color Meas-
urement and Spectral Reconstruction, Journal of Imaging Science and Technology, Vol. 54 (2010), No. 3, pp.
30501-1-30501-10, ISSN 1943-3522.
7. MAILE, F. J., PFAFF, G. & REYNDERS, P.: Effect pigments—past, present and future, Progress in Organic Coat-
ings, Vol. 54 (2005) pp.150–163, ISSN 0300-9440.
8. ASTM E2539 - 12: Standard Practice for Multiangle Color Measurement of Interference Pigments, ASTM In-
ternational, West Conshohocken, 2014. Available from: http://www. astm.org; Accessed: 2015–06–11.
9. TOMIĆ I. et al.: Camera Characterization for Colorimetric Assessment of Goniochromatic Prints, Journal of
Imaging Science and Technology, Vol. 61 (2017) No. 2, pp. 20502-1-20502-15, ISSN 1943-3522.

Corresponding author:
Ivana TOMIĆ
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21 000 Novi Sad, Serbia
phone: +381 21 485 2626, fax: +381 21 485 2620, e-mail: [email protected]
http://www.ftn.uns.ac.rs/1417835183/ivana-tomic

Co-author(s):
Sandra DEDIJER
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21 000 Novi Sad, Serbia
phone: +381 21 485 2623, fax: +381 21 485 2620, e-mail: [email protected]
http://www.ftn.uns.ac.rs/1417834029/sandra-dedijer

Dragoljub NOVAKOVIĆ
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21 000 Novi Sad, Serbia
phone: +381 21 485 2620, fax: +381 21 485 2620, e-mail: [email protected]
http://www.ftn.uns.ac.rs/n1062641918/dragoljub-novakovic

Ivan PINĆJER
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21 000 Novi Sad, Serbia
phone: +381 21 485 2624, fax: +381 21 485 2620, e-mail: [email protected]
http://www.ftn.uns.ac.rs/1417835088/ivan-pincjer

8th Conference on Information and Graphic Arts Technology 52

 Oral presentations

DIFFERENCES BETWEEN DESIGNERS’ AND USERS’ PERCEPTION

OF INSTRUCTIONS DESIGN

Dorotea KOVAČEVIĆ, Maja BROZOVIĆ

University of Zagreb, Faculty of Graphic Arts, Getaldićeva 2, 10000 Zagreb, Croatia

ABSTRACT: Although professional designers are trained to create appealing visualizations, their perception
of attractiveness does not necessarily have to match users’ perception. This study investigated the disagree-
ment between those who design instructions and those who use them. We manipulated different samples of
instructions designs to test their attractiveness by two groups: professional graphic designers and laypeople.
Attractiveness was measured by Likert scale responses. The results indicated that two groups evaluated instruc-
tions differently. While designers preferred linear and simple instructions design, laypeople were attracted by
the colourful designs made of planes. Implications for creating visually appealing instructions were discussed.

Keywords: Instructions, attractiveness, designer, user.

1 INTRODUCTION

Large amount of consumer products comes with instructions for use, whether printed on the packaging or
as inserted leaflet. Information provided by these instructions enhances the effectiveness of handling the
product and, consequently, increases users’ satisfaction. Although graphic presentation of instructions should
encourage unhindered information processing, unfortunately, it happens that these messages sometimes fail
to communicate relevant messages due to poor legibility or design-related features. Visual attractiveness of
instructions design is important. Making instructions aesthetically appealing increases the possibility that
users will pay attention to them. Even more, attractive design can play significant role in evoking positive
impressions about the product (Magnier, 2016; Wang, 2013).
By deliberate application of design principles, experts are able to make aesthetically pleasing and easy to
follow instructions (Agrawala, 2003). Professional graphic designers are trained to create visually appealing
graphic presentations that are legible and easy to interpret. However, their process of designing is often guided
by their own preferences. This is not surprising, since every kind of graphical presentation depends not only
on the characteristics of presented information and the audience, but also on the presenter’s objectives and
her or his preferences, as reported by Tractinsky and Meyer (1999). Thus, it is likely that designers’ perception
of attractiveness does not necessarily have to match users’ perception of attractiveness. Quispel and Maes
(2014) demonstrated this in their study of data visualizations. Some other studies also showed discrepancy
between designers’ and users’ perception. For example, Hsu et al. (2000) investigated the differences between
designers and users in perceiving telephone design. Their results showed that designers prefer elegant style,
while users appreciate modern design. In the study of Vogt and Magnussen (2007), different pictures were
viewed by two groups of participants – artists and artistically untrained people. The results indicated two
groups used different viewing patterns, which suggests that experience in art (or design) affects the way ob-
server perceive visual messages. Another study of Bonnardel et al. (2011) also demonstrated some differences
between designers of websites and users. This motivated us to investigate the possible disagreement between
those who design instructions and those who consult the instructions.

2 EXPERIMENTAL

Our investigation included two types of participants’ responses. One part of experiment was subjective eval-
uation of different samples of instructions design. Another part was reporting the design features mentioned
by the participants while describing each of the instructions designs.

2.1 Participants

40 volunteers participated in the study. Their ages ranged from 21 to 41 years of age (M = 28.983, SD = 6.02). 65%
of the participants were female and 35% were male. One group consisted of 20 professional graphic designers
while another group included 20 laypeople without any kind of design experience. The selection criteria for
choosing designers required that participants have formal design education and have at least one experience
with the design of instructions.

8th Conference on Information and Graphic Arts Technology 53

 Oral presentations

2.2 Samples

We used four different samples of instructions to test their attractiveness by the two groups of participants.
Tested instructions differed according to two factors: graphical elements and a theme. Both factors varied at
two levels. Graphical elements used for the instructions design were whether lines (linear design) or planes
(flat design). Themes referred to the type of consumer product: pasta or cleaning spray. Samples of instruc-
tions are shown on Figure 1.

Figure 1. Designs of instructions used as stimuli: 1 – linear/pasta,

2 – linear/spray, 3 – flat/spray, 4 – flat/pasta.

2.3 Procedure

The experiment was conducted in a laboratory cabinet to ensure constant viewing conditions across the par-
ticipants. Each of the participants evaluated the samples individually. Samples were presented one by one.
The presentation time was not time-limited. The samples of instructions were presented on screen, on Lenovo
computer display (model LEN L1900pA) with the viewing distance of 60 (+/–1) cm. Using 7-point Likert scale
(7 = very attractive, 1 = unattractive), participants evaluated the attractiveness of each instructions design.
This type of evaluation is commonly used in testing the attractiveness (Barlow, 1991; Monk, 2007). Participants
were also encouraged to mention which design features made instructions especially attractive or unattractive.
The experimenter recorded the answers and later categorised them in a table.

8th Conference on Information and Graphic Arts Technology 54

 Oral presentations

3 RESULTS

Although designs with planes were rated as slightly worse (M = 5.26, SD = 1.49) than liner design (M = 5.33,
SD = 1.09), the repeated-measures ANOVA analysis showed no significant difference among two structural
elements, F(1,79) = 0.10, p = 0.75. The repeated-measures ANOVA analysis with respect to theme of instructions
found no significant difference between instructions for preparing pasta and instructions for using a cleaning
spray F(1,79) = 0.89, p = 0.35.
A paired t-test showed that the two groups of participants rated instructions designs differently t(79) = 2.64,
p < 0.01.The mean rate given by the group of non-designers was M = 5.60, SD = 1.15, while the designers were
more critical with the mean rate M = 4.98, SD = 1.38.
Group means and standard deviations of participants’ rates are shown in Figure 2. The group mean results
indicate that designers evaluated linear instructions for spray as the best design (M = 6.00, SD = 0.1.12), while
non-designers preferred planar design of instruction for pasta (M = 6.50, SD = 0.76).
Table 1 summarizes the design features mentioned by the participants during the evaluation process. Each
feature is put in the tables only if mentioned more than twice.

Figure 2. Mean rates for the instructions design across the participant groups.

Table 1. The most common design features mentioned by the participants while describing each of the in-
structions designs.
Non-designers Designers
Instructions
design Attractive Unattractive Attractive Unattractive
features features features features
clean
Linear - pasta simple outdated simplified
appealing
indistinct clean unnoticeable
Linear - spray light
plain elegant
attractive
uniformed
Flat - pasta modern cluttered
cohesive
stylish
modern
dark
Flat - spray eye-catching
cluttered
colourful

8th Conference on Information and Graphic Arts Technology 55

 Oral presentations

4 DISCUSSION

The results showed that the product type did not affect the attractiveness evaluation. Participants’ responses
were not influenced by graphical elements also. More interesting finding is that there was a significant differ-
ence in the attractiveness perception between designers and non-designers. First of all, designers were more
severe than users while evaluating the samples, and this was not surprising. While designers preferred linear
and simple instructions design, laypeople were attracted by the colourful designs made of planes. As noted by
Tractinsky and Meyer (1999), people prefer simple graphical presentations that facilitate efficient information
processing. Generally, well-trained designers rely on this concept in their work, so this probably guided their
evaluation in our study. On the other side, laypeople show a preference for visual embellishments which are
not essential to understanding the information (Bateman 2010), and this might resulted with their higher
scores for colourful instructions.
The results of the evaluations were in line with the verbal explanations of participants. Most of the designers
mentioned “clean” and simplified form as characteristics of attractive visualizations, while laypeople appre-
ciated the power of colour and the modern look of more complex designs.
Our finding about the gap between designers and users is in line with other studies that showed differences
between this two groups of participants. Quispel (2014) investigated how people differ in their evaluation of
data visualizations, and the results showed that designers rated the attractiveness of samples differently than
laypeople. This is in accordance with the study focused on product design, conducted by Hsu et al (2000). They
also reported about different preferences across the participants while evaluating telephone design.

5 CONCLUSIONS

Overall, our results suggest that designers prefer simple and sophisticated designs, while laypeople like em-
bellished colourful visualizations. The study demonstrated notable mismatch between these two groups of
respondents. In the light of our findings, we suggest that every graphical presentation of instructions for
product use should be tested on the group of end-users. Testing prototypes is common practice in design
projects, so it should not be the exception in instructions design.
Our study have limitations. We measured only the attractiveness of instructions design, without evaluating
their understandability. Future studies should investigate both measures in order to get more data about
functional and aesthetic aspects of instructions design. Furthermore, line and plane are not the only graphical
elements commonly used for instructions design, so other elements (such as dot) should be manipulated as
independent variable in future research.

6 REFERENCES

• Agrawala, M., Phan, D., Heiser, J., Haymaker, J., Klingner, J., Hanrahan, P., Tversky, B. 2003. “Designing Effective
Step-by-Step Assembly Instructions” ACM Transactions on Graphics (TOG) 22 (3): 828-837
• Barlow, T., and Wogalter, M.S., 1991. “Increasing the Surface Area on Small Product Containers to Facilitate
Communication of Label Information and Warnings.” Proceedings of Interface 91(7): 88-93
• Bateman, S., Mandryk, R. L., Carl Gutwin, C., Genest, A., McDine, D., and Brooks, C. 2010. “Useful junk?: The
effects of visual embellishment on comprehension and memorability of charts” Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems, 2573-2582. ACM
• Bonnardel, N., Piolat, A., and Le Bigot, L. 2011. "The impact of colour on Website appeal and users’ cognitive
processes." Displays 32(2): 69-80
• Hsu, S.H., Chuang, M. C., and Chang, C. C. 2000. "A Semantic Differential Study of Designers’ and Users’
Product Form Perception" International Journal of Industrial Ergonomics 25(4): 375-391
• Magnier, L., Schoormans, J., and Mugge, R. 2016. "Judging a Product by Its Cover: Packaging Sustainability
and Perceptions of Quality in Food Products." Food Quality and Preference 53: 132-142
• Monk, A. and Lelos, K., 2007. “Changing Only the Aesthetic Features of a Product Can Affect Its Apparent
Usability.” In Home Informatics and Telematics: ICT for The Next Billion, edited by Alladi Venkatesh, Timothy
Gonsalves, Andrew Monk and Kathy Buckner, 221-233. Boston: Springer
• Quispel, A. and Maes, A. 2014. "Would You Prefer Pie or Cupcakes? Preferences for Data Visualization Designs
of Professionals and Laypeople in Graphic Design." Journal of Visual Languages & Computing 25(2): 107-116
• Tractinsky, N. and Meyer J. 1999. “Chartjunk or Goldgraph? Effects of Presentation Objectives and Content
Desirability on Information Presentation.” MIS Quarterly 23(3): 397-420

8th Conference on Information and Graphic Arts Technology 56

 Oral presentations

• Vogt, S. and Magnussen, S. 2007. “Follow results: “Expertise in Pictorial Perception: Eye-Movement Patterns
and Visual Memory in Artists and Laymen” Perception 36(1): 91-100
• Wang, E. S. T. 2013. "The Influence of Visual Packaging Design on Perceived Food Product Quality, Value, and
Brand Preference." International Journal of Retail & Distribution Management 41(10): 805-816

Corresponding author:
Dorotea KOVAČEVIĆ
University of Zagreb, Faculty of Graphic Arts, Department of Graphic Design and Imaging
Getaldićeva 2, 10000 Zagreb, Croatia
phone: +385 1 237 10 80/226, fax: +385 1 237 10 77, e-mail: [email protected]

Co-author(s):
Maja BROZOVIĆ
University of Zagreb, Faculty of Graphic Arts, Department of Graphic Design and Imaging
Getaldićeva 2, 10000 Zagreb, Croatia
phone: +385 1 237 10 80/213 • fax: +385 1 237 10 77 • e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 57

 Oral presentations

Electron beam Curing of printing inks in packaging industry

Jakub GRIGER1, Dávid ŠARISKÝ1, Zuzana ŠTROMAJER1, Michal ORAVEC2,

Michal ČEPPAN2 & Viera JANČOVIČOVÁ2

1
Chemosvit Fólie, a.s., Štúrová 101, 059 21 Svit, Slovak Republic
2
Department of Graphic Arts Technology and Applied Photochemistry, Institute of Natural and Synthetic
Polymers, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Rad-
linského 9, 812 37 Bratislava, Slovak Republic

ABSTRACT: The aim of this work was to find a suitable methodology for the evaluation of the cure degree
of offset inks using electron beam (EB) curing by FTIR spectroscopy and characterization of printing EB inks
and printed substrates properties. The effect of EB inks curing was analyzed by comparing the spectra of the
uncured and the cured colors. For the quantitative characterization, the method of decomposition of the spec-
trum into peaks and program Origin utilizing the Fit Multi-peaks function were used. For comparison among
spectra and calculation of conversion of peaks, we determined the internal standards. To assess the cure rate,
peaks representing double bonds were used. The influence of the EB reactor settings on the stability of the
printed film and the hardening of the black ink were studied. All of the examined samples showed required
cure degree

Keywords: electron beam curing; curing degree; EB inks, FTIR spectroscopy, colorimetry.

1 INTRODUCTION

Due to the ever-increasing demands on food quality, there is an increasing need for innovation in their pack-
aging. Considering the vast range of packaged goods, packaging is made from a wide range of different ma-
terials. In order to achieve the best properties and the low cost of packaging, it is often necessary to combine
different types of materials to create multilayer or intelligent packaging.
The most commonly used materials for the production of flexible polymeric packages are polyolefins. (Izdeb-
ska, 2015). Their advantage is their low cost and easy processability Alternative options are also biodegradable
polymers and polymers produced from renewable sources (Brigham, 2018). Packaging today employs all the
printing techniques that, besides paper and cardboard, can print all kinds of films, plastic or metal, glass and
other materials. It is important that the packaging does not contain ingredients that could migrate and affect
the quality of the packaged goods. Therefore, perfect curing of used print inks is necessary. The aim of this
work was to find a suitable methodology for the evaluation of the degree of curing offset inks using electron
beam by FTIR spectroscopy and characterization of the prints of the EB inks and printed substrates as well as
the study of the influence of the EB reactor on the quality of the curing.

2 EXPERIMENTAL

The samples were prepared by offset printing machine which combines offset printing units with flexo print-
ing unit, uses electron beam curring inks which are crosslinked by accelerated electrons. The biaxial oriented
polypropylene foil (BOPP) was used as printed substrat. The rheological measurements of inks, measurement
of free surface energy using the contact angle measurements of the sessile drop, colorimetric and gravimetric
evaluation of the ink layer printed on the foil (Table 1) and spectroscopic measurements for all samples in
infrared region were performed. The effect of EB inks curing was analysed by comparing the spectra of uncured
and EB cured inks. For the quantitative characterization, the method of deconvolution of the spectrum into
peaks (Origin Fit Multi-peaks function) was used.

8th Conference on Information and Graphic Arts Technology 58

 Oral presentations

Table 1. Printing ink deposition on the BOPP foil.

Ink Ink deposition [gm-2]
White 2.786 ± 0.357
Cyan 1.216 ± 0.270
Magenta 1.096 ± 0.145
Yellow 0.430 ± 0.292
Black 1.458 ± 0.275

The curing process was evaluated by IR spectroscopy (FTIR spectrophotometer EXCALIBUR SERIES Digilab FTS
3000 NX, USA). The spectra of uncured samples were measured on aluminium plates using simple reflexion, the
spectra of cured films on foils were measured with ATR technique. The degree of conversion of the cured film
was determined according to the amount of double bond (twisting vibration at 810 cm−1, stretching vibration
at 1610 – 1640 cm−1) by a baseline method. In order to eliminate the influence of scatter in layer thickness, we
used as internal standard a peak at 1730 cm−1 or 2900 cm-1 (the choice depends on the layer thickness). The
degree of conversion X was calculated from equations (1) (Jančovičová, 2013) which was modified according
to the standard peak

(1)

where A0(l) is the absorbance of monomers C=C bounds of uncured ink measured at chosen wavelength (810,
1410 or 1640 cm-1), A0(IS) is the absorbance of internal standard of uncured ink (1730 or 2900 cm-1), At(l) and a
At(IS) are the values of absorbance of inks cured by EB on BOPP foil (printed foil) at the same wavenumber as
by uncured samples. The absorbance was calculated by the base line method.

3 RESULTS & DISCUSSION

3.1 Characterization of samples by FTIR spectroscopy

Uncured inks and printed foils were characterized using with FTIR spectroscopy. In the white ink spectrum (Fig-
ure 1) we can see the peaks in the region of 3000 - 2800 cm-1 representing the CHX groups, the most significant
peak at 1730 cm-1, characterizing the carbonyl group and the peaks at 1640, 1410 and 810 cm-1 corresponding to
the double bond (C = C). The spectra of all inks (C, M, Y, K) are very similar to the spectrum of white ink, due
to the content of the same binder. In FTIR spectra of inks mainly binders are visible because of low contents
of pigments in the composition.

1.0 1730

0.9

0.8

0.7
2900
absorbance

0.6
1410
0.5 1295

0.4
985 810
0.3

0.2 1640

0.1

0.0

4000 3500 3000 1800 1600 1400 1200 1000 800 600
-1
wavenumber [cm ]

Figure 1. Normalised FTIR spectra of uncured (red line) and cured white ink (black line) on the foil.

8th Conference on Information and Graphic Arts Technology 59

 Oral presentations

3.2 The development of the methodology for the evaluation of the cure degree of inks

The degree of curing of offset EB inks printed on foils was evaluated by FTIR spectra analysis. Before the
analysis, it was necessary to correct the measured spectra. Since the infrared spectra of uncured inks were
measured by the simple reflection method, and the FTIR spectra of the cured inks were measured by the ATR
method, the ATR correction of the spectra measured by the ATR technique is required for their direct compar-
ison (the penetration depth of radiation depends on on its wavelength – radiation penetration depth at 2800
cm-1 is around 1 μm, at 1400 cm-1 about 2 μm and at 810 cm-1 to 3.5 μm). ATR correction adjusts the absorbance
values ​​depending on the depth of radiation penetration into the surface layer of the material. The spectra
were subsequently normalized to peak at 1730 cm-1, which is the highest peak in most spectra of uncured and
cured inks (Fig. 1).
In order to compare the height of the absorption bands between the uncured and cured inks, it is necessary
to relate their values to the value of the internal standard. As a value of the internal standard, the height of
the peak at 1730 cm-1 characterizing the carbonyl group was choosen, which is used as the internal standard
for UV curing (Jančovičová, 2013), and the average height of the absorption bands 3000-2800 cm-1 (herein-
after referred to as the band at 2900 cm-1) characterizing CHx (CH, CH2, CH3) groups on the assumption that
EB curing does not cause their disappearance but only the change in their ratios. The choice of the internal
standard is related to the thickness of the layer; for thinner layers, the more suitable band appears at 1730,
for thicker band at 2900 cm-1.
In Figure 1 we can see that curing results in a significant decrease in the bands characterizing the C = C link-
age (1640, 1410 and 810 cm-1). Since the radiation at 810 cm-1 also penetrates into the foil (the thickness of the
colored layer estimated based on the values in ​​ Table 1 ranged from 0.4 to 2.5 μm), and the band at 1640 cm-1
is significantly affected by the environment as best suited to determining the double bond conversion that
closely correlated with the degree of cure, we evaluated the band at 1410 cm-1. By using the ratios of the these
peaks to the internal standards, the conversion was determined to characterize the electron beam cure rate. For
the quantitative characterization of peaks firstly we used the zero line method and the peaks deconvolution
using the function in the Origin program. The results achieved by both methods were similar, so we preferred
a simpler and faster zero line method, especially when there was a need for rapid analysis in practice.
Conversion values calculated using equation (1) are shown in Table 2 (using internal standard 1700 cm-1)
and Table 3 (using internal standard 2900 cm-1). It was confirmed that the best agreement of the results was
observed using the peak at 1410 cm-1. From Tables 2 and 3 we can see that all conversions calculated on the
basis of the absorbance at this wavelength reached ≥ 70%. In the literature, a reference conversion value for
good cured EB inks was not found. UV inks are good cured by double bond conversion in the range 70 – 80%.
So we can say that the studied colors are cured enough, the problem could be only in the case of magenta.

Table 2. Conversion values of EB cured inks on foil using internal standard 1730 cm-1.
Wavenumber Conversion [%]
Sample
[cm-1] C M Y K
1640 48 61 63 68
Foil 1 1410 92 70 84 85
810 64 74 57 30
1640 40 80 79 67
Foil 2 1410 93 70 76 81
810 67 78 49 35

8th Conference on Information and Graphic Arts Technology 60

 Oral presentations

Table 3. Conversion values of EB cured inks on foil using internal standard 2900 cm-1.
Wavenumber Conversion [%]
Sample
[cm-1] C M Y K
1640 81 79 87 66
Foil 1 1410 97 84 93 93
810 87 86 85 64
1640 75 88 92 88
Foil 2 1410 97 81 91 93
810 86 87 81 76

3.3 Influence of EB reactor setting on the foil and ink curing

Considering that the printed foils are used as packaging material, they need to preserve the properties im-
portant for the packaging, in particular good optical and barrier properties. Since these properties can be
negatively affected by the degradation of the printed substrate (BOPP treated with corona discharge), it is
necessary to determine whether the effect of EB does not degrade the substrate. We found that the EB reactor
did not cause damage (degradation) of the BOPP film used, even at the maximum reactor setting.
On foils printed with black ink, we tested the effect of EB reactor parameters setting (voltage, dose, beam cur-
rent) on ink curing. Using the method described in part 3.2, we calculated the conversion of the double bonds.
Due to the fact that it was a relatively thin ink layer, we used the peak at 1730 cm-1 as standard to calculate the
conversion of double bond at 1410 cm-1. We evaluated the ink curing in the center of the film and on its edge.

Table 4. The influence of the voltage on the curing of black ink.

Beam Conversion [%]
Sample Voltage [kV] Dose [kGy]
current [mA] in the center on the edge
EB1 110 34.6 115 90 89
EB2 100 34.6 115 83 83
EB3 90 34.6 115 78 78

As can be seen from Table 4, decreasing the value of the voltage also results in a decrease in the conversion of
the double bond, the conversions obtained at different positions of the printed foil are the same. The lowest
conversion achieved at the 90 kV voltage was 78%, its value dropped by 12% compared to the value achieved
at 110 kV.

Table 5. The influence of the dose on the curing of black ink.

Beam Conversion [%]
Sample Voltage [kV] Dose [kGy]
current [mA] in the center on the edge
EB1 EB1 110 34.6 115 90
EB2 EB4 110 20 60 76
EB3 EB5 110 10 32 64

Table 5 compares the conversion dependent on radiation dose, with dose decreasing the conversion of the dou-
ble bond decreased significantly more than in the case of a voltage decrease. The lowest achieved conversion
(64 and 61%) is relatively low, and it can be assumed that the ink was not sufficiently cured at this setting of
the EB reactor. According to the results obtained, the radiation dose should not decrease under 20 kGy.

4 CONCLUSIONS

The procedure for evaluating of the electron beam curing of inks on a polymeric foil using FTIR spectroscopy
was optimized. At the maximum setting of EB reactor, all inks were sufficiently cured. The influence of the
EB reactor settings on the stability of the printed film and the curing of the black ink were studied. We have
found that even at maximum reactor power (voltage 110 kV, radiation dose 34.6 kGy, beam current 115 mA) at

8th Conference on Information and Graphic Arts Technology 61

 Oral presentations

a print speed of 100 m/min no damage of the printed films (BOPP) occurred. Thus, we can conclude that EB has
no significant degradation effect on the printed film. However, the reduction of the radiation dose to 10 kGy
and the beam current to 32 mA may have a negative impact on ink curing, since the double bond conversion
achieved in this case was only 61% (Griger, 2017).

ACKNOWLEDGMENT
This work was supported by COST Action FP 1405.

5 REFERENCES

• Brigham C. 2018. “Chapter 3.22 - Biopolymers: Biodegradable Alternatives to Traditional Plastics.”In Green
Chemistry, edited by Béla Török and Timothy Dransfield, 753−770. Amsterdam: Elsevier.
• Griger, J. 2017.“Spectral and Optical Characteristics of Inks and Foils in Offset Printing Using EB Curing.”Di-
ploma Thesis, Slovak University of Technology in Bratislava.
• Izdebska, J., Sabu, T. 2015. Printing on polymers. Fundamentals and Applications. Amsterdam: Elsevier.
• Jančovičová, V., Mikula, M., Havlínová, B., and Jakubíková, Z. 2013. “Influence of UV-curing conditions on po-
lymerization kinetics and gloss of urethane acrylate coatings.” Progress in Organic Coatings 76 (2-3): 423−438.

Corresponding author:
Viera JANČOVIČOVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers, Department of Graphic Arts Technology
and Applied Photochemistry
Radlinského 9, 812 37 Bratislava, Slovak Republic
phone: +421 259 325 227, e-mail: [email protected]

Co-author(s):
Jakub GRIGER
Chemosvit Fólie, a.s.,
Štúrová 101, 059 21 Svit, Slovak Republic
phone:+421 905 860 666, e-mail: [email protected]

Dávid ŠARISKÝ
Chemosvit Fólie, a.s.,
Štúrová 101, 059 21 Svit, Slovak Republic
phone: +421 917 914 518, e-mail: [email protected]

Zuzana ŠTROMAJER
Chemosvit Fólie, a.s.,
Štúrová 101, 059 21 Svit, Slovak Republic
phone: +421 907 353 924, e-mail: [email protected]

Michal ORAVEC
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers, Department of Graphic Arts Technology
and Applied Photochemistry
Radlinského 9, 812 37 Bratislava, Slovak Republic
e-mail: [email protected]

Michal ČEPPAN
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers, Department of Graphic Arts Technology
and Applied Photochemistry
Radlinského 9, 812 37 Bratislava, Slovak Republic
phone: +421 259 325 224, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 62

 Oral presentations

EQUIPPING THE ARS VIVA INSTITUTE WITH WEB COMMUNICATION CHANNELS

Anja ŠKERJANC1, Jure VIDMAR2, Primož PREVC2 & Matija MAROLT3

1
Faculty of Natural Sciences and Engineering, Snežniška 5, 1000 Ljubljana
2
Faculty of Electrical Engineering, Tržaška cesta 25, 1000 Ljubljana
3
Faculty of Computer and Information Science, Večna pot 113, 1000 Ljubljana

ABSTRACT: This contribution presents the results of the project entitled Equipping the Ars Viva Institute with
target communication channels of new media, co-financed by Public Scholarship, Development, Disability
and Maintenance Found, Ministry of Education, Science and Sport of the Republic of Slovenia and European
Social Fund. The main purpose of the project was the development of clear and efficient online communication
channels and accessible web media based on the activities and the pursued goals of the Institute. The goal
was to introduce new digital strategies into the functioning of the Institute and to strengthen the already
existing online channels. The results of the project were planned to increase the Institute’s recognisability in
the local and wider environment and to help the disadvantaged groups better integrate into society. Experi-
mental work involved the user-centred design of a new user-friendly website and the optimisation of the use
of social networks.

Keywords: web communications channels, website design, user experience, social networks.

1 INTRODUCTION

The Web is becoming one of the key elements of communication and organization in the lives of all people,
so it is important that the information and services it offers are accessible. Certain vulnerable groups are faced
with many problems in the use of the Internet. The Web allows vulnerable groups access to information and
more opportunities for interaction. It is important that the web can be perceived and understood by all, so
that they can navigate through it, interact with it, and co-create it (Demšar, 2015). Web accessibility means
that people with disabilities can perceive, understand, navigate, and interact with the Web, and that they
can contribute to the Web. Web accessibility also benefits others, including older people with changing
abilities due to aging. Tips for Web Accessibility introduce some basic considerations for making a website
more accessible to people with disabilities. Tips are grouped by activities in Designing tips, Writing tips and
Developing tips (W3C, 2018).

The basic concept of user-centred design is: Take the user into account in every step of the development of
the product. This design approach includes five planes (James Garrett, 2011):
1. The Strategy Plane: involves the understanding of what the users want from the product and how that
fits in with the context and other goals they have
2. The Scope Plane: the strategy is translated into scope through the creation of functional specifications
and content requirements.
3. The Structure Plane: the scope is given structure through interaction design, in which we define how the
system behaves in response to the users and information architecture.
4. The Skeleton Plane: the skeleton plane breaks down into three components: information design, interface
design and navigation design.
5. The Surface Plane: here, the finished product (functionality-oriented product or an information resource)
creates the sensory experience.

Moreover, while it cannot be denied that accessibility and usability are two qualities that interact with each
other it has always been difficult to define the scope and extent of this relationship. In fact, if accessibility
and usability are not properly integrated, Websites can turn out to be either accessible but barely usable, or
usable but barely accessible (Aizpurua, 2016).
The Ars Viva Institute is a culture centre and incubator active in the field of social issues, integration of disad-
vantaged groups into society and awareness of local cultural and natural heritage. The Youth Hostel Ars Viva
operates within the Ars Viva Institute and is the first accommodation object in Slovenia which is completely
adapted for the physically impaired individuals. Due to the remoteness and lack of recognisability of the in-
stitute on the national level, there was a need for clearer representation of activities and the Institute’s vision
through media and visual communication.

8th Conference on Information and Graphic Arts Technology 63

 Oral presentations

2 EXPERIMENTAL

2.1 Website

The website design began with the analysis and evaluation of the current Institute’s website at the time. Online
communication channels of other, competitive non-profit organisations and providers of accommodation in
the region were also analysed. The final step of this phase was the definition of target users. Target users were
split into two groups namely locals and tourists. Locals are people who live in the region and are interested in
Institute’s activities such as workshops and events. This group includes young people, parents, elderly people
and people with disabilities who have difficulty integrating into society. On the other hand, the second group
includes tourists who prefer different experiences and love nature and culture. The needs of the boughs groups
were taken into account while designing.

In the second phase, web-content types, web-functionalities and the needs of the Institute were determined.
This was followed by the development of information architecture and interaction design.
A simple card sorting method was used to create information architecture. Content for the new website was
written on cards which were then distributed in to different groups. Prior to the design of wireframes, the
trends in website graphic design and recommendations for accessibility were examined. It is very important
to create a user-friendly and interesting website. After the review, wireframes were designed in Adobe Illus-
trator. The design of wireframes began with raw schematic models and gradually proceeded to the final form.
Graphic design began with the determination of the colour scheme and fonts used on the website. This was
based on the Institute’s logo which kept the consistency over the entire graphic identity. By upgrading the
wireframes with graphic elements and content, final design was gradually reached using Adobe Illustrator,
Adobe InDesign, and Adobe Photoshop.

Depending on the needs and the knowledge of the representatives of the Institute a decision was made for
the website to be managed in Content management system WordPress. This makes it easier for administrators
to manage content and the website in general.

Furthermore, the new website was tested with the founder of the Institute, who is disabled. The feedback
obtained in this way has helped us improve the accessibility of the website.

2.2 Social networks

The workflow for social networks optimisation involved the analysis of social networks in which mistakes that
affected attainment of target groups and consequently, the recognisability of the Institute were discovered.
Optimisation was based on the needs of target groups and focused on social networks, such as Facebook,
YouTube and Google+. On Facebook, the main problem was the improper posting that did not reach target
users. For this purpose a model was designed for targeted promotion. Interesting, informative posts and posts
with local content were created to teach the representatives of the Institute how to properly prepare the posts
and how to communicate with the target users through social networks.

The model contained ten posts featuring mascot Bruno the bear, which was developed as part of the project.
Bruno helped inform about the region and the Institution. Promotion started with a post announcing that the
Institute, in cooperation with students, is working on a project within which the mascot is being developed.
The following posts were of informative nature, containing interesting content about bears (behaviour, eating
habits, bear encounters in the wild), local sights and events organised by the Institute. In the end of targeted
promotion the new website was presented. All the posts were evenly distributed over a period of two months
to obtain as many followers as possible.
The goals that were set during the planning of the strategy were for posts to trigger responses (comments,
likes and shares) and to increase the number of followers.

Throughout the targeted promotion action, responses to each post were monitored and compared with re-
sponses to other posts that were not part of promotion action.

8th Conference on Information and Graphic Arts Technology 64

 Oral presentations

3 RESULTS & DISCUSSION

3.1 Website

Analysis of the old website revealed poor user experience due to badly structured information architecture.
The Institute and the hostel had two different websites which were badly connected and confusing to the
user due to their poor organisation and visual layout. Other design flaws were also found, such as typography
without caron letters, various typography in news, small images and fonts, overlapping images and text and
poorly visible typography due to inappropriate colours. Photos in the photo gallery were small and opaque.
Furthermore, there was no link on the website leading to the Institute’s social networks.

The competition was analysed from the aspects of the Institute and the hostel in which the competition of the
Institute was represented by other non-profit organisations with similar activities in the region. The analyses
was based on online communication channels. Two thirds of the analysed organisations did not have web-
sites or did not have user-friendly websites. The rest had websites with good user experience. Almost all of
the non-profit organisations in the region were present on Facebook. The analysis of the hostel was done in a
similar way. The competition of the hostel was represented by other accommodation providers in the region.
About 80% of analysed accommodation providers did not have websites or did not have a user-friendly web-
sites. The rest had modern and user-friendly websites. The competitive advantages that can be highlighted are
the presence on the web, an inspiring story, the enthusiasm of all generations and interesting events such as
workshops, exhibitions and concerts. A big disadvantage was the unclear connection between the Institute
and the hostel and poor information structure.

The main need of the locals as a target group was clear announcement of the events, as well as the ability to
view those events online in a gallery of pictures and videos. The possibility of accommodation is the most
important information for tourists, so it was necessary to establish a clear link between the Institute and the
hostel.

The information architecture, presented in Figure1, displays clearly divided content on the website with all
the features and touch points. In this part of website design it was important to create a link between all so-
cial networks and the website and to create networking. The needs and wishes of the Institute regarding the
content and functionality were taken into account.

Figure 1. The information architecture.

The site is designed to be simple and easy to use with alternating layout which makes the site dynamic. The
main elements on the site are images through which the Institute narrates its story. The design was derived
from the elements and colours of the overall graphic identity of the Institute. The logo of the Institute is
included in the website’s navigation, as home button. The image of old house, which is also an important
element of the identity, is included in the footer. The primary colour used is orange, which is combined with
gray, white and blue. Two fonts are used, linear font Open sans and serif Playfair Display for titles. On the
left side of the site are buttons with links to the social networks of the Institute. In Figure 2 a wireframe and
a design of subpage in presented.

8th Conference on Information and Graphic Arts Technology 65

 Oral presentations

Figure 2. Wireframe and design of subpage About the Institute.

During the whole planning, design and programming of the website, recommendations for accessibility were
taken into account. At each stage of website design, representatives of the Institute were consulted. The ulti-
mate site also included the ability to adjust the contrast for the visually impaired and a side slider, for those
who cannot scroll with the mouse.

As a result, the Institute was equipped with a freshly developed, responsive and accessible website that enables
simplified communication between the representatives of the Institute and target users and high level of online
communication. Finally, the sustainability of the project results was achieved with the transfer of knowledge
from web developers to the representatives of the Institute.

8th Conference on Information and Graphic Arts Technology 66

 Oral presentations

3.2 Social networks

Facebook page analysis revealed that the page had a relatively small number of followers. It was also found that
the founder of the Institute posted content from his private Facebook account instead of the Institute’s page.
Content was then shared on the Institute’s page which is the reason why posts could not reach a larger crowd.
The results of the target promotion showed that Bruno’s posts and Bruno himself were well received (Figure
3). Responses to Bruno’s posts were on average better than responses to other posts, when sponsored posts
are not included.

Figure 3. Responses to posts target promotion.

With the optimisation of the Facebook page and implementation of a model for targeted promotion of this net-
work, the number of followers was increased and the informative and engaging level of content was improved.

Figure 4. Numbers of followers target promotion.

YouTube channel review and analysis revealed that the Institute had an inoperable channel due to problems
with access (problems with the e-mail administrator). The YouTube channel of the Institute was therefore
transferred to another e-mail from which the representatives can access the YouTube channel. Video data was
also optimised by editing it with descriptions and keywords.

4 CONCLUSIONS

When it comes to website design it is important to include target users and their needs at every step. That
is the way to provide an interesting website with good experience for all users. Most websites are designed
without the consideration of disadvantaged groups and other minorities which is why they might struggle
when navigating online. Our website design was based on the needs of all groups including the disadvantaged.
Analysis and constant communication with the target users was the key to reaching our final goal.
The results of the project were of a great importance and a benefit for the representatives of an Institution
Ars Viva and also for local and regional environment. The project results’ dimensions included networking,
informing, introducing and especially connecting target groups with disabilities.
Moreover, the sustainability of the results were achieved with the transfer of the knowledge to the reresenta-
tives of the Institute that through accessible web media can continue with transparent, modern and engaging
comunication.

8th Conference on Information and Graphic Arts Technology 67

 Oral presentations

ACKNOWLEDGMENT
The presented research work was financed by Public Scholarship, Development, Disability and Maintenance
Fund of the Republic of Slovenia, Ministry of Education, Science and Sport (Republic of Slovenia) and Euro-
pean Social Fund. The authors of this contribution would like to thank the supporters of this project and all
the dedicated participants involved (the Ars Viva Institute).

5 REFERENCES

• Aizpurua A., Harper S., Vigo M. 2016. “Exploring the relationship between web accessibility and user expe-
rience.” Int. J. Human-Computer Studies 91 (2016): 13-23
• Demšar, D., Krajnc, A., Vesel, A., Pochyla, D., Klemenčič, A., Smerdel, A., & Lištvanova, L. 2015. Dostopnost
spletnih strani. Ljubljana: Beletrina.
• James Garrett, J. 2011. The Elements of User Experience: User-Centered Design for the Web and Beyond,
Second Edition. Berkeley: New Riders
• Web Accessibility Initiative, W3C. Introduction to Web Accessibility: What is Web Accessibility? URL: https://
www.w3.org/WAI/intro/accessibility.php (last accessed on 10. 1. 2018)

Corresponding author:
Anja ŠKERJANC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Chair of Information and Graphic Arts Technology
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 31 556 977, e-mail: [email protected]

Co-author(s):
Jure VIDMAR
University of Ljubljana, Faculty of Electrical Engineering
Tržaška cesta 25, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Primož PREVC
University of Ljubljana, Faculty of Electrical Engineering
Tržaška cesta 25, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Matija MAROLT
University of Ljubljana, Faculty of Computer and Information Science
Večna pot 113, 1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 68

 Oral presentations

EXTAND OF PAPER COATING cracking

Klemen MOŽINA 1, Jure AHTIK 1, Špela BEČAJ 1, Urška KRAJNC 1, Matej PIVAR 1,
David RAVNJAK 2 & Diana Gregor SVETEC 1
1
University of Ljubljana, Faculty of Naturals Sciences and Engineering
2
B & B, Papirnica Vevče, Labels and Flexpack

ABSTRACT: In the paper mill Papirnica Vevče, one priority remains, i.e. to develope coated papers with as little
defects as possible, especially to reduce the number of coating pits and the extent of coat craking at fold at
the end use applications (Brigl & Bergmeister, 2018). Among paper properties their basic phyisical (thickness,
mass, grammage, specific volume, moisture and ash content), optical (specular gloss, ie. the coated side of
a paper), surface (roughness), printing (print penetration) and morphological properties (SEM) were deter-
mined. To determine which coated paper has a greater tendency towards cracking, prints in offset printing
techique were prepared. Paper samples were printed on the coated side with the black offset printing ink, just
to ensure high contrast. Printed papers were than folded and the place of fold was observed under the optical
microscope. Pictures taken with the optical microscope were processed using the ImageJ program. The extened
of cracking at the fold was evaluated as the area where the printing ink has detached.

Key-words: paper, coating, SEM, optical microscope, ImageJ.

1 INTRODUCTION

Base paper, when it comes from the paper machine is rough, and needs further treatment, such as coating, in
order to improve optical and surface properties and also to optimize the printing characteristics of paper. With
coating a uniform paper surface is created, as the irregularities of the paper are smoothed out, which results in
high printing quality. Generally, the quality of a coated paper is in a close correlation to a base paper quality,
which has to have appropriate mechanical properties, adequate optical and surface characteristics (Možina,
2017). Nevertheless, the mayor influence has the type of coating technique, performed either on- or off-line,
coating speed and coating composition. Depending on the intending end use of coated paper, the coating con-
sists of pigments together with other ingredients and binding agents, such as latex, starch, optical whitener,
dyes and other additives, which give the paper desired rheological behavior and improve the performance of
coated paper (Novak, 2004; Kimpimäki, 1998).

2 EXPERIMENTAL

Analyzed papers were one side coated, each with different recipe and different coat weight. The difference was
also in grammage of base paper and therefore consequently coated papers also differ in the grammage. Two
of the paper samples are used as a wet-strength papers that are alkali resistant label papers, with a grammage
of 68 and 70 g/m2. The other two paper samples with the grammage of 70 and 80 g/m2 are non-wet strength
and non-alkali resistant label papers. The third pair of paper samples included in our research are flexible
packaging papers with a grammage of 60 to 70 g/m2.

3 RESULTS & DISCUSSION

In experimental part, the main focus of our research was the evaluation of tendency of coating layer towards
cracking. Therefore, the correlation among the studied papers was noticed. The lowest measured value of
specular gloss was on S3 (52,65%), on which the highest cracking area, i.e. 4,15% (Table 2) was measured. The
highest amount of ash residue has S6, which has also the highest grammage and consequently due to the
quantity of added CaCO3, the highest specular gloss (Table 1). The highest measured value of print penetration
(Table 1) was measured on S2, i.e. 121 mm, which indicates that the specimen 2 has the lowest print absor-
bent capabilities. Overall evaluation of print penetration and air permeability of coated papers are collated.
The paper surface is closed, with a few minor micro defects caused by the air bubbles caught in the coating
mixture (Table 3).
Paper coating cracking occurred on specimens at a different extent, ranging from 1,13 (S1, CD to 4,15% for S3,
MD) and it can be clearly noticed with an eye. While the studied papers are produced for packaging purpos-
es, they are really implemented as white papers. Since packaging needs to be printed, the paper cracking of

8th Conference on Information and Graphic Arts Technology 69

 Oral presentations

coating layer becomes an issue. Even more is the problem emphasized when the majority of paper surface is
covered with paint, e.g. black or dark blue. S4 has the lowest tendency towards cracking in either measured
way, i.e. MD, 2,04% or CD, 1,53% (Table 2).

Table 1. Grammage, thickness, ash content, print penetration, roughness and specular gloss.
S1 S2 S3 S4 S5 S6
Grammage [g/m ] 2
66,6 60,9 69,1 70,2 69,6 78,4
Thickness [mm] 0,053 0,051 0,057 0,055 0,060 0,064
Ash content [%] 20,81 18,47 17,45 23,58 23,37 25,51
Print penetration [mm] 115 121 108 118 114 116
Roughness [ml/min] 13 14 18 22 14 18
Specular gloss [%] 66,85 72,05 52,65 78,80 54,60 81,55

Table 2. Measured values for air permeability (AP), average roughness (Ra) and cracking area (CA).
S1 S2 S3 S4 S5 S6
MD CD MD CD MD CD MD CD MD CD MD CD
AP [ml/min] 0 0 0 0 0 0 0 0 0 0 0 0
Ra [μm] 1,97 1,99 1,50 2,03 1,49 1,87 1,45 2,05 1,25 1,63 1,46 1,70
CA [%] 2,30 1,13 3,47 2,44 4,15 3,02 2,04 1,53 3,01 2,09 2,74 2,87

8th Conference on Information and Graphic Arts Technology 70

 Oral presentations

Table 3. SEM pictures of coated side of the papers.

Coated surface Air bubbles Conglomerate
100 x 3000 x 5000 x

V1

V2

V3

V4

V5

V6

8th Conference on Information and Graphic Arts Technology 71

 Oral presentations

Table 4. Pictures of folds, captured with optical microscope.

MD CD

V1

V2

V3

V4

V5

V6

8th Conference on Information and Graphic Arts Technology 72

 Oral presentations

4 CONCLUSIONS

In this study, extend of paper coat cracking was determined by using an optical microscope and pictures
captured were than evaluated with ImageJ, to establish the extensiveness and range of papers coating layer
towards cracking and the following conclusions can be made:
• cracking of coated papers has a negative impact on a aesthetics of the packaging and at costumer raises a
question in products quality;
• the size of coat cracking depends on the base paper and the chemistry composition of paper coating;
• to achieve the highest possible quality of coated packaging paper, within the reasonable tolerance limits,
the S4 has the best performance, i.e. lowest CA, among all six studied specimens.

5 REFERENCES

• Možina, K. 2017. “Viskoelastične lastnosti grafičnih papirjev.” Doktorska disertacija., University of Ljubljana.
• Novak, G. 2004. Grafični materiali. Ljubljana: Naravoslovnotehniška fakulteta, Oddelek za tekstilstvo.
• Kimpimäki, T. 1998. ”Dispersion coating and product application.” In Paper and Paperboard Converting,
edited by Antti Savolainen, 81−120. Helsinki: Fapet Oy.
• Brigl & Bergmeister. URL: htttp:// https://www.brigl-bergmeister.com/home// (last accessed on 20. 1. 2018).

Corresponding author:
Klemen MOŽINA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 99, e-mail: [email protected]

Co-author(s):
Diana GREGOR SVETEC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 72, e-mail: [email protected]

Špela BEČAJ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design

Urška KRAJNC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design

David RAVNJAK
Paper Mill Vevče, d. o. o.,
Papirniška pot 22, 1261 Ljubljana Dobrunje, Slovenija
phone: +386 1 587 72 00, e-mail: [email protected]

Jure AHTIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 86, e-mail: [email protected]

Matej PIVAR
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 73, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 73

 Oral presentations

FABRICATION OF CAPACITIVE SLIDERS

Tanja PLEŠA 1, Urška KAVČIČ 1, Gregor LAVRIČ 1 & Deja MUCK 2

1
Pulp and Paper Institute, Ljubljana, Bogišićeva 8, 1000 Ljubljana, Slovenia
2
Faculty of Natural Sciences and Engineering, Snežniška 5, 1000 Ljubljana

ABSTRACT: In this research different designs for printed capacitive sliders were made. In preliminary study
the distance between sliders electrodes was examined and the optimal distance was determined. According to
results of preliminary study 8 designs for big and 2 designs for small sliders were made. Slider designs were
examined according to the distance between the electrodes, the shape of a space between the electrodes and
the embraced conductive line around the electrodes. Based on this results the most optimal design for small
and big sliders was determined.

Key-words: printed electronics, printed sensors, printed sliders, printing, screen printing.

1 INTRODUCTION

Printed electronics is a technology that has paved the way for many interesting new uses in electronics and is
steadily gaining traction in the market. Printed electronics can be applied on packaging (smart packaging),
posters (smart posters), etc. It brings new application options for electronics that conventional electronics
cannot reach (Mraovič et al., 2014).

After years of research, printed functionalities were developed using conductive, semi-conductive and dielec-
tric materials for printing. These materials need to have the ability to be formulated into an ink. The printing
technologies used for printed electronics allow us to use flexible and rigid substrates that are not suitable for
conventional electronics. Electronics can be printed on plastic foil, thick plastic substrates, paper, cardboard,
glass, ceramic, metal and other materials. Conventional printing techniques are used to create printed elec-
tronics: inkjet, screen printing, gravure, flexography, pad printing, offset gravure and offset printing. After
printing, the electronics is treated with drying or sintering. The result of using conventional printing substrates
and printing technologies is high volume production of low-cost electronics. This kind of electronics has also
some disadvantages such as low level of integration, requires large surface of electronic components and slow
switching compared to the production of conventional electronics (Starešinič and Muck, 2010).

Printed electronics are increasingly being used to make different kinds of sensors. Sensors are devices that
respond to changes in the environment and give measurable responses.

Printed sensors have some significant advantages over their traditionally fabricated counterparts. Conven-
tional electronics is made using the photolithographic process, which is expensive, mostly because of the high
capital costs involved. On the other hand, it is cheap to fabricate printed electronics because of the simplified
production process, lower capital costs and the high speed of production. Only rigid substrates can be used
with the photolithographic process, but with printing technologies we can make printed electronics on flexi-
ble substrates. Photolithography is a high temperature process – unlike the press, which is a low temperature
process. On top of this, all printed electronics bring several new application opportunities such as packaging
integration (Pleša, 2017 and Kubersky, 2014).

Printed sensors have already started to make serious inroads into the market. By expanding the network of
the Internet of Things, the market for printed sensors is also on the rise. The market requires high volume
production and low costs, which is precisely what printed electronics offer.

Touch sensitive sensors are just one of the many sensors available. They are particularly interesting because
they can be used as a switch. Touch sensors that operates on the principle of a change of capacitance are di-
vided into three groups (Figure 1):
• zero-dimensional sensors, e.g. buttons – these sensors have only one touch point;
• one-dimensional sensors, e.g. sliders, wheels – these sensors detect the linear movement of a finger;
• two-dimensional sensors, e.g. touch screens – these sensors detect the movement of a finger on two axes
(Atmel, 2015).

8th Conference on Information and Graphic Arts Technology 74

 Oral presentations

Figure 1. Three groups of touch sensors (Atmel, 2015).

Capacitive sensors has nominal capacitance in a normal inactive state. However, when the surface of the sen-
sor is touched, the capacitance increases. The change in capacitance can only be measured using a chip (IC).

The aim of research is to fabricate a prototype of a printed capacitive slider based on a line sensor. That pro-
totype is presenting a hybrid electronics that combine conventional elements (IC and programming) with
printed conductive layers. The goal of the research was to:

• determine and optimize the process of making printed sliders

• fabricate a prototype of a printed slider

2 EXPERIMENTAL

2.1 Materials and methods

In this research printed circuit for sliders were screen printed on three different paper substrates: (1) recycled
paper, (2) special paper for printed electronics and (3) synthetic paper. For printing circuits silver conductive
ink CRSN 2442 (SunChemical) was used and for coating SG 70/15 (Coates Screen) was used. Regarding ink
properties screen printing mesh 120 l/cm was used for printing. All the samples were dried for 225 s at 120°C.

2.2 Printing of test capacitors

In preliminary study 10 simple capacitor designs with various distances between electrodes from 0.1 to 1 mm
were printed using screen printing (Figure 2). Capacitors have simple square electrodes dimensions 1 x 1 cm.
From every electrode 0.5 mm line leads to round contact surface which was used for measuring the capaci-
tance. Capacitor capacitance was measured with multimeter and the influence of the distance between the
capacitors electrodes was studied.

Figure 2. Test capacitors with various distances between electrodes.

8th Conference on Information and Graphic Arts Technology 75

 Oral presentations

2.3 Designing and printing of capacitive sliders

Based on preliminary research results 10 different designs of a printed circuit for sliders were designed: 8 de-
signs for large capacitors and 2 designs for small capacitors (Figure 3). According to literature (Atmel, 2015)
small sliders (between 21 and 26 mm long) should be designed with straight space between the electrodes.
In this research two small slider designs were tested: one with three electrodes one by another (design 6 in
Figure 3) and the other where highest channel electrode hugs the other two electrodes (design 5 in Figure 3).
For large sliders (between 26 and 60 mm long) a zigzag space between the electrodes is recommended so there
were made four modifications of “zigzag sliders”: one with 0.5 mm distance between the lines (design 1B in
Figure 3), other with 1 mm between the lines (design 2B in Figure 3), third where highest channel electrode
hugs the other two electrodes and the distance between the electrodes is 0.5 mm (design 4B in Figure 3) and
the fourth where all three electrodes are embraced with other conductive line (design 3B in Figure 3). All four
modifications were made also on “straight line sliders” (designs 1A, 2A, 3A and 4A in Figure 3). The designs
with highest channel electrode hugging the other two electrodes and the designs with all three electrodes
are embraced with conductive line were made to improve the electric field which effects the slider accuracy.
After printing circuit was coated and so protected from external influences and kept electrically isolated. Ca-
pacitance was measured on printed samples using LCR-300 Voltcraft multimeter with and without touching
them with capacitive touch pen and different designs were compared.

Figure 3. Different slider designs with labels.

8th Conference on Information and Graphic Arts Technology 76

 Oral presentations

3 RESULTS & DISCUSSION

Measurements of capacitance of capacitors show that the capacitance is higher when electrodes are closer
together. Capacitor with the distance 0.2 mm between the electrodes has the highest capacitance and the ca-
pacitor with the distance between electrodes 1 mm has the lowest capacitance (Figure 5). The smallest distance
between the electrodes that was tested (0.1 mm) is proven to be too small for good print quality with chosen
ink and selected screen mesh. Comparing capacitance measured on capacitors printed on different printing
material revealed that capacitance of the capacitors is lower when printed on special paper for printed elec-
tronics, but on the other hand almost the same when printed on recycled paper and synthetic paper.

Figure 4. Capacitor capacitance printed on different substrates.

Nevertheless, special paper for printed electronics had the lowest capacitance the measurements of the change
in capacitance when the slider is touched are comparable to recycled and synthetic paper. The chip is measuring
the change in capacitance that is why low nominal capacitance is not an issue. When measuring the change in
capacitance the influence of printing material is small and it is hard to determine a correlation between them.
Differences between the materials can also be attributed to the measurement error.
When measuring the change of capacitance on printed sliders which is essential for slider functioning it was
discovered that design 1B is the most responsive, right behind it there are 2B and 3B designs. All of these three
designs have zigzag line between the electrodes. After that the same slider designs with different shaped spaces
between the electrodes were compared: 1A vs. 1B, 2A vs. 2B, 3A vs. 3B, 4A vs. 4B. The zigzag space has proved
better in all the cases except in 4A vs. 4B comparison, where A4 with straight line space had higher capacitance
than 4B with zigzag space. That means the zigzag line on large slider design is more optimal. The same designs
with different distances between the electrodes were also compared: 1A vs. 2A, 1B vs. 2B. The highest change
in capacitance is on designs 1A and 1B where electrodes are closer together, the exception is a design with
zigzag space on paper for printed electronics where the capacitance is lower where the electrodes are closer
together. Designs 3A, 3B, 4A and 4B with highest channel electrode hugging the other two electrodes and the
designs with all three electrodes are embraced with conductive line do not show improved slider response.

8th Conference on Information and Graphic Arts Technology 77

 Oral presentations

Figure 5. Change of capacitance on different slider designs.

4 CONCLUSIONS

Higher capacitance of a capacitor is achieved with smaller distance between the electrodes, but the question
is how fine space can be achieved with printing conductive materials on various printing materials with differ-
ent surface properties. For printing capacitors the smallest distance between the electrodes is recommended
according to the substrate, ink and printing technique properties. The roughness of the substrate, the size of
conductive ink particles and a printing technique influence the precision of print and should be considered
when determining the distance between the electrodes.

The research proves that printed sliders can be made by printing conductive inks on different printing sub-
strates. However, on sliders performance printing substrate, shape and size of the slider have affect. For small
sliders, design with space between the electrodes in a shape of a thin line should be used. Better response of
a small slider is achieved if the highest channel electrode hugs the other two electrodes – this increases the
intensity of the electric field and higher capacitance is achieved. For large sliders, the design with zigzag space
is more optimal. Sliders with smaller distance between the electrodes have better response than sliders with
greater distance. Unlike small sliders, for large sliders it is not important if the highest channel electrode hugs
the other two electrodes – the slider response is more or less the same.

ACKNOWLEDGEMENTS
The work was carried out within the RDI project Cel.Cycle: »Potential of biomass for development of advanced
materials and bio-based products« (contract number: OP20.00365), co-financed by the Republic of Slovenia,
Ministry of Education, Science and Sport and European Union under the European Regional Development
Fund, 2016–2020.

5 REFERENCES

• Atmel. QTAN0079 Buttons, Sliders and Wheels Sensor Design Guide [accessible online]. Atmel documenta-
tion, published 2015 [cited 20. 7. 2015]. Accessible online: <http://www.atmel.com/products/TouchSolu-
tions/bsw/default.aspx?tab=documents>.
• KUBERSKÝ P., SYROVÝ T., HAMÁČEK A., NEŠPŮREK S. in SYROVÁ L. Fully printed electrochemical NO2 sensor.
Kidlington: Elsevier, Procedia Engineering, Volume 87, 2014, str. 1043-1046
• MRAOVIĆ, M., MUCK, T., PIVAR, M., TRONTELJ, J. and PLETERŠEK, A. Humidity sensors printed on recycled
paper and cardboard. Basel: Sensors, no. 14, 2014, p. 14.
• PLEŠA, T. Izdelava tiskanih stikal, diplomsko delo. Ljubljana: Univerza v Ljubljani, Naravoslovnotehniška
fakulteta, Oddelek za tekstilstvo, grafiko in oblikovanje, 2015
• PLEŠA, T. Fabrication of printed switches. Lancashire: PITA, Paper Technology, Volume 57, No. 2, 2016, p. 34-37.

8th Conference on Information and Graphic Arts Technology 78

 Oral presentations

• PLEŠA, T. Fabrication of Capacitive Sliders. Gradz: Cellulose Materials Doctoral students conference 2017,
2017, p. 45
• Sekitani, T., Someya T. Human-friendly organic integrated circuits. Kidlington: Elsevier, Materials today,
September 2011, Volume 14, št. 9, 2011, str. 398-407
• STAREŠINIČ, M. and MUCK, T. Large-area, Organic & Printed Electronics Convention. Ljubljana: Grafičar, no.
5, 2010, p. 8−9.

Corresponding author:
Tanja PLEŠA
Pulp and Paper Institute, Print and Packaging Department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Co-author(s):
Urška KAVČIČ
Pulp and Paper Institute, Print and Packaging Department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Gregor LAVRIČ
Pulp and Paper Institute, Print and Packaging Department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Deja MUCK
Chair of Information and Graphic Arts Technology
Faculty of Natural Sciences and Engineering
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 84, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 79

 Oral presentations

FORMATION AND CHARACTERIZATION OF SHELLAC FILMS

Urška VRABIČ BRODNJAK

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Aškerčeva 12, SI-1000 Ljubljana

ABSTRACT: The aim of our research was to prepare and characterize properties of shellac as a potential pack-
aging film. Advantages and disadvantages of prepared films have been discussed and presented.
This research shows preparation and characterisation of biopolymer, using different amounts of plasticizers.
Our research is focused on film properties that are important for packaging materials. The aim of this work is
to investigate moisture barrier properties and tensile properties of de waxed shellac flakes, using 2 different
amounts of glycerol and poly (ethylene) glycol as plasticizers, which could be used as packaging materials.
The results of the research has shown that shellac has certain barrier properties that could be used as film in
the packaging filed as replacement of plastic film barriers.

Keywords: shellac, barrier, properties, Kerria Lacca, packaging.

1 INTRODUCTION

These days, plastics are still widely used as packaging material, due to their ability to preserve the sensory
properties and nutritional values in food products (Byun et al., 2012). On the other hand, there are high
demands for the utilisation of natural polymer sources as food packaging materials (Byun et al., 2011). Nat-
ural polymers are conserved as an alternatives to synthetic plastics and many research has been done on the
application on biopolymers in many fields (Vrabič Brodnjak, 2017). Since food packaging materials require
non-toxicity, good moisture, water, mechanical and chemical properties, many biopolymers cannot conquer
to plastic materials, such as polyethylene, polypropylene etc. Therefore a good selection, production and
preparation of biopolymers should be performed and implemented as food packaging. As a natural polymer,
which is already used in many fields (wood preservation, finishes, adhesives for wood, component in cosmetics
such as nail polish and many more) as well as in packaging, is Shellac. It is a natural polymer, obtained from
purified resinous secretion, by the insects Kerria Lacca (Kerr) Lindinger (Coccideae). This species is the most
important lac insect, being a main source of lac, for the production of shellac. The insect is mostly cultivat-
ed on host trees in Thailand, India and Myanmar (Soradech et al., 2012). Shellac is one of the thermosetting
resins of animal origin. It is a natural bioadhesive polymer and is chemically similar to synthetic polymers,
and thus can be considered a natural form of plastic. It can be turned into a moulding compound when mixed
with wood flour and moulded under heat and pressure methods, so it can also be classified as thermoplastic.
Its chemical structure is composed of hard and soft resin of polyesters and single esters containing hydroxyl
and carboxyl groups (Farang, 2009; The, 2008). Shellac has excellent film forming and barrier properties. It
is soluble in alcohol and alkaline solutions (Soradech, et al., 2012). Therefore it has been widely used in the
food and agro industries for gas, moisture, water and microbial protection of food products (Soradech et al.,
2017). The goal of our research was to prepare shellac films, which could be used as packaging films. Different
amounts of plasticizer have been used. Therefore, the research was focused on film properties (moisture barrier,
tensile properties etc.), which change with the amount of plasticizer and are important for packaging films.
The results has shown, that shellac has a potential as a packaging film, most of all it could be used as barrier
and replacement of plastic film barriers.

2 EXPERIMENTAL

The shellac solution was prepared by dissolving 2 g of shellac flakes in 100 ml ethanol and PEG 200 was added as
a plasticiser (2 and 4%). The solution was mixed room temperature for 30 minutes until flakes was not dispersed.
After that, the film solution was filtered through a polyester screen (mesh no. 140 with mesh opening 160 µm)
with aspiration to remove small lumps in the solution. After the aspiration and the treatment, the Shellac solu-
tion was casted onto petri dishes (50 ml), spread thinly, uniformly and dried at 55°C for 10 hours. After the films
were peeled off from the dishes, they were cooled at room temperature (23°C; 55% RH).
When films were prepared thickness, water vapour permeability, tensile and surface properties, and thermal
stability were determined and analysed.

8th Conference on Information and Graphic Arts Technology 80

 Oral presentations

The thickness of films was measured with a precision digital micrometre Mitutoyo Corporation, Japan, to the
nearest 0.0001 mm at 5 random locations on each film.

To determine the WVP of films and moisture content, the ASTM E96 standard desiccant method was used (Srin-
ivasa et al., 2004). The test cups were filled with silica gel (RH = 16% in the cup), where a sample was placed
between the cup and the ring cover of each cup (Srinivasa et al., 2004). There was an air gap of 11 mm between
the silica gel and the underside of the placed film. To ensure the best results of WVP, a silicone sealant was
applied around the cup edge. The films with the exposed area of 50 cm2 were tested at 50 ± 2% RH and 40 ±
2°C for 24 hours. Two replicas per film were tested.

Thermal stability of shellac films was made using Mettler apparatus, with heating plate Hot stage FP 82 HT.
Measurement conditions were: starting temperature 25°C, heating speed: 2°C/min and end temperature 150°C.
For each sample 25 measurements have been done.

Tensile strength (TS) and elongation at break (E) of the films were determined on a tensile testing machine
Instron 6022. The samples were analysed in standard atmosphere 23°C ± 1°C of temperature and 55% ± 2% of
relative humidity. The cross speed head was 0.15 mm/s. The films of 6 cm in length and 0.7 cm in width were
used, and a minimum of five probes for each sample was tested. During the sample stretching, several load
and elongation data per second were recorded until a break of the sample occurred.

The SEM micrographs of film surfaces were taken with a scanning electron microscope (JSM -6060 LV). The
instrument was operated at 10 kV, at the magnification 400×.

3 RESULTS & DISCUSSION

Shellac is one of the thermosetting resins of animal origin. The functional properties of films were investigat-
ed as function of thickness, water vapour permeability, tensile properties and plasticizer content (Table 1).
The thickness of the films influences water vapour properties. For food packaging materials it is great im-
portance to achieve water vapour permeability (WVP) as low as possible, where high WVP determines poor
moisture and water barrier properties. In our research, the thickness (mean values) of analysed samples were
used in the calculations for WVP.

Table 1. Thickness, moisture content, water vapour permeability (WVP) and thermal stability of Shellac films,
with two different percentage of PEG in Shellac films.
Percentage of
Moisture con-
plasticizer PEG Thickness WVP Thermal
Sample tent
200 (µm) (g /m2∙day) stability (°C)
(%)
(%)
Shellac_4 4 233 ± 3.0a 7.13 ± 0.6a 4.78 ± 0.2b 149.5 ± 2.6c
Shellac_2 2 229 ± 1.5a 9.66 ± 0.5a 6.98 ± 0.7b 102.8 ± 3.7c
a
Means of five replicas ± standard deviation; b Means of five replicas ± standard deviation
c
Means of ten replicas ± standard deviation

As expected, the thickness of the sample with higher amount of plasticizer, was higher compared, to the other
sample. The water vapour permeability of films for food packaging should be as low as possible, where high
WVP determines poor barrier properties. The thickness of films influences water vapour properties. In our
research, the thickness (mean values) of films was used in the calculations for water vapour properties. Table
1 shows a comparison of films with different ratio of added plasticizer, and it can be seen that film with 4% of
PEG had lower WVP compared to sample with 2% PEG. From the obtained results, it can be seen that plasticizer
improved water and moisture barrier properties.

Thermal stability of both samples has shown that over 100°C is the temperature at which films degrade. Film
sample with 2% of PEG has melted at 102°C, but 4% of PEG shown improvement in thermal stability of Shellac.
With higher amount the stability increased, which has proven, that it could be a good solution.

Great mechanical properties such as tensile strength are very important for packaging films, due to the han-
dling and shipping of packaged products. At the same time, elongation at break – flexibility is also very im-
portant. For such products, high tensile strength is required, but deformation, elongation should be adjusted

8th Conference on Information and Graphic Arts Technology 81

 Oral presentations

according to the characteristics required for certain film applications. Typical stress-strain profiles for each
film formulations have been determined. Results from the shellac film demonstrated (Figure 1) that higher
amount of plasticizer increased the elongation at break, but the tensile strength decreased. Nevertheless, the
elongation at break was better because of the increase in solubility and homogeneity of all film solutions.

Figure 1. Tensile strength and elongation at break for both samples with 2% and 4% of PEG.

SEM micrographs of films are presented in Figure 2. The 10kV voltage was used since at higher voltage, film
samples would get damaged very quickly and the determination of the surface would not be correct. The same
was at 400× magnification, as higher magnifications degrade the surface of analysed films. Form the figures (a)
and (b) it can be seen, that sample with 2% has not so even surface, compared to sample with higher amount
of plasticizer. The surface micrographs revealed a smoother, even surface at this film sample. Lower concen-
tration of plasticizer caused less uniform, smooth and dense structure

Figure 2. SEM micrographs of Shellac samples with (a) 2% and (b) 4% of PEG.

4 CONCLUSIONS

Shellac has an excellent film forming and barrier properties. It is soluble in alcohol and alkaline solutions.
Therefore it has been widely used in the food and agro industries for gas, moisture, water and microbial pro-
tection of food products. The nature of films from biopolymers is mostly hydrophilic; therefore, the thickness
influences water barrier and mechanical properties.
In our research, Shellac films were successfully prepared. The results have shown that PEG had good impact on
film properties. With higher amount elongation increased, as well as thermal stability and moister barriers.
In addition, in blend films, a decrease in water vapour permeability was detected and this could be explained
with the addition of higher PEG content.

8th Conference on Information and Graphic Arts Technology 82

 Oral presentations

One of the issues to be taken into account in further research on enhancing water barriers and tensile proper-
ties of shellac films, to prepare bi-layer or laminated films with other bio polymer components, which ensure
more effective barriers against water transfer and better mechanical properties.

ACKNOWLEDGEMENT
The author would like to thank Mr. Manfred Penning and Mr. Stephen Hall from the A. F. Suter &Co Ltd. Com-
pany, Essex, United Kingdom for the Shellac flakes.
This research was supported by the COST action CA15216 (ENBA), and University of Ljubljana, Faculty of Nat-
ural Sciences and Engineering, Department of Textiles, Graphic Arts and Design, Ljubljana, Slovenia.

5 REFERENCES

• Byun, Y., Ward, A. and Whiteside, S. 2012. “Formation and characterization of shellac-hydroxypropyl methyl-
cellulose composite films. “ Food hydrocolloids 27(2): 364-370.
• Farag, Y. and Leopold, C. S. 2009. “Physicochemical properties of various shellac types“. Dissolution Tech-
nology 16: 33-39.
• Soradech, S., Nunthanid, J., Limmatvapirat, S., et al. 2012. “An approach for the enhancement of the mechan-
ical properties and film coating efficiency of shellac by the formation of composite films based on shellac
and gelatin“. Journal of Food Engineering 108(1): 94-102.
• Srinivasa, P. C., Ramesh, M. N., Kumar, K. R., et al. 2004. “Properties of chitosan films prepared under different
drying conditions”. Journal of Food Engineering 63:79-85.
• The, D. P., Debeaufort, F., Luu, D., et.al. 2008. “Moisture barrier, wetting and mechanical properties of shellac/
agar or shellac/cassava starch bilayer bio-membrane for food applications. “ Journal of Membrane Science
325(1): 277-283.
• Vrabič Brodnjak, U. 2017. “Influence of ultrasonic treatment on properties of bio-based coated paper.” Pro-
gress in organic coatings 103: 93-100.

Corresponding author:
Urška VRABIČ BRODNJAK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 89, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 83

 Oral presentations

MONTE CARLO DENOISING OF COMPUTER GENERATED IMAGES

BY NON-LOCAL MEANS FILTERING

Helena GABRIJELČIČ TOMC, Aleš HLADNIK

University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia

ABSTRACT: In the present study Monte Carlo noise that had been introduced to various computer generat-
ed images was reduced by implementing the non-local means algorithm. Its performance, i.e. the degree of
noise reduction, was assessed using three image quality metrics in addition to performing subjective visual
evaluation of the denoised images.

Keywords: computer generated imagery, rendering, Monte Carlo, denoising, non-local means filtering.

1 INTRODUCTION

In the computer graphics pipeline, rendering is the process of creating a photorealistic or stylised final im-
age, where computations of objects and scene geometry, textures and shading algorithms, direct and indirect
illumination, camera and specific effects are performed. Final visualisation result is typically a 2D bitmap
image, in which pixels represent the complexity and correlation of all input parameters. Path tracing is a ren-
dering method that includes Monte Carlo (MC) integration (Zwicker, 2015) and considers indirect and global
illumination in their full potential. Integration implies the gathering of all possible light rays from different
directions and interacting with the surface of the 3D object. In a computer simulated light transport, all pos-
sible solutions of light travelling cannot be computed, consequently with the equiprobability the random
directions of the light are in the rendering workflow defined with samples. Here, two main interdependent
problems arise. First, for the sake of rendering realism the number of samples has to be extremely high in
order to generate a noise-free output image. Second, the amount of noise over the image is non-uniform, i.e.
spatially-variant. In the past, the challenge of removing the MC noise from CG renderings by preserving the
details of the image and sharp edges was approached by applying various filtering methods on bitmap images
and adopted to the pipeline of CG imagery generation (Rousselle, 2013). Recently, the researchers introduced
machine learning and artificial intelligence methods in MC rendering workflow. Kalantari et al. presented
an intelligent system based on artificial neural networks and a nonlinear regression model (Kalantari, 2015).
Chakravarty et al claimed that their novelty including the variation of deep convolutional networks could
solve the challenges of MC integration for real-time rendering (Chakravarty, 2017).
The objective of our study was a reduction of MC noise in a number of computer generated (CG) images – only
two will be shown – by implementing the non-local means (NLM) algorithm. Its performance was assessed
through computing three image quality metrics, as discussed in the Materials and methods section.

2 EXPERIMENTAL

2.1 Non-local means filtering

Non-local means (NLM) algorithm (Buades, 2005) is one of the more successful recent attempts to reduce noise
in an image while at the same time preserving important information, such as image structures, details and
edges. Many of the traditional image denoising attempts – Gaussian-, anisotropic-, bilateral-, Wiener filtering,
wavelet thresholding, total variation minimization and others – make several assumptions about the noisy
image being processed (Evans, 2005) that can eventually result in its undesirable blurring. These methods try
to separate the image into the smooth part – true image – and the part with the noise by removing the higher
frequencies from the lower ones. But in practice images are rarely completely smooth since they also contain
high-frequency fine details and structures. When high frequencies are removed, not only the noise will be
eliminated, but the high-frequency content of the true image as well. In addition, low-frequency noise might
still remain in the image.

8th Conference on Information and Graphic Arts Technology 84

 Oral presentations

Figure 1. Self-similarity in an image: neighbourhoods of pixels p and q1 are similar,

while those of p and q2 are not.

Unlike these methods, NLM is based on the concept of self-similarity appearing in an image. As shown in
Fig. 1, neighbourhoods of pixels p and q1 are very similar, while neighbourhoods of pixels p and q2 are not.
Consequently, pixel q1 will have a stronger effect on the denoised value of p compared to q2, i.e., the weight
w(p,q1) will be much greater than w(p,q2). So instead of replacing the value of p by averaging locally within
its neighbourhood only, the idea here is to look for pixels with similar neighbourhoods across the whole image
and take average (mean) values of these regions in order to calculate a new value of p.
NLM algorithm can be controlled by three adjustable parameters (Python, 2018). Patch size (PS) is the radius
of the neighbourhoods (patches) used to find the similarity between the pixel values. If this parameter is set
too high, no similar neighborhoods will be found, if it is too low, too many similar neighborhoods will be
detected. Patch distance (PD) is the radius of a search window, in other words, the maximum distance where
to search for patches. The third parameter, H, is the weight-controlling parameter; a higher value results in a
smoother image with less noise, but at the expense of blurring image features.

2.2 Image quality assessment

In the study several images (renderings) produced with Cycles, a path tracing Blender rendering engine were
investigated. Renderings were prepared with noise (no. of samples = 50) and without it (no. of samples = 2000).
Images differed in brightness, contrast, the number of salient structures and the amount of detail. Each of the
three NLM parameters (factors) – PS, PD and H – was investigated at three levels: PS at 1, 6 and 11, PD at 2, 7
and 12 and H at 0.02, 0.06 and 0.1. These values were chosen empirically based on extensive experimentation
with various CG images and settings.
In addition to a visual inspection of the resulting images containing reduced amounts of noise, three objective
image quality metrics were computed: PSNR, SSIM (Wang, 2004) and PSNR-HVS-M (Ponomarenko, 2007). Unlike
PSNR that compares the true, noise-free, and the resulting image pixelwise, SSIM and PSNR-HVS-M are based
on the properties and limitations of the human visual system and should according to the literature corre-
spond more closely to our visual experience and assessment of image quality. For each of the three metrics,
the higher its value, the better the image quality, i.e., the higher the resemblance of the noise-reduced image
to the ground truth (image without any noise) one.
The obtained results from a non-replicated 3 × 3 × 3 factorial design were interpreted using a three-way anal-
ysis of variance (ANOVA).

3 RESULTS & DISCUSSION

Grayscale versions of two ground truth images coded as ‘01’ and ‘05’ are, together with their noisy versions on
which NLM algorithm was implemented, displayed in Fig. 2. The degree of elimination or reduction of noise
artefacts using 27 combinations of three NLM parameters was measured by PSNR, SSIM and PSNR-HVS-M
metrics (Fig. 3). While PSNR and PSNR-HVS-M trends are for most of the images quite similar, SSIM results
for several parameter combinations show some striking differences compared to those obtained by the other
two metrics. This is e.g. the case with the images PS1_PD12_Hxx_01 and PS1_PD7_Hxx_01. These denoised
images (see Figs. 4a and b) that are obviously not at all affected by their actual level of the H parameter, are
characterised by very low SSIM and at the same time extremely high PSNR values. One possible explanation for
this unusual behaviour can be that the SSIM metric that tries to take into account some aspects of our visual
system heavily penalizes severe blurring that is evident in these images. On the other hand, the fact that noise
in these images was almost completely removed by the denoising algorithm was rewarded via a high mark
by a simple pixel-by-pixel metric such as PSNR. Note also that SSIM values for all ‘05’ denoised images were
higher than those for ‘01’ images, probably due to the fact that SSIM, unlike PSNR, takes into account image
contrast, which is more pronounced in ‘05’ images.

8th Conference on Information and Graphic Arts Technology 85

 Oral presentations

‘01’ and ‘05’ denoised images with the highest SSIM values were PS11_PD2_H0.06_01 and PS11_PD2_H0.06_05,
respectively (Figs 4c and d). It can be seen that the details (i.e. the sharpness) in these two images are pre-
served, while there is still a fair amount of noise present especially in the parts of the objects directly facing
the camera (image ‘01’) and in the shadows (image ‘05’).

01 01_noisy 05 05_noisy

Figure 2. Two ground truth (noise-free) and noisy images investigated in the study.

Factor means plots (Fig. 5) summarize rather complex relationships between the three NLM parameters and
their effect on image denoising performance. Results for PSNR-HVS-M are quite similar to those for PSNR and
are not displayed in this paper. The plots on the diagonals show the overall level means, i.e. main factor effects,
while the off-diagonal plots show the means for each pair of factor levels. These plots nicely reveal the two-way
interactions as demonstrated in Fig. 5e where an interaction between PS and H (see the top-right plot of Fig.
5c) is shown once again, this time with actual factor settings and PSNR values. With PS = 1, PSNR is entirely
independent of the level of the parameter H (PSNR = 31.2 = constant), while at the other two PS settings – PS
= 6 and PS = 11 – the denoising performance (= PSNR value) is strongly influenced by the corresponding level
of H, i.e. whether it is set to 0.02, 0.06 or 0.1. In general, significant interactions between two of the NLM pa-
rameters exist in all those cases (plots) where the three corresponding curves are not parallel to each other.

8th Conference on Information and Graphic Arts Technology 86

 Oral presentations

Figure 3. Image denoising results: SSIM (a), PSNR (b) and PSNR-HVS-M (c).

a) PS1_PD12_H0.06_01 b) PS1_PD7_H0.06_01 c) PS11_PD2_H0.06_01 d) PS11_PD2_H0.06_05

Figure 4. Four typical denoised images with extreme PSNR and SSIM values; see text for an explanation.

8th Conference on Information and Graphic Arts Technology 87

 Oral presentations

It should be noted that the SSIM factor means plots (Fig. 5a and b) are very similar for both investigated im-
ages while this cannot be said for the corresponding PSNR plots (Figs. 5c and d). This fact has to be somehow
related to the intricate differences in both images, such as contrast, the number of edges and other important
features, etc. but further work is clearly required to understand these discrepancies better.
ANOVA tables for SSIM and PSNR metrics are shown in Tables 1 and 2, respectively. They contain information
about the magnitude of main effects and contributions of two-factor interactions (see F-Ratio). A contribu-
tion (effect) is statistically significant if its P-value is less than 0.05 at the 95% confidence level. Similarity in
SSIM results (Table 1) for both images – significant effects of parameters PD and PS of similar strength, much
lower importance of the remaining four contributions – corresponds to the above mentioned resemblance of
Figs. 5a and b. With PSNR metric (Table 2), the situation is far more ambiguous: while for the image ‘01’ the
most significant effect is that of parameter PS, followed by H (their mutual interaction is also important – see
Fig. 5e), denoising of image ‘05’ seems to be most heavily influenced by the H settings, but PS-H and PS-PD
interactions are also not negligible.

Vertical scale = 0.0 to 0.16 Vertical scale = 0.0 to 0.4

Level Level
1 1
PS 2 PS 2
3 3

PD PD

H H

PS PD H PS PD H

Vertical scale = 20.0 to 32.0 Vertical scale = 24.0 to 32.0

Level Level
1 1
PS 2 PS 2
3 3

PD PD

H H

PS PD H PS PD H

32 PS
1
30 6
11
28
PSNR

26

24

22

20
0.02 0.06 0.1
H

Figure 5. Factor means plots; SSIM results for images ‘01’ (a) and ‘05’ (b),
and PSNR results for images ‘01’ (c) and ‘05’ (d).

Table 1. SSIM results: ANOVA table for images ‘01’ (left) and ‘05’ (right).

8th Conference on Information and Graphic Arts Technology 88

 Oral presentations

Table 2. PSNR results: ANOVA table for images ‘01’ (left) and ‘05’ (right).

4 CONCLUSIONS

Results show that the three NLM parameters affect the algorithm’s noise removal capability to various degrees,
depending on the characteristics of each investigated image. SSIM metric appears to correspond more closely
to subjective visual perception of image quality compared to PSNR since it rewards sharp edges and object
contours in the images. Further investigations are necessary to better understand the mutual interactions of
the NLM parameters and their influence on image quality metrics, such as SSIM and PSNR.

5 REFERENCES

• Buades, A., Coll, B. and Morel J. 2005. “A non-local algorithm for image denoising.” IEEE International Con-
ference on Computer Vision and Pattern Recognition 2(2): 60−65.
• Chaitanya Alla, C., Kaplanyan, S., Schied, C., Salvi, M., Lefohn, A., Nowrouzezahrai, D. and Alia, T. 2017. “In-
teractive Reconstruction of Monte Carlo Image Sequences using a Recurrent Denoising Autoencoder.” ACM
Transactions on Graphics 36(4): Article 98.
• Evans, W. 2005. “Image denoising with the non-local means algorithm.”URL: http://www.cs.wisc.edu/~ev-
answes/cs766.html. (last accessed on 20. 01. 2018).
• Kalantari Khademi, N., Bako, S. and Sen, P. 2015. “A Machine Learning Approach for Filtering Monte Carlo
Noise.” ACM Transactions on Graphics (TOG) 34(4): Article 122.
• Ponomarenko, N., Silvestri, F., Egiazarian, K., Carli, M., Astola, J. and Lukin, V. 2007. “On between-coefficient
contrast masking of DCT basis functions” (in German), CD-ROM Proceedings of the Third International
Workshop on Video Processing and Quality Metrics for Consumer Electronics VPQM-07, 25.–26. January 2007
• Python. “Scikit learn documentation.” URL: http://scikit-image.org/docs/dev/api/skimage.restoration.htm-
l#skimage.restoration.denoise_nl_means. (last accessed on 20. 01. 2018).
• Rousselle, F., Manzi, M. and Zwicker, M. 2013. “Robust Denoising using Feature and Color Information.” Com-
puter Graphics Forum (Proceedings of Pacific Graphics) 32(7), URL: https://pdfs.semanticscholar.org/442b/
f9c8b6c0aa5c8f8efd810bd080124c4c8735.pdf. (last accessed on 18. 01. 2018).
• Wang, Z., Bovik, A.C., Sheikh, H.R. and Simoncelli, E.P. (2004). “Image quality assessment: from error visibility
to structural similarity”. IEEE Transactions on Image Processing 13 (4): 600–612.
• Yoon, S.E. 2015. “Recent Advances in Adaptive Sampling and Reconstruction for Monte Carlo Rendering.”
Computer Graphics Forum 34(2): 667–681.
• Zwicker, M., Jarosz, W., Lehtinen, J., Moon, B., Ramamoorthi, R., Rousselle, F., Sen, P., Soler, C. and Yoon, S.E.
2015. “Recent Advances in Adaptive Sampling and Reconstruction for Monte Carlo Rendering.” Computer
Graphics Forum 34(2): 667–681.

Corresponding author:
Aleš HLADNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +3862003290, fax: +3862003270, e-mail: [email protected]
website: https://sites.google.com/site/aleshladnik/

Co-author(s):
Helena GABRIJELČIČ TOMC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 89

 Oral presentations

RECONGNITION OF WORDS WITH JUMBLED LETTERS

Nace PUŠNIK1, Primož WEINGERL1, Vladimir DIMOVSKI2 & Irma PUŠKAREVIĆ2

1
University of Ljubljana, Faculty of Natural Sciences and Engineering
2
University of Novi Sad, Faculty of Technical Sciences

ABSTRACT: Visual word recognition reflects the efficiency of word interpretation. Numerous studies propose
models to explain the underlying mechanisms of this phenomenon. Adding to this field of research are the
findings of one particular study that have circulated the internet and occupied the scholars’ interest for some
time. These findings provided support for the notion that jumbled letters in a word (apart from the first and
the last letter) provide no influence on reading speed, recognition speed, and reading comprehension. The
findings have been labelled as a hoax by scholars. We decided to conduct a research with the aim to investi-
gate the differences in reading performance between the words with non-jumbled and jumbled letters for two
languages (Slovene and English) across two background colours (white and grey) through the analysis of eye
movements. The findings of the study indicate that longer reading time is needed for the words with jumbled
letters in both languages and that more complex words provide greater number of fixations.

Keywords: reading comprehension, fixation frequency, reading speed, word recognition.

1 INTRODUCTION

Reading effectiveness is influenced by perceptual and cognitive processes such as visual word recognition. Cog-
nitive psychologists suggest numerous models for visual word recognition: whole word recognition approach
(Haber & Schindler, 1981; Reicher, 1969), serial letter recognition (Sperling, 1963), parallel letter recognition
(McClelland & Rumelhart, 1981; Rayner, Pollatsek, Ashby, & Clifton Jr, 2012). The currently acknowledged model
suggests that words are identified on the basis of their compositional elements–letters, where the letters are
not identified sequentially but parallelly. Additionally, letter processing i.e. letter identification efficiency
is interwoven with typeface design (Pušnik, Možina, & Podlesek, 2016) on the basis of its legibility through
familiarity (Beier, 2009; Nedeljković, Novaković, & Pinćjer, 2017).
In connection with this topic, scholars have explored the visual word recognition when letters in words are
transposed i.e. jumbled. These investigations were the result of a certain statement that circulated on the
internet suggesting that the order of the letters in a word is not relevant to the recognition process as long
as the first and last letter preserve the right place. An explanation of this phenomenon can be found in an
online article called The “Cambridge Effect” (Wren, 2005). The author of the article explains that so-called
“Cambridge Effect” is only partly true. Namely, it is easy to read and understand short words (e.g. hmuan;
human) with jumbled letters. However, it would be far too complicated or impossible to read longer words in
this way (e.g. recasreh; research). Even though the statement was later found to be a hoax, some researchers
took an interest in the matter. For example, Rayner et al. (Rayner, White, & Liversedge, 2006; White, Johnson,
Liversedge, & Rayner, 2008) report findings of their research on reading text with transposed letters and con-
clude that some variations of text in this regard are easier to read while others are not. Namely, their research
shows that reading speed of the text where words have transposed letters declines as transposition moves from
beginning letters, across final or ending letters, to internal letters. In comparison to the normal text, the text
with transposed letters will always have a cost to reading time and it will significantly vary on the position of
the letters. However, the choice of the typeface and text–background was not controlled in previous research.
Therefore, the purpose of this article was to investigate the effect of letter transposition on reading perfor-
mance in different languages and against different background colour through examination of eye movements.

2 EXPERIMENTAL

2.1 Participants

Eighty students from the University of Ljubljana participated in the experiment. Participants consisted of 50
women (62.5%) and 30 men (37.5%) with the average age of 22.1 years. All of them reported normal or correct-
ed-to-normal vision (32 participants reported the use of either glasses or contact lenses).

8th Conference on Information and Graphic Arts Technology 90

 Oral presentations

2.2 Stimulus design

Eight tests were performed that were divided in two groups. Four tests were conducted in the English and
four in the Slovene language. The tests in both languages had the same number of lines (12 for the English
language; 10 for the Slovene language). The content was the same for all the tests and it showed an extract
from the book The Little Prince. The only difference was in the length of the individual words that occupied
paragraphs in different languages. The line length of the paragraphs for both languages was kept constant and
was formulated as suggested by relevant sources, containing up to 75 characters (spaces included) (Bringhurst,
1992; Dyson, 2004; Ruder, 1981). All the tests were set in Georgia Bold typeface style with 29 px size (Arditi &
Cho, 2005; Franken, Podlesek, & Možina, 2014; Lund, 1997). The colour of the text in paragraphs was set in black
and the background colour in grey or white (Baetens, 2008; Dzulkifli & Mustafar, 2013; Hall & Hanna, 2004).
Table 1 shows the setups of the performed tests.

Table 1. The setups of the performed tests.

Size Background Typeface Number Characters/line
Test Typeface Language
[px] colour colour of lines with spaces
A grey 12
B grey 12
64.9 ENG
C white 12
D white 12
Georgia 29 black
E grey 10
F grey 10
67.7 SLO
G white 10
H white 10

2.3 Measures

The most common approach for measuring reading rate is tracking the eye movements (Rayner, 1998) due to
its reliability over the linkage to shifts in attention (Deubel & Schneider, 1996; Wedel, 2013). For this study,
two variables for measuring eye movements were chosen: number of fixations (critical parameter in deter-
mining search times) and total fixation duration (indicating cognitive effort), alongside the reading speed.
The collection of data on eye movements was executed with Tobii X120 eye tracker device. Eye movements
were assembled with 120Hz frequency and subsequently processed for calculation of eye fixation frequency
and fixation duration.

2.4 Procedure

Participants were asked to read silently from the screen and afterwards to answer series of questions which
were designed to verify if the text was actually read (Aberson & Bouwhuis, 1997; Beymer, Russell, & Orton,
2008; Weisenmiller, 1999). At the end of the experiment, there was a question which was a control question
for the content of the text. The role of the question was to check for comprehension to see whether the par-
ticipants silently read the text. The question referred to the word telescope in the sentence. All 80 participants
reported having acknowledged the control word. Therefore, the comprehension check-up showed that all the
participants understood the content of the text.

3 RESULTS & DISCUSSION

The investigation of the reading time of paragraphs of words with non-jumbled and jumbled letters revealed
that the overall shortest reading time (40.43 s) was for the text in Slovene language in the test where the colour
background was white (Table 2, Test H). The difference in reading speed between the words with non– jumbled
and jumbled letters was from 13 to 25 seconds apart in the favour of the non–jumbled text. The lowest differ-
ence, 13.2 seconds, was for words in English language where black text was presented on the grey background.
14.93 seconds was the difference for Slovene text (black typeface on white background). Next difference was
15.4 second for Slovene text; black letters on white background. The highest difference was noticeable for
English text where black letters were presented on the white background (24.2 seconds). Generally, texts with
non–jumbled letters were read faster in both languages but were read with different pace across different

8th Conference on Information and Graphic Arts Technology 91

 Oral presentations

backgrounds (SLO grey = 47.70, SLO white = 40.43; ENG grey = 58.46, ENG white = 52.60). On the other hand,
the data on reading rate for the texts with jumbled letters revealed a slightly different trend. The text in Slo-
vene language was read faster when set against the white background (SLO grey = 63.10, SLO white = 55.36)
whereas the text in English language was read faster when set against the grey background (ENG grey = 71.48,
ENG white = 76.80). The difference in reading rate between words with jumbled and non–jumbled letters for
both languages was expected due to the findings of previous studies. However, it was found that reading of
jumbled letters was faster in Slovene language. These results are understandable since all the participants
were Slovene native speakers and English language was their second language.
The analysis of reading speed revealed that the fastest reading rate was for the text in Slovene language with
non–jumbled letters set against the white background. The number of fixations, however, was not the smallest
in this case. Namely, the number of fixations was, expectedly, higher for words with jumbled letters. But the
results also show that the number of fixations was higher for the text set against the white background for
both languages (jumbled letters: SLO grey = 148, SLO white = 163; ENG grey = 173, ENG white = 250 / non–
jumbled letters: SLO grey = 106, SLO white = 125; ENG grey = 135, ENG white = 193). Even though the reading
rate was faster for texts set against the white background, the number of fixations was greater in this case in
comparison to the number of fixations for the texts set against the grey background. This can be explained
by the higher contrast in colour between the text and background. In this case eyes receive more input (more
detail) and use more fixations to search through details of the stimuli.
When the results of the total fixation duration were analysed, it was observed that the shortest fixation dura-
tions were for the texts in Slovene language regardless of the background. In regards to reading comprehension,
we could observe that comprehension was high for both languages regardless of the background colour and
jumbling of the letters. All the participants answered correctly when asked about the content of the read text.

Table 2. An average number of fixations, reading speed and total fixation duration for each test.
Language Letter order Colour NF RS TFD
Test
ENG SLO Jumbled Non-jumbled Background Typeface N [s] [s]
A • • grey 173 71.48 1.50
B • • grey 135 58.46 1.08
C • • white 250 76.80 1.14
D • • white 193 52.60 0.70
black
E • • grey 148 63.10 1.50
F • • grey 106 47.70 0.80
G • • white 163 55.36 0.90
H • • white 125 40.43 0.50
* NF. – Number of fixations; RS. – Reading speed; TFD. – Total Fixation Duration

Additional analysis of variance was performed through the series of one–way ANOVA tests to compare the
effect of letter transposition on reading performance in different languages and on different backgrounds.
The outcome variables were found to be normally distributed and equal variances are assumed based upon
results of Levene’s test and robust test on the equality of means. Analysis of variance that was conducted to
compare the effect of letter transposition on the number of fixations, in the case of English language and grey
background, showed a statistical difference between condition means [F(1,18) = 10.836, p = 0.004], effect size r
= 0.37. The main effect of the letter transposition on the number of fixations, in the case of Slovene language
and grey background, was also found statistically significant [F(1,18) = 17.332, p = 0.001], effect size r = 0.49.
There was no statistical difference in group means in the cases of English and Slovene on white background.
The results indicate that letter search is significantly greater for the words with jumbled letters when they are
presented on the gray background. This can be explained by the lack of contrast between the text–background
colour where eyes use more fixations for the search of incongruent patters in comparison to when the same
stimuli is presented against the white background.
Analysis of variance that was conducted to compare the effect of letter transposition on reading speed, in the
case of English language and white background, showed a statistical difference between condition means
[F(1,18) = 36.275, p = 0.000], effect size r = 0.67. The main effect of the letter transposition on the reading speed,
in the case of Slovene language and white background, was also found statistically significant [F(1,18) = 31.192,
p = 0.000], effect size r = 0.63. There was no statistical difference in group means in the cases of English and
Slovene on grey background. These results can be explained by suggesting that greater contrast between the
colour of the text and background influences faster pattern recognition i.e. eyes detect letter shapes faster.

8th Conference on Information and Graphic Arts Technology 92

 Oral presentations

Analysis of variance for the search of the effect of letter transposition on the third variable, total fixation du-
ration, showed a statistical difference between group means in all four conditions (p < .05) with large effect
size (r > 0.2). The results indicate that, nomatter the conditions, words with non–jumbled letters are processed
faster. These results are consistent with previous findings (Rayner et al., 2006; White et al., 2008).
Further analysis of the data shifted the focus of the initial aim of the paper towards the analysis of the reading
rate of longer words with jumbled letters (in comparison to those with non–jumbled letters). From all the
data gathered on the number of fixations of all the participants, an average number of fixations was calculat-
ed. Specifically, all the fixations from each test were combined with the heat map of each test. In this way, it
was possible to observe where the centres of the interest were and determine where in text participants had
difficulties in reading (either they had to return to the same point, or the word was not easily recognizable
because of the jumbled letters, or maybe the word was generally complicated even though it was presented
with non–jumbled letters). Figures 1 to 4 present centres of fixations (on the left jumbled letters, on the right
non–jumbled letters).

Figure 1. Main fixations in jumbled and non–jumbled letter order

for English text set against the grey background.

Figure 2. Main fixations in jumbled and non–jumbled letter order

for English text set against the white background.

Figure 3. Main fixations in jumbled and non–jumbled letter order

for Slovene text set against the grey background.

Figure 4. Main fixations in jumbled and non–jumbled latter order

for Slovene text set against the white background.

8th Conference on Information and Graphic Arts Technology 93

 Oral presentations

In this way, the maximum number of important fixations from each test was gathered (Table 3). For tests with
words set with jumbled letters, the number of average fixation points was 7 or greater and for tests with words
set with non–jumbled letters the number of average fixation points was 5 or fewer. Table 3 presents all the
words were the average fixation point in the test was.

Table 3. An average mean of the number of fixations in non–jumbled and jumbled letter order for all tests.
Test Fix. Word
A 7 taht, raelly, platens, Veuns, trhuogh, nmae, sireous
B 4 Others, telescope, throuhg, planet
C 8 scarcely, Mras, wihch, trehe, ohters, discevors, astromoner, tehm
D 4 Earth, hundreds, which, does
E 7 Verena, obsjata, durgih, mahnji, štelivko, inema, mali
F 5 Jupiter, teleskop, astronom, planet, asteroid
G 7 prenesetilo, pbsjata, durgih, mahnji, tevažo, sokzi, inema
H 4 Presenetilo, drugih, teleskop, asteroid

Table 3 shows were the word fixations were made. It can be observed that in the Test A participants had dif-
ficulties with words of different length–from four to seven letter words. Within the test where letters had
non–jumbled order (Test B) words were longer, consisting of more letters–from six to nine. Both tests were
performed with black typeface on the grey background. Test C shows the highest number of fixation centres.
Fixated words had from four to ten letters. Again, it can be observed that there was no connection between
the fixation centres and number of letters. The complexity of the word was found to be more important i.e. the
sequence of the jumbled letters. A similar trend, as in Test B, was observed in the Test C. Four fixation centres
appeared but this time the word length had greater range (from four to eight letters). Similar observations
were noticeable for the text presented in Slovene language. Fewer fixation centres could be found in texts with
non–jumbled letters while text with jumbled letters needed more fixation centres. On average, all the letters
in Slovene language, where fixation centres appear, were longer compared to English words.

4 CONCLUSION

The goal of this study was to investigate the reading performance of the words with jumbled and non–jumbled
letters, set in two different languages, against two different backgrounds by analysing the data gathered with
an eye tracking device. The findings indicate that reading rate of the text when words have jumbled letters in
both Slovene and English language is slower in comparison to the text when words have non–jumbled letters.
Through this investigation, it was revealed that texts, regardless of the transposition of the letters, were faster
read when set in Slovene and against the white background. Further investigation of the fixation centres within
words showed that there was no connection between fixation centres and the number of letters and that the
complexity of the word had a greater impact on reading rate.

5 REFERENCES

• Aberson, D. H. A., & Bouwhuis, D. G. (1997). Silent reading as determined by age and visual acuity. Journal
of Research in Reading, 20(3), 184–204.
• Arditi, A., & Cho, J. (2005). Serifs and font legibility. Vision Research, 45(23), 2926–2933.
• Baetens, J. (2008). Colour as a visual signifier in screen typography:“less means more.” Visual Studies, 23(3),
267–274.
• Beier, S. (2009). Typeface Legibility: Towards defining familiarity. The Royal College of Art. Retrieved from
http://researchonline.rca.ac.uk/id/eprint/957
• Beymer, D., Russell, D., & Orton, P. (2008). An eye tracking study of how font size and type influence online
reading. In Proceedings of the 22nd British HCI Group Annual Conference on People and Computers: Culture,
Creativity, Interaction-Volume 2 (pp. 15–18). BCS Learning & Development Ltd.
• Bringhurst, R. (1992). The elements of typographic style. Hartley & Marks.
• Deubel, H., & Schneider, W. X. (1996). Saccade target selection and object recognition: Evidence for a common
attentional mechanism. Vision Research, 36(12), 1827–1837.
• Dyson, M. C. (2004). How physical text layout affects reading from screen. Behaviour & Information Tech-
nology, 23(6), 377–393.

8th Conference on Information and Graphic Arts Technology 94

 Oral presentations

• Dzulkifli, M. A., & Mustafar, M. F. (2013). The influence of colour on memory performance: A review. The
Malaysian Journal of Medical Sciences: MJMS, 20(2), 3.
• Franken, G., Podlesek, A., & Možina, K. (2014). Eye-tracking Study of Reading Speed from LCD Displays: In-
fluence of Type Style and Type Size. Journal of Eye Movement Research, 8(1), 1–8.
• Haber, R. N., & Schindler, R. M. (1981). Error in proofreading: Evidence of syntactic control of letter process-
ing? Journal of Experimental Psychology: Human Perception and Performance, 7(3), 573.
• Hall, R. H., & Hanna, P. (2004). The impact of web page text-background colour combinations on readability,
retention, aesthetics and behavioural intention. Behaviour & Information Technology, 23(3), 183–195.
• Lund, O. (1997). Why serifs are (still) important. Typography Papers, 2, 91–104.
• McClelland, J. L., & Rumelhart, D. E. (1981). An Interactive Activation Model of Context Effects in Letter Per-
ception: Part I. An Account of Basic Findings. Psychological Review, 88(5), 375–407. http://doi.org/10.1016/
B978-1-4832-1446-7.50048-0
• Nedeljković, U., Novaković, D., & Pinćjer, I. (2017). Detecting universal structure and effects of typefaces.
Tehnicki Vjesnik - Technical Gazette, 24(2), 557–564. http://doi.org/10.17559/TV-20150831131738
• Pušnik, N., Možina, K., & Podlesek, A. (2016). Effect of typeface, letter case and position on recognition of
short words presented on-screen. Behaviour & Information Technology, 35(6), 442–451. http://doi.org/10.1
080/0144929X.2016.1158318
• Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological
Bulletin, 124(3), 372.
• Rayner, K., Pollatsek, A., Ashby, J., & Clifton Jr, C. (2012). Psychology of reading. Psychology Press.
• Rayner, K., White, S. J., & Liversedge, S. P. (2006). Raeding wrods with jubmled lettres: There is a cost.
• Reicher, G. M. (1969). Perceptual recognition as a function of meaningfulness of stimulus material. Journal
of Experimental Psychology, 81(2), 275.
• Ruder, E. (1981). Typographie: A Manual of Design. Hastings House Publishers.
• Sperling, G. (1963). A model for visual memory tasks. Human Factors, 5(1), 19–31.
• Wedel, M. (2013). Attention Research in Marketing: A Review of Eye Tracking Studies. Robert H. Smith School
Research Paper No. RHS, 1–28.
• Weisenmiller, E. M. (1999). A Study of the Readability of On-Screen Type. Virginia Tech.
• White, S. J., Johnson, R. L., Liversedge, S. P., & Rayner, K. (2008). Eye movements when reading transposed
text: The importance of word-beginning letters. Journal of Experimental Psychology: Human Perception
and Performance, 34(5), 1261.
• Wren, S. (2005). The “Cambridge Effect.” Retrieved from http://www.balancedreading.com/cambridge.html

Corresponding author:
Nace PUŠNIK
University of Ljubljana,
Faculty of Natural Sciences and Engineering, Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 77, e-mail: [email protected]

Co-author(s):
Primož WEINGERL
University of Ljubljana,
Faculty of Natural Sciences and Engineering, Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 69, e-mail: [email protected]

Irma PUŠKAREVIĆ
University of Novi Sad,
Faculty of Technical Sciences, Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
Phone +381 21 485 26 03, e-mail: [email protected]

Vladimir DIMOVSKI
University of Novi Sad,
Faculty of Technical Sciences, Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
Phone +381 21 485 26 25, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 95

 Oral presentations

THE DEVELOPMENT OF THE GENERATIVE BRANDS’ CATEGORIZATION

Anže KAVČIČ, Helena GABRIJELČIČ TOMC

University of Ljubljana, Faculty of Natural Sciences and Engineering

ABSTRACT: Modern computer programming is constantly progressing and some communication methods with
computers are developing to such extent that it is easier also for non-programmers to understand them, learn
them and create. More comprehensive syntax and semantic enable the designers exploring new approaches
to present brand identities. One of these approaches is generative brand identity, a subgroup of dynamic
brand identities. These generative brand identities are made with computer languages, and to present iden-
tity elements, an external variable is used. But they need at least an element that remains the same in order
to be recognisable and coherent. Now the companies are encouraged to use brand identities and logos that
are not limited to traditional graphic visual expression but aimed to promote the development of interactive
performances of generative brand identity and logos. This study gathers and compares two categorizations of
dynamic brand identities that are later used for creation of a new one, dedicated solely to generative brand
identities.

Keywords: Generative design, graphic design, programming, logo, brand identity.

1 INTRODUCTION

For the last several years generative design has created considerable excitement among professional designers.
Through the interplay of complex information with graphic design and programming, new and fascinating
visual worlds are emerging where the coincidental is shaped to help correlations become visible. The creator
can discover new capabilities of the design itself, creates something unique and different with the code that
cannot be achieved otherwise (Massingham, 2011). The possibilities of programming languages, mainly simpli-
fication of syntax and semantics, will change the role of the designer since computer art allows the audience
or user to co-create the final product, which attracts and awakens interest in the work itself (Ish, 2013). We
are expecting a paradigm shift in design that will lead to realms of visual imagery. Until now, designers have
used tools that programmers have developed for them, which has forced designers to adjust to their systems
(Bohnacker, Groß, Laub and Lazzeroni, 2012).

Generative logos use a generative way of presenting elements of identity, which means that they are constantly
changing. Visual symbols as a business trademark can be visually changing and vibrant. Traditional forms of
media are unable to stand out and attract the necessary attention of readers today. With the progress of digital
media techniques, which change people’s behavior in information collection and reading, corporations are
encouraged to use brand identity and trademarks that are not limited to traditional graphic visual expression
but aimed to promote the development of interactive performances of dynamic brand identity and trademarks
(Hsu, 2013).

In the theory of generative design, we explore the programming and the various options offered by the pro-
gramming languages. With the acquired knowledge, the designers became active in the production of pro-
grams and programming languages, and at the same time allow other designers easier access to the software.

Generative art, on the one hand, does not interfere with aesthetics, but rather it intends to deepen into the
disciplinary approach to it. This includes a human factor that outlines the implementation process (Curralo,
2014). Algorithms produce interactive and generative logos, which find a variety of solutions based on the
parameters that are set, which preserves the visual identity.

These brand identities should be adapted to more and more devices, and there are increasingly complex ideas
for presenting businesses, products, individuals, etc., it is difficult to present a multi-layered overall image
with just one logo. One solution to this problem are the generative logos and associated brand identities.
With their logical and systematic design structure, generative logos vary according to different environments.
Although they are more complex to design, the effect is stronger and more impressive. Generative logos are an
established, but not quite common practice in graphic design, which has not yet been thoroughly researched
and represent a new way of displaying the overall image.

8th Conference on Information and Graphic Arts Technology 96

 Oral presentations

The aim of the research was to review two existing generative brands’ categorizations that could enable a more
comprehensive understanding and use of these media and the proposal of one new approach. The experiment
included the selection of generative brands’ categorization, followed by analysing each one with the aim to
identify the (predominant) input variables that define a category and the definition of the used methods for
categorisation. Next, a new categorisation with six groups was preliminary introduced, based on analysis of
generative logos collected from the web, books and articles and new observations about their dynamic and
generative aspects.

2 OVERVIEW OF LITERATURE

Because of their relative recent introduction to the world of graphic design, the review of the references
reveals that there have not been many in-depth research analyses aimed at the performances of generative
brand identities and logos. There has been research made in the field of the use of a dynamic brand identity
(Felsing, 2010) that has shown that they are preferred by the organizations of the public and cultural sector.
They are attracted by the interaction between organization, content and visualization, which affects the quality
of the overall image. Apart from this one and the general categorization (Lin, 2007; Williams, 2001; Liu, 2009;
Krasner, 2008; Taylor, 2011; van Nes, 2012; Hsu, 2014), questions about the authorship of the final product
(Cox, McLean, Ward, 2000, McCormack, 2014, Paul, 2016), the aesthetics of randomness (Schönleib, Schubert,
2009) and raising questions about copyright protection (Postigo, 2012; Birdy, 2012) more detailed studies of
generative brand identities and logos were not found.

3 CATEGORIZATION OF DYNAMIC BRAND IDENTITIES

The main purpose of categorizing is to show the various ways how dynamic brand identity develops into the
final product. Research provides examples that can also be used to solve problems in design and other areas.
These can serve designers as inspiration for further work. New technologies allow to leap from paper to the
screen, where the use of dynamic brand identities is the widest.

The review of the refferences shown that there are only two relevant categorisation of dynamic brand identi-
ties, i.e. Hsu’s categorisation according to the time frame and Van Nes’s categorisation by use of the dynamic
identity.

3.1 Hsu’s categorisation according to the time frame

In the scientific paper, Visual Expression and Design Principles for Dynamic Brand identities (2014) Hsu pro-
poses that the expression forms are divided into four types based on four different timeframes of motion:

• Limited timeframe: Moving images are played within a limited timeframe such as films, animation, etc. This
way of presenting a logo is mostly adopted by the industry companies of communication media. But, similar
forms are gradually introduced to all industry companies due to the emerging technology.
• Rotational timeframe motion: Moving graphics are played rotationally. Only a small portion of companies
are using this way of presenting their dynamic logos.
• Non-fixed rotational timeframe motion: Moving images are randomly generated and played rotationally. A
non-fixed rotational motion graphic is a dynamic trademark that generates no fixed form, found randomly
under mobile carriers.
• Motionless logos: Trademarks present motionless static and dynamic images both. These logos contain no
limited timeframe animations or moving images with both static and dynamic trademarks that are versatile
and widely distributed in different industries.

This categorization aimed to analyse the visual expression forms of dynamic brand identities and their design
concept. Forty-four international dynamic brand identities were analysed using a case study method; in ad-
dition, the types of dynamic brand identities, visual expression forms, and design concepts were established
through discussion forums held three times by focus groups with five to seven participants per discussion.
Collected data was then analysed and put into categories.

The only criterion for this categorization of logos is the time, therefore, this categorization considers all the
dynamic brand identities, and there are no specific categories for generative logos. Rotational and linearly
developing logos between the generative ones cannot be traced because these logos are managed by the al-
gorithm. In this case, generative logos would fall only in the non-linear or static category.

8th Conference on Information and Graphic Arts Technology 97

 Oral presentations

3.2 Van Nes’s categorisation by use of the dynamic identity

Van Nes divides them into six types in her book Dynamic Identities: How to create a living brand (2012), where
she named them based on their behaviour:

• Container: a most obvious way of creating a dynamic identity, where with only one changeable variable
(colour, image ...) a great recognisable variety of logos can be created.
• Wallpaper: a common example of dynamic identities where a variable is placed behind the constant logo.
The shape can vary, but impression is still a single identity.
• DNA: the brand identity has several key elements that are interchangeable, thus creating a new image every
time.
• Customised: this allows the client to interact, be a part of the brand and customize its logo. This is the first
step towards letting identity reflect a certain sense of community, creating an emotional bond.
• Formula: where the whole system is a constant, whether this is a grid or a set of rules it forms a system that
brings everything together.
• Generative: lets the identity be influenced by external data (weather, news, tweets, number of visitors…)
and puts it in real time. The identity can adapt in response to its input.

For this categorization van Nes dissected the identity into six components: a logo, colour, typography, graphic
elements, imagery and language. Together they form a system that builds the identity of the brand and each
component helps sharpen the identity of the brand it represents. She researched nighty-three dynamic brand
identities and put them into six categories based on the analysis of visual behaviour of their components and
case studies of the identities. Within van Nes’ explanation, the functionality of the flexible elements receives
more attention than the actual strategic foundation of the visual identity.

This system allows a simple categorization of most the dynamic brand identities and gives designers new ideas
in finding solutions. However, it does not go into enough detail, especially for more technologically-supported
identities that are influenced by data or are completely automatic. Some categories, in a simple and obvious
way, change static visual image into dynamic ones. It is precisely this category of generative brand identities
that gives the need for categorization, as it mentions and does not define all the ways in which companies
can represent their employees or use a more personal way or get in touch with users.

In Figure 1 the procedure for identification of a new brand identity categorisation is shown.

Figure 1. The procedure for identification of a new brand identity categorisation.

4 RESEARCH METHODS AND CATEGORIZATION OF GENERATIVE BRAND IDENTITIES

In this study, we want to analyse and categorize generative logos and their conceptual background based on
the type of input variable that affects the change of the logo. The purpose of the study, based on thirty-six
generative logos collected from web, books and articles (van Nes, 2013; Bohnacker, Groß, Laub, Lazzeroni,
2012; Jochum, 2013; Hsu, 2014), is to present the importance of these logos in contemporary graphic design.
A case study of each logo was made then by the help of a discussion group consisting of four participants:
three professional designers and one linguist, aged 30-40, three men and one female, and a citation analysis.
Analytically presented research and development in this field, as well as the introduction of before mentioned

8th Conference on Information and Graphic Arts Technology 98

 Oral presentations

methodologies the collected data was analysed depending on the structure of the code and the input variable,
contributing to the new categorization of generative logos in this research field, which provides a thorough
understanding of this development field of generative art.

The acquired data has been put into six groups:

• Picture
Logos draw sources from colours and shapes found in provided images.
• Sound
Logos are responsive to music or voice and the changes occur according to the given parameters.
• Path
The solutions include the tracking devices and tracking movement of people or animals, and thus create a
logo with the route in given area.
• Human input
This topic is the most diverse, as the ways of entering information for the creation of logos vary, including
also instances of interactive logos where the person directly influences the change.
• Actual time
The logos of this group share the property that everything is happening in real time.
• Random
Here, the logos are randomly generated and have no input variable.

Some of the logos are multi-modular, especially in the case of human input where some logos have an addi-
tional variable. In these cases, the dominant variable was considered which determined into what category
the logo was put into.

The demonstrated results present many ways of influencing the final image and reveal various ways in which
the computer affects the final product. The presented categorization can serve to designers and programmers
as an inspiration for their work and facilitate communication between the designers and the users. Moreover,
the result of this analytical approach facilitates the users’ understanding of the impact of programming lan-
guages on the ultimate look of the product.

5 CONCLUSION

Today, generative logos are more accessible than ever because of less demanding software tools for their
production. They are widely used in television and internet adverts, where their ability to express uniqueness,
the involvement of users in the creative process and the ability to attract the viewer’s attention is used. Web
standards are evolving and becoming more and more powerful and advanced, and the capabilities of even the
most basic knowledge of programming languages are ​​ becoming incomparable with those five years ago. In
the research the review of three dynamic brand identities’ categorisations are presented. Hsu’s categorisation
according to the time frame and Van Nes’s categorisation by use of the dynamic identity were presented and
discussed. Based on some critical observation about the existing approaches, a categorisation specifically for
generative brand identity was proposed. The use of categorisation form (type) is not unambiguous, yet accord-
ing to their release in the research and professional publication, they certainly present researchers’ observation
according to methodology used and presumably also the evolving trends of generative art.

6 REFERENCES

• Birdy, A. 2012. »Coding Creativity: Copyright and the Artificially Inteligent Author.« Stanford Technology
Law review 5: 2-28.
• Bohnacker, H., Groß, B., Laub, J., Lazzeroni, C. 2012. Genertive Design. USA: Princeton Architectural Press.
• Cox, G., Mclean, A., Ward. A. 2000. »The Asthetics of Generative Code«. Paper presented at The International
Conference on Generative Art.
• Curalo, A. F. 2014. »Communication Design in the Information Age«. Paper presented at The International
Conference on Generative Art
• Hsu, C.-M. 2013. »Annotations of dynamic identities in interactive aesthetics«. Advances in Journalism and
Communications 1 (4): 41–49
• Hsu, C.-M. 2014. »Visual Expression and Design Principles for Dynamic Brand Identities.« The international
Journal of Visual Design 7 (2): 6-23

8th Conference on Information and Graphic Arts Technology 99

 Oral presentations

• Ish, A. 2013. »Re-Generate«. India: National Institute of Design

• Jochum, E. 2013. »Dynamic branding - Thesis.« MA Thesis, Vienna university.
• Kostner, M., Schönleib, C.-B., Schubert, F. 2009. »Chaos, Noise, Randomness and Coincidence as Constitutional
for Generative Art.« Bridges 2010: 467-470
• Krasner, J. 2008. Motion Graphic Design: Applied History and Aesthetics. New York: Focal Press.
• Lin, Y. A. 2007. »The research of animated logo«. Taipei: Fu Jen Catholic University.
• Liu, P. 2009. »Three expression forms of logo design-hyper-plane, dynamic and interaction«. Computer-Aided
Industrial Design & Conceptual Design IEEE 10th International Conference. November 26-29.
• Massingham, C. 2011. »Autorship & Control in Generative Design«. Great Britain: Craig Massingham
• Mccormack, J., Brown, O., Dorin, A., Mccabe, J., Monro, G., Whitelaw, M. 2014. »Ten Questions Concerning
Generative Computer Art« Leonardo 47 (2): 135-141.
• Paul, C. 2016. A Companion to Digital Art. Great Britan: John Wiley and Sons.
• Postigo, H. 2012 The Digital Rights Movement: The Role of Technology in Subverting Digital Copyright,
Cambridge: MIT Press
• Taylor, A. 2011. Design essentials for the motion media artists. Burlington: Focal Press.
• Van Nes, I. 2013. Dynamic identities - How to create a living brand. Nederlands: BIS Publishers.
• Williams, R. 2001. The Animator’s Survival Kit. New York: Faber and Faber.

Corresponding author:
Anže KAVČIČ
University of Ljubljana,
Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 40 171 998, e-mail: [email protected]

Co-author(s):
Helena GABRIJELČIČ TOMC
University of Ljubljana,
Faculty of Natural Sciences and Engineering,
Department of textile, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 91, fax: +386 1 200 32 70, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 100

 Oral presentations

THE EFFECTS OF TYPEFACE AND IMAGE COMPLEXITY ON CONSUMER' VISUAL

ATTENTION AND ATTITUDINAL RESPONSES IN ADVERTISING

Irma PUŠKAREVIĆ1, Uroš NEDELJKOVIĆ1, Ivan PINĆJER1, Gregor FRANKEN2 & Nace PUŠNIK2
1
University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design

ABSTRACT: The effects of semantic typeface design properties have been a subject of research mainly found
in the field of advertising. Unlike the functional properties, the semantic properties of typefaces can t: rigger
a cognitive and emotional reaction. These properties are often correlated with the complexity of the form.
When put in the context of print advertisements, the complex forms of typefaces are paired with different
levels of image complexity. Previous findings provide insufficient guidelines on this matter. Therefore, the aim
of this paper is to test the effects of typeface vs image complexity in print advertisements by analysing eye
movements and attitudinal responses. An experiment 2 × 2 between subjects’ factorial design was conducted.
The results indicate that the relationship between low image complexity and high typeface complexity can
positively influence consumer’s visual attention and attitude.

Keywords: typeface complexity, eye tracking, advertising, visual attention, visual complexity.

1 INTRODUCTION

Typefaces are determined by their functional and semantic properties (Henderson, Giese, & Cote, 2004). The
functional properties of typefaces relate to the attributes of the form that are universal i.e. the properties
that help distinguish one letterform from another (see Pušnik, Podlesek, & Možina, 2016). The effects of the
functional properties of typefaces have been extensively explored through the studies on the legibility of
type forms (Beier, 2009; Nedeljković, Novaković, & Pinćjer, 2017; Rayner, 1998). On the other side, the semantic
properties of typefaces relate to graphic descriptions which trigger cognitive and emotional reactions (G. R.
Morrison, 1986). In such instances, the formal attributes of a typeface establish semantic associations which
are often correlated with the complexity of the form (see Beier, Sand, & Starrfelt, 2017). The effects of such
typeface properties that are more elaborative and complex have been the interest of research in the field of
advertising effectiveness (Childers & Jass, 2002; Doyle & Bottomley, 2006). Still, a vast body of research in this
field focuses on the effects of the image i.e. pictorial part of the print advertisement (which has proven to be
the most frequently used medium in the print commercial propaganda). Particularly, existing studies have
addressed the effect of visual rhetoric of images in advertising on comprehension and elaboration (Gkiouzepas
& Hogg, 2011; McQuarrie & Mick, 2003; Scott, 1994), as well as originality and familiarity on visual attention
(Pieters, Warlop, & Wedel, 2002). However, the effects of the semantic properties of typefaces in relation to
the pictorial part of an advertisement have not yet been fully explored.

A particular feature of the form–that of visual complexity, embodies a range of implications that invite re-
searchers to investigate effects of visual complexity on consumer reactions (B. J. Morrison & Dainoff, 1972).
Namely, advertisements that contain more detail are considered more visually complex. Pieters et al. (2010)
propose the concept of feature complexity in advertisements which suggests that the image is more complex
when there are more detail and variations in the basic visual features such as colour, luminance and edges.
Their findings indicate that higher feature complexity may hurt visual attention. However, their inferences are
based on the sample of tested advertisements that were not controlled for typeface design. Therefore, their
research lacks concrete implications for typeface application in relation to image complexity.

Considering that findings of previous studies indicate that the effect of print advertisements depends on the
semantic properties of typefaces, which is defined by the complexity of visual elements (McCarthy & Mothers-
baugh, 2002; Puškarević, Nedeljković, Dimovski, & Možina, 2016), as well as feature complexity of images
(Pieters et al., 2010), the purpose of this paper is application of eye tracking methodology and attitudinal
responses to identify the effects of typeface vs image complexity in print advertisements.

8th Conference on Information and Graphic Arts Technology 101

 Oral presentations

2 EXPERIMENTAL

An experiment 2 (typeface complexity: low vs high) × 2 (image complexity: low vs high) between subjects’
factorial design was conducted. In total, 120 students from the University in Ljubljana took part in the study.

2.1 Independent variables

According to visual complexity theory (Donderi, 2006) and methodology of Pieters et al. (2010), measurements
of image complexity were collected. The starting point was the assumption that more detail and variations
of basic elements will need more computer memory to store the image. DSLR camera Canon EOS 5D Mark III
was used to make 20 images of various content. Images were shot in RAW, unstructured format. In the Camera
Raw Conversion software Adobe RGB 98 colour profile and 8 bits per channel were chosen. After this step,
images were saved as TIFF file. The JPEG algorithm is considered a standard for image compression (Wallace,
1992) and it was used as the measure of feature complexity in the study. In order for JPEG algorithm to pro-
vide relevant results, all the images in the sample needed to be of the same pixel dimension (e.g. 1920x1280).
Image compression resulted in image complexity scale with two endpoints: low feature complexity (656kb)
and high feature complexity (2255kb). Once the scale was determined, target images i.e. images to be used
in target ads (see Puškarević et al., 2016; Voss, Spangenberg, & Grohmann, 2003) were composed following
above–mentioned procedure. The target images depicting realistic visual structure across two levels of feature
complexity are shown in Figure 1.

According to Frutiger’s (1989) common skeleton (i.e. letter matrix) and Dixon’s (2008) description framework
for typeface classification, samples of typeface stimuli were created. The starting point was the letter matrix
which was found in the surface area of overlapped typefaces (Garamond, Baskerville, Bodoni, Excelsior, Times,
Palatino, Optima, Helvetica). Graphic application Adobe Illustrator was used in this process. Once the common
skeleton was determined, components under the label »formal attributes« in Dixon’s description framework
were used to make the typeface complexity scale. Eight formal attributes were superimposed on the letter
matrix in succession in the following order: weight, modelling, proportion, key characters, terminals, construc-
tion, shape and decoration. This method provided subjective typeface complexity scale that was correlated
with measurements from ImageJ software. Results from both methods aligned. Subsequently, combinations of
several formal attributes were used for typography design for the target advertisements. The target typography,
depicting low typeface complexity and high typeface complexity, is shown in Figure 1.

Figure 1. Target images.

The design of the target advertisements which is based on the objective measures of visual complexity i.e. the
image feature complexity and typeface complexity is shown in Figure 2.

8th Conference on Information and Graphic Arts Technology 102

 Oral presentations

Figure 2. Stimulus design–target advertisements. From left to right depiction of realistic visual structure
with: low image feature complexity and low typeface complexity; low image feature complexity and high
typeface complexity; high image feature complexity and low typeface complexity; high image feature com-
plexity and high typeface complexity.

2.2 Measures

Since eye movements (fixations) are considered indicators of visual attention (Deubel & Schneider, 1996), in-
formation on individual fixations was obtained through fixation frequency (spatial distribution of fixations)
and total fixation duration. For collecting data on eye movements Tobii X120 eye tracker device was used. Ad
(items: likable, favourable, interesting) and brand (items: good, favourable) attitude, as well as conative (items:
probably, possibly) attitude, were measured with a five-point Likert scale (Mackenzie, Lutz, & Belch, 1986).

2.3 Procedure

Participants were randomly assigned to each of the experimental conditions. There were four groups of par-
ticipants of which the two were the control group (observed low typeface complexity) and the two were the
experimental group (observed high typeface complexity). Before the start of the experiment, each participant
adjusted to the lighting conditions in the room after which he/she was exposed to 9-point calibration on
the screen. Introduction pages were then presented containing the test ad. After the participants confirmed
that he/she understood the procedure, a collection of four advertisements (one target ad and three filler ads)
followed. After each advertisement, Likert scales were displayed and participants were asked to mark their
responses. The participants viewed the collection of advertisements at their own pace.

3 RESULTS & DISCUSSION

To analyse data gathered through Likert scales (effects of typeface complexity on attitude), Mann-Whitney U
test was used. First, data on advertisements with low image feature complexity will be presented. The control
group rated advertisements lower on the likable dimension for the attitude toward the ad scale (Mdn = 4; Mean
rank = 27.07) than participants in the experimental group (Mdn = 4; Mean rank = 35.66), and Mann- Whitney
U value was found to be statistically significant U = 347 z = –1.976, p = 0.048. Also, on the favourable dimen-
sion the control group rated advertisements lower in the control group (Mdn = 4; Mean rank = 26.35) than in
the experimental group (Mdn = 4; Mean rank = 36.33). Mann- Whitney U value was found to be statistically
significant U = 325 z = –2.362, p = 0.018. The attitude toward the brand on the good dimension also showed
statistically significant result: control group Mdn = 3; Mean rank = 26.20; experimental group Mdn = 4; Mean
rank = 36.47; U = 321, z = –2.386, p = 0.017. Additionally, the purchase intention on the probability dimension
was found to be statistically significant: control group Mdn = 4; Mean rank = 38.36; experimental group Mdn
= 3; Mean rank = 27.80; U = 356, z = –2.321, p = 0.02. Main effects for other dimensions are not interpretable and
will not be discussed. The data of attitude ratings towards typeface complexity on advertisements with high
image feature complexity reveal no statistical significance. Descriptive statistics of the ratings on attitudinal
scales of all groups are shown in Table 1.

8th Conference on Information and Graphic Arts Technology 103

 Oral presentations

Table 1. Means (M) and Standard Deviation (SD) for Attitude Dimensions for advertisements in both control
(low typeface complexity) and experimental (high typeface complexity) groups across levels of image com-
plexity (low vs. high image feature complexity).
Image L Image H
Typeface L Typeface H Typeface L Typeface H
M SD M SD M SD M SD
Likable 3.67 0.88 4.13 0.83 3.69 1.01 3.72 1.07
Aad Favourable 3.73 0.82 4.22 0.66 3.79 0.98 3.76 0.95
Interesting 3.37 1.16 3.63 1.04 3.48 1.02 3.45 1.12
Favourable 3.93 0.78 3.91 0.73 3.76 0.91 3.66 0.89
Ab
Good 3.53 0.82 4.03 0.74 3.62 0.82 3.66 0.81
Probable 3.40 1.07 3.97 0.82 3.14 1.16 3.24 1.27
Pi
Possible 3.50 1.01 3.63 0.94 3.31 1.20 3.38 1.17
Image L – Low image feature complexity; Image H – High image feature complexity; Typeface L – Low typeface complexity; Typeface H – High
typeface complexity.

Series of One–way ANOVAs were executed for the analysis of the effect of typeface complexity on visual
attention. In the case of the advertisements with low image feature complexity, the main effect of typeface
complexity was found on both fixation frequency (F(1,55) = 8.678, p = 0.005)) and total fixation duration (F(1,50)
= 63.549, p = 0.000)). In the case of the advertisements with high image complexity, the effect of typeface
complexity was found not to be statistically significant. Descriptive statistics of the eye movements through
measures of fixation frequency and total fixation duration of all groups are shown in Table 2.

Table 2. Means (M) and Standard Deviation (SD) of Attention Measures for ROI typography.
FF TFD FF TFD
M SD M SD M SD M SD
Image H
Image L

control 11.68 6.63 2.79 1.63 control 16.86 10.89 4.70 3.35

experimental 18.00 9.30 8.48 3.25 experimental 19.58 10.93 5.58 3.23

FF–Fixation frequency (Number of Fixations); TFD–Total Fixation Duration (Total Looking Time in Seconds)

These results indicate that viewers pay more attention to advertisements that are composed of less complex
images and more complex typography. This relationship has proven to be beneficial for forming the positive
attitude toward the ad and brand, as well as the purchase intention. These findings are in line with the ex-
isting research on the effects of image complexity and typeface semantics in advertising. Also, the current
findings contribute to the existing pool of research. Namely, previous findings of Pieters et al. (2010) indicate
that image feature complexity can hurt visual attention, implying that low image feature complexity can help
visual attention. However, detailed guidelines on how to effectively pair the image and typography, as the
two decisive executional elements of every advertisement, have yet to be updated. The results from this study
contribute to these guidelines to a certain extent. The heat map analysis (Figure 3) shows a slight indication
that advertisements with high image complexity might also benefit from the more complex typography. To
pursue this further, an experiment with different levels of typeface complexity and a larger number of par-
ticipants would be needed.

8th Conference on Information and Graphic Arts Technology 104

 Oral presentations

Figure 3. Heat maps comparing the eye movements of the control and experimental groups.

4 CONCLUSION

The goal of this study was to examine the effect of typeface complexity in relation to image complexity in
print advertisements on visual attention and attitudinal scales. These effects were evaluated based on eye
movements, attitudes and intentions of the viewers under incidental exposure to an advertisement. According
to the findings, we can infer that typeface complexity has a significant effect on visual attention and attitude
of consumers in certain conditions. When paired with an image that is low in feature complexity i.e. contains
fewer detail and variations of basic elements, typeface complexity can influence consumer’s visual attention
and positively shape attitudinal responses. However, when paired with an image that is objectively complex
i.e. contains more objects and details, typeface complexity will not have any influence due to the extensive
visual “clutter”. Future research might consider testing the effects of typeface complexity when different types
of products are advertised or when an image in advertisement depicts unrealistic visual structure.

5 REFERENCES

• Beier, S. (2009). Typeface Legibility: Towards defining familiarity. The Royal College of Art.
• Beier, S., Sand, K., & Starrfelt, R. (2017). Legibility implictations of embellished display typefaces. Visible
Language, 51(1), 112–133.
• Childers, T. L., & Jass, J. (2002). All Dressed Up With Something to Say: Effects of Typeface Semantic Associa-
tions on Brand Perceptions and Consumer Memory. Journal of Consumer Psychology, 12(2), 93–106. http://
doi.org/10.1207/S15327663JCP1202_03
• Deubel, H., & Schneider, W. X. (1996). Saccade target selection and object recognition: Evidence for a common
attentional mechanism. Vision Research, 36(12), 1827–1837.
• Dixon, C. (2008). Describing typeforms: a designer ’s response. InfoDesign Revista Brasileira de Design Da
Informação, 21–35.
• Donderi, D. C. (2006). Visual complexity: a review. Psychological Bulletin, 132(1), 73–97. http://doi.
org/10.1037/0033-2909.132.1.73
• Doyle, J. R., & Bottomley, P. a. (2006). Dressed for the Occasion: Font-Product Congruity in the Perception
of Logotype. Journal of Consumer Psychology, 16(2), 112–123. http://doi.org/10.1207/s15327663jcp1602_2
• Frutiger, A. (1989). Signs and symbols: their design and meaning. Van Nostrand Reinhold Company.
• Gkiouzepas, L., & Hogg, M. (2011). Articulating a new framework for visual metaphors in advertising. Journal
of Advertising, 40(1), 103–120. http://doi.org/10.2753/JOA0091-3367400107
• Henderson, P. W., Giese, J. L., & Cote, J. A. (2004). Impression management using typeface design. Journal of
Marketing, 68(4), 60–72.
• Mackenzie, S. B., Lutz, R. J., & Belch, G. E. (1986). The Role of Attitude Toward the Ad as a Mediator of Adver-
tising Effectiveness : A Test Competing Explanations. Journal of Marketing Research, 23(2), 130–143.
• McCarthy, M. S., & Mothersbaugh, D. L. (2002). Effects of typographic factors in advertising-based persua-
sion: A general model and initial empirical tests. Psychology and Marketing, 19(7–8), 663–691. http://doi.
org/10.1002/mar.10030
• McQuarrie, E. F., & Mick, D. G. (2003). Visual and verbal rhetorical figures under directed processing versus
incidental exposure to advertising. Journal of Consumer Research, 29(4), 579–587.
• Morrison, B. J., & Dainoff, M. J. (1972). Advertisement Complexity and Looking Time. Journal of Marketing
Research, 9(4), 396–400.
• Morrison, G. R. (1986). Communicability of the emotional connotation of type. ECTJ, 34(4), 235–244. Retrieved
from http://link.springer.com/article/10.1007/BF02767404

8th Conference on Information and Graphic Arts Technology 105

 Oral presentations

• Nedeljković, U., Novaković, D., & Pinćjer, I. (2017). Detecting universal structure and effects of typefaces.
Tehnicki Vjesnik - Technical Gazette, 24(2), 557–564. http://doi.org/10.17559/TV-20150831131738
• Pieters, R., Warlop, L., & Wedel, M. (2002). Breaking Through the Clutter: Benefits of Advertisement Orig-
inality and Familiarity for Brand Attention and Memory. Management Science, 48(6), 765–781. http://doi.
org/10.1287/mnsc.48.6.765.192
• Pieters, R., Wedel, M., & Batra, R. (2010). The stopping power of advertising: Measures and effects of visual
complexity. Journal of Marketing, 74(5), 48–60. Retrieved from
• Puškarević, I., Nedeljković, U., Dimovski, V., & Možina, K. (2016). An eye tracking study of attention to print
advertisements: Effects of typeface figuration, 9(5), 1–18. http://doi.org/10.16910/jemr.9.5.6
• Pušnik, N., Podlesek, A., & Možina, K. (2016). Typeface comparison− Does the x-height of lower-case letters
increased to the size of upper-case letters speed up recognition? International Journal of Industrial Ergo-
nomics, 54, 164–169.
• Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological
Bulletin, 124(3), 372.
• Scott, L. M. (1994). Images in Advertising : The Need for a Theory of Visual Rhetoric. Journal of Consumer
Research, 21(2), 252–273.
• Voss, K. E., Spangenberg, E. R., & Grohmann, B. (2003). Measuring the Hedonic and Utilitarian Dimensions
of Consumer Attitude. Journal of Marketing Research, 40(3), 310–320.
• Wallace, G. K. (1992). The JPEG still picture compression standard. IEEE Transactions on Consumer Electronics,
38(1), xviii–xxxiv.

Corresponding author:
Irma PUŠKAREVIĆ
University of Novi Sad, Faculty of Technical Sciences, Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2603, e-mail: [email protected]

Co-author(s):
Uroš NEDELJKOVIĆ
University of Novi Sad, Faculty of Technical Sciences, Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2624, e-mail: [email protected]

Ivan PINĆJER
University of Novi Sad, Faculty of Technical Sciences, Department of Graphic Engineering and Design
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2624, e-mail: [email protected]

Gregor FRANKEN
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 65, e-mail: [email protected]

Nace PUŠNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000, Ljubljana, Slovenija
phone: +386 1 200 32 77, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 106

 Oral presentations

THE INFLUENCE OF BACTERIAL NANOCELLULOSE ADDITION

ON MECHANICAL PROPERTIES OF PAPER

Gregor LAVRIČ, Igor KARLOVITS, Tanja PLEŠA & Mija SEŽUN

Pulp and Paper Institute, Ljubljana, Bogišićeva 8, 1000 Ljubljana

ABSTRACT: Nanocellulose is in last period increasingly present in papermaking, mainly due to its positive
effects on the different properties of various products. In this research, the influence of bacterial nanocellulose
(which is less presented in papermaking) on mechanical properties of paper was analyzed. Nanocellulose was
firstly chemically and mechanically processed and then added into the laboratory paper sheets together with
a cationic starch as a retention agent. Nanocellulose has a positive effect on the tensile index of the samples.
The add-in has improved tear index up to 35% and had no significant impact on burst index and elongation
at break of samples.

Keywords: nanocellulose, bacterial nanocellulose, papermaking, paper properties.

1 INTRODUCTION

Nanocellulose is the novel form of the most ancient and natural based polymer on earth: cellulose (Mani-
kkam, 2018). It has exceptional strength characteristics on a par with Kevlar, a lightweight material used to
manufacture high-strength, durable composites. However, in contrast to Kevlar and other materials based on
fossil fuels, nanocellulose is completely renewable.

Especially because of its properties it is a topic of numerous researches in different fields. In last period it is
also increasingly used in the pulp and paper industry, where it is incorporated in various types of papers and
coatings in order to improve mechanical and other properties of products (Seppänen, 2014). Nanocellulose
can be obtained in many different ways. With chemical procedures (Saito et al., 2009), mechanical procedures
(Correia et al., 2016) or with the use of appropriate microorganisms (Jozala et al., 2016). Nanocellulose ob-
tained with the use of microorganisms is environmentally friendlier since it does not require large quantities
of chemicals or energy to be produced.

Bacterial nanocellulose is a nanofibrillar polymer produced by certain bacteria such as Gluconacetobacter

xylinus, as a bottom-up process of polymerization of glucose into cellulose. The bacteria excrete the formed
thick cellulose polymer gel consisting of fine-structured cellulose fibrils outside its cell wall. The width of
the bacterial cellulose fibrils is around 20 - 100 nm and they consist of even finer cellulose nanofibrils with
width around 2 - 4 nm. The purity of bacterial nanocellulose, its molar mass, the degree of polymerization and
crystallinity are high. The mechanical strength is typically very high, but at the same time, it is very elastic
and formable. Compared to cellulose nanomaterials originating from plants the water retaining of bacterial
cellulose is excellent, due to its highly porous structure and large specific surface area (Gama et al. 2016).

As far as the influence of bacterial nanocellulose on the mechanical properties of papers is concerned, the lit-
erature review showed that there are not many articles related to this topic. Basta and El-Saied (2009) studied
the efficiency of bacterial nanocellulose as a flame retardant and evaluating its behavior in paper production.
They found that adding bacterial nanocellulose leads to an improvement in the breaking length and burst
index of the paper. The research of Gao et al. in 2011 showed that with the increase of bacterial nanocellulose
dosage, the properties of the tensile index, tear index, burst index, and stiffness greatly improve, while the
porosity and the relative water absorption decrease. These promising results were the reason that we studied
this area in more detail.

8th Conference on Information and Graphic Arts Technology 107

 Oral presentations

Figure 1. (A) Surface scanning electron micrograph of bacterial cellulose

and (B) corresponding cross-section (Gama et al. 2016).

2 EXPERIMENTAL

In the first part of our study, bacterial nanocellulose was chemically and mechanically processed in order to
be integrated into the paper. Initially, it was thermally treated in various media in order to remove impurities
and extractives. Later mechanical treatment with a homogenizer and an ultrasonic probe was performed. All
samples were prepared with the same procedure. Only the homogenization time was varied (sample BC 4; 4
min., BC 6; 6 min., BC 8; 8 min).

Laboratory paper sheets were produced on Rapid - Köthen apparatus, in accordance with ISO 5269-2:2004.
For the laboratory paper sheets production commercially available bleached, eucalyptus fibers obtained by
sulfate procedure were used. Fibers were refined in accordance with ISO 5264-2:2002 in laboratory PFI mill
(1500 revolutions). Cationic starch as retention agent was added to the pulp in an amount of 0.5 % by weight
of absolutely dry mass of basic fibers.

Paper sheets without bacterial nanocellulose were produced as a reference (sample V1). Other sheets contained
20% of bacterial nanocellulose by weight of dry matter of basic fibers (samples BC 4, BC 6 and BC 8). All paper
samples were conditioned according to ISO 187:1990 before testing. All tests were carried out under 50% rela-
tive humidity at a temperature of 23°C. In accordance with the standards we determined the following charac-
teristics: grammage (ISO 536:2012), thickness (ISO 534:2011), density (ISO 534:2011), tear index (ISO 1974:2012),
burst index (ISO 2758:2014), elongation at break (ISO 1924-2:2008) and tensile index (ISO 1924-2:2008).

3 RESULTS & DISCUSSION

As it can be seen from Table 1, which shows the basic properties of laboratory-produced paper, the reference
sample V1, which had no added bacterial nanocellulose, reached the lowest values ​​of grammage and density
among samples. The values of​​ sample thickness do not differ significantly all deviations are within the toler-
ance of the methods. With bacterial nanocellulose addition, grammage and density of samples were under-
standably increased. The highest increase was observed in the BC 4 sample, to which the least mechanically
treated nanocellulose was added. Larger, less-treated particles were largely retained in the paper sheet and
consequently increased these two parameters.

Table 1. Basic properties of laboratory-produced papers.

Sample grammage [g/m2] thickness [mm] density [kg/m3]
V1 66.7 0.133 501
BC 4 79.0 0.128 616
BC 6 77.1 0.126 609
BC 8 77.8 0.134 579

From the Figure 2, it can be seen the addition of bacterial nanocellulose has a positive effect on the tensile
index of the samples. The add-in has improved fiber links and increased this value in the case of the BC 6
sample by approximately 17%.

8th Conference on Information and Graphic Arts Technology 108

 Oral presentations

Figure 2. Tensile index of samples.

A similar trend can also be observed in the case of tear index of samples shown in Figure 3. Bacterial nanocellu-
lose has improved fiber links and increased sheets density. This led to higher values of
​​ the tensile index, in the
case of BC 8 sample for approximately 35%. It is clear from the picture that the mechanically more processed
nanocellulose has improved the tear index to a greater extent.

Figure 3. Tear index of samples.

The bacterial nanocelluloses addition did not have a drastic effect on the burst index of the samples. This is
evident from Figure 4.

Figure 4. Burst index of samples.

8th Conference on Information and Graphic Arts Technology 109

 Oral presentations

Bacterial nanocellulose addition has slightly decreased only the values ​​of elongation at break, as it can be
seen from Figure 5. Due to stronger connections between short eucalyptus fibers, the stiffness of the paper
most likely increased slightly. This is the reason for slightly lower values especially at samples BC 4 and BC 8.

Figure 5. Elongation at break of samples.

4 CONCLUSIONS

Based on the results of the research it can be concluded that bacterial nanocellulose can improve characteristics
of the paper, since it had a positive effect on practically all the measured mechanical properties, except elon-
gation at break. It was found that the mechanical properties of the paper in which bacterial nanocellulose is
added are enhanced with more intensive mechanical treatment. While retention of nanocellulose in the paper
sheet is one of the key parameters, which effect on its function. For this reason, a relatively large amount of
nanocellulose was added during this research. With the use of a more efficient and tailored retention agent,
this percentage could be reduced, and it is expected that similar results would be achieved.

ACKNOWLEDGMENTS
The work was carried out within the RDI project Cel.Cycle: »Potential of biomass for development of advanced
materials and bio-based products« (contract number: OP20.00365), co-financed by the Republic of Slovenia,
Ministry of Education, Science and Sport and European Union under the European Regional Development
Fund, 2016–2020.

5 REFERENCES

• Basta, A., El-Saied, H. (2009). “Performance of improved bacterial cellulose application in the production of
functional paper”. Journal of Applied Microbiology 107(6): 2098-107.
• Correia, V., dos Santos, V., Sain, M., Santos, S., Leão, A. and Savastano Junior, H. (2016). “Grinding process
for the production of nanofibrillated cellulose based on unbleached and bleached bamboo organosolv
pulp”. Cellulose, 23(5): 2971-2987.
• Gama, M., Dourado, F., Bielecki, S. 2016. Bacterial nanocellulose: from biotechnology to bioeconomy. Am-
sterdam, Netherlands: Elsevier.
• Gao, W., Chen, K., Yang, R., Yang, F., & Han, W. (2010). Properties of bacterial cellulose and its influence on
the physical properties of paper. BioResources, 6(1), 144-153.
• Jozala, A., de Lencastre-Novaes, L., Lopes, A., de Carvalho Santos-Ebinuma, V., Mazzola, P., Pessoa-Jr, A., Grot-
to, D., Gerenutti, M. and Chaud, M. (2016). “Bacterial nanocellulose production and application: a 10-year
overview”. Applied Microbiology and Biotechnology, 100(5): 2063-2072.
• Manikkam, Vasambal. Nanocellulose: Nano in size, tremendous in strength and endless in applications.URL:
https://prescouter.com/2018/01/nanocellulose-applications/(last accessed on 05. 04. 2018).
• Saito, T., Hirota, M., Tamura, N., Kimura, S., Fukuzumi, H., Heux, L. and Isogai, A. (2009). “Individualization
of Nano-Sized Plant Cellulose Fibrils by Direct Surface Carboxylation Using TEMPO Catalyst under Neutral
Conditions”. Biomacromolecules 10(7): 1992-1996.
• Seppänen, Rauni. Use of nanocellulose for high performance papermaking products. URL: http://www.kcpk.
nl/algemeen/bijeenkomsten/presentaties/20140930-rauni-seppanen (last accessed on 16. 01. 2018).

8th Conference on Information and Graphic Arts Technology 110

 Oral presentations

Corresponding author:
Gregor LAVRIČ
Pulp and Paper Institute, Print and Packaging department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Co-authors:
Igor KARLOVITS
Pulp and Paper Institute, Print and Packaging department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Tanja PLEŠA
Pulp and Paper Institute, Print and Packaging department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Mija SEŽUN
Pulp and Paper Institute, Laboratory for Materials and Technologies in Papermaking
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 18, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 111

 Oral presentations

THE POSSIBILITY OF USING ARDUINO BASED WATER LEVEL SENSOR FOR TISSUE
PAPER ABSORPTION RATE AND CAPACITY MEASUREMENT

Igor KARLOVITS, Urška KAVČIČ, Tanja PLEŠA & Mija SEŽUN

Pulp and paper Institute, Bogišićeva 8, 1000 Ljubljana, Slovenia

ABSTRACT: The absorption rate and absorption capacity are one of the most important properties of the
tissue paper. The interaction between the liquids and the tissue paper structure is determined by the surface
angle and the rate of absorption which is defined by the porous structure of the paper. For some time there are
several published studies, standards and methods like the ISO 12625-8:2010, ASTM D-4250, INDA Test Method
IST 10.1. which are used by the industry besides pure experimental studies. In our research, we have assem-
bled an Arduino based water level sensing device, which responses were calibrated with a specific amount of
water. We have tested two 2 and one 3 ply tissue papers where we acquired the dynamics time and absorption
weight and compared them with results obtained by the ISO 12625-8:2010 Part 8: Water-absorption time and
water-absorption capacity, basket-immersion test method and the PDA – Penetration Dynamics Analyzer –
ultrasound method.

Keywords: tissue paper, absorption, penetration dynamics analysis, Arduino.

1 INTRODUCTION

The absorption rate and absorption capacity are one of the most important properties of the tissue paper. As
a speciality paper for wiping off liquids, it has to have high absorbance and bulk structure to pick up a large
quantity of water or other liquid. The interaction between the liquids and the tissue paper (towel or wipes)
are defined by the surface angle and the rate of absorption which is defined by the porous structure of the
paper. In absence of external pressure, the absorption rate is mainly determined by the capillary pressure,
liquid viscosity and density. The capillary pressure, in turn, depends on wettability, liquid surface tension and
pore structure (Thorman et. al, 2012). An ideal tissue is a compromise between a large pore structure with
high porosity to take in a large volume of liquid rapidly and a micro-fine pore structure that can lift liquids
to a large height and retain it against opposing forces. The largest source of porosity in a tissue is from the
volume between the tissue and a surface to be wiped, or between plies of a multi-ply tissue. An extensive
study of the characterization of absorbent flow rate in towel and tissue was researched by (Beuther et. al,
2010). In their study, they used Sherwood ATS radial absorbency tester and X-ray unit to find absorption rates
and direction. and their results indicate that the X-ray data confirms the validity of the radial test method,
and demonstrates that by defining an experimental procedure, accurate and repeatable data can be measured
by the more simple radial test method. Another testing method was proposed by a group of authors which
used horizontal gravimetric method (Loebker and Sheenan, 2011). A Capacity and Rate Tester (radial orifice
wicking instrument) was constructed where the sample is placed over a monofilament and centred over water
supply tube. Water is applied and the sample weight is recorded in the function of time. The results obtained
in this experiment had better repeatability that the rate and capacity measurement proposed by ATS and
CGATS methods. The apparatus was able to detect differences in plyes, but the basis weight influence had to
be improved. In a PhD thesis by Fabritius (2007) he developed develop an optical method for measuring liq-
uid penetration into the paper in all three directions. Underlying the optical approach was the idea that the
scattering properties of wetted and dry paper differ strongly, and optical measurement methods enable the
detection of dynamical changes in scattering properties. Two different optical devices were used streak-camera
and optical coherence tomography and the obtained measurements showed that the effect of three different
coexistent subprocesses related to paper wetting can be detected. Only the effect of liquid migration along
fibres could not be distinguished by the used methods. In his master thesis Hedberg (2015) developed a new
method for measuring the dynamic absorption properties in the thickness direction of a single sheet of tis-
sue paper which uses a camera to record the position of the front of a water flow passing on one side of the
sample and when it is absorbed through the paper sample. The absorption rate is estimated by determining
the time difference when the water passes the same position on both sides of the sample. Gabrielsson et al
(2014) used yet another method to investigate the dynamic liquid absorption in kitchen towels. Their method
has the basis in thermography where they have determined the absorption rate and the speed of absorption
for different ply and embossed towel kitchens. Their results indicate that the embossing structures had clear
effects, as well the sample compression. The spreading type of the conventional samples was oval shaped. For

8th Conference on Information and Graphic Arts Technology 112

 Oral presentations

some time there are several published studies, standards and methods like the ISO 12625-8:2010, ASTM D-4250,
INDA Test Method IST 10.1 which are used by the part of the industry. Some of the proposed methods were
changed due to measurement uncertainty. As a new low-cost approach, we have tried using Arduino based
water level sensing module to acquire paper penetration dynamics and absorption rate. The water level sensing
can be done by several methods like the float sensor method or by using capacitance or conductivity sensors
or probes. There are also other methods like ultrasound, optical method for the continuous measurement of
liquid level. In our research, we have used a contact current amplification method as a low-cost alternative
which can easily be assembled and used for quick water absorption determination.

2 EXPERIMENTAL

To test the different water (liquid) absorption measurement we have measured three different tissue paper
towels which are commercially available. Two towels were made from 2 ply materials and out of pure cellu-
lose, while the third sample was a 3 ply mixture tissue paper which contained 20% of recycled fibres. All of the
samples had different embossing and creeping which was taken as an intrinsic value between different meas-
urement methods and procedures. For the ultrasound method, we have used the Emtec PDA device which has a
built-in module for tissue paper absorption evaluation. The method is based on the ultrasound signal intensity
reading which is put through the water which is immersed in demineralized water.We have used a 2Mhz signal
and the value T (impregnation time) was evaluated. Also, the dynamics of the impregnation for a 30s period
of time was calculated to gain the dynamics curve. For the second method, we have used the ISO 12625-8:2010
Water absorption time and water absorption capacity basket-immersion test method. In this method, a 5,0±0,2
g test piece with a width of 76±1mm is put in a cylindrical basket with the weight of 3,0±0,1g made of stainless
steel wired. The basket with the sample is submerged into water container and the absorption time is evaluated
visually with a stopwatch (results in seconds), while the water absorption capacity Wa is calculated in grams
of water per gram of each test piece (basically difference before and after of limited time water absorption).
For the third method, we have used a water level sensor which was connected to Arduino Mega 2560 R3 board.
This module uses current amplification by a transistor. When the liquid level is high enough to conduct the
current between the base and the positive power supply, a certain amount of current is generated between
the base and the emitter. And in a meanwhile, an electric current is produced in a certain amplification factor
between the collector and the emitter and applied to the resistant in the emitter to produce a voltage. Then,
this voltage will be collected by an AD converter. The operating voltage is 2.0-5.0V (we have used 5.0V) and
the detectable depth is 40 mm. To get the amount of water absorbed we have calibrated the sensor response
signal of Arduino with the amount of water which was added in precisely defined dosage. After calibration, we
have immersed the tissue paper samples with 5 g for 30 seconds into the water tank in which the sensor was
deposited and the readings were memorized. From the difference in start and stop signal values the amount
in ml/grammes was calculated. From the individual readings, the absorption dynamics curve was constructed.
For all tests, 10 samples were used and the mean value and standard deviation was calculated.

3 RESULTS & DISCUSSION

The immersion time value T for the sample measured with the ultrasound method are presented in Table 1.
while the absorbtion curves are presented in Figure 1.

Table 1. T (immersion time) values for the samples.

Sample name Ply Material T mean Tδ
Happy Casa 2 100% cellulose 0,213 0,078
W5 2 100% cellulose 0,212 0,068
Floralys 3 Cellulose +20% recycled fibres 0,179 0,073

As we can observe the T mean value is very similar for the 2 ply samples even if they have different creeping
and embossing pattern, while the sample with 3 plys had the smallest T values.

8th Conference on Information and Graphic Arts Technology 113

 Oral presentations

Figure 1. Dynamics of the paper towels absorption measured by ultrasound.

From Figure 1. we can observe that the Floralys 3 ply sample had intensity curve values during the time of
absorption while the Happy casa 2 ply sample had the lowest intensity values and the sample W5 2ply has a
similar absorption like the 3 ply sample for the first 2,5 seconds after which the 2 ply sample has lower inten-
sity value from the 3 ply sample. This indicates that the Happy Casa sample absorbs very quickly the liquid
in comparison with other samples.

For the ISO method, we have used the prescribed procedure and the water absorption capacity and absorption
time is presented in Table 2.

Table 2. Water absorption capacity and absorption time determined by ISO 12625-8:2010.
Wa Wa δ At mean At δ
Sample name Ply Material
mean (g) (g) (s) (g)
Happy Casa 2 100% cellulose 10,37 0,32 10,64 3,67
W5 2 100% cellulose 6,04 0,39 19,80 2,21
Floralys 3 Cellulose +20% recycled fibres 8,014 0,4 21,37 3,29

From Table 2 we can see that there is a big difference in water absorption capacity and absorption time between
samples. The quickest absorption time had the 2 ply Happy Casa sample (twice quicker from other samples)
while it also had the largest water absorption capacity with a value of 10,37. As we can notice one operator
measured the time, but still large variations occur as a visual and stopwatch method is prescribed by the ISO
standard. No possibilities of the dynamics are possible as there is no way of collecting data without sensing
device.

To come around this problem we have tried and approach of modifying the ISO method with water level sens-
ing device. We have used the same grammage of paper to have reference value regarding the number of paper
sheets which absorb the water and have used the same time of immersing (the 30s) but without the draining.
We focused on as in real life application to the quantity of water which is taken off/from the surface as it is
what really counts. The dynamics of the absorption were determined from the sensor reading till it visually
was fully wetted and the amount of taken up water was calculated from the difference between the initial
water level sensed by the module and the level after lifting out the soaked sample. The dynamics of the paper
absorption is presented in Figure 2. while the calculated value of water uptake is presented in Table 3.

8th Conference on Information and Graphic Arts Technology 114

 Oral presentations

Table 3. Water absorption capacity and absorption time measured by water level sensor.
Mean Δ Calculated
Sample Wa δ At mean At δ
Ply Material of sensing Wa mean
name (g/ml) (s) (g)
level (g/ml)
Happy Casa 2 100% cellulose 36 4,22±0.86 0,32 6.22 0.66
W5 2 100% cellulose 22.5 2,64±0,42 0,39 7.44 0.52
Cellulose +20% recycled
Floralys 3 32 3.73±1.13 0,4 4.1 0.99
fibres

We can see that the mean time of full absorption (determined by hand feel) the Floralys 3 ply sample had the
quickest time and W5 2 ply had the slowest time. We have made a comparison with the ISO test method values
and as we can observe there is a very good correlation between the samples regarding absorption capacity
as the W5 had the smallest amount of absorbed water in ISO method and the other samples also matched
regarding ratios (higher/lower) with high R2 value of 0,93. With this, we can calculate with good accuracy the
amount acquired by the ISO basket immersion technique even that there is a difference in our method as the
samples are not dried off for 60 s, but just taken out of the water. For times comparison there are differences
for the Floralys sample which could be most probably due to the immersion technique. For the ISO method,
there was the free fall method in a horizontal position (length of the basket) while we pushed the tissue paper
from the bottom side to certain water level in our glass and applied forced immersion. The dynamics of the
absorption is presented in Figure 2.

Figure 2. Dynamics of the paper towels absorption.

From Figure 2., we can see that the W5 2ply sample had a slight rise due to immersion and almost linear curve
of absorption. The 2ply (Happy Casa) and 3 plies (Floralys) samples had a sharp drop after the first two sec-
onds with lowest values around 5 seconds after which the dynamics of paper absorption was rising for both
samples. The 2 ply Happy Casa sample had the sharpest drop from all the samples and in the 30th second had
almost the same value as the W5 2 ply sample. There is a difference between the time of full wetting (3 ply
casa had the quickest time) and lowest sensor value (2 ply Happy Casa) which is due to the amount of water
taken and needed to reach full wetting of the sample.

4 CONCLUSION

From the presented results regarding water absorption dynamics and capacity measurement, the proposed
method with the water level sensing using Arduino based platform is very useful and viable solution. There
is a high correlation with the basket immersion method regarding quantity and additional information re-
garding dynamics is obtained which is not possible in the ISO method. The ultrasound method, on the other
hand, does not give absorption capacity information. Regarding the low cost of the proposed system, we see
it very useful for quick laboratory checks, as other proposed methods are either expensive regarding needed
equipment, or are not very useful except in laboratory conditions.

8th Conference on Information and Graphic Arts Technology 115

 Oral presentations

ACKNOWLEDGMENTS
The work was carried out within the RDI project Cel.Cycle: »Potential of biomass for development of advanced
materials and bio-based products« (contract number: OP20.00365), co-financed by the Republic of Slovenia,
Ministry of Education, Science and Sport and European Union under the European Regional Development
Fund, 2016–2020.

5 REFERENCES

• Beuther, P., Veith, M., Zwick, K. 2010. “Characterization of Absorbent Flow Rate in Towel and Tissue” Journal
of Engineered Fibers and Fabrics 2/5 (2010):1-7
• Fabritius T., 2007. “Optical method for liquid sorption measurements in paper ”, PhD diss., University of Oulu
• Gabrielsson A., Vomhoff H., Tysén A. 2014. “Investigation of the Dynamic Liquid Absorption properties of
kitchen towels .” In Papercon 2014 Proceedings, 3137-3160, Nashville: TAPPI
• Hedberg S., 2015. “Development of a method for measurement of the dynamic absorption properties in the
thickness direction of tissue products. ”, Msc Thesis, KTH School of Chemical Science and Engineering (CHE)
• Loebker, D., Sheehan, J. 2011. “Paper Towel Absorptive properties and Measurement using a Horizontal Gravi-
metric Device.” In PaperCon 2011 Proceedings, TAPPI, 1210−1218. Covington, Kentucky: Curran Associates, Inc.
• Thorman, S., Ström, G., Hagberg A., Johansson, P. 2012. “Uniformity of liquid absorption by coatings - Tech-
nique and impact of coating composition.” Nordic Pulp and Paper 2/27 (2012): 459−465

Corresponding author:
Igor KARLOVITS
Pulp and Paper Institute, Print and Packing Department
Bogišićeva 8, 1000, Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Co-author(s):
Urška KAVČIČ
Pulp and Paper Institute, Print and Packing Department
Bogišićeva 8, 1000, Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Tanja PLEŠA
Pulp and Paper Institute, Print and Packing Department
Bogišićeva 8, 1000, Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

Mija SEŽUN
Pulp and Paper Institute, Laboratory for materials and technologies in papermaking
Bogišićeva 8, 1000, Ljubljana, Slovenia
phone: +386 1 200 28 18, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 116

 Oral presentations

USE OF IMAGE SEGMENTATION FOR WOVEN FABRICS’ POROSITY

AND ALPHA MAP DETERMINATION

Tanja Nuša KOČEVAR, Živa ZUPIN

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design

ABSTRACT: Correct visualisation of porous material in 3D computer renderings is important while it can re-
veal a material positioned underneath and thus significantly affects the rendered model’s appearance. Besides
complex models, accurate cloth appearance can be visualised with simple techniques including maps (normal,
bump, alpha, specular). Textile porosity and pore parameters can be defined with image segmentation and
measured with image analysis. The aim of the research was the definition of most suitable method of image
acquisition and segmentation for creation of the alpha map for 3D visualisation of porous woven textile
structure. In the results it is demonstrated that all tested factors (weave type, threads’ density, illumination
and threshold algorithm) influence open area and pore´s size of investigated woven fabrics, i.e. segmentation
data in thresholded images and consequently textile visualisations. Statistical analysis also proved significant
interactions between tested parameters (factors) and their influence on renderings.

Keywords: transmitted illumination, reflective illumination, porosity, image analysis, alpha map, 3D visuali-
sation.

1 INTRODUCTION

Realistic 3D visualisation of textiles plays important role in many industries, such as fashion, interior design,
entertainment, etc. One of the visualisation methods is the record of cloth’s appearance and its representation
with various 2D maps. Alpha map pictures the porous structure of the model. In textiles, all spaces filled with
air are considered as pores. Fabric porosity is in general defined as physical characteristic of fabrics which
shows the portion of air in the total volume of fabrics (Zupin, 2012). According to mathematical definition
and the theory of porosity, 2D and 3D models describing porous structures, can be applied. The simplest 2D
model of representation for horizontal porosity is pure geometry of the projection of yarns diameter and yarn
density and is calculated from the cover factor. The third dimension also called vertical porosity is defined
with vertical pores, which are formed in places of longer, floating parts, i.e. non-interlaced segments of warp
and weft yarns. The calculation of three-dimensional pores includes yarns diameters, distance between warp
and weft yarns and length of floating yarns, fabrics thickness (Dubrovski, 2000).
Textile porosity can also be measured with image analysis, where with the implementation of appropriate
algorithms on digitised images, the data about the area covered by pores and other pore parameters (i.e. pe-
rimeter, shape) can be calculated. These procedures are relatively demanding due to the fabric deformability.
Nevertheless, with the use of corresponding equipment and image acquisition procedures (illumination with
reflection and transmission), and with the development of a suitable thresholding algorithm, the methods can
be automated, and satisfactory results can be obtained (Kang, 2001; Cardamone, 2002; Tàpias, 2010; Aydilek,
2002).

1.1 Modelling the appearance of textiles

Properties that are important for textiles’ visual appearance and that are included in appearance-modelling are:
optical properties (reflection, scattering, transmission and absorption); porosity; colour (optical properties of
fibres and yarns and constructional parameters); texture and relief (type of weave, fibre and yarn construction
parameters, finishing); and specific properties (anisotropy, yarns and fibres with special effects and higher
translucency). The different types of properties are correlating and do not influence the final appearance in
isolation. The review of literature shows that there are many different approaches that can be used for mod-
elling the appearance of textiles. The most complex are
mathematical appearance models that were extensively studied and reviewed by Schröder et al. (Schröder, 2012).
These physically-based appearance models are divided into three main types of approaches: surface-based,
volumetric and explicit models. The cues that were defined and analysed in the study were translucency,
silhouette, light diffusion, the possibility of real-time rendering, scalability, integration scale and viewing
distance. On the other hand, in game, web, augmented (AR) and virtual (VR) reality solutions of textiles’ and

8th Conference on Information and Graphic Arts Technology 117

 Oral presentations

cloths’ performance, less expensive techniques are usually applicable, i.e. maps. These techniques are firmly
established in 3D animation workflow and in the production of static visualisations that include many objects
on the scene, which can be viewed by close, medium and far distance (Kočevar, 2017).
The aim of the research was to define the influence of defined parameters (weave type, threads’ density, illu-
mination during acquisition and threshold algorithm) on alpha map preparation and the definition of most
suitable method of image acquisition and segmentation for creation of the alpha map for 3D visualisation of
porous woven textile structure. Methods of transmitted and reflected light microscopy were used and com-
pared for image acquisition and three different algorithms for image thresholding were applied. The results
were analysed with multifactor statistical methods.

2 EXPERIMENTAL

For the research purpose, 8 different woven fabric samples were produced. Cotton warp and weft yarns were
of the same linear density for all fabrics 17 × 2 tex. Four different densities of warp and weft were chosen,
warp densities 22 and 29.3 ends/cm, and weft densities 15 and 20 picks/cm, which resulted in four different
groups of woven fabrics with densities 22/15 (samples d1), 22/20 (samples d2), 29.3/15 (samples d3) and
29.3/20 (samples d4) threads/cm. Two different weave types were chosen, plain weave (PL) and twill 1/3 (T).
All samples were woven with two threads in reed dent.
Fabrics’ images were acquired with optical microscope (Nikon SMZ800 with the limits of magnification of
1–8x and ratio of magnification 8 : 1) and photo camera (Nikon D600, with settings 1/25 and ISO whiteness
value 200). Images were acquired with software digiCamControl. Two different illumination techniques were
implemented, i.e. under and above the sample, consequently image data were captured with transmitted and
reflected illumination. Number of image samples for each technique and for each density of chosen weave
type was 10. Workflow of image segmentation and analysis was performed in ImageJ, where image processing
included scale setting, histogram equilibration and thresholding. Three different threshold algorithms were
used for image segmentation, i.e. Minimum, Otsu and Yen. These three algorithms were selected due to their
optimal performance in preliminary testing among all the proposed algorithms of the ImageJ software. With
image analysis pore size and the area covered with pores were calculated. Open area of used samples was
analysed. The experimental results were statistically processed by multifactor ANOVA at 0.05 significance
level. For the statistical analysis of image analysis of woven fabrics five independent variables – factors – were
chosen as displayed in Table 1. Each factor has two or three levels.

Table 1. Experimental design diagram.

Influencing parameter Factor Level
Plain weave (PL)
Weave W
Twill weave (T)
22
Warp density D1
29
15
Weft density D2
20
Transmitted (T)
Illumination L
Reflected (R)
Minimum (MIN)
Threshold algorithms Alg Otsu (OTSU)
Yen (YEN)

8th Conference on Information and Graphic Arts Technology 118

 Oral presentations

3 RESULTS & DISCUSION

In Table 2 the results of statistical analysis by multifactor ANOVA are presented for open area and average
pore’s size.

Table 2. Impact of factors on open area and on pore size of woven fabrics.
Open area Pore´s size
Source SS Df MS F-Ratio P-Value SS MS F-Ratio P-Value
MAIN EFFECTS
A:W 10,77 1 10,77 6,44 0,012 4,6E-05 4,6E-05 54,25 0,000
B:D1 3553,02 1 3553,02 2126,4 0,000 1,0E-03 1,0E-03 1220,06 0,000
C:D2 1081,57 1 1081,57 647,29 0,000 4,1E-04 4,1E-04 477,05 0,000
D:L 1065,50 1 1065,50 637,67 0,000 2,1E-05 2,1E-05 24,64 0,000
E:Alg 698,52 2 349,26 209,03 0,000 2,1E-04 1,0E-04 122,48 0,000
INTERACTIONS
AB 19,22 1 19,22 11,5 0,001 1,4E-05 1,4E-05 16,55 0,000
AC 2,45 1 2,45 1,47 0,227 2,8E-08 2,8E-08 0,03 0,857
AD 0,01 1 0,01 0,01 0,926 9,1E-08 9,1E-08 0,11 0,744
AE 12,02 2 6,01 3,6 0,029 1,5E-05 7,5E-06 8,85 0,000
BC 96,76 1 96,76 57,91 0,000 6,8E-06 6,8E-06 8 0,005
BD 310,77 1 310,77 185,99 0,000 5,8E-05 5,8E-05 68,67 0,000
BE 195,67 2 97,84 58,55 0,000 2,9E-05 1,5E-05 17,07 0,000
CD 12,58 1 12,58 7,53 0,006 4,1E-10 4,1E-10 0 0,983
CE 47,71 2 23,85 14,28 0,000 1,3E-05 6,5E-06 7,66 0,001
DE 143,74 2 71,87 43,01 0,000 4,3E-05 2,2E-05 25,31 0,000
ABC 10,99 1 10,99 6,57 0,011 1,2E-05 1,2E-05 13,89 0,000
ABD 0,45 1 0,45 0,27 0,606 6,7E-08 6,7E-08 0,08 0,780
ABE 0,58 2 0,29 0,17 0,840 2,4E-06 1,2E-06 1,39 0,250
ACD 0,08 1 0,08 0,05 0,825 4,2E-06 4,2E-06 4,88 0,028
ACE 27,42 2 13,71 8,2 0,000 9,0E-06 4,5E-06 5,29 0,006
ADE 7,30 2 3,65 2,18 0,114 1,0E-05 5,1E-06 5,99 0,003
BCD 29,44 1 29,44 17,62 0,000 1,7E-05 1,7E-05 19,59 0,000
BCE 2,72 2 1,36 0,82 0,443 5,2E-06 2,6E-06 3,04 0,049
BDE 42,17 2 21,09 12,62 0,000 3,7E-05 1,8E-05 21,7 0,000
CDE 47,22 2 23,61 14,13 0,000 2,9E-05 1,5E-05 17,21 0,000
RESIDUAL 548,06 328 1,67 2,8E-04 8,5E-07
TOTAL 13797,50 364 3,3E-03

In Table 2 statistical analysis multifactor ANOVA shows that all five tested parameters (weave (W), warp (D1)
and weft density (D2), illumination (L) and thresholding algorithm (Alg)) are significant factors and that
they actually influence open area of woven fabrics. The factors by statistically importance are warp density
(D1), weft density (D2) and illumination (L), algorithm (Alg) and at least important weave (W). Factors weft
density (D2) and illumination during image acquisition are a little more than 3-times less important factors
than warp density (D1), while thresholding algorithms are 10-times less important than the most important
factor D1 and factor weave is 330 times less important than warp density.
Besides, statistical analysis demonstrates that many interactions between and among factors influence open
area of analysed samples. The most evident is the statistical significance between results of warp threads
density (D1) and illumination (L), between warp and weft density (D1 and D2) and between thread’s density

8th Conference on Information and Graphic Arts Technology 119

 Oral presentations

(warp and weft) and threshold algorithms (Min, Otsu and Yen). Namely, the statistical results show that all
three algorithms impact the open area and there are statistically significant differences between them.
Pore’s size is also influenced by all five factors that were analysed in the experimental. The most prominent
factor is warp density (D1) followed by weft density (D2), then all three thresholding algorithms (Alg) and
the type of illumination during image acquisition (L). The weakest influence on the pore’s size is manifested
by the two weaves (W), i.e. plain weave and twill, the same as with open area. The analysis of the interactions
between and among the factors show that many of them are statistically relevant, however their significance
is much lower than the significance of the main influential factors.

In Figure 1 the statistical significance of differences between levels of the factors and their influences is pre-
sented for open area: a.) weave (W); b) warp density (D1); c.) weft density (D2); d.) illumination (L) during
acquisition and e.) threshold algorithm (Alg).

Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals

8,2 13,2 11,1

10,1
8 11,2

9,1
Open area

7,8

Open area
Open area
9,2
8,1
7,6 7,2
7,1

7,4 5,2
6,1

7,2 3,2 5,1

PL T 22 29 15 20
W D1 D2

a) b) c)
Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals

10,2 11,8

9,2
9,8
Open area

8,2
Open area

7,8
7,2

6,2 5,8

5,2
3,8
R T
MIN OTSU YEN
L
Alg

d) e)
Figure 1. Five influencing factors on Open area of woven fabrics.

In Figure 2 the statistical significance of differences between levels of the factors and their influences is pre-
sented for average pore’s size: a.) weave (W); b) warp density (D1); c.) weft density (D2); d.) illumination (L)
during acquisition and e.) threshold algorithm (Alg).

Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals

(X 0,0001) (X 0,0001) (X 0,0001)

57 76 74

55
66
64
53
Pores size
Pores size

Pores size

56
51 54
46
49
44
36
47

45 26 34
PL T 22 29 15 20
W D1 D2

a) b) c)
Means and 95,0 Percent LSD Intervals Means and 95,0 Percent LSD Intervals

(X 0,0001) (X 0,001)
56 7

54
6
Pores size

52
Pores size

5
50

4
48

46 3
R T MIN OTSU YEN
L Alg

d) e)
Figure 2. Five influencing factors on average pore´s size of woven fabrics.

8th Conference on Information and Graphic Arts Technology 120

 Oral presentations

Analysing 3D visualisations of woven samples using alpha maps generated with image segmentation process,
the influence of all five factors (illumination, weave, warp and weft density and use of various algorithms for
their creation) is identifiable. The visualisations are shown in Table 3. The influence of illumination can be
recognised through revelation of details. Small details, such as interlaced fibres in the void spaces of pores
are more emphasized on thresholded images, consequently on visualisations, recorded using reflected illumi-
nation. Differences between plain weave and twill are shown through distribution of pores on the visualised
surface. The “grid” of plain weave pores and the shape of pores on the plain weave fabric are more even than
in the case of twill samples, due to longer floating threads and different types of pores present on the surface
of twill. The influence of varied densities of warp and weft threads is also very noticeable; pores’ sizes are
always larger when density is smaller, regardless weave, illumination and algorithm used were captured for
thresholding images. Considering the influence of algorithms used for alpha maps preparation, the variances
between sizes of pores and their shape are perceptible and the level of visible details as well. Pores are the
smallest when algorithm Min is used, and the largest when algorithm Yen is used.

Table 3. 3D visualisations of plain weave and twill 1/3 samples with various warp and weft densities (d1,
d3), where reflected and transmitted light illuminations for image acquisition were used and three different
threshold algorithms were applied for creation of the alpha maps.

4 CONCLUSION

In the presented research we confirmed the influence of the weave and the density of warp and weft threads
on porosity of the fabric respectively on pore’s size and on open area. We also established the importance
of illumination technique for the optical image acquisition and the use of chosen algorithm for image seg-
mentation for alpha maps preparation. Firstly, statistical analysis showed that all five investigated factors
influence open area and pore´s size of studied woven fabrics, i.e. segmentation data in thresholded images
(open area and pore’s size). Secondly, it was proven there are statistically significant interactions between
factors. The findings were correspondingly evaluated on 3D visualisations that confirmed evident differences
in the computer generated formations of textile structure. Further, the conclusions can also be summarised
with the next findings:
• The most important factor when analysing open area and pore´s size is warp density (D1).
• The weakest influence on open area and pore´s size has weave.
• All three threshold algorithms impact the open area and pore’s size and there is statistically significant
difference between all three algorithms (Min, Yen and Otsu).

8th Conference on Information and Graphic Arts Technology 121

 Oral presentations

• Between the two different illumination techniques used for image acquisition (transmitted and reflected
light microscopy) there is also statistically significant difference when analysing open area and pore´s size
of woven fabrics.
• Particularly at images visualised using alpha maps generated with images recorded with reflected illumina-
tion method and thresholded with algorithm Min, the level of details revealed is significantly higher than
at algorithms Otsu and Yen.

5 REFERENCES

• Aydilek, H.A., Oguz, H.S. and& Edil, B.T. 2002. “Digital Image Analysis to Determine Pore Opening Size. Dis-
tribution of Nonwoven Geotextiles.” Journal of Computing in Civil Engineering 16(4): 280–290.
• Cardamone, J.M., Damert, W.C., Phillips, J.C. and& Marmer, W.N. 2002. “Digital Image Analysis for Fabric
Assessment.” Textile Research Journal 72(10): 906–916.
• Dubrovski Dobnik P. 2000. “Volume Porosity of Woven Fabrics.” Textile Research Journal 70(10): 915–919.
• Kang, T.J., Choi, S.H., Kim S.M. and& Oh K.W. 2001. “Automatic Structure Analysis and Objective Evaluation
of Woven Fabric Using Image Analysis.” Textile Research Journal 71(3): 261–270.
• Kočevar, T. N., Gabrijelčič Tomc, H. 2017. “Modelling and visualisation of the optical properties of cloth.”
In Computer simulation, Computer and Information Science, Computer Science and Engineering, edited by
Dragan Cvetković, 45 – 65. Rijeka: InTech., Cvetković, D. (Ed.). Rijeka, 45 – 65.
• Schröder, K. , Zhao, S., Zinke, A. 2012. Recent Advances in Physically-Based Appearance Modeling of Cloth.
ACM SIGGRAPH Asia 2012: Course Notes, Nov. 2012. AvailableURL: http://cg.cs.uni-bonn.de/en/publica-
tions/paper-details/schroeder_clothcourse_2012/ (last accessed: 15.4.2018).
• Tàpias, M., Ralló, M., Escofet, J., Algaba, I. and& Riva, A. 2010. “Objective Measure of Woven Fabric’s Cover
Factor by Image Processing.” Textile Research Journal 80(1): 35–44.
• Zupin, Ž., Hladnik, A. & Dimitrovski, K. 2012. “Prediction of one-layer woven fabrics air permeability using
porosity parameters.” Textile Research Journal 82(2): 117–128.

Corresponding author:
Živa ZUPIN
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežiška 5, 1000 Ljubljana, Slovenija
phone: +386 1 200 32 17, e-mail: [email protected]

Co-author(s):
Tanja Nuša KOČEVAR
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežiška 5, 1000 Ljubljana, Slovenija
phone: +386 1 200 32 49, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 122

 Oral presentations

USER EXPERIENCE OF WAIT-ANIMATION PROGRESS INDICATORS

Jesenka PIBERNIK, Jurica DOLIĆ, Roberta STANIĆ & Lidija MANDIĆ

University of Zagreb Faculty of Graphic Arts, Getaldićeva 2, 10 000 Zagreb, Croatia

ABSTRACT: The research aim is to determine design parameters of animated progress indicators and present a
user study on how they are perceived in the context of mobile application. Users’ experiences of loader anima-
tion are generally more satisfactory when more real feedback is given. Informing users of the current working
state makes the process more tolerable to the user by reducing uncertainty. User’s perceived duration of time
passing can be manipulated by altering a speed and acceleration of the progress animation, altering display of
accuracy of status, altering perception with the application / brand design, using humour to alleviate anxiety,
using text to explain the process, etc.. . This paper research questions are: (1) Does providing feedback during
the wait prolong users’ tolerable waiting time in the context of shopping app? (2) How is user’s perception
of time correlated to the design and speed of the progress indicator?

Keywords: Progress bar, Animation, User Experience, Waiting Time.

1 INTRODUCTION

Animated progress indicators functions are: to reassure the user that the system is working and reduce the
user’s uncertainty, give the user something to look at while waiting, offer a reason to wait for the system to
finish and reducing users’ perception of time (Head, 2016). A user’s willingness to wait for a system to complete
an action depends on several factors, including: the urgency and complexity of the goal or task in mind; the
context of use, be it spending time in line on a mobile phone, or hurrying to get an important project uploaded
and users’ expectations based on prior experience with the app or similar processes (Scherwin, 2014).

Users’ experiences of loader animation are generally more satisfactory when more real feedback is given.
Informing users of the current working state makes the process more tolerable to the user by reducing uncer-
tainty. Informative loader can convey one or more messages that make waiting period much more pleasant, es-
pecially when it fosters the sense of anticipation. If loader’s animation and design are customized and intended
to metaphorically provoke a meaning, those meanings usually indicate future actions, product’s function or
brand identification. Viget’s experiment confirmed that branded loading experiences hold participants’ on the
loading page for longer, and have lower abandon rates than the non-branded, generic experiences (Tate, 2015).

User’s experience of the wait animation is also influenced by the mobile app micro-animation design, namely
the animation of the transition between previous and the next screen. The previous research results show that
in the transition between two images, bringing up the next image earlier dominates the perception of a fast
transition over other variables examined in the study (Huhtala at al, 2010). New content should be brought
up rather earlier than later despite of the effects of transition or overall duration.

User’s perceived duration of time passing can be manipulated by altering a speed and acceleration of the
progress animation, altering display of accuracy of status, altering perception with the application / brand
design, using humour to alleviate anxiety, using text to explain the process, etc. By making a loading screen
less generic and more novel, designer might be able to make time seem to pass more quickly or even make the
wait more pleasant. This paper research questions are: (1) Does providing feedback during the wait prolong
users’ tolerable waiting time in the context of shopping app? (2) How is user’s perception of time correlated
to the design and speed of the progress indicator?

2 EXPERIMENTAL

The research was done at Faculty of Graphic Arts University of Zagreb as part of master’s theses (Stanić, 2017).
On-line survey of 37 smartphone users, out of whom 63,2% women and 37,8% male, was used as user research
and user testing tool. 8,1% of participants fit into the category of age less than 20; 56,8% of participants fit
into the category of age 21 to 25 and 35,1% fit into the category of age 26 to 30.

8th Conference on Information and Graphic Arts Technology 123

 Oral presentations

Animated design of determinate and indeterminate loaders with different style and speed were created and
uploaded in video formats. The aim was to create animations that were comparable to the common generic
animation which populate the applications, like the spinning pinwheel or loading bar. Based on the previous
research, the animation’s duration was set to five seconds (Nah, 2004). The research was divided in four sec-
tions. Collected feedback was analysed to prove research hypotheses.

Section one of the experimental was set to determine whether the speed of loading animation has an influ-
ence on perceived wait time. Participants were presented with five loading animations of the same design and
duration (five seconds) but different speed. They were not aware that the animation’s duration was the same
and they had to rank animations according to the perceived time of duration.

In section two of the experimental, participants were presented with five animations of the same duration
(five seconds) but different speed and different design (Figure 1). The task remained the same: they had to
rank animations according to the perceived time of duration (Table 1.). Participants were also asked to rank
animations according to the aesthetic appeal and visual interest. Animation D was ranked as the most appeal-
ing and visually interesting.

Figure 1. Loader animation stimuli from the section 2 of the research.

From left to right: Animation A, Animation B, Animation C, Animation D, Animation E.

Table 1. Section 2 research question: Rank the animations according to the perceived duration.
longest shortest
speed 2 3 4
duration 1 duration 5
Animation A 24,3% 13,5% 21,6% 27,0% 13,5%
Animation B 5,4% 21,6% 43,2% 10,8% 18,9%
Animation C 21,6% 18,9% 27,0% 16,2% 16,2%
Animation D 8,1% 18,9% 24,3% 32,4% 16,2%
Animation E 27,0% 37,8% 13,5% 8,1% 13,5%

Section three of the experimental was designed to determine whether the perceived duration of animation
is different for determinate loading animation than for indeterminate (Figure 2). Participants were presented
with three groups of animations. The loaders in three groups differ in design. Each group consisted of one
determinate and one indeterminate loading animation. Indeterminate animations were designed to have
higher speed. Participants were asked to judge which loader animation takes longer time to finish. Overall,
the indeterminate loaders were perceived to take longer time to finish, compared to determinate.

8th Conference on Information and Graphic Arts Technology 124

 Oral presentations

Figure 2. Loader animation stimuli from the section 3 of the research: Group 1, time lapse of determinate
loading animation (left) and indeterminate loading animation (right) stimuli.

Table 2. Section 3 research question: Which loader animation takes longer time to finish?
Speed

Determinate Indeterminate The same

Group 1 6,3% 43,2% 40,5%
Group 2 19% 43,2% 37,8%
Group 3 24,6% 29,7% 45,9%

In section four of the experimental participants were presented with three groups of animated videos designed
as prototypes of shopping applications.
Videos had the same length (5 seconds). Each group consisted of two examples of user interface for the same
(shopping) application. One of the examples, shown in all three groups, was a loading spinner commonly used
for this kind of application. The second example was designed as more interesting, fluid and novel solution.
The speed of more interesting design was slower Participants were asked to choose a loader design which is
more visually interesting and appealing (Table 3). Also, they had to judge which loading time lasted longer
(Table 4) and which design is more suitable for the application.

Table 3. Section 4 research question: Which loader animation is more appealing and visually interesting?
Group1 Group2 Group 3
Common Interesting Common Interesting Common Interesting
design design design design design design
Appeal 18,2% 81,8% 29,7% 70,3% 27,0% 73,0%
Visual
5,4% 94,6% 24,3% 75,7% 16,2% 83,8%
interest
More suitable
51,4% 48,6% 48,6% 51,4% 48,6% 51,4%
for this app

Table 4. Section 4 research question: Which loader animation takes longer time to finish?
Speed

Common design Interesting design The same

Group 1 29,7% 45,9% 24,3%
Group 2 29,7% 43,2% 27,0%
Group 3 35,1% 27,0% 37,8%

8th Conference on Information and Graphic Arts Technology 125

 Oral presentations

3 RESULTS & DISCUSSION

The results of the section one of the experimental have proven that the speed of loading animation has an
influence on perceived wait time. Despite some individual variation, the quickest animation was ranked as the
one with shortest duration and the slowest animation was ranked as the longest.

In the second part of experimental, where participants were asked to rank different animations according to the
perceived duration, animation D achieved the best results. Animation D was also judged as the most visually
interesting and appealing. The findings bring to the conclusion that human perception of time is influenced
not only by speed but also by the design of animated loader.

The third’s section results have proved that the indeterminate loaders were perceived to take longer time to
finish, compared to determinate. Even though the speed of indeterminate loaders was faster than the speed
of determinate, the participants have declared that the animation’s duration of determinate loaders were
shorter. On the bases of results we can conclude that participants would be more willing to wait if presented
with determinate loaders than with indeterminate.

The fourth section results have showed that loader’s animation design is not the main contributing factor in
terms of participant’s perception of duration. Faster animations were perceived to have shorter duration even
if designed as less visually interesting and appealing. When questiond which design is more suitable for this
kind of applicatons, the participant’s answers were divided in half.

4 CONCLUSIONS

Visual feedback in the format of loading indicator helps users understand that system is working on their
request. Mobile applications utilize visually interesting animations when loading it’s content to prevent us-
ers from geting boared or uncertant. Loading animations keep the user occupied with visual feedback and as
a result, users perceive a shorter wait time.This research’s results are in line with previous research findings
about human perception of time (Harrison, 2010). The study made on Carnegie Mellon University proposed
and evaluated variations on two visual designs for progress bars that alter users’ perception of time passing,
and “appear” faster when in fact they are not. As a baseline, it used standard, solid-color progress bars, preva-
lent in many user interfaces. In a series of direct comparison tests, the study ranked how these augmentations
compare to one another. On the bases of results, the conclusion was made that progress bars with animated
ribbing that move backwards in a decelerating manner proved to have the strongest effect.

The study has proven that animations can be a good tool for making users perceive things as faster than they
really are. Faster animation of generic spinners create a sense of shorter waiting time. Users prefer determi-
nate progress indicators over indeterminate, even if the later are moving faster. Contrary to the expectation,
animation’s speed is perceived as more important factor for a positive user experience than visually interesting
and appealing animation design.

5 REFERENCES

• Harrison Chris, Yeo Zhiquan, Hudson Scott E. (2010): „Faster Progress Bars: Manipulating Perceived duration
with visual augmentations”, CHI ‘10 Proceedings of the SIGCHI Conference on Human Factors in Computing
Systems, pp 1545-1548.
• Head Val (2016): “Designing interface animation”, Rosenfeld, pp.118-123.
• Huhtala Jussi at al (2010): “Animated UI transitions and perception of time: a user study on animated effects
on a mobile screen”, CHI ‘10 Proceedings of the SIGCHI Conference on Human Factors in Computing Systems,
pp. 1339-1342.
• Nah Fui-Hoon Fiona (2004): “A study on tolerable waiting time: how long are Web users willing to wait?”
Behaviour & Information Technology, Vol. 23, No. 3.
• Scherwin Katie (2014): “Progress Indicators Make a Slow System Less Insufferable”, URL: https://www.nn-
group.com/articles/progress-indicators/, (last accessed on 12. 01. 2018).
• Stanić Roberta (2017): Design of UI animation for mobile applications. Master’s theses, University of Zagreb
Faculty of Graphic Arts.
• Tate Ian (2015): “Experiments in Loading - How Long Will You Wait?”, URL: https://www.viget.com/articles/
experiments-in-loading-how-long-will-you-wait, (last accessed on 12. 01. 2018)

8th Conference on Information and Graphic Arts Technology 126

 Oral presentations

Corresponding author:
Jesenka PIBERNIK
Department of graphic design and image processing,
University of Zagreb, Faculty of Graphic Arts
Getaldićeva 2, 10 000 Zagreb, Croatia
e-mail: [email protected]

Co-author(s):
Jurica DOLIĆ
Department of graphic design and image processing,
University of Zagreb, Faculty of Graphic Arts
Getaldićeva 2, 10 000 Zagreb, Croatia
e-mail: [email protected]

Lidija MANDIĆ
Department of graphic design and image processing,
University of Zagreb, Faculty of Graphic Arts
Getaldićeva 2, 10 000 Zagreb, Croatia
e-mail: [email protected]

Roberta STANIĆ
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 127

 Poster
section
 Poster section

AVAILABILITY OF E-TEXTBOOKS ON THE SLOVENIAN MARKET

Grega TOMAŽIN1, Urška STANKOVIČ ELESINI2

1
University of Ljubljana, Academy of Fine Arts and Design, Ljubljana, Slovenia
2
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Ljubljana, Slovenia

ABSTRACT: The introduction of ICT in the educational process means not only the introduction of new
tools, but also changes to the roles of the learner and teacher, the use of new teaching methods (e.g. flipped
learning), and consequently the new role of the textbook. The objectives of our research was to analyse the
Slovenian e-learning materials market, in particular e-textbooks. According to the results 39 i-textbooks have
been developed and approved for use in primary and secondary schools. Several publishers develop e-learning
materials that come in different forms such as digitalised printed versions of textbooks or workbooks having
multimedia and interactive elements. E-learning materials are also offered on different platforms. Some ma-
terials are active, while others are not. As we noticed through the research and as was communicated to us
through interviews, there are the problems associated with the continued use of e-textbooks, however those
problems are more pedagogical than technical.

Keywords: e-learning material, e-textbook, Slovenian market, publisher.

1 INTRODUCTION

The textbook represents the basic teaching material for the achievement of the educational goals and stand-
ards of knowledge defined in a specific curriculum. A textbook may take printed, electronic or printed and
electronic form. [1] The main difference between traditional and e-textbooks lies in the digitalisation of the
entire learning environment. In addition to traditional elements (text and graphical presentations), the latter
also include interactive elements (didactic applications, games, animations, etc.) and multimedia elements
(simulations, videos, sounds, etc.). As a rule, e-textbooks are accessible on the web regardless of time and place,
while that accessibility is dependent on the user’s internet connection. An e-textbook is a digital teaching
source or an asset stored on electronic media that can be used to replace printed media. Technical tools are,
however, required for the use of an e-textbook [2].

Pesek, Zmazek and Mohorčič [3] categorise e-textbooks to three levels. The first level comprises so-called dig-
italised textbooks or d-textbooks, which are electronic copies of printed textbooks in PDF or EPUB2 format.
D-textbooks can be read/displayed by applications that facilitate the addition of bookmarks, records and work
with interactive tables. The second level includes so-called rich textbooks or r-textbooks, which are d-text-
books upgraded with sounds and videos. Because some r-textbooks also have simple integrated questions that
provide simultaneous feedback to learners, they are quite popular among publishers. Interactive or i-textbooks
comprise the third level of e-textbooks. The content of such textbooks is adapted to human-computer inter-
action. I-textbooks include interactive elements (e.g. video, soundtracks, computer animations, hyperlinks
and, in the future, most likely augmented reality, which will be popular in presentations of practical samples)
and interactive examinations (instant and significantly improved feedback). They also facilitate the storage of
responses, the analysis of effectiveness and the monitoring of users.

The introduction of electronic learning materials (hereinafter: e-learning materials) in Slovenian primary and
secondary schools has proceeded under the auspices of governmental organisations. The Guidelines for the
evaluation of e-learning materials, which were linked with the Trubar catalogue for learning materials, were
prepared in 1995. The first tenders for smaller projects addressing the preparation of online e-learning mate-
rials were released after 1997. E-learning materials were prepared by certain individual teachers or groups of
teachers in the scope of the aforementioned projects [4].

In 2006, an action plan for continuing scholastic informatisation [5] was presented. The plan stated that the
role of learners is changing from passive to active, and that all ICT tools can be used. The role of teachers is
changing from that of knowledge “provider” to tutor and education coordinator, as well as moderator for
the evaluation of information. The existing services of the school system are changing, while new services
are being developed. The system is thus taking on a greater role in terms of the assessment and application

8th Conference on Information and Graphic Arts Technology 129

 Poster section

of development results, and in terms of connecting the system with the outside world (e.g. the economy).
According to the aforementioned plan, it was necessary to expand the offer of e-learning materials, and to
raise the level of those materials to modern, qualitative and (publicly) accessible e-materials that exploit the
opportunities presented by the media (e.g. interactivity and multimedia). From 2006 to 2008, the first calls
were made for projects addressing the preparation of extensive online e-learning materials. Those materials
were drawn up in accordance with relevant curricula. They included interactive and multimedia elements and
were freely accessible. [6]

Slovenia’s Development strategy for the information society [7] stated that the main subject of the informati-
sation of the learning and teaching process is the learner, while the teacher is a critical factor for success. The
teacher must accept advanced ICT, which will not replace or eliminate traditional learning, but will offer an
additional opportunity to change the learning and teaching process, making it more effective and attractive,
user-friendly and can be used anywhere at any time, and facilitates examinations, team work (web classrooms),
research work and remote learning.

In the period 2009 to 2013, the first extensive project was implemented under the name E-education. [8], [9]
Two smaller sub-projects were implemented in the scope of the aforementioned project: E-competent teacher
and E-support. After the project was completed, teachers and schools tested and evaluated e-learning materials
and e-services via two additional projects: E-textbooks for science classes in primary school, the objectives
of which were to upgrade existing e-materials to modern, freely accessible e-textbooks, to evaluate their
usefulness in the teaching and learning process and to identify requirements for their proper use (ICT infra-
structure, specific training for teachers, etc.) [10]. A new project under the name E-Schoolbag [11] was kicked
off after the successful implementation of the E-textbooks for science classes in primary school project. The
main objective of the E-Schoolbag project was to develop i-textbooks for the field of social sciences for the
eighth and ninth grades of primary school, and for the first year of secondary school (gymnasium). Another
objective was to ensure accessibility to and support for newly developed e-services and e-content. The project
was based on the three pillars of the 21st century school: the establishment of an e-learning environment, the
development of appropriate e-content and the training of e-competent teachers. In order to test e-content and
e-services, the participating teachers and learners were equipped with tablets, while all participating schools
were equipped with a wireless network. The E-Schoolbag project was completed in November 2015. In January
2016 [10]. In 2014, 44 schools were chosen to participate in a pilot project under the name Testing e-learning
contents and e-services. Learners were equipped with tablets or computers provided by their schools. During
testing, the main emphases were on planning lessons, focusing on the added value of achieving objectives
through the use of ICT, and on evaluating learners’ achievements.

Projects relating to the e-learning environment in primary and secondary schools have been mostly completed.
However, the search continues for efficient and lasting systemic solutions for the editing and maintenance
of existing e-materials and e-services, while technical support will also be required as new ideas emerge and
are implemented [12]. Consequently, new projects have already been launched this year. As was concluded
in a study by Hopson, Simms and Knezek [13], the positive effects of using technologies in teaching lead to
learners’ desire to continue using those technologies, ever after a project is completed.

In terms of higher education in Slovenia, the faculty members of universities have implemented several indi-
vidual and integrated projects that mainly involved the installation of the ICT infrastructure. Some members
have participated in various (particularly European) projects in which they secured funding for the estab-
lishment of key elements of e-learning. Some members were also involved in the aforementioned projects as
advisers and to develop specific activities. According to the Development strategy for the information society
until 2020 [14], support should be given to projects aimed at the modernisation of higher educational didac-
tics, through the requisite use of ICT and the transition to digital education. Innovative and flexible forms of
teaching should also be established. Consequently, a call was issued in the year 2017 for the inclusion of ICT
into the higher education pedagogical process for a period of four years.

2 EXPERIMENTAL

Our research was focused on the Slovenian e-learning materials market, in particular e-textbooks. For that
purpose, we searched the internet sources/pages of different publishers, ministries, organisations, etc., and
contacted some in search of reliable answers. During our research, we analysed 10 different publishers, e-text-
books issued by the National Education Institute Slovenia and different e-platforms containing e-learning
materials. In this research, some publishers were also contacted to get exact answers about their e-materials.

8th Conference on Information and Graphic Arts Technology 130

 Poster section

After we analysed e-material from mentioned sources, new questions arose about the real use of e-textbooks,
evidence and a vision for the future. Answers were obtained through interviews with people responsible for
Slovenian e-content projects.

3 RESULTS & DISCUSSION

According to the Rules on the approval of textbooks [1], textbooks may take printed, electronic or printed and
electronic form. According to the aforementioned rules, e-textbooks are divided into two levels: d-textbooks
that are digitalised editions of printed textbooks and include only text and figures; and i-textbooks that in-
clude interactive elements, constructions and interactive tasks with multiple feedback to the text, and the stor-
ing of results and monitoring of users. If a textbook corresponds to the curriculum for which it is intended, the
Council of Experts for General Education approves it in accordance with the procedure described in the rules.

Among the numerous textbooks approved for primary and secondary school, 27 i-textbooks [15] have been
approved for primary school and 12 for secondary school (Fig. 1). All i-textbooks were issued by the National
Education Institute Slovenia in the E-Schoolbag and E-textbooks emphasising natural sciences in primary
school projects in 2014 and 2015. Many authors, illustrators, animators, reviewers and other experts from dif-
ferent fields were included in the development of i-textbooks, together with Slovenian companies that offered
support during the development process. I-textbooks are freely accessible, and may be used in schools and at
home, free of charge. They can be used on computers and tablets that run Windows, Android and iOS operat-
ing systems, and can be used online or downloaded using different clients and used offline on mobile devices.
I-textbooks that have already received official approval cover the following subjects: the Slovene, English and
German languages, mathematics, chemistry, natural sciences and techniques, physics, geography, music and
fine arts. An i-textbook for sport is currently in the approval process [16].

Figure 1. Web portal iUčbeniki (www.eucbeniki.sio.si).

In addition to approved i-textbooks, a list of other e-learning materials developed through tenders issued by
the Ministry of Education, Science and Sport are collected on the SIO.si portal [17]. In our research, we analysed
the e-learning materials found on this portal, and discovered that many are inactive. For example, the E-um
portal (www.e-um.si), where mathematical tasks for primary and secondary schools are available, was last
updated in 2015. The same problem was identified with the Nauk.si web portal (www.nauka.si), which was
developed for mathematics (for secondary school), and for physics, logic and computer sciences (for primary
and secondary schools). The portal was developed by the NAUK Group with the main ideas to put teachers
“back in the game”, as their role in developing e-materials is often neglected [18].

The e-gradiva web portal (www.egradiva.si) appears more active and offers e-content from different subject
areas, while different news is also available on their Facebook page. A great deal of the e-learning materials
developed by different schools and companies are unavailable, or web portals do not exist. All those e-portals
were developed in the scope of different projects. Unfortunately, it seems that maintaining existing e-learning
materials was overlooked.

E-portal that should be mentioned and is freely accessible is Učiteljska.net (www.uciteljska.net) (Fig. 2). This
e-platform was edited in 2003 and is still active. It is intended for teachers, students and other pedagogical
workers. It offers the free exchange and use of school learning materials, preparations, seminar papers, tests,
exercises, articles, etc., all of which are edited voluntarily by teachers.

8th Conference on Information and Graphic Arts Technology 131

 Poster section

Figure 2. Web portal Učiteljska.net (www.uciteljska.net).

E-learning materials are also developed and offered by different Slovenian publishers. E-learning materials are
usually available on web portals, the most popular of which are i-Modrian (www.modrian.si), Lilibi.si (www.
lilibi.si) (Fig. 3), Radovednih pet (www.radovednih-pet.si) (Fig. 4), iRokus (www.irokus.si), iRokus+ (www.
irokusplus.si), Vedeževo e-okolje (www.vedez.dzs.si), Naša ulica (www.nasaulica.si) and UČIMte.com (www.
ucimte.com).

Figure 3. Web portal Lilibi.si (www.lilibi.si).

Figure 4. Web portals Radovednih pet (www.radovednih-pet.si).

E-textbooks, e-workbooks and other different multimedia and interactive materials can be accessed on those
portals. E-textbooks are substantially and formally identical to printed editions, as those textbooks are ap-
proved by the Council of Experts for General Education. E-textbooks and e-workbooks on some of the afore-
mentioned portals have a simple toolbar, which allows a user to scroll through the pages, write and draw on
the pages, tag words, enter notes, add web links and enlarge individual parts of a page. Some e-materials are
also created as multimedia-interactive pages enriched with videos, 3D models, sound records, animations,
interactive tasks and other add-ons. Some platforms also include an LMS, through which teachers can monitor
certain activities relating to learners or groups of learners, create their own tests and surveys, create and follow
activities using a working calendar, select additional media materials, and track statistics regarding the success
and progress of an individual or group of learners. On the other hand, success in solving tasks in relation to an
entire group is also available to learners. Some platforms are freely accessible, while others require payment.
On the latter, teachers and learners receive access to portals at the beginning of the scholastic year, typically in
conjunction with a learning kit that includes printed versions of textbooks, workbooks and other accessories.

A few companies offer web portals with interactive tasks for specific subjects. One of them is Moja matema-
tika (www.moja-matematika.si) (Fig. 5), where interactive mathematical tasks are available from the first to
fifth grades of primary school. Although this portal requires payment, it has functioned successfully for more

8th Conference on Information and Graphic Arts Technology 132

 Poster section

than six years, mainly because tasks are interesting, useful and regularly updated with the new challenges.
In addition, each learner can trace their own success in solving tasks and also in relation to other learners.

Figure 5. Web portal Moja matematika (www.moja-matematika.si).

Textbooks for students in higher education are usually published by a university faculty member. Some text-
books are already published as d-textbook, however, no i-textbook was found on the market.

As is evident from the data obtained, learners and teachers in Slovenian primary and secondary schools have
significant opportunities to use the e-learning materials that have been developed over the last few years.
Since the 2007/2008 scholastic year, teachers autonomously select textbooks from a catalogue, which lists
textbooks approved by the Council of Experts for General Education. It is thus the teacher’s responsibility to
choose the most suitable learning materials for their learners. According to an analysis performed during the
2012/2013 scholastic year, the most effective and most frequently chosen combination by teachers today is
the use of a printed textbook and e-learning materials [19].

4 CONCLUSIONS

According to the results of an analysis of the Slovenian market, 39 i-textbooks have been developed and
approved for use in primary and secondary schools in Slovenia. All of those i-textbooks have been issued by
the National Education Institute Slovenia and are freely accessible. In addition, several publishers develop
e-learning materials. Those materials come in different forms such as digitalised printed versions of textbooks
or workbooks, which in some cases have multimedia and interactive elements. E-learning materials are also
offered on different platforms, some of which require payment and some of which are free. Some materials
are active, while others are not. As was communicated to us through interviews, the problems associated with
the continued use of e-textbooks are more pedagogical than technical. Open issues include: how to change
current ways of teaching; how to use e-textbooks as an effective tool for teachers and learners; how to instruct
teachers on changing their teaching plans, etc. On the other hand, [6] believe that the lifespan of e-learning
materials and e-services in Slovenia, despite proven their effectiveness, is currently equal to the duration of
the associated projects. They suggest that it is necessary to find effective and lasting systemic solutions for
editing and maintaining existing e-learning materials and e-services, and for providing technical support for
new ideas as they emerge and for the implementation thereof.

5 REFERENCES

1. Rules on approving textbooks. Official Gazette of the Republic of Slovenia. 15 5 2015, 34, pp. 3771-3773.
2. KREUH, N., KAČ, L., MOHORČIČ, G. Izhodišča za izdelavo e-učbenikov. Ljubljana : Zavod RS za šolstvo,
2011, ISBN 978-961-03-0000-7.
3. PESEK, I. et al. Od e-gradiv do i-učbenikov. Slovenski i-učbeniki. Ljubljana : Zavod Republike Slovenije za
šolstvo, 2014, pp. 9-17.
4. PESEK, I. et al. Projekt e-učbeniki pri naravoslovnih predmetih v osnovni šoli. [ed.] I., Zmazek, B. & Milekšič,
V. Pesek. Slovenski i-učbeniki. Ljubljana : Zavod Republike Slovenije za šolstvo, 2014, pp. 17-27.
5. LESJAK, D. Akcijski načrt nadaljnjega preskoka informatizacije šolstva. Ljubljana : Programski svet za
informatizacijo šolstva, Ministrstvo za šolstvo in šport, 2006.
6. FLOGIE, A., ČUK A. Kaj nam prinaša projekt e-Šolska torba? [ed.] A. & Čuk, A. Sambolić Beganović. Kranjska
Gora : Zavod RS za šolstvo, 2015. pp. 19-69.
7. Strategija razvoja informacijske družbe v Republiki Sloveniji si2010. Vlada republike Slovenije. Ljubljana :
Ministrstvo za visoko šolstvo, znanost in tehnologijo, Direktorat za informacijsko družbo, 2007. p. 67.
8. STANKOVIČ, J. Bilten E-šolstvo. Ljubljana : E-središče v okviru projekta E-šolstvo, 2010, ISSN 1855-9743.

8th Conference on Information and Graphic Arts Technology 133

 Poster section

9. Projekt e-šolstvo. Ministrstvo za izobraževanje, znanost in šport. Available from: www.mizs.gov.si/si/delov-

na_podrocja/direktorat_za_investicije/ikt_v_solstvu/sredisce_za_e_solstvo /; Accessed: 2016-03-08PESEK,
I. Projekt E-učbeniki s poudarkom naravoslovnih predmetov v osnovni šoli. Available from: http://www.
zrss.si/projektiess/gradiva/Projekt_9_eucbeniki.pdf; Accessed: 2016-06-16.
10. E-šolska torba. Slovensko izobraževalno omrežje. Available from: http://projekt.sio.si/e-solska-torba/;
Accessed: 2016-05-21.
11. SAMBOLIĆ BEGANOVIĆ, A., ČUK, A. Kaj nam prinaša e-Šolska torba II. Primeri obetavnih praks in evalvacija
projekta. Zavod RS za šolstvo. Ljubljana, 2016, ISBN 978-961-03-0324-4.
12. HOPSON, M.H. et al. Using a Technology-Enriched Environment to Improve Higher-Order Thinking Skills.
Journal of Research on Technology in Education, 2001, Vol. 34, no. 2, pp. 109-120.
13. Digitalna Slovenija 2020 - Strategija razvoja informacijske družbe do leta 2020. Ljubljana : Republika
Slovenija, 2016.
14. I-učbeniki - Spletno mesto interaktivnih učbenikov. Zavod Republike Slovenije za šolstvo. Available from:
https://eucbeniki.sio.si; Accessed: 2016-03-18.
15. Vključimo v pouk e-učbenike. National Education Institute Slovenia. Available from: http://www.zrss.si/
project/vkljucimo-v-pouk-e-ucbenike; Accessed: 2017-05-14.
16. Zbirka gradiv na portalu SIO. SIO Slovensko izobraževalno omrežje. Available from: http://portal.sio.si/
gradiva; Accessed: 2016-11-15.
17. LOKAR, M. Re-using teaching material. International Journal for Technology in Mathematics Education,
2010, Vol. 16, no. 1, pp. 37-42.
18. POTOČNIK, N., LOGAJ, V. Priporočila za izbiro učbenikov in kompletov učnih gradiv za šolsko leto 2017/18.
Ljubljana : National Education Institut Slovenia, 2017.

Corresponding author:
Urška STANKOVIČ ELESINI
University of Ljubljana
Faculty of Natural Sciences and Engineering
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1210 Ljubljana, Slovenia
phone: +386 1 200 32 29, fax: + 386 1 200 32 70, e-mail: [email protected]

Co-author(s):
Grega TOMAŽIN
University of Ljubljana
Academy of Fine Arts and Design
Ljubljana, Slovenia
Phone: +386 1 4212 530, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 134

 Poster section

BLUEPRINTING - AN ANCIENT BUT STILL LIVING WAY OF PRINTING TEXTILES

Veronika KRAVCOVÁ1, Ondrej PANÁK1, Jiří HEJDUK1 & Petra FORTE TAVČER2
1
University of Pardubice, Faculty of Chemical Technology, Department of Graphic Arts and Photophysics,
Technological Pavilion Doubravice 41, 533 53 Pardubice, Czech Republic
2
University of Ljubljana, Faculty of Natural Science and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia

ABSTRACT: Dyeing with indigo is one of the oldest ways of dyeing and printing of textiles in the world. In
this study, samples made by classical blueprint processes prepared by one of Czech workshops were tested. A
100% cotton fabric was used. Basic parameter such as mass per unit area and thickness were analyzed before
and after dyeing. Colour fastness of dyed samples to washing, light and rubbing (crock test) were measured.
Colour fastness to washing was measured at 40°C, 60°C and 90°C, respectively. For the light fastness test the
sample was exposed to the artificial light for 72 hours. Wet and dry crock fastness were measured. It was found
out that the fastness to washing and light of blueprinted cotton fabric are very good, whereas the blueprint
is very sensitive to rubbing.

Keywords: indigo, blueprint, textile dyeing, cotton fabric, colour fastness.

1 INTRODUCTION

The blueprint started to be used in Slovakia and Czech Republic in 18th century and has remained until today.
Few blueprint workshops are still producing fabrics by traditional techniques. There are two companies in
Czech Republic and just one in Slovakia. Historically, indigo was a natural dye extracted from the leaves of
certain plants. A large percentage of indigo dye produced today is synthetic. Several thousand tons are syn-
thesized each year to produce denim cloth and blue jeans [1].

Indigo is a vat dye and it is not soluble in water. To be dissolved, it must undergo a chemical change (reduction).
Reduction converts indigo into water-soluble leuco-indigo. When a submerged fabric is taken out from the
dye bath, the leuco-indigo quickly react with oxygen in the air and the insoluble, intensively colored indigo
is restored [2].

Difference between blueprinting and dyeing with indigo solution, is that blueprint uses reserve to create pat-
tern and then is dyed in the indigo solution. It means that blueprint is type of resist printing [3]. The name of
technique is derived from the colour of the final product and from the fact that the pattern was printed using
wooden models. The formation of pattern on the fabric is a two-step process. First, the reserve is printed by
hand onto pre-starched fabric with wooden forms. Nowadays also screen printing technique can be employed
in the process of reserve application. Reserve consists of several chemicals and each workshop usually has its
unique formulation. When the reserve is completely dry (it takes about 7 days), the dyeing can start. Textile is
dyed in big barrels with indigo solution. One dyeing takes 5-10 minutes. Then it is taken out and left for few
minutes in the air to oxidize [1]. Textile must be completely immersed into the indigo solution. Reserve pro-
tects printed places on the textile from dyeing producing a white pattern. The resulting shade depends on the
number of dyeing steps. The more dyeing steps, the darker is the shade of the fabric. After dyeing the fabric
must be washed. First with clean water to remove redundant indigo and other dyeing additives. Second wash
is then in weak sulfuric acid solution to remove reserve from the fabric.

2 EXPERIMENTAL

100% cotton fabric supplied by a blueprint company Arimo spol, s. r. o. - Strážnice from the Czech Republic,
was used. Before dyeing the fabric was starched. Mass per unit area of the fabric is 156 g/m2.

2.1 Printing and dyeing

Before dyeing fabric was flat-screen printed by hand with a resist printing paste. Screen mash was 120 threads
per cm. One squeegee stroke was enough for applying reserve. Printed samples were left on the table to dry
in the air. General recipe of the resist printing paste is shown in Table 1. Figure 1 shows the printing pattern.

8th Conference on Information and Graphic Arts Technology 135

 Poster section

The pattern consists of lines with different thickness, dots with varying resolution and density, text in Arial
with Different font size and a Siemens star. This pattern was evaluated only visually.

Fabric was dyed in blueprint company Arimo spol, s. r. o. - Strážnice. Fabric was immersed into indigo dyeing
bath for five minutes. Then it was pulled out for a five minutes to oxidize. The dyeing consisted of seven of
these cycles. After the dyeing, the fabric was washed. First washing process was done in solution of weak
sulfuric acid to remove the reserve from the fabric. Second washing was done with clean water to remove
residual indigo and dyeing additives.

Table 1. Approximate recipe of the resist printing paste and indigo dyeing bath [4].
Recipe of resist printing paste Recipe of indigo dyeing bath
Lead (II) nitrate indigo dye
Copper (II) sulfate sodium hydroxide
Lead (II) acetate sodium hydrosulphite
Kaolin
Arabic gum

Figure 1. Design of the print pattern (left), reserve printed on the cotton fabric (middle),
finished sample after printing, dyeing and washing (right).

2.2 Analysis

Colour fastness to washing

The colour fastness to washing of dyed samples was tested in accordance to standard ISO 105-C10: 2006 at
40, 60 and 90°C. Washing was carried out for 30 minutes at 40 rpm, 1, 3 and 5 times. A gray scale was used to
evaluate the samples.

Light fastness - exposure to artificial lights

Samples were illuminated for 72 hours in Xenostest Alpha instrument by artificial xenon lamp. The test was
performed in accordance to standard ISO 105-B02: 1994. A blue scale was used to evaluate the samples.

Crockfastness
Testing was carried out in accordance to standard ISO 105: X12: 1993 to determine abrasion resistance for dry
and wet rubbing. Testing was done for 10 rubbing steps. The coloured layer was captured on the surface of the
abrasion fabric at 9N. The Electronic Crockmeter M238BB instrument was used for the measurement.

Change of mass per unit area and thickness

Ten samples were used for measurement of mass per unit area before and after dyeing. The samples were 4×4 cm
in size and area of one sample is about 0.0016 m2. Values were averaged and interpreted in terms of g/m2.

8th Conference on Information and Graphic Arts Technology 136

 Poster section

The thickness of dyed and raw fabric was measured at 5 spots. Values were averaged.

Colour measurements
To obtain CIELAB values spectrophotometer Datacolor Spectrum Flash® SF600 PLUS-CT was used. Measure-
ment geometry was de: 8 with circular aperture of 9 mm in diameter. Colorimetric values were captured by
Datacolor Datamaster software (standard D65 illumination and 10°standard observer). MS Excel was used for
evaluation of colour different ΔE* according to the Equation (1).

∆E*=√(L*s-L*m)2 + (a*s-a*m)2 + (b*s-b*m)2 (1)

where L*s , a*s , b*s are values of reference, and L*m , a*m , b*m are values for measured sample.

3 RESULTS & DISSCUSION

Colour scale
The final colour shade of fabric depends on the numbers of dyeing steps. To illustrate the change of colour
shade after each dyeing a colour scale was made. Table 2 shows L*a*b* values of scale samples and colour
difference ΔE* between dyeing cycles showing how much the shade was changed. ΔE*1 is calculated between
each dyeing cycles, ΔE*2 is calculated between first and following dyeing cycles. With each dyeing cycle the
colour of the fabric becomes darker and less saturated. The highest colour difference is between first and
second dyeing cycle. The final dying cycle differs from the first for 24 units ΔE*. Figure 2 shows colour scale
and colour differences after each dyeing cycle.

Table 2. L*a*b* values of colour scale and colour differences ΔE* between samples dyed at different cycles.
Dyeing cycle L* a* b* ΔE*1 ΔE*2
1st 33.73 -2.50 -23.98 0 0
2nd 27.56 0.06 -22.10 6.94 6.94
3th 24.21 0.75 -19.59 4.24 10.98
4th 21.89 1.90 -16.87 3.76 14.49
5th 19.93 3.01 -13.99 3.66 17.91
6th 18.10 3.53 -9.75 4.65 21.98
7th 17.13 3.8 -7.96 2.05 23.91
ΔE*1 difference between each dyeing cycle, ΔE*2 difference between respective dyeing cycle and first dyeing

Figure 2. Simulation of colour scale (right side), colour of white fabric.

8th Conference on Information and Graphic Arts Technology 137

 Poster section

Colour fastness to washing

Table 3 shows grey scale rating for washed samples after first, second and fifth washing. For 40 and 60°C the
colour fastness of samples is considered to be very good. After washing, the viscose samples were stained
more than cotton samples.

Table 3. Grey scale rating for colour fastness to washing.

Grey scale rating
Temperature Wash Alt a
Sta-Cob Sta-Cvc
1st 5 5 3
40°C 2 nd
5 5 3/4
5 th
5 4/5 4
1 st
4/5 4/5 3/4
60°C 2nd
5 5 4/5
5 th
5 4/5 4
1st 4/5 4/5 4/5
90°C 2nd 4 4/5 4/5
5th 2/3 5 4/5
a
Alteration of colour, Staining of cotton adjacent fabric, Staining of viscose adjacent fabric
b c

Light fastness - exposure to artificial lights

The cotton fabric dyed with indigo can be considered as very stable when exposed to light. On the blue-scale
from 1 (bad) to 7 (good) samples were rated by 7.

Crock fastness
Staining of white fabric at dry rubbing was evaluated by a grade 2/3 and at wet rubbing by the grade 3. The
results indicate that the resistance to rubbing is low.

Change of mass per unit area and thickness before and after dyeing
The mass per unit area of the fabric increased by 0.14 g/m2 after dyeing. The thickness of the fabric increased
by 0.08 mm after dyeing. Both results indicate that the dyeing and printing process do not have significant
impact on these parameters. The original mechanical properties of the fabric remained unchanged.

4 CONCLUSION
Samples which were made by traditional blueprint technique were tested regarding their colour fastness. The
results of colour fastness to light indicate that cotton fabric dyed by indigo can be considered as very good. The
washing test shows no colour change of dyed samples, although the adjacent fabrics were stained remarkably.
Viscose samples were stained more than cotton samples. All of viscose samples had light blue shade. Worst
results were found for washing at 90°C. The best colour fastness to washing was found for washing at 40°C.
The crockfastness test shows that samples were not resistant to rubbing. A lot of dyes transferred to the white
fabric during rubbing. The reason for this is the surplus of the dye on the surface of the fabric, which arise from
seven dyeing cycles. Measuring of weight and thickness shows only negligible increase of values after dyeing.

5 REFFERENCE
1. Perkins, W.S., 1996. Textile Coloration and Finishing, Carolina Academic Press.
2. Miles, L.W.C., 2004. Textile printing, Society of Dyers and Colourists, Bradford.
3. Plajh, L. et all., 2015, “Digital printing of blue-printed textile exhibits replicas.” Industria textilæa, 66(2),
67-73.
4. Danglova, O., 2014. Blueprint in Slovakia, Ústredie ľudovej umeleckej výroby, 9788089639120, Bratislava.

Corresponding author:
Veronika KRAVCOVÁ
Department of Graphic Arts and Photophysics, Faculty of Chemical Technology,
University of Pardubice, Technological Pavilion Doubravice 41,
533 53 Pardubice, Czech Republic
[email protected]

8th Conference on Information and Graphic Arts Technology 138

 Poster section

COLOUR AND ODOUR PERCEPTION AT COSMETIC PACKAGING

Kaja JOKIĆ, Barbara BLAZNIK & Sabina BRAČKO

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, SI-1000 Ljubljana

ABSTRACT: The aim of our research was to study the relation between colour and fragrance of a cosmetic
product. A group of observers described the selected colour samples with colour characteristics which were
chosen in accordance with previous investigations and reflected six pairs of opposite characteristics: ac-
tive-passive, warm-cool, hard-soft, heavy-light, clean-dirty, male-female. In the second part, the observers
described the colour samples with the characteristics of odour which represented some of the most common
scents. The results have shown that colours exhibit their typical character and odour characteristics, but the
responses of male and female observers were not always the same. According to the results, the colours which
are positioned closely in the CIELAB colour space also exhibit similar associations. In conclusion, our senses
are capable of connecting the visual perception with other influences from the environment and this should
be taken into consideration when designing different products and packaging.

Key words: colour perception, odour, CIELAB colour space, cosmetic packaging.

1 INTRODUCTION

Colour is a subjective perception, generated in the brain and triggered by the light, entering the eyes of the
observer (Golob, 2001). Beside their esthetic value colours also exhibit biological, psychological, sociolog-
ical and cultural values (Kovačev, 1997). The influence of colours on our physical and emotional response is
increasingly investigated.

A certain colour can often be connected with a specific odour (Kim, 2013, Plümacher, 2007). A group of re-
searchers performed an experiment to find out whether the linking between the colours and different odours
is universal or whether it originates from the cultural factors such as language. As it shows, the connections
are not always consequent. The scent of almond, for example, is connected with red in Canada, whereas with
blue in Australia (Chang, 2014). The interactions between colour and fragrance family were investigated (Kim,
2013) and the results confirmed the existence of cross-modal correspondences between vision and sense of
smell, which are differently interpreted by men or women.

For the majority of the people, the sense of smell along with the sense of flavour represents the biggest and
most direct influence on emotions (Wrzesniewski, 1999). By connecting the perception of colour and odour,
more distinctive impact of a product could be achieved (Barkat, 2003). Different authors tried to explain the
relationship between the characteristics of colours and the perception of other sensations, such as smell and
touch (Ou, 2004, Jraissati, 2016), but several dilemmas remain unsolved. In our research, we aimed at establish-
ing the connections between the colours and fragrances of cosmetic products. We investigated the differences
between male and female observers in Slovenia.

2 EXPERIMENTAL

The research was based on the experiment, in which the selected colour samples were described by a group
of observers, firstly with their colour characteristics and secondly with typical odour characteristics. 20 col-
our samples of size 65 mm × 65 mm were prepared, with Konica Minolta bizhub C364 printer and Color Copy
(Mondi) paper with grammage 160 g/m2. CIELAB values of the samples were measured (Figure 1) in accordance
with ISO 13655 : 2009 by using the spectrophotometer Eye One (X-Rite).
The investigation included 10 observers, 5 male and 5 female, of age 20-30 years. The testing was performed
under constant and controlled lighting conditions. The surrounding was isolated from additional odours.
The research was divided in two parts. In the first part, the observers were asked to describe the samples with
selected colour characteristics, which were chosen in accordance with previous investigations and to describe
six pairs of opposite characteristics (Ou, 2004): active-passive, warm-cool, hard-soft, heavy-light, clean-dirty,
male-female. In the second part, the observers described the colour samples with the characteristics of odour,
which were selected in accordance with previous investigations and represent some of the most common
scents: floral, fruity, woody, bitter, fresh, sweet, mouldy, youthful and herbal (Levitan, 2014).

8th Conference on Information and Graphic Arts Technology 139

 Poster section

Figure 1. Colour samples in the a*b* (left) and a*L* plane (right) of CIELAB colour space.

3 RESULTS & DISCUSSION

The bubble charts in the a*b* plane of the CIELAB colour space (Figure 2) show typical colour characteristics
for twenty colour samples. Six pairs of opposite characteristics (Ou, 2004) were used to describe each sample
and the diameters of circles represent the number of selections.

a) b)

c) d)

e) f)
Figure 2. Colour samples in the a*b* plane of CIELAB colour space, representing selected opposite proper-
ties of colour: a) active/passive b) warm/cool c) hard/soft d) heavy/light e) clean/dirty f) male/female.
● positive properties: active, warm, soft, light, clean, female
● negative properties: passive, cool, hard, heavy, dirty, male
○ neutral colours.

According to the results shown in the Figure 2a, yellow, red and green were described as the most active col-
ours. Grey, brown and olive green, on the other hand were described as the least active, which confirms the

8th Conference on Information and Graphic Arts Technology 140

 Poster section

results of previous investigations (Ou, 2004). The activity of colours increases when saturation is increased.
The colours described as warm can be found in the red-yellow area of the CIELAB colour space whereas cool
colours are located mostly in the green-blue-violet area. The results show that when saturation is increased,
warm colours are described as warmer and cool colours as colder (Figure 2b). Brown, grey and purple were de-
scribed as heavy colours (Figure 2d), while unsaturated colours, such as pink and light purple, were described
as light. Those findings are in accordance with results of previous investigations (Ou, 2004).

We also investigated the differences in the response of male and female observers. Figure 3 shows the results
for those colours where the disagreements were the most obvious: magenta, blue, violet, red and green. Men
described magenta as warm and active colour, women, on the other hand, perceive it as hard and heavy. Light
blue was described as active and clean colour by men but as passive and dirty by the women. The opposite
opinions were also given for the samples of pink and unsaturated green colour: the male observers described
those colours as passive and women as active. Surprisingly, men believe that red is soft and masculine colour,
women on the other hand perceive it as hard and feminine.

Figure 3. Description of colour samples by male (□) and female (○) observers.

Figure 4 represents the arrangement of colour samples in the fragrance wheel which is divided into four parts
according to its author Michael Edwards (Kim, 2013): floral, oriental, woody and fresh tones. According to our
group of observers, floral fragrances are associated mostly with different shades of purple and fresh scents with
blue and yellow hues. Oriental fragrances are described mostly with purple, brown and red and the group of
wooden fragrances with different hues of brown, grey and olive green. Yellow colour causes several different
associations: it is perceived as sour, floral, bitter, youthful, fresh and sweet. The magenta was connected with
fruity and sweet scents, different shades of blue with sports tones. In accordance with previous research (Kim,
2013), our group of observers connected floral scents with warm colours with higher lightness and wooden
scents with cooler colours of lower lightness.

Figure 4. Description of colour samples regarding the selected odour properties (left)
and by properties of colour (right).

Beyond the cultural influences, the sensation of odour is influenced by the age, gender and past experiences
(Levitan, 2014). According to the results of our study (Figure 5), there are some differences between male and
female observers in Slovenia. Floral scent, for example is connected with different hues of purple and yellow
by the men, but women connect it with a purple sample. Men connect fruity scent with red or orange colour,
women on the other hand with magenta and light yellow. There were also some differences in the description

8th Conference on Information and Graphic Arts Technology 141

 Poster section

of a sweet fragrance. A more uniform response was observed at wooden and herbal fragrances. Also, a dark
blue was connected with a sports scent and a light blue with a fresh scent.

floral fruity woody oriental sport sour bitter fresh sweet musty youthful herbal
Male

Female

Figure 5. The perception of odour and colour: the differences between male and female observers.

The bubble chart (Figure 6) shows the colours representing six selected odours (floral, fruity, youthful, fresh,
herbal, bitter) placed in the a*b*plane of the CIELAB colour space. The diameter of each circle represents the
number of selections. The comparison of Figs 6a and 6b shows that floral and fruity scents provoke very similar
responses of the observers regarding the colour selection, as the positioning of the samples in the a*b* plane
is very similar. These colours can mostly be found in the red-yellow region, at a considerable distance from
the origin of the CIELAB space, thus their saturation is very high.

The groups of fresh and youthful scents also show some similarities. These scents are mostly connected with
colours of low saturation, i.e. neutral tones. The majority of those samples is positioned along the b* axis of
the CIELAB colour space, therefore yellow and blue hues prevail (Figures 6c and 6d). Some similarities can
also be found between the groups of herbal and bitter scents. They are mostly associated with yellow-green
colours (Figures 6e and 6f).

a) b)

c) d)

e) f)
Figure 6. Colour samples in the a*b* plane of CIELAB colour space, representing selected odours:
a) floral, b) fruity, c) fresh, d) youthful, e) herbal, f) bitter.

8th Conference on Information and Graphic Arts Technology 142

 Poster section

4 CONCLUSIONS

In the present research, we investigated the relation between the colour and fragrance of cosmetic products.
The results have shown that colours exhibit their typical character and odour characteristics, but the responses
of male and female observers were not always the same. Despite this fact, different colour groups can be de-
fined by fundamental colour attributes. Colours with higher lightness and higher saturation were described
as active, soft, feminine, light and clean. The opposite characteristics were connected with darker and less
saturated colours. The feeling of activity increases with the saturation of colour. When chroma is increased,
warm colours are perceived as even warmer and cool colours as cooler. Floral scents were connected with warm
and brighter colours, wooden fragrances on the other hand with darker and cooler colours. According to the
results, the colours which are positioned closely in the CIELAB colour space also exhibit similar associations
regarding odour and colour characteristics. In conclusion, our senses are capable of connecting the visual
perception with other influences from the environment and this should be taken into consideration when
designing different products and packaging.

5 REFERENCES

• Barkat, S., Thomas-Danguin, T., Bensafi, M. et al. 2003. »Odor and color of cosmetic products: Correlations
between subjective judgement and autonomous nervous system response.« International Journal of Cosmetic
Science 25 (6): 273-283.
• Chang, B. Which colors do you smell? URL: https://psmag.com/social-justice/color-smell-85662 (last accesed
on 15.9.2017).
• Golob, V. 2001. »Merjenje in vrednotenje barve.« In Interdisciplinarnost barve I, edited by Slava Jeler and
Marko Kumar, 199-230, Maribor: DKS.
• ISO 13655: 2009. Graphic technology- Spectral measurement and colorimetric computation for graphic arts
images.
• Jraissati, Y., Slobodanyuk, N., Kanso, A. et al. 2016. »Haptic and tactile adjectives are consistently mapped
onto color space.« Multisensory research 29 (1-3): 253-278.
• Kim, Y. J. 2013. »Can eyes smell? Cross-modal correspondences between color hue-tone and fragrance family.«
Color Research and Application 38 (2): 139 – 156.
• Kovačev, A. N. 1997 Govorica barv. Ljubljana: Prešernova družba.
• Levitan C.A. et al. What color is that smell? Cross-cultural color-odor associations. URL: http://journals.
plos.org/plosone/article?id=10.1371/journal.pone.0101651/ (last accesed on 15.9.2017).
• Ou, L.C., Luo, M.R., Woodcock, A. and Wright, A. 2004. »A study of colour emotion and colour preference.
Part I. colour emotions for single colours.« Color Research & Application 29 (3): (232-240).
• Plümacher M. and Holz P. Speaking of colors and odors. URL: https://books.google.si/books?id=zjTI1sF-
CzA4C&pg=PA17&lpg=PA17&dq=plumacher+holz+speaking+of+colors&source=bl&ots=0QZEiT6NnC&sig=fs-
J4Vrm8_n9IvGTI26qvvCbadjI&hl=sl&sa=X&ved=0ahUKEwjHjP3q36bWAhVLPxoKHdDCA-MQ6AEITDAK#v=o-
nepage&q=plumacher%20holz%20speaking%20of%20colors&f=false (last accesed on 15.9.2017).
• Wrzesniewski, A., McCauley, C. and Rozin, P. Odor and affect: Individual differences in impact of odor on
liking places, things and people. https://academic.oup.com/chemse/article/24/6/713/320341/Odor-and-Af-
fect-Individual-Differences-in-the (last accessed on 15.9.2017).

Corresponding author:
Sabina BRAČKO
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 38, e-mail: sabina.brač[email protected]

Co-authors:
Kaja JOKIĆ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, SI-1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 143

 Poster section

Barbara BLAZNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 22, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 144

 Poster section

COLOR OF TASTE

Gregor FRANKEN
Faculty of Natural Sciences and Engineering
Department of Textiles, Graphic Arts and Design

ABSTRACT: The present study focused on the impact of taste on the choice of corresponding colors, and
the correlation between observation and the choice of individual color. In the experiment, we used the Tobii
X-120 eye-tracking device. There were 48 subjects participating in the study. After calibrating the eye-tracking
device for each subject, the subjects selected a color they associated with each individual taste (salty, sweet,
sour, bitter and umami) from among 24 different color samples shown on the screen. After the measurements
for each individual taste were obtained, we marked Areas of Interested (AOI), based on the results visualized
as Heat Maps, and analyzed each AOI in detail. Within each AOI, we obtained the number of fixations and
fixation duration. We thus acquired an assortment of six colors that the subjects associated most with each
individual taste.

Keywords: taste, color, eye movements, eye-tracking.

1 INTRODUCTION

The color of the food and packaging of food products can convey an extremely strong message. Color is also
the main aspect in our decision to pick a certain packaging in favor of another. Prior to our direct contact with
its contents it should try to induce in us various feelings of specific tastes (ISO 5492). While we initially tend
to associate its color with the odor, it is the flavor itself that comes last: it may trigger feelings of satisfaction
in us or perhaps even shock if the color led us to anticipate a different flavor. In tasting food and beverages
we therefore have a multi-sensoric perception of taste, where our sensors must be synchronized in order to
offer us perfect satisfaction and pleasure (Chylinski, 2015). Color constitutes one of the most salient of visual
cues converning the likely sensor properties of food and drink (Delwiche, 2012).

Most mammals are dichromats and can only distinguish between two dimensions of colour: bright versus
dark and blue versus yellow. In contrast, humans are trichromats, our extra class cone cells enabling us to
discriminate between reds and greens which would otherwise appear identical. Trichromacy evolved as an aid
to frugivory. This notion is particularly attractive, as many fruits gradually turn yellow, red or orange during
ripening. These colours are strikingly visible to trichromats, but dichromats have difficulty distinguishing
them from a dappled background of green leaves (Wolf, 2002).

Under most everyday conditions, consumers have the opportunity to inspect food and bewerage color before
deciding on whether or not to purchase or taste it (Spence, 2015). Throughout the history of gastronomy,
there has always been an understanding that food is both a human necessity as well as being a great source
of pleasure. What happens to our senses when the taste is different than expected? But what exactly happens
in our mind at this point? These questions require a thorough understanding of what the term “flavor per-
ception” really means, starting with its very definition. It turns out that what most people have in mind when
they talk about “taste” is really “flavor”; that is, the result of the integration of inputs from several different
senses (Piqueras-Fiszman, 2016). By themselves, the gustatory receptors on the tongue only provide informa-
tion about the so-called basic tastes, such as sweet, salty, bitter, sour, and umami (Umami or savory taste, is
one of the five basic tastes (together with sweetness, sourness, bitterness, and saltiness). It has been described
as savory, and characteristic of broths and cooked meats (Stuckey, 2012). In fact, according to one oft-quoted
statistic, somewhere between 75% and 95% of what most people commonly describe as the “taste” of food
actually comes from inputs detected in the nose (Piqueras-Fiszman, 2016). Trigeminal inputs contribute to our
perception of food and drink, giving rise to hot, cold, tingling, burning, and electric sensations. Over the last
80 years or so, researchers have investigated the psychological impact of food color (Spence, 2016).

1.1 Taste and color

Most consumers expect a more intense food color, associating it with greater taste intensity. In the last 50 years,
various studies of food and beverages established that the intensity of coloring gives people a more intense

8th Conference on Information and Graphic Arts Technology 145

 Poster section

sense of taste (Spence, 2016, Johnson, 1982, Calvo, 2001). Perhaps the most robustly demonstrated effect of
adding (or changing) food coloring has been on the ability of people to identify the flavor of food or, more
commonly, drink (Spence et al., 2015). Studies of the effect of color intensity in yoghurts with strawberry,
orange and fruit-of-the-forest flavors on the sweetness established that greater color intensity led to a greater
perception of sweetness only in the yoghurts with fruit-of-the-forest flavor. For the rest of the flavors, the
perception of sweetness was not impacted by the color and remained unchanged (Calvo et al., 2001). Studies
of non-fizzy beverages with green, red and orange colorant showed that color masking dramatically decreased
flavor identification of fruit‐flavored beverages, while atypical colors induced incorrect flavor responses that
were characteristically associated with the atypical color. In addition, the color level of beverages had signif-
icant effects on their overall acceptability, acceptability of color and of flavor, as well as on flavor intensity
(DuBose 1980). In Oram et al.’s (1995) study, the subjects (children and adults) tested four different flavors
(chocolate, orange, pineapple, and strawberry) in four different colors (brown, yellow, orange and red). The
subjects tried to guess the correct taste of the beverage. For the typical drinks, the selection of the correct
name was greater than 80% for all ages. Subsequent research by demonstrated that adult participants were
often confused by the addition of inappropriate color to a range of fruit-flavored soft drinks. The accuracy
of participants’ flavor discrimination performance was significantly lower when the solutions were colored
inappropriately than when they were colored appropriately (Zampini et al., 2017) Maga’s (1974) study focusing
on colorant adding (red, green, yellow) for basic flavors showed that adding colorant to food influences taste
threshold sensitivity. Adding green colorant decreased sour taste sensitivity and increased sweet taste sensitiv-
ity. Adding yellow decreased both sour and sweet taste sensitivity, while red decreased bitter taste sensitivity.
Recent research revealed a strong association between salty and white and, to a lesser extent, between salty
and blue (Spence, 2015). In O’Mahony’s (1983) study, most subjects chose white for salty taste, red for sweet
taste, yellow for sour taste and green for bitter taste. In Heller’s (2004) study, most subjects chose white for
salty taste, red for sweet taste, yellow for sour taste and purple for bitter taste. Koch and Koch’s (2003) study
established that most subjects chose white for salty taste, red for sweet taste, yellow for sour taste and black
for bitter taste. Wan et al.’s (2014) study compared tastes, and addressed the issue of color correspondences
across cultures: it was established that white was associated with salty taste, pink with sweet taste, green
with sour taste, black with bitter taste and brown with umami taste. Lavin and Lawless (1998) investigated
the influence of varying the intensity of food coloring on people’s ratings of sweetness intensity for two pairs
of strawberry-flavored drinks. One pair light- and dark-red, the other pair light- and dark-green. The drinks
were equally physically sweet and varied only in terms of their color intensity. The adults who took part in this
study rated the dark-red and lightgreen drinks as tasting sweeter than the light-red and dark-green samples,
respectively (Lavin, 1989). In Franken et al.’s (2012) study, the subjects chose from among 15 different colors
of fruit juice, shown on screen; using eye-tracking technology, it was established that the subjects chose from
among different shades of yellow in most cases. Jantathai et al. (2013) also employed eye-tracking technology;
however, their aim was to investigate how the subjects choose from among the different colors of food prod-
ucts regardless of the taste. Most of the above studies did not employ eye-tracking technology; we therefore
decided to test the tastes, and search for a suitable color using the eye-tracking method.

In addition to color, other factors also influence the taste. Lufthansa uses 1.7 million liters of tomato sauce
a year (Jackson, 2014, Engeber, 2016). Researchers at the Fraunhofer Institute investigated why tomato sauce
tastes better than we are used to at altitudes of approximately 9000 meters: they found that due to the low air
humidity, the receptors for taste and smell function differently. In fact, a parallel can be drawn with having a
cold (Schroeder, 2010). Our sense of smell is 30% less effective; our sweet and salty perceptions are diminished.
The salty perception is diminished by 20-30% and the sweet perception by 15-30%. The sour and fruity percep-
tions remain virtually unaffected. While the taste known as umami is heightened. Thirsty passengers may find
they yearn specifically for something rich and savory, and they frequently choose tomato juice (Eitner, 2010).

1.2 Marketing in food color

Changing the color of food and beverages, and adding colorants have long been used as effective marketing
moves. (Spence, 2016, Jantathaia 2014). Margarine producers colored margarine yellow because they wanted
to highlight its similarity to butter. Butter producers tried to stop this at all costs (Masurovski, 1939). Today,
margarine remains yellow. A similar case is familiar from the beverage industry: all the colas have a color
similar to that of Coca-Cola. They also try to imitate it by using red in their packaging design, with the notable
exception of Pepsi, where blue is the dominant color. Moreover, there are well-known examples of changing
the color of drinks in established brands, including Coca-Cola. Several years ago, a colorless Coca-Cola was
launched; however, it was a complete marketing failure. The Italian apéritif producer Campari is a similar ex-
ample: the attempt to replace its characteristic red color with a colorless version was a failure (Sertl, 2007). In

8th Conference on Information and Graphic Arts Technology 146

 Poster section

2017 the Spanish launch a blue wine which shocks both wine connoisseurs as well as ordinary buyers (Bruner,
2017). With blue wine, it is not about producing an artificial wine substitute, but about isolating anthocy-
anin, a blue pigment, from red wine, and adding indigo pigment (Mateus et al., 2003, Schrieberg, 2017). The
color itself makes it virtually impossible for people to choose blue wine for reasons other than curiosity and
momentary trend.

2 EXPERIMENTAL

2.1 Apparatus

To track the eye movement we used the Tobii X120 eye tracking device and the Tobii Studio 3.4.4 software. The
eye tracker tracks the movement of the eye by following the reflection of the image from the cornea. This re-
flection is generated when the infrared illuminators at the front side of the eye tracker create patterns of light
reflecting from the cornea. Eye tracker contains an infrared-sensitive camera that tracks the individual’s eye
movements and fixations. Prior to the measurement, each individual adapted to the lighting conditions of the
room for 5 minutes and then underwent a 5-point screen-based calibration. For calibration procedure, subjects
had to follow a red dot on the screen with their eyes. The dot was moving from the screen center to the corners.
This procedure allows the eye tracker to adjust the analysis algorithm to the specific eye configuration of the
tested individual. We measured the number of fixation and fixation duration of each taste by each individual.
The color samples were displayed on 25-inch LCD screen with a resolution 2560 × 1440 pixels (pixel size 0.216
mm) at a 60 Hz refresh rate on light grey backgrounds. The tested individuals were positioned 60 (± 3) cm
from the screen according to the recommendations of the ISO 9241 standard (2012). Each test started with
an instruction text and text describing the taste on the screen. After completing the eye-tracking procedure,
subjects were asked to look at color images on the display showing the differently colored samples (green,
pink, and yellow) and to choose the most preferred color variant for each taste. The texts describing the taste
were displayed in the middle of the screen. Consecutive instructions were invoked by successive mouse clicks.

2.2 Preliminary measurments

There were 48 participants, 24 were male and 24 female, aged 19 to 51, with an average of 24.76 years (±6.52);
all participants had normal or corrected-to-normal vision. The participants were divided into 4 groups of 12
participants. We have tried to better understand our color-induced sensory expectations by facing individuals
with various flavors, which they tried to match to individual colors presented on a LCD screen. The participants
within a single group observed presentation of 256 different color samples for 9 second for each taste. For
color samples we used 256 web safe colors. In our preliminary measurements we have chosen only the basic
colors, which were later supplemented by additional shades. The color samples were shown to all 4 groups in
different orders, using the so-called Latin square design. Before the main experiment, we attempted to select
the most frequently chosen colors for each individual taste from among 256 colors (Web safe colors) shown
on the screen. he time of showing of the colors was limited to 9 seconds. For each individual taste, we marked
the areas of interest (AOI) for individual colors using the most frequent fixations and the Heat Maps obtained.

2.3 Main measurments

These preliminary measurements enabled us to choose 24 color samples, which we then showed to each indi-
vidual subject separately for each taste. The time of showing of the colors was limited to 9 seconds. As with
the preliminary measurements, we also marked the AOI using gaze density presented as a Heat Maps (Figure
1), and analyzed the time and duration of fixations (Figure 2). There were also 48 participants, 24 were male
and 24 female, aged 18 to 51, with an average of 25.24 years (±7.32). The participants were divided into 4 groups
of 12 participants.

8th Conference on Information and Graphic Arts Technology 147

 Poster section

   Figure 1. Heat maps for sweet taste.   Figure 2. Fixations from one tested person.

3 RESULTS & DISCUSIONS

This study investigated the effect of taste on using eye-tracking technology and the correlation between gazing
behavior and color choice decision. In presenting the results we have focused on the number of fixation and
fixation duration On average, the subjects had a fixation duration of 20.86 seconds over the entire surface. The
average time of all the pauses was 7.68 seconds. In the presentation and evaluation of the results, we only focus
on the 6 most common colors, where the subjects’ fixation duration was longest, separately for each taste.

3.1 Salty taste

For salty taste, the subjects generally tended to have the most fixations in the area of white #FFFFFF, specifically,
7.4 fixations with the average fixation time of 2.92 seconds. The maximum fixation time was 1.27 seconds, and
the minimum was 0.05 second. The area of turquoise #00FFCC had the least fixations, specifically, 1.2 fixations
with the average fixation time of 0.23 second (Table 1).

Table 1. Salty taste.

Color #FFFFFF #CCCCCC #00FFFF #999999 #009999 #00FFFF
No. of fixations 7,40 4,57 3,57 2,17 1,60 1,20
Fixation duration in sec. 2,92 1,64 1,55 0,81 0,72 0,23
Max. fix. duration in sec. 1,27 1,02 2,01 0,57 0,91 0,25
Min. fix. duration in sec. 0,05 0,05 0,16 0,01 0,22 0,23

3.2 Sweet taste

For sweet taste, the subjects generally tended to have the most fixations in the area of light pink #FFCCFF,
specifically, 3.29 fixations with the average fixation time of 1.18 second. The maximum fixation time was 1.17
second, and the minimum was 0.07 second. The area of red #FF0000 had the least fixations, specifically, 1.67
fixations with the average fixation time of 0.73 second (Table 2).

Table 2. Sweet taste.

Color #FFCCFF #CC00FF #FF00FF #8000FF #FFFFFF #FF0000
No. of fixations 3,29 2,50 2,25 2,13 1,88 1,67
Fixation duration in sec. 1,18 0,77 0,78 0,67 0,78 0,73
Max. fix. duration in sec. 1,17 1,05 0,63 0,60 0,60 0,56
Min. fix. duration in sec. 0,07 0,17 0,16 0,12 0,05 0,17

3.3 Sour taste

For sour taste, the subjects generally tended to have the most fixations in the area of spring bud color #99FF00,
specifically, 3.90 fixations with the average fixation time of 1.51 second. The maximum fixation time was 1.13
second, and the minimum was 0.91 second. The area of olive #999900 had the least fixations, specifically, 1.80
fixations with the average fixation time of 0.95 second (Table 3).

8th Conference on Information and Graphic Arts Technology 148

 Poster section

Table 3. Sour taste.

Color #99FF00 #009900 #FFFF00 #006600 #009999 #999900
No. of fixations 3,90 3,00 2,88 2,17 1,80 1,80
Fixation duration in sec. 1,51 1,34 1,15 1,12 0,98 0,95
Max. fix. duration in sec. 1,13 1,02 1,53 0,91 0,90 0,86
Min. fix. duration in sec. 0,91 0,13 0,1 0,05 0,04 0,13

3.4 Bitter taste

For bitter taste, the subjects generally tended to have the most fixations in the area of olive color #999900,
specifically, 7.50 fixations with the average fixation time of 3.60 second. The maximum fixation time was 0.80
second, and the minimum was 0.17 second. The area of blue #0000FF had the least fixations, specifically, 2.13
fixations with the average fixation time of 0.65 second (Table 4).

Table 4. Bitter taste.

Color #999900 #999933 #996633 #996600 #009999 #0000FF
No. of fixations 7,50 4,33 4,00 2,83 2,43 2,13
Fixation duration in sec. 3,60 1,93 1,23 1,14 0,98 0,65
Max. fix. duration in sec. 0,80 0,77 0,44 0,42 0,42 0,35
Min. fix. duration in sec. 0,17 0,24 0,15 0,15 0,13 0,13

3.5 Umami taste

For umami taste, the subjects generally tended to have the most fixations in the area of golden brown color
#996600, specifically, 6.25 fixations with the average fixation time of 2.11 second. The maximum fixation time
was 1.61 second, and the minimum was 0.6 second. The area of light orange #FF9900 had the least fixations,
specifically, 2.00 fixations with the average fixation time of 0.65 second (Table 5).

Table 5. Umami taste.

Color #996600 #996633 #666633 #999900 #990000 #FF9900
No. of fixations 6.25 4.33 3.40 3.14 2.25 2.00
Fixation duration in sec. 2.11 2.02 1.48 1.14 0.69 0.65
Max. fix. duration in sec. 1.61 1.27 1.08 1.02 0.95 0.93
Min. fix. duration in sec. 0.6 0.11 0.05 0.05 0.05 0.08

The results showed that the method is appropriate for determining a suitable color for different tastes. For
salty taste, our subjects generally chose white, which is very much in accordance with previous research by
Spence et al. (2015) and Wan et al. (2014). Similarly, Heller (2004) also claims that the subjects in her studies
generally chose white for salty taste, as do Koch and Koch (2003).
In our studies, sweet taste is associated with pink, magenta and their various shades. In Wan et al.’s (2014)
study, sweet taste also corresponded to pink. For sweet taste, red color was in the sixth place, although red
is the closest to sweet taste according to Heller (2004). O’Mahony (1983), on the other hand, maintains that
yellow has the strongest correspondence to sweet taste. It is interesting that Maga’s (1974) in Lavin’s (1989)
studies suggest that it is green that increases sweet taste.
With sour taste, our subjects mostly chose between different shades of green, which is in accordance with the
results of Wan et al.’s (2014) study, and yellow, which is in accordance with Heller’s (2004) findings that her
subjects chose yellow for sour taste.
In our study, bitter taste was most strongly associated with olive green and light brown, which may remind us
of the taste of beer; this corresponds to the findings of O’Mahony (1983). Heller (2004) and Wan et al. (2014)
report that black is associated with bitter taste. We did not test for black because black is virtually non-existent
in the food and beverage industry.
Our findings about the choice of brown and light brown for the umami taste correspond to the findings of
Wan et al. (2014).

8th Conference on Information and Graphic Arts Technology 149

 Poster section

4 CONCLUSIONS

Using eye-tracking technology for successfully monitoring eye fixations and determining the colors corre-
sponding to the individual tastes, we obtained results similar to those obtained in studies not employing
eye-tracking technology. There was not much divergence for the individual tastes and colors among individ-
ual subjects. When the measurements were carried out, it became evident that the samples should have been
placed wider apart. The actual presentation of the color samples should have been more aligned with the
individual tastes. It was often the case that the subjects only passed a certain sample when searching for the
right color; nevertheless, the eye tracker recorded their fixation on the color they passed. The sample invoked
no associations for a given taste, yet it was recorded in the measurements. In future research, we would limit
the number of samples shown and increase the spacing between the samples. The measurements obtained by
eye-tracking were shown to be precise, and will continue to be used in research on the interaction of tastes
and colors, as well as research on the interaction of aromas, colors and tastes. The research was carried out
consistently under the same conditions; however, we know that when tasting food, other factors, including
temperature, humidity, noise and last, but not least, our personal mood, also influence food tasting.

5 REFERENCES

• Alley, R. L. and Alley, T. R. 1998. “The influence of physical state and color on perceived sweetness.” Journal
of Psychology: 132(5): 561–568.
• Auvray, M. and Spence, C. 2008. “The multisensory perception of flavor: Assessing the influence of color cues
on flavor discrimination responses.” Food Quality and Preference 18(7): 975−984.
• Bruner, Raisa. You Might See Blue Wine Popping Up on Social Media Soon: Here’s What to Know About It.
URL: http://time.com/4973301/blue-wine-taste/ (last accessed on 10. 04. 2018).
• Calvo, C., Salvador, A., and Fiszman, S. 2001. “Influence of colour intensity on the perception of colour and
sweetness in various fruit-flavoured yoghurts.” European Food Research and Technology, 213, 99–103.
• Chylinski, M., Northey, G., and Ngo, L. V. 2015. “Cross-modal interactions between color and texture of food.”
Psychology & Marketing, 32, 950–966.
• Delwiche, J. F. 2012. “You eat with your eyes first.” Physiology & Behavior, 107, 502–504.
• Delwiche, J. 1996. “Are there ‘Basic’ Tastes.” Trends in Food Science & Technology 7 (12): 411-415.
• DuBose, C. N., Cardello, A. V., & Maller, O. 1980. “Effects of colorants and flavorants on identification, per-
ceived flavor intensity, and hedonic quality of fruit-flavored beverages and cake.” Journal of Food Science,
45, 1393–1399.
• Eitner, Janis. Rätsel um Tomatensaft im Flugzeug gelöst. URL:https://www.ibp.fraunhofer.de/de/Presse_
und_Medien/Presseinformationen/Raetsel_um_Tomatensaftgeloest.html/ (last accessed on 18. 04. 2018).
• Engeber, Daniel. Why do people drink so much tomato juice on airplanes? http://www.latimes.com/opin-
ion/op-ed/la-oe-engber-tomato-juice-airplanes-20160911-snap-story.html/ (last accessed on 18. 04. 2018).
• Heller, E. 2004. Wie Farben wirken: Farbpsychologie - Farbsymbolik - Kreative Farbgestaltung, Reinbek:
Rowohlt Taschenbuch Verlag.
• ISO, (2012). Standard 9241-303: Ergonomics of human-system interaction – Part 303: Requirements for elec-
tronic visual displays, Gèneve: International Organization for Standardization.
• ISO, (2008). Standard 5492: Terms relating to sensory analysis, Gèneve: International Organization for Stand-
ardization.
• Jackson, Steven. Why do we drink tomato juice on planes? URL: https://whyy.org/segments/why-do-we-
drink-tomato-juice-on-planes/ (last accessed on 18. 04. 2018).
• Johnson, J., and Clydesdale, F. M. 1982. “Perceived sweetness and redness in colored sucrose solutions.” Journal
of Food Science, 47, 747–752.
• Jantathai, S., Danner, L., Joech, M., Dürrschmid, K. 2013. Gazing behavior, choice and color of food: Does
gazing behavior predict choice? Food Research International 54(2): 1621-1626.
• Jantathaia, S., Sungsri-in, M., Mukprasirt, A., & Dürrschmid, K. 2014. Sensory expectations and perceptions
of Austrian and Thai consumers: A case study with six colored Thai desserts. Food Research International,
64, 65–73.
• Koch, C., and Koch, E. C. 2003. “Preconceptions of taste based on color.” The Journal of Psycholog, Interdis-
ciplinary and Applied 137(3): 233–242.
• Kildegaard, H., Olsen, A., Gabrielsen, G., Moller, P. and Thybo, A.K. 2011. “A Method to Measure the Effect
of Food Appearance Factors on Children’s Visual Preferences.” Food Quality and Preference, 22(8): 763-771.
• Lavin, J. G., and Lawless, H. T. 1998. “Effects of color and odor on judgments of sweetness among children
and adults.” Food Quality and Preference, 9, 283–289.

8th Conference on Information and Graphic Arts Technology 150

 Poster section

• Levitan, C., Zampini, M., Li, R. and Spence, C. 2008. “Assessing the role of color cues and people’s beliefs
about color–flavor associations on the discrimination of the flavor of sugar-coated chocolates.” Chemical
Senses, 33(5): 415–423.
• Maga, J. A. 1974. “Influence of color on taste thresholds.” Chemical Senses and Flavor, 1, 115–119.
• Mateus, N., Silva, A. M. S., Rivas-Gonzalo, J. C. Santos-Beluga, C. and de Freitas V. 2003. “A New Class of Blue
Anthocyanin-Derived Pigments Isolated from Red Wines.” Journal of Agricultural and Food Chemistry 51,
1919−1923.
• Masurovsky, B. I. 1939 How to obtain the right food color. Food Industries, 11, 55–56.
• O’Mahony, M. 1983. “Gustatory Responses to Nongustatory Stimuli.” Perception 12(5): 627-633.
• Oram, N., Laing, D. G., Hutchinson, I., Owen, J., Rose, G., Freeman, M., et al. 1995. “The influence of flavor and
color on drink identification by children and adults.” Developmental Psychobiology, 28, 239–246.
• Piqueras-Fiszman, B., Spence, C. 2012. “The Influence of the Color of the Cup on Consumers’ Perception of a
Hot Beverage.” Journal of Sensory Studies 27(5): 324-331.
• Piqueras-Fiszman, B., Spence, C. 2016. “Introducion.” Multisensory Flavor Perception, From Fundamental
Neuroscience Through to the Marketplace, 1–13.
• Schrieberg, Felipe. Blue Wine From Spain Runs Afoul of E.U Law. URL: https://www.forbes.com/sites/felipe-
schrieberg/2017/01/20/blue-wine-from-spain-declared-illegal-for-being-the-wrong-color/#7cda62d76212/
(last accessed on 10. 04. 2018).
• Schroeder, Mariana. Lufthansa investigates the science of airline food. URL: http://www.dw.com/en/
lufthansa-investigates-the-science-of-airline-food/a-6114748 (last accessed on 18. 04. 2018).
• Sertl William. Campari: Good And Bitter. URL: https://www.saveur.com/article/Wine-and-Drink/Campari-
Good-and-Bitter/ (last accessed on 18. 04. 2018).
• Spence, C. 2015. “Eating with our ears: Assessing the importance of the sounds of consumption to our per-
ception and enjoyment of multisensory flavour experiences.” Flavour, 4 (1): 1–14.
• Spence, C. 2015. “On the psychological impact of food colour.” Flavour, 4, 21.
• Spence, C., Piqueras-Fiszman, B. 2016. “Food Color and Its Impact on Taste/Flavor Perception.” Multisensory
Flavor Perception, From Fundamental Neuroscience Through to the Marketplace, 107–132.
• Spence, C., Wan, X., Woods, A., Velasco, C., Deng, J., Youssef, J., et al. 2015. “On tasty colours and colourful
tastes? Assessing, explaining, and utilizing crossmodal correspondences between colours and basic tastes.”
Flavour, 4, 23.
• Stuckey, B. 2012. Taste what you’re missing: The passionate eater’s guide to why good food tastes good.
London, UK: Free Press.
• Taylor, A. J. and Roberts, D. D. 2004. Flavour Perception. Oxford: Blackwell Publishing.
• Thomson, D. M. H. 2016. “Sensory Branding: Using Brand, Pack, and Product Sensory Characteristics to
Deliver a Compelling Brand Message.” In Multisensory Flavor Perception, From Fundamental Neuroscience
Through to the Marketplace, edited by Betina Piqueras-Fiszman and Charles Spence, 313–336. Duxford:
Elsevier.
• Wan, X., Woods, A. T., van den Bosch J., McKenzie, K.J., Velasco, C., Spence, C. 2014. Cross-cultural differences
in crossmodal correspondences between tastes and visual features. Frontiers in Psychology, 1-13.
• Wolf, K. 2002. “Visual Ecology: Coloured Fruit is What the Eye Sees Best.” Current Biology 12, 253–255.
• Zampini, M., Sanabria, D., Phillips, N., & Spence, C. 2007. “The multisensory perception of flavor: Assessing
the influence of color cues on flavor discrimination responses.” Food Quality & Preference, 18, 975–984.
• Wan, X., Woods, A. T., van den Bosch J., McKenzie, K.J., Velasco, C., Spence, C. 2014. Cross-cultural differences
in crossmodal correspondences between tastes and visual features. Frontiers in Psychology (5): 1-13.

Corresponding author:
Gregor FRANKEN
Faculty of Natural Sciences and Engineering, Department of TextilesGraphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 94, fax: +386 1 200 03 270, e-mail: [email protected]
web: https://www.ntf.uni-lj.si

8th Conference on Information and Graphic Arts Technology 151

 Poster section

COMPARISON OF DIGITAL AND TRANSFER PRINTING OF COTTON FABRIC

Michaela ČESÁKOVÁ1, Ondřej PANÁK1, Jiří HEJDUK1 & Petra Forte TAVČER2
1
University of Pardubice, Faculty of Chemical Technology,
Department of Graphic Arts and Photophysics, Studentská 95, 532 10 Pardubice, Czech Republic
2
University of Ljubljana, Faculty of Natural Science and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia

ABSTRACT: The aim of this work was to examine the printing of the fabric by two different methods of print-
ing. We compared the direct digital printing of textile with pigmented inks and indirect transfer printing on
textile with a foil. Both methods were applied to 100% cotton fabric. The fabric was directly printed on a Kornit
Digital printing machine, which uses water-based pigmented inks. For the transfer printing Turbo print 4036
matt foil was used. The print motif was printed on the foil using a Mimaki JV33-160 digital printer, which
uses Mild solvent printing inks. Only magenta colour was used for printing. Printed samples were subjected
to rubbing, artificial light ageing, washing and heat. The resistance to rubbing was done according to the
ISO 105-X12: 1993 standard at 50, 250, 500 and 1000 abrasion steps. The lightfastness test was performed in
accordance to ISO 105-B02: 1994. A further test was the washing test, which was made in accordance to ISO
105-C10: 2006. Washing was performed 1x, 3x and 5x at 40, 60 and 90°C. The results were interpreted by visual
assessment and by CIELab coordinates. To characterise the printed cotton fabric, the weight, the thickness and
the air permeability were measured.

Keywords: cotton fabric, digital printing, transfer printing, foil, fastness properties.

1 INTRODUCTION

Nowadays almost every material can be printed. This is valid also for textiles. Textile printing can be described
as an industrial art, having a long history and with a steady future perspective. It requires a knowledge of
science and technology, but will always be a multidisciplinary activity. Different techniques and different
colorants are used for fibres of different chemical structures and different appearance. The choice of the
appropriate fabric, appropriate colorant system and appropriate printing technology is very important for
achieving appropriate quality of the final product [1].
Textile can be printed directly or indirectly. Screen printing and ink-jet printing are the predominant direct
printing techniques nowadays. Different colorants are applied directly onto the surface of the fabric. Mainly
pigment colorants are employed in printing processes. They are easy to use, offer a solid range of colouring
with a very good end-use properties at reasonable cost on a wide range of base fabrics [2].
Indirect printing is also called transfer printing. The motive is first printed with special inks on a paper and
then transferred from the paper to the textile with the help of high temperature and high pressure in a transfer
press. In the case of sublimation transfer, the paper is printed with dispersive dyes, which are then transferred
to polyester fibres. In the case of film-release transfer the print motive can be cut out from a special foil or
a motive can be printed on a foil and subsequently cut out from it. In both cases, the motive can be finally
transferred to almost all kinds of textile materials, from natural to synthetic fibres [1, 3].
The special foil is made of several layers: a white foil on which the printing is made, a layer of adhesive and a
carrier foil that protects the adhesive layer. The printed motif on the foil is cut on a plotter cutter after print-
ing, and excessing foil (foil that does not carry the printed image) is removed from the carrier by peeling. The
print motif ready for transfer to the fabric is left on the print pad.
The benefit of the indirect transfer printing is that it is possible to transfer the print motif first to another
medium (paper or foil) and to apply it to the textile material only when needed.
Proper fabric printing depends on many aspects including proper printing conditions and application of suit-
able printing inks. Important factor influencing the printing is the fabric to be printed. The emphasis put on
the fabric is mainly regarding its physical and chemical properties, which are different for each fabric. These
properties are influenced by their chemical composition and method of their preparation and subsequent
treatment prior to printing [4]. Cotton fabric was chosen in our research because of its frequent use in the
manufacture of clothing [4]. Cotton is widely used fabric because of its high liquid absorption, lower binding
properties of greasy dirt and pleasant touch. It is a natural material of plant origin.
Digital printing with pigmented inks and transfer printing with a foil onto a 100% cotton fabric were compared
in presented research. Differently printed samples were subjected to rubbing, artificial light ageing, washing

8th Conference on Information and Graphic Arts Technology 152

 Poster section

and heat. The results were interpreted by visual assessment and by CIELab colorimetry. To technically charac-
terise the printed materials the weight, the thickness, and the air permeability were measured.

2 EXPERIMENTAL

A 100% cotton fabric, 150 cm wide, from KOH-I-NOOR, a.s., was used for printing. The fabric has the following
characteristics:
• weight – 144 g/m2,
• thickness – 0.26 mm,
• air permeability – 70.7 l/min (at an area of 10 cm2 and a pressure of 1000 Pa).

2.1 Printing

Digital printing
Digital printing was made using a digital printer Kornit Digital using water-based pigmented inks [5]. Fabric
printing was done directly in the printing machine. The fabric was first placed in the instrument and mounted
in a special frame so it did not move, preventing the blurring of the test pattern during printing. Subsequently,
a pre-conditioning fixation agent containing 5% acetic acid solution was applied to the fabric. The purpose of
this treatment is to fill the voids between the individual mesh fabrics. Printing was done on a wet fabric. The
fabric was printed with a water-soluble Kornit V223 magenta ink. The fabric was placed in the hot-air tunnel
and let to dry. During the first washing, the pre-conditioning agent is removed and the ink stays on the fabric.
Figure 1 represents the ink-jet printed sample.

Figure 1. Direct digital printing on cotton fabric.

Transfer printing
The transfer print was made using a Mimaki JV33-160 digital printer and a special white Turbo print matt foil
from POLI-TAPE GROUP – TAPE, Ltd. [6]. This digital printer uses solvent-based inks for printing. Only magen-
ta ink was used. The test pattern was subsequently cut out on a cutting device. Excessive foil parts without
printed parts that did not create the print motif were removed by hand. The transfer of the test pattern was
done via the DIGITAPE402 transfer foil from ALPHASET d.o.o. [6]. Test pattern was finished in the hot press.
In the press the material are stacked in order – fabric, foil with printed motif, transparent transfer foil with
a glue layer. The transfer was carried out at 150°C for 20 seconds and pressure of 3 Pa. Figure 2 represent the
sample of the transfer print.

8th Conference on Information and Graphic Arts Technology 153

 Poster section

Figure 2. Indirect transfer printing on cotton fabric.

2.2 Analysis

Colour fastness to washing

Printed samples were tested for colour fastness to washing in accordance to standard ISO 105-C10: 2006 [8] at
40, 60 and 90°C. Washing was carried out for 30 minutes at 40 rpm, 1, 3 and 5 times.
The evaluation of colour change and staining of adjacent fabric was done according to the gray scale with
5 grades. The fabric labelled with grade 5 is resistant to the tests and shows no signs of staining. The fabric
labelled with grade 1 is the least resistant to the test.

Crock fastness
Testing was carried out in accordance to standard ISO 105: X12: 1993 [9] to determine the resistance to dry and
wet rubbing. Testing was done for 10 rubbing steps with a frictional force of 9 N. The Electronic Crockmeter
M238BB instrument was used for the measurement.

Resistance to rubbing
Resistance to rubbing was also performed on Martindale apparatus. Samples were subjected to 50, 250, 500,
and 100 abrasion steps at 9 kPa.

Light-fastness – exposure to an artificial light

Samples were illuminated for 72 hours in Xenostest Alpha instrument, where they were subjected to exposure
of an artificial xenon lamp. The test was performed in accordance with ISO 105-B02: 1994 [10].

Colour measurements
A Datacolor Spectrum Flash® SF600 PLUS-CT spectrophotometer was used with the 8° illumination included
geometry to capture colorimetric values (D65 standard illuminant and 10° standard observer). The measured
area was of a circular shape with 9 mm diameter. The measured values were recorded by Datacolor Damaster
software and MS Excel was used for further evaluation. Equations (1-4) were used to calculate colour differ-
ence – ΔE*ab, difference in chroma – ΔC, difference in lightness and darkness – ΔL and difference in hue – ΔH.

ΔE*ab= √((L*s-L*m )2 + (a*s-a*m )2 + (b*s-b*m )2 ) (1)

where are values for standard (test pattern before tested) and are values for measured samples

ΔC= C*s-C*m (2)

where is value for standard (test pattern before tested) and is value for measured sample

ΔL= |L*s- L*m | (3)

where is value for standard (test pattern before tested) and is value for measured sample

ΔH= √(ΔE2-ΔL2-ΔC2 ) (4)

8th Conference on Information and Graphic Arts Technology 154

 Poster section

The evaluation of colour differences and differences in colour attributes was carried out according to [11],
which specifies the deviation and how much the sample is affected before and after the tests.

3 RESULTS & DISSCUSSION

Colour fastness to washing

Samples were washed at 40, 60 and 90°C, once, three and five times. Samples with transfer printing were washed
only twice at 90°C, because the printed foil was released from the fabric during washing. Table 1 shows the
results of colour fastness to washing after the fifth wash. The results are evaluated using the grey scale grade.

Table 1. Colour fastness to washing at 40, 60 and 90°C.

Sample Alta Sta-Cob Sta-Cvc
Temperature [°C] 40 60 90 40 60 90 40 60 90
Digital print 5 5 4 5 5 5 5 5 4/5
Transfer print 5 5 – 5 5 – 5 5 –
a
Alteration of colour, b Staining of cotton adjacent fabric, c Staining of viscose adjacent fabric

The results show that almost the same fastness to washing is obtained by both printing technologies at 40 and
60°C. However, the digital printing with pigmented inks allows also the washing at 90°C, while the transfer
printed foil peeled of the fabric after the second wash.

Crockfastness
Table 2 shows the evaluation of the crock-test by the grey scale grade.

Table 2. Crockfastness evaluation.

Sample Digital print Transfer print
Dry rubbing Wet rubbing Dry rubbing Wet rubbing
Evaluation 4 4/5 4/5 4/5

The results show that the fabrics printed with both printing methods have good abrasion resistance. The
rubbing resistance is not perfect, as stains of printed inks appear on the white adjacent fabric.

The fabric resistance of the Martindale test is shown in Table 3. The test results are evaluated in terms of
CIELab colour differences

Table 3. Rubbing fastness on Martindale instrument

Sample name / Digital print Transfer print
number of rubbing steps ΔE*ab ΔL ΔH ΔC ΔE*ab ΔL ΔH ΔC
50 6 3 1 5 6 3 1 5
250 7 3 1 7 4 2 0 4
500 5 0 0 5 4 2 0 3
1000 5 1 0 5 4 3 0 3

Table 3 shows that the samples changed their shade during rubbing (ΔE*ab), which can be also seen by eye.
There was a big difference in chroma (ΔC), while lightness (ΔL) changed negligible and the colour hue (ΔH)
was preserved.

Colour fastness to light

The results of the blue-scale evaluation of samples exposed to light are shown in Table 4.

Table 4. Colour fastness to light.

8th Conference on Information and Graphic Arts Technology 155

 Poster section

Digital print Transfer print

Evaluation 4 7

The results show that the ink-jet printed sample is more sensitive to light than the transfer printed one. After
72 hours of exposure to light, the ink-jet printed sample showed signs of degradation (area exposed to light
had visible light-stripe), while the transfer printed samples retained its original colour.

Characteristics of the printed fabric

Table 5 lists the common characteristics of printed fabric; weight of fabric, thickness and air permeability
before and after printing, respectively.

Table 5. Properties of the printed fabric.

Property Unprinted fabric Digitally printed fabric Transfer printed fabric

Mass per unit area [g/m2] 144 155 246

Thickness [mm] 0.26 0.33 0.29
Air-permeability [l/min] 70.7 25.3 0

4 CONCLUSIONS

Two techniques, direct inkjet printing and indirect thermal transfer printing on cotton fabric, were tested
in this study. Printed fabrics were exposed to light, washing and rubbing. Results show that the fabric with
transfer printed foil is slightly more resistant than the fabric with digitally printed pigmented inks. The ex-
ception is the resistance to washing at 90°C, where the foil separated from the fabric after the second wash.
Huge differences appear at testing of the physical-mechanical properties of the fabrics. The increasing of the
weight of the fabrics because of the additional ink layer made by digital technology is almost negligible. In
the case of transfer printing the weight of the fabrics significantly increases. The thickness of printed fabrics
at both printing methods is comparable. The air-permeability drops from 70.7 l/min to 25.3 l/min for digital
printing and to 0 l/min for transfer printing. The original fabric properties are better retained at digital print-
ing with pigments than at transfer printing with foil.

5 REFERENCES

• Miles, L.W.C., 2004, Textile Printing , Bradford, Society of Dyers and Colourists
• Freire, E. M., Ujiie, H., 2006. Ink jet printing technology (CIJ/DOD), Digital printing of textiles, Woodhead
Publishing Limited, North America, 29–52
• Forte Tavčer, P., Badjura, S., 2006. “Influence of transfer printing conditions on fastness and covering ability
of pigmented plastisols on cotton knitwear.” Tekstilec, 49(4/6), 81-88.
• Holme, I. 1999. “Adhesion to textile fibres and fabrics.” International Journal of Adhesion & Adhesives 19(6),
455‒463.
• Kornit Digital. URL: https://www.kornit.com/k-solution/cmyk-ink/ (last accessed 10.04.2018)
• POLI-TAPE GROUP – TAPE, Ltd. URL: https://www.poli-tape.de/1/home/ (last accessed 10.04.2018)
• ALPHASET, d.o.o. URL: http://www.alphaset.cz/ (last accessed 10.04.2018)
• ISO 105-C10 Textiles – Tests for colour fastness – Part C10: Colour fastness to washing with soap or soap and
soda, 2006
• ISO 105-X12 Textiles – Tests for colour fastness – Part X12: Colour fitness to dubbing, 1993
• ISO 105-B02 Textiles – Tests for colour fastness – Part B02 :Colour fitness to artificial light: Xenon arc fading
lamp tests, 1994
• Electronics For Imaging, Inc. Delta E, Delta H, Delta T: What Does It Mean?. URL: http://help.efi.com/fieryxf/
KnowledgeBase/color/Delta%20E_H_T.pdf (last accessed 23.04.2018)

Corresponding author:

8th Conference on Information and Graphic Arts Technology 156

 Poster section

Michaela ČESÁKOVÁ
Department of Graphic Arts and Photophysics, Faculty of Chemical Technology
University of Pardubice
Studentská 95, 532 10 Pardubice, Czech Republic
e-mail: [email protected]

Co-authors:
Ondřej PANÁK

Jiří HEJDUK

Petra Forte TAVČER

8th Conference on Information and Graphic Arts Technology 157

 Poster section

COMPARISON OF SCREEN PRINTED AND ARDUINO-BASED SENSORS

FOR ALCOHOL DETECTION

Urška KAVČIČ1, Igor KARLOVITS1 & Matej PIVAR2

Pulp and Paper Institute, Bogišićeva 8, 1000 Ljubljana
1

2
Faculty of Natural Sciences and Engineering, Snežniška 5, 1000 Ljubljana

ABSTRACT: In the field of printed sensors, gas sensors are new coming technology which could be useful in
our daily life. They consist of two printed electrodes that are coated with an active sensing layer, which changes
its resistivity when exposed to the analyte.
In our research study, the possibility of using screen printed gas sensors for alcohol detection in comparison
to a commercial MQ-3 gas sensor was analysed. Screen printed electrodes were printed with conductive sil-
ver printing ink and coated with two polymers (1) ethyl cellulose and (2) poly(ethyl methacrylate) as active
sensing layers in two different thicknesses. As a comparison, a commercial Arduino-based MQ-3 gas sensor
that uses metal-oxide SnO2 as a gas sensing layer was used. The sensors were evaluated with ethanol gas of
200 ppm and 1000 ppm concentrations and appropriate correlation with the type of sensing layer is presented.

Keywords: Screen printed sensors, gas sensor, alcohol detection, Arduino.

1 INTRODUCTION

Screen printing as one of the conventional printing techniques offers printers a possibility to print different
printed matters that are used in everyday life. When using electrically conductive functional printing ink
instead of conventional one, different functional elements can be printed. Different electrically conductive
printing inks printed on various flexible, thin and lightweight printing materials are very promising also for
printing low-cost and disposable printed electronics. Researchers recognized the potential in printing with
functional printing inks for electronic components; especially in the field of printed RFID antennas, batter-
ies and sensors. The last mentioned, printed sensors on flexible substrates, are presenting a growing market
according to IDTechEx who also predicts that the next generation of printed sensors will enable applications
from human-machine interfaces to environmental sensing. (Chansin, 2017)

In the field of printed sensors, gas sensors play a vital role among the most important technologies in our
daily life. Typical applications of gas sensors include the detection of toxic analytes and explosive gases for
the purpose of public and domestic safety, industrial processes, and monitoring of environmental pollution
and air quality. (Zhang, 2016)

Printed gas sensors consist of two printed electrodes that are coated with an active sensing layer. During expo-
sure, the analyte interacts with the active sensing layer and the resistance between the electrodes changes. The
interaction between gas molecules and materials mainly takes place on the surface of an active sensing layer,
hence the selection of active sensing layer has a crucial influence on the sensor performance. (Zhang, 2016)
Gas sensing methods (Liu, 2012) based on detection of electrical variation normally use metal-oxide semicon-
ductors, polymers, carbon nanotubes or moisture absorbing material for active sensing material. Most often
sensors that use metal-oxides have been used, also in different commercially available gas sensors. Those
sensors are affordable and have high sensitivity. The sensitivity of the metal-oxides sensor is correlated with
a principle of operation, that is redox reaction between the target gas and the oxide surface. (Yamazoe, 2008)
And for that, a high operating temperature is required which is a drawback of that kind of sensors.

Many types of researches have been done in the field of printed resistive and capacitive gas sensors (Mattana,
2016), also in the field of ethanol gas sensors but the majority of them are not printed and do not work at
room temperature. Asgari (Asgari, 2017) presents SnO2 decorated SiO2 sensor for ethanol and acetone which
have high selectivity at temperature 70°C to 370°C. The sensor is prepared on alumina and only the active
sensing layer was printed over. Similarly, also Lee (Lee, 2017) presents microelectromechanical system (MEMS)
heater-based SnO2 gas sensor, that works at 344°C and is only partially printed. Lakhane (Lakhane, 2016) pre-
sented screen printed sensors for alcohol using Ca-stilbite and Mg-stilbite that work at temperatures higher
than 65°C.

8th Conference on Information and Graphic Arts Technology 158

 Poster section

Besides metal-oxides, for alcohol detection also enzymes are used. Inkjet printed breathalyser on paper, with
the enzyme alcohol dehydrogenase and its cofactor that detects ethanol as a solution and as vapour, is present-
ed by Bihar (Bihar, 2016). Rama (Rama, 2012) compared electrochemical response of different alcohol enzyme
sensors based on different commercial screen printed carbon electrodes. Gaidan et al. (Gaidan, 2017) present
screen printed sensor for propanol using different compositions of TiO2 and ZnO mixed into polyvinyl butyral
which made the sensor operating at room temperature which is novel, while other metal-oxide sensors work
at higher temperatures.

To overcome this limitation, and to detect gas at room temperature, sensors use polymers as an active sensing
layer. Polymers used for gas sensors can be conductive or non-conductive, but the mechanism of the reaction
of the gases on the surface of the films is still not well understood. (Gaidan, 2017, Liu, 2012) According to
Ruchika (Ruchika, 2015) gas-solid interactions may be of the form of adsorption or chemical reactions.

In our research study, the comparison of a commercial MQ-3 (Hanwei Electronics) gas sensor and screen print-
ed gas sensors for alcohol detection has been done. The sensors were evaluated with ethanol gas of different
concentrations and appropriate correlation with the type of sensing layer and coated thickness was done.

2 EXPERIMENTAL

The interdigitated comb-like electrodes were firstly designed and printed using a semi-automatic screen
printing machine. Electrodes were printed on PVC foil using silver conductive thermal drying printing ink
(SunChemical) with monofilament polyester plain weave mesh 120 l/cm. After printing, samples were dried
10 min at 150°C.

As active sensing layers ethyl cellulose (EC) and poly(ethyl methacrylate) (PEMA) were used. Firstly both
polymers (EC and PEMA), that have been purchased as a powder, were dissolved. A solution of EC and PEMA
have been printed over electrodes in two thicknesses (using screen mesh of 62 l/cm and 100 l/cm).

The preparation of gaseous ethanol followed. Ethanol was diluted with water to get a concentration in the
gas phase of 200 ppm and 1000 ppm (according to Henry’s law (Sander, 2015)).

After that, each sensor element (with different active sensing layer and different thicknesses) was exposed to
200 ppm and 1000 ppm of gas ethanol and the resistance was measured using multimeter Fluke. The sensors’
resistance was measured for 2 minutes, 1 min in ethanol gas and 1 min after the sensor was removed from
ethanol gas.

The same procedure was taken for the MQ-3 sensor.

Besides that FT-IR spectroscopy has been done to get the information about possible changes in chemical
structure of active sensing layer (EC and PEMA) when exposed to ethanol gas.

3 RESULTS & DISCUSSION

Even though the compared gas sensors (printed and the tested commercial one) do not work on the same
principle, all sensors give an immediate response to ethanol gas. Printed sensors use polymers as active sensing
layer and work at room temperature without heating, while the tested cheap commercial one uses metal-oxide
(SnO2) that needs the heater to provide the necessary working condition.

The comparison of a change in resistance of sensors (with different active sensing layer of different thickness
and exposed to different ethanol concentration) is shown in Figures 1, 2 and 3. It is obvious, that all samples
did not conduct when not exposed to ethanol gas, and started to conduct in first 3 seconds after exposure
to ethanol gas. The resistance dropped immediately and stabilised after approximately 30 seconds. When the
samples were taken out of the ethanol gas, samples stopped conduct in less than 15 seconds. Oppositely, the
commercial MQ-3 gas sensor provides intrinsic values of the sensor ranging from 27 in the air to almost 275
when exposed to 1000 ppm of ethanol gas (Figure 4) and it takes more than a minute to get back to the initial
value.

Comparing active sensing layers (Figure 1), one can observe that PEMA gives a better response (the change
in resistance in one minute of exposure to ethanol gas is larger) than EC. It is seen that PEMA reaches lower

8th Conference on Information and Graphic Arts Technology 159

 Poster section

values of resistance than EC within 1 minute, other obvious differences are not noticeable. We suppose that
the reason for that could be moisture itself that can change the resistance of the sensor as well.

The comparison of response in different gas concentration (Figure 2) reveals, that there is not as huge differ-
ence as one could expect. The change in resistance of printed sensors, on the other hand, does not change
proportionally to analyte concentration but the difference in response of the MQ-3 sensor presents the con-
centration much more.

The difference in thicknesses of active sensing layer does not have great influence on the resistance change in
our case, only slight delay in response when testing sensors with the thicker active sensing layer.

Also the design of experiments (DoE) showed that the greatest impact on change in resistance after 30 second
of exposure to ethanol gas has the selected active sensing layer. Thickness of the active sensing layer and gas
concentration have same, much lower impact and do not show a significant difference in resistance change.

Figure 1. Resistance of the sensors with EC (left)

and PEMA (right) active sensing layer.

Figure 2. Resistance of the sensors when exposed to 200 ppm (left)

and 1000 ppm (right) concentration of ethanol.

Figure 3. Resistance of the sensors with active sensing layer printed with screen 62 l/cm (left)
and 100 l/cm (right).

When testing printed sensors, that were printed under constant conditions, it was noticed that there are also
major fluctuations among samples, especially if the same sample was exposed to ethanol gas more times.

8th Conference on Information and Graphic Arts Technology 160

 Poster section

Therefore the FT-IR analysis of active sensing layer has been done. In Figure 4 one can observe, that the in-
tensity of peaks at different wave numbers fall when EC is exposed to ethanol gas and increase when PEMA
is exposed to ethanol gas in comparison to samples that were not exposed to ethanol gas. From the results of
FT-IR analysis, it can be assumed that there was a minor chemical change on the surface of the active sensing
layer and consequently that the samples of sensors are not completely reversible, which means that that kind
of printed sensors is not reusable.

Figure 4. ATR FT-IR spectre of samples with EC (left) and PEMA (right) active sensing layer exposed
(red line) and not exposed (blue line) to ethanol gas.

On the other hand, the commercial MQ-3 sensor (Figure 5) could be used many times while it has high repeat-
ability, high sensitivity to alcohol, fast response, is stable and has a long life expectancy. Its response is fast
and it is highly dependent on gas concentration. Contrary to printed sensors, that has very low recovery time
(few seconds), the commercial MQ-3 sensor uses more than a minute to recover.

Figure 5. Response of the MQ-3 sensor.

Comparing sensors’ recovery time, it is obvious, that printed sensors have shorter recovery time than commer-
cial one. Besides that, printable sensors could be used immediately, while commercial one needs for about 10
minutes to heat up and to work properly.

4 CONCLUSIONS

A good sensor has to be sensitive, stable, reversible, selective, reliable and has to response quickly. According
to that, it can be concluded that all sensors (printed and commercial one) are sensitive to ethanol at room
temperature and have a quick response, but it can not be claimed that sensors (printed and commercial one)
are selective. All sensors are dependent at least on humidity and temperature too. They are all reversible at
first sight, but as it was seen from the FT-IR spectre, there are some chemical changes left on the surface of the
active sensing layer of printed sensors after exposure to ethanol. In the future, the active sensing layer (before
and after exposure to the analyte) should be analysed in detail and the operation principle should be studied
deeply. The stability of the MQ-3 sensor is regular and the amount of resistance change is proportional to the
concentration of the analyte. Printed sensors, on the other hand, are still in research phase, as stability and
reliability are questionable.

ACKNOWLEDGEMENT
The work was carried out within the RDI project Cel.Cycle: »Potential of biomass for development of advanced
materials and bio-based products« (contract number: OP20.00365), co-financed by the Republic of Slovenia,

8th Conference on Information and Graphic Arts Technology 161

 Poster section

Ministry of Education, Science and Sport and European Union under the European Regional Development
Fund, 2016–2020.

5 REFERENCES

• Asgari, M., Fahimeh, H. S., Yadollah M., et al. 2017. “SnO2 decorated SiO2 chemical sensors: Enhanced sensing
performance toward ethanol and acetone.” Materials Science in Semiconductor Processing 68: 87-96.
• Bihar, E., Deng, Y., Miyake, T., et al. 2016. “A Disposable paper breathalyzer with an alcohol sensing organic
electrochemical transistor.” Scientific reports. URL: ehttps://www.nature.com/articles/srep27582#supple-
mentary-information (last accessed on 13. 3. 2018)
• Chansin, G. Printed and Flexible Sensors 2017-2027: Technologies, Players, Forecasts. URL: https://www.idte-
chex.com/research/reports/printed-and-flexible-sensors-2017-2027-technologies-players-forecasts-000504.
asp. (last accessed on 30. 1. 2018).
• Gaidan, I., Brabazon, D., Ahad, I. U. 2017. “Response of a Zn2TiO4 Gas Sensor to Propanol at Room Temper-
ature.” Sensors (Basel, Switzerland), 17(9): 1995.
• Lakhane, M. A., Choudhari, A. L., Khairnar, R. S., et al. 2016. “Alcohol Sensor Based on Mg-STI zeolite Thick
Films.” Procedia Technology, 24: 595–602.
• Lee, D. H., Sun, K. K., Yusin P., et al. 2017. “Transfer of preheat-treated SnO2 via a sacrificial bridge-type ZnO
layer for ethanol gas sensor.” Sensors and Actuators B: Chemical no. 2 (255): 70-77.
• Liu, X., Sitian, C., Hong L., et al. 2012. “A survey on gas sensing technology.” Sensors (Basel, Switzerland) no.
12 (7): 9635-9665.
• Mattana, G., Briand, D. 2016. “Recent Advances in Printed Sensors on Foil.” Materials Today, 19(2): 88−99.
• Rama, E. C., Biscay, J., González García, M. B., et al. 2012. “Comparative study of different alcohol sensors
based on Screen-Printed Carbon Electrodes.” Analytica Chimica Acta no. 728: 69-76.
• Ruchika, K. A. 2015. “Performance analysis of Zinc oxide based alcohol sensors.” Int. Journal of Applied
Sciences and Engineering Research no. 4 (4): 427-436.
• Sander, R. 2015. “Compilation of Henry’s law constants (version 4.0) for water as solvent.” Atmos. Chem.
Phys. no. 15 (8): 4399-4981.
• Yamazoe, N., and Kengo S. 2008. “Theory of power laws for semiconductor gas sensors.” Sensors and Actu-
ators B: Chemical no. 128 (2): 566-573.
• Zhang, J., Liu, X., Neri, G., Pinna, N. 2016. “Nanostructured Materials for Room-Temperature Gas Sensors.”
Advanced Materials, 28(5): 795−831.

Corresponding author:
Urška KAVČIČ
Pulp and Paper Institute, Print and Packaging Department
Bogišićeva ulica 8
1000 Ljubljana, Slovenia
phone: +386 1 200 28 47
e-mail: [email protected]
web: www.icp-lj.si

Co-author(s):
Igor KARLOVITS
Pulp and Paper Institute, Print and Packaging Department
Bogišićeva uliva 8
1000 Ljubljana, Slovenia
phone: +386 1 200 28 47
e-mail: [email protected]
web: www.icp-lj.si

Matej PIVAR
Faculty of Natural Sciences and Engineering
Snežniška 5
1000 Ljubljana, Slovenia
phone: +386 1 200 32 44
e-mail: [email protected]
web: www.ntf.uni-lj.si

8th Conference on Information and Graphic Arts Technology 162

 Poster section

CREATION OF INTERACTIVE BOARD FOR LEARNING ABOUT TYPOGRAPHY

Tevž ZRINSKI, Klementina MOŽINA & Dejana JAVORŠEK

University of Ljubljana, Faculty of Natural Sciences and Engineering, Department of Textiles, Graphic Arts
and Design, Chair of Information and Graphic Arts Technology, Snežniška 5, 1000 Ljubljana, Slovenia

ABSTRACT: Typography is all around us, as it can be seen on every step. But understanding of typography
requires interaction between the observer and typefaces, that standard learning methods cannot provide.
The main goal of our research was to design interactive board that will help in understanding of typography.
In the introduction, standard and interactive teaching methods were described along with the advantages
of interactive methods. In the experimental part, the possibilities of preparing letters for 3D printing using
Blender, were investigated. A comparison between different types of communication was made. All the com-
ponents used for communication were tested and then assembled into a functioning whole. The commands
with which the Arduino microcontroller manages all the components were designed and tested. Our final
product includes different fields of profession (3D printing, electronics, typography) that create functional
and visually attractive interactive board for the purpose of interactive teaching of typography.

Keywords: 3D print, Arduino, Blender, interactive board, typography.

1 INTRODUCTION

Educational games improve the acquisition of knowledge when compared to traditional methods, and increase
the sense of satisfaction that occurs. The use of educational games led to changes in relation to work due to
the high level of involvement in the process. Positive feelings that students feel when solving the game do
not contribute fully to improving the knowledge, but it increases the desire for progress (Giannakos, 2013). By
using interactive games, students also improve, the challenge solving skills (Green, Batool, 2017).
There are increasing initiatives for the transition from traditional to interactive methods of learning, since
they represent fewer restrictions and open the door to a new world of discoveries and learning process (Pamuk
et al., 2013). In the last 20 years, interactive games have been developing and continually incorporating new
technologies to enhance gaming experience.
Interactive games reflect changes in the media culture as they bring together various media with many pur-
poses. They represent a new direction in education by combining new media, learning, and playing. Optimal
use of interactivity allows the user a sufficient degree of control, where there is also the possibility of failure.
By doing this, the user is constantly attentive to the individual steps that should be performed. Communi-
cation between the user and the game is crucial because it is an intermediate link that keeps the user on the
right track (Booth, 2016). The use of interactive technologies suggests that simple use and interactivity have
influence on emotional processes (authenticity, cognitive activity) that affect memory. The results show that
when users perceive interaction with technology as intuitive and interactive, they experience a higher level
of understanding and involvement (Pallud, 2017). An interactive educational game based on cooperation, im-
prove the learning process, raise motivation, improve student’s achievements and self-efficacy (Sung, Hwang,
2013). Learning process and motivation do not depend only on the entertainment side of the game, but on the
ability to constantly transfer the knowledge to the student. In this context, precisely instruction and guidance
through the game provide a successful path of education. It would be very interesting to include more edu-
cational games into the work program in different fields of education, since it is well known that the use of
educational games leads to a higher level of motivation, enjoyment and knowledge acquisition (Erhel, Jamet,
2013). An interactive learning board is a fun learning tool that helps students to recognize the advantages and
disadvantages of their knowledge. Additionally, it adds a new dimension to teaching and learning.
Interactive boards are based on a variety of low cost devices that allow the rapid and efficient development
of prototypes. Such an example of the device is the Arduino microcontroller (Irigoyen, Larzabal, Priego, 2013)
that is easy to use, compatible with many electronic components and has broad support in the Arduino com-
munity (Jamieson, Herdtner, 2015).
An example of use of the Arduino microcontroller is an interactive science board where Arduino Nano controls
the operation of the LCD screen, MP3 player, keys and individual LED lights. The components were connected
to each other using a circuit with the exact connection plan (Make: DIY Projects nad Ideas for Makers, 2016).
Method that we used to produce components of interactive boards is 3D printing technology, also called as
the pioneer of the new industrial revolution. In comparison with other technologies 3D printing technology

8th Conference on Information and Graphic Arts Technology 163

 Poster section

allows creation and adaptation of items to the specific requirements (Weller, Kleer, Piller, 2015). 3D technol-
ogies can be used in various fields, with the printing of various materials (Muck, Križanovskij, 2015).
The main goal of our research was to design and create visually attractive interactive board that will help in
understanding of typography.

2 EXPERIMENTAL

2.1 Materials

The main part of the interactive board is the Arduino Uno microcontroller (Arduino), which enables com-
munication with all the components used on the interactive board. Other components include: LCD display
(Winstar), MCP chip (Microchip Technology Inc.), LEDs (Kingbright), RGB LEDs (Optosupply), electric wires,
metal plates, self-adhesive magnetic tapes and plexiglass.
For the preparation and printing of 3D letters, we used an open source program Blender (Blender Foundation)
and CubePro (3D Systems).
The application Blender was used for 3D modeling of the letters. Then, we set the appropriate settings for 3D
printing and exported them in CubePro. We printed the letters with the Cube Duo Pro (3D Systems) printer
using ABS material.

2.2 Methods

Design of interactive board: Program code for recognising letters and their order was created and implemented.
Each letter of the typeface had only one correct position on the interactive board. This would be implemented
by specifying a unique communication that would make the letters recognizable to controler, by using the path
of an electric current. A properly-placed letter would trigger a response that would be interpreted by microcon-
troller and converted to user recognisable output. Any correct layout of the whole set of letters would turn on
the green LED. By correctly placing all the letters on the interactive board, the user would receive a LED signal
that would indicate the correctly completed game. The interactive board would have a on/off power switch.

Choosing 12 typefaces for the interactive board: When choosing the typefaces for the interactive board, several
factors were taken into account, such as historical importance, purpose and style of diversity. We chose the
typefaces from six styles: Humanist, Garald, Transitional, Modern, Slab-serif and Sans serif. Based on certain
factors, we made a selection of 12 typefaces, two from each style.

Restriction of the set to twenty-five letters: We compared all the lower-case letters of the Slovene alphabets.
From the set of lower-case letters we eliminated all letters consisting of two parts, such as letters with carons
and dots. We also removed letters that due to their simplicity do not contain enough information about type-
face family. We chose among letters with many different strokes. Our final choice was three letters consists
of x-height and upper case, x-heignt, and x-heignt and lower case. The layout of letters on the support box
according to the basic position of the base line is shown in Figure 1.

Figure 1. The layout of letters on the support box according to the basic position of the base line.

8th Conference on Information and Graphic Arts Technology 164

 Poster section

Defining the text that describe appropriate letter: We have defined the description of the letter image, the
style of letter, and the historical information about the typeface. With terms, we created letter descriptions,
which help us to determine the position of the letters on the interactive board. The terms were then included
on the interactive board under each set of letters of one typeface.

Selection of 3D printing material: The selection of various materials for printing was performed by compar-
ing the properties and purpose of using both materials for 3D printing. We compared the more flexible and
durable ABS and a PLA material that is more rigid.

Defining the properties of the printed letters: The properties of the printed letters were determined by com-
paring the different type sizes and errors that occur during printing. The size of the letters was determined
by testing the layout of the letters on the carrier box. By calculating the material usage and the relative costs,
we compared two methods of printing – printing of hollow letters that do not contain fillings and full-letter
printing. The final decision was based on cost savings.

3 RESULTS & DISCUSSION

In the first step of our research – when preparing the letters in the Blender program, we wanted to found out
whether it is possible to print the letters of all the styles with the 3D printing technique.
Letters when imported in Blender were created with many poligons and were not suitable for further manipula-
tion in 3D program. If they were used as such they produced uneven surfaces, duplicate edges and random dots.
Most errors were represented by intersect faces. The number of cross sections was too large to be manually
removed. We tried alternative solutions to create space in letters using the method of counting the volume
of characters from the volume of the letter, but the methods did not solve the challenges of the complexity
of the letters network. Often, the problem has become even worse and more difficult to solve. We also tried
to import the border of letters created in Adobe Illustrator into the Blender program and build a three-di-
mensional image of the letters, but this method also encountered similar problems. We have found out that
the errors originate mainly from the fonts themselves, which are designed in a way that it is not suitable for
editing in the Blender. The errors were not present when solid leters were printed, but it resulted in greater
use of material.
Better method would be to print holow letters because the printed walls do not extend beyond the outer edge
of the letters. We also found out that the overhang faces were acceptable for printing if the errors were at the
bottom of the letter. Cross-sections were not affected in the print since they were no longer present in the
structure after exporting the .stl file, they were fused together in one image and they did not extend beyond
the image of the letters.
With the experimental prints of the letters of different styles, we wanted to determine whether all the type-
face styles can be correctly printed. Without errors hollow letters of Slab-serif and Sans-serif typefaces were
printed. In some cases, the printing of Humanist, Garalde and Transitional typefaces caused problems, while
in the Modern typefaces we did not manage to avoid printing errors. We have found out that if we want to use
only the typefaces that were printed without error, we should greatly narrow the typeface selection, which
would consequently mean that the interactive board would not contain selected typographic styles. Moreover,
it would contain only typefaces with small differences and thus would lower the diversity. It would be also
difficult to identify typefaces and find the correct position on the interactive board. However, as the full-let-
ter printing did not allow the correction of errors in the printing of letters with thin strokes, we decided to
use a set of all selected typefaces on the interactive board, and manually repaired the errors that occurred
during (Figure 2). Subsequently, the letters were sandwiched on parts where we casted the material and thus
eliminated numerous irregularities.

8th Conference on Information and Graphic Arts Technology 165

 Poster section

Figure 2. Letters "d” and "g” with corrected errors of uncompleted surfaces.

When writing code for the operation of the Arduino microcontroller, we found out that the components can
be managed with different commands. Through the code, for each digital pin that allows communication, we
determined the direction of the electric current and the command that is made when changing it. By doing
so, we have set the conditions on which different commands are executed, depending on the status recorded.
The result was a set of settings and commands, tested with different methods of work.
With the preparation and 3D printing, we managed to create 3D image formats representing different typeface
styles. We managed to print the details of the letters, so the letters retained their basic images. From the set of
different components we managed to select those who added value to the interactive board and enabled the
communication between the user and the board. We have prepared a sequence of commands that allow the
interactive operation of the board using the Arduino microcontroller. By creating a working interactive table
(Figure 3), we achieved the goal, since we created a learning aid for learning about typography.

Figure 3. Interactive board during operation.

8th Conference on Information and Graphic Arts Technology 166

 Poster section

4 CONCLUSIONS

During the creation of the interactive board, we had to select the typefaces that are an important part of the
interactive board and prepared them for 3D printing with the Blender program. With the study of different
typeface styles, we determined the criteria by which we prepared the letters of individual styles for 3D print-
ing. With the help of the Arduino microcontroller and appropriate components, we tested different ways of
communication between the user and the interactive board. Our final product is a fusion of different fields,
such as 3D printing, electronics, typography and graphic design. Together, they create functional and visually
attractive board, which serves the purpose of interactive teaching of typography.

5 REFERENCES

• Booth, P. 2016. “Board, game, and media: Interactive board games as multimedia convergence.” The Interna-
tional Journal of Research into New Media Technologies, 22(6): 647–660.
• Erhel, S. and Jamet, E. 2013. “Digital game-based learning: Impact of instructions and feedback on motivation
and learning effectiveness. ” Computers & Education, 67: 156–167.
• Giannakos, M. N. 2013. “Enjoy and learn with educational games: Examining factors affecting learning per-
formance.” Computers & Education, 68: 429–439.
• Green, Z. A. and Batool, S. “Emotionalized learning experiences: Tapping into the affective domain.” Evalu-
ation and Program Planning, 62: 35–48.
• Irigoyen, E., Larzabal, E. and Priego, R. 2013. “Low-cost platforms used in Control Education: An educational
case study.” IFAC Proceedings Volumes, 46 (17): 256–261.
• Jamieson, P. and Herdtner, J. More Missing the Boat - Arduino, Raspberry Pi, and Small Prototyping Boards
and Engineering Education Needs Them. Frontiers in Education Conference (FIE), 2015 IEEE. URL: <http://
www.users.miamioh.edu/jamiespa/html_papers/fie_2015_ar.pdf> (last accessed on 15. 6. 2017).
• Muck, T. and Križanovskij. I. 2015. 3D-tisk : [--- tehnologije 3D-tiska, priprava 3D-modelov za tisk, pojmovnik,
---]. Ljubljana: Pasadena.
• Pamuk, S., Çakir, R., Ergun, M., Yilmaz, H. B. and Ayas, C. 2013. “The Use of Tablet PC and Interactive Board
from the Perspectives of Teachers and Students: Evaluation of the FATİH Project.” Educational Sciences:
Theory & Practice, 13 (3): 1815–1822.
• Pallud, J. 2017. “Impact of interactive technologies on stimulating learning experiences in a museum. ” In-
formation & Management, 54: 465–478.
• Sung, H. and Hwang, G. 2103. “A collaborative game-based learning approach to improving students’ learning
performance in science courses.” Computers & Education, 63: 43–51.
• These Middle Schoolers Built a Custom Arduino Board Game to Explore Ancient Egypt. Make: DIY Projects
nad Ideas for Makers. URL: <http://makezine.com/2016/04/13/these-middle-schoolers-built-custom-ardui-
no-board-game-explore-ancient-egypt/>(last accessed on 15. 6. 2017).
• Weller, C., Kleer, R. and Piller, K. T. 2015. “Economic implications of 3D printing: Market structure models in
light of additive manufacturing revisited.” Int. J. Production Economics, 164: 43–56.

Corresponding author:
Dejana JAVORŠEK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Chair of Information and Graphic Arts Technology
Snežniška c. 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 37, e-mail: [email protected]

Co-author(s):
Tevž ZRINSKI
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Chair of Information and Graphic Arts Technology
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 37, e-mail: [email protected]

Klementina MOŽINA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Chair of Information and Graphic Arts Technology
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 42, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 167

 Poster section

DESIGN OF A GIFT CERTIFICATE USING PROTECTIVE ELEMENTS IN A STANDARD TRIAD

Nadezhda SAVCHUK, Olga NOVOSELSKAYA

Byelarussian State Technological University / Informatics and Web-Design Dept.,
Minsk, Republic of Belarus

ABSTRACT: Features of graphic design for reproduction a gift certificate are considered in the article. Despite
the development of electronic payment systems and electronic document management systems, cash tickets
and documents on paper are widely used at present and there is no significant reduction in their number in the
future. One of the main peculiarities of gift certificates is the possession of copywriting and the presence of
some currency value equivalent adopted in the state of the issuer. In addition, an integral part of the standard
gift certificate should have beautiful design and high quality of printing. The examples of gift certificates with
full color screenless images by the technology formulate above are given in the work.

Keywords: rose elements, protection, design, gift certificate, printing production.

1 INTRODUCTION

Practically in any beauty salon, boutique or supermarket you can buy a beautifully decorated document for
a certain amount of money - a gift certificate. A gift certificate (gift card) is usually a card (certificate) or a
plastic card that has some form of copy protection, which carries a certain equivalent value in the currency
accepted in the issuer’s country.

Gift certificates are not only a way to profitably sell goods and services. This is also a way to attract new cus-
tomers. After all, a certificate can be given to a person who has never heard of a company, and, since the services
have already been paid, they will definitely use them. So the gift certificate will also play an advertising role.

The main requirement for the design of the gift certificate is the availability of the name of the organization
and the amount for which the purchase can be made. In addition, the design should meet the expectations of
the audience, as well as give birth to images, associations and emotions.

Most companies make a unified design of the gift card, regardless of denomination.

Figure 1. Examples of gift certificates.

Another majority of companies change the design or color scheme of the gift certificate depending on the
denomination.

8th Conference on Information and Graphic Arts Technology 168

 Poster section

Figure 2. Examples of gift certificates.

At all times, there have often been cases of making fakes of various printing products. To avoid this, a lot of
special protections for printing products were invented, many of which are used today.
All produced printing products have a certain set of protective elements, which corresponds to the level of
value of the manufactured product and the wishes of the client.

A set of protective elements of printing products can consist of a whole complex of means of protecting the
product from counterfeits. Such means of protection can include the following:
• Rose elements;
• latent (latent) elements;
• relief grids;
• special protective rasters;
• microtext;
• graphic elements of protection from photocopying and others.

2  EXPERIMENTAL

Autotyped color synthesis - obtaining shades of color on the print by combining raster or line images printed
with colors of different colors, for example, triad colors: cyan, magenta and yellow (CMY). Images consist of
strokes having the same thickness of the ink layer on the impression. Microelements have a different area and
constant frequency. In this case, the total color halftone image is formed using three colors. When overlaid
raster elements of the image on the impression during printing. Autotype color synthesis has a mixed addi-
tive-subtractive character.

Previously developed elements were analyzed using various scanning techniques. Originally imprinted on a
digital printing machine, the image was scanned at a low resolution. The result of the scan showed that the
boundary contours cease to be perceived as composite, and reproduced in the form of a raster. An example
of the result of scanning and determining the color coordinates on the border zones is shown in Fig. 4. The
coordinates of point 1 (pure yellow on the Yellow-Magenta border) are equal to Cyan 4%, Yellow 64%. Point
2 (pure purple on the border of Magenta -Yellow) - Magenta 44%, Yellow 38%. Point 3 (neutral color between
the lines of magenta and yellow) - Magenta 26%, Yellow 47%. Point 4 (pure purple on a white background) -
Magenta 64%, Yellow 4.

Figure 3. The result of scanning and determining the color coordinates.

8th Conference on Information and Graphic Arts Technology 169

 Poster section

An analysis of the data shows that on the boundary zones the sum of the coordinates gives shades of red.
Pure colors give a 64% sensation of a plicate yellow and magenta color with a small content (4%) of cyan and
yellow, respectively. That is, no point is reproduced with a given color coordinate. If you do not know about
the parameters of the color guilloche element initially, then it will be impossible to reproduce the similar
effect of the blurred technology in the CMY system.

Another task was to determine the maximum color coverage of the CMY system. Color coverage is the number
of shades that can be rendered using a certain set of colors.

To establish the maximum color coverage by scaling, the square-shaped elements shown in Figure 4 were
created. To paint the outline, all possible combinations of three colors were used. The results were printed on
an ordinary white sheet and on protected paper, adding figures dyed with clean spot colors.

Figure 4. Elements for establishing the maximum color gamut.

The color information was read using the X-Rite 500 spectrodensitometer. Conversion of the color coordinates
to the xyZ system allowed us to graphically display the CMYK standard colors in the form of dots of uniform
colors and reproduced using patterns on the chromaticity diagram xy. The result is shown in Figure 5.

Protected paper
White paper

Figure 5. Color gamut.

8th Conference on Information and Graphic Arts Technology 170

 Poster section

Figure 5 shows that there is a shift in color coverage. Pure spot colors are more shifted in the red color zone.
Colors obtained as a result of autotypic synthesis in the zone of blue flowers. Displacement occurs because of
the lack of whiteness of the protected paper, manufactured without special bleaching agents. Some points on
the chart are duplicated. You can see that these are the coordinates of the colors that were created by different
combinations of all three elements of the CMY system, from which it follows that the color is practically in-
dependent of the sequence of arrangement of the color contours. It also follows from the graph that the area
of ​​color coverage obtained as a result of autotypic synthesis is 5 times smaller than the area of color
​​ coverage
of pure spot colors, which proves an increase in the degree of protection from falsification when using this
method of protection.

The displacement of the axis of the locus is due to the fact that all the elements were not located closely (the
outline to the outline), but with a certain step that gave the color tone bleach, the color shift and its rotation,
i.e. the color of the substrate changes the sense of color tone.

For the characteristics of information, units of measurement are used, which allows some piece of information
to ascribe numbers-the quantitative characteristics of information. At the moment, the following methods for
measuring the amount of information (measure of the amount of information) are the most known: structural
(volume), entropic, algorithmic.

The most common methods for determining the information capacity for text and raster graphics are based
on the Hartley formula:
I=l×log2h,

where h — the base of the number system (the number of states that an element holding a given number can
take); l — number of elements.

As for vector graphics, well-known techniques for determining the information capacity were not revealed.
Therefore, the task was to determine what the information volume of the vector graphic file depends on. For the
initial analysis, the files created in the CorelDRAW graphics editor were taken. The volumes of an empty file, a
file with a colorless object, files with black and colored objects, a file with several objects were analyzed. The
following formula for the determination of the information volume of a vector file is derived by experience:

h=Vб+Nв∙512+Vцв,

где – base file size * .cdr with built-in color profile; – number of vertices; – the volume occupied by the color.
Also similar dependence is observed in files of * .svg format.

3 RESULTS & DISCUSSION

One way to protect is applying rose elements. Rose elements are continuous interlacing narrow lines forming
complex geometric figures; Used as a background on the print. Means of protection against forgery of bank-
notes and securities.

In CorelDRAW, such elements can be created from the simplest shapes by turning and scaling. From the rose
elements, replacing the black and white contours with colored ones, we obtain diffraction patterns. Combining
cyan, yellow and magenta can create a large number of shades.

Figure 6. Creature of rose elements.

8th Conference on Information and Graphic Arts Technology 171

 Poster section

Based on the previously created diffraction patterns, a version of the gift certificate was developed.

Figure 6. Gift certificate with rose elements and standard gift certificate.

The calculation made it possible to determine that with the use of rose elements the information capacity of
the gift certificate increases 4.5 times.

4 CONCLUSIONS

Analysis shows that the most effective protections are hidden protections, which can be identified only in
conditions of professional environment (in expert laboratories and equipped certificate centers). The applica-
tion of this form of protection is most appropriate for a documentary group of product. To create a protected
product with a unique design it is proposed to use rose elements that provide copywriting and better visu-
alization of printing production.

Rose elements are built on the laws of symmetry which allows them to be introduced into software products by
their design. Color rose elements create imitation of rainbow printing. In this case, the feature of the element
is the preservation of the bar-like elements in multi-color printing even with the use of standard printing
equipment. Composition of designed rose elements is based on without screening technology. The effect of
iridescence is realized due to mixing of subtractive and additive types of synthesis. Subtractive synthesis is
used due to printing technique, and additive one is a result of visual intersections of rays from ink reflections
of CMYK triad base. Program realization of this method helps to include such elements in any type of printing
production and facilitate the process of its designing.

Work include the justification for the method of protection. The technology for assessing the degree of pro-
tection is based on determining the structure of file size and then equation the information capacity index for
a vector image, taking into account the features of its structure, the complexity of the curves being formed,
and the quantitative analysis of the points forming the basis of the vector image and the levels of symmetric
reproduction.

An alternative option for calculating the information capacity index extends the methods for assessing the
results of scientific research and allows them to be applied in quantitative description of digital and graphic
information, to systematize data on the structure of graphic security elements and to bring its mathematical
description.

5 REFERENCES

• Koshin, A. A. 1999. Protection of polygraphic products from falsification. M: Sinus.

• Meresin, VM Protected polygraphy: reference book / VM Meresin. - Moscow: Flint, 2012. - 640 p.
• Medyak, DM Protection of polygraphic products: educational-methodical manual / DM Medyak, MI Kulak.
- Minsk: BSTU, 2013. - 86 p.
• Osokin, AN Information theory: textbook / AN Osokin, AN Malchukov - Tomsk: Tomsk Polytechnic Univer-
sity, 2014. - 208 p.
• Savchuk, N., Novoselskaya, O. 2017. “Rose Elements as a Protection for Printing Production.” Proceedings
of the 7th International Scientic Conference “Printing Future Days 2017” October 04 - 06, 2017. Chemnitz,
Germany.

8th Conference on Information and Graphic Arts Technology 172

 Poster section

• Sharifulin M. Defense first of all // Publish, 2000. - № 7. - P. 21.

• Larionov VG, Skripnikova MN How to protect yourself against forgery. / / Marketing in Russia and abroad,
2001. - № 3. - P. 25.

Corresponding author:
Nadezhda SAVCHUK
Informatics and Web-Design Department
Faculty of Information Technology, Belarusian State Technological University
13a, Sverdlova str.
220006, Minsk, Republic of Belarus
phone: +375 29 241 34 72
e-mail: [email protected]

Co-author(s):
Olga NOVOSELSKAYA
Informatics and Web-Design Department
Faculty of Information Technology, Belarusian State Technological University
13a, Sverdlova str.
220006, Minsk, Republic of Belarus
phone: +375 29 273 70 97
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 173

 Poster section

DEVELOPMENT OF dispersion barrier coating FOR GREASE,

MINERAL OIL AND WATER VAPOUR

Jelena VASILJEVIĆ1, Barbara GOLJA1, Ivan JERMAN2, Ervin ŠEST2, Matic ŠOBAK2, David RAVNJAK4, Mateja
KERT1, Marija GORJANC1, Brigita TOMŠIČ1, Tea KAPUN3, Diana GREGOR SVETEC1 & Barbara SIMONČIČ1
1
University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia
2
National Institute of Chemistry, Slovenia
3
Pulp and Paper Institute, Slovenia
4
Paper Mill Vevče d.o.o., Slovenia

ABSTRACT: The repulpable and more sustainable grease, mineral oil and water vapour barriers for the safe
paper-based packaging were created by dispersion coating. All created composite barrier coatings were wa-
ter-based and prepared from dextrin, latex and two barrier components B1 and B2. The created barriers enabled
the achievement of the highest KIT value equals to 12 and decrease of the Cobb-Unger oil absorbency of 0.14
g/m2, i.e. decrease of about 90%. Simultaneously, these barriers decreased the water vapour transmission
rate for about 80%. The prepared multifunctional barrier coatings can be considered as repulpable and more
sustainable, which is of high importance for the development of the safe paper-based packaging as alternative
to currently used bioaccumulative plastic counterparts.

Keywords: dispersion coating, paper, grease and mineral oil barrier, water vapour barrier.

1 INTRODUCTION

Packaging provides protection against environmental influences, i.e. light, water vapour, oxygen, microorgan-
isms, aromas as well as against grease and oils. Typical plastic packaging materials or barrier coatings used in
the industry are based on the synthetic polymers (polyolefins, PET, polyamides, PLA, EVOH, PVDC, etc.) and
their combinations in the laminated structures. However, due to the high stability of these petroleum-derived
polymers, the latter negatively contributes to the global environmental pollution and the preservation of the
limited petroleum resources.
Therefore, the current research is directed towards the development of new green alternative composite coat-
ings, which should be prepared from more sustainable and renewable resources, and should be more repulpa-
ble and recyclable, biodegradable, and suitable for direct food contact (Lagaron et al. 2017). These composite
coatings are created by dispersing different micro- and nano-sized materials in the continuous matrix (Wang
and Jing, 2017; Weizman et al. 2016; Ovaska et al. 2016). Consequently, the chemistry of the continuous phase
and the surface of the dispersing phase dictate the interface interactions, which influence the properties and
tortuosity factor of the created composite coatings.
The objective of our study was to develop more sustainable repulpable water-based dispersion coating for
the paper-based food packaging. Oil and grease resistance, and resistance to water vapour transmission were
selected as the main barrier properties aimed to be achieved. The evaluation of the viscosity of the prepared
dispersions and the barrier properties of the coated papers enabled selection of the barrier coatings with the
best performance.

2 EXPERIMENTAL

2.1 Materials and application procedure

Two different non-wet-strength label papers P1 (80 g/m2) and P2 (55 g/m2) were provided by Paper Mill Vevče
d.o.o. (Ljubljana, Slovenia). The coating dispersions were made from dextrin, latex and two barrier components
B1 and B2 of three increasing concentrations c1, c2 and c3. The coatings were prepared on the back side of the
papers by using the RK Multicoater K303 (RK Print Coat Instruments) (Figure 1). The samples’ codes and the
composition of the coatings are presented in Table 1.

8th Conference on Information and Graphic Arts Technology 174

 Poster section

Figure 1. Application of the barrier coating on the RK Multicoater K303.

Table 1. Sample codes and the composition of the coatings.

Sample code Composition of the coating
P1 Without barrier coating
P1-B1-c1 Dextrin, latex, barrier component B1 of concentration c1
P1-B1-c2 Dextrin, latex, barrier component B1 of concentration c2
P1-B1-c3 Dextrin, latex, barrier component B1 of concentration c3
P1-B2-c1 Dextrin, latex, barrier component B2 of concentration c1
P1-B2-c2 Dextrin, latex, barrier component B2 of concentration c2
P1-B2-c3 Dextrin, latex, barrier component B2 of concentration c3
P2 Without barrier coating
P2-B1-c1 Dextrin, latex, barrier component B1 of concentration c1
P2-B1-c2 Dextrin, latex, barrier component B1 of concentration c2
P2-B1-c3 Dextrin, latex, barrier component B1 of concentration c3
P2-B2-c1 Dextrin, latex, barrier component B2 of concentration c1
P2-B2-c2 Dextrin, latex, barrier component B2 of concentration c2
P2-B2-c3 Dextrin, latex, barrier component B2 of concentration c3

2.2 Characterisation methods

Viscosity of the barrier suspensions was determined by the dynamic rheological measurements.
Morphology of the sample surfaces and cross-sections was investigated by scanning electron microscopy
(SEM).
Grease resistance of the papers was determined by the TAPPI grease resistance KIT test T 559 cm-12 (KIT value
from 1 to 12).
Oil absorbency was determined by the Cobb-Unger method SCAN-P 37:77.
Water vapour transmission rate (WVTR) was determined according to the Standard ISO 2528 (gravimetric
dish method).
Coat weight of the coatings was determined as follows:

Coat weight= grammage1 - grammage2 (g/m2) (1)

where grammage1 belongs to the paper with the barrier coating and grammage2 to the paper without the
barrier coating.

8th Conference on Information and Graphic Arts Technology 175

 Poster section

3 RESULTS & DISCUSSION

3.1 Viscosity of the barrier suspension

The viscosity of the prepared dispersions was measured at 100 rpm, i.e. at a constant shear rate of 1.6667 s-1
and the results are presented in Figure 2. It can be seen that the viscosity of the barriers suspensions varied
from 6500 to 10000 mPa×s. The highest values were obtained for dispersions with the lowest concentration
(c1) of the barrier components irrespective to their composition. The increase of the barrier components con-
centrations decreased the suspensions viscosity.

Figure 2. Viscosity, h, of the barrier suspensions versus time, t,

at the constant share rate equals to 1.6667 s-1.

3.2 Surface morphology

SEM images of the surface and the cross section of the representative paper samples without and with barrier
coating are presented in Figures 3 and 4. Figure 3 shows that the barrier coating completely covered the surface
of the paper, closed the paper structure and decreased the surface roughness. This resulted in the increase of
the paper gloss. SEM images of the paper cross sections in Figure 4 revealed that there is no sharp boundary
between the surface of the paper and the barrier coating (continued 10 microns thick), which indicates their
compatible structural composition.

P1 P1-B1-c3
Figure 3. SEM images of the surface of the representative paper samples without and with barrier coating.

8th Conference on Information and Graphic Arts Technology 176

 Poster section

P1 P1-B1-c3
Figure 4. SEM images of the cross section of the representative paper samples without and with barrier coating.

3.3 Coat weight

The coat weight values are presented in Figure 5. The results show that the coat weight values of the created
coatings were between 5.5 g/m2 and 7.5 g/m2 in the case of the P1-samples and between 5.3 g/m2 and 8.4 g/m2
in the case of the P2-samples. It should be stressed that the barrier coatings did not increase the grammage of
the paper P1 by more than 9% which is very low. A comparison of Figures 2 and 5 also show that the highest
coat weight values were obtained by the use of the barrier suspension B2-c1 with the highest viscosity. How-
ever, in other samples, the viscosity of the barriers suspensions did not significantly influence the coat weight.

Figure 5. Coat weight of the created coatings.

3.4 Barrier properties

The barrier properties of the coatings are presented in Table 2. The results show that the paper samples without
the barrier coatings, P1 and P2, are not resistant to grease and oils because even Castrol oil penetrates into both
papers. The presence of all barrier coatings significantly increased the grease resistance of the papers resulting
in the increase of the KIT values to 11-12 or even 12 which is the highest value. The results also revealed that
the barrier component B2, irrespective to its concentration, provided excellent barrier for grease. In the case
of the coatings including the barrier component B1, the KIT values were slightly lower.
Furthermore, the presence of all barrier coatings importantly decreased the Cobb-Unger oil absorbency of
both papers. The oil absorbency was not significantly influenced by the structure of the barrier component,
but it was in general decreased by the increase of the component concentration.
Both paper samples P1 and P2 were porous and provided high transmission of the water vapour. The presence
of the barrier coatings decreased the water vapour transmission rate for more than 3-times in the case of P1
and more than 2.2-times for P2. Whereas both barrier components provided similar barrier for the water vapour
on P1, the barrier component B1 was more effective on P2 in comparison with B1.

8th Conference on Information and Graphic Arts Technology 177

 Poster section

Table 2. KIT value, oil absorbency, water vapour transmission rate (WVTR) for paper samples without and
with barrier coatings.
Sample code KIT value Cobb-Unger oil absorbency (g/m2) WVTR (g/(m2 24h))
P1 non rating 2.40 430
P1-B1-c1 12 0.33 140
P1-B1-c2 11-12 0.17 130
P1-B1-c3 11-12 0.20 130
P1-B2-c1 12 0.27 130
P1-B2-c2 12 0.14 130
P1-B2-c3 12 0.18 120
P2 non rating 1.44 440
P2-B1-c1 11-12 0.16 200
P2-B1-c2 11-12 0.19 120
P2-B1-c3 12 0.16 120
P2-B2-c1 12 0.18 180
P2-B2-c2 12 0.14 170
P2-B2-c3 12 0.17 180

4 CONCLUSIONS

In this research, new effective water-based dispersion barrier coatings for grease, mineral oil and water vapour
resistance were developed and successfully applied to the back side of two papers of different grammage, i.e.
paper P1 with 80 g/m2 and paper P2 of 55 g/m2. Dispersions with higher concentrations of the barrier compo-
nents displayed the best performance. The barrier coatings pass the TAPI T559 standard as well as the Cobb-Un-
ger oil absorbency test on both papers and significantly decreased the water vapour transmission through
the papers. However, to improve the water vapour barrier, the porosity of the coatings should be decreased.
This remains our major challenge. Considering the importance of the development and implementation of
the repulpable and more sustainable barrier coatings on the industrial level, these barrier coatings will be
optimised in order to meet the requirements regarding the runnability of the application process.

ACKNOWLEDGEMENTS
This work was supported by the EU Programme Potential of biomass for development of advanced materials
and bio-based products, acronym CEL.CYCLE, founded by the European regional development found and the
Slovenian Research Agency, Slovenia (I0-0026 Infrastructural Centre RIC UL NTF).

5 REFERENCES
• Lagaron, J.M., Cabedo, L. and Fabra, M.J. 2017. “Barrier Nanomaterials and Nanocomposites for Food Pack-
aging.” In Nanotechnology in Agriculture and Food Science, edited by Monique A. Axelos, Marcel H. Van de
Voorde, 167−175. John Wiley & Sons.
• Ovaska, S.-S., Hiltunen, S., Ernstsson, M., Schuster, E., Altskär, A. and Backfolk, K. 2016. “Characterization of
rapeseed oil/coconut oil mixtures and theirpenetration into hydroxypropylated-starch-based barrier coat-
ingscontaining an oleophilic mineral.” Progress in Organic Coatings 101: 569–576.
• Wang, S. and Jing, Y. 2017. “Effects of formation and penetration properties of biodegradable montmoril-
lonite/chitosan nanocomposite film on the barrier of package paper.” Applied Clay Science 138: 74−80.
• Weizman, O., Dotan, A., Nir Y., Ophir, A. 2017. “Modified whey protein coatings for improved gas barrier
properties of biodegradable films.” Polymers for Advanced Technologies 28(2): 261−270.

Corresponding author:
Barbara SIMONČIČ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 33, fax: +386 1 200 32 70, e-mail: [email protected];
homepage: https://www.ntf.uni-lj.si/toi/en/employee/barbara-simoncic-2/

8th Conference on Information and Graphic Arts Technology 178

 Poster section

Co-author(s):
Jelena VASILJEVIĆ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Barbara GOLJA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Ivan JERMAN
National Institute of Chemistry, Department of Materials Chemistry
Hajdrihova 19, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Ervin ŠEST
National Institute of Chemistry,
Department of Materials Chemistry
Hajdrihova 19, 1000 Ljubljana, Slovenia
e-mail: [email protected]

David RAVNJAK
Paper Mill Vevče d.o.o.
Dobrunje, 1261 Ljubljana, Slovenia
e-mail: [email protected]

Mateja KERT,
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Marija GORJANC
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Brigita TOMŠIČ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles Graphic Arts and Design
Aškerčeva 12, 1000 Ljubljana, Slovenia
e-mail: [email protected]

Tea KAPUN
Pulp and Paper Institute
Bogišićeva 8, 1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 179

 Poster section

ENZYMATIC DEINKING OF FLEXOGRAPHIC PRINTED PAPER

Mija SEŽUN1, Gašper DOKL2, Matej PIVAR3 & Gregor LAVRIČ1

1
Pulp and Paper Institute, Bogišićeva 8, 1000 Ljubljana
2
Biotechnical faculty, Jamnikarjeva 101, 1000 Ljubljana
3
Faculty of Natural Sciences and Engineering, Snežiška 5, 1000 Ljubljana

ABSTRACT: Flexographic printing ink removal is a current and urgent issue to be solved by CEPIPRINT and its
members. The reason that this topic came to their attention is the spread of flexographic for packaging print-
ing. Extended range of flexographic prints could pose a major threat to European recycling targets (INGEDE,
2008). In our study, we tested the performance of different types of enzymes in comparison with the chemicals
in deinking of flexo printed paper samples. The flexo samples were printed with water based and solvent based
flexo inks on 150g/m2 matte coated paper. The samples were determined by the thickness, grammage and their
optical properties (ISO whiteness, CIE Whiteness, with and without UV component).

Keywords: chemicals, deinking, enzymes, flexo ink, recycling paper.

1 INTRODUCTION

In the European paper industry recovered paper plays an important role in achieving goals. It is observed that
a large part of waste paper belongs to graphic papers, which are problematic in the deinking process. Deink-
ing is a complex process and it depends on many factors (e.g. quality of the paper, type of printing process,
properties of the printing ink, etc). Numerous studies include process optimization, but unfortunately, we have
not noticed an actual progress in the deinking process (CEPI, 2009). Today, huge quantities of chemicals are
used in the deinking process such as NaOH, Na2SiO3, Na2CO3, H2O2, chelating agent and surfactants (Pathak et
al., 2010). In any case, it is very important to reduce chemicals and replaced them with environmentally more
friendly substances, for example, enzymes. Bajpai (2014) reports that the use of enzymes has been suggested
as an environmentally friendly alternative to supplement conventional deinking chemicals. Enzymes could
reduce the demand for chemicals and would lower the process costs and the environmental impact. These en-
zymes include cellulases, pectinases, amylases, lipases, esterases and laccases. If the proportion of hydrophilic
inks is large, it represents a major issues for deinking process, because flexographic and other hydrophilic
inks are not able to be removed by flotation. This means that it is very important to find a suitable solution
for removing this type of ink (Josephson & Krishnagopalan, 2015). Such a challenge involves optimization of
pulping conditions, including use of different type of enzymes. But however, existing technologies have not
been able to fully reduce the negative impact of pigmented inks on deinking of recycled paper (Hsieh, 2012;
Kemppainen et al., 2011; Bhattacharyya e tal., 2009; Miller, 2009). Also, Lee et al. (2013) were not effective in
their research, because they have not been shown to remove hydrophilic inks such as inkjet and flexographic
inks to an acceptable degree.
During our study, we have shown that the efficiency of enzyme treatment is comparable to the efficiency of
chemical treatment. First we have determined the optical properties, which are considered by the point that
they are the most important, because of removing ink particles. In the following, more attention will be paid
to mechanical properties.

2 EXPERIMENTAL

In our study, we compared different deinking processes by using chemicals and specific enzymes (cellulase,
enzyme mixture). We investigated which treatment was more efficient by determining the optical properties
of the laboratory paper sheets after.

2.1 Samples
Samples were printed with water based and solvent based flexographic inks on 150 g/m2 matte coated paper
(Figure 1).

8th Conference on Information and Graphic Arts Technology 180

 Poster section

1a 1b

Figure 1. Flexographic prints – water (1a) and solvent based (1b) samples.

2.2 Deinking process

Samples were merged and cut to pieces of 2x2 cm. We also determined dry matter, according to the standard
EN 14346: 2007, and the loss on ignition according to the standard DIN EN 12879: 2001. Each sample weighed
75 g of absolutely dry matter.

The deinking process was carried out by slightly modified INGEDE 11 method.

Chemical deinking process

During the experiment we used the following chemicals: sodium hydroxide (NaOH), collector (Nopco fleet),
sodium silicate (Na₂SiO₃) and hydrogen peroxide (H₂O₂). All these chemicals were obtained from the Slovenian
paper mill Vipap Videm Krško d.d..
100 mL of H₂O₂ was added into the solution of chemicals (400 mL) and the mixture was diluted with water to
a volume of 1500 mL. 75 g of absolutely dry sample was added to the mixture of chemicals during constant
stirring at 3000 rpm in a thermostat spreader (T = 45°C, time of mixing 20 minutes). The temperature was
maintained by using a water bath. Once the dissolution was completed, the mixture was stabilized in a water
bath for 60 minutes at a constant temperature of 45°C. Based on the standard, the recommended pH value
was 9.48. Next step was flotation of the previously dissolved sample, which took place in a flotation cell in
the presence of water (18 L). Substance concentration during flotation was 0.42%. After flotation process lab-
oratory paper sheets were prepared.

Enzymatic deinking process

75 g of absolutely dry sample was added to 1200 mL of water during constant stirring at 3000 rpm in a ther-
mostat spreader (T = 45°C, time of mixing 5 minutes). The temperature was maintained by using a water bath.
Once dissolution was completed, we added 100 mL of enzyme dilution we prepared (99 mL of water was added
1 mL of enzyme, at temperature 55°C), for 20 min at 55°C and occasional stirring. At the end of the enzyme
treatment collector (Nopco fleet) was added to the mixture.

Analysis of optical properties

ISO brightness and CIE L*a*b* coordinates of laboratory paper sheets were measured using WinPaper Elrepho
450 X Datacolor spectrophotometer. CIE76 delta E formula was used to determine the quality of the deinking
process. Imaginary white surface with L*a*b* coordinates 0, 0, 100 was used as the reference in calculations.

3 RESULTS & DISCUSSION

3.1 Optical properties

Higher ISO brightness was achieved by enzyme deinking of water based flexographic print in comparison
with chemical deinking: + 0.51% with cellulase and + 2.61% with enzyme mixture treatment. ISO brightness of
solvent based flexographic print was also higher by enzyme deinking in comparison with chemical deinking:
+ 0.14% with cellulose and + 1.69% with enzyme mixture treatment.

8th Conference on Information and Graphic Arts Technology 181

 Poster section

Table 1. ISO brightness after deinking process.

ISO brightness (%)
Enzyme treatment Enzyme treatment
Sample Chemical treatment
(cellulase) (mixture)
Flexographic print - water 80.71 81.22 83.32
Flexographic print - solvent 83.04 83.18 84.73

Delta E value of flexographic print with solvent based is lower in comparison with water based, which cor-
responds to higher ISO brightness values of solvent based inks, as seen in Table 1. We assigned it as the con-
sequence of different inks used.
The lowest delta E value was achieved with mixture enzyme treatment, following cellulase treatment and
chemical treatment (Figure 4 and 5).

Figure 4, 5. Delta E after deinking process.

Based on the results we can conclude, that deinking with enzymes is at least same efficient as chemical deink-
ing in the case of flexographic ink removal. In a few years’ time, we can expect even more studies in this field
especially because of especially because of environmental legislation that is increasingly strict.

3.2 Reduction of pulping time

During the study we succeeded shorter processing time by enzymes. Duration time during enzyme treatment
(pulping) was 25 minutes and 80 minutes by chemicals.

4 CONCLUSIONS

In our study, we focused on enzyme treatment, with the aim to replace chemicals using during the deinking
process. We determined the optical properties, and on the basis of these we determined the efficiency of the
ink removal. The issue was also to reduce the pulping time, because this is very important in terms of process
costs. The processing time with enzymes was reducing by about 3 times. The fact that we have proved that
the chemicals can be at least partially replaced with enzymes is definitely a big contribution for papermaking
industry, since the impact on the environment considerably lower.

ACKNOWLEDGEMENTS
We would like to thank Vipap Videm Krško for cooperation and Igor Karlovits for help in the experimental
work (optical properties). This study was carried out within the RDI project Cel.Cycle: »Potential of biomass
for development of advanced materials and bio-based products« (contract number: OP20.00365), which is
co-financed by the Ministry of Education, Science and Sport of the Republic of Slovenia, and the European
Union, as part of the European Regional Development Fund 2016–2020.

8th Conference on Information and Graphic Arts Technology 182

 Poster section

5 REFERENCES

• Bajpai, P. 2014. “Recycling and Deinking of Recovered Paper.”, Pages 139–153

• CEPI. 2009. “Guide to an optimum recyclability of printed graphic paper.” URL:http://www.cepi.org/system/
files/public/documents/publications/recycling/2009/GuidetoandOptimumRecyclabilityofPrintedGraphic-
Paper.pdf (last accessed on 07.04.2018).
• Heise, O.U., Unwin, J.P., Klungness, J.H., Fineran, W.G., Sykes, M. and Abubakr, S. 1996. “Industrial scaleup of
enzyme-enhanced deinking of nonimpact printed toners.” Tappi Journal 79 (3): 207-212.
• Hsieh, J.S. 2012. “Deinking of inkjet digital nonimpact printing.” Tappi Journal 11 (9): 9-15.
• Jeffries, T.W., Klungness, J.H., Sykes, M.S. and Rutledgecropsey, K.R. 1994. “Comparison of enzyme-enhanced
with conventional deinking of xerographic and laser-printed paper.” Tappi Journal 77 (4): 173-179.
• Josephson, W.E. and Krishnagopalan, G.A. 2005. “Deinking of furnishes containing flexographically printed
old newsprint.” Appita Journal 58 (6): 470-474.
• Kemppainen, K., Korkko, M. and Niinimaki, J. 2011. “Fractional pulping of toner and pigment-based inkjet
ink printed papers - ink and dirt behaviour.” BioResourses 6 (3): 2977-2989.
• Lee, C.K., Ibrahim, D. and Omar, I.C. 2013. “Enzymatic deinking of various types of waste paper: Efficiency
and characteristics.” Process Biochemistry 48 (2): 299-305.
• Lee, D.T. 2014. “Mehanism and novel deinking methods for non-impact printed paper.” MS diss., Georgia
Institute of Technology.
• Miller, N. 2009. “Sustainable Digital Print Solutions: Deinkable Inks, Papers, and Optimizes Deinking Pro-
cesses.” presented at the PaperCon, St. Louis, Missouri.
• Ng, H.T., Bhattacharyya, M.K., Mittelstadt, L.S. and Hanson, E.G. 2009. “Deinking of HP digital commercial
prints: Effect of chemicals and their loadings on deinkability.” International Conference on Digital Printing
Technologies: 173-176.
• Nyman, K. and Hakala, T. 2011. “Decolorization of inkjet ink and deinking of inkjetprinted paper with lac-
case-mediator system.” BioResources 6 (2): 1336-1350.

Corresponding author:
Mija SEŽUN
Pulp and Paper Institute,
Laboratory for Materials and Technologies in Papermaking
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 18, e-mail: [email protected]

Co-author(s):
Gašper DOKL
Biotechnical faculty
Jamnikarjeva 101, 1000 Ljubljana, Slovenia
phone: +386 41 982 967, e-mail: [email protected]

Matej PIVAR
Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 44, e-mail: [email protected]

Gregor LAVRIČ
Pulp and Paper Institute, Print and Packaging department
Bogišićeva 8, 1000 Ljubljana, Slovenia
phone: +386 1 200 28 47, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 183

 Poster section

EVOLUTION OF NATURAL PATTERNS from RANDOM FIELDS

Lovrenc ŠVEGL1, Igor GRABEC2

Faculty of Natural Sciences and Engineering,
1

2
LASIN - Faculty of Mechanical Engineering, University of Ljubljana, Slovenia

ABSTRACT: In the article a transition from pattern evolution equation of reaction-diffusion type to a cellu-
lar automaton (CA) is described. The applicability of CA is demonstrated by generating patterns of complex
irregular structure on a hexagonal and quadratic lattice. With this aim a random initial field is transformed
by a sequence of CA actions into a new pattern. On the hexagonal lattice this pattern resembles a lizard skin.
The properties of CA are specified by the most simple majority rule that adapts selected cell state to the most
frequent state of cells in its surrounding. The method could be of interest for manufacturing of textiles as well
as for modeling of patterns on skin of various animals.

Keywords: random field transformation, cellular automata, labyrinthine pattern, lizard skin.

1 INTRODUCTION

Many technological problems require physical characterization of developing fields by a quantitative field
variable S(r,t) denoting measureable properties like material composition, surface roughness, color, etc. This
variable generally depends on the position r and time t. The development of the field can be physically de-
scribed by the evolution equation [1]:

(1)

Here O(r) indicates a properly selected surrounding of the point r and G(…) denotes a nonlinear field generator
function that generally includes differential and integral operators. Eq. 1 has already been utilized in the study
of various rather complex phenomena leading to formation of patterns in technical and natural environments
[1-7]. Among them the generation of patterns by various reaction-diffusion processes in chemical reactors,
plasma and biological environments are the most outstanding [2,7,8]. Fig.1 shows the field of ionization waves
in a turbulent plasma developed from random initial conditions as determined by the numerical solution of
a nonlinear integro-differential reaction-diffusion equation of type Eq. 1. Surprisingly, it resembles a charac-
teristic pattern of a muscular structure.

Figure 1. Distribution of turbulent ionization waves in a plasma of a glow discharge in a tube of argon.
The distribution was determined by solving a nonlinear integro-differential reaction-diffusion equation of
type Eq. 1 using random initial conditions and periodic boundary conditions [2]. The horizontal coordinate
corresponds to time and the vertical to the axial coordinate of the discharge tube.

For the numerical treatment a discrete joint variable s = (r,t) is usually introduced into evolution equation (1)
and then it is transformed into the mapping relation [2,4]:

8th Conference on Information and Graphic Arts Technology 184

 Poster section

(2)

This relation maps the old field distribution into the new one: S(so)®S(s). In such a case Eq. 2 represents a
general form of multi-dimensional cellular automaton (CA) [8]. In order to apply it, one has to specify the
generator function G(…), a discrete lattice of cells, initial and boundary conditions, as well as to solve Eq.
2 numerically. The generator function G(...) can be specified either theoretically, based upon physical laws
governing the treated phenomenon [8], or empirically from experimental records of the field S(s) [4,6]. In the
last case a neural network can be applied to express the generator function in terms of recorded data [1,3,4,5].
Eq. 2 is convenient for modeling of patterns in graphic art and technology as well as in biology [1-7].

Figure 2. A characteristic pattern on a lizard skin.

2 SPECIFICATION OF CELLULAR AUTOMATON

The goal of the present article is to demonstrate applicability of Eq. 2 by modeling patterns of complex irreg-
ular form. For this purpose we first consider transformation of an initially random field by a two-dimensional
cellular automaton into a new pattern resembling a lizard skin. This example is selected since it could be of
interest for manufacturing of textiles as well as for modeling of patterns on skin of various animals [6]. For
the specification of the CA structure we examine the pattern on a sample of lizard skin shown in Fig. 2. The
pattern is comprised of scales representing approximately hexagonal cells, and consequently, we first accept a
hexagonal lattice as the basis for the operation of our two-dimensional CA. In this case each cell is surrounded
by six neighbor cells as shown in Fig.3.

Figure 3. Two characteristic examples of center cell adaptation to its surrounding by the majority rule,
left: without change, right: with a change of state.

3 EXAMPLES

In the first example we apply the hexagonal lattice. The initial random field distribution is shown in Fig. 4 A
while the distribution after 1st , 2nd , and 10th action are shown in Figs. 4 B, C, D respectively. In the first CA
action many cells change the color, but the number of changes ∆N in a single action is decreasing with the
number A of CA actions and ceases after several steps. This property is demonstrated in Fig. 5. It shows the total
number of state changes N in dependence of the number A of CA actions. It is evident that our rule removes
single jumps in the field distribution and yields more smooth variation in the final pattern. In spite of this

8th Conference on Information and Graphic Arts Technology 185

 Poster section

smoothing effect the irregular character of the labyrinthine pattern is still preserved. This property does not
depend on particular properties of the initial random field sample and resembles the effect of a digital filtering
or changing of patterns by convolutional neural networks. Comparison of the final distribution in Fig. 4 D
with the distribution on a real lizard skin shown in Fig. 2 reveals surprising similarity of characteristic features.
Similar properties as on the hexagonal lattice exhibits CA also on a quadratic lattice. Figs. 6 and 7 show results
corresponding to Figs. 4 and 5. In this case the lines of the final pattern appear more cornered and the number
of changes caused by CA actions is higher as in the hexagonal case. At some places a cyclic changing of state
caused by successive CA actions can take place that leads to permanent increasing of N with A.

A B

C D

Figure 4. Distribution of the initial random field – A and distributions after 1st - B,
2nd - C,and 10th – D action of CA on the hexagonal lattice.

8th Conference on Information and Graphic Arts Technology 186

 Poster section

Figure 5. Number of changes N caused by A actions of CA on the hexagonal lattice.

A B

C D
Figure 6. Distribution of the initial random field – A and distributions after 1 - B,
st

2nd - C, and 10th – D action of CA on the quadratic lattice.

8th Conference on Information and Graphic Arts Technology 187

 Poster section

Figure 7. Number of changes N caused by A actions of CA on the quadratic lattice.

4 CONCLUSIONS

Our first example indicates applicability of the hexagonal cellular automaton for modeling of irregular com-
plex patterns resembling the lizard skin. The second example reveals that the structure of CA lattice influences
the roughness of the lines in the generated pattern. Since we consider just the nearest cells on the hexagonal
or quadratic lattice, the lines of the generated patterns are rather cornered. By taking into account also more
distant cells this weakness can be improved. At the specification of the generating function G(…) we here
apply just the most simple majority rule. However, by changing this function, as well as the basic lattice, the
properties of patterns can be well adapted to given samples [1,4,5]. For this purpose the function G(…) can
be determined from given field records as has been described elsewhere [4,6].

5 REFERENCES

1. GRABEC, I. & MANDELJ, S.: Experimental modeling of chaotic fields, Progr. Theoret. Phys., (2003) 150: 81-88.
2. GRABEC, I.: Nonlinear properties of high amplitude ionization waves, Phys. of Fluids, (1974). 17: 1834-1840.
3. GRABEC, I. & SACHSE, W.: Synergetics of Measurement, Prediction and Control, Springer, Berlin (1997).
4. BORŠTNIK-BRAČIĆ, A., GRABEC, I. & GOVEKAR, E.: Modeling spatio-temporal field evolution, Eur. Phys. J.,
(2009) B 69: 529–538.
5. MANDELJ, S., GRABEC, I. & GOVEKAR, E.: Nonparametric statistical modeling of spatio-temporal dynamics
based on recorded data, Int. J. Bifurcation & Chaos, (2004) 14(6): 2011-2025.
6. MANUKYAN, L., MONTANDON, S. A., FOFONJKA, A., SMIRNOV, S. & MILINKOVITCH, M. C.: A living meso-
scopic cellular automaton made of skin scales, Nature, (2017) 544: 173-179.
7. TURING, A. M.: The Chemical Basis of Morphogenesis, Phil. Trans. Royal Soc., (1952) B237: 37–72.
8. WOLFRAM, S.: A New Kind of Science, Wolfram Media Inc. (2002).

Corresponding author:
Lovrenc ŠVEGL
Faculty of Natural Sciences and Engineering
Home: Poklukarjeva 66
SI-1000, Ljubljana, Slovenia
phone: +386 70 691 383, e-mail: [email protected]

Co-author(s):
Igor GRABEC
Faculty of Mechanical Engineering
Aškerčeva cesta 6
1000 Ljubljana, Slovenia
phone: +386 41 678 182, email: [email protected]

8th Conference on Information and Graphic Arts Technology 188

 Poster section

EYE TRACKING ANALYSIS OF INTERNAL FACIAL FEATURES

IN OBSERVATION AND RECOGNITION PROCESS

Andrej ISKRA, Helena GABRIJELČIČ TOMC

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Chair of Information and Graphic Arts Technology, Snežniška 5 Ljubljana

Abstract: The aim of our research was to examine how people observe facial images for different observation
times. A single test consisted of two parts: observation and recognition. Observation tests were divided accord-
ing to the different observation times (1, 2, 4 and 8 seconds). First analysis of our testing was focused on the
recognition test where we measured correct and incorrect recognition. In second investigation we measured
observation times of the internal facial features (eyes, nose, mouth).
Our main goal was to find possible correlation between observation times of main three facial features and
recognition results. We found out that 4 second is a turning point where participants looked all face (all facial
features) and their gaze returned back to the area of internal facial features. Correct and incorrect recognition
has also turning point at 4 seconds test. That was also confirmed by gazeplot obtained from eye tracker.

Keywords: eye-tracking, face recognition, face image, facial features, region of interest.

1 INTRODUCTION

The use of facial images is currently very widespread, and features in many different aspects of our lives, e.g.
web presentations, criminalistics, security, psychology, neuroscience, advertising, marketing, etc. It is also
important that the people working for various institutions, associations and companies are represented ac-
curately by their facial image online. How we look at and remember faces had been part of a many research
studies for several decades.

During the face observation and recognition process, the most important parameter is the length of time that
the face is presented during the observation phase (Reynolds, 1992, Leyk, 2008; Iskra, 2016)

Currently, there are two main directions that have been taken in facial image research. One group focuses on
emotional facial expression (Malcom, 2008; Cangöz, 2013). Most researchers define six main face expressions
(beside neutral) which are happy, sad, angry, disgust, surprise and fear. Some researchers add also annoyed
and grumpy. The second area in facial image research that has been extensively explored is the different angles
with which a face can be presented (Brooksô, 2007, Brielmann, 2014). Both areas are connected by exploring
how people see different facial features (Want, 2003; Henderson, 2005). Our goal was to find any possible
connection between face recognition success and procedure of face observing by investigating portion of
observation for certain face features. Due to the extensive number of parameters we limited our experiment
by use of only frontal face images with neutral face expression.

2 EXPERIEMENTAL

As mentioned above, we deal with several parameters that can influence the way participants observe and
recognize faces. In our investigation, we focused only on frontal facial images with a neutral facial expression.
We used the faces of Caucasian race and a similar age to our participants and considered them to be typical
faces. We also attempted to implement the real circumstances we are faced with in everyday situations.

2.1 Participants

The test participants were our students, all of whom were aged between 19 and 23 years old with normal vision.
The participants were divided into four groups according to four different observation times. There were 41
participants. Since a Tobii eye tracking detection rate of 90% was required (Tobii Studio user manual, 2012),
we performed our tests in such a way to ensure that there was an equal number of participants in each test
group that passed that criteria. Eigth participants were allocated to each group, which meant that the results
for 32 participants were analysed in total.

8th Conference on Information and Graphic Arts Technology 189

 Poster section

2.2 Stimuli
The facial images were obtained from the Minear and Park facial database (Minear, 2004), which was created
by photographing volunteers in controlled conditions. The two main categories for organizing this database
were age and emotion expression and include all four main races (Caucasian, Black, Asian and Indian). We took
the facial images of 20 male and 20 female Caucasian people aged 18 to 29 years old. This was done in order to
have similar test participants and stimuli, and to avoid the influence of others parameters (race, age, emotion)
on recognition performance. It was proven that people have different learning processes and recognition per-
formance for faces of different races (Goldinger, 2009). All images were prepared in dimensions 1280 × 960 px.

2.3 Apparatus

All of the tests were performed using the Tobii X-120 eye tracker. The monitor was 24’’ with a resolution of
1440 × 900. Although a higher resolution could be set on the monitor, we were forced to set lower resolution
due to the restriction of Tobii studio 3.4.4 software which was used to collect and analyse the testing data.
The distance between the participants and the monitor was around 60 cm (Goldinger, 2009; Cangöz, 2013).

2.4 Procedure

We designed four tests according to the different observation times of facial images presented. All of the facial
images in the tests were the same. Each participant performed one of the four tests. As mentioned above we
investigated correlation between observation time, recognition success and portion of observation time for
main three face elements. Observation times were set for 1 second, 2 seconds, 4 seconds and 8 seconds. The test
for the participants was divided into two parts. The first part was the observation test and the second part was
the recognition test. Calibration of the eye tracking system was carried out for each participant at five control
points and was performed at the beginning of the test. The observation part consisted of 20 images (10 male
and 10 female images). This testing procedure was automatic and is shown in Figure 1.

Figure 1. Procedure of the observation test.

After the initial instructions participants clicked mouse button for the first facial image. After the observation
time of facial image (depending on the test), there was a 2 second pause with dark screen. The purpose was to
neutralize the position of the eyes (Cangöz, 2013). In that case last fixation of the previous facial image had
no influence on the first fixation on the next facial image.

The second part of our testing was focused on the recognition test, which was controlled by the participants
themselves. The test comprised 40 facial images, 20 from observation test and 20 new faces from the same group
as before (Caucasian race, neutral expression, aged 18–29). These 20 new faces were also equally divided into
10 male and 10 female facial images. After the instructions were provided, the first facial image was displayed.
The participants were required to answer YES or NO if they saw that face image in the observation test. The
answers were then recorded manually. After providing each answer, the participants clicked the mouse and the
next image appeared. The order of the images presented was previously prepared (not random), but was the
same for all participants. This made analyses of the results much easier to carry out than if the facial images
presented to every participant would be completely random. Since there was no time limit, each participant
had as much time as they needed to think about each image, but usually answers were said relatively quickly
after the appearance of the face image. The recognition test procedure is presented in Figure 2.

Figure 2. Recognition test procedure.

8th Conference on Information and Graphic Arts Technology 190

 Poster section

2.5 Analysis 1

First analysis was focused on recognition success. We measured the effectiveness of recognition in relation to
the observation time of the facial image. We defined two terms:
• CR (correct recognition). The facial image was in the observation test and the participants confirmed this
by answering YES.
• IR-FA (incorrect recognition – false alarm). The facial image was not included in the observation test, but
the participant stated that he/she saw it.

2.6 Analysis 2

The focus of this analysis was on portion of observation time for three main facial features (eyes, nose and
mouth). For this purpose, we drew an AOI for each face image manually. An example of an AOI is shown in
Figure 3.

Figure 3. Setting of AOI on facial image.

Defining AOI meant that we got number of fixation, average fixation duration and total observation time for all
AOI’s for all facial images in the observation tests. We calculated average observation time for all participant
together for each facial feature for different time tests (1s, 2s, 4s and 8s). The final results were presented as a
percentage of the total observation time for every facial feature.

3 RESULTS & DISCUSSION

3.1 Analysis 1

Recognition test gave us results for correct recognition for all four observation times and can be seen in Figure 4.

Figure 4. Recognition success (%) and incorrect recognition (%) in dependence of observation time.

8th Conference on Information and Graphic Arts Technology 191

 Poster section

Figure 4 show the predicted relationship between observation time and recognition success (correct recogni-
tion and incorrect recognition). Longer presentation of facial images meant better recognition performance.
The curve is steep for a short observation time, which means that recognition success improves significantly as
the observation time increases until the 4 second presentation time, which is when the curve starts to flatten
out. Recognition success level got is some kind of saturation. We also analysed incorrect recognition, which is
commonly also referred to as False Alarm. In this case, the participant responded that that face was presented
in the observation test, despite this not being the case (the participant made a mistake). As we expected, incor-
rect recognition gave us opposite curve shape from the correct recognition. Longer observation time for facial
images meant less number of incorrect recognition. Here the turning point is again at 4 seconds observation
time, because at this point the level of incorrect recognition dropped significantly from the level at 2 seconds
test. On the other hand, 8 seconds test did not improve incorrect recognition much. Also 1 second test had
the same incorrect recognition level as 2 second test. This showed that 4 second face presentation in sufficient
enough that incorrect recognition level is satisfactorily small. Our results of recognition success confirmed
previous researches (Henderson, 2005; Iskra, 2016).

Results of second analysis gave portion of time spent on main three facial features (eyes, mouth, nose) in
dependence of observed time. Tobii Studio gave us the total time spent on these AOI and we calculated the
relevant portions for all the facial features. The results are presented in Figure 5.

Figure 5. Portion of observation time for the main three facial features and the cumulative portion.

We attempted to discover the pattern of how participants looks at facial images when they are presented at
different observation times. For observation time 1 second the percentage of observation of three main face
elements were quite high. All three main facial features took over ¾ of the total observation time (77.9%). In
these short observation time, the participants’ gaze mostly remained directed towards the centre of the face.
These three main facial features are also referred to as internal or central facial features (Henderson, 2005;
Hills, 2014).

This changed at observation time 2 seconds. Here participants had enough time to start observing also other
facial feature (chin, cheeks, forehead and ears), so portion of observation for main three facial features dropped
(eyes from 42.5% to 36.4%, mouth from 18,9% to 16,9%, nose from 16,6% to 12,4%). These trends continued also
for 4 seconds observation time (eyes 37,9%, mouth 9%, nose 7,4%) where observation percentage for mouth
and nose dropped hugely. Absolute time for observing mouth and nose were almost the same for 2 seconds
and 4 seconds tests (mouth 0,34 s and 0,36 s; nose 0,25 s and 0,3 s). Percentage of eye observing already slightly
increased (participant’s gaze returned back to eyes). For observation time 8 seconds these “return” effect was
even more obvious. The percentage of observing main three facial features (internal) again increased signifi-
cantly. In total percentage from 54,2% to 72,7%. In terms of absolute observation time, this meant an increase
from 2,17 seconds to 5,81 seconds (Figure 5). Therefore, from additional 4 seconds of observation facial image
(difference from 4 seconds and 8 seconds tests), 3.64 seconds was spent on the three central facial features. In
absolute time, a huge increase was shown in the mouth (from 0,36 s to 1,52 s) and nose (from 0,3 s to 1,09 s)
area. All of these results and the scan path of the participants’ gazes revealed that sequence followed when
looking at the face was the eye → mouth → nose → external facial features → internal facial features. A time
period of approximately 4 seconds was sufficient to observe the whole face, before reverting back to the in-
ternal facial features. This pattern can be seen in Figure 6.

8th Conference on Information and Graphic Arts Technology 192

 Poster section

Figure 6. Gazeplots for different observation times (1 second, 2 seconds, 4 seconds and 8 seconds).

First face image shows gaze mostly at internal facial features, then participants start looking also other facial
features (2 seconds test), at 4 seconds they start returning to the eyes area (mouth and nose portion are at the
lowest value), at 8 seconds observation time internal facial features again had most of the fixations.

4 CONCLUSION

Results of both analyses showed us that in process of observing face image had similar pattern for most
participants. First they looked at eyes, then moved their gaze to other two internal facial elements (mouth
and nose), after that we observe other less important facial elements (forehead, chin, cheek). All this pro-
cedure happened in app. 4 seconds, when participants returned to the internal face elements. This 4 seconds
observation time was also breaking point for memorizing faces, which was shown at recognition tests. Cor-
rect recognition reached sufficient level at 4 seconds test and longer observation time did not significantly
improve recognition success. That breaking point was even more obvious at incorrect recognition, when level
of wrong recognition dropped significantly at 4 seconds test. We can conclude that it took 4 seconds for the
participants to see the whole face, and that this time was sufficient for all the facial features to be seen. For
longer observation times, the participants just went back to observing the internal facial features, which did
not improve their recognition performance.

5 REFERENCES

• Brielmann, A. A., Bülthoff, I. and Armann, R. 2014. “Looking at faces from different angles: Europeans fixate
different features in Asian and Caucasian faces.” Vision Research, 100: 105-112.
• Brooksô, K. R. and Kemp, R. I. 2007. “Sensitivity to feature displacement in familiar and unfamiliar faces:
Beyond the internal/external feature distinction.” Perception 36: 1646–1659.
• Cangöz, B., Altun, A., Aşkar, P., Baran, Z. and Mazman, S. G. 2013. “Examining the visual screening patterns
of emotional facial expressions with gender, age and lateralization.” Journal of Eye Movement Research,
6(4): 1-15.
• Goldinger, S. D., He, Y. and Papesh, M. H. 2009. “Deficits in cross-race face learning: Insights from eye move-
ments and pupillometry.” Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(5):
1105-1122.
• Hills, C., Romano, K., Davies-Thompson and J. Barton, J. J. S. 2014. “An adaptation study of internal and ex-
ternal features in facial representations.” Vision research 100: 18-28.
• Henderson, J. M., Williams, C. C. and Falk, R. J. 2005. “Eye movements are functional during face learning.”
Memory & Cognition 33(1): 98-106.
• Leyk, D., Sievert, A., Heiss, A., Gorges, W., Ridder, D., Alexander, T., Wunderlich, M. and Rüther, T. 2008. “Val-
idation of a short-term memory test for the recognition of people and faces.” Ergonomics, 51 (8): 1125-1136.
• Iskra, A. and Gabrijelčič Tomc, H. 2016. “Eye-tracking analysis of face observing and face recognition.” Journal
of Graphic Engineering and Design 7(1): 5-11.
• Malcom, G. L., Lanloy, L. J., Fugard, A. J. B. and Barton, J. J. S. 2008. “Scan patterns during the processing of
facial expression versus identity: An exploration of task-driven and stimulus-driven effects.” Journal of
Vision, 8(2): 1-9.
• Minear, M. and Park, D.C. 2004. “A lifespan database of adult facial stimuli.” Behavior Research Methods,
Instruments, & Computers. 36(4): 630-633.

8th Conference on Information and Graphic Arts Technology 193

 Poster section

• Reynolds J. K. and Pezdek K. 1992. “Face Recognition Memory: The Effects of Exposure Duration and Encoding
Instruction.” Applied Cognitive Psychology, 6(4), 279-292.
• Want, S. C., Pascalis, O., Coleman, M. and Blades, M. 2003. “Recognizing people from the inner or outer
parts of their faces: Developmental data concerning ‘unfamiliar’ faces.” British Journal of Developmental
Psychology 21: 125–135
• (2012) User Manual Tobii Studio version 3.2. Tobii Technology Danderyd, Sweden.

Corresponding author:
Andrej ISKRA
University of Ljubljana, Faculty of Natural Sciences and Engineering
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 22, fax: +386 1 200 32 70, e-mail: [email protected]

Co-author(s):
Helena GABRIJELČIČ TOMC
University of Ljubljana, Faculty of Natural Sciences and Engineering
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 91, fax: +386 1 200 32 70, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 194

 Poster section

HOW COLOR SHAPES BRAND IDENTITY

Urška STANKOVIČ ELESINI1, Karla MRKOBRADA1 & Jani TOROŠ2

University of Ljubljana, Faculty of Natural Sciences and Engineering,
1

Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia
2
Innovative Management Institute, Dunajska 125, Ljubljana, Slovenia

Abstract: The objective of this research was to analyse logotypes of the banks on the Slovenian market by
colours and to compare colours with their values and principles and with the colour scheme made by the
Towergate Insurance. The banks were divided into two groups by origin into the well-known old banks and
the newcomers. As will be shown in the research, the most preferable colour in the logotypes of the Slovenian
banks is blue, however the logotypes of the newcomers are more colourful, composed of green, red, black
and blue. The colours are in most banks related to bank values and principles, although some additional col-
ours like yellow and especially green are suggested to be used in the logotypes the well-known old banks. In
research an additional quiz was performed related to the recognizability of the bank logotypes according to
the colour. The results were impressive, since the logotypes of some banks scored almost 80% recognisability
according to their logotype colours.

Keywords: logotype, colour, colour scheme, bank, values.

1 INTRODUCTION

A logotype is not a brand, but rather a form of a brand’s expression; it has a significant impact because it
represents a promise. A logotype must be something completely new, but not overly classical, so it might
remain popular for a longer period (Wiedemann, 2005). Macnab (2008) suggested that a logotype should be
natural and harmonious, comprising simple numbers and forms, and more focused on quality than on quantity.
The symbols used must be meaningful and recognisable, regardless of the spoken language. The elements of
a logotype can be pictorial, typographic or a combination of both. However, thought must be given to the
selection of all elements, and the appropriate colours must also be chosen (Walch, 1995), as they may evoke
both positive and negative feelings in people (Trstenjak, 1996). Chen and Chen (2007) claimed that, besides the
symbols and text used, colour is also a function of communication, identification and symbolism, and that its
meaning is often forgotten. The right colour or colour combination of a logotype appeals to the appropriate
target audience.

The interesting observation about suitable colour for the business/institution was introduced by Jacobs et al.
(1991) in the research about cross-cultural meanings of colours. They found that some colours had universal
meaning while others did not. For example, the results showed that in China and Japan colour grey was associ-
ated with word inexpensive, while in the USA this colour was associated with expensive as well as high quality
and dependable. On the other hand, colour blue connoted high quality in all tested societies (China, Japan,
USA and South Korea) and had meaning of sincere and trustworthy in South Korea. All cultures identify colour
red with love and adventure, and with good-tasting (except for China where colour red is associated also with
good luck). Colour purple was associated with expensive in the Asian culture, but inexpensive in the United
States. Colour black was associated with powerful and expensive in all four countries and as dependable in
both China and Japan. The Chinese associated black also with being trustworthy and high quality. According
to the research, colour brown was the most often associated with restaurants in USA, China, Korea and Japan;
colour grey was associated with hospitals, governmental buildings, factories, and museums, while school were
most often associated with colour green. An interesting conclusion was that companies with a monopoly or a
superior product may not need to be concerned with colour. As long as product has little direct competition,
colour may not be significant. And vice versa, when company introduces product into highly competitive for-
eign market it should investigate the meaning and significance of the selected colours in the target market.

In the research of Grimes and Doole (1998), remarkable similarities were noticed among UK and Taiwanese
cultures in terms of the emotions associated with colours. Also in this research it was concluded, that colour
associations are likely to be similar across cultures and that the standardised use of colour can and should be
a very powerful tool in creating international brand identity and awareness. They also emphasized that colour
is a tool of evolution, not revolution.

8th Conference on Information and Graphic Arts Technology 195

 Poster section

In the research performed by Madden et al. (2000), it was revealed that although respondents from eight coun-
tries (Austria, Brazil, Canada, Colombia, Hong Kong, China, Taiwan and the USA) shared similar associations
with colours, they have different tendencies about matching colours in a logo design. Thus, choosing colours(s)
for logotypes should be done with understanding of how colours and their combinations are perceived in
each culture. However, some cultural differences must be considered and thus for example, the logotype of the
company McDonalds in Israel was redesigned from the traditional yellow arches and red background into blue
and white, with the word “kosher” alongside for their restaurants where the kosher food is served. (Chang,
2010; McDonalds changes logo in Israel to show that its food is kosher, 2006)

Besides cultural differences, age and gender of target audience must also be held in mind when choosing
logotype colour(s). In the research of Dittmar (2001) it was revealed the change of colour preferences during
the life span i.e. the preference for colour blue decreased steadily from age 19 to age 90 years, while the popu-
larity of green and red increased with ages in both gender. As it was noticed by De Bortoli and Maroto (2001),
younger children were more attracted by strong, warm, and intense colours, while over the years those colours
became less violent with softer shades.

A lot of studies have been focused on the selection of colours for certain business (Towergate Insurance, 2017;
A Guide to Choosing Colors for Your Brand, 2017; Color Emotion Guide, 2013; MacDonald, 2018). From those
studies can be seen that certain colour is not tied to any particular business. For example, colour red in the log-
otypes of fast-food companies (i.e. McDonalds, Burger King, KFC, DQ etc.) catches the attention of customers,
but is not pleasing to look at for an extended period, meaning customers soon leave the restaurant and leave
space for others. (Chiazzari, 2000) On the other hand, colour red also denotes feelings of power, energy, pas-
sion, love, motivation, aggression or danger and thus it can also be found in logotypes of different beverages
(e.g. Coca-Cola, Pepsi, Red Bull), automotive and racing industries (e.g. Toyota, Mitsubishi, Honda, Formula
1), broadcasting (e.g. CNN, BBC News), entertainment/music groups (e.g. Rolling Stones, AC/DC, Epica, Iron
Maiden), footwear (e.g. Puma, Diesel) etc. Depends on emotions and feelings which those logotypes should
evoke, they include smaller or higher percentage of different shades of colour red.

Although logotypes mostly follow some “standardise” colour schemes, some deviations from those norms can
be found on the market. In this article, the emphasis is given to the logotypes of the banks on the Slovenian
market, thus some information on topic from different sources are given in the continuation. MacDonald
(2018) has noticed, that “banks have tried to differentiate themselves by choosing a simple memorable colour
scheme.” From the Towergate Insurance survey (2017) can be seen that banking predominately uses blue to
promote smart, trust and dependability and also red to promote confidence and ambition (Fig. 1). In smaller
shares, banking also uses colour black in their logotypes denoted maturity and balance, colour green which
is associated with wealth, initiative and growth, colour orange denoted with affordability and confidence and
colour yellow associated with happiness and optimism. De Bortoli and Maroto (2001) explained that colour
orange is sometimes associated with cheapness, colour grey with strength, exclusivity, and success and blue
with wealth, trust and security. These associations explain why banks are more likely to use blue and grey for
their logotypes rather than orange. Although “classic” and almost conservative colour schemes are usually
used for bank logotypes, colours must be adjusted to the market, target audience, values and principles of
companies, etc.

1. Restaurants 12. Department Store

2. Banking 13. Electronic Utilities
3. Beverage 14. Food Processing
4. Airlines 15. Heavy Equipment
5. Comms Equipment 16. Home Improvement
6. Computer Services 17. Hotels & Motels
7. Consumer Electronics 18. OIl & Gas Food Retail
8. Apparel & Accessories 19. Pharmaceuticals
9. Auto & Truck Manufacturers 20. Publishing
10. Broadcasting & Cable 21. Specialised Chemicals
11. Consumer Finance 22. Comms Services
23. Courier

Figure 1. Colour scheme of certain industries made by Towergate Insurance (2017).

8th Conference on Information and Graphic Arts Technology 196

 Poster section

The aim of our research was a) to analyse colours of the bank logotypes which can be found on the Slovenian
market, b) to compare gained colour scheme with the scheme made by the Towergate Insurance (2017), c) to
determine if colours correspond to values and principles of each bank and d) to investigate, if banks could be
recognised through the logotype colours.

2 EXPERIMENTAL

On the Slovenian market, 12 banks were chosen and divided into two groups by origin i.e. six well-known old
Slovenian banks (some with more than 100 years tradition) and six newcomers. For each bank, colours of their
logotypes were analysed in whole and according to symbols and typography in program application Adobe
Photoshop CC 2018. Logotype colours were taken from their corporate identity (CI) guidelines or in the case
when CI was not available, colours were taken from logotypes found on their internet pages. After colours
were analysed, results were compared with colour scheme made by the Towergate Insurance (2017). The values
and principles of each bank were identified from their web pages and compared with their logotype colours.

During the research, we were playing with an idea, how to determine the recognisability of banks according
to their logotype colours. Thus, we performed an online quiz in which respondents had to determine which
colour combination is related to the specific bank. The colour combinations of bank logotypes were design in
Adobe Illustrator CC 2018 and are shown in Fig. 2. The quiz was completed by 156 respondents (75.5% female,
24.5% male) mostly students (80.7% were younger than 30 years and 19,3% were older).

Figure 2. Colour combinations of twelve analysed logotypes.

3 RESULTS WITH DISCUSSION

3.1 Comparison of logotype colours by the Towergate Insurance colour scheme

The colour analysis of 12 different logotypes revealed some interesting results. The logotypes were analysed
as a whole (Fig. 3) and separately by their symbols and typography.

Figure 3. Colours of logotypes of well-known old banks (left) and newcomers (right).

From Fig. 3 can be seen that dominant colour in logotypes of well-known old banks was blue (55.5%), following
by colour red (19.1%) and colour gray (10,6%). In smaller percentage also colours orange, white and yellow
were presented in those logotypes. In contrary, logotypes of newcomers on the Slovenian market were more
colourful by dominant colour green (31.2%), red (27.9%), black (22.2%) and blue (13.6%). In smaller percentage
colours white, orange and yellow involved in logotypes. Comparing those results with the colour scheme of
the Towergate Insurance company in which dominant colours for banking are blue, red and black, our research

8th Conference on Information and Graphic Arts Technology 197

 Poster section

gave slightly different results for both type of banks. In six well-known old bank logotypes blue and red are
dominant colours, however colour black was not detected in any logotype. In logotypes of newcomers, colour
green was incorporated in higher percentage in two logotypes, although this colour was presented in colour
scheme of the Towergate Insurance in smaller extend. Colours blue, red and black were also presented in
logotypes of newcomers.

In continuation of the research, colours of symbols and typography in logotypes were also analysed. From
Fig. 4 can be seen that colours in symbols of well-known old banks are slightly more colourful with domi-
nant red blue and white colour, while colours blue, white and gray were used for typography. The symbols of
newcomer’s logotypes were mostly in red and green, while other colours were represented in smaller extend.
Typography of those logotypes were in colours black, white, green and blue.

Figure 4. Colours of logotype symbols of well-known old banks (left) and newcomers (right).

3.2 COMPARISON OF LOGOTYPE COLOURS BY THE BANK VALUES AND PRINCIPLES

The logotype colours were also analysed by the values and principles, which were listed on the internet pages
of each analysed bank. The most common values were responsibility, respect, confidence and trust, honesty,
discretion, excellency, integrity, loyalty and effectiveness. According to the dominant colour blue in four log-
otypes of well-known old banks, the communicating values through the logotypes were trust and security
which were also promoted through their internet pages. On the other hand, blue is also the colour of assurance
and composure, dependability, smartness and stability in performance (Brown, 2016), which could be reflected
through their logotypes. Colour blue in logotypes of well-known old banks was usually in combination with
gray (as just small part of the symbol or rather typography). Two of six Slovenian well-known old banks had
colours red and orange in their logotypes. Colour red and orange both promote confidence, while red also
promotes ambition, and orange promotes affordability, innovation and modernity (Brown, 2017) which are
not explicitly stated as values of any analysed bank. Innovation and modernity are values which could not be
forgot nowadays. Both colours, especially red, were used in five logotypes of newcomers. Colour red promotes
power of banks, maturity (Brown, 2016a), passion and balance (Brown, 2016), especially in combination with
black. In the case of colour orange, De Bortoli and Maroto (2001) emphasised that this colour could promotes
cheapness, thus it could be used with certain degree of caution. Colour green was presented in logotypes of
two newcomers but not in our well-known old banks. This colour is associated with initiative, wealth and
growth (Brown, 2016), ethical and freshness and beside that it is associated with Slovenia by the most Slove-
nian citizens. Since colours of bank logotypes are often associated with their national colours (Brown, 2017),
green should be more often chosen by Slovenian banks. Among all banks, only two (one well-known and one
newcomer) had colour yellow in their logotypes. Although this colour represents optimism and friendliness,
their share in logotypes were rather small.

3.3 Recognisability of banks according to their logotype colour

Results of the quiz by which recognition of banks was analysed according to their logotype colours are shown
in Fig. 5. Although we expected, that logotypes just by the colours will be very difficult to determine, it showed
up, that some of the logotypes were quite recognizable according to their colours. The most recognizable were

8th Conference on Information and Graphic Arts Technology 198

 Poster section

well-known old Slovenian banks, especially those with longer tradition. Among newcomers the most recog-
nizable were those who are more advertised in different media (especially TV) and those who were primarily
Slovenian but lately bought by foreign banks.

Figure 5. Percentage of respondents who recognized bank according to the colour of logotype.

4 CONCLUSIONS

Colours in logotypes of well-known old banks and newcomers follow the colour scheme made by the Tower-
gate Insurance. Bathi (2017) stated: “Every colour counts (as long as it’s blue)« and colour blue is thus “the
safest” colour for the bank (Brown, 2016a). The colour blue is the most common colour also in the logotypes
of banks in Slovenia. The second most preferable colour was red, however green and black were swapped. The
difference in logotype colours among well-known old banks and newcomers exist, regardless if logotypes
were analysed in whole or separately by the symbols or typography. Colours of logotypes correspond to the
promised values and principles of banks; however, some well-known Slovenian banks could risk more with
colours red and especially orange which promote innovation and novelty, both nowadays popular values.
Seven banks of twelve were recognized according to the colour by most respondent (more than 50 percent)
which applies that colour is an important and powerful element of logotypes.

5 REFERENCES

• A Guide to Choosing Colors for Your Brand. URL: http://www.usabilitypost.com/2008/09/29/a-guide-to-

choosing-colors-for-your-brand/ (last accessed on 24. 1. 2018).
• ALLEN, J.: Bank logos and their nationalistic colours. URL: http://www.colorobjects.com/en/color-columns/
jason-allen/item/268-bank-logos-and-their-nationalistic-colours.html (last accessed on 15. 2. 2018).
• BATHI. The logo colors of accounting. Available from: https://99designs.com/logo-design/business-of-
color/accounting (last accessed on 16. 2. 2018).
• BROWN, E. 2016. 20 Top-Notch Banks & Their Logo Palettes. URL: https://www.designmantic.com/blog/
infographics/20-banks-and-logo-palettes/ (last accessed on 20. 11. 2018).
• BROWN, E. 2016a. Branding Your Bank: Logo Maker vs Custom Logo. URL: www.designmantic.com (last
accessed on 5. 11. 2017).
• BROWN, E. 2017. What Color Should Your Logo Design Be: 3 Questions To Ask!. URL: https://www.design-
mantic.com/blog/infographics/what-color-should-your-logo-be/ (last accessed on 22. 3. 2018).
• CHANG, W.L., LIN, H.L. 2010. The impact of color traits on corporate branding. African Journal of Business
Management. Vol. 4, No. 15, pp. 3344-3355.
• CHEN, Y., WEI, B. & CHEN, M. 2007. The effect of colour guidance on design communication. Tunghai, IASDR.
• CHIAZZARI, S. 2000. Barve. Ljubljana, Slovenska knjiga.
• Color Emotion Guide. URL: https://thelogocompany.net/blog/infographics/psychology-color-logo-design/
(last accessed on 4. 12. 2017).
• DE BARTOLI, M., MAROTO, J. 2001. Translating colours in web site localisation, Proceedings of the European
Languages and the Implementation of Communication and Information Technologies (Elicit) conference,
University of Paisley.
• DITTMAR, M. 2001. Changing Colour Preferences with Ageing: A Comparative Study on Younger and Older
Native Germans Aged 19–90 Years. Gerontology, Vol. 47, pp. 219–226.

8th Conference on Information and Graphic Arts Technology 199

 Poster section

• GRIMES, A. and DOOLE, I. 1998. Exploring the Relationship Between Colour and International Branding: A
Cross Cultural Comparison of the UK and Taiwan. Journal of Marketing Manadement, Vol. 14, pp. 799-817.
• JACOBS, L., KEOWN, C., WORTHLEY, R. 1991. Cross-cultural Colour Comparisons: Global Marketers Beware!
International Marketing Review, Vol. 8, No. 3, pp. 21-30.
• MACDONALDS, J.I.: Connecting illusion, brand recognition, colour meaning, and multiple intelligence. URL:
http://jamesmac.co/blogging/connecting-illusion-brand-recognition-colour-meaning-and-multiple-intel-
ligence/ (last accessed on 23. 01. 2018).
• MACNAB, M. 2008. Decoding Design: Understanding and Using Symbols in Visual Communication. Cincin-
nati: HOW books.
• MADDEN, T.J., HEWETT, T., ROTH, M.S. 2000. Managing Images in Different Cultures: A Cross-National Study
of Color Meanings and Preferences. Journal og International Marketing, Vol. 8, No. 4, pp. 90-107.
• McDonalds changes logo in Israel to show that its food is kosher. 2006. URL: http://www.worldjewishcon-
gress.org/en/news/mcdonalds-changes-logo-in-israel-to-show-that-its-food-is-kosher (last accessed 15
01 2018).
• Towergate Insurance. Colour in branding. What does it say about the industry? URL: https://www.towergate-
insurance.co.uk/liability-insurance/professional-indemnity-insurance/colour-in-branding (last accessed
on 23. 01. 2018).
• TRSTENJAK, A. 1996. Psihologija barv. Ljubljana: Inštitut Antona Trstenjaka za psihologijo, logoterapijo in
antropohigieno.
• WALCH, M., HOPE, A. 1995. Living colours: the definitive guide to colour palettes through the ages. San
Francisco: Chronicle books.
• WIEDEMANN, J. 2006. Logo design. Köln: Taschen.

Corresponding author:
Jani TOROŠ
Innovative Management Institute
Dunajska 125, 1000 Ljubljana, Slovenia
phone: +386 41 734 800, e-mail: [email protected]

Co-author(s):
Urška STANKOVIČ ELESINI
University of Ljubljana
Faculty of Natural Sciences and Engineering
Department of Textiles, Graphic Arts and Design
Snežniška 5
1210 Ljubljana, Slovenia
phone: +386 1 200 32 29, fax: + 386 1 200 32 70, e-mail: [email protected]

Karla MRKOBRADA
University of Ljubljana
Faculty of Natural Sciences and Engineering
Department of Textiles, Graphic Arts and Design
Snežniška 5
1210 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 200

 Poster section

IMPACT OF AGE OF PIGMENT PRINTING PASTE ON PRINT QUALITY

Petra FORTE TAVČER

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia

Abstract: The purpose of this work was to determine the impact of the age of the pigment printing paste
on the quality of printed fabric. Two printing pastes, light red and dark red, were prepared and they were
printed onto a cotton fabric by screen printing technique on a magnetic printing machine Mini MDF R350
(J. Zimmer). Printing was performed at a predetermined time intervals, which were 1, 2, 3, 4, 7, 14, 30 and 60
days after preparation of the printing pastes. Printed samples were air dried and cured at 150°C in an Ernst
Benz dryer. The printing pastes were air-sealed and stored in the refrigerator at the constant temperature of
8°C. The quality of the prints was monitored by measuring the color depth (K/S) and CIELAB color values of
printed samples. The color fastness of the prints to washing at different temperatures, resistance to rubbing
and resistance to light were measured as well. The changes of the viscosity of the printing pastes was measured
on a Rheolab QC (Anton Paar). It was found out that, the quality of the prints shows minimum deviation with
time and that the pastes were still appropriate for printing after 60 days.

Keywords: pigment printing, cotton fabric, color value, color fastness, durability.

1 INTRODUCTION

Textile printing is one of the most versatile and important methods used for introducing color and design to
textile fabrics. Several techniques and different colorants are available for textile printing. More than 50% of
all printed textiles are printed with pigments. The main factors leading to pigment printing being so widely
used are the quality of prints, application to almost every kind of fibre or mixture, and the ability to avoid any
washing processes after the curing. The process is simple, economical and environment friendly [1]. A large
color gamut is available. Insoluble pigments, which have no affinity for the fibre, are fixed onto a textile with
binding agents in the pattern required. The pigment printing pastes consist of pigments, binders, synthetic
thickeners and printing auxiliaries. The wide variety of binders and auxiliaries available allow the printer to
achieve almost any requisite fastness properties. A conventional thermal curing technique is generally used to
bind the pigments onto the textile surface. The heat energy starts the crosslinking of a binder, which binds the
pigment particles onto the surface [1]. Flat or rotational screen printing is still the prevailed technique in the
printing of pigments; however, the interest in the ink-jet printing technology is constantly on the increase [2].

Three required components of a pigment print system are pigments, binder and thickener, all of which are
dispersed in water. Most of organic and inorganic pigments are used in the form of pigment dispersions. They
usually contain pigment, dispersant, wetting agents, preservatives, thickener and chemicals to adjust pH.
Thickeners are mostly synthetic polyacrylatic polymers and binders are most commonly acrylic copolymer
emulsions [3]. Most of these components have low biodegradability and it is important to use as much of
prepared paste as possible to prevent the disposal of unused chemicals. A higher amount of the printing paste
can be used if the paste is stable enough to be used in a longer time interval.

The aim of the research was to determine the impact the age of the pigment printing paste has on the quality
of a printed fabric. Two printing pastes, light red and dark red, were prepared and they were printed onto a
cotton fabric with the screen printing technique in time intervals from 1 to 60 days. The quality of prints was
monitored by measuring color depth (K/S) and CIELAB color values of printed samples and by determining
different fastness properties.

8th Conference on Information and Graphic Arts Technology 201

 Poster section

2 EXPERIMENTAL

100% cotton fabric from Tekstina, Ajdovščina with the mass per unit area 156g/m2 was used in the research.

2.1 Printing

Table 1 presents the recipe for stock paste and Table 2 the recipes for two printing pastes, for a light and a dark
shade. Pigment Bezaprint Rot KGC from Bezema, Switzerland was used for printing. All chemical components
of the printing pastes were a product of Minerva, Italy.
Flat screen printing was performed on the laboratory magnetic printing machine MINI MDF R 390, Johannes
Zimmer AG (Austria). The area coverage of the printing paste on the cotton cloth was approximately 5 x 15
cm. The printing, drying and curing conditions are presented in Table 3.
Printing, drying and curing was performed 1, 2, 3, 4, 7, 14, 30 and 60 days after the preparation of the pastes.
The pastes were kept sealed in the refrigerator at constant temperature of 8°C.

Table 1. Stock paste recipe.

Component Quantity [g]
EUTSCHAUMER W. KONZ 4
MINERPRINTBINDER WST 150
CLEAR D/27 26
FIXATOR L/F 10
WEICHMACHER A/95 10
Distilled water 800
Sum 1000

Table 2. Printing pastes recipes.

Component Quantity [g]
Light shade Dark shade
Bezaprint rot KGC 6 20
Stock paste 994 980
Sum 1000 1000

Table 3. Printing, drying and curing conditions.

Phase Conditions
Flat screen stencil: mesh 43 threads/cm
Printing speed: 80%
Printing Squeegee diameter: 6 mm
Magnet pressure: level 3
No. of passes: 2
Drying Air drying
Ernst Benz TKF 15-M500 drier,
Curing
T = 150°C, t = 5 min

2.2 Analysis

After printing the samples were kept in a dry, dark place. All measurements were performed at once at the end
of all printings. All samples were conditioned before measuring according to standard SIST EN 20139:1992 at
20°C and 65% humidity for 24 hours.
Color properties were determined by a Datacolor Spectraflash® SF 600 PLUS-CT spectrophotometer, under
illuminant D65 using the 10° standard observer, d/8º measurement geometry and a measurement area of 20
mm in diameter. Ten measurements were done on each sample. CIE L*a*b*, C*, h and color differences (ΔE*)
were calculated using Datacolor Datamaster software.

8th Conference on Information and Graphic Arts Technology 202

 Poster section

The color strengths (expressed as K/S value) were calculated with the Kubelka-Munk Equation (1):

(1)

where K represents the absorption, S the scattering and R the reflection of light.

Color fastness to washing at 40°C and at 60°C was tested according to SIST EN ISO 105-C01: 1985 (E)
Crockfastness was measured according to SIST EN ISO 105-X12:1993 on a CM-5 Crockmeter AATCC Atlas (USA).
Ten measurement were done for dry and wet crockfastness tests.
Color fastness to artificial light was measured according to SIST EN ISO 105-B02:1999 on a Xenotest Alpha,
produced by ATLAS Material Testing Technology BV (USA).
The viscosity of printing pastes was measured using the rotational rheometer Rheolab QC (Anton Paar, Aus-
tria) at 25°C.

3 RESULTS & DISSCUSSION

Figure 1 shows the viscosity of printing paste in the dependence of shear rate on the 4th and 60th days after
preparation. The diagram shows a shear thinning flow behavior, which is typical for such polymeric solutions.
It is clearly seen that viscosity of all printing pastes is quite similar. Neither the concentration of the pigment
nor the age of the printing paste influence on the flow behavior. This meant that we could print with all pastes
under the same printing conditions and we could expect the same behavior.

Figure 1. Viscosity (h) of printing pastes versus shear rate (g) at 4th and 60th day after preparation.

Figure 2 shows the color strength, K/S, of light and dark samples, printed at different time intervals. K/S val-
ues of dark prints are higher than K/S values of light prints. There are no differences in K/S values between
samples printed at different times. We can conclude that the same color value can be obtained with the fresh
or 60 days old printing paste.

8th Conference on Information and Graphic Arts Technology 203

 Poster section

Figure 2. Color depth, K/S, of samples printed at different time intervals.

Tables 3 and 4 show the CIELAB color values of dark and light samples printed with pastes of different age. The
sample printed on the 1st day represents the standard for calculation of color differences, ΔE*. At the light
colored paste, most color differences are smaller than 1, which means that they are invisible for the human eye.
They do not increase with the age of the printing paste. The color differences are higher at dark color printing
paste, some even exceed the value 2, which is visible to human eye. Even at dark color, there is no increasing
of color differences with the age of the paste. From CIELAB values, we can conclude again, that the color of
the printed samples is adequate when they are printed with 60 days old printing paste.

Table 3. CIELAB color values of light colored samples.

Paste age (day) L* a* b* C* h ΔE*
1 44.73 60.19 30.19 67.34 26.64 /
2 44.99 60.26 29.56 67.12 26.13 0.69
3 44.99 60.23 29.54 67.08 26.12 0.71
4 45.24 60.45 29.06 67.07 25.68 1.27
7 44.88 60.53 30.29 67.68 26.58 0.39
14 45.38 60.23 29.84 67.21 26.36 0.74
30 45.00 60.90 30.06 67.91 26.27 0.77
60 44.83 60.83 30.56 68.07 26.67 0.74

Table 4. CIELAB color values of dark colored samples.

Paste age (day) L* a* b* C* H ΔE*
1 52.44 58.46 19.13 61.51 18.12 /
2 51.88 59.87 20.87 63.41 19.22 2.31
3 52.41 59.02 19.88 62.28 18.62 0.94
4 52.79 59.12 19.71 62.32 18.44 0.94
7 52.11 60.05 20.85 63.57 19.15 2.36
14 52.42 59.67 20.69 63.15 19.13 1.97
30 52.52 59.56 20.34 62.94 18.86 1.64
60 53.03 59.00 19.69 62.20 18.46 0.97

For the evaluation of the prints, their fastness properties are also important. They were investigated by measur-
ing the color fastness to washing at 40 and 60°C (Table 5), the fastness to light (Table 6) and the color fastness
to rubbing with the crock test (Table 7).
It can be seen that all prints have good fastness to washing at both temperatures. All values are 5, regardless
the color of the prints or age or printing paste.

All fastness to light are also good. All values are between 4/5 and 5.

8th Conference on Information and Graphic Arts Technology 204

 Poster section

Crock test shows that the samples have worse wet than dry fastness to rubbing, which is normal for pigment
printing. Also normal is that better values are obtained for light than for dark samples. The values do not
differ according to the age of the printing paste.
The obtained values of different fastness properties appointed that the samples are equally resistant to dif-
ferent impacts when they are printed with fresh or 60 days old printing paste.

Table 5. Color fastness to washing of light samples at 40°C and 60°C.

Grey scale rating for color change at washing
Light shade Dark shade
Paste age (day)
40°C 60°C 40°C 60°C
1 4/5 5 5 5
2 5 5 5 5
3 5 5 5 5
4 5 5 5 5
7 5 5 5 5
14 5 5 5 5
30 5 5 5 5
60 5 5 5 5

Table 6. Color fastness to light.

Blue scale rating for color change at exposing to light
Paste age (day) Light shade Dark shade
1 4/5 4/5
2 4/5 5
3 5 4/5
4 4/5 5
7 4/5 5
14 4/5 4/5
30 5 4/5
60 4/5 5

Table 7. Rubbing fastness.

Grey scale rating for color change at rubbing
Paste age (day) Light shade Dark shade
Dry Wet Dry Wet
1 5 2/3 45 2
2 4/5 2 4/5 2
3 5 3 4/5 2
4 5 2/3 4/5 2
7 5 3 4/5 2
14 4/5 4 4/5 2/3
30 5 2 4/5 2
60 5 2/3 4/5 2

8th Conference on Information and Graphic Arts Technology 205

 Poster section

4 CONCLUSIONS

In the research, it was investigated how long it is possible to use a pigment printing paste in order to obtain
quality prints on a cotton fabric. The period from 1 to 60 days was chosen from practical reasons. It was estab-
lished that quality prints are obtained even 60 days after the preparation of the printing paste if it is properly
prepared, sealed and stored in a cool and dark place. The color strengths and CIELAB values of samples printed
60 days after the preparation were the same as of those printed on the first day. The prints were equally resistant
to washing, light and rubbing. The pastes kept their rheological properties and pH value.

5 REFERENCES

• Miles, L.W.C., 2004, Textile printing. Bradford: Society of Dyers and Colorists.
• Mikuž, M., Šostar Turk S., Forte Tavčer, P. 2010. “Properties of ink-jet printed, ultraviolet-cured pigment prints
in comparison with screen-printed , thermo-cured pigment prints.” Coloration Technology, 126, 249-255.
• Cardozo, B., 1995. Pigment Printing Handbook: A problematic approach to pigment printing, Research Tri-
angle Park, N.C. 27709, AATCC.

Corresponding author:
Petra FORTE TAVČER
University of Ljubljana
Faculty of Natural Science and Engineering
Department of textile, graphics and design
Snežniška 5
1000 Ljubljana, Slovenia
email: [email protected]

8th Conference on Information and Graphic Arts Technology 206

 Poster section

INFLUENCE OF OPTICAL PAPER PROPERTIES OF RECYCLED AND NON-RECYCLED

PAPERS FOR DIGITAL PRINTING ON COLOUR REPRODUCTION

Ivana JURIČ, Dragoljub NOVAKOVIĆ, Nemanja KAŠIKOVIĆ, Sandra DEDIJER, Neda MILIĆ & Jelena VASIĆ
Trg Dositeja Obradovića, 6, 21000 Novi Sad, Serbia / University of Novi Sad, Faculty of Technical Sciences,
Department of Graphic Engineering and Design

ABSTRACT: The main aim of the present research was to examine whether the papers made of 100% waste
paper could replace papers made of virgin fibres as substrates in InkJet printing. We chose several different
papers with different paper properties which were printed on InkJet digital printing machine. The optical paper
properties (whiteness and yellowness) were measured with spectro-densitometer SpectroDens (Techkon). The
test chart used in the paper contained primary (cyan, magenta, yellow and black) and secondary (red, green
and blue) process colours. Quality of the reproduced colour was also measured with spectro-densitometer
SpectroDens (Techkon). According to the obtained results, optical paper properties such as whiteness and
yellowness are inversely proportional. Results showed that some recycled papers have even the same whiteness
as non-recycled papers. The experiment shows that still recycled papers cannot be compared to non-recycled
regarding reproduced colour in inkjet printing.

Keywords: recycled paper, InkJet printing, colour reproduction, optical paper properties.

1 INTRODUCTION

The paper is still the substrate mostly used for the image reproduction in digital printing technique. Hence,
knowledge of the influence of paper on print quality is needed to understand and improve the digital print-
ing performance. Paper, which is the receiving substrate, is an important component in determining the final
print quality, especially for colour printing. Paper can be described as sheet material that is largely composed
of cellulosic fibers, which become joined together in a hydrogen-bonded structure as a consequence of the
evaporation of water (Hubbe et al, 2008).
To ensure the best quality and productivity on digital printing, some requirements for paper properties must
be met. While many types and grades of paper can be run in digital printing systems, there are a common set
of characteristics and quality criteria that must be built into papers that are optimized for performance in
digital printing systems – either monochrome or colour. While some offset sheets may also function, they may
not perform as reliably, or produce the same print quality as papers specifically designed for use in digital
printing systems (Xerox, 2004).
Printing companies participate in choosing the paper on which something is to be printed. Besides the quality
of paper, environmental consideration is one of the factors that influence the choice and therefore they often
choose to print on recycled paper. Demand for the use of recycled paper has increased recently. The original
motivation for paper recycling was primarily economic. The intention was to use collected waste paper in the
manufacturing of paper. That was primarily a concern in countries without indigenous forests, which could
ensure their production of newsprint and packaging material. Today environmental concerns are the main
motivation.

The primary material in the production of paper is pulp which consists of cellulose fibers. The fibers are either
fresh fibers from the wood of coniferous and deciduous trees or recycled fibers from collected paper (Silfver-
stolpe, 2008). There are two principal types of pulp, depending on the method of the production: chemical or
mechanical. Mechanical pulp is produced primarily from spruce wood, and during the production, the fibers
are liberated only through mechanical processing of the wood. In the production of the chemical pulp, cel-
lulose fibers are extracted by boiling with chemical additives. The chemical pulp consists both of long-fiber
pulp from coniferous trees and of short-fiber pulp from deciduous trees.
Recycled paper is a type of paper that completely or partially consists of recycled fibers. These fibers can have
very different origins and therefore also very different characteristics when it comes to being a component of
the new paper. When producing graphic papers based on recycled fibers, the paper mill needs as high-quality
raw materials as possible. This includes as little ink as possible to facilitate deinking, or preferably not having
deinking at all. Unlike the preparation of the pulp from virgin fibers, the fibers obtained from printed recov-
ered paper must have the ink removed. Large particles of ink that are left in the pulp will result in blemished
papers with visible specks of ink. If the ink particles are less than 40 μm in size, the eye cannot resolve them

8th Conference on Information and Graphic Arts Technology 207

 Poster section

(Thompson, 2004). Recycled paper also includes a large proportion of the paper from chemical pulp, which
produces the strongest fibers (Silfverstolpe, 2008). The concerns with recycled paper are usually about the
fiber qualities. However, the increasing demands and improvements in technology have been perfecting the
recycled paper quality, as well as driving the cost to be in the same range compared to the virgin fiber paper
(Wingkono et al, 2011). It is said that recycled paper is more expensive then non-recycled paper. It is wrong
because the recycled paper has aligned with conventional paper prices in recent years due to the increase in
demand, refinement of the papermaking process and the significant reduction in energy and water necessary
to make recycled paper compared to conventional paper (Jurič et al, 2013).
In previous research (Grilj et al, 2011; Jurič et al, 2013), it has been proven that there is a difference in prints on
recycled and non-recycled papers in electrophotography. Therefore there was an idea to investigate further
whether there is a possibility that recycled papers could replace regular papers in ink jet printing technique.

2 MATERIALS AND METHOD

For this research, we used several different papers with different paper properties (six recycled and eight
non-recycled papers). In Table 1 are presented papers with their names and grammage (papers marked with
green are recycled). All papers were printed on InkJet digital printing machine - Canon PIXMA MP230 with
the same print settings. The optical paper properties (whiteness (WCIE) and yellowness (G1925), which are
quality parameters very important for the paper industry) were measured with spectro-densitometer Techkon
SpectroDens using the standard observer of 10° and D50 illuminant. The test chart used in the paper contained
primary (cyan, magenta, yellow and black) and secondary (red, green and blue) process colours. Quality of the
reproduced colour was also measured with spectro-densitometer Techkon SpectroDens, the standard observer
was set at 2° and D50 illuminant. Lab values were measured five times for each colour, and only average values
are shown in the results.

Table 1. Papers used in this research,

Samples Paper Grammage [g/m2]
1 Nautilus Superwhite 250
2 Nautilus Superwhite 160
3 Nautilus Superwhite 70
4 Nautilus Universal 250
5 Nautilus Universal 160
6 Nautilus Universal 70
7 HP mat 220
8 HP semi gloss 220
9 HP high glossy 220
10 Fabriano 200
11 Fabriano 220
12 Perfect Proof Semi Matte Lite 155
13 Perfect Proof Semi Matte 185
14 Perfect Proof PhotoGloss 175

3 RESULTS & DISCUSSION

Obtained results for whiteness and yellowness of papers are presented in Table 2 and Figure 1. Based on the
results, it can be concluded that the whiteness and the yellowness are inversely proportional and in strong
correlation (R2 = 0.95). By increasing the whiteness of the paper, the yellowness decreases, and vice versa. Pos-
itive values of G1925 indicate a yellowish tint of paper, while negative values indicate a bluish tint of paper.

8th Conference on Information and Graphic Arts Technology 208

 Poster section

Table 2. Optical paper properties.

Samples Whiteness - WCIE Yellowness – G1925
1 114,44 -13,56
2 112,42 -12,36
3 107,40 -13,16
4 77,63 0,84
5 74,13 1,22
6 75,44 -1,67
7 109,56 -8,62
8 112,34 -12,43
9 115,33 -14,01
10 129,34 -19,31
11 129,21 -19,77
12 99,79 -7,96
13 97,70 -6,38
14 103,51 -7,78

It is clear that recycled papers (samples 4, 5 and 6) have the smallest whiteness (and greatest yellowness), but
the first group of recycled paper (samples 1, 2 and 3 – Nautilus Superwhite) can counter non-recycled papers.
Fabriano papers (samples 10 and 11) are the whitest with the smallest yellowness.

Figure 1. Optical paper properties of samples used in the experiment.

In Figures from 2 to 8 are presented colorimetric values (L*a*b*) of primary and secondary colors. Based on
the results, it can be noted that there is a difference in the printed colour on recycled and non-recycled papers.
Smaller lightness is obtained for cyan and yellow on recycled papers, while reddish tones (magenta and red)
are darker on recycled paper. For black, green and blue, the lightness values are more or less the same, the
only difference is noticed on Fabriano papers (samples 10 and 11). Chromatic values (a* and b* coordinates)
differ in all colors between different groups of papers.

8th Conference on Information and Graphic Arts Technology 209

 Poster section

a) b)
Figure 2. Lab coordinates for cyan, a) L* values and b) a* and b* values.

a) b)

Figure 3. Lab coordinates for magenta, a) L* values and b) a* and b* values.

a) b)

Figure 4. Lab coordinates for yellow, a) L* values and b) a* and b* values.

8th Conference on Information and Graphic Arts Technology 210

 Poster section

a) b)
Figure 5. Lab coordinates for black, a) L* values and b) a* and b* values.

a) b)

Figure 6. Lab coordinates for red, a) L* values and b) a* and b* values.

a) b)

Figure 7. Lab coordinates for green, a) L* values and b) a* and b* values.

8th Conference on Information and Graphic Arts Technology 211

 Poster section

a) b)
Figure 8. Lab coordinates for blue, a) L* values and b) a* and b* values.

4 CONCLUSIONS

Recycled papers are slowly becoming more and more used for printing and could replace nonrecycled papers
for less demanding and lower quality prints. We were interested in whether there was a difference in the
optical properties of recycled and nonrecycled papers, and whether these properties affect the final quality
of the digital prints (inkjet). According to the results, some recycled papers have the similar whiteness as
regular papers, while for some recycled papers whiteness is much lower, and the yellowish is larger. But the
printed colour varies considerably on these papers, regardless of their similar optical properties. Some tones
are brighter (cyan and yellow) while the other (magenta and red) are darker which are printed on recycled
papers. Also, there are major differences in the chromaticity of primary and secondary colours. Therefore, based
on the results obtained, it can be concluded that recycled papers can not yet be replaced by recycled material
if the colour reproduction and the quality of the printed colour are significant.

5 REFERENCES

• Grilj, S., Muck, T., Hladnik, A. and Gregor-Svetec, D. (2012) Recycled papers in everyday office use, Nordic
Pulp and Paper, Research Journal, Volume 27, 2012, Issue No. 4, pp. 739-749
• Hubbe, M. A., Pawlak, J. J. and Koukoulas, A. A. (2008). Paper’s appearance: A review, BioResources 3 (2),
pp. 627-665
• Jurič, I., Novaković, D., Karlović, I., Tomić, I. (2013) The possibility of using recycled paper in digital printing.
Advanced technologies 3(1), UDC 676:655.3:628.477.6, pp. 26-32
• Silfverstolpe, M. (2008) Recycling of printed products, The environmental council of the Swedish printing
industries, Belgium
• Techkon (2010) Manual – Spectro-denistometer SpectroDens. [Online] Avaiable at: http://www.partxpres.
com/files/spectrodens_manual.pdf, (last visited: 16.7.2015.)
• Thompson, B. (2004): Printing materials: Science and technology: A Pira International printing guide, Pira
International, Leatherhead
• Wingkono, G., Oswald, T. R. and Stoffel, J. (2011) Surface Treatment to Improve Print Quality on Recycled
Paper, NIP 27 and Digital Fabrication 2011, Technical Program and Proceedings
• Xerox Corporation (2004): Helpful facts about the paper [Online] Available at : http://www.xerox.com/
downloads/usa/en/s/supp_lib_Helpful_Facts_About_Paper.pdf (last visited: 16.12.2012.)

Corresponding author:
Ivana JURIČ
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 25 92 e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 212

 Poster section

Co-author(s):
Dragoljub NOVAKOVIĆ
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 26 20 e-mail: [email protected]

Nemanja KAŠIKOVIĆ
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 26 22 e-mail: [email protected]

Sandra DEDIJER
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 26 23 e-mail: [email protected]

Neda Milić
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 25 92 e-mail: [email protected]

Jelena VASIĆ
Department of Graphic Engineering and Design
Faculty of Technical Sciences, University of Novi Sad
Trg Dositeja Obradovića 6
21000 Novi Sad, Serbia
phone: +381 21 485 25 92 e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 213

 Poster section

POLYLACTIC ACID FIBRES CONTAINING FLUORESCENT PIGMENTS

Viera JANČOVIČOVÁ, Barbora ĎURICOVÁ, Veronika HRABOVSKÁ & Katarína HABEROVÁ

Institute of Natural and Synthetic Polymers, Faculty of Chemical and Food Technology,
Slovak University of Technology in Bratislava, Radlinského 9, 812 37 Bratislava, Slovak Republic

ABSTRACT: Fibres are globally considered as dynamic, highly developing materials. Research is currently
engaged in the study of special fibres, which have interesting properties. This work is focused on the char-
acterization and study of light and thermal stability of PLA fibres containing yellow fluorescent pigment.
Spectroscopic and colorimetric measurements were used to characterize and study sample stability. Samples
were found to be more subject to light aging; with significant color changes and fluorescence disappeared
completely after 4 days. Changes in the PLA structure using FTIR spectroscopy were not observed.

Keywords: fibres, polylactic acid, fluorescent pigment, colorimetry, spectroscopy.

1 INTRODUCTION

Fibres are generally considered as dynamic, highly developing materials. Materials made from natural and
chemical fibres have a wide use in all areas of life. Synthetic fibres are resistant to atmospheric influences
and are inexpensive therefore mass production of these materials contributes to accumulation of the waste
in environment, in the better case in landfills. The solutions of these serious environmental problems require
replacement of the synthetic polymers by biodegradable polymers and polymers from renewable resources
(Sin, 2013). Research is currently engaged on the study of special fibres, which are interesting from the point
of view of their properties (biodegradability, nontoxicity, security). Such fibres are used in the manufacturing
of protective or children’s clothing and in various medical applications. This work is focused on the charac-
terization and study of light and thermal stability of polylactic acid (PLA) fibres with different contents of
fluorescent pigments.

2 EXPERIMENTAL

The yellow fluorescent pigment RADGLO EA – 30: Chartreuse (0.1 – 0.8 wt. %) was added directly to the polymer
matrix (PLA) in order to prepare fibres with multifunctional properties containing the pigment directly in its
structure. The pellets of PLA were mechanically mixed with pigment and mixture was used for undrawn PLA
fibres preparation by classical spinning from melt at the temperature 190°C using a laboratory pilot line. The
undrawn PLA fibres were drawn on the drawing ratio 1.5.
Two types of accelerated aging – light aging in the Q-sun Xenon test chamber (type Xe-1-S light intensity at
420 nm: 1.1 W/m2; illumination 89000 Lx; black point temperature 65°C, window glass filter, light intensity in
the region 300 – 800 nm approx. 494 W/ m2) and thermal aging in the dark at 80°C - were used to study the
light and heat stability of the prepared samples [2].
UV Vis spectra of samples were monitored by UV Vis spectrometer CE3055 (Cecil Instruments). Original reflec-
tance (R) spectra were transformed into optical density (O.D.) using Eq. (1):

O.D. = -log R (1)

The FTIR spectra were measured with the FTIR spectrophotometer Excalibur Digilab FTS 3000MX, USA, using
the ATR adapter with diamond window.
The colorimetric coordinates of samples (L*, a*, b*) were obtained by means of Spectrophotometer Spectro-
dens (Techkon, illumination D50, standard observer 2°). CIE L*a*b* system was used to evaluate the color
changes. Value L* represents the lightness of color spot, chromatic coordinates a* and b* range from green to
red and from blue to yellow colours, respectively.
The total color difference was calculated from Eq. (2) (Hunt, 1995),

(2)

where values DL*, Da* and Db* are the differences between relevant values attributed to aged and non-aged
samples.

8th Conference on Information and Graphic Arts Technology 214

 Poster section

3 RESULT & DISCUSSION

3.1 Characterization of samples

Two spectroscopic methods (UV Vis, FTIR) and colorimetry measurements were used to characterize samples
as well as to study the effects of accelerated aging [1]. Figure 1 shows FTIR spectra of PLA fibers with varying
contents of fluorescent pigment RADGLO CHART EA-30. As the pigment was added in very small amounts, the
spectra correspond mainly to PLA. In the FTIR spectra of PLA typical bands at 2960 – 2880 cm-1 (-CH3, -CH2, -CH
group), 3590 – 3650 cm-1 (-OH group), 1720 – 1740 cm-1 (-C = O group) can be observed. The absorption bands
of the pigment are visible only at 2855 and at 1600 cm-1, where a slight increase in absorbance with increasing
pigment content can be observed. UV Vis spectra of PLA fibers with various content of pigment are shown in
Fig. 2. Since the PLA is white, it absorbs only in the UV range (235 and 300 nm). In the UV Vis spectra the max-
ima at 335 and 430 nm and the minimum at 310 nm increase with the increasing pigment content (Figure 2).

0.15
Absorbance [ - ]

0.10

0.05

0.00
4000 3500 3000 2000 1800 1600 1400 1200 1000 800 600
Wavenumber [cm-1]

Figure 1. FTIR spectra of PLA fibres with various pigment concentration

(0.1 wt. % red line, 0.5 wt. % blue line, 0.8 wt. % green line, without pigment black line).

0.6
Optical density [ - ]

0.4

0.2

0.0
200 300 400 500 600
Wavelength [nm]

Figure 2. UV Vis spectra of PLA fibres with various pigment concentration

(0.1 wt. % red line, 0.5 wt. % blue line, 0.8 wt. % green line, without pigment black line).

Colorimetric coordinates L*, a*, b* were also measured on all samples (Table 2). With the increasing pigment
content, the values of the coordinates L*, a*, b* were changed. PLA had the L*, a*, b* coordinates typical for
white, but when the pigment was added, the coordinate L* (darkening of the samples) and the coordinate a*
(shift to green) decreased, while the coordinate b* increased (shift to yellow). (Panák, 2008)

8th Conference on Information and Graphic Arts Technology 215

 Poster section

Table 1. The colorimetric coordinates L*, a*, b* of samples with various pigment concentration
Pigment content [%] L* a* b*
0 96 ± 4 1.2 ± 0.9 1.5 ± 0.5
0.1 63 ± 4 -10.1 ± 0.7 9±2
0.5 65 ± 2 -23 ± 1 29 ± 3
0.8 66 ± 2 -28 ± 1 37 ± 2

3.2 Light and temperature stability of samples

Subsequently, the color stability of the pigmented fibres was analysed. The samples were subjected to light
and thermal aging, the changes were evaluated by UV Vis spectroscopy and colorimetry.
After 3 days of accelerated light ageing (Q-sun chamber) the significant changes in UV Vis spectra were ob-
served (Figure 3). In pigment free PLA fibres (Figure 3A), after the first day of light aging, the absorbance in the
visible region of the spectra increased, indicating a darkening of the sample. In UV Vis spectra of pigmented
samples (Figure 3B), the peak at 430 nm decreased and at the same time decreased the minimum at 335 nm
associated with the disappearance of fluorescence due to pigment degradation. Based on the above mentioned,
we can assume that even the sample with the highest pigment content (0.8%) did not show fluorescence after
the first two days of accelerated light aging.
Colorimetric measurements (Table 2) showed that during light aging of pigment free PLA fibres the coordinate
L* decreased (samples darkening), the changes of coordinates a* and b* were negligible, their values are in
the area of achromatic colors. In samples containing pigment we observed the increase of the coordinate L*
(samples fading) and a* (green loss) and the significant decrease of the coordinate b* (yellow loss), which is
associated with significant decoloration of samples, as evidenced by the high ΔE*ab values, which increase
with increasing pigment content.

0.7

0.6
A 0.6 B
Optical density [ - ]

Optical density [ - ]

0.5
0.4
0.4

0.3
0.2

0.2

0.0
0.1 200 300 400 500 600
200 300 400 500 600
Wavelength [nm]
Wavelength [nm]

Figure 3. UV Vis spectra of pigment free PLA fibres (A) and PLA fibres with 0.8 wt. % of pigment (B) during
light ageing (1 day red line, 2 days blue line, 3 days green line, unaged samples black line).

The samples were subjected to thermal aging without light access, where the aging time was 28 days. During
thermal aging much smaller and slower changes than during light aging of samples (except PLA without
pigment) (Table 1) were observed. Changes of the shape of the UV Vis spectra were small, and the “minimum”
at 310 nm remained present, although with time decresed (Figure 4). Thus, we can assume that fluorescence
of pigmented fibres under thermal aging decreased, but did not disappear. During light aging in the Q-SUN
chamber the fluorescence of the samples with pigment content of 0.8 wt. % completely disappeared after 2
days, the effect of thermal aging was significantly smaller.
In the case of pigment-free fibers, the more significant color changes (mainly the L* coordinate decrease) were
observed at thermal ageing (80°C) as at light ageing, with the presence of the pigment in the amount of 0.1%
already increased the PLA thermal stability.

8th Conference on Information and Graphic Arts Technology 216

 Poster section

0.6 A 0.6 B

Optical density [ - ]
Optical density [ - ]

0.4 0.4

0.2 0.2

0.0 0.0
200 300 400 500 600 200 300 400 500 600
Wavelength [nm] Wavelength [nm]
Figure 4. UV Vis spectra of pigment free PLA fibres (A) and PLA fibres with 0.8 wt. % of pigment (B) during
thermal ageing (1 day red line, 3 days blue line, 28 days green line, unaged samples black line).

Table 2. Comparison of color changes of PLA fibers with pigment due to light and heat aging
DEab* after 3 days
Pigment content [%]
light ageing thermal ageing
0 15.2 42.7
0.1 14.8 10.1
0.5 36.7 5.0
0.8 47.5 7.1

After four days of light aging in the Q-SUN the samples were stored and after one year their characteristics
were re-measured. Subsequently, they were light aged for 7 days. During storage and additional aging, there
were no significant changes in the FTIR and UV Vis spectra. So we can say that the most important changes
occurred during the first days of aging.
In colorimetric measurements of the samples, which were after 3 days of light ageing one year stored in dark
and subsequently aged for seven days in the Q-SUN chamber, there were no significant changes in the coor-
dinates L*, a*, b* and ΔEab* values are in the range of measurement errors.

4 CONCLUSIONS
In conclusion, the color stability of PLA fibers with fluorescent pigments is significantly influenced by light,
as all samples were significantly decolorised and lost their fluorescence after 3 days of accelerated light age-
ing (corresponding approximately to 3 years of natural ageing), whereas the color changes achieved after 28
days of aging at 80 °C in the dark were much smaller and fluorescence decreased only about 50 %. Potential
modification of fluorescent pigments could be increase its stability in PLA fibers.

ACKNOWLEDGEMENTS
This work was supported by the Slovak Research and Development Agency under the contract No: APVV-14-
0175.

5 REFERENCES

• Sin, T.L. et al. 2013. Polylactic Acid; PLA Biopolymer Technology and Applications (Oxford, UK: Elsevier),
ISBN: 978-1-43-77-4459-0.
• Ďuricová, B. 2017. “Special colorants used in the fibers.” Bachelor´s Thesis. Slovak University of Technology
in Bratislava.
• Hunt, R.W.G. 1995. Measuring Colour, Ellis Horwood Limited, London, United Kingdom.
• Panák, J., Čeppan, M., Dvonka, V., Karpinský, Ľ., Kordoš, P., Mikula, M., Jakucewicz, S. 2008. Polygrafické min-
imum. Bratislava: Typoset.

8th Conference on Information and Graphic Arts Technology 217

 Poster section

Corresponding author:
Viera JANČOVIČOVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers,
Department of Graphic Arts Technology and Applied Photochemistry
Radlinského 9,
812 37 Bratislava, Slovak Republic
phone: 00421259325227; e-mail: [email protected]

Co-author(s):
Barbora ĎURICOVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers,
Department of Graphic Arts Technology and Applied Photochemistry
Radlinského 9,
812 37 Bratislava, Slovak Republic
e-mail: [email protected]

Veronika HRABOVSKÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers,
Department of Plastics, Rubber and Fibres
Radlinského 9,
812 37 Bratislava, Slovak Republic
e-mail: [email protected]

Katarína HABEROVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers,
Department of Graphic Arts Technology and Applied Photochemistry
Radlinského 9,
812 37 Bratislava, Slovak Republic
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 218

 Poster section

PREVENTION OF STARCH DEGRADATION IN PAPER MAKING PROCESS

FOR RECYCLED PACKAGING PAPER

Dimitrina TODOROVA, Ventsislav BLYAHOVSKI & Spas LADZHOV

University of Chemical Technology and Metallurgy, Sofia, Bulgaria

ABSTRACT: In this report were observed the possibilities of prevention of starch degradation in papermaking
processes for corrugated paper production by using zinc-based product that inhibits amylase enzyme. The
examinations were carried out at zero and 24 hours’ downtime of the paper suspension. The consumption of
the amylase enzyme inhibitor additive was 0,5%; 1%, 1,5% and 2% of o.d.f. Laboratory paper samples (90g/
m2) were obtained. Dewatering time, turbidity and conductivity were determined. Starch amount was meas-
ured using iodine staining. The investigated strength properties were - burst strength test index, compressive
strength test index and tear index. The results show positive effect of the enzyme inhibitor over the dewatering
effect and turbidity of the white waters together with constant strength properties. Most interesting is the
result for the starch concentration in the white waters, which are in the range of 0,05% even after the 24 hours’
downtime of the paper suspension.

Keywords: recycling, paper, starch, enzyme inhibition, corrugated packaging, strength .

1 INTRODUCTION

The demand for packaging and container boards has grown both globally and in Europe in the past few years.
The annual demand is expected to grow at about 3-6% in emerging markets and with 1 % in Europe over the
next five years and will amount to almost 115 million tonnes of converted material by 2019 (Smithers Pira,
2015). Global demand for container boards will continue to increase due to the increase online shopping and
trading with a significant market share. At the same time, board manufacturers are looking for opportunities
to decrease the production costs and to improve lightweightening of the paper together with environmental
and sustainability requirements without compromising the most critical quality parameter – strength.
The papermaking process consists of several major steps: stock preparation, sheet forming, pressing, dry-
ing, and surface finishing (Maurer, 2009). Starch is an important component of many paper grades. Starch
consumption by weight in papermaking and paper conversion processes ranks third after cellulose fibre and
mineral pigments. Starch is used as a flocculant and retention aid, as a bonding agent, as a surface size, as a
binder for coatings, and as an adhesive in corrugated board, laminated grades, and other products (Vrabic
Brodnjak, 2017). The major starch sources are corn, potato, waxy maize, wheat, and tapioca. Refined starches
are supplied in powder form or as slightly aggregated pearl starch. Unmodified (native) starch is rarely used
in the paper industry, except as a binder for laminates and in the corrugating process. Most starches for use in
papermaking are specialty products that have been modified by controlled hydrolysis, oxidation, or derivati-
zation (Maurer, 2009). When choosing starch for paper production there are some important aspects such as:
application requirements for starch, dispersion of starch, environmental aspects of starch, and use of starch
in papermaking furnish, is it for surface sizing of paper or as a coating binder, as adhesive in paper conversion
or in newer specialty papers Starch analysis in paper is having the same importance as the other wet-end or
surface chemical additives (Neimo, 1999).
It is also well known that the typical composition of the main paper industry microflora consists of aerobic
bacteria and anaerobic bacteria, which are two types, spore forming and non-sporulating, fingi – moulds and
yeast and algae – blue-green or green (Blanco, 1996). Due to this favourable conditions preventing or reducing
deposits in pulp and paper making processes is having significant importance for the overall papermaking
runability together with the need of increased savings. One opportunity for preventing or reducing deposits in
pulp and paper making processes is by including a cyclodextrin during or after pulp formation in an amount
effective for preventing or reducing the deposits (US 2006O124266A1).
Actual trends in papermaking concern on the use of recycled fibres and minimizing effluent volume by closing
water circuit. Beside environmental and economic advantages, both trends can lead to serious problems in
papermaking processes due to contaminants introduced with recycled fibres and their accumulation in closed
water system. A key factor to avoid disturbances in the papermaking is to prevent the accumulation of det-
rimental compounds. In this respect, first it is necessary to find out methods to characterize and control the
effects of chemical additives for different papers grade.

8th Conference on Information and Graphic Arts Technology 219

 Poster section

Amylase is an enzyme produced by many bacteria and fungi. The α-amylase belongs to a family of endo-amyl-
ases that catalyses the initial hydrolysis of starch into shorter oligosaccharides through the cleavage of α-D-
(1-4) glycosidic bonds (Souza, 2010). This enzyme breaks down starch polymers into smaller sugars which
can be seen as high COD values in process waters and increased microbial activity due sugar nutrients (Fig.1).
Significant improvement of the enzyme inhibition could be achieved by appropriate microbiological control
followed by starch stabilization which stopes the enzymatic degradation of starch. Retention and strength
management to retain starch and boost strength and productivity is also beneficial.

Figure 1. Starch in papermaking process.

Two different mechanisms occur while using enzyme inhibition additive, one mechanism inhibits the existing
amylase activity and the other mechanism prevents the production of new amylase by microorganisms, giving
a synergistic impact (US 9278874 B2). A special amylase enzyme inhibitor prevents the enzymatic degradation
of starch polymers, and fine-tuned synthetic polymer program ensures that this secondary starch is retained
within fibres (Fig.2). Less starch degradation means less pH drop and that more starch is present with the re-
cycled fibres (Thorn, 2009). Other benefits are improved runability due to stable wet-end chemistry as well as
optimized retention, better dewatering ability of paper suspensions, improved turbidity of the white waters,
increased strength properties and more efficient and sustainable raw material usage due to improved yield of
starch and reduced need of added starch.

Figure 2. Starch in the paper suspension.

In this report were observed the possibilities of prevention of starch degradation in recycling and papermaking
processes for corrugated paper grades production by using zinc-based product that inhibits amylase enzyme,
which has benefits both on paper suspension properties and paper strength properties of the produced paper
grades. All the examinations were carried out at zero and 24 hours’ downtime of the paper suspension.

2 EXPERIMENTAL

The used raw material was secondary fibre material for the production of testliner for corrugated cardboard
grades delivered from DS Smith Bulgaria SA. The secondary fibre material was prepared in laboratory con-
ditions by hydropulper and the end refining degree was 35ºSR. Paper slurry contained the regularly used
chemical additives for the production of the corrugated liner paper grades: Carboxymethyl Cellulose - 16
kg/t, Starch - 30kg/t and Retention Additive - 0,136l/t. As enzyme inhibitor was used Fenno Spec 1200, deliv-
ered by Kemira. It is a zinc-based product (zinc sulphate 28-30%, zinc chloride <=0,58%) that inhibits amylase
enzyme. As retention additive was used Fenno Pol K 2130T - modified poly acryl amide with cationic charge,
delivered by Kemira.

8th Conference on Information and Graphic Arts Technology 220

 Poster section

The amylase enzyme inhibitor additive was used at consumption of 0,5%; 1%, 1,5% and 2% of o.d.f. Paper sus-
pensions were prepared in accordance with the following order of adding the components:
1. Only pulp (P)
2. Pulp and 0,1% retention additive (PR)
3. Pulp, 0,5% Feno Spec 1200, 0,1% retention additive (0,5% FS)
4. Pulp, 1% Feno Spec 1200, 0,1% retention additive (1% FS)
5. Pulp, 1,5% Feno Spec 1200, 0,1% retention additive (1,5% FS)
6. Pulp, 2% Feno Spec 1200, 0,1% retention additive (2% FS)

Dewatering time (T700) of the resulting paper suspensions were determined followed by measuring Turbidity,
NTU acc. ISO 7027 and Conductivity, µS acc. ISO 7888 of the white waters. The refining degree was determined
by the Schopper Riegler Value ºSR, acc. to ISO 5267-1/AC: 2004. The dewatering time was determined by
Shopper Riegler apparatus (Germany). The measuring conditions are the same as for determination of beating
degree but the central vertical out-pipe is closed. The concentration is 0.2% (2g o.d.f in 1000ml water). In the
current experiment the dewatering time was measured for 700ml filtrate. Starch amount was measured using
iodine staining. For determining the starch concentration in the white waters, the absorbance of the iodine in
the white waters and in the three standardized starch solutions, were determined by using UV/Visible Scan-
ning Spectrophotometer UV 3300 PC of BIOBASE Co., Ltd, by the method of standard calibration curve and
the concentration of the starch in the white waters was determined in percent starch. All paper samples were
prepared on paper laboratory machine (Rapid-Kothen, Germany) acc. ISO 5269-2:2005, with a grammage of
90g/m2, with drying conditions of - 96°C and duration of 7 minutes. The investigated strength properties of
the obtained paper samples: burst strength test index (BI) according to ISO 2759:2014, compressive strength
test index (SCT) accord. to ISO 9895:2008 and tear index (TI), accord. to ISO 1974:2012 were analysed in the
standard atmosphere at 23 °C of temperature and 50% of relative humidity.

3 RESULTS & DISCUSSION

Usually, the dewatering time gives us indirect information about the flocculation ability of the paper sus-
pension together with the drainage ability. The accelerated dewatering and increased retention as a result
of flocculants effect means purer waters in paper mill as well. After preparation of each pulp mixture, the
dewatering time has been measured (Fig.3). As it is seen from the chart, the dewatering time is improved for
all paper suspensions, compared to that from secondary fibre material only (P).

Figure 3. Dewatering time T700, s of the paper suspensions.

Adding of retention additive (PR) improves the dewatering ability with 10 s, but best result at 0h downtime was
observed for the paper mixture with 0,5% of o.d.f. enzyme inhibitor (0,5%FS). Increasing the FS consumption
from 1% to 2% had no significant effect and dewatering time is practically unchanged. The tendencies after
24h downtime of the paper suspension are different. The dewatering of suspension consisting only retention
additive aggravated significantly with about 8 seconds, but still had better dewatering than that of pulp only.

8th Conference on Information and Graphic Arts Technology 221

 Poster section

Paper suspensions consisting Feno Spec 1200 were with improved dewatering time. At 0,1% consumption 700
ml filtrate was obtained for less than 25s and this result is only 1s higher than that at 0,5% FS at 0h downtime
of the paper suspension. The optimal FS consumption regarding the dewatering time, both at 0h and 24h
downtime is 0,1% of o.d.f. The improvement of the dewatering time probably is due to the synergic effect of
the retention additive and the zinc based additive, which has also positive charge 9not so strong in electricity
as Al3+) and acts as a coagulant.

Figure 4. White waters turbidity, NTU and conductivity, µS.

Measuring the electric conductivity gives an indication of the total concentration of electrolyte in the liquid
phase, while the turbidity indicates indirectly the overall retention effect. Regarding the turbidity and con-
ductivity of the white waters, the results which are presented in Fig. 4 are interconnected. With increasing the
enzyme inhibition additive from 0,5%FS to 2%FS the turbidity is decreased, while the conductivity is increased
and this effect is bigger while using paper suspensions directly after its preparation. At 24h downtime the
clarification effect is not that strong. It is visible from the chart, that consumption of 2% Fenco Spac 1200 is
sufficient for inhibition of the amylase enzyme and results before and after the paper suspensions downtime
are very similar. The curves for the conductivity before and after the suspensions downtime are very close and
presence of FS increased the white waters conductivity, both at 0h and 24h suspension downtime due to the
increased retention of fines, starch and inorganic substances.
As shown in Table 1, as the concentration of the enzyme inhibitor increased and at 24h suspensions downtime,
the starch concentration in the white water decreased. Because fresh starch was not used and the studies were
conducted only with the starch in available secondary paper, which is 30kg/t, the amount of starch in the
white waters ranged from 0.04-0.05%. Results show that the effectiveness of the additive used can be judged
with great accuracy only when fresh starch is incorporated. These experiments are provided in the next stage
of the study, along with industrial experiments.

Table 1. White waters adsorption and starch concentration.

Starch, % Adsorption
Adsorption P PR 0,5%FS 1%FS 1,5%FS 2%FS
0,03 0,173
0,04 0,244 0h 0,396 0,319 0,318 0,314 0,311 0,31
0,05 0,35 24h - 0,269 0,285 0,284 0,286 0,285

The obtained results for the strength and dynamic parameters of the investigated laboratory obtained paper
samples are shown in Fig.5 and Fig.6. Before the testing process paper sheets were conditioned at tempera-
ture of 23°C ±1°C and relative humidity 50% ± 5% for 24 hours. Bursting strength is particularly important for
packaging papers, as it determines how much pressure paper can tolerate before it ruptures. The results for
the burst strength test - burst index (BI, kPa.m2/g) and tear strength - tear index (TI, mN.m2/g) are shown in
Fig. 5. It can be seen that both parameters are relatively constant. Burst index is more sensitive with regard

8th Conference on Information and Graphic Arts Technology 222

 Poster section

to the downtime of the paper suspension, which is probably due to the decrease of interfiber bonds, mainly
hidrogen bonds togeter with increase in the mobility of the paper structure elements and the increase in the
flexibility of the fibers themselves as a result of the inhance starch and overal retention. The tear resistance
determines the force applied during the tearing procedure. It indicates the behavior of paper and it is impor-
tant for packaging papers, where the toughness and shock absorption are very important factors. Tear index is
more stable and after 24h of papersuspension downtime the results are practically unchainged but still slightly
higher than the zero sample (pulp without any additives). This parameter depends primarily on the strength
and nature of the fibers themselves, that’s why the influence of the amount of additives and the downtime of
the paper suspension is insignificant.

Figure 5. Burst index and Tear index of the paper samples.

In the short-span compression test (SCT) the compressive strength of paper in compression mode is deter-
mined. Because only a very short length of paper (0.7 mm) is used in the test, greater account is taken of the
load-bearing fiber portion of the material than with the traditional test methods (ring crush test, corrugated
crush test or linear crush test).

Figure 6. SCT of the paper samples.

Determination of the compressive resistance has now become established as an alternative to the ring crush
test. The effect of the enzyme inhibition additive over the compressive strength test index (SCT) before and

8th Conference on Information and Graphic Arts Technology 223

 Poster section

after the paper suspensions downtime is shown of Fig.6. The SCT parameter for the investigated paper samples
is with constant values. The presence of Feno Spec 1200 has no negative strength effect.

4 CONCLUSIONS

The amylase enzyme inhibitor prevents the enzymatic degradation of starch polymer and the results from the
paper suspension analysis show positive effect of the enzyme inhibitor over the dewatering effect and turbid-
ity of the white waters. Investigated strength and deformation properties of the obtained paper samples are
constant before and after 24 h of paper suspension downtime. Result for the starch concentration in the white
waters are in the range of 0,05% even after the 24 hours’ downtime of the paper suspension.

ACKNOWLEDGEMENTS
The authors would like to thank University of Chemical Technology and Metallurgy, Sofia, Bulgaria for the
financial support under the project 11780. We are also thankful to DS Smith, Bulgaria and Kemira for providing
materials for this work.

5 REFERENCES

• Hans W. Maurer, Chapter 18 – Starch in the Paper Industry, Starch, Chemistry and Technology, A volume in
Food Science and Technology, 2009, Pages 657–713, Academic Press
• Hui Xu, Jing Luo, Kim Bloomfield, Julie Clemmons, Jun. 15, US 2006O124266A1 Patent Application Publication,
USE OF CYCLODEXTRINS FOR REDUCING DEPOSITS DURING PAPER PRODUCTION, 2006
• I.Thorn, C.O.Au (Eds.), „Applications of Wet-End Paper Chemistry”, 2-nd edition, 2009, Springer
• Leo Neimo , “Book 4 Papermaking Chemistry”, Fapet Oy, 1999, Finland
• M. A. Blanco · C. Negro · I. Gaspar · J. Tijero, “Slime problems in the paper and board industry”, Appl Microbiol
Biotechnol 46, 1996, 203 – 208, Springer-Verlag
• Paula Monteiro de Souza; Perola de Oliveira e Magalhaes, APPLICATION OF MICROBIAL α-AMYLASE IN
INDUSTRY – A REVIEW, Brazilian Journal of Microbiology, 2010, 41: 850-861, ISSN 1517-8382
• Smithers Pira, http://www.smitherspira.com, “Paper and Board Market in Central and Eastern Europe to
grow by 2% annually from 2013 to 2019”, January 2015
• U. Vrabic Brodnjak, D. Muck, Printing quality of chitosan-rice starch coated packaging paper, Bulgarian
Chemical Communications, Volume 49, Special Issue L, 2017, pp. 86-92
• US 9278874 B2 Prevention of starch degradation in pulp, paper or board making processes using zinc ions
and an oxidizing biocide US 14/348,070, PCT/EP2012/069228, 8. March 2016

Corresponding author:
Dimitrina TODOROVA
University of Chemical Technology and Metallurgy, Faculty of Chemical Technology,
Department of Pulp, Paper and Graphic Art
bld. 8 Kl. Ohridski / University of Chemical Technology and Metallurgy
1756, Sofia, Bulgaria
phone: +359 88 736 44 74, e-mail: [email protected]

Co-author(s):
Ventsislav BLYAHOVSKI
University of Chemical Technology and Metallurgy, Faculty of Chemical Technology,
Department of Pulp, Paper and Graphic Art
bld. 8 Kl. Ohridski / University of Chemical Technology and Metallurgy
1756, Sofia, Bulgaria
phone: +359 88 576 31 44, e-mail: [email protected]

Spas LADZHOV
University of Chemical Technology and Metallurgy, Faculty of Chemical Technology,
Department of Pulp, Paper and Graphic Art
bld. 8 Kl. Ohridski / University of Chemical Technology and Metallurgy
1756, Sofia, Bulgaria
phone: +359 89 962 06 66, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 224

 Poster section

Printing braille with special expandable screen printing inks

Karin KUKEC, Raša URBAS

University of Ljubljana, Faculty of Natural Sciences and Engineering, Department of Textiles, Graphic Arts
and Design, Chair of Information and Graphic Arts Technology, Slovenia

ABSTRACT: Printing braille and other tactile elements for blind and visually impaired is distinguished by
specific characteristics, which are not always required in prints intended for normally sighted. For achieving
good readability and tactile recognition of prints adequate height of prints, size and dimensions of braille
and other tactile elements, suitable printing material and printing ink need to be fulfilled. Last but not least a
proper, suitable printing technology needs to be selected. This paper presents results of the conducted research
in which braille was printed with screen printing technique using two different 3D expandable water-based
printing inks.
Concluding from the results of the research we could establish that good readability and tactile recognition
can be achieved with printing only in one layer if we choose the proper 3D expanding printing ink with smaller
microspheres and modify the screen printing mash in a way that it enables the deposition of thicker layer.
Analysis has also shown that better durability and quality of prints needs to be achieved by addition of thin
coating varnish layer.

Key-words: braille, blind and visually impaired, tactile elements, screen printing, 3D expandable printing
inks.

1 INTRODUCTION

Each printing process possesses its certain specifics and demands for achieving proper quality, readability
and aimed content recognition, which consists of several factors. Printing braille is probably one of the most
complex printing methods. It is intended for a group of people whose specific needs dictate their way of per-
ception. Preparing all printing material and selecting proper technique is therefore complex.
Braille and other tactile elements are nowadays printed with different printing techniques, most often by
embossing and UV ink-jet printing, though in some cases other techniques are also used (e.g. flexo printing,
thermo-vacuum technique, screen printing etc.) (Chang, 2012; Kukec, 2015; Urbas, 2016; Zhang, 2011). Achieve-
ment of specific characteristics – adequate height of prints, dimensional ratios of individual elements (Braille,
International Council on English, 2013; Deutsches Institut für Normung e. V., 2009; Fajdetič, 2015; Slovenski
inštitut za standardizacijo; 2014) demand accurate – precise printing in several layers, where printing surface
of printed braille as well as other tactile elements needs to be smooth and rounded (without sharp edges), is
very time consuming. Beside mentioned, printing substrate needs to have specific properties. Printing substrate
should have higher mass (grammage) (between 100 and 300 g/m2) and thickness for preventing mechanical
damages (folds, cracks etc.) and surface abrasions, also it should possess relatively high smoothness for achiev-
ing easier tactile reading and preventing damages of fingertips. Though embossing is not using printing ink/s,
inks of other printing techniques need to offer smooth, soft and touch-pleasant prints that during reading
do not cause abrasions and fingertip skin injuries. Namely, UV inkjet printing often offers hard, rigid prints
of braille and other tactile elements, while embossing frequently causes either during the production (if
unsuitable printing substrate is being used) or during it use (reading) cracking of the tactile surface. In both
cases prints are therefore not preferable for reading by blind and visually impaired.
Due to the fact that specific dimensions of braille and other tactile elements need to be considered, prints
made for blind and visually impaired, compared with the prints indented for normally sighted, are much
larger and due to the used printing techniques (e.g. embossing) and readability requirements mainly printed
only one sided (Kukec; 2015). Therefore, efforts have been made in trying to find a suitable technique, which
would enable one layer both sided printing that would enable all demanded specifics of prints intended for
blind and visually impaired (Fajdetić, 2015; Kron, 2004; Kukec, 2015; Urbas, 2016).
This paper presents a study in which modified classic screen printing technique that allows, with certain mod-
ifications of screen printing form, one-layer printing with the use of special 3D expandable printing inks. The
use of these special 3D expandable inks reduces the consumption of the printing ink and printing substrate,
as well as the amount of work, which leads to saving money and time. In some cases, the use of these special
inks enables a possibility for printing braille not only on one but on both sides of paper format.

8th Conference on Information and Graphic Arts Technology 225

 Poster section

2 MATERIALS AND METHODS

2.1 Materials and printing method

The main goal of our research was to produce tactile recognizable prints, which would follow the standard
specifications for braille and other tactile elements with the use of modified conventional screen printing
technique.
For that purpose, two special 3D expandable screen printing inks have been used – Admiral (indicated as ADM)
and Minerfoam SR (indicated as MFSR) (both Achitex Minerva S.p.A., Italy), which, if applied and treated prop-
erly on selected printing substrate provide stable, adequately high and recognizable tactile prints. Printing
inks contained of printing base with microspheres that consisted of a thermoplastic polymer shell and liquid
core. Liquid core under certain conditions (according to the producer, temperature around 150°C and 3-minute
exposure) expand and therefore increase the thickness of prints. The size of microspheres, according to the
manufacturer range from 5 to 10 μm (Achitex Minerva, 2014).
Printing was performed manually using a screen printing PET mash with density of 32 threads/cm and mono-
filament diameter 100 μm on coated cardboard Kromopak (Količevo Karton, d.d., Slovenia), with grammage of
260 g/m2. In order to ensure a sufficient transfer of the printing inks to the printing substrate the screen was
coated with 10 layers (one on the inside, nine on the outside) of photoemulsion.
Selected braille text was printed on coated cardboard with smooth surface for easier tactile readability in one
and in two layers. The printing ink ADM was applied in two layers (wet on wet method), while MFSR was
printed only in one. Drying and expansion conditions of both selected printing inks differed slightly – ADM
printing ink was directly dried and expanded for 3 minutes at 130°C, while MFSR was firstly dried for 45 sec-
onds in the drying tunnel at 100°C and later expanded for 3 minutes at temperature of 150°C.
For achieving higher printing paper smoothness and improved durability of prints some printed samples were
coated with a thin one (1L) or two layers (2L) of water based Suncoat 9260/55s matt varnish, using the same
screen printing mash. Varnish was dried for 135 seconds at 100°C in the drying tunnel.

2.2 Testing methods

For conducting the results of printing braille with selected 3D expandable printing inks firstly the analyses
of printing material were performed (grammage, thickness, surface structure). Later, height of prints – braille
dots was measured with micrometer and SEM analysis (e.g. Scanning electron microscope) and measurements
were compared. For assessing the shape, surface structure and quality of prints SEM and OM (optical micro-
scope) images were investigated. Printed braille was also tested with blind and visually impaired, where pre-
pared prints were given to a group of 10 blind people which gave us their opinion on readability and tactile
recognition of prints.

3 RESULTS AND DISCUSION

The research of the printing substrate – coated cardboard showed that its grammage was 263,90 g/m2 (slight-
ly higher than defined by the producer), while its average thickness measured with micrometer was 0,434
mm and 0,521 mm measured with SEM. Both values were used later for the calculation of printed braille dot
height (calculated braille dot height = height of print – thickness of printing substrate). Image analysis has
shown that the printing substrate in average had relatively smooth surface, though it was on certain areas a
bit uneven (Figure 1).

Figure 1. Surface of coated cardboard printing substrate (SEM; 1.500× magnification).

Dot heights measured with micrometer and SEM (image analysis) are showed in Table 1. Heights that are
measured by the micrometer are as expected significantly lower, due to the fact that micrometer compresses

8th Conference on Information and Graphic Arts Technology 226

 Poster section

the samples during measurements with the load of 0,5 kg in the area of 1,0 cm2. Comparing the samples coated
with or without varnish can see that better results are achieved in samples coated with two layers of varnish
(2L) making them more resistant to pressure. SEM measured values of thickness were much higher showing
same trends of results as with micrometer.

Table 1. Measured thickness of prints and printing substrate on micrometer (M) and SEM together with cal-
culated braille dot height.
Thickness of prints Thickness of printing substrate Calculated braille dot height
Sample [mm] [mm] [mm]
M SEM M SEM M SEM
ADM 0,500 0,882 0,066 0,361
ADM-1L 0,539 0,894 0,105 0,373
ADM-2L 0,553 0,907 0,119 0,386
0,434 0,521
MFSR 0,524 0,971 0,090 0,450
MFSR-1L 0,566 0,983 0,132 0,462
MFSR-2L 0,570 0,995 0,136 0,474

SEM image analysis has shown that printed braille dots differ according to the used printing ink (Figure 2).
On dots printed with both ADM and MFSR printing inks the mesh structure of the used screen can clearly
be seen. Even the expansion of the microspheres in the printing ink did not completely diminish the effect.
Later application of one or two layers of the water-based varnish slightly smoothened the clear filament lines
on the prints though they can still clearly be seen. Though we have to emphasize that the SEM imaged were
taken at 70× magnification and that the effect is not visible when being observed without the magnification
(Figure 3).
Impression of the screen mash was more visible in the samples printed with ADM printing ink, which we
contributed to the fact that MFSR possessed larger number of smaller microspheres which expanded evenly
and therefore filled the surface more uniformly. Smaller expanded microspheres also adhered better to the
surface of the printing substrate.

ADM ADM-1L ADM-2L

MFSR MFSR-1L MFSR-2L

Figure 2. Surface of printed braille dots with ADM and MFSR printing ink coated with one (1L) or two (2L)
layers of water-based varnish (SEM; 70× magnification).

8th Conference on Information and Graphic Arts Technology 227

 Poster section

Figure 3. Printed sample in braille with selected text.

Printed braille samples were tested with blind and visually impaired, which gave us the feedback on readability
and tactile recognition of prepared braille prints obtained with both selected printing inks. All interviewees
were able to read the printed text though the readability was slightly worse it the case of ADM printing ink.
They liked the smooth and soft surface of prints and no one detected the impressions of the screen mash in
braille dots. Final conclusion was that the samples, coated with one or two layers of varnish enabled smooth-
er and easier tactile fingertip transitions during printing where even the smoother surface of the printing
substrate was recognized.

4 CONCLUSIONS

Results have shown that with both selected special 3D expanding screen printing inks ADM and MFSR we
can achieve very good results – adequate height and sharp shape of the printed braille dots and suitable tac-
tile impressions of the printed surface. Interestingly, higher prints of braille dots were achieved with MFSR
printing ink, which was applied in only one and not two layers as ADM. Therefore, we can conclude, that by
using this kind of 3D expandable printing ink with smaller microspheres we can enable the lowering of the
printing costs and spent time of production. Beside mentioned we have concluded that using varnish as a
coating layer increases the durability of prints. Testing in practice with blind and visually impaired showed
that the tactile surface seemed smooth despite of its rough surface shown (impressions of screen mash in
prints) in image analyses.
Results of the research have proved that with the use of modified conventional screen printing and right se-
lection of special screen printing ink we can enable simple, fast and primarily economic printing of braille,
that is adequate to prints produced by embossing, and satisfying to touch.

5 REFERENCES

• Achitex Minerva (Italy): Minerfoam SR – Saftey data sheet, 3. 5. 2017.

• Achitex Minerva (Italy): Admiral – Saftey data sheet, 9. 6. 2017.
• Braille, International Council on English. The rules of united English braille (2013). Available from: <http://
www.iceb.org/Rules%20of%20Unified%20English%20Braille%202013.pdf>; Accessed: 2018–1–26.
• Chang, Y. 2012. “Braille ink-jet printing apparatus and printing method thereof.” Patent WO 2012163212 A1.
• Deutsches Institut für Normung e. V. 2009. “Packaging – Braille on packaging.” DIN 55561.
• Fajdetić, A. Standardisation of Braille in the EU and other European Countires. Available from: <http://bib.
irb.hr/datoteka/551906.Standardization_of_Braille_in_EU_and_other_European_Countries.pdf>; Accessed:
2015–6–29.
• Kron, A., Nordin, O., Nilsson, S. and Berglund, C. 2004. “Microspheres”. EP 1 592 733 B1: 2–6.
• Kukec, K. 2015. “Printing braille with screen printing and expanding inks.” BSc diss., University of Ljubljana.
• Slovenski inštitut za standardizacijo. 2014. “Embalaža – Braillova pisava na embalaži za zdravila.” SIST EN
17351:2014.
• Urbas, R., Rotar, B., Hajdu, P. and Stankovič Elesini, U. 2016. “Evaluation of the modified braille dots printed
with the UV ink-jet technique.” Journal of Graphic Engineering and Design, vol. 7, no. 2, pp. 15–24.
• Zhang, Y. 2011. “Inkjet printing device for braille.” Patent CN 202071517 (U).

8th Conference on Information and Graphic Arts Technology 228

 Poster section

Corresponding author:
Raša URBAS
University of Ljubljana, Faculty of natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 83, e-mail: [email protected]

Co-author(s):
Karin KUKEC
University of Ljubljana, Faculty of natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Snežniška 5
1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 229

 Poster section

SCIENTIFIC ILLUSTRATIVE DESIGNS IN GRAPHIC PRODUCTS

Tadej PREZELJ, Urška VRABIČ BRODNJAK

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Aškerčeva 12, SI-1000 Ljubljana

ABSTRACT: In this research, the definition of illustration, presentation of its characteristics and wide range
of usage are presented. It was important to briefly describe the development of illustration in Slovenian space
and to mention its most influential illustrators (Gaspari, 2017).
Because the brand creates the visual presentation of their products or services, we can witness realistic illus-
trations being a better choice, compared to photographic material. In conclusion, illustrations of diverse styles
can drastically improve the aesthetic look and recognisability of the product. The aim of our research was to
present the wide usage of illustration, meaning and influence of illustrative design in graphic products and
design a graphic image presented on an educational product, with the addition of realistic illustrations. The
end result is a graphic image product with the addition of realistic illustrations of a selected animal as well
as an interactive prototype design.

Keywords: illustration, scientific illustrations, design, reindeer, graphic product.

1 INTRODUCTION

Quality illustrations are meant to convince the spectator in this case that they are more informative and
articulate than photographs. Graphic products appear as accompanying objects in our everyday life, each dif-
ferentiates from another in purpose of usage, physical and graphic design (Rupel, 2013). Beside the product’s
quality services, graphic design or visual appearance also play a very important role. Addition of illustrations
in graphic image can most certainly provide or enhance its uniqueness and recognisability. As with other
forms of painting or art, visual arts are also important in illustration: line, color, light and shadows, shapes,
structure, typography and composition. A quality illustration can be very simple or complex, drawing or paint-
ing, realistic or abstractly colored. Therefore illustration of an author differs from the illustrations of other
authors, although they are of the same motive and areas. The wide usage of illustrations in graphic products
were researched and were later narrowed down to usage of only realistic illustrations.
Realistic illustrations are often found on products for educational purposes, but can also appear in the role
of advertisement on products in various markets (Gošnik Godec, et al., 2005). That can often appear on pack-
aging’s, mostly on packaging of food industry products (Hodges, 2017; Poynor, 2017).
The practical part covers the process of designing a board prototype as an interactive graphic product for client
ZOO Ljubljana, Slovenia, with the addition of illustration with an educational role for a selected animal-rein-
deer. The process from the design of concept up to final prototype product version is described and presented.
An extensive research regarding the selected animal was necessary for accurate illustrations. At the same time
it was important to select the right material for the printing quality, process and for the interactive prototype.

2 EXPERIMENTAL

Traditional drawing included the largest part of our research. A series of illustrations had to be made
until the illustrations, which could be suitable for use, has been made. Illustrations were made on pho-
to, colored carton with Faber Castell color pencils. After the selection, the illustrations had to be digitized
and processed (Adobe Lightroom, Adobe Photoshop and Adobe Illustrator) for proper use on the graph-
ics products. The final part of the task was also aimed at the graphic design of the entire product, which
besides the illustrations also contains the necessary graphic elements that gave the product a whole.
Since several illustrations of the reindeer were produced in the process, a thematically complementary product
was also produced, with general educational information content.

3 RESULTS & DISCUSSION

In cooperation with the ZOO Ljubljana, the concept of a graphic product, where the included illustrations
would have the main importance for education, have been designed. The purpose of the work was to create
a prototype of an interactive table as a graphic product that the ZOO will need for the animal – reindeer

8th Conference on Information and Graphic Arts Technology 230

 Poster section

introduction. Initially, it was necessary to design and agree on the conceptual concept of the board, which
would provide the audience with interactivity and information of an educational nature. Figure 1 and 2 present
detailed illustrations of some body parts of the reindeer.

      Figure 1. Illustration of a reindeer eye.      Figure 2. Illustration of a reindeer leg.

It was necessary to understand the proportions and correct the reindeer from the aspect of anatomy. With
the colorings, it was necessary to give the appropriate texture to the reindeer, and at the end, notice the dif-
ference between the sharper horns and the softer figure of the animal. Figure 3 presents the illustration of
whole animal, after several studies.

Figure 3. Illustration of the reindeer.

The end graphic product, with reindeer illustrations is complemented by a quarter-cut round plate (Figure
4), which had to be adequately graphically solved. The panel and the end product was written in Slovenian
language due to demands of the ZOO.

Since the panel covers a large area, an additive with an inadequate solution could spoil the product or reverse
the illustrations of the bottom panel. Since the panel rotated, the only option was, that the panel would not
be empty, using a texture. In order to feel a separate surface immediately, there is a slight contrast in color
between the plates. The pattern formed is in a hazy blue and brown color, enough to give a slight contrast to
the lower plate. The pattern on the panel is consequently in circular shapes, and an arrow is added to supple-
ment the product, which gives information about the rotation and its direction as a clockwise direction. The
shape of the panel had to be cut with the cutting machine through the contour.

8th Conference on Information and Graphic Arts Technology 231

 Poster section

Figure 4. Graphic product-prototype with included illustrations in Slovenian language.

4 CONCLUSIONS

For the Ljubljana ZOO Ljubljana we have created a prototype of an educational interactive board that will
provide visitors an information, for educational purpose about their animals. Because the illustration is a type
of fine arts, the inclusion of realistic illustrations of the selected animal gives the product a higher value. For
the entire concept of the product, we started with the client in the foreign literature and produced a prototype,
where the graphic elements complement the product, give it interactivity and illustrations to give the product
higher value. The prototype is made of corrugated cardboard and plate. The graphic image in the cardboard is
laminated on the substrate as a label, but on a plate it was printed with digital printing technique.
We have produced a thematically complementary product, offering a general education on the selected an-
imal, thus providing the basis for the training of specific characteristics and it’s presented with the product
of the horn cycle.

ACKNOWLEDGEMENT
The authors would like to thank ZOO Ljubljana for cooperation.

5 REFERENCES

• Gaspari, M. SloArt. URL: http://www.sloart.si/m-52-gasparimaksim.aspx (last accessed on 15. 12. 2017)

• Gošnik Godec, A., Manček, M. and Čoh, Z. 2005. Album slovenskih ilustratorjev. Ljubljana: Mladinska knjiga.
• Hodges, E.R.S. Science illustration: Adaped from the Guild handbook of science illustration. URL: https://
gnsi.org/science-illustration (last accessed on 15. 12. 2017).
• Poynor, R. PrintMag: The missing critical history of illustration. URL: http://www.printmag.com/article/
the-forgotten-history-of-illustration (last accessed on 10. 11. 2017).
• Rupel, N. 2013. »Ilustracija: Od grafike h grafiki. Avtorska slikanica : diplomsko delo«, University of Ljubljana.

Corresponding author:
Urška VRABIČ BRODNJAK
Department of Textiles, Graphic Arts and Design
Faculty of Natural Sciences and Engineering, University of Ljubljana
Snežniška 5
SI-1000, Ljubljana, Slovenia
phone: +386 1 200 32 89, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 232

 Poster section

SELECTING THE TRADEMARK COLOUR DEPENDING

ON THE COMPANY’S BUSINESS AREA

Ivan PINĆJER, Ivana TOMIĆ, Irma PUŠKAREVIĆ & Gojko VLADIĆ

University of Novi Sad, Faculty of Technical Sciences, Department of graphic engineering and design

ABSTRACT: The paper investigates the influence of colour on a trademark design in regards the company’s
profile or business area. The aim was to assess the influence of colour on consumer’s attitude and to test which
colour best matches particular business area. The trademark design should meet certain criteria in accordance
with the company’s commercial milieu. Different colours have different effects and, therefore, only appropri-
ately defined colour should be used for specific company trademark. Hence, tests in this study were designed
to help determine which colour is appropriate for a particular area of business.

Keywords: colour, trademark, business area, semantics.

1 INTRODUCTION

A trademark is a noticeable sign used by an individual, business organization or other legal entity to identify
their products and distinguish them from competitors. The trademark should visually illustrate the charac-
teristics of the company it represents.

In all areas and types of graphic design, the trademark is one of the most complex projects, because the idea
and the graphic-artistic solution itself should include many requirements in a simple, concise and sublimated
form.

One way of differentiation of the trademark in relation to the other is the use of colour. Colour has the power
to create an appropriate emotion. For thousands of years, colour plays an important role in the creation of a
certain state of mind and its balance. For example, learning and treatment environments are no longer filled
with cold colours; now they are mostly coloured with colours that stimulate the mind.

As a powerful form of communication, the colour is irreplaceable. By using different colours for a product or
logo, different reactions can be caused. Colours can send subliminal messages, which play a key role in success
or failure of the product.

The problem of selecting appropriate colours, depending on, and in accordance with the activities of the com-
pany, can be resolved by examining the psychological effects of colour on consumers. The colour experience
varies depending on the age of the respondents, so it has been decided that the focus groups should include
students, as more research has shown that the colour affects young people more than the elderly. This is es-
pecially true when it comes to marketing and the choice of colour for a product. The business fields covered
by this research are medicine, electronics, transport, construction and food industry. For the colour palette,
primary and secondary additive colour mixing were selected.

2 EXPERIMENTAL

In order to examine theoretical claims in the modern context, an internet survey in the form of a simple web
application was conducted. The test contained a trademark design, a company name and a colour palette.
The company’s area could be determined by its name, while the trademark design was such as not to reveal
any association. The respondents were instructed to observe black and white trademarks and determine their
colour from the pallet in accordance with their preferences.

The business areas covered by this research are medicine, electronics, transport, construction (civil engineer-
ing) and food industry. The colours used are the red, purple, blue, green and yellow-orange, covering the
primary colour range. Each colour is offered in three shades. The font used for all trademarks is also simple,
without any additions that could be associated with the business area, so that the focus is exclusively on colour.

8th Conference on Information and Graphic Arts Technology 233

 Poster section

There were 130 participants in the experiment. Out of 130 results, 123 were processed, as 7 of the participants
provided contradictory results. Of the 123 participants, 64 were graphic design students, while 59 had no back-
ground in graphic design. As colour has a greater impact on young people, the focus group was respondents
aged 20 to 25.

Figure 1. The appearance of the test that was used in the survey.

As shown in Figure 1, the names of the companies are changed, while the logo and font remain unchanged.

Figure 2. Different trademarks for business fields:

medicine, electronics, transport, construction and food industry.

3 RESULTS & DISCUSSION

First, graphics with all colours are displayed individually, so that it can accurately see which colour prevailed
for a particular business area. After that, the shades of the same colour were grouped.

Figure 3. Results for the area of medicine. Figure 4. Results for the area of electronics.

Figure 5. Results for the area of transport. Figure 6. Results for the area of construction.

8th Conference on Information and Graphic Arts Technology 234

 Poster section

Figure 7. Results for the area of the food industry.

Based on the results of all the tests, the following results were obtained: for medicine, the association is red, for
electronics blue and green, for transport blue, for the construction yellow-orange, for the food industry orange.

The results of the tests carried out by persons engaged in or related to the graphic design are as follows: for
medicine red and blue, for electronics blue and green, for transport blue and green, for the yellow-orange
construction, for the yellow and orange food industry. The results of tests performed by participants who
are not related to the design are: for medicine - red, for electronics - blue and green, for transport - blue and
yellow, for the construction - yellow-orange and yellow and orange for the food industry.

Figure 8. Answers of the participants related to Figure 9. Answers of the participants with no
design for the field of medicine. background in design for the field of medicine.

Figure 10. Answers of the participants related Figure 11. Answers of the participants with no
to design for the field of electronics. background in design for the field of electronics.

8th Conference on Information and Graphic Arts Technology 235

 Poster section

Figure 12. Answers of the participants related Figure 13. Answers of the participants with no
to design for the field of transport. background in design for the field of transport.

Figure 14. Answers of the participants related Figure 15. Answers of the participants with no
to design for the field of construction. background in design for the field of construction.

Figure 16. Answers of the participants related Figure 17. Answers of the participants with no
to design for the field of food industry. background in design for the field of food industry.

4 CONCLUSIONS

Based on the obtained results it can be concluded that certain colours are associated with the certain business
area by respondents. If more than one colour is needed for the trademark design, results are showing what
those colours are. Also, the results show the difference between two groups of respondents. In choosing the
appropriate colour for the trademark design for the specific business area of the company, designer responded
differently than respondents with no graphic art background.
Colour is the basic element of many forms of communication, and a good designer needs to take full advan-
tage of the possibilities that this huge colour range offers, to do unique works. Shape along with colour can
represent a strong message. Each colour has its character and causes different associations, so sometimes
even small errors in choosing the right colour can cause negative results. The same shape, trademark or logo,
when displayed in different colours, can cause different effects with the viewer. The choice of colour should
be carefully studied and determine which colours can be used in the certain business area, and which not.

8th Conference on Information and Graphic Arts Technology 236

 Poster section

ACKNOWLEDGEMENT
The research is supported by the Ministry of Education, Science and Technology Development of the Republic
of Serbia, project number: 35027 “Development of software model for scientific and production improvement
in graphic industry”.

5 REFERENCES

• Akcay, O., Dalgin, H., Bhatnagar, S. 2011, “Perception of Color in Product Choice among College Students:
A Cross-National Analysis of USA, India, China and Turkey”, International Journal of Business and Social
Science, Vol. 2 No. 21 [Special Issue-November].
• “Intellectual Property Flashcards | Quizlet.” Insert Name of Site in Italics. N.p., n.d. Web. 08 Apr. 2018 <https://
quizlet.com/222950876/intellectual-property-flash-cards/>.
• Klausbernd, V. 2005. „Das große Buch der Farben“ (Königsfurt Verlag), Krummwisch
• Krippendorff, K., 2000. “Propositions of human-centeredness; A philosophy for design”, In D. Durling & K.
Friedman (Eds.), Doctoral education in design: Foundations for the future: Proceedings of the conference,
La Clusaz, France (pp. 55-63)
• Nedeljković S. M. 1998. “Elementi za formiranje grafičkih komunikacija”, Beograd: Zavod za udžbenike i
nastavna sredstva
• Skinner, V. J. 2001. “Cashing in on the “Simple Magic” of Color”
• Stone, T., Morioka, N., Adams, S. 2006. “Logo Design Workbook: A Hands-On Guide to Creating Logos”,
Gloucester, Rockport Publishers

Corresponding author:
Ivan PINĆJER
Department of Graphic Engineering and Design,
Faculty of Technical Sciences, University of Novi Sad,
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2624, e-mail: [email protected]

Co-author(s):
Ivana TOMIĆ
Department of Graphic Engineering and Design,
Faculty of Technical Sciences, University of Novi Sad,
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2626, e-mail: [email protected]

Irma PUŠKAREVIĆ
Department of Graphic Engineering and Design,
Faculty of Technical Sciences, University of Novi Sad,
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2603, e-mail: [email protected]

Gojko VLADIĆ
Department of Graphic Engineering and Design,
Faculty of Technical Sciences, University of Novi Sad,
Trg Dositeja Obradovića 6, 21000, Novi Sad, Serbia
phone: +381 21 486 2622, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 237

 Poster section

SELF-FOLDABLE KNITTED BAGS FOR HOUSEHOLD STORAGE

OF BAKERY PRODUCTS BASED ON BIOMIMETICS

Darja RANT1, Alenka PAVKO-ČUDEN2

1
Klobbka, d.o.o.
2
University of Ljubljana, Faculty of natural sciences and engineering,
Department of textiles, graphic arts and design

ABSTRACT: Textiles have multifunctional food storage potential. They can mechanically protect food against
impact, they are air-permeable, washable, biodegradable, foldable, pleasant to touch and decorative. Knitted
fabrics are distinguished by their stretchability and flexibility, the possibility of seamless production, high
efficiency of production, possible self-folding and the possibility of complex mechanical and chemical func-
tionalization. The use of weft knitted self-foldable structures for packaging has been an unexplored research
field. The purpose of the research was to design, produce and test visually interesting, aesthetic and multifunc-
tional weft knitted fabrics usable for household storage of baked goods/breads. Experimental work in design
was inspired by natural foldable and permeable structures and based on the systematic development of 3D,
textured and foldable structures with auxetic potential, made from various materials. The technological exper-
imental work consisted of testing the foldable knitted bags for the ability to preserve the freshness of bread.

Keywords: textiles, knitted fabrics, packaging, biomimetics, foldability.

1 INTRODUCTION

Industrial design including textile design most often draws ideas from nature. The underlying assumption in
is that nature performs a function with the least amount of energy, uses the commonest materials, and is the
most reliable. Biomimetics is a relatively young study embracing the practical use of mechanisms and func-
tions of biological science in engineering, design, chemistry, electronics and so on. (Julian, 2007). Biomimetics
offers competitive advantages to suppliers of materials, processes and components, and to the makers and
brand owners of finished products. Product designers represent one important and welldefined channel for
dissemination (Hollington, 2007).

Figure 1. Beetroot peel - sparingly packaged seeds [Powers, 2000].

Nature’s designs, materials, processes and structures have always inspired packaging. Packaging has a design, a
shape, a structure, a concept, a finish and a decoration or print. Packaging is alongside the product, the driver
to attract consumers. It is the first item the consumer sees, feels, smells, touches and (maybe) tastes. People
are used to natural structures. Nature’s solutions have stood the test of time. (Pointevin, 2007). One of the first
and most critical roles of packaging is to contain the contents. Faced with changing volumes, numbers, sizes
or shapes of items over time, nature has evolved some innovative and adaptive packaging solutions. Flexible
containers are popular in nature since the size or number of the contents isn’t always predictable. Flexible
containers can expand and collapse. Nature therefore optimizes packaging on many scales. For example, the
folding strategies from leaves have been mimiched into deployable structures for folding into the smallest
space possible. These solutions could filter down to consumer application just as origami has influenced Jap-
anese consumer packaging (Jedlička, 2009).

8th Conference on Information and Graphic Arts Technology 238

 Poster section

Efficient packaging materials and technology ensure the proper safety and quality of our food products from
processing and manufacturing through handling and storage and ultimately to consumption. The polymers and
materials used for food-packaging today consist of a variety of petroleum-derived plastic materials, metals,
glass, paper and board, or combinations hereof. With the exception of paper and board, all of these packag-
ing materials are actually based on non-renewable materials, implying that at some point, more alternative
packaging materials based on renewable resources have to be found (Weber et al, 2002).

2 PRESERVATION OF FOOD FRESHNESS

Preserving the freshness of food is an important contribution to healthy nutrition and reducing the amount
of discarded food. Food conservation including packaging is important for preserving the freshness of food.
For household storage of foods, a variety of materials are used in modern times, of which textiles represent a
smaller share. Textiles have multifunctional food storage potential. They can mechanically protect food against
impact, they are air-permeable, thus preventing the formation of molds, they are washable and therefore re-
peatedly usable, biodegradable and thus sustainable, foldable, pleasant to the touch and decorative, i.e. suit-
able for production in different colors, sizes and shapes. Among the textiles, knitted fabrics are distinguished
by their stretchability and flexibility, the possibility of seamless production, high efficiency of production,
possible self-folding and the possibility of complex mechanical and chemical functionalization. Knits have
been mostly used for packaging in the form of nets, produced on warp knitting machines. The use of weft
knitted self-foldable structures for packaging has been an unexplored research field.

Bread is one of the most perishable food products. Its 40% concentration of water leading to a water activ-
ity of 0.96 makes susceptible to mould attacks. The micro-organisms that degrade the quality of the bread
originate in the flour where they get from the cereals opportunist micro-flora, often accompanied by other
microbes arising during milling or flour storing. By means of current preservation technologies and systems,
the optimum duration for bread preservation on shelves is of 2-3 days (Cioban et al, 2010). Storage stability or
the shelf-life of baked products could be defined as maintenance of the sensory and physical characteristics
associated with freshness (Baixauli et al., 2008).

3 EXPERIMENTAL

The purpose of the research was to design, produce on an industrial weft knitting machine and test visually
interesting, aesthetic and multifunctional weft knitted fabrics usable for household storage of baked goods/
breads. Experimental work in design was inspired by natural foldable and permeable structures and based on
the systematic development of 3D, textured and foldable structures with auxetic potential, made from various
materials. The technological experimental work consisted of testing the foldable knitted bags for the ability
to preserve the freshness of bread. Also, a visual analysis of the visually attractive and functionally useful
packaging material for household storage of baked goods/breads was performed.
The suitability of foldable seamless knits for the storage of bread and bakery products has been studied. It
was assessed by testing the antibacterial properties of the selected foldable knitted structures made from
various raw materials.

3.1 Microbiological deterioration testing

First, the soil burial test according to SIST EN ISO 11721-1:2001 standard was performed for determining the
resistances of foldable knitted fabrics made from various yarns to microbiological deterioration. Apart from
a basic single knitted structure, a zigzag links-links knitted structure with 4 × 4 square repeating unit cell
was selected. The selected yarns were made of: 100% combed cotton, 100% cotton with added polyamide
filament, flax/viscose blend, natural bamboo/modal blend, carded 100% cotton, 100% lyocell, 100% viscose,
wool/viscose blend, lyocell/viscose blend, lyocell/cotton blend, cashmere/polyamide/viscose and 100% wool.
The time of exposure was 12 days. Afterwards the knitted samples were carefully removed from the soil. The
samples were rinsed in water, sterilised by soaking in 70% ethanol at room temperature for 30 minutes and
dried. Initially, the deterioration of samples was visually assessed. After that, the rate of biodegradation of
the examined samples was determined by colour measurement with a spectrophotometer SPECTRAFLASH 600
PLUS (Datacolor International, USA) using the CIELAB colour system. ΔL* values of the buried and unburied
samples were determined and compared.

8th Conference on Information and Graphic Arts Technology 239

 Poster section

3.2 Antibacterial properties testing

In order to achieve antibacterial activity, Bioshield Excalibur (Izinova Ltd., Bled) finishing agent was selected,
which is chemically alkyl dimethyl (3-trimethoxysilylpropyl) ammonium chloride. Its antibacterial activity is
based on bio-barrier formation mechanism. Antibacterial activity of the examined knitted samples was esti-
mated by determination of bacterial reduction according to the ASTM E 2149− 01 standard method. Bacterial
reduction of the samples was evaluated against Gram-negative bacteria Escherichia coli (ATCC 25922).

3.3 Bread freshness preservation testing

To study the freshness preservation potential, two foldable links-links structures were knitted: a zigzag struc-
ture with a 24 × 24 repeat having a width of 12 ribs and a zigzag angle of 45°, and a zigzag structure with a
repeat having a width of 2 ribs and the same inclination as the previous sample. The 24 × 24 structure was
well folded while the 4 × 24 structure was less folded and therefore had a lesser auxetic potential. From each
material and structure, three bags were made for bread storage.

White sandwich bread (Pekarna Jurčkova, Ljubljana) was chosen for testing the freshness preservation po-
tential. When it was cooled after baking, it was cut into 2.5 cm wide slices, weighed and stored in prepared
textile bags. Every 12 hours the bread samples were weighed and the percentage of the bread mass reduction
was calculated (Požrl, 2017).

To test the bread freshness change during storage in textile packaging the moisture loss was determined by
weighing. For packaging, materials with foldable ability and good antibacterial properties were selected: 100%
liocel, liocel/cotton blend, bamboo/modal blend, flax/viscose/elastane blend and 100% bamboo cellulose
viscose. In spite of good folding ability and good antibacterial properties, keratin fibers have been discarded
as the reusable textile food packaging is expected to be repeatedly washable at higher temperatures while the
keratin materials do not meet this requirement. For comparison, two conventional materials in which bread
is stored in households were selected: a plastic bag and a sandbag paper, both from Drogerie Markt. From the
testing results it is apparent that bread keeps the softness and humidity the longest if it is stored in a polyvinyl
bag, but after two days it is possible to notice a change in the surface of the core. Paper bags give the worst
results when storing bread because the bread is drastically dried out.

3.4 Product design

3D-knitted packaging bags have an attractive, full/compact look and good folding ability. In the folded and
passive state, the bag takes up little space, especially in comparison with the bread box. Furthermore, it is
decorative, washable and flexible. In the open and active state, it is well-adapted to its content regarding shape
and size. The links-links weft knitted structure allows the redistribution of the volume is. Due to its ability to
adapt to the content size and shape, the image of the packaging bag changes all the time.

4 RESULTS & DISCUSSION

The results of the microbiological deterioration study showed that the ΔL* values of the foldable knitted
structures were lower than for the single knitted structures. The visual assessment of the samples also showed
that in most cases, the foldable knitted structures were less deteriorated than the single knitted structures. The
results proved that the foldable links-links knitted structures have better microbiological resistance compared
to the single structures.

According to the results, foldable links-links knitted samples from flax/viscose blend, natural bamboo/mod-
al blend, lyocell/viscose blend, 100% lyocell, cashmere/polyamide/viscose and 100% wool were selected for
further investigation.

The results showed that the selected antibacterial finish was fully effective, reflecting in complete growth
reduction of the tested bacteria. On the other hand, untreated foldable links-links knitted samples exhibited
rather important differences in the reduction of E. coli growth. The samples made from lyocell/cotton blend
exhibited 55% reduction, the reduction of cashmere/polyamide/viscose blend was 11%, while the reduction
of the rest of the samples was less than 10%. Better results for the samples made from natural bamboo/modal
yarn were expected due to the original antibacterial properties of the natural bamboo fibres.

8th Conference on Information and Graphic Arts Technology 240

 Poster section

Textile bags, made from cellulose fibers, showed almost the equivalent results: they were better than paper
bags and are much worse than plastic ones in bread freshness preservation. The tested bread dried the most
during the first 24 hours, then the drying process slowed down. The results indicated that the bread was the
least dried out in the bag made of 100% bamboo viscose. Both structure exhibit minimal moisture-retaining
potential, less than 1%.

Figure 2. Knitted bag for bread storage in passive state (folded and empty).

Figure 3. Knitted bag for bread storage in active state (unfolded and full).

5 CONCLUSIONS

Foldable knitted structures are multifunctional and widely usable. They can be produced in a variety of struc-
tures, qualities and dimensions: in panels, fully-fashioned or seamless. They exhibit a supreme aesthetics
and have a big potential for the use in multiple areas. Some of the foldable knitted structures exhibit auxetic
properties which have lately become a subject of extensive research. Foldable knitted structures, links-links
knits among them, can be considered a promising development line of sustainable hi-tech knitting technology
and design, especially if combined with other technologies. The development of sustainable, re-usable and
up-cyclable, genuine self-folding knitted collapsibles should be encouraged.
All tested cellulosic textile materials are suitable for bread storage. Taking into account biodegradability and
antibacterial properties, it would be reasonable to make bread bags for household bread storage from a mix-
ture of lyocell/cotton or from 100% lyocel yarn. Due to the foldability which enables the adoption of shape
hape of the packaging to its content, it is recommended to use the foldable links-links zigzag structure with
a larger square repeat.

6 REFERENCES

• Baixauli R., Salvador A. and Fiszman S.M. 2008. Textural and colour changes during storage and sensory shelf
life of muffins containing resistant starch. European Food Research and Technology 226: 523-530.
• Cioban, C., Alexa, E., Sumalan, R., Merce, I. 2010. Impact of packaging on bread physical and chemical prop-
erties. Bulletin UASVM Agriculture. 67 (2): 212-271.
• Hollington G., Biomimetics and product design. In: Biomimetics : strategies for product design inspired by
nature – a mission to the Netherlands and Germany. Report of a DTI global watch mission, Department of
Trade and Industry, 2007. URL: http://www.catedrasimonetti.com.ar/ attachments/article/278/ Biomimet-
ics_report_final_version[1].pdf. (last accessed on 30.01.2018).

8th Conference on Information and Graphic Arts Technology 241

 Poster section

• Jedlička, W., 2009. Innovation toolbox. In: Jedlička, W. Packaging sustainability: tools, systems and strategies
for innovative package design, 267-317. Hoboken: John Wiley & Sons.
• Julian V., Background to miomimetics. In: Biomimetics : strategies for product design inspired by nature – a
mission to the Netherlands and Germany. Report of a DTI global watch mission, Department of Trade and
Industry, 2007. URL: http://www.catedrasimonetti.com.ar/ attachments/article/278/ Biomimetics_report_fi-
nal_version[1].pdf. (last accessed on 30.01.2018).
• Pointevin P., Examples of biomimetic applications: biologically inspired packaging. In: Biomimetics : strat-
egies for product design inspired by nature – a mission to the Netherlands and Germany. Report of a DTI
global watch mission, Department of Trade and Industry, 2007. URL: http://www.catedrasimonetti.com.ar/
attachments/article/278/ Biomimetics_report_final_ version[1].pdf. (last accessed on 30.01.2018).
• Powers, A., 2000. Natur und Design : Inspirationen für Architektur, Mode und angewandte Kunst. Bern;
Stuttgart; Wien : P. Haupt, 160 str.
• Požrl, T., 2017. Personal communication, Ljubljana, 21.6.2017.
• Webery, C. J., Haugaard, V., Festersen, R. and Bertelsen, G. 2002. Production and applications of biobased
packaging materials for the food industry. Food Additives and Contaminants. 19, Supplement: 172-177.

Corresponding author:
Alenka PAVKO-ČUDEN
Department of textiles, graphic arts and design
Faculty of natural sciences and engineering, University of Ljubljana
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 00, fax: +386 1 200 32 70, e-mail: [email protected]

Co-author(s):
Darja RANT
Lobbka, d.o.o., Spodnja Luša 25, 4227 Selca, Slovenia

8th Conference on Information and Graphic Arts Technology 242

 Poster section

STUDY OF ACCELERATED AGING OF BLACK INKJET PRINTS

Michal ORAVEC, Michal ČEPPAN & Patrícia PARAČKOVÁ

Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers, Radlinského 9, Bratislava, Slovakia

ABSTRACT: The purpose of this work is the study of accelerated aging of prints. This work is focused on the
study of the fast and non-destructive approach to investigate the spectral changes of black inkjet prints during
heat accelerated aging, which is currently interested from the point of view of forensic analysis of questioned
documents. In this work, we analyzed the set of 10 accelerated aged samples printed using the office printer
on the office paper. The process of accelerated aging was performed for 29 days using glass bottles closed with
the Viton® sealing ring in drying oven heated to 100°C according to the ISO 5630 – 1. The samples were meas-
ured using Near Infrared (NIR) optic fibre device and processed by the Liner Discriminant Analysis (LDA). The
spectra were also measured using Fourier transform (FT) Raman device. The FT-Raman spectra were interpreted
separately with regard to observing the changes and trends of spectra shapes and slopes.

Keywords: NIR spectroscopy, Forensic science, Chemometry, Document.

1 INTRODUCTION

Ubiquitous a large number of printed documents that have to be examined in forensic to detect potential
falsifications requires the fast and reliable approach to reveal the origin of the questioned documents. For
this reason, numerous destructive and non-destructive methods were applied.
The method allowed repeatable and reliable investigation of black inkjet inks extracted from printouts were
realized using capillary electrophoresis coupled to mass spectrometry with electrospray ion source and time
of flight analyses (CE-ESI-TOF-MS) (Kula, 2014). However, for forensic aspect the non-destructive approach
is required since the document in question can be analyzed by using other methods as key evidence. Non-de-
structive techniques like NIR spectroscopy are usually preferred since they are simpler to perform and usually
require minimal manipulation of the analyzed material to preserve the integrity of the document in question
(Materazzi, 2017). Infrared spectroscopy proved to be a useful technique in forensic science (Ferrer, 2017). The
well-known chemometric methods play an important role in the forensic purpose of sample analysis. In par-
ticular, the NIR spectroscopy coupled in tandem with chemometrics represent the ability to study substances
for qualitative and quantitative examination of the matrices (Gál 2015), (Reháková, 2017). The suitability of
FT-Raman spectroscopy to investigate some black inkjet prints was described in (Oravec, 2018).
This work is focused on the non-destructive approach of spectral examination, which is currently considerably
required, in forensic analysis of questioned documents.

2 EXPERIMENTAL

NIR measurements
All NIR reflection spectra were measured with an Ocean Optics fibre optics spectrophotometer consisting of Hi-
Res spectrometer NIR 256-2.5 spectrometer with NIR light source HL-2000-FHSA (Ocean Optics, Inc. Dunedin,
FL USA). The spectra were acquired in the range 850 – 2630 nm. Each sample was measured 9 times. The dataset
was embedded in the software The Unscrambler X 10.5 (CAMO Software AS., Oslo, Norway) (Unscrambler) for
following data processing and design LDA.

Liner Discriminant Analysis (LDA)

The NIR spectroscopy was combined with LDA method to classify the samples based on the time of accelerated
aging. The 9 spectra of each sample were divided into Calibration set (Cal) and Validation set (Val) in the ratio
of 6:3. For each sample, 3 separate LDA models were created. The selection of spectra into Val was performed in
3 different manners: the Val 123 represent first 3 measurements of the 9 in total. Val 456 represents the group of
next 3 acquired spectra and Val 789 comprise the remaining 3 measurements. The Cal was varied respectively.
This variation was done to improve statistical significance of the experiment.

8th Conference on Information and Graphic Arts Technology 243

 Poster section

FT-Raman measurements
FT-Raman spectroscopy of 10 samples was performed to study differences in the shapes and trends of the spec-
tra. The spectra were acquired using Bruker FT-Raman Spectrometer MultiRAM (Bruker Optics Inc. Billerica,
MA USA) with FT-Raman microscope. The device was equipped with Nd:YAG laser, diode pumped by P = 1000
mW. All spectra were acquired using device adjustment to the power of 1000 mW, 64 scans and resolution of
4 cm-1. The measuring was performed in the region 50 – 3800 cm-1. The FT-Raman spectra were acquired with
the OPUS 6.5 software (Bruker Optics Inc., Billerica, MA USA). The spectra were treated using manual baseline
correction with OriginPro (OriginLab Corporation, Northampton, MA USA), (Origin). The two parameters of
the most prominent peak (around 1600 cm-1) were recorded: the Full Width at Half-Maximum (FWHM) and
the peak intensity of the band.

Sampling
In this study were investigated 10 prints of the same office printer EPSON L210 (the ink supplies reference is
C13T66414A) on the same office paper (80 g/m2). The printer was set up on normal quality and the black ink
was used only. Each of the samples consists of 3 completely filled equal squares (Fig. 1) with dimensions of
1.5 cm x 1.5 cm.
The accelerated aging process was performed using closed glass bottles with Viton® sealing ring in drying
oven heated to 100°C according to the ISO 5630 – 1. Before the samples were placed into the drying oven, they
were treated with the conditioned atmosphere (T = 24°C and RH = 50%) for 24 hours. In this way, 10 samples of
the different aging time were prepared. Heat treatment of the samples was carried out in constant conditions:
relative humidity and temperature during 1, 3, 6, 8, 10, 13, 16, 20, 24 and 29 days. The names of the samples
were derived from days of aging i.e. sample E1 was 1 day accelerated aged, E3 3 days respectively.

Figure 1. The view of a print file printed on the office paper.

The experiment was focused on 3 completely filled squares.

3 RESULTS & DISCUSSION

With the aim of classifying the black inkjet prints with different accelerated aging time, the analytical strat-
egy consisting of the chemometric LDA analysis of NIR reflection spectra and the interpretation of FT-Raman
spectra.

LDA analysis of NIR reflection spectra

The 90 NIR spectra were embedded into The Unscrambler to realize LDA. The aim of this method was to classify
spectra from Val in the model built from Cal for each of 10 samples. The results of LDA are summarized in the
table (Table 1). If the Val spectra of the examined sample were in accordance with the model (all 3 spectra of Val
were right classified) than the sample was marked using “Y” symbol. If all 3 Val spectra were wrong classified,
then the examined sample was marked using “N” symbol. The average accuracy of right response was 45.6%.
As follows from the Table 1, LDA model is not reliable enough to classify the NIR spectra of aged inkjet prints.

8th Conference on Information and Graphic Arts Technology 244

 Poster section

Table 1. The interpretation of LDA classification.

Sample Val 123 Val 456 Val 789
E1 - Y -
E3 Y - N
E6 - N -
E8 - Y Y
E10 N N -
E13 N N N
E16 Y Y Y
E20 N - N
E24 - - Y
E29 - N N

FT-Raman method
The FT-Raman spectra of the aged prints without baseline correction are on Figure 2 and the spectra with
manual baseline corrections are on Figure 3.
The trend of intensity during aging is not certain in both cases: neither in untreated nor in baseline corrected
spectra (Figure 2, 3).

Figure 2. The FT-Raman spectra of E1, E8, E20 and E29 samples. The band of 1590 cm-1 represents
the most prominent peak of the spectrum and belongs to the G-modes of a disordered carbon.

Figure 3. The comparison of FT-Raman spectra after manual baseline correction.

8th Conference on Information and Graphic Arts Technology 245

 Poster section

The investigation of prints aging using FT-Raman spectra is based on the properties of a relatively significant
band of the spectrum. This strong Raman band at about 1590 cm-1 can arise from the G-modes of a disordered
carbon. The plots of intensity and FWHM (band of 1590 cm-1) against the time of accelerated ageing are on
Figure 4.
The trends in both plots are not monotonic, and hence they do not allow direct classification of the unknown
aged sample. However, in the combination of these parameters can be useful to distinguish which one the
model sample is the best match for the unknown sample.

Figure 4. The intensity and FWHM of the baseline corrected FT-Raman spectra of prints during ageing.

4 CONCLUSIONS

This work is focused on the non-destructive approach of spectral examination of aged inkjet prints.
This investigation shows the interpretation of acquired spectra: the NIR spectroscopy combined with su-
pervised classification method LDA and FT-Raman two parameters identification. In general, the trends of
the spectra are not unambiguous. However, the potential to classify the samples based on similarities with
the model are presented. The FT-Raman spectroscopy in tandem with NIR spectroscopy combined with LDA
could be considered for applications in the forensic analysis of documents in question when the conventional
methods are not effective.

5 REFERENCES

• Ferrer. N. 2017 Forensic science, applications of IR spectroscopy: Reference module in chemistry, molecular
sciences and chemical engineering encyclopedia of spectroscopy and spectrometry (third edition)., pp.
695-706.
• Gál, L., Oravec, M., Gemeiner P., and Čeppan, M. 2015. “Principal component analysis for the forensic dis-
crimination of black inkjet inks based on the Vis–NIR fibre optics reflection spectra.” Forensic Science In-
ternational 257: 285-292.
• Kula, A., Król M., Wietecha-Posłuszny R., Woźniakiewicz, M. and Kościelniak, P. 2014. “Application of CE-MS
to examination of black inkjet printing inks for forensic purposes.” Talanta 128: 92-101.
• Materazzi, S., Risoluti, R., Pinci, S. and Romolo, F. S. 2017. “New insights in forensic chemistry: NIR/Chemo-
metrics analysis of toners for questioned documents examination.” Talanta 174: 673-678.
• Oravec M., Sasinková, V., Tomanová, K., Gál, L., Parciová, S. and Huck, C. W. 2018 “In-situ surface-enhanced
Raman scattering and FT-Raman spectroscopy of black prints.” Vibrational Spectroscopy 94: 16-21.
• Reháková M., Gál L., Belovičová, M., Oravec, M., Dvonka, V., Stojkovičová, D. and Čeppan, M. 2017. “Identi-
fication of iron-gall inks in historical drawings by Fibre Optics Reflection Spectroscopy – Extension to the
NIR spectral range” Journal of Cultural Heritage 27: 137-142.

Corresponding author:
Michal ORAVEC
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology
Radlinského 9, 812 37, Bratislava, Slovak Republic
phone: +421 2 59 325 224. fax: +421 2 52 493 198, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 246

 Poster section

Co-author(s):
Michal ČEPPAN
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology
Radlinského 9, 812 37, Bratislava, Slovak Republic
phone: +421 2 59 325 224, fax: +421 (2) 52 493 198, e-mail: [email protected]

Patrícia PARAČKOVÁ
Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology
Radlinského 9, 812 37, Bratislava, Slovak RepublicPostal code, city, State
phone: +421 2 59 325 224, fax: +421 (2) 52 493 198, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 247

 Poster section

STUDY OF CHANGES OF WRITING INKS ON THE PAPER DOCUMENTS

Silvia STAŠKOVÁ, Milena REHÁKOVÁ & Aleš HÁZ

Slovak University of Technology in Bratislava, Faculty of Chemical and Food Technology,
Institute of Natural and Synthetic Polymers, Radlinského 9, 812 37, Bratislava, Slovak Republic

ABSTRACT: Examining of writing inks is a part of analysis of documents which could help forensic experts
to detect a crime. Nowadays many researches are focused on studying ink changes on the paper documents,
because falsification of documents is becoming more common and more frequent.
The most common volatile solvent in ballpoint pens is 2-phenoxyetanol (2-PHE). At the moment when the
ink is deposited on the paper, alcohol starts to evaporate. The amount of 2-PHE in the extract is therefore
dependent on the application time on the surface. Ink extracts can be analysed using a combination of gas
chromatography with mass spectrometry (GC–MS). The time of creation of written record was identified by
comparing the amount of 2-PHE in fresh samples and samples subjected to artificial aging.

Keywords: Writing inks, document, forensic examination, evaporation, 2-phenoxyethanol.

1 INTRODUCTION

Falsification and modifying of writings in documents – personal documents, checks, agreements and bills are
part of economic crime.
The analysis of the ink composition as part of documents for criminological and forensic purposes is dating
to 1988. In this year, the chemical analysis of the ink was the first time used on the questioned document as
evidence in the court in the State of New York. This interest in examining ink naturally continued to evolve
with the development of writing instruments. Analysing the age of the questioned document based on an
examination of the inks present very complex process (Parciová, 2016). The degradation of the ink begins im-
mediately after the ink has been applied to the substrate. Successive chemical reactions dependent also on
the composition of the substrates and specially on the storage conditions (humidity, air flow, light, exposure
to air pollutants, etc. (Menzyk, 2015).
Examining of writing inks is a part of analysis of documents which could help forensic experts to detect a
crime. Nowadays many researches are focused on studying ink changes on the paper documents, because falsifi-
cation of documents is becoming more common and more frequent. Analysis of individual ink components can
determine if the record has changed, or if the period of the creation of document has modified (Jones, 2013).
Anyway, three different actions are in progress during the degradation. The first action is a migration of dyes
or pigments, the second one is a polymerization of resin and the third process is the evaporation of the vol-
atile solvent. When writing ink is placed on a substrate, a drying process begins. This process is dependent
on the composition of the ink, paper and on the storage conditions – humidity, light and thermal conditions
(Cantú, 2012).

2 EXPERIMENTAL

Chemicals and materials

Samples preparation
Blue ballpoint pens used in the experiment were purchased in SEVT, a.s. Slovakia. The width tip of each pen
was 0.5 mm. Inks were applied onto office paper in the form of a thin lines. The paper: First class (80 g/m2),
Mondi SCP a.s. (Ružomberok, Slovakia). The samples were stored in the at laboratory temperature (20 – 25°C).
Relative humidity during storage was in the interval 45 – 70%.

Extraction procedure
The experiment tested two series of samples. The first one was fresh written record placed in laboratory condi-
tions (non-aged). The second one was artificially aged (the condition of ageing: 70±5°C for 1 hour). All samples
had form of 5 cm long lines written by ballpoint pen on the paper. They were cut to small pieces and put into
micro-vials (2 ml). Subsequently extraction agents with an internal standard were added into samples. The
extraction agent was acetonitrile (99.8%), Sigma-Aldrich® (St. Louis. MO, USA) and the internal standard was
diphenylamine (DFA, ≥ 99%), Sigma-Aldrich® (St. Louis. MO, USA). Concentration of DFA was 0,0025 mg/ml.

8th Conference on Information and Graphic Arts Technology 248

 Poster section

GC–MS analysis
The solvents were analysed on a gas chromatograph – Agilent Technologies – model 7890A with mass detec-
tor 5975C (MSD) operating in the electron impact mode of ionization. Parameters of the column HP5-MS:
length 30m, internal diameter 250 µm and the film thickness of the stationary phase 0,25 µm. Used thermal
program: initial kiln temperature: 50 °C (±1 °C), heating 15 °C/min to 280 °C, final step: 1 min at a constant
temperature of 280 °C, total length of analysis: 17,33 min, dosage amount: 1µm. The Split-less mode was used
for dosing the sample.

3 RESULTS & DISCUSSION

For a detection of 2-phenoxyetanol in a fill of ballpoint pens it is necessary to know a retention time of the
analysed components. Retention time is the time which flows between the feed of sample and the moment
when the component leaves the column at the maximum concentration.
The chromatographic peak associated with 2-PHE should occur approximately at RT = 7.16 min (± 0,43 min)
and peak for internal standard at RT = 10.65 min (± 0,61 min). These retention times are important because
they are one of the main parameters to indicate the accuracy of the experiment layout (Luczak, 2002). Tables
1 and 2 show the compounds in two sample sets together with data on their retention times and peak areas.

We analyzed 2 sets of samples: lengths of 5 cm and 1 cm. A sample of 5 cm line was analyzed to optimize the
setting of the device anf for better handling. The confirm of the presence of 2-phenoxyethanol was easier to
detect in a larger sample. A sample of 1 cm line is the maximum quantity that can be taken from a document.
At the same time, it seems that is the minimum amount from which it is possible to extract a detectable amount
of 2-PHE. From a real handwriting it is possible to take the sample from various places of the signature in parts
(in the total length of the line 1 cm). Basically, to determine the period of the record creation, it is necessary
to take 2 cm of line on paper because 1 cm is unheated and the other cm is subjected to the artificial aging.

Table 1. Chemical compounds represented in the sample set 1.

The name of the Mole weight
Order RT [min] Peak area 10-6
compound [g/mol]
1 7,10 ± 0,42 8,47 ± 0,12 2-phenoxyethanol 138,07
2 10,59 ± 0,51 0,78 ± 0,04 diphenylamine 169,09

Table 2. Chemical compounds represented in the sample set 2.

The name of the Mole weight
Order RT [min] Peak area 10-6
compound [g/mol]
1 7,26 ± 0,49 3,34 ± 0,03 2-phenoxyethanol 138,07
2 10,70 ± 0,63 0,62 ± 0,01 diphenylamine 169,09

The method of calculation of the percentage dependence R was realized using equations 1-3.

R[%]= ((Pw-Ps)/Pw) 100% (1)

Pw= (Peak area 2 - PHE (non aged))/(Peak area DFA) (2)

Ps= (Peak area 2 - PHE aged))/(Peak area DFA) (3)

On the basis of the above mathematical relations, the relevant parameters Pw, Ps and R were calculated. These
parameters are shown in Table 3 for sample set 1 and in Table 4 for sample set 2.
The process of decreasing the concentration of 2-PHE in one of the aged sample long 5 cm to the non-aged
sample is shown in the Figure 1. The process of decreasing the concentration of 2-PHE in one of the aged
sample long 1 cm to the non-aged sample is shown in the Figure 2.

8th Conference on Information and Graphic Arts Technology 249

 Poster section

Table 3. Characteristic values for sample set 1.

Non aged sample Aged sample
Peak area 10 -6
Peak area 10 -6
Peak area 10-6 Peak area 10-6 Pw Ps R/%
(2-PHE) (DFA) (DFA) (2-PHE)
14,24 ± 0,49 0,65 ± 0,02 0,36 ± 0,01 0,34 ± 0,01 21,98 0,96 95,65

Table 4. Characteristic values for sample set 2.

Non aged sample Aged sample
Peak area 10 -6
Peak area 10 -6
Peak area 10-6 Peak area.10-6 Pw Ps R/%
(2-PHE) (DFA) (DFA) (2-PHE)
63,16 ± 0,94 3,72 ± 0,03 6,33 ± 0,09 23,38 ± 0,23 17,00 3,70 78,26

Sample in laboratory condition

Sample artifically aged
Relative intensity .106

Time [min]

Figure 1. Chromatogram of sample from set 1 (5 cm line).

Sample in laboratory condition

Sample artifically aged
Relative intensity .106

Time [min]

Time [min]

Figure 2. Chromatogram of sample from set 2 (1 cm line).

4 CONCLUSIONS

2-phenoxyetanol is the most common volatile solvent presents in the ballpoint pens. Nowadays the analysis
of 2-phenoxyetanol by GC-MS is the most frequent method to determine of period of creation the record.
Analysis of the period of the record creation was made by calculating the percentage dependence R, which was
calculated by share of the peak area of the 2-phenoxyethanol aged and non-aged to peak area of the internal
standard (diphenylamine).

8th Conference on Information and Graphic Arts Technology 250

 Poster section

Crucial findings of our experiment were:

• There is no direct relationship between R and the length of the analysed line. While percentage amount of
2-PHE (R) was 95% for the line of 5 cm length, R was 78% for the line of 1 cm length, using an equivalent
amount of solvent in both samples.
• Accurate amount of analysed sample (ink line) is very important factor. We must determine the time depend-
ence of R (considering the time pass since the ink was applied to the paper) at defined accurate amount of
analysed material and specific conditions of analysis. This experiment is currently running and the results
are expected in form like in the publication (Luczak, 2002):
R < 20%: the result cannot be interpreted (writing is older than 18 month / did not contain 2-PHE)
R > 20%: record is probably younger than 18 months.

ACKNOWLEDGEMENT
This work was supported by Slovak Scientific Grant Agency VEGA No 1/0888/15, STU Grant Scheme for Support
of Excellent Teams of Young Researchers for financial assistance under contracts no. 1663. This publication
is the result of the project implementation: ITMS: 26240120034 supported by the Research and Development
Operational Programme funded by the ERDF.

5 REFERENCES

• Cantú, A., 2012. “A study of the evaporation of a solvent from a solution—Application to writing ink aging”,
Forensic Science International 219 (2012) 119–128.
• Jones, R., McClelland, J., 2013. “Analysis of writing inks on paper using direct analysis in real time mass
spectrometry”, Forensic Science International 231 (2013) 73–81.
• Luczak, R., Krawczyk, W., 2002. “Metodyka badań wieku dokumentów“, Problemy kryminalistyki 236 (2002),
18 – 22.
• Menzyk, A., Sajewicz, M., 2015. “Physicochemical analysis of ink − dating and establishing the sequence of
intersecting lines of ink entries”, Problems of Forensic Sciences, 104, 279–302
• Parciová, S., 2016. “Štúdium starnutia záznamových prostriedkov v dokumentoch” (“Study of aging writing
inks in documents”), Diploma thesis, Slovak University of Technology in Bratislava (2016), pp. 75.

Corresponding author:
Silvia STAŠKOVÁ
Department of Graphic Arts Technology and Applied Photochemistry,
Institute of Natural and Synthetic Polymers
Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava,
Radlinského 9, 812 37 Bratislava, Slovak Republic
phone: +421 904 032 05 80, e-mail: [email protected]

Co-author(s):
Milena REHÁKOVÁ
Department of Graphic Arts Technology and Applied Photochemistry,
Institute of Natural and Synthetic Polymers
Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava,
Radlinského 9, 812 37 Bratislava, Slovak Republic
e-mail: [email protected]

Aleš HÁZ
Department of Wood, Pulp and Paper, Institute of Natural and Synthetic Polymers
Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava,
Radlinského 9, 812 37 Bratislava, Slovak Republic
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 251

 Poster section

SYNTHESIS, COATING AND EVALUATION

OF ANTIMICROBIAL MICROCAPSULES ON PAPER

Bojana Boh PODGORNIK1, Boštjan ŠUMIGA1, Barbara GOLJA1, Matej ŠUŠTARŠIČ2,

Barbara ŠUMIGA2 & David RAVNJAK3
1
University of Ljubljana, Faculty of Natural Sciences and Engineering, Ljubljana, Slovenia
2
Pulp and Paper Institute, Ljubljana, Slovenia
3
Papirnica Vevče d.o.o., Ljubljana – Dobrunje, Slovenia

ABSTRACT: An antimicrobial formulation for coating of papers, based on microencapsulation technology,

was developed. Essential oil of Cymbopogon citratus (citronella oil) was microencapsulated by complex co-
acervation of gelatine and carboxymethylcellulose, and by in situ polymerization of melamine-formaldehyde
prepolymers with a polyacrylic acid modifier. Both methods resulted in a container type pressure-sensitive
microcapsules with a distinct liquid core and solid wall. Minimal inhibitory concentrations of non-activated
citronella oil microcapsules were determined for Bacillus subtilis, Escherichia coli, Pseudomonas aeruginosa
and Saccharomyces cerevisiae. The in situ polymerization microcapsules had better suspension stability and
formulation properties, and were more appropriate for the coating process, due to smaller microcapsule di-
mensions. Pressure-activation tests of coated papers suggested the possibility of a prolonged use with gradual
release of citronella oil at repeated pressures. The research will continue to optimise the quantity and/or type
of antimicrobial active substance to reach higher antimicrobial activities of coated papers.

Keywords: microcapsules, synthesis, coating, paper, antimicrobial.

1 INTRODUCTION

Interest in the bio-efficacy of essential oils and their uses as natural antimicrobials in environmentally-friendly
products has increased in recent years. Several innovations in smart packaging food systems included essential
oils (Rodrigez et al., 2007; Manso et al., 2013; Atarés and Chiralt, 2016; Ribeiro-Santos et al., 2017a,b). As the
components of essential oils are volatile and subject to degradation under environmental influences, different
encapsulation techniques may be applied to provide protection and prolonged or controlled release: micro-
capsules and microspheres, nanoparticles and nanospheres, liposomes, micro-emulsions, gels and molecular
inclusion (Xiao et al., 2014; Majeed et al., 2015; El Asbahani et al., 2015; Bakry et al, 2016; Rodríguez et al., 2016).
The purpose of our research was to develop a pressure-sensitive controlled-release microencapsulated form
of an essential oil with antimicrobial effects, which would be applicable as a functional additive in paper and
cardboard secondary packaging systems for food and pharmaceuticals. After the analysis of scientific literature,
the lemongrass (Cymbopogon citratus) essential oil, also known as the citronella oil (CAS RN 8000-29-1), was
selected for microencapsulation (Leimann et al., 2009; Wang, 2018). The citronella oil is known for its anti-
microbial properties (Naik et al., 2010) and pleasant aroma, and is widely used in food, pharmaceutical and
cosmetic products, as well as a repellent and a mild natural pesticide in agriculture and horticulture. The oil
consists predominantly of monoterpenes, such as citral (synonym geranial) (39-48%), neral (32-35%), mircen
(11-15%) and geraniol (3-5%) (Sacchetti et al., 2005; Matasyoh et al., 2011; Bassolé et al., 2011; Fadli et al., 2016)
and has a boiling point at 200°C, specific weight 0.897, and the oral LD50 for rat 7200 mg/kg (ScienceLab, 2013).
In the experimental part, our research focussed on the following steps: (a) synthesis of microcapsules by two
different methods – complex coacervation and in situ polymerization, (b) morphological characterisation
of produced microcapsules, (c) antimicrobial testing of microcapsules, (d) preparation of formulations and
coating on paper, (e) weight-activation of pressure-sensitive microcapsules on papers, and (f) antimicrobial
evaluation of paper matrices and coated papers.

2 EXPERIMENTAL
A non-diluted citronella oil (Sigma-Aldrich) was microencapsulated as a model antimicrobial component
by two processes taking place in oil-in-water emulsions. In the complex coacervation process, acid-treated
gelatine, type A (Sigma) and low viscosity carboxymethylcellulose, type 7L (Hercules) were used as macro-
molecular colloids with opposite charges for the formation of microcapsule walls. In a reactor vessel with a
turbine impeller the temperature was maintained at T = 50°C. Emulsification took place at 800 rpm, 10 min. To
form microcapsules, coacervation was initiated by dilution with water and by lowering pH to 4.3 (2 hours, T
= 50°C, 800 rpm). After cooling the system to a room temperature (1°C / min, 800 rpm), coacervate walls were

8th Conference on Information and Graphic Arts Technology 252

 Poster section

crosslinked by glutaraldehyde (Acros); the process was completed after 2 hours. In the in situ polymerization
microencapsulation process, melamine-formaldehyde precondensate (Melamin) and the polyacrylic acid mod-
ifier (BASF) were used. The laboratory reactor was equipped with a 1000 ml vessel, a heating/cooling system,
and a Heidolph mixer with adjustable speed 0-2000 rpm. An aqueous solution of the modifier was prepared,
the citronella oil was emulsified to form an oil-in-water emulsion, and a melamine precondensate was added
as the wall material. The polymerisation was initiated and maintained at an elevated temperature (T = 70-
80°C) for 1 hour, following by cooling to terminate the reaction. Antimicrobial effects of microcapsules were
determined as the minimum inhibitory concentration (MIC) in a liquid medium on Bacillus subtilis, Escher-
ichia coli, Pseudomonas aeruginosa and Saccharomyces cerevisiae, with 2,3,5-triphenyltatrazolium chloride
(Merck) added as an indicator. The MICs were determined after 24 hours of incubation at 37°C. Microcapsule
suspensions were formulated with water soluble starch or with carboxymethylcellulose (ratio 1:1 and 1:2) and
coated on paper (label paper, 70 gsm, and paper for flexible packaging, 50 gsm), 2 g formulation / m2 (for 1:1
formulation 0.086 g/m2 citronella oil within coacervation microcapsules, or 0.275 g/m2 citronella oil within in
situ microcapsules). Pressure-sensitive microcapsule walls were activated by pulling the coated paper under
a 5 kg metal weight. The morphology of microcapsules and the wall rupture after activation was examined
under the light transmission microscope (LTM) and scanning electron microscope (SEM). To evaluate the coated
papers, a method for the determination of aerobic mesophilic bacteria in paper was used for uncoated paper
matrices, and for papers coated with microcapsule containing formulations.

3 RESULTS & DISCUSSION

Both microencapsulation methods resulted in a container type microcapsules with a distinct liquid core and
solid wall. The coacervate microcapsules were larger (a diameter of 10 to 60 μm), with elastic natural polymer
walls (Figure 1), while the in situ polymerization microcapsules were smaller (diameter of 1 to 8 μm), with
spherical, rigid and impermeable synthetic polymer walls (Figure 2).

Figure 1. Coacervation microcapsules. Left LTM, 100×. Right: SEM, 500 ×.

Figure 2. In situ polymerization microcapsules. Left LTM, 100×. Right: SEM, 850×.

8th Conference on Information and Graphic Arts Technology 253

 Poster section

The results of antimicrobial testing of non-activated microcapsules in a liquid medium were collected after 24
hours incubation at 37°C. The MIC values within the range 0.07%< MIC >0.014% were determined for the intact
non-activated coacervation microcapsules with three tested bacterial species (B. subtilis, E. coli, P. aeruginosa),
and 1.72%< MIC >0.34% was determined for the yeast S. cerevisiae. The observed antimicrobial activity can be
attributed to partially permeable walls of the coacervation microcapsules. The obtained MIC values are com-
parable to the results by Wang et al. (2018), who produced citronella oil microcapsules with hydroxyapatite
/ quaternary ammonium salt of chitosan / sodium alginate walls, and determined the MIC 0.0625% for E coli
and S. aureus. For E. coli and B. subtilis, the results also coincide with the literature data for the non-capsu-
lated citronella oil - MIC 0.06% was determined by Naik et al. (2010). In contrast, in the case of non-activated
in situ polymerization microcapsules, the range 1.72%< MIC >0.34% was determined for B. subtilis, and MIC
>1.72% for the other three tested microorganisms. The results suggested that the in situ polymerization mi-
crocapsule walls were impermeable, and there was no residual non-encapsulated citronella oil present in
the suspension. Based on these results it can be assumed that coacervate citronella oil microcapsules can be
more suitable for applications with gradual release of the essential oil without mechanical activation, and
the in situ polymerization microcapsules for products where a long-term retention is desired, and a planned
release of active substance by a mechanical pressure. The combination of both microcapsule types could have
a double effect - a slow sustained release of the citronella oil, and a planned quick pressure-activated release
when used after a prolonged storage.

Microcapsules of both types withstood air drying when applied in a layer on a glass plate for LTM, and the
vacuum during the SEM microscopy. In the paper coating process, the in situ polymerization microcapsules
showed better formulation characteristics and were more suitable for the coating process, primarily due to
smaller microcapsule sizes, better suspension stability, and formulation mixing (Figure 3).

Figure 3. Coated paper with microcapsules. Left coacervation microcapsules (SEM, 1000×),
right in situ polymerization microcapsules (SEM, 1000×).

SEM photographs of coated papers confirmed a partial pressure-activation of in situ polymerization micro-
capsules in the 5 kg weight pulling test. Microcapsules with broken walls and empty cores were visible, while
some microcapsules remained intact (Figure 4). This indicated that not all of the essential oil was released
from microcapsules at once, and suggested the possibility of prolonged use - gradual release of citronella
oil at multiple pressures. We estimated that the mechanism of essential oil release by mechanical pressure
was suitable for the intended application. However, the question remained whether the amount of released
citronella oil would be sufficient for the antimicrobial effect. The preliminary antimicrobial testing - determi-
nation of aerobic mesophilic bacteria in paper on standard count agar (for E. coli, P. aeruginosa, B. subtilis )
and determination of yeast on sabouraud 2% glucose agar (for S. cerevisae) revealed no significant difference
between the uncoated paper matrices and papers coated with a formulation containing in situ polymerization
microcapsules with citronella oil in the applied concentration range (2g of coating formulation / m2), without
or with pressure-activation. The result could be expected, as the coating 2 g/m2 only contained approximately
0.275 g of encapsulated citronella oil per m2, and the 1 cm2 testing sample on agar plate only contained 0.00275
g of citronella oil in microcapsules. Therefore, after the successful microencapsulation and coating step, our
research will continue to optimise the antimicrobial activity on coated papers, first by increasing the amount
of citronella oil microcapsules in the coating composition, and then by integrating a larger proportion of mi-
crocapsules into the paper mass during the production of paper. Another possibility is to encapsulate antimi-

8th Conference on Information and Graphic Arts Technology 254

 Poster section

crobial compounds with stronger antimicrobial activities, and/or to prepare mixtures of several antimicrobial
compounds targeting Gram-positive and Gram-negative bacteria as well as yeasts and moulds.

Figure 4. Coated paper with in situ polymerization microcapsules after the 5 kg weight pulling test.
Left SEM 1000×, right SEM 3300×.

4 CONCLUSIONS

The aim of our research was to develop a pressure-sensitive controlled-release formulation with antimicro-
bial properties, to be used in packaging systems for food or pharmaceuticals. After a literature analysis, the
essential oil of Cymbopogon citratus (citronella oil) was chosen as the antimicrobial compound, due to its
antimicrobial activity, pleasant aroma, and applicability in food, pharmaceutical, cosmetic and agricultural
products. To reduce its volatility and prolong the activity, the citronella oil was microencapsulated. Two pro-
cesses were successfully applied: the complex coacervation of acid-treated gelatine and low viscosity carbox-
ymethylcellulose, and the in situ polymerization of melamine-formaldehyde precondensates with polyacrylic
acid modifying agent. Both methods resulted in a container type microcapsules with a distinct liquid core
and solid wall. However, the properties differed: the coacervate microcapsules were larger (10 to 60 μm), with
elastic partially permeable walls, while the in situ polymerization microcapsules were smaller (1 to 8 μm),
with rigid and impermeable walls. Minimal inhibitory concentrations of microencapsulated citronella oil were
determined for B. subtilis, E. coli, P. aeruginosa and S. cerevisiae. The results indicated that the antimicrobial
activity depended on microcapsule type and wall permeability. Coacervate citronella oil microcapsules would
be suitable for applications with gradual release, while the in situ polymerization microcapsules could provide
a long-term retention and a release by a mechanical pressure. The in situ polymerization microcapsules showed
better formulation characteristics (suspension stability and formulation mixing), and were more appropriate
for the coating process, due to smaller dimensions. Pressure-activation of coated papers indicated that some
microcapsules opened during the 5 kg weight pulling test and released the citronella oil, while a part of mi-
crocapsules remained intact, suggesting the possibility of a prolonged use with gradual release at repeated
pressures during the product use. However, the amount of microcapsules or the concentration of antimicrobial
agent still needs to be optimised, as the applied coating did not result in a significant antimicrobial activity,
compared to uncoated papers. This will remain the goal for further research in the continuation of the project.

ACKNOWLEDGEMENTS
The work was carried out within the RDI project Cel.Cycle: »Potential of biomass for development of advanced
materials and bio-based products« (contract number: OP20.00365), co-financed by the Republic of Slovenia,
Ministry of Education, Science and Sport and European Union under the European Regional Development
Fund, 2016–2020.

5 REFERENCES

• Atarés, L., and Chiralt, A. 2016. “Essential oils as additives in biodegradable films and coatings for active
food packaging“. Trends in food science & technology, 48, 51-62.
• Bakry, A. M., Abbas, S., Ali, B., Majeed, H., Abouelwafa, M. Y., Mousa, A., and Liang, L. 2016. “Microencapsu-
lation of oils: a comprehensive review of benefits, techniques, and applications“. Comprehensive Reviews in
Food Science and Food Safety, 15(1), 143-182.

8th Conference on Information and Graphic Arts Technology 255

 Poster section

• Bassolé, I. H. N., Lamien-Meda, A., Bayala, B. O. L. C., Obame, L. C., Ilboudo, A. J., Franz, C., and Dicko, M. H.
2011. “Chemical composition and antimicrobial activity of Cymbopogon citratus and Cymbopogon giganteus
essential oils alone and in combination“. Phytomedicine, 18(12), 1070-1074.
• El Asbahani, A., Miladi, K., Badri, W., Sala, M., Addi, E. A., Casabianca, H., and Elaissari, A. 2015. “Essential
oils: from extraction to encapsulation“. International journal of pharmaceutics, 483(1-2), 220-243.
• Fadli, M., Pagès, J. M., Mezrioui, N. E., Abbad, A., and Hassani, L. 2016. “Artemisia herba-alba Asso and Cym-
bopogon citratus (DC.) Stapf essential oils and their capability to restore antibiotics efficacy“. Industrial
Crops and Products, 89, 399-404
• Leimann, F. V., Gonçalves, O. H., Machado, R. A., and Bolzan, A. 2009. “Antimicrobial activity of microen-
capsulated lemongrass essential oil and the effect of experimental parameters on microcapsules size and
morphology“. Materials Science and Engineering: C, 29(2), 430-436.
• Majeed, H., Bian, Y. Y., Ali, B., Jamil, A., Majeed, U., Khan, Q. F. and Fang, Z. (2015). “Essential oil encapsula-
tions: uses, procedures, and trends“. RSC Advances, 5(72), 58449-58463.
• Manso, S., Cacho-Nerin, F., Becerril, R., and Nerín, C. 2013. “Combined analytical and microbiological tools
to study the effect on Aspergillus flavus of cinnamon essential oil contained in food packaging“. Food Con-
trol, 30(2), 370-378.
• Matasyoh, J. C., Wagara, I. N., and Nakavuma, J. L. (2011). “Chemical composition of Cymbopogon citratus es-
sential oil and its effect on mycotoxigenic Aspergillus species“. African Journal of Food Science, 5(3), 138-142.
• Naik, M. I., Fomda, B. A., Jaykumar, E., and Bhat, J. A. (2010). “Antibacterial activity of lemongrass (Cymbopo-
gon citratus) oil against some selected pathogenic bacterias“. Asian Pacific Journal of Tropical Medicine, 3(7),
535-538.
• Ribeiro-Santos, R., Andrade, M. and Sanches-Silva A. 2017a. “Application of encapsulated essential oils as
antimicrobial agents in food packaging“. Current Opinion in Food Science, 14, 78-84.
• Ribeiro-Santos, R., Andrade, M., de Melo, N. R., and Sanches-Silva, A. 2017b. “Use of essential oils in active
food packaging: Recent advances and future trends“. Trends in food science & technology, 61, 132-140.
• Rodriguez, A., Batlle, R., and Nerin, C. (2007). “The use of natural essential oils as antimicrobial solutions in
paper packaging. Part II“. Progress in Organic Coatings, 60(1), 33-38.
• Rodríguez, J., Martín, M. J., Ruiz, M. A., and Clares, B. (2016). “Current encapsulation strategies for bioactive
oils: From alimentary to pharmaceutical perspectives“. Food Research International, 83, 41-59.
• Sacchetti, G., Maietti, S., Muzzoli, M., Scaglianti, M., Manfredini, S., Radice, M., and Bruni, R. 2005. “Compar-
ative evaluation of 11 essential oils of different origin as functional antioxidants, antiradicals and antimi-
crobials in foods“. Food chemistry, 91(4), 621-632.
• ScienceLab, 2013. Citronella oil - Material Safety Data Sheet, https://www.sciencelab.com/msds.
php?msdsId=9923497 (last accessed on 29.03.2018).
• Wang, J., Li, X., Chen, M., Chen, Z., Wu, H., Zhang, P., and Hu, Y. 2018. “Fabrication of sustained‐release and
antibacterial citronella oil‐loaded composite microcapsules based on Pickering emulsion templates“. Journal
of Applied Polymer Science, DOI: 10.1002/APP.46386.
• Xiao, Z., Liu, W., Zhu, G., Zhou, R., and Niu, Y. 2014. “A review of the preparation and application of flavour
and essential oils microcapsules based on complex coacervation technology“. Journal of the Science of Food
and Agriculture, 94(8), 1482-1494.

Corresponding author:
Bojana BOH PODGORNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Aškerčeva 12, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 57, fax: +386 1 200 32 70, e-mail: [email protected]

CO-AUTHOR(S):
Barbara GOLJA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Aškerčeva 12, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 30, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 256

 Poster section

David RAVNJAK
Papirnica Vevče d.o.o.
Papirniška pot 25, SI-1261 Ljubljana Dobrunje, Slovenia
phone: +386 1 587 73 62 , fax: +386 1 5285450, e-mail: [email protected]

Barbara ŠUMIGA
Pulp and Paper Institute, Ljubljana
Bogisiceva ulica 8, SI-1000 Ljubljana, Slovenia
phone: 386 1 200 28 48, e-mail: [email protected]

Boštjan ŠUMIGA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design
Aškerčeva 12, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 00, e-mail: [email protected]

Matej ŠUŠTARŠIČ
Pulp and Paper Institute, Ljubljana
Bogisiceva ulica 8, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 28 15, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 257

 Poster section

TLC AND GC/MS ANALYSIS OF INK JET PRINTING INKS

Barbara BLAZNIK1, Franci KOVAČ2 & Sabina BRAČKO1

1
Faculty of Natural Sciences and Engineering
2
Faculty of Chemistry and Chemical Technology

ABSTRACT: At the moment when ink hits the printing material, the chemical structure of ink changes and
several chemical processes occur, which enable fast drying or penetration of ink into the printing material.
For a comprehensive understanding of the process that occurs during ink-substrate interaction, it is necessary
to be aware of the components of the ink as well as the substrate. Therefore, the aim of our study was the
determination of components that were needed to achieve required colour as well as what components for
achieving appropriate physical and chemical properties of the printing ink were used. The thin-layer chro-
matography (TLC) analysis gave us the insight into the colour system of printing inks. Regardless of some
limitations of gas chromatography (GC) the GC/MS chromatography enabled the overview on the additives
that were added to the printing ink.

Keywords: ink-jet, printing inks, TLC, GC/MS, chromatography.

1 INTRODUCTION

Ink jet technology is advancing rapidly. This technology can be encountered in the workplace as well as in
our homes (Hudd, 2010). The interest of average end user is mostly quality print and maybe even a fastness of
prints. Therefore, the manufacturers do not see the importance of spreading confidential information about
components that our home ink jet printers use. The aim of ink manufacturers seems to be meeting the re-
quirements of printer manufacturers, which is mainly limited to the physical and chemical properties of the
ink (Agarwal, 2016). Consequently, they do not hesitate to replace ink components with more cost-effective
components (Sharma, 2014) as long as they provide the same physical and chemical properties of the ink. At
the moment when ink hits the printing material, the chemical structure of ink changes and several chemical
processes occur, which enable fast drying or penetration of ink into the printing material. Meaning that iden-
tification of printing inks is complicated due to methods of application, non-transparent selection of different
chemical structures, and commercial names (Lewis, 2009).
Therefore, when we observe prints, we usually do not think only about colour itself, but we should also take
into consideration the printing material that accompanies the print. However, at the same time, we forget
about components that were or are still present in ink.
The interest of our study was the determination of components that were needed to achieve required colour
as well as what components for achieving appropriate physical and chemical properties of the printing ink
were used.
Therefore, in the experimental part of our study, the determination of ink components was undertaken with
a relatively simple and useful method – chromatography. The use of thin-layer chromatography (TLC) for
the separation of dyes has proved to be the ideal solution due to the wide selection of stationary phases and
unlimited mobile-phase mixtures (Wall, 2000). For the first, polar or stationary phase the silica gel with a
fluorescent indicator on aluminium plate was selected. The second or mobile phase was chosen according to
the previous tests. According to the experience of other researchers (Agarwal, 2016; Lewis, 2009; Wall, 2000)
using a single solvent combination for mobile phase the complete separation of ink components will not be
possible. Due to non-volatile dyes, a gas chromatography (GC) has limitations in the field of characterisation
of inks (Agarwal, 2016). However, GC/MS chromatography enabled the overview on the additives that were
added to the printing ink.
In the field of dye separation, we can find several analytical methods which allow more extensive understand-
ing of ink jet ink components, each of them having specific advantages and importance in analysing printing
inks. With those methods, it is possible to confirm some assumptions and find some answers to the questions
that appear during experimental part.

8th Conference on Information and Graphic Arts Technology 258

 Poster section

2 EXPERIMENTAL

2.1 Printing inks

For the experimental part, products of the two ink manufacturers (Canon and Epson) were carefully selected.
The conditions for selecting those two manufacturers were:
• dye based cyan, magenta, yellow and black ink jet printing inks;
• use of printing inks in desktop ink jet printers;
• printing inks should be in a liquid state.
Before the analysis, using a needle, the ink was removed out of the cartridge. The ink samples were stored in
a dark and cool place in tightly closed glass vials.
Due to the sensitivity of the measuring device, the dilution of ink and methanol was required at the concen-
tration of 1:3000.

2.2 Thin-layer chromatography (TLC)

For TLC analysis the silica with a 254 nm fluoresecence indicator coated aluminum plates (Sigma-Aldrich, DE)
were used. For mobile phase we selected four different mixtures of solvents:
• ethylmethylketone:acetone:distilled water (7:5:3 v/v);
• ethyl acetate:ethanol:distilled water (70:35:30 v/v);
• butanol-2:ethanol:distilled water, (10:2:3 v/v);
• acetonitrile:distilled water (2:1 v/v).
The samples were placed as spots on a line drawn 1 cm above the bottom of coated aluminium plate and ver-
tically placed into the glass chamber containing the mixture of solvents. The process of migration and sepa-
ration occurred when the mobile phase was moving upward the stationary phase. After a certain period, the
plate was removed out of the chamber and dried in the air. Under the visible light, the separated spots were
marked and the retention factor (Rf) was calculated according to equation 1, where ds is migration distance
of substance and dsf is the migration distance of solvent front.

Rf = ds/dsf (1)

2.2 Gas chromatography – Mass spectrometry (GC/MS)

For GC/MS analysis the Agilent HP 6890 (Hewlett Packard, USA) instrument was used. As a carrier gas (mo-
bile phase) the helium was used. The parameters for GC/MS analysis was, previously heated of the injector
to the 250°C and heating chamber to the temperature of 80°C. During the analysis, the temperature in the
heating chamber gradually raised (20°C per minute) to the final temperature of 270°C and maintained for 20
minutes. Only a small amount (2 μl) of the sample was required for the analysis. The final amount of sample
was assessed by a split ratio of 10:1, which means that only 0.2 μl of sample passed through the GC column.
The mass detector performed scanning of molecular weight from 15 to 700.

3 RESULTS & DISCUSSION

3.1 Thin-layer chromatography (TLC)

According to the previous researches (Agarwal, 2016; Lewis, 2009; Wall, 2000), the complete separation of
the ink will not be possible by using only a single solvent combination for mobile phase. According to the
tests of four solvent mixtures we have decided that for the first mobile phase the solvent mixture of ethyl-
methylketone:acetone:distilled water (7:5:3 v/v) and for the second mobile phase the solvent mixture of ethyl
acetate:ethanol:distilled water (70:35:30 v/v) was used. Referring to chromatograms the retention factors (Rf)
for each sample were calculated. Rf factor provided a qualitative estimation of components that were needed
to achieve required colour in printing ink.
From TLC chromatogram (Figure 1) we can observe that the printing inks consist of two or more coloured
components. Figure 1 shows more or less successful ink separations on the aluminium plate with the first and
second solvents mixture. In the case of cyan and black printing inks, we have achieved incomplete separation of
printing inks. Since the part of the ink component was kept at the starting point as well as some components
had the same Rf values and were not completely separated during the process. It can be assumed that some
ink components may contain a salt-based ink as the salt-forming dye is superior in dissolution stability with
time and fastness properties (e.g. light and water resistance) and it is applied in combination with a basic dye

8th Conference on Information and Graphic Arts Technology 259

 Poster section

(Ono, 1994), which prevent the component to run upward over the stationary phase. Inorganic salts are often
added as extenders, so different batches have the same dyeing potential (Wall, 2000). Also, for black inks an-
other undesirable result became apparent, several components had pretty much the same polarity. Therefore,
some components did not separate completely from each other during the analysis process.

1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 C 0.6
M
0.5 0.5

Rf
Rf

Rf (1) Rf (1)
0.4 0.4
Rf (2) Rf (2)
0.3 0.3
0.2 0.2
0.1 0.1
0 0
CC3 EC3 CC3 EC3 a) CM3 EM3 CM3 EM3 b)
ethylmethylketone:acetone:water ethyl acetate:ethanol:water ethylmethylketone:acetone:water ethyl acetate:ethanol:water

1 1
0.9 0.9
0.8 0.8
0.7 0.7 K
0.6 Y 0.6 Rf (1)
0.5 Rf 0.5
Rf

Rf (1) Rf (2)
0.4 0.4
Rf (2) Rf (3)
0.3 0.3
Rf (4)
0.2 0.2
0.1 0.1
0 0
CY3 EY3 CY3 EY3 c) CK3 EK3 CK3 EK3 d)
ethylmethylketone:acetone:water ethyl acetate:ethanol:water ethylmethylketone:acetone:water ethyl acetate:ethanol:water

Figure 1. Display of calculated retention factor (Rf) of ink jet inks with first and second mobile phase for
individual printing inks. (a) Cyan (CC3 – Canon Cyan and EC3 – Epson Cyan); (b) Magenta (CM3 – Canon
Magenta and EM3 – Epson Magenta); (c) Yellow (CY3 – Canon Yellow and EY3 – Epson Yellow); (d) Black
(CK3 – Canon Black and EK3 – Epson Black).

3.2 Gas chromatography – Mass spectrometry (GC/MS)

Due to non-volatile dyes, a gas chromatography (GC) has limitations in the field of characterisation of inks
(Agarwal, 2016). However, GC/MS chromatography enabled the overview on the additives that were added to
the printing ink. The results of GC/MS analysis revealed the insight into the ink components which manufac-
turers add to the printing inks to provide some specific physical and chemical properties of the inks. According
to the results, the preparation of ink formulations varies according to the manufacturer as well as according
to the colour. However, some components appear in both cases. We have found the following compounds of
glycerol, divinyl sulfone, 1,5-pentanediol, 2-pyrrolidone, ethanol, and in some cases the presence of 2-imi-
dazolidone as well as glycerin and triethylene glycol. Aliphatic glycols, among which is also glycerol, are in
water-based inks to behave as a carrier fluid (Foucher, 1999). The purpose of alcohols in printing inks is to
regulate the surface tension or to control the drying of the ink on the substrate and to improve the resistance
of the prints against water (Foucher, 1999; Pekarovicova, 2015). To prevent the evaporation of water and the
sedimentation of coloured components the 2-pyrrolidone (Foucher, 1999) is used. 1,5-pentanediol is used to
improve curling effect (Foucher, 1999). Moreover, the presence of divinyl sulfone in the GC/MS chromatogram
could be associated with reactive dyes, which are also used in ink-jet printing (Novak, 2004).

4 CONCLUSION

The analysis of different inks is useful for a comprehensive understanding of the process that occurs during
the ink-substrate interaction, and it is of great interest in forensic science. Detected components can improve
the understanding of the degradation process of prints with consideration of the ratio that those components
contribute to the degradation process of prints.
Despite the fact that TLC and GC/MS chromatography are relatively simple methods we have obtained a
significant amount of useful data on the composition of ink jet printing inks. For example, with specific im-

8th Conference on Information and Graphic Arts Technology 260

 Poster section

provements in the selection of the mobile phase mixture of solvents better separation of dyestuff could be
achieved in the TLC chromatography process. GC/MS has proved to be an important analysing method for
determination of volatile organic compounds in ink samples.

5 REFERENCES

• Agarwal, A., Sharma, N., Negi. Y. S. 2016. “Review:Techniques for the Characterization of Inks.” IOSOR-Journal
of Applied Chemistry 9 (10): 76–96.
• Foucher, D. A., Sacripante, G. G., Wong R. W., Breton, M. P. 1999. “Ink Compositions.” Canada Patent 5,969,003.
• Hudd, A. 2009. “Inkjet prining technologies.” In The Chemistry of Inkjet Inks, edited by Shlomo Magdassi,
3–18. Xennia Technology Limited, UK.
• Lewis, S. W. 2009. “Analysis of dyes using chromatography.” In Identification of textile fibres, edited by Max
Houck, 203–223. Cambridge, Woodhead Publishing.
• Novak, G. 2004. Grafični materiali. Ljubljana: Univerza v Ljubljani, Naravoslovnotehniška fakulteta, Oddelek
za tekstilstvo.
• Ono, T., Yagyu, T., Akase, A. 1994. “Salt-forming type dye and ink composition.” US Patent: 5281264.
• Pekarovicova, A., Husovska, V. 2015. “Printing ink Formulations,” In Printing on polymers: Fundamentals and
applications, edited by Joanna Izdebska, Sabu Thomas, 41–55. Oxford: Elsevier Inc.
• Sharma, R., Baggi, T. R., Chattree, A., Gupta, A. K. 2014. “Application of High Performance Thin Layer Chroma-
tography in Examination of Magenta Coloured Printed Matter.” IOSR Journal of Applied Chemistry (IOSR-
JAC) 7 (7): 48–52.
• Wall, P. E. 2000. “Thin-Layer (Planar) Chromatography.” In Encyclopedia of separation science, edited by I.
D. Willson, M. Cooke, C. F. Poole, E. R. Adlard, 2619–2631. New York: Academic Press.

Corresponding author:
Sabina BRAČKO
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Chair of Information and Graphic Arts Technology
Snežniška 5
1000 Ljubljana, Slovenia
phone: +386 1 200 32 38, fax: +386 1 200 32 70, e-mail: [email protected]

Co-authors:
Barbara BLAZNIK
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Chair of Information and Graphic Arts Technology
Snežniška 5
1000 Ljubljana, Slovenia
phone: +386 1 200 32 22, fax: +386 1 200 32 70, e-mail: [email protected]

Franci KOVAČ
University of Ljubljana, Faculty of Chemistry and Chemical Technology
Večna pot 113
1000 Ljubljana, Slovenia
e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 261

 Poster section

TYPEFACE AND ECOLOGICAL DESIGN

Klementina MOŽINA, Alenka TOMAŠIČ

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Chair of Information and Graphic Arts Technology,
Snežniška 5, SI-1000 Ljubljana, Slovenia

ABSTRACT: Within the context of ecological design, typographers have started designing typefaces that re-
quire less ink for printing. Nevertheless, these typefaces need to be legible. The typeface Epika in three weights
was made by taking into account the design rules for better legibility of a typeface, and an ecological design
suggestion for spending less ink during printing. The conventional Arial typeface was used for comparison.
For the purpose of the research, eight different texts were selected. The texts were printed by using the ink
jet printing technology and electrophotography. We printed onto two different paper grades, i.e. uncoated
office and recycled papers. The typographic tonal density of all printed typefaces was measured. The legibility
of printed texts was measured with the rate of work method. The results showed that the typeface Epika had
lower typographic tonal density than Arial. The texts in the typeface Epika were more legible.

Keywords: desktop publishing, ecological design, legibility, typeface.

1 INTRODUCTION

The modern way of life represents a big burden for the nature; therefore, the importance of environmental
protection is nowadays on the increase. The field of ecology was at the beginning a part of biology, and it only
later became an interdisciplinary science which covers natural sciences and human behaviour that influences
the world changes (Vuk, 2000). Ecology is influenced by worldwide environmental protection. When thinking
about ecological design (ISO 14001, 2004; ISO/TR 14062, 2002), the solution of raw materials and materials that
consume energy should be taken into consideration as well (Možina, 2006). Environmental friendliness has
become a necessity in the modern way of life (Radonjič, 2008; Zakon o varstvu okolja, 2015). Furthermore, the
development of ecological design is of the essence (ISO 14001, 2004; ISO/TR 14062, 2002). Within the context
of ecological design, typographers also design typefaces that spend less ink during printing. Nevertheless,
these typefaces need to be legible.
The communication through a page or a screen requires from the reader to translate symbols into meaning.
Legibility refers to how easily this process is performed. To make reading possible, the text must be character-
ised by three properties (Reynolds, 1979; Možina, 2001):
• The property of visibility, so that a clear image of adequate size is received by the retina. The visibility re-
quirements depend on the viewer’s eyesight.
• The recognisability or perceptibility of letters and words making up the text. This is affected by the factors
such as type style and form as well as by the reader’s reading skills.
• The property of comprehensibility. Comprehension is affected not only by the content of the text but also
by its visibility and perceptibility, and by the verbal capacity and intelligence of the reader.
Legibility has been studied using different methods (Reynolds, 1979; Možina, 2001), e.g. visibility, perceptibility
at a distance, perceptibility in the peripheral vision, speed of perception, ocular movement, rate of work and
other criteria. The rate of work is the most satisfactory measure for the legibility of a continuous text. The
methods which were applied include the speed of reading aloud, the speed of reading silently, the speed of
skimming or scanning the text (e.g. subjects are asked to locate specific items in the list of food ingredients or/
and to locate specific target words in a text), and the rate of comprehension (e.g. subjects are asked to answer
open-ended or dichotomous questions). The rate of work is also useful in studying the presentation of more
complex materials (e.g. timetables and directories) (Reynolds, 1979; Možina, 2001; Možina, Likar, Muck, 2016).
A large number of studies on legibility points to its importance. There is a big discrepancy in understanding
what makes a text legible. However, it is possible to determine some general guidelines that can help create
a legible text. There are some typographic characteristics to be observed to make a text more legible. For a
small type size, it is known that the differences in stroke weight and typographic tonal density (or typographic
tonality) are significant (Možina et al, 2010; Rat et al, 2011; Možina, Likar, Muck, 2016), since they influence text
legibility. Furthermore, a number of other typographic characteristics needs to be observed in order to make
a text more legible, i.e. distinctive character features (counter shape), x-height, ascender, descender, serifs,

8th Conference on Information and Graphic Arts Technology 262

 Poster section

contrast (stroke weight), set width, type size, leading (i.e. space between lines) etc. (Reynolds, 1988; Možina,
2001; Gaultney, 2001; Tracy, 2003). The contrast depends on the difference between thick and thin strokes of
a given letter. The set width defines the width of a letter (e.g. the same typeface style and the same size can
have different letter width) (Bringhurst, 2002; Možina, 2003). In this research, we focused on the designing of
the typeface Epika in three weights (regular, bold and light), where the design rules for better legibility of a
typeface, and an ecological design suggestion for spending less ink during the printing (Bolanča Mirković,
Možina, 2009) were taken into account. We also checked the legibility of our ecological typeface in comparison
with a conventional, non-eco, widely used typeface (i.e. Arial).
The purpose of the research was to determine the eco typeface legibility in print media compared with the
conventional typeface. For the purpose of the research, eight different texts were selected. Attention was paid
to the length of the text. We put each text into one of the two typefaces. The typefaces were 10 points in size.
Moreover, the typographic tonal density was taken into consideration as it refers to the relative blackness or
shades of grey of type on a page. It can be expressed as the relative amount of ink per square centimetre, pica
or inch (Keyes, 1993). The changes in various type features can create variations in typographic tonal density
(Reynolds, 1988; Možina, 2001; Možina et al, 2007; Možina, Likar, Muck, 2016). By using the ink jet printing
technology and electrophotography (Kipphan, 2001; Kumar, 2009), we printed the texts onto two different
paper grades, i.e. ordinary uncoated office and recycled papers. The legibility of printed texts was tested with
the time of reading and a dichotomous question at the end of reading.

2 EXPERIMENTAL

The differences between different typefaces, used papers and printers, and legibility were researched with
legibility testing and the analysis of typographic properties.

2.1 Methods and Materials

In the research, the legibility of the typeface Epika, which was designed according to the ecological guidelines,
was compared with a conventional, non-eco, widely used typeface (i.e. Arial). The typeface Epika was designed
with the program FontLab Studio. The test form was designed with the program Adobe InDesign and was
used as a PDF file. We used two different papers, i.e. ordinary uncoated office paper (80 g/m2), marked S1, and
recycled paper (80 g/m2), marked S2. Black prints were made with the ink jet printing technology, using the
printer Epson Stylus SX425W (marked P1), and with electrophotography, using the printer HP 1018 (marked P2).
The differences in the typographic tonal density of tested typefaces were measured with an image analysis (Im-
ageJ). This software gives the opportunity to measure, analyse and provide output values, e.g. area, number of
particles, circularity and percentage of coverage (National Institutes of Health, Research Services Branch, 2017).

2.2 Legibility

Different texts from the Slovenian edition of the journal National Geographic were printed in different type-
faces, with different printers onto different papers. Eight different texts contained between 100 and 110 words.
The observers (N = 20) were people aged between 20 and 30 years with a normal or corrected-to-normal vision.
They read the texts at the same conditions of lighting measured with a spectrophotometer X-rite EyeOne,
and viewing distance between 300 and 400 mm. Each observer read all eight combinations. Among different
observers, the texts were presented in random order contributing the Latin square design, which was used
for counterbalancing the order of the texts, i.e. to randomise the measurement process and to eliminate pos-
sible fatigue effects. Each second observer received the texts in a different order. For each tested observer, we
measured the reading time.

3 RESULTS & DISCUSSION

3.1 Typeface Epika

The typeface Epika in three weights, i.e. regular (cf. Figure 1), bold (cf. Figure 2) and light (cf. Figure 3), was
made taking into account the design rules for better legibility of a typeface and an ecological design sug-
gestion for spending less ink during the printing. This typeface was designed in a sans-serif style (McLean,
1996; Možina, 2003). It has a higher x-height. Letters and digits are more open. In the contacts of strokes, we
made ink traps to reduce the blackness of the characters. The strokes of the bold version of the typeface are
wider by 16%, and the strokes of the light version of the typeface are thinner by 26%, compared to the regular
version of the Epika typeface.

8th Conference on Information and Graphic Arts Technology 263

 Poster section

Figure 1. Regular typefaces Epika.

Figure 2. Bold typeface Epika.

Figure 3. Light typeface Epika.

3.2 Typographic properties of prints

The typographic tonal density (TTD) in the regular version of each typeface, 10 pt in size, on both used paper
graders, printed with both printers was measured (cf. Table 1).
The results show lower TTD at the typeface Epika which results from the thinner stroke width and bigger
counter size of letters. Due to the used printing technology, lower TTD was expected on the prints made with
electrophotography (P2) than on the prints made with the ink jet technology (P1). The lowest TTD was ob-
served at the typeface Epika printed on the ordinary office paper (S1) with the printer P2. There is a difference
in the value of TTD between the used papers. The value of TTD is noticeably higher on the recycled paper (S2).

8th Conference on Information and Graphic Arts Technology 264

 Poster section

Table 1. Average value of typographic tonal density (TTD) of tested typefaces according to used printers (P1,
P2) and used papers (S1, S2).
TTD [%] of printed samples
Typeface
S1-P1 S2-P1 S1-P2 S2-P2
Epika 15.72 15.81 12.35 13.30
Arial 17.69 17.92 14.15 15.00

3.3 Legibility of prints

The observers (average age 25.83 years) read different texts on both tested papers, in both typefaces and printed
with both printers. 32% of observers were male and 67% female. While reading the texts, the average lighting
in the room was 556.30 lux.
Table 2 shows the influence of the used typefaces, papers and printers on the speed of reading. On average,
the reading speed was higher at the prints on the ordinary office paper (S1). Since electrophotography gives
prints of better quality than the ink jet technology, higher reading speed was expected (P2) than on the prints
made with the ink jet technology (P1). At all examples, the texts printed with the Epika typeface were read
the fastest, regardless of the used papers and printers. In previous research (Možina, Likar, Muck, 2016), it was
established that eco-friendly designed typefaces are read faster than the conventional ones. The interaction
between lower TTD and faster reading was noticed (cf. Table 1 and Table 2). A larger counter size gave a lower
TTD value and consequently resulted in faster reading.

Table 2. Average reading time of tested typefaces according to used printers (P1, P2) and used papers (S1, S2).
Reading time [s] of different texts
Typeface
S1-P1 S2-P1 S1-P2 S2-P2
Epika 26.80 27.91 24.13 26.89
Arial 27.69 28.75 26.40 28.15

4 CONCLUSIONS

The results of the research show that the eco-friendly designed typeface Epika is read faster than the conven-
tional one, i.e. Arial. On the other hand, it was observed that all typefaces printed with the ink jet technology
use more ink than the typefaces printed with electrophotography and that they result in lower reading speed.
The typographic tonal density of the eco-friendly designed typeface is lower from the non-eco typeface. If
some other eco-friendly designed typefaces along with conventional typefaces were taken into consideration
as well, we would be able to give further recommendations about the usability of eco-friendly designed
typefaces.

5 REFERENCES

• Bolanča Mirković, I., Možina, K. 2009. “Paper recycling efficiency in function of the typeface.” In DAAAM
International scientific book 2009, edited by Branko Katalinic, 599–612. Vienna: DAAAM International Vienna.
• Bringhurst, R. 2008. The Elements of Typographic Style. Point Roberts: Hartley & Marks.
• Gaultney, V. 2001. “Balancing Typeface Legibility and Economy: Practical Techniques for the Type Designer”,
research essay, University of Reading.
• ISO 14001: Environmental management systems – Requirements with guidance for use. 2004.
• ISO/TR 14062: Environmental management systems – Integrating environmental aspects into product design
and development. 2002.
• Keyes, E. 1993. “Typography, color, and information structure.” Technical Communication 4: 638–654.
• Kipphan, H. 2001. Handbook of print media: technologies and production methods. Berlin: Springer.
• Kumar, M. 2009. Tehnologija grafičnih procesov. Ljubljana: Srednja medijska in grafična šola Ljubljana.
• McLean, R. 1996. The Thames and Hudson Manual of Typography. London: Thames and Hudson.
• Možina, K. 2001. “Zgodovinski razvoj knjižne tipografije.” PhD diss., University of Ljubljana.
• Možina, K. 2003. Knjižna tipografija. Ljubljana: University of Ljubljana.
• Možina, K. 2006. “Paper in digital revolution” In Proceedings, 10th international conference on printing,
design and graphic communication Blaž Baromić, edited by Zdenka Bolanča, Miroslav Mikota, 175–180.

8th Conference on Information and Graphic Arts Technology 265

 Poster section

Zagreb: Faculty of Graphic Arts, University of Zagreb; Senj: Ogranak Matice hrvatske Senj; Ljubljana: Pulp
and Paper Institute.
• Možina, K., Černič, M., Demšar, A. 2007. “Non-destructive methods for chemical, optical, colorimetric and
typographic characterisation of a reprint.” Journal of Cultural Heritage 8: 339–349.
• Možina, K., Medved, T., Rat, B., Bračko, S. 2010. “Influence of Light on Typographic and Colorimetric Properties
of Ink Jet Prints.” Journal of Imaging Sciences and Technology 54 (6): 060403-1–060403-8.
• Možina, K., Likar, K., Muck, D. 2016. “Legibility of eco fonts.” In Proceedings, 8th International Symposium
on Graphic Engineering and Design GRID, edited by Živko Pavlović, 387–393. Novi Sad: Faculty of Technical
Sciences, Department of Graphic Engineering and Design.
• National Institutes of Health, Research Services Branch. URL: http://rsb.info.nih.gov/ij/ (last accessed on
3. 1. 2017).
• Radonjič, G. 2008. Embalaža in varstvo okolja: zahteve, trendi in podjetniške priložnosti. Maribor: Založba
Pivec.
• Rat, B., Možina, K., Bračko, S., Podlesek, A. 2011. “Influence of Temperature and Humidity on Typographic
and Colorimetric Properties of Ink Jet Prints.” Journal of Imaging Sciences and Technology 55 (5): 050607-
1–050607-8, 2011.
• Reynolds, L. 1979. “Progress in Documentation – Legibility Studies: Their Relevance to Present-Day Docu-
mentation Methods.” Journal of Documentation 35 (4): 307–340.
• Reynolds, L. 1988. “Legibility of Type.” Baseline: International Typographic Journal 10: 26–29.
• Tracy, W. 2003. Letters of Credit: A View of Type Design. Boston: David R. Godine.
• Vuk, D. 2000. Uvod v ekološki management. Ljubljana: Založba moderna organizacija.
• Zakon o varstvu okolja. URL: http://arhiv.mop.gov.si/si/zakonodaja_in_dokumenti/veljavni_predpisi/
zakon_o_varstvu_okolja (last accessed on 27. 12. 2017).

Corresponding author:
Klementina MOŽINA
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Chair of Information and Graphic Arts Technology,
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 42, fax: +386 1 200 32 70, e-mail: [email protected]

Co-author(s):
Alenka TOMAŠIČ
University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design,
Chair of Information and Graphic Arts Technology,
Snežniška 5, SI-1000 Ljubljana, Slovenia
phone: +386 1 200 32 42, fax: +386 1 200 32 70, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 266

 Poster section

USE OF BIOCOMPOSITE FILAMENTS FOR 3D PRINTING

Maruša ROPRET, Deja MUCK

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia

ABSTRACT: In our researche we focused on printing with biocomposite materials based on PLA and various
additives, such as (wood, hemp, and nanocellulose fibers with and without lignin). We have tested the influ-
ence on various agents such as water, oil and detergent according to the standard SIST ISO 2836: 2004 on the
printed objects. The prints were also analyzed on lightfastness and thermal aging in different temperature
intervals. After all mentioned analysis, colour differences were calculated. The results presented in our research
showed that filaments with nanocellulose fibers were higher sensitive on water, oil, detergent than filaments
with wood and hemp fibers. The similar results were achieved after lightfastness and thermal aging.

Keywords: biocomposite filaments, 3D printing, influence of agents, abrasion resistance, thermal aging.

1 INTRODUCTION

Over the last two decades, the interest in biocomposite materials has increased in various fields as mechan-
ics, medicine, architecture, construction, thanks to their availability, competitive quality, sustainability and
environmentally friendly features.
For many years the industry has shown an interest in the use of natural fibers, for example, flax, jute, industrial
cannabis, cotton and sisal for the production of a variety of end products, which are ultimately cheaper, have
less weight and less negative effect for the environment (Padney, 2015).
Biocomposites are composite material, which structure is composed of two or more starting materials. The
advantages of biocomposits are improvement of mechanical, chemical and physical properties after the start-
ing materials were combined together. Basically, these materials are composed of matrix, polymer foundation
often from renewable resources (mostly biodegradable), and natural fibers (biofibres) from wood, hemp, cot-
ton, bamboo etc. General classification of biocomposites defines three main groups: porous materials, fibrous
composites and articulated composites.
Material extrusion is one of the seven categories of additive technologies defined by the standard (ISO, 2015).
The most widely used technologies based on material extrusion, are those with thermoplastic materials, i.e.
thermoplastic extrusion.
Also, in the field of 3D printing new materials are already made from a mixture of materials, for example,
wood fibers and thermoplastic matrix (Muck, 2015). Most biocomposite materials are made from natural
fibers in combination with PLA thermoplastic matrix (polyacetic acid), which is also organic, and therefore
biodegradable and consequently friendly to the environment (Rosainvest, 2013).
At 3D printing with biocomposite materials, there are several problems compared with conventional thermo-
plastics materials, but there are very little researches on the runability, printability and mechanical properties
of mentioned materials (Tisserat, 2015). Besides, the review of literature also reveals the lack of researches
regarding the chemical properties and resistance to selected substances of the objects printed from biocom-
posite materials.
In the case of wood biocomposite materials, the final properties of the printed object depend on the orien-
tation of the model in the process of printing, the porosity of the basic filament (which increases with the
diameter of the filament) and the hygroscopicity of biocomposite materials.
The tensile strength of the objects printed with thermoplastic extrusion is therefore smaller compared to the
objects produced by, for example, injection molding, where high pressures improve the mechanical properties
of finished products (Le Duigou, 2016).
Thus, the question arises - how the addition of natural fibers in a biocomposite material changes the final
properties of the printed objects.
In addition to the basic polymer, polymer matrix and additives, which dictate the desired properties of the
final printed object, the mechanical properties are influenced by compatibilizers, which contribute to better
interactions between substances in composite polymer blends, and can also affect the final mechanical prop-
erties of the material (Yu, 2006).

8th Conference on Information and Graphic Arts Technology 267

 Poster section

2 EXPERIMENTAL

2.1 Materials

In our researches we focused on commercial and non-commercial biocomposite thermoplastic materials in the
shape of filaments for 3D printing. All mentioned materials were based on PLA matrix.
As commercial biocomposite materials, filaments with 20% wood (PLA_wood) and hemp (PLA_hemp) additives
were used. Pure PLA filament was used as a reference (PLA).
From non-commercial biocomposite three different filaments were used: PLA with 5% nanofibrilated cellulose
fibers (PLA_5NCF), PLA with 5% nanofibrilated cellulose fibers and 5% lignine (PLA_5NCF_5L), PLA with 5%
nanofibrilated cellulose fibers and 5% polimerized lignine (PLA_5NCF_5PL)

2.2 Methods

All samples were fabricated with ZMorph multitool 3D printer at the temperature of thermal heating plate
at 80°C and extruder printing head at 195°C.

Influence on various agents

The influence on various agents such as water (tap water), oil (vegetable sunflower) and detergent (1% dish-
washing detergent) according to the standard SIST ISO 2836: 2004 were tested on printed samples with di-
mension of 20 × 50 × 3 mm (w × l × t). Some samples were exposed to water for 24h, some to detergent for 3h
and some to oil for 48h. Before and after testing the L* a* b* values were measured and at the end ΔE was
calculated on the base of equation (1).

∆E= √(L*2 + a*2 + b*2) (1)

Light resistance, lightfastness

The testing was performed using a Xenotest Alpha (Atlas, USA) source in accordance with ISO 12040 standard.
The testing samples were printed with dimensions of 45 × 15 × 2 mm (w × l × t) and exposed to Xe-light for 72
h. After the lightfastness test the colour differences were determined based on the equation (1).

Thermal aging / artificial aging

Thermal aging at different temperature intervals (80°C, 110°C and 130°C) was performed only on samples
printed using commercial filaments. The printed samples were 75 mm long in eprouvette shapes type 1BA
according to standard ISO 527-2. After 7 days of thermal aging at each mentioned temperature intervals, the
colour differences were measured based on the equation (1).

3 RESULTS & DISCUSSION

Influence of various agents

In Figures 1, 2 and 3 water, detergent and oil resistance of commercial and non-commercial filaments. For
commercial three and for non-commercial two samples are presented.

Figure 1. Water resistance of commercial and non-commercial filaments.

8th Conference on Information and Graphic Arts Technology 268

 Poster section

Figure 2. Detergent resistance of commercial and non-commercial filaments.

Figure 3. Oil resistance of commercial and non-commercial filaments.

All filaments, commercial and non-commercial were highly resistant to water and detergent (Figures 1 and 2).
The measurements showed that samples printed with non-commercial filaments were more sensitive to water
and detergent. Visible differences in colour were found in resistance to oil (Figures 3). From all printed samples
PLA-wood, PLA_5NFC_5L and PLA_5NFC_5PL, the latter showed on average the highest colour changes. It is
also obvious that NFC as a sole additive improved oil resistance to PLA filament in comparison to additives
wood and hemp, however when it is used as an additive together with lignin and polymerized lignin the results
showed obvious decrease of resistance to oily agent.

Light resistance, lightfastness

In Figure 4 light resistance of commercial and non-commercial filaments is shown.

Figure 4: Light resistance of commercial and non-commercial filaments.

From the calculated colour differences it can be seen that samples printed with commercial filaments are more
resistant to light than samples printed with non-commercial filaments (Figure 4). At commercial filaments the
addition of wood fibers (PLA_wood) affected the colour degradation and it was very evident also with the

8th Conference on Information and Graphic Arts Technology 269

 Poster section

naked eye. The measured colour difference was more than 6. PLA and the filament with the addition of hemp
fibers (PLA_hemp) were more stable. Their measured colour differences vary between 1 and 3.

The results achieved after Xenotest for printed samples with non-commercial filaments showed higher colour
differences with the exception of samples printed with PLA_5NFC. The final achieved colour difference was
approx. 5. The value was somewhere between samples printed with commercial filaments (PLA_hemp and
PLA_wood). Samples printed with filaments on the base of PLA and the addition of lignin and polymerized
lignin (PLA_5NFC_5L and PLA_5NFC_5PL) showed higher colour differences. The highest measured value which
was almost 10 was achieved at PLA_5NFC_5PL.

Thermal aging / artificial aging

Thermal aging was tested only on samples printed with commercial filaments. As shown on Figures 5, 6 and
7 the colour degradation correlated with increasing temperature. Samples which were 7 days exposed at 80°C
did not show obvious visible colour degradation and colour differences were measured in average from 2 to
4,43. The highest value was measured on samples PLA_hemp. After the increasing temperature up to 110°C the
samples printed with PLA and added wood fibres (PLA_wood) drastically changed the colour while samples
printed with PLA and PLA_hemp filaments stayed the same as they were at 80°C. The colour differences at
samples printed with PLA_wood achieved final colour differences from 8,87 to 13,69. The highest temperature
that was used for temperature aging was 130°C. After 7 days samples didn’t change only in colour, but also
in their shape. From all samples, PLA_wood degradated in very high rate. The highest measured colour dif-
ferences was 31,01.

Figure 5. Thermal resistance of commercial filaments at 80°C.

Figure 6. Thermal resistance of commercial filaments at 110°C.

8th Conference on Information and Graphic Arts Technology 270

 Poster section

Figure 7. Thermal resistance of commercial filaments at 130°C.

4 CONCLUSIONS

On the basis of all the results we can summarize the following; if we use 3D printed items based on commer-
cial or non-commercial filaments as final products (decorative items, toys) we have to be careful not to come
in contact with oil, items should not be exposed to direct light and excessive temperatures above 80°C. In the
long-term contact with oil the samples printed with filament PLA_5NFC were shown as the most resistant.

If products printed on the basis of commercial or non-commercial filaments would be exposed to light for
the long period of time, the degradation of colour may occur. The most light-resistant samples were printed
with PLA and PLA_hemp filament. After 72 hours of exposure to Xenon light samples printed with PLA_wood,
PLA_5NFC_5L and PLA_5NFC_5PL showed visible color changes.

Elevated temperatures above 80°C greatly affected not only on colour changes, but also on the deformed
shape of the samples. The highest temperature stability showed samples printed with PLA filament without
any additives. It is noteworthy that additives of hemp or wood fibers greatly reduce temperature resistance.

For samples printed on the basis of PLA with added hemp fibers (PLA_hemp) the deformed shapes were more
obvious than color change. The worst thermal stability was shown on the samples printed with PLA_wood
filament. In conclusion the commercial filaments have proven to be more resistant to all of the performed
tests than the non-commercial filaments. The addition of fibers to the basic matrix PLA mostly reduced both
light and temperature resistance.

5 REFERENCES

• ISO/ASTM 52900:2015(en). ISO Online Browsing Platform (OBP). URL: https://www.iso.org/obp/ui/#iso:st-

d:iso-astm:52900:ed-1:v1:en/ (last accessed on 20. 2. 2017).
• Padney, K. 2015. Natural Fibre Composites For 3D Printing: Degree thesis. Theseus.fi Open Repository of the
Universities of Applied Sciences. URL: https://www.theseus.fi/bitstream/handle/10024/96363/Pandey_Ka-
pil.pdf?sequence=1/ (last accessed on 15.2.2017).
• Muck, T. and Križanovskij, I. 2015. 3D-tisk. Ljubljana : Pasadena.
• Podjetje Rosainvest 3D tisk. 3D tisk. Splošne lastnosti PLA termoplatičnega materiala. 2013. URL: http://
www.3dtisk.si/pla-material-fdm-spajanje-slojev/ (last accessed on: 4. 8. 2015).
• Tisserat, B., Liu, Z., Finkenstadt, V., Lewandowski, B., Ott, S. and Reifschneider, L. 2015. 3D printing biocom-
posites. Plastics Research Online. URL: http://www.4spepro.org/pdf/005690/005690.pdf/ (last accessed on:
15. 2. 2017).
• Le Duigou, A., Castro, M., Bevan, R. and Martin, N. 2016. 3D printing of wood fibre biocomposites: From
mechanical to actuation functionality. Materials and Design 96: 106–114.
• Yu, L., Dean, K., Li, L. 2006. Polymer blends and composites from renewable resources. ScienceDirect.
Obnovljeno 6. 6. 2006 [citirano 21. 2. 2017]. URL: http://www.sciencedirect.com/science/article/pii/
S0079670006000414/ (last accessed on: 21. 2. 2017).

8th Conference on Information and Graphic Arts Technology 271

 Poster section

Corresponding author:
Deja MUCK
Department of Textiles, Graphic Arts and Design
Faculty of Natural Sciences and Engineering, University of Ljubljana
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 84, fax: +386 12003270, e-mail: [email protected]

Co-author(s):
Maruša ROPRET
Department of Textiles, Graphic Arts and Design
Faculty of Natural Sciences and Engineering, University of Ljubljana
Snežniška 5, 1000 Ljubljana, Slovenia
phone: +386 1 200 32 11, fax: +386 12003270, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 272

 Poster section

USING UAV RGB CAMERA AS A COLOUR MEASURING DEVICE

Andrej UČAKAR, Aleš HLADNIK

University of Ljubljana, Faculty of Natural Sciences and Engineering,
Department of Textiles, Graphic Arts and Design, Snežniška 5, 1000 Ljubljana, Slovenia

ABSTRACT: Today unmanned aerial vehicles (UAV) – drones – are very popular and widespread in various
applications. Traditional multichannel systems for acquiring spectral images are often used only in labora-
tories, museums, as a part of research equipment and are very expensive and inflexible to be used on drones.
Our experiments show that reconstructed spectral reflectances from RGB pictures, taken by camera-drone are
quite promising, but are very much dependent on the choice of the training sets of colour chart patches. The
best results are achieved when the imaging colours correspond to the training samples of the colour charts.
We proved that the adoption of a drone-based camera is an efficient way to achieve good results in terms of
colour accuracy, drone flexibility, ease of use, usability for a wide range of applications, way of capturing
inaccessible areas where the use of commercial means of transport is not possible.

Keywords: unmanned aerial vehicles (UAV), RGB camera, spectral reflectance reconstruction, colour accura-
cy.

1 INTRODUCTION

Today unmanned area vehicles (UAV) – drones – are very popular and widespread in various application areas
such as aerial photo and video, remote sensing, animals tracking, etc (Klemas, 2015). In a few cases drones are
an indispensable part of the equipment, such as in the filming industry, military, exploration of the areas dan-
gerous for humans - volcanoes, natural disasters: fires, floods, spillage of hazardous substances, observation
of electric power lines, research of the consequences of avalanches and elsewhere (Madden, 2015).

Traditional multichannel systems for acquiring spectral images are still mostly implemented in research,
such as for restoration and conservation in the field of cultural heritage, remote sensing, forestry and geol-
ogy, biomedical research, quality control in industry, medicine in determining the presence of melanin and
hemoglobin, astronomy, in a printing workflow with high precision color rendering (Hardeberg, 1999). They
are often used only in laboratories, museums, as a part of research equipment and are very expensive and
inflexible to be used on drones.

The acquisition of spectral images with a conventional spectral camera is slow, and the mobility of the equip-
ment is poor. These reasons lead us to the estimation of spectral images from the RGB image acquired by
drone-based camera. With this method, collecting spectral data of pictures is much more efficient and easier
because of the RGB camera’s good mobility and fast image acquiring time (Miyata, 2003).

The aim of our study was to prove that drones can also be used for effective and reliable colour measure-
ments. Nowadays drones have the ability to use high quality cameras with very good characteristics such as
large sensor size, high resolution, very good optics. By applying appropriate data filtering algorithms, such
as Wiener estimation procedure, the images captured by a drone equipped with a standard 3-channel RGB
camera will be used to reconstruct spectral reflectance of an imaging surface with the camera signals as an
input (Cheung, 2005, DiCarlo, 2003).

2 MATERIALS AND METHODS

Our platform for capturing images consists of a quadcopter DJI Mavic Pro Platinum with 3-axis stabilized
gimbal system and 3-channel RGB camera equipped with Sony Exmor 1/2,3’’ sensor with 12,76 illion pixels.
Despite its compact size, a drone contains (in addition to a camera) a large number of sensors and high degree
of complexity, providing smart functions such as avoiding obstacles and collisions, tracking system, etc. It is
equipped with 24 high-performance CPU cores, a new transmission system with a 7km range, 5 vision sensors
which ensures redundancy. The maximum image size used for capturing is 4000 x 3000 pixels and the images
are saved in Adobe’s DNG raw format.

8th Conference on Information and Graphic Arts Technology 273

 Poster section

Starting with an acquisition the linearized camera responses have to be transformed into a high dimension-
al device-independent space (a spectral or multi-illuminant space). Various methods have been developed
for this purpose, which can be classified into two basic categories: Target-based methods and model-based
methods. The basic strategy of all these methods is to use as much information as possible of the underlying
capture process. Target-based methods use captured colours with known reflectances in order to construct a
response-to-reflectance transformation and apply this transformation on other captured images (Imai, 1999,
Zhao, 2007, Li, 2005). The training target needs to be chosen very carefully and has to include representative
spectra of the processed images. The main advantage of this class of methods is that a prior knowledge of the
acquisition illuminant and device model parameters is not necessary. The main drawback is the high depend-
ency of the spectral and colorimetric accuracy on the target choice (Mohammadi, 2004).

3 EXPERIMENTAL

In our research the estimation of the spectral image was done using the Wiener estimation method (Stigell,
2007), which was based on the use of a priori knowledge. Xrite ColorChecker Digital SG Colour Chart was
attached on a neutral grey background and illuminated under 45° with Rotolight LED Anova Pro 2 Bi-Color
110° light, set to colour temperature of 3200 Kelvin (Figure 1).

Out of 140 pads of the color chart we randomly chose 93 (i.e. two thirds) patches as training samples. These
patches were measured with the spectrophotomer Xrite i1Pro2 to get spectral data and also captured with
the drone-based RGB camera in Adobe’s DNG raw format (and transformed using Adobe Camera Raw 10.0
converter to TIFF format without compression, without colour profile attached and colour temperature set to
3200K Kelvin as set on LED lights) to get RGB data. Wiener estimation algorithm was applied to this data to
get spectral estimations for selected patches.

RGB data was also acquired for the remaining 47 colour pads – test samples – for which spectral estimations
via Wiener method and spectral measurements were obtained as well. Later we compared estimated and
measured values of 47 test colour pads and also calculated RMSE (Root Mean Square Error), PSNR (Peak sig-
nal-to-noise ratio) and dE2000 values to see how well the estimated curves fit the measured spectral data and
how pronounced are the colour differences.

8th Conference on Information and Graphic Arts Technology 274

 Poster section

Figure 1. Workflow of the study.

4 RESULTS & DISCUSSION

4.1 Performance metrics (RMSE, PSNR and dE2000)

For the algorithm performance evaluation, we specifically calculated PSNR, RMSE and dE2000 between meas-
ured and estimated reflectance for several monochromatic, saturated and skin tones of the color chart pads
(Figure 2).

Figure 2. Patch No. 64 with the highest dE2000 (left) and patch No. 45 with the lowest dE2000 (right).

8th Conference on Information and Graphic Arts Technology 275

 Poster section

Figure 3. Patch No. 95 with the lowest RMS and highest PSNR.

When calculating CIE dE2000, first step was to compute XYZ values from measured spectra, second step was
to compute XYZ values from estimated spectra of patches. CIE dE2000 formula was then applied to get CIE
dE2000 color difference values. XYZ and Lab values were calculated using D50 as a light source and CIE standard
observer defined in 1931 with a 2° field of view.

4.2 RMS, PSNR and dE2000 on monochromatic, saturated and skin tones

Table 1. PSNR, RMS and dE2000 values for 10 patches of ColourChecker Digital SG chart.
Patch no. RMS PSNR dE2000 R G B colour
64 0,123 66,3 12,30 198 10 34
74 0,113 67,1 11,84 255 225 44 SATURATED
94 0,062 72,3 10,89 105 167 222
45 0,057 73,0 1,79 245 245 246
MONO -
65 0,105 67,7 5,57 205 207 210
CHROMATIC
95 0,012 86,6 9,15 46 47 50
58 0,064 71,8 2,57 221 191 170
87 0,086 69,5 10,67 181 125 8 SKIN
97 0,037 76,8 3,79 225 185 149 TONES
98 0,026 79,7 2,33 227 191 154

From calculated RMS, PSNR and CIE dE2000 values it is possible to conclude that the highest dE2000 colour
difference between the measured and the estimated patches occurs with the saturated colors, followed by the
patches of skin tones and monochromatic patches. It is also possible to see that patches with low RMSE and
high PSNR values tend to have low dE2000 colour differences.
Our experiments show that if most of the parameters are carefully controlled, spectral reflectance results are
quite promising, but are very much dependent on the choice of the training sets of colour chart patches. The
best results are achieved when the imaged samples’ colours correspond to the training samples of the colour
charts.

5 CONCLUSION

With improvements in the applied algorithm we could obtain a better performance and after optimization
employed, the reconstruction method can be used in other applications as well (Shen, 2006, 2008). We proved
that the adoption of a drone-based camera is a very efficient way to achieve good results in terms of colour
accuracy, drone flexibility, ease of use, usability for a wide range of applications, way of capturing inaccessible
areas where the use of commercial means of transport is not possible.
In our further research for better colour measurements and certain applications we could use drone with
camera that do not include IR CUT filter, or use other techniques such as model based methods which are
dependent on the light source and camera sensor spectral sensitivities. It would also be interesting to make

8th Conference on Information and Graphic Arts Technology 276

 Poster section

measurements under different conditions such as different light sources with various colour temperature and
using different ColorCharts with much higher number of color patches.

6 REFERENCES

• Cheung, V., Westland, S., Li, C., Hardeberg, J. and Connah, D., 2005. “Characterization of trichromatic color
cameras by using a new multispectral imaging technique,” J. Opt. Soc. Am. A 22, 1231-1240.
• DiCarlo, J. M. and Wandell, B. A., 2003. “Spectral estimation theory: beyond linear but before Bayesian,” J.
Opt. Soc. Am. A, 20, 1261-1270.
• Hardeberg, J. Y., 1999. “Acquisition and reproduction of color images: colorimetric and multispectral ap-
proaches,” Ph.D. dissertation (Ecole Nationale Superieure des Telecommunications).
• Li, C., Luo, M. R. “A novel approach for generating object spectral reflectance functions from digital cameras.”
In IS&T/SID, pages 99–103, Scottsdale Ariz., 2005.
• Imai , F. H. and Berns, R. S., “Spectral estimation using trichromatic digital cameras.” In Intl. Sym. Multispec-
tral Imaging and Color Reproduction for Digital Archives, pages 42–49, Chiba University, 1999.
• Klemas, V., 2015. “Coastal and Environmental Remote Sensing from Unmanned Aerial Vehicles: An Overview”.
Journal of Coastal Research: Volume 31, Issue 5: pp. 1260 – 1267.
• Madden, M., Jordan, T., Cotten, D., O’Hare, N. and Pasqua, A.,. “The Future of Unmanned Aerial Systems
(UAS) for Monitoring Natural and Cultural Resources”, URL: http://www.ifp.uni-stuttgart.de/publications/
phowo15/350Madden.pdf (last accessed on 10. Januar. 2017). Wichmann/VDE Verlag, Belin & Offenbach,
369-385, 2015.
• Mohammadi, M., Nezamabadi, M., Berns, R. S, and Taplin, L. “Spectral imaging target development based on
hierarchical cluster analysis.” In IS&T/SID, 12th Color Imaging Conference, pages 59–64, Scottsdale Ariz., 2004.
• Miyata, K., Kurushima H., “Digital Systems for Historical Documents–Improvements of Legibility by Frequen-
cy Filters,” in Proceedings IS and T’s 2003 PICS Conference (Rochester NY, USA, 2003), pp. 291–295.
• Shen, H.-L., Cai, P.-Q., Shao, S.-J. and H. Xin, John. Reflectance reconstruction for multispectral imaging by
adaptive Wiener estimation. . URL: https://www.osapublishing.org/DirectPDFAccess/EE9D2387-9F02-4A7F-
30C5BC9B76E24075_144761/oe-15-23-15545.pdf?da=1&id=144761&seq=0&mobile=no/ (last accessed on 15.
Januar. 2017). OPTICS EXPRESS, 2007 Optical Society of America, Vol. 15, No. 23, November 2007
• Shen, H. L., Xin, J. H., 2006. “Spectral characterization of a color scanner based on optimized adaptive esti-
mation,” J. Opt. Soc. Am. A 23, 1566-1569.
• Stigell, P., Miyata, K. and Hauta-Kasari, M. 2007, “Wiener estimation method in estimating of spectral reflec-
tance from RGB images.“ Pattern Recognition and Image Analysis, 2007, Vol. 17, No. 2, pp. 233–242. Pleiades
Publishing, Ltd., 2007
• Zhao,Y. and Berns, R. S. “Image-based spectral reflectance reconstruction using the matrix R method.” Color
Research and Application, 32(5):343–351, 2007.

Corresponding author:
Andrej UČAKAR
Chair of Information and Graphic Arts Technology
Faculty of Natural Sciences and Engineering, University of Ljubljana
Snežniška 5
SI-1000 Ljubljana, Slovenia
phone: + 386 1 200 32 79, fax: + 386 1 200 32 70, e-mail: [email protected]

8th Conference on Information and Graphic Arts Technology 277