A

B
C
D
E
F
G
UNI VERSI TY OF OULU P. O. Box 7500 FI - 90014 UNI VERSI TY OF OULU FI NLAND
A C T A U N I V E R S I T A T I S O U L U E N S I S
S E R I E S E D I T O R S
SCIENTIAE RERUM NATURALIUM
HUMANIORA
TECHNICA
MEDICA
SCIENTIAE RERUM SOCIALIUM
SCRIPTAACADEMICA
OECONOMICA
EDITOR IN CHIEF
EDITORIAL SECRETARY
Professor Mikko Siponen
Professor Harri Mantila
Professor Juha Kostamovaara
Professor Olli Vuolteenaho
Senior Assistant Timo Latomaa
Communications Officer Elna Stjerna
Senior Lecturer Seppo Eriksson
Professor Olli Vuolteenaho
Publication Editor Kirsti Nurkkala
ISBN 951-42-8213-2 (Paperback)
ISBN 951-42-8214-0 (PDF)
ISSN 0355-3213 (Print)
ISSN 1796-2226 (Online)
UNI V E RS I TAT I S OUL UE NS I S
ACTA
C
TECHNICA
OULU 2006
C 256
Erkki Alarousu
LOW COHERENCE
INTERFEROMETRY AND
OPTICAL COHERENCE
TOMOGRAPHY IN PAPER
MEASUREMENTS
FACULTY OF TECHNOLOGY,
DEPARTMENT OF ELECTRICAL AND INFORMATION ENGINEERING,
INFOTECH OULU,
UNIVERSITY OF OULU
C

2
5
6
A
C
T
A

E
r
k
k
i

A
l
a
r
o
u
s
u
C256etukansi.fm Page 1 Tuesday, November 14, 2006 1:26 PM
ACTA UNI VERS I TATI S OULUENS I S
C Te c h n i c a 2 5 6
ERKKI ALAROUSU
LOW COHERENCE
INTERFEROMETRY AND OPTICAL
COHERENCE TOMOGRAPHY
IN PAPER MEASUREMENTS
Academic dissertation to be presented, with the assent of
the Faculty of Technology of the University of Oulu, for
public defence in Auditorium TS101, Linnanmaa, on
November 24th, 2006, at 12 noon
OULUN YLI OPI STO, OULU 2006
Copyright 2006
Acta Univ. Oul. C 256, 2006
Supervised by
Professor Risto Myllyl
Reviewed by
Professor Erik Vartiainen
Professor Dmitry Zimnyakov
ISBN 951-42-8213-2 (Paperback)
ISBN 951-42-8214-0 (PDF) http://herkules.oulu.fi/isbn9514282140/
ISSN 0355-3213 (Printed)
ISSN 1796-2226 (Online) http://herkules.oulu.fi/issn03553213/
Cover design
Raimo Ahonen
OULU UNIVERSITY PRESS
OULU 2006
Alarousu, Erkki, Low coherence interferometry and optical coherence tomography in
paper measurements
Faculty of Technology, University of Oulu, P.O.Box 4000, FI-90014 University of Oulu, Finland,
Department of Electrical and Information Engineering, Infotech Oulu, University of Oulu, P.O.Box
4500, FI-90014 University of Oulu, Finland
Acta Univ. Oul. C 256, 2006
Oulu, Finland
Abstract
This thesis describes the application of Low Coherence Interferometry (LCI) and Optical Coherence
Tomography (OCT) in paper measurements. The developed measurement system is a combination of
a profilometer and a tomographic imaging device, which makes the construction versatile and
applicable in several paper measurement applications. The developed system was first used to
measure the surface structure of paper.
Different grades of paper were selected to provide maximum variation in surface structure. The
results show that the developed system is capable of measuring grades of paper from rough base paper
to highly coated photo printing paper.
To evaluate the developed system in surface characterization, the roughness parameters of five
laboratory-made paper samples measured with the developed system and with a commercial optical
profilometer were compared. A linear correlation was found with roughness parameters R
a
and R
q
.
Next, the surface quality of paper was evaluated using LCI, a Diffractive Optical Element Based
Glossmeter (DOG), and a commercial glossmeter. The results show linear correlation between R
a
and
gloss measured with the commercial glossmeter. The roughness R
a
and averaged gloss measured with
the DOG didn't give such a correlation, but a combination of these techniques provided local
properties of gloss and surface structure, which can be used to evaluate the local surface properties of
paper.
In the next study, determination of the filler content of paper using OCT is discussed. The
measurement results show clear correspondence of the slope of the averaged logarithmic fringe signal
envelope and the filler content.
The last studies focus on 2D and 3D imaging of paper using OCT and begin with imaging of a
self-made wood fiber network. The visibility of the fibers was clear. Next, several refractive index
matcing agents are studied by means of light transmittance and OCT measurements to find the best
possible agent for enhancing the imaging depth of OCT in paper. Benzyl alcohol was found to have
the best possible combination of optical, evaporation, and sorption characteristics, and it is applied in
2D and 3D visualizations of copy paper.
Keywords: filler, gloss, imaging, refractive index, roughness, wood fiber
Acknowledgements
The work reported in this thesis was carried out at the Optoelectronics and Measurement
Techniques Laboratory, Department of Electrical and Information Engineering, Faculty of
Technology, University of Oulu, during 2001-2006. I wish to express my deepest
gratitude to Professor Risto Myllyl for supervising this work.
This research was done partly in collaboration with the Department of Physics,
University of Joensuu, and I wish to thank Professor Kai Peiponen and M.Sc Mikko Juuti
for the fruitful discussions concerning paper III.
I wish to thank Professor Dmitry Zimnyakov and Professor Erik Vartiainen for
reviewing this thesis. Thanks also to Mr. Keith Kosola for revising the English of the
manuscript and Mr. Rauno Varonen for revising the English of the original papers. I also
wish to thank the Infotech Oulu graduate school for their financial support and travel
grants. Financial support from Seppo Synjkankaan tiedesti, Tekniikan
edistmisti, Tauno Tnningin sti, and Kaupallisten ja Teknillisten tieteiden
tukisti is gratefully acknowledged.
I thank all my fellow workers at the Optoelectronics and Measurement Techniques
Laboratory and in the Workshop of the Department, especially Mr. Vesa Kaltio for the
considerable mechanical work he did with the interferometer, DI Tuukka Prykri for his
assistance in most of the measurements and for the great discussions, DI Leszek Krehut
for developing the electronics for the measurement system, DI Tapio Fabritius for the
research and collaboration in paper V, and Dr. Jukka Hast for his support and technical
advice in the early stage of this work.
Finally, I would like to thank my wife Katja and my son Julius for their great support
and patience, and my mother and father for their support and encouragement during this
project.
Oulu, August 2006 Erkki Alarousu

List of terms, symbols, and abbreviations
2D Two-dimensional
3D Three-dimensional
ACF Autocorrelation function
AD Analog to digital
ASCII American standard code for information interchange
BS Beam splitter
DAQ Data acquisition
DOCT Doppler optical coherence tomography
DOG Diffractive optical element based glossmeter
FD-OCT Fourier domain optical coherence tomography
GU Gloss unit
LCI Low coherence interferometry
LSQ Least squares
NA Numerical aperture
OCDR Optical coherence domain reflectometry
OCT Optical coherence tomography
OLCR Optical low coherence reflectometry
PC Personal computer
PCC Precipitated calcium carbonate
PCI Partial coherence interferometry
PSD Power spectral density
ROI Region of interest
SC Supercalendered
SLD Superluminescent light emitting diode
TOF Time of flight
WLI White light interferometry

x Transversal resolution
c Speed of light
C Coherence function
d Beam diameter
d
g
Geometric depth in the medium
E
r
Reference field
E
s
Sample field
E

s
Path length-resolved sample field density
f Frequency
f
0
Center frequency
f
d
Doppler frequency
f
l
Focal length
i Number of data points in the transversal direction
I
d
Intensity of the photodetector
I
r
Intensity from the reference arm
I
s
Intensity from the sample arm
I
signal
Signal carrying intensity
k Specific absorption coefficient
L
c
Coherence length
L
c,m
Coherence length in a dispersive medium
l
r
Geometric length of the interferometer reference arm
L
r
Optical length of the interferometer reference arm
l
s
Geometric length of the interferometer sample arm
L
s
Optical length of the interferometer sample arm
n Refractive index
n Number of effective reflecting layers
n
g
Group refractive index
R Path length-resolved diffuse reflectance
R
0
Reflectance from a single sheet
R

Reflectance relative to a perfect diffuser

R
a
Roughness average
R
q
Rms roughness average
R
ku
Kurtosis; a distribution of spikes above and below the mean line
R
p
Distance between the highest peak and the surface position average
R
sk
Skewness; symmetry of depth distribution
R
t
Distance between the highest peak and deepest pit
R
v
Distance between the deepest pit and the average surface position
R
z
Average value of distance between the five highest peaks and five deepest
pits
s Specific scattering coefficient
S Power spectral density
S
a
Average slope
S
q
Rms slope
std Standard deviation
t Time
t Layer transmittivity
r Layer reflectivity
V
mc
Normalized mutual coherence function
V
tc
Temporal coherence function

x Transversal coordinate
y Transversal coordinate
Z
k
Surface position at transversal coordinate k
Z
kl
Surface position at transversal coordinates k and l
Z
avg
Average surface position
z
r
Rayleigh range

Correlation length

0
Center wavelength of the light source
Full width at half the maximum wavelength range of the light source
Time delay
Phase difference


List of original papers
This thesis is a summary of the work published in the following six papers
I Alarousu E, Myllyl R, Gurov I & Hast J (2001) Optical Coherence Tomography in
Scattering Material for Industrial Applications. Proc. of SPIE 4595: 223-230.
II Alarousu E, Krehut L, Myllyl R & Hast J (2004) Optical coherence tomography
device for paper characterization. Proc. of SPIE 5475: 48-55.
III Peiponen K E, Alarousu E, Juuti M, Silvennoinen R, Oksman A, Myllyl R &
Prykri T (2006) Diffractive optical element based glossmeter and low coherence
interferometer in assessment of local surface quality of paper. Opt. Eng. 45, 043601.
IV Alarousu E, Prykri T, Hast J & Myllyl R (2004) Low Coherence Interferometry in
Paper Surface and Near-surface Characterization. Proc. of ODIMAP IV: 36-41.
V Fabritius T, Alarousu E, Prykri T, Hast J & Myllyl R (2006) Characterization of
Optically Cleared Paper by Optical Coherence Tomography. J. Quant. Electron.
36(2): 181-187. (DOI: 10.1070/QE2006v036n02ABEH013121)
VI Alarousu E, Krehut L, Prykri T & Myllyl R (2005) Study on the use of optical
coherence tomography in measurements of paper properties. Meas. Sci. Technol. 16:
1131-1137. (www.iop.org/journals/mst)
The research work described in this thesis was carried out during 2001-2006 in the
Optoelectronics and Measurement Techniques Laboratory, Department of Electrical and
Information Engineering and Infotech Oulu, University of Oulu, Finland.
Paper I describes the early development of Optical Coherence Tomography for
industrial applications. It introduces the idea of using OCT for paper measurements.
Paper II discusses the device used in the measurements in more detail and gives a first
example of a 2D cross-sectional image of the internal structure of paper. It also gives the
surface profiles of two different paper samples to provide an example of the multipurpose
use of the method.
Paper III discusses the evaluation of the surface quality of paper using three different
devices: a diffractive optical element based glossmeter (DOG), a commercial glossmeter,
and a low coherence interferometer (LCI). The surface topography measurements were
performed by the author and the paper was jointly written by the author, K. Peiponen, and
M. Juuti.
Paper IV deals with measuring the topography and roughness parameters of five
laboratory-made sheets of paper. The results measured using LCI are compared with
results measured with a commercial optical profilometer. A 3D wood fiber network was
measured to show the capabilities of the OCT technique in paper imaging.
Paper V introduces the use of refractive index matching liquids for 2D and 3D
imaging of paper. The author was responsible for the OCT measurements and the paper
was jointly written by the author and T. Fabritius.
Paper VI gives on overview of the latest results and provides an insight into the
capabilities of LCI in evaluating various kinds of paper surfaces, discusses the possibility
of measuring the filler content of paper, and finally shows the first 3D image of
commercial paper measured using OCT.

Contents
Abstract
Acknowledgements
List of terms, symbols, and abbreviations
List of original papers
Contents
1 Introduction ................................................................................................................... 15
1.1 Background and motivation of the work ................................................................ 15
1.2 Contribution of the thesis ....................................................................................... 16
1.3 Contents of this work.............................................................................................. 17
2 Paper structure and properties ....................................................................................... 18
2.1 Geometry and dimensions of a single wood fiber .................................................. 18
2.2 Paper geometry and composition............................................................................ 19
2.3 Optical properties of paper ..................................................................................... 20
2.3.1 Interaction of light with paper ......................................................................... 21
2.3.2 Models for simulating paper properties........................................................... 22
2.3.3 Paper gloss....................................................................................................... 22
2.4 Structural properties of the paper surface ............................................................... 23
3 Low Coherence Interferometry and Optical Coherence Tomography........................... 25
3.1 Introduction to Low Coherence Interferometry...................................................... 25
3.2 From Low Coherence Interferometry to Optical Coherence Tomography............. 29
3.3 Spatial resolution of an Optical Coherence Tomography imaging system............. 30
4 Measurement system..................................................................................................... 32
4.1 Interferometer ......................................................................................................... 33
4.2 Analog signal processing and control unit .............................................................. 34
4.3 Data acquisition...................................................................................................... 35
4.4 Digital signal and image processing....................................................................... 36
5 Low Coherence Interferometry and Optical Coherence Tomography in paper
measurements................................................................................................................ 38
5.1 Surface analysis of paper........................................................................................ 38
5.2 Surface quality evaluation of paper using LCI and a DOG.................................... 45
5.3 Filler content evaluation of paper ........................................................................... 48
5.4 2D and 3D structural imaging of paper .................................................................. 50
6 Discussion ..................................................................................................................... 57
6.1 Measurements......................................................................................................... 58
6.2 Future research ....................................................................................................... 60
7 Summary ....................................................................................................................... 61
References
Original papers




1 Introduction
1.1 Background and motivation of the work
Optical Coherence Tomography (OCT) was first introduced in 1991 (Huang et al. 1991).
It launched the development of the technique in various applications in the field of
medicine, and especially in opthalmology. OCT is based on low coherence interferometry
(LCI), also known as White Light Interferometry (WLI), Optical Coherence Domain
Reflectometry (OCDR), Optical Low Coherence Reflectometry (OLCR), and Partial
Coherence Interferometry (PCI). The principle was first introduced by Sir Isaac Newton
and later again by Fercher et al, Danielson et al., Youngwist et al., and Takada et al.,
among others (Danielson et al. 1987, Fercher 1986, Fujimoto 2002, Fujimoto 2003,
Masters 1999, Takada et al. 1987, Younguist et al. 1987, Kubota et al. 1987). An
important milestone in the development of these techniques was the emergence of high
brightness semiconductor broadband light sources like superluminescent light emitting
diodes (SLDs) in the early 1980s (Wang et al. 1982).
As stated above, after the discovery of OCT in the early 1990s, the main application
area has been in medicine. Industrial applications of the OCT technique are surprisingly
few, even though the first studies of its basis, LCI, dealt with testing of optical wave
guides (Hee et al. 1995).
Because of the strong forest industry base in Finland, it was a clear choice to explore
the capabilities of LCI and OCT in industrial applications. There is a growing need for
improved methods for diagnosing and characterizing manufactured products, especially
in the modern paper industry. By and large, paper testing tends to be focused on the goal
of ensuring that the produced product meets the end-user specifications at the lowest
possible cost (Hahn 2002, Levlin 1999).
Research on paper and pulp began in 1972 in the Optoelectronics and Measurement
Techniques Laboratory. A Masters thesis by Hirsimki studied the possibilities of
measuring the flocculation of fiber suspensions by using focused ultrasound (Hirsimki
1972). The results of one of the first studies in which an optical approach was applied
were published in 1995. In that study, papermaking pulp properties were estimated by
using time of flight (TOF) measurements (Karppinen et al. 1995). The research was
16
continued by Saarela et al. in 2000, when they did an experiment with changes in time of
flight in thermomechanical pulp (TMP) with a streak camera (Saarela et al. 2001). Paper
was measured for the first time in 2001, when Saarela et al. studied the change in time of
flight to determine paper porosity (Saarela et al. 2001). The first experiments with white
light interferometry were conducted in the Optoelectronics and Measurement Techniques
Laboratory in the late 1970s. A commercially available white light interferometer based
on a halogen light source was applied to plastic sheet thickness measurement.
The first experiment in which OCT was briefly applied to paper measurement was
published in 2000 (Fercher et al. 2000). In that experiment, a tungsten halogen low
voltage lamp was used as a light source in an OCT setup and two depth-resolved signals,
i.e. a-scans, were presented from a sheet of white paper, but nothing more. After that, no
other group, excluding our group, has presented any results on measuring paper using the
OCT technique.
The combination of a non-invasive, non-contact measurement method for imaging the
surface structure and internal structure of paper and simultaneous calculation of structural
parameters was the ultimate goal of this work. The motivation came from the industry,
where there is a need for a fast and cost-effective way to study the structure of paper.
Several conventional methods are separately used to determine the quality of paper.
1.2 Contribution of the thesis
The need to analyze and characterize paper is continually increasing in the modern paper
industry. Current measurement methods tend to be either slow, labor-intensive, expensive
or invasive. A combination of a profilometer and an imaging device would be a great tool
for the researcher in the structural characterization of paper. Such a device could be one
of the strongest alternatives to conventional techniques, or it could replace them in the
future as part of the standard methods used even in on-line quality control.
The industrial applications of the OCT technique are surprisingly few. Especially in
the case of paper, 2D and 3D imaging pose a great challenge because of the structural
complexity of paper, which unavoidably leads to optical complexity in the form of
complex photon migration in paper. The methods described in this thesis suggest that the
internal structure of paper is visible to OCT when the sample is pretreated by enhancing
its optical properties with appropriate agents.
The basic idea of this thesis is that a new industrial application is introduced in the
field of OCT research, and a new method is introduced for paper researchers. The aim of
this thesis is to prove that a combination of LCI and OCT can be applied to characterize
the paper surface and to image the internal structure of paper and determine parameters
like roughness and filler content on the basis of these measurements.
17
1.3 Contents of this work
In this thesis project, Low Coherence Interferometry and Optical Coherence Tomography
were applied to measure paper properties like roughness and filler content, and to image
the 2D and 3D structure of paper, with refractive index matching agents employed to
increase the imaging depth of the developed OCT system, and with a Michelson-type
free-space interferometer used as the core of the system.
Chapter 2 provides information about the structure and properties of paper; single
fiber properties, paper geometry and composition, and finally optical properties and
photon migration inside paper. This chapter also discusses different models that describe
the structure of paper.
Chapter 3 presents the theory of Low Coherence Interferometry and Optical
Coherence Tomography in detail. The spatial resolution of the system is also discussed
briefly.
Chapter 4 introduces the measurement system in detail. The system is divided into
four parts: (1) the interferometer, (2) an analog signal processing and control unit, (3)
data acquisition, and (4) digital signal and image processing. All four parts, the
measurement sequence, and communication between the parts are discussed in detail to
give an overview of how the measured signal is converted into the final image data.
The mesurements in chapter 5 begin with measurement of the paper surface. Several
topographies of various types of paper are presented. The roughness values of the
samples are compared with values measured using a commercial optical profilometer.
This chapter provides an overview of the capabilites of the LCI technique in measuring
highly different grades of paper.
Next, two methods, low coherence interferometry and diffractive optical element
based glossmetry (DOG), were utilized to give joint transversally localized information
on surface roughness and gloss that helps papermakers in their research and development
of optimal paper surface quality, which is crucial for optimal ink absorption in the
printing process. The commercial glossmeter was also used to give an average value for
the gloss.
Filler content evaluation is discussed next. Three samples with a different filler content
were measured by using the slope of the LSQ-fitted line to determine the amount of filler.
The last part of chapter 5 introduces the capabilities of OCT in determining the 2D and
3D structure of paper. Beginning from single a-scans of paper, the results show a 3D
image of a simulation network of paper, followed by 2D images of commerical copy
paper treated with different refractive index matching agents, and finally a 3D image of
copy paper treated with Benzyl alcohol as the matching agent. More detailed descriptions
of the measurements can be found in the original papers I-VI at the end of this thesis.
The results of the work and future research are discussed in chapter 6, and a summary
of the work is given in chapter 7.

2 Paper structure and properties
Paper is a stochastic network of fibers, but since the fibers are much longer than the
thickness of a paper sheet, the network can be treated as planar and almost two-
dimensional (Niskanen et al. 1998). Fiber is one of the basic units from which paper is
composed, and it is natural to begin the discussion of the structure of paper from the
individual fiber itself (Kolseth et al. 1986).
2.1 Geometry and dimensions of a single wood fiber
The dimensions of a wood fiber depend on the pulping process and raw materials used. A
rule of thumb is that the length of a softwood fiber is 100 times the fibers diameter. For
hardwood this ratio is in the order of 30-70. Additionally, the thickness of the fiber wall is
approximately one tenth of the fibers diameter. However, the cross-sectional dimensions
vary within one species, due to seasonal variations of earlywood and latewood fibers, and
also due to age and the growing environment of the stem (Heikkurinen 1999).
The most important softwood species for the Finnish paper industry are domestic pine
(Pinus) and spruce (Picea). These species are also the most common materials in the
worlds papermaking industry. The fibers of softwoods are closed at both ends.
Earlywood fibers have a thin wall and their cross-sectional shape is nearly square.
Latewood fibers have a thicker wall and a smaller lumen than earlywood fibers, and they
have a rectangular cross-section (VTT Automation et al. 2005).
One of the most important species of hardwoods for the paper industry is northern
birch (Betula pubescens and Betula verrucosa). Because of its excellent fiber properties,
it is one of the most suitable species for paper manufacturing. European manufacturers
prefer to use hardwoods such as eucalyptus, beech, and birch in pulp production due to
their higher availability (M-real Digital imaging 2004). However, their dimensions are
very different from those of softwood species; their fibers are shorter and less porous.
The dimensional variation between fibers of different species is larger in hardwood than
in softwood, but there is no clear seasonal variation in fiber dimensions. The fiber
dimensions of the most common softwood and hardwood species used in paper
manufacturing are presented in Table 1 (Heikkurinen 1999, VTT Automation et al. 2005).
19
Table 1. Typical cross-sectional dimensions of wood fibers (VTT Automation et al. 2005).
Birch Eucalyptus Pine,
earlywood
Pine,
latewood
Spruce,
earlywood
Spruce,
latewood
Fiber length
(mm)
1.1 1.0 2.9 2.9 2.9 2.9
Fiber diameter
(m)
22 16 35 20 33 19
Fiber wall
thickness (m)
3 3 2.1 5.5 2.3 4.5
2.2 Paper geometry and composition
As stated earlier, paper can be considered a continuous three-dimensional network of
solid material like fibers, fiber fragments, and possible fillers (Holmstad et al. 2002). In
the same manner the pore structure inside paper forms a continuous three-dimensional
network of voids (Samuelsen et al. 2001). Most of the fiber material is aligned along the
plane of the paper. In machine-made papers, the fiber orientation is anisotropic, which
means more fibers are aligned closer to the machine direction than to the cross-machine
direction (Niskanen et al. 1998). In addition, most of the fibers lose their original shape
and collapse, leading to the flat shape of their cross-section (Carlsson et al. 1995).
The arrangement of the fibers in the z-direction can be layered or felted. A layered
structure will form if the fibers land on the wire one after another and form an ordered
sequence in the z-direction. In a felted structure, the sequence is not clear (Niskanen
1998). Both cases are shown in Fig. 1.

Fig. 1. Layered (top) and felted (bottom) sheet cross-section (Niskanen 1998).
The cross-section shown in Fig. 1 is highly simplified. Paper consists of several smaller
particles in addition to fibers.
Fines are fragments of fibers. In addition to whole fibers, pulp always contains a
certain amount of smaller particles. In the pulpmaking process and in mechanical
treatment, fibers are damaged and surface layers, bunches of fibrils, and single fibrils are
20
loosened (VTT Automation 2005). A common definition of fines is that material that
passes through a 200 mesh is considered fines (Hiltunen 1999). In chemical pulp, the
amount of fines is about 5-10% and in mechanical pulp, about 30% (Heikkurinen 1999).
An increase in fines content will improve surface smoothness and the light scattering
coefficient (Nesbakk et al. 2001).
Most grades of paper also contain fillers, which are fine, white pigment powders. They
fill in the spaces between the fibers and smooth the paper surface. They also improve the
evenness of formation, printability, opacity, dimensional stability, gloss, and brightness,
but as a drawback, they reduce strength. Savings in manufacturing costs is also a
tempting aspect of using fillers; filler material is cheaper than fiber raw material
(Krogerus 1999). The most important fillers used in the papermaking process are: clay
(Al
4
Si
4
O
10
(OH)
2
), talc (Mg
3
Si
4
O
10
(OH)
2
), calcium carbonate (CaCO
3
), gypsum (CaSO
4
-
nH
2
O), precipitated calcium carbonate, titanium oxide (TiO
2
), and synthetic silicate
(Krogerus 1999, VTT Automation et al. 2005).
Base paper forms the framework of paper. In most cases paper is coated with pigments
to enhance its surface properties, like printability, appearance, and water impermeabilty
(Aaltio 1969). The coating layer is also a complex structure; in addition to pigment
particles, the coating material contains binders, thickeners, and additives (Chinga et al.
2002, VTT Automation 2005). The main pigments used in coating paste are clay, calcium
carbonate, talc, and gypsum (VTT Automation 2005).
2.3 Optical properties of paper
Optical properties like opacity, brightness, transmittance, gloss, and color usually
characterize the optical properties of paper (Borch 2002, Leskel 1998, Van Der Reyden
1992). These optical properties are not discussed in detail in this thesis, with the
exception of gloss in section 2.3.3. Gloss and its transversal distribution were one of the
parameters compared in parallel with the surface structure measured using LCI to
evaluate the surface quality of paper. In the case of LCI and OCT, the most important
features are light interaction and photon migration in paper. The complexity of light
propagation and estimation is discussed in section 2.3.1. Variations in scattering and
absortion coefficients in paper are high due to the large differences in paper composition;
fillers and coating colors vary greatly in different grades of paper, and they have an effect
on light interaction in paper. The models used to estimate these two coefficients are
discussed in section 2.3.2. Light propagation in paper is highly dependent on the
refractive indices of the paper materials. The refractive indices of the most common
constituents are shown in Table 2 (Aschan 1969, Krogerus 1999, Lee et al. 2005). The
values vary depending on the reference used.
Table 2. Refractive indices of the most common constituents of paper.
Material Cellulose Clay Talc Gypsum PCC Titanium oxide
Refractive
index
1.55 1.55-1.60 1.55-1.57 1.52 1.58-1.63 2.50-2.70
21
2.3.1 Interaction of light with paper
The interaction of light with paper is a complex process that depends on the structure and
composition of the paper. Four types of interactions are presented in Fig. 2 (Vaarasalo
1999). Part of the light hitting the paper surface is reflected back and the remainder
penetrates the sheet. After passing into paper, light can be transmitted through, scattered
back, or absorbed to form heat (Leskel 1998).

Fig. 2. Interaction of light with paper.
It is practically impossible to obtain information about the actual light scattering
phenomena in paper due to its complex structure and composition, which inevitably
cause multiple scattering events. Interferences in multiple scattering are solvable by using
either Maxwells equations or radiative transfer theory. The first choice is practically
intractable because of the random nature of the paper structure and the second one gives
simplified solutions because it doesnt take into account the superposition of
electromagnetic fields, but only the superposition of scattered intensities (Leskel 1998).
Carlsson et al. have modeled light propagation through paper using Monte-Carlo
simulations, but in their model the paper structure was highly simplified; fibers, fines,
fillers, and pores randomly appeared in the path along which the photon was traveling
(Carlsson et al. 1995). They also did experiments with time-of-flight measurement
through paper and suggested that the path length of the transmitted photon is
approximately ten times the thickness of a paper sheet (Carlsson et al. 1995). The
simplifications in the simulation models make estimation of light propagation in real
paper practically impossible, and for that reason most of the models typically dont
concentrate on giving exact information on the actual light scattering phenomena in
paper, but rather specify material properties like light scattering and absorption
coefficients (Leskel 1998). The absorption coefficient in most grades of paper is much
smaller than the scattering coefficient because most of the papermaking raw materials are
ineffective absorbers (Leskel 1998), and it is the scattering coefficient which mainly
defines the optical properties of paper.
Specular
reflection
(gloss)
Transmission Scattering
Absorption
(heat)
22
2.3.2 Models for simulating paper properties
The Kubelka-Munk theory and its extensions are the most widely used methods for
describing the optical properties of paper (Kubelka et al. 1931, Kubelka 1948, Kubelka
1954). The theory was originally developed for paint films, but it was later applied to
paper. The basic idea of the model is the assumption that the interaction between diffuse
light and the material can be described using phenomenological parameters: the specific
scattering coefficient s and the specific absorption coefficient k of the material calculated
from the reflectance values of R
0
and R

, which are the reflectance from a single sheet

and the reflectance relative to a perfectly reflecting diffuser, respectively (Karppinen
2004, Leskel 1998, Pauler 1986). A homogenous slab of even thickness was used in the
original model, where the material contains randomly distributed particles that scatter and
absorb light, but it was later extended by Kubelka to multilayer structures (Borch 2002,
Kubelka 1948, Kubelka 1954, Leskel 1998). The Kubelka-Munk theory is limited to
optically thick materials in which multiple scattering occurs and transmittance is low. The
material becomes optically thick when half of the light is reflected and 20% of the light is
transmitted. Grades of paper differ greatly in composition, so in many cases the
requirements of the Kubelka-Munk theory are not met and the theory can only be used
with caution (Pauler 1986). The fact is that no information on actual light scattering
events is obtained with the Kubelka-Munk theory. Another aspect of the theory is the
assumption that the material is homogenous or a stack of homogenous slabs that is far
from reality; spatial changes in paper properties are high due to the additives that are
added to enhance the paper properties. The Kubelka-Munk theories dont take into
account the scattering particle size, the refractive index, or the absorption coefficient
(Borch 2002).
Another model of paper was developed by Scallan and Borch (Scallan et al. 1972). In
this model paper is modeled as a stack of separate parallel layers that form a stack of
bonded fibers. Light scattering is assumed to take place mainly in fiber-surrounding
media interfaces. The parameters of the Scallan-Borch model are the number of effective
reflecting layers n, layer transmittivity t, and layer reflectivity r. The Scallan-Borch
model is limited to sheets containing only fibers, but no additives, so its also highly
simplified (Borch 2002, Karppinen et al. 2004, Leskel 1998).
The Scallan-Borch model was further modified by Leskel (Leskel 1993). His model
took into account additives like fillers and coatings in the paper model.
2.3.3 Paper gloss
Optical properties like gloss define what the paper looks like. The evaluation of gloss is
often made by human eye (Bland et al. 2000). When light hits the paper surface, a
portion of the light is specularly reflected. The gloss of paper is its ability to reflect light,
i.e. the degree to which the paper surface simulates a perfect mirror in its capacity to
reflect incident light, so that the reflected light has the same angle to the surface normal
as the incident light (Lee et al. 2003, VTT Automation et al. 2005). The physical
definition of gloss is the intensity ratio of specularly reflected light to incident light
23
(Leskel 1998). Coating color and surface smoothness are considered the main
contributors to overall gloss. A higher coating weight usually gives higher gloss. The
gloss of coated paper is mainly affected by particle size, size distribution, and shape (Lee
et al. 2003). The effect of surface roughness on gloss is not so straightforward; there can
be two surfaces with different texture that have the same average roughness, but different
gloss. The optical properties of surfaces and the orientation of pigment particles can
cause this effect. Not only overall gloss, but also transversal gloss distribution affects the
appearance of paper. The distribution of coating color and local surface roughness have a
great effect on local gloss, i.e. the distribution of gloss over the measured area (Bland et
al. 2000, Borch 2002).
2.4 Structural properties of the paper surface
Surface properties like roughness are among the most important factors affecting the
quality of paper. Roughness refers to the uneven surface of paper or board, and it is
especially important in printing papers, graphical boards, and many packaging boards
(Kajanto et al. 1998). Surface properties must be selected according to the application; in
tissues, a rough, soft surface is usually desired. In printing papers, the surface must be
rather smooth and homogenous to accept, retain, and present ink in an optimum manner
to the reader of the print (Bristow 1986).
Roughness can be divided into three categories: 1. macroroughness (>100 m), 2.
microroughness (1-100 m), and 3. optical roughness (<1 m). Macroroughness is a
result of formation, microroughness is a result of fibers/fines position and shape, and
optical roughness is a result of fiber and pigment particle properties (Kajanto et al. 1998,
VTT Automation et al. 2005). Roughnesses are presented as R parameters, which were
originally developed for two-dimensional stylus-type profiling applications, but were
later adapted to three-dimensional use like optical profilers, as well. Two of the most
common measures of roughness are 2D and 3D roughness averages:
1
1
j
a k avg
k
R Z Z
i
=
=

, (1)
1 1
1
j i
a kl avg
k l
R Z Z
ij
= =
=

. (2)
They present the arithmetic means of the absolute values of surface departures from the
mean plane area, where i and j are the number of data points in the transversal directions,
Z
avg
is the average plane, and Z
k
and Z
kl
are the surface heights at the transversal
coordinates of k and l. 2D and 3D root mean square (rms) averages are also widely used:

24
( )
2
1
1
i
q k avg
k
R Z Z
i
=
=

, (3)
( )
2
1 1
1
j i
q kl avg
k l
R Z Z
ij
= =
=

. (4)
They are obtained by squaring each height value in the data set and then taking the square
root of the mean. (Gadelmawla et al. 2002, Veeco Instruments Inc. 2002).
In addition to amplitude parameters R
a
and R
q
and the parameters described in the list
of terms, symbols, and abbreviations, there are roughness parameters which describe the
surface in the transversal direction. One of them is the autocorrelation function (ACF),
which describes the characteristics scale of surface height fluctuations in the transversal
direction. It is a quantitive measure of the similarity between a laterally shifted and
unshifted version of the profile (Gadelmawla et al. 2002, Vorburger et al. 1981,
Whitehouse 1994). A correlation length is used to describe the correlation
characteristics of the ACF. Its value defines the shortest distance in which the ACF drops
to a certain fraction (Gademawla et al. 2002). Information similar to the autocorrelation
function is provided by the power spectral density (PSD) function. In the case of
periodicity in the surface, a spike is seen in the PSD function in the corresponding spatial
frequency (Vorburger et al. 2002). The power spectral density function is actually a
Fourier transform of the autocorrelation function (Bhushan et al. 1985). The average
slope S
a
and rms slope S
q
parameters represent the slope of the profile over the
assessment length (Gademawla et al. 2002). They are important parameters from the
optical point of view because they control light reflectance from the surface and have an
effect on properties like gloss. There are a number of other parameters as well, but they
are not considered in detail in this thesis because only the main amplitude parameters
were discussed in the measurements.


3 Low Coherence Interferometry and Optical
Coherence Tomography
3.1 Introduction to Low Coherence Interferometry
Interference is the superposition of two or more waves, resulting in a new wave pattern.
When two light beams are combined, their fields add and produce interference. Low
coherence interferometry measures the field of an optical beam rather than its intensity
(Fujimoto 2002). A simplified schematic of a low coherence interferometer is shown in
Fig. 3.

Fig. 3. Simplified diagram of a Michelson interferometer. BS=beam splitter.
Lightsource
Detector
Sample
Reference
mirror
Lr = 2*n*lr
Ls = 2*n*ls
BS
26
The principle of low coherence interferometry can be analyzed in terms of the theory of
two-beam interference for partially coherent light. The light field from a low temporally
coherent but high spatially coherent light source is directed onto a beam splitter that
divides the beam into a reference beam and a sample, i.e. a measurement beam.
Assuming that the sample in Fig. 3 is a perfectly reflecting mirror and the polarization
effects of light are ignored, the scalar complex functions ( ) c / L t E
s s
and ( ) c / L t E
r r

represent the sample beam reflected from the specimen under measurement and the
reference beam from the reference mirror, respectively.
s
L and
r
L are the corresponding
optical path lengths of the arms of the interferometer and c is the speed of light. If we
assume that the photodetector collects all the light from the arms, the resultant intensity
d
I can be written as
[ ][ ]

+ + + + = ) t ( E ) t ( E ) t ( E ) t ( E ) ( I
r s r s d
, (5)
where the angular brackets denote the time average over the integration time at the
detector and c / L = is the time delay corresponding the round-trip optical path length
difference between the two arms, i.e. ( )
r s r s
l l n 2 L L L = = , and the asterisk symbol
means the complex conjugate operation. 1 n is the refractive index of air, and
s
l and
r
l are the geometric lengths of the arms. Because intensities ( ) ( ) t E t E I
s s s

= and
( ) ( ) + + =

t E t E I
r r r
, the resultant intensity becomes
( ) ( ) 2 Re
d s r s r mc
I I I I I V = + +

, (6)
where the first two components
s
I and
r
I are the detected intensities backscattered by the
sample and the reference arm, respectively, and the last term gives the amplitude of the
interference fringes, which depends on the optical time delay set by the position of the
reference mirror and which carries the information about the structure of the sample. This
equation is termed as the generalized interference law for partially coherent light. The
normalized mutual coherence function ( )
mc
V in equation (6) gives the degree to which
the temporal and spatial characteristics of
s
E and
r
E match, and it can be written as
( )
( ) ( )
r s
r s
mc
I I
t E t E
V
+
=

. (7)
If the complexity of the spectral components is ignored, the phase difference with respect
to optical path delay can be written as
=
0
f 2 , (8)
where
0
f is the center frequency of the light source. Furthermore, fields
s
E and
r
E
originate primarily from a single wave front, and spatial coherence can be neglected and
the complex mutual coherence function reduces to self-coherence and equation (6) can be
rewritten into the form (Pan et al. 1995, Schmitt 1999, Wang et al. 2004)
27
( ) + + =
0 tc r s r s d
f 2 cos V I I 2 I I ) ( I . (9)
According to the Wiener-Khintcine teorem, the temporal coherence function ( )
tc
V is
actually the Fourier transform of the power spectral density ( ) f S of the light source,
which is fully characterized by its shape, its spectral bandwidth, and its center
wavelength (Akcay et al. 2002, Gilgen et al. 1989, Schmitt 1999, Wang 1999), as,
( ) ( ) ( )df 2 j exp f S V
0
tc

= . (10)
This relationship reveals that the shape and width of the emission spectrum of the light
source are important variables, because they influence the sensitivity of the low
coherence interferometer to the optical path difference of the sample and the reference
arm. The equation (9) can be rewritten in a form that gives the intensity
d
I as a function
of L
( ) L k cos L V I I 2 I I ) L ( I
0 tc r s r s d
+ + = , (11)
where
0 0
/ 2 k = is the average wave number and the relation
0 0
f / c = can be used
to transform from the time domain to the path domain (Pan et al. 1995, Schmitt 1999).
In the derivation of equation (11), the sample was treated as a perfectly reflecting
mirror that induces a time delay, but leaves the amplitude and coherence of the sample
beam unchanged. In reality, this is an unrealistic situation. Actually, the light reflected
from the sample can be categorized into least backscattered light, i.e. the light that
undergoes only single scattering or very little scattering, and diffuse backscattering when
the light has scattered numerous times. In principle, least backscattered light maintains its
coherence and multiple scattered light loses it. If we take into account the scattering path
inside the media, the total round-trip path length of the sample arm is
'
s 0 s s
L L L + = , (12)
where
0 s
L is the round-trip path length to the sample surface and
'
s
L is the total round-trip
path length inside the sample, i.e.

=
i s
'
s
l n L , (13)
where
s
n is the refractive index of the medium and
i
l the scattering path of light inside
the medium. Then the intensity given in equation (5) can be written in the form
28
( )
*
r
L
s s
'
s r
L
s s
'
s d
) t ( E dL ) L , t ( E ) t ( E dL ) L , t ( E I
0 s 0 s

+ +

+ + =


, (14)
where ) L , t ( E
s
'
s
is the path length-resolved field density. Equation (14) yields
( ) ( )

+ + = LdL k cos L V L R I I 2 I I ) L , L ( I
0 tc s r s r s r s d
, (15)
where ( )
s
L R is the normalized path length-resolved diffuse reflectance, i.e. the
normalized derivative of the intensity depth distribution of the measuring wave,
representing the fraction of power reflected from the layer located at position
s
L within
the object. The signal-carrying interference term, i.e. the interference modulation in
equation (15), can be expressed as a convolution
( ) ( ) ( ) [ ]
r s s r s r s signal
L L C L R I I 2 L , L I = , (16)
where ( )
r s
L L C is the coherence function, i.e. the interferometer response in the ideal
case with a mirror in both arms multiplied by the cosine term. It is also called a point
spread function PSF of the light source. denotes the convolution operation. As a
conclusion, low coherence interferometry traces out variation in the path length-resolved
reflectances defined by ( )
s
L R (Kulkarni et al. 1997, Pan et al. 1995, Pan et al. 1996,
Schmitt 1999, Yung el al. 1999). In the case of a turbid media like paper, the above
description of the path length-resolved diffuse reflectance includes all the photons falling
in the time window defined by the coherence length of the light source. That means the
useful signal is masked by the diffuse component creating signal-degrading speckles
which are formed by the out-of-focus light that scatters multiple times and returns within
the time delay set by the difference between the optical paths in the two arms of the
interferometer (Schmitt et al. 1999). The effect of the diffuse component and how to
reduce it in paper measurements are discussed later in section 5.4.
The signal-carrying component is typically separated from dc by modulating the
optical time delay between the arms by translating the reference mirror by a constant
speed, which shifts the interference signal to the corresponding Doppler frequency
defined by
0
d
v 2
f

= , (17)
where v is the speed of the moving mirror and
0
is the center wavelength of the light
source. Shifting to the Doppler frequency facilitates the removal of the dc background
and low-frequency noise during demodulation (Schmitt 1999). To extract this signal-
carrying component, the detection circuit has three main components: (1) a
29
transimpedance amplifier, (2) a bandpass filter centered at f
d
, and (3) an amplitude
demodulator to extract the envelope of the interferometric signal (Hee 2002).
As stated earlier, ( )
r s
L L C is the function which determines the interferometer
response to path length-resolved reflectance variations. This means the signal is evident
at the detector only when the reference arm distance matches the optical length of the
reflective path through the sample within the coherence length of the light source, which
is typically around 10 m. The coherence length is the parameter which determines the
resolving ability of backscattering or reflecting sites inside the sample (Hee 1995). The
resolving ability is discussed in more detail in section 3.3.
3.2 From Low Coherence Interferometry to Optical
Coherence Tomography
Being a one-dimensional ranging technique, however, Low Coherence Interferometry has
its limitations. This is where Optical Coherence Tomography comes in, by offering the
capacity to reconstruct cross-sectional images of an object from its projections. The term
tomography is used whenever two-dimensional data is derived from a three-dimensional
object to construct a slice image of the object's internal structure. In OCT, multiple
parallel LCI scans are performed to generate the two-dimensional image. The method is
similar to ultrasound B-mode imaging, except that infrared light waves rather than
acoustic waves are used, and the achieved axial resolution is up to 100 times higher,
ranging from a few tens of micrometers to less than one micrometer. In contrast to time-
domain techniques, optical coherence tomography can be performed with continuous-
wave light without the need for ultra-short pulse laser sources, which are used mainly in
the laboratory environment (Ballif et al. 1997, Ettl et al. 1999, Fercher 1996, Schmitt
1999, Tearney et al. 1997, Yoshimura et al. 1995).
A typical measurement system consists of a Michelson interferometer illuminated by a
low temporally coherent light source, such as a superluminescent diode. In OCT, the
object to be measured is placed in one arm of the interferometer. A measurement beam
emitted by the light source is reflected or scattered from the object with different delay
times, depending on the various optical properties of the different layers within the
object. A longitudinal profile of reflectivity versus depth is obtained by translating the
reference mirror of the interferometer and synchronously recording the magnitude of the
intensity of the resulting interference fringes. A fringe signal is evident at the detector
only when the optical path difference in the interferometer is less than the coherence
length of the light source. Locating the maximum fringe visibility position allows one to
select internal cross-sections of the object with a resolution of 1-2 m (Hee et al. 1995,
Masters 1999, Vabre et al. 2000).
30
3.3 Spatial resolution of an Optical Coherence Tomography
imaging system
The spatial resolution of an OCT imaging system can be divided into two parts: the axial
resolution and the transverse resolution, which are independent from each other. The
axial resolution depends on the characteristics of the light source and the transverse
resolution is determined by the optics of the imaging device. In most cases the
autocorrelation function of the OCT device has a Gaussian shape. It is caused by the
emission spectrum of the light source, which in most cases has a Gaussian shape or
resembles it. In the case of a Gaussian-shaped autocorrelation function, the coherence
length
c
L is

=
2
0
c
2 ln 2
L , (18)
where is the FWHM (full width at half maximum) wavelength range of the light
source (Brezinski 1999).
c
L is a common value for the axial resolution, and for the
coherence length, the equation above is the most used. The definition of
c
L is sometimes
a bit confusing. The coherence length in equation (18) is the round-trip coherence length,
but it is also adopted as a definition for the longitudinal coherence length, which is
actually twice as large as the value given by equation (18) (Akcay et al. 2002,
Baumgartner et al. 1997, Baumgartner et al. 1998, Fercher et al. 1999, Hitzenberger et al.
2002, Laubscher 2004, Tomlins 2005, Zhang et al. 2001). In papers II and IV, a different
value for coherence length was adopted, but then it was mentioned that the resolution of
the system is then described by dividing the coherence length by 2. The equation (18) is
valid only in a vacuum. The usual case is that the sample is dispersive, which means the
refractive index of the sample material depends on the wavelength. The effect is
significant when broadband sources are used (Sampson 2004). The distances in low
coherence interferometry are optical distances. This means the geometric distance is
derived by dividing the optical distance by the group refractive index
g
n of the media

=
d
dn
n n
g
. (19)
In real materials, not only is the refractive index n a function of wavelength, but also the
group index
g
n
.
The result is a broadening of the interferograms and an increase in
coherence length, leading to a corresponding reduction in resolution. In a dispersive
medium, the coherence length
m , c
L can be calculated by
2
g
g 2
c m , c
d
d
dn
L L

+ = , (20)
31
where
g
d is the geometric depth in the medium (Drexler et al. 1998, Hitzenberger et al.
1998, Hitzenberger et al. 1998, Hitzenberger et al. 1999).
As stated earlier, the transversal resolution is determined by the optics of the imaging
device and doesnt correlate with the axial resolution. The selection of optics is a trade-
off between the transversal resolution and the imaging depth range, i.e. the depth of
focus. The transversal resolution x can be written as

=
d
f 4
x
l
, (21)
where
l
f is the focal length of the focusing lens and d is the light beam diameter on the
lens aperture. It can be seen from equation (21) that a large numerical aperture NA gives
a small spot size and a high resolution, but another aspect which has to be taken into
account is that the imaging depth range, i.e. two times the Rayleigh range, becomes
shorter. The Rayleigh range
r
z is

=
4
x
z
2
r
. (22)
The Rayleigh range gives the distance from the focal plane to the point where the light
beam diameter has increased by a factor of 2 (Clements 2004, Drexler 2004, Fujimoto
2002). In most biological applications, the imaging depth should be in the millimeter
scale, but in the case of a paper sheet, there is no need for such high imaging depths, and
a high NA can be used.


4 Measurement system
A schematic of the free-space OCT setup used in the experiments is shown in Fig. 4. This
experimental model is highly modifiable and can be used for several low coherent
purposes: LCI, OCT, and DOCT. The main drawback of the device is that it needs a
moderately skilled operator to change the procedure of the imaging sequence. The free-
space model is introduced in detail in the next subsections.

Fig. 4. Schematic of the OCT system: (1) SLD; (2) photodiode; (3) preamplifier; (4) bandpass
filter; (5) demodulator; (6) collimator; (7) mirror; (8) axial scanner; (9) sample; (10)
transversal scanner; (11) data acquisition; (12) display; (13) computer; (14) PZT driver; (15)
ramp generator; (16) motion controller.
The system in Fig. 4 can be divided into four parts: (1) the interferometer, (2) an analog
signal processing and control unit, (3) data acquisition, and (4) digital signal and image
processing. The first of these, the interferometer, is introduced in subsection 4.1, where
the components and corresponding specifications are described in detail. The analog
signal processing and control unit is described in subsection 4.2. It gives an overview of
the blocks of the unit: amplifiers and a sweep generator for PZT. The data acquisition
33
procedure is described in subsection 4.3. Subsection 4.4 describes how the acquired
signals are converted to images and gives an overview of the color maps used in the
images.
4.1 Interferometer
The schematic of the experimental measurement setup presented in Fig. 4 uses a
superluminescent diode (
1
Superlum Ltd. Model SLD-380-MP3-TOW2-PD or
2
Superlum
Ltd. Model SLD-380-HP2-TOW2-PD) as a low coherent light source. With a peak
emission power of
1
8.25 mW /
2
50 mW, the diode illuminates a Michelson-type
interferometer with a free-space configuration shown in Fig. 5.


Fig. 5. Free-space OCT measurement setup.
Having a center wavelength of
1
822 nm /
2
832 nm and a FWHM spectral width of
1
20.2
nm /
2
19.7 nm, the achieved axial resolution in air is
1
14.8 m /
2
15.5 m. Both diode
models were used in this thesis project. Two different Current and Temperature
Controllers were used; Superlum models PILOT-2 and PILOT-3-1. They hold the diode
current stable by monitoring the output power and temperature of the diode.
A diverging beam from the SLD is collimated by using a triplet lens collimator
(Melles Griot B. V. Model 06GLC003) and split equally by a cube beam splitter designed
for near-IR (Melles Griot B. V. Model 03BSC025) into the two arms of the
interferometer: the reference arm and the sample arm. Antireflection-coated plano convex
diode laser lenses (Melles Griot 06LXP007/076) are used for focusing in the arms.
34
According to equation (21), the size of the focused beam (defined by the 1/e
2
radius of
the Gaussian beam) is then ~8.5m, which defines the transversal resolution. The first of
the arms, the reference arm, contains an axial scanner for producing interference
modulation, and it shifts the signal to a Doppler frequency that depends on the speed of
the scanner. Scanning can be performed by either a piezoelectric scanner (a Physik
instrumente P-783.ZL PZT element with an E-662.LR controller) or a servo motor (a
Newport VP-25XA servo stage with an ESP300 controller). For best accuracy and
linearity, a piezoelectric scanner is used in a closed loop mode. The object to be measured
is placed in the other arm, the measuring arm, which contains stepper motor driven xy
stages (Newport DM11-25 actuators with an ESP300 controller) for scanning the object
in transversal directions to achieve 2D and 3D images. All the stages are computer-
controlled and synchronized with each other via an RS232 or a GBIP bus. The reflected
and scattered beams are focused on the detection arm with a plano convex lens (Melles
Griot 06LXP007/076) and combined on the silicon photodetector (Melles Griot
13DSI005) to produce an interferometric signal containing information about the internal
structure of the sample. This particular detector has a responsivity of 0.45 A/W for a
wavelength of 820 nm.
4.2 Analog signal processing and control unit
After detection of the signal in the interferometer, the signal is passed to a
transimpedance amplifier, which is designed and optimized especially for OCT purposes
by using appropriate, low-noise operational amplifiers. The architecture of the single-
stage transimpedance amplifier ensures a high signal-to-noise ratio, providing a very
strong signal already at the output of the first stage without unnecessary DC bias.
Transimpedance can be adjusted from 1 to 2 M by using an 8-position switch on the
front panel of the analog signal processing unit shown in Fig. 6.

Fig. 6. Analog signal processing and control unit.
When no further analog signal processing is needed, the signal is passed to the direct
output (DIRECT OUT) shown in Fig. 6. This output is used when digital signal
processing algorithms are used in parallel with analog processing.
The signal from the transimpedance amplifier is then passed to the next block, which
is the bandpass filter forming an 8
th
order Butterworth filter with a 3 kHz pass bandwidth.
35
The task of this filter is to remove all the unnecessary noise coming from external light
sources, vibrations, and the electronics itself. The frequency and bandwidth of the useful
signal are closely related to the parameters of the light source and the velocity of the
reference mirror. The signal of this output (OCT OUT) is used whenever the fringe signal
amplitude and phase are needed.
The third ouput (DOCT OUT) is used when Doppler measurements are performed.
This output stage also has an 8
th
order Butterworth filter, but it also has a 72 kHz pass
bandwidth, which allows us to measure the velocity of flowing/moving particles.
There is usually no need to record the phase of the interferometric signal, and only the
magnitude of the signal is recorded. That is also a practical way to reduce the sampling
frequency and data storage space needed. The fourth output (ENVELOPE OUT) gives us
the envelope of the fringe signal by using a demodulator that consists of two sub-blocks:
a precision rectifier and a lowpass filter. With the rectifier sub-block, signals as low as 2
mV
p-p
can be rectified without significant distortions. The second sub-block, the lowpass
filter, is of the third order Butterworth type with the cutoff frequency set to effectively
remove the carrier frequency but leave the envelope undistorded.
The fifth output of the unit is the logarithmic envelope output (LOGARITHMIC
OUT), which amplifies the envelope signal logarithmically, which is a useful option in
detecting signals with a very wide dynamic range. This was the output used in most of the
measurements made during this thesis project.
The internal sweep generator in the E-662.LR controller was found to give bad
linearity and stability when the speed was increased. The bandpass of the OCT filter was
centered at 9.2 kHz, which corresponds to scanner speeds of 3.78 mm/s and 3.83 mm/s
with 822 nm and 832 nm light sources, respectively. An external sweep generator, i.e. a
control signal generator, was constructed to give a nice slope for the E-662.LR. The
repetition rate, slope, and amplitude of this control signal can be adjusted from the front
panel of the sweep generator unit. The control signal is then amplified in the E-662.LR,
which drives the PZT element and follows the movement by means of a feedback loop
(Krehut 2003).
4.3 Data acquisition
Data acquisition is performed by using a National Instruments (NI PCI-6070E) DAQ card
with a 12-bit AD converter connected to the PCI bus of the PC. The measured signals and
triggering pulses are connected to the DAQ card via an interface card (NI BNC-2110)
with coaxial connectors. The DAQ card is contolled by a Labview
TM
progam which
handles the signal acquisition and saving. The detailed measurement sequence of the
system is:
1. The Labview program is started, where parameters like the number of scans, sampling
frequency, and number of data points/scan are defined.
2. The measurement program is downloaded from the PC to the ESP300 controller via a
serial bus, and it defines the movement parameters of the transversal scanning stepper
motors, like speed, acceleration, motion increments, and movement procedure. It also
36
controls the triggering of the depth scanning start pulse, which is guided to the sweep
generator of the PZT control unit.
3. The trigger pulse causes the sweep generator to give a sweep signal to the PZT
controller, which amplifies the control signal to the desired level of PZT element. At
the same time the Labview program starts to acquire the signal of the selected output
of the analog signal processing and control unit. Samples are taken at pre-selected
sampling intervals, so that the total amount of samples covers the whole scanning
range of the PZT element.
4. After the scanner has reached its maximum displacement, it is returned to its original
position by the sweep generator and the sequence is started again from phase 3.
The acquired signals are saved in ASCII or binary format. The ASCII format was used in
early measurements, but the binary format was adapted later because of its higher speed
of saving. Thus there is a need to convert it to ASCII later for data processing.
4.4 Digital signal and image processing
The process of signal to image conversion is quite a straightforward process, and the
amount of processing depends on the type of image formed. Three types of images were
used in this thesis project: topography, tomography, i.e. 2D images, and 3D images. Each
of these needs a slightly different procedure. All the processing was perfomed by using a
Matlab
TM
program.
The first of these, the topography image, is the easiest one to form; there is no need to
save the whole measurement vectors in Labview
TM
, but only the maximum amplitude
position originating from the surface of the sample, i.e. from the first air-sample interface.
It means that a matrix of these values is measured and then combined in Matlab
TM
to
form a 3D matrix that gives us surface z-positions as a function of transversal positions.
The set of position values is not very informative when we discuss about topography,
thus some processing of the data set is needed. Matlab
TM
has built-in functions for
combining a set of data points to form a surface. Every data point on this surface is
marked with a different color depending on its depth value. The most common color
maps are grayscale and false color maps. The grayscale map contains a defined number
of gray levels from white to black. However, the human eye is not very sensitive to
separate a number of gray levels, which sometimes makes the topography details
invisible. The solution is to use a false color map that employs multiple colors scaled to
the depth data. There is no standard false color map for topography, and many variations
can be found. A false color map called jet was used in some of the experiments done in
this thesis project. In the jet map, color varies from dark blue to bright red, and
sometimes this map gives a better view of the structure of the sample. The precision of
the system for surface position detection in toporeconstruction is 0.2 m when the
maximum amplitude position is recorded. The limiting factor is the quantization error that
causes drift in the position calculation. The recorded maximum amplitude position was
used for toporeconstruction throughout the measurements done in this thesis project. If
the center of mass of the envelope is used for position determination, the precision is 40
nm.
37
The second option, the tomography image, gives a 2D slice image of the objects
internal microstructure. The whole measurement vectors are now saved and because there
is no need for further processing of separate signals in Matlab
TM
anymore, the
logarithmic envelope output of the analog signal processing and control unit is usually
used., Matlab
TM
also has its own function for slice images, called imagesc. This function
transforms the z-data to pixels. Each pixel has its own color selected from the color map
used. The color depends on the pixels amplitude, i.e. the amplitude of the signal. A
number of these vectors are placed in parallel so that they form a matrix of image pixels.
Color map scaling and thresholding are the only actions done as postprocessing of the
images. The resolution of the 2D slice images is determined by the focusing optics and
the coherence length of the light source.
The last option, a 3D image, gives the complete structure of the sample in three
dimensions. This image type is the most complex to form and needs a powerful computer
to handle all the data. Also in this case the whole measurement vectors are saved from
the logarithmic envelope output of the analog signal processing and control unit. The data
set has four values for each pixel; a position in three dimensions x, y, and z, and a signal
amplitude value at that particular position, and it can be stated as a volume data set. A
Matlab
TM
function called isosurface connects the adjacent points in the volume data
matrix if their amplitude value crosses the specified value defined in a function call,
which in other words means the selection of a ROI (Region of Interest). This can be
easily understood by taking a 3D image of a round tube as an example. If we take OCT
images of this tube, we have a number of parallel 2D images which have the round
shaped cross-section of the tube. Selection of a ROI is performed by thresholding the
image data so that the cross-section is separated from the surrounding medium. The edges
of these cross-sections are connected by a surface, and we get a 3D presentation of the
object. An isocaps function in Matlab computes the isosurface end cap geometry for the
volume data, which in the case of a tube corresponds to the images of both ends of the
tube. For 3D images, color map scaling and 3D smoothing are usually performed to
enhance image quality. The resolution of the 3D images is determined by the focusing
optics and the coherence length of the light source.

5 Low Coherence Interferometry and Optical Coherence
Tomography in paper measurements
This chapter presents the results of using Low Coherence Interferometry and Optical
Coherence Tomography for paper measurements. In section 5.1, LCI is used to evaluate
the surface structure of paper. Some of the roughness values are compared with the
values measured with the commercial profilometer. These measurements are discussed in
papers II, V and VI. In section 5.2, the surface structure, roughness, average gloss, and
local gloss of paper surfaces are discussed in parallel. A detailed analysis can be found in
paper III. Section 5.3 gives an example of determining the filler content of paper by
using OCT, introduced originally in paper VI. Sections 5.4 and 5.5 discuss 2D and 3D
imaging of paper, respectively. The results are collected from papers I, IV, V and VI.
Except for the measurement of the line profiles in Fig. 7 and 8 and the a-scans in Fig. 18,
the Superlum Ltd. model SLD-380-HP2-TOW2-PD was used in all the experiments as
the light source of the interferometer.
5.1 Surface analysis of paper
Several methods and standards are used to analyze the surface structure of paper in
modern paper research and in-line quality control processes. Nevertheless, some of these
methods are non-scientific; they all yield slightly different values and typically require a
number of separate devices to cover all types of paper.
Surface analysis, including roughness estimation, is particularly important in printing
papers, graphical boards, and packaging boards, because parameters like roughness affect
such optical properties of paper as gloss and ink absorption. Knowing the structure of the
base paper (paper before further processing, like coating) surface is also an important
characteristic when estimating the amount of coating color required to make the paper
printable or otherwise more suitable for a particular application (Kajanto et al. 1998).
Low Coherence Interferometry enables measurement of various kinds of paper
surfaces. In the papermaking industry, air-flow methods for surface roughness evaluation
still play an important role and are widely used. Two air-leak methods are used to
39
measure the roughness of paper samples: 1. the Bendtsen air-leak method, in which a
hard ring is pressed on top of the paper surface, and the ensuing air leak escaping under
the ring is measured in unit time, and 2. the PPS method (Parker Print Surf method),
which uses a soft ring instead of a hard one and responds remarkably well to small
surface imperfections.
However, air-leak methods are prone to failures, and paper sheets often have holes
penetrating through them which, naturally, affect the measurement results. This does not
happen with LCI. Moreover, LCI measurements do not require multiple devices for
simultaneous imaging of the surface structure and measurement of roughness parameters.
As a result, both time and money are saved.
All the experiments presented in this section were measured using the setup described
in section 4.1. The surface position was determined by recording the maximum amplitude
position originating from the surface of the paper sample.
In the first experiment, two fine paper surfaces were measured to define the surface
roughness parameters R
a
and R
q
. The first of these, code HW 75/25 F30 fine paper, is
manufactured in a test paper machine by Metso Paper. It contains 75% birch, 25% pine
and 30% PCC (CaCO
3
precipitated calcium carbonate) filler, and the second one is
typical commercial copy/printing paper whose exact constituents were unknown. The
step increment between the transversal pixels was 1 m. The surface line profiles
measured from the samples are presented in Fig. 7 and 8 (Paper II).
Fig. 7. Surface line profile of fine HW 75/25 F30 fine paper.
Fig. 8. Surface line profile of commercial copy/printing paper.
-80
-60
-40
-20
0
20
40
0 100 200 300 400 500
Lateral position [m]
D
e
p
t
h

p
o
s
i
t
i
o
n

[

m
]
-80
-60
-40
-20
0
20
40
0 100 200 300 400 500
Lateral position [m]
D
e
p
t
h

p
o
s
i
t
i
o
n

[

m
]
40
The surface roughness parameters R
a
and R
q
for the line profiles in Fig. 7 and Fig. 8 can
be calculated using equations (1) and (2). They give R
a
=4.6 m and R
q
=5.9 m for HW
75/25 F30 paper and R
a
=5.5 m and R
q
=6.9 m for copy paper. When we compare the
figures above, there is a clear difference in surface line profiles. The difference between
the roughness values is yet only a scant micrometer, but the difference is clearly evident
when the paper surface is physically touched. This is because, in the profiles measured
above, we deal with local roughnesses, i.e. short transversal distances compared with
structural changes in paper in the transverse directions. In the following surface structure
measurements, a larger area was measured to give a more representative view of the
surface and a value for roughness.
As stated earlier, LCI can be applied to various grades of paper to measure the surface
structural properties. In the next experiment, three widely disparate grades of paper were
measured to demostrate the capabilities of the LCI technique in visualizing the surface
structure of various types of paper and in calculating the corresponding roughness values
according to equations (3) and (4). Fig. 9-11 show the results of this experiment. The
sizes of the images are 500 x 500 pixels with 5 m transversal step increments, which
yields 2.5 x 2.5 mm images. The smaller images in the top right corner show zoomed 0.5
x 0.5 mm images from the top left corner of the samples. Fig. 9 shows the topography of
a laboratory-made paper sheet. As seen, the sample containing long pine fibers with 15%
PCC filler is quite rough; R
a
=6.9 m. Furthermore, it is an uncoated, uncalendered
sample with low gloss and high porosity. The visibility of individual fibers is high,
because no machining has been performed on the surface and no coating materials have
been used to smooth it.
Fig. 9. Topography of a laboratory-made paper sheet. The size of the larger image is 2.5 x 2.5
mm and the smaller one is 0.5 x 0.5 mm.
41
The second sample, HW 75/25 F30 fine paper, is actually the same used in the previous
experiment presented in Fig. 7. The topography of this sample is seen in Fig. 10. In
comparison with the laboratory-made paper, this sample is also uncoated and
uncalendered. Other characteristics include moderate gloss and porosity, and a roughness
of Ra=4.6 m. Manufactured in a test paper machine by Metso Paper, the visibility of the
individual fibers is quite poor, due to manufacturing processes such as milling and the
addition of a fairly large amount of filler in the pulp.
Fig. 10. Topography of an uncoated fine paper sheet. The size of the larger image is 2.5 x 2.5
mm and the smaller one is 0.5 x 0.5 mm.
The third sample, presented in Fig. 11, is also an uncalendered fine paper sheet, but now
with high gloss and low porosity. It is highly coated to smooth the surface for photo
printing and has a roughness of Ra=0.8 m. As in the case of HW 75/25 F30, the fiber
visibility is more or less zero because of the coating on the surface. The amount and type
of filler in this sample were unknown.
The preceding measurements show that LCI is capable of measuring the surface
topographies and roughness values of several grades of paper. Papers from rough base
paper to smooth photo printing papers can be measured using the same setup. One
important feature is that the structure of the surface and the roughness values can be
obtained simultaneously. Compared with air-flow methods, there are less sources of error
present in the measurement event. One of these is an air leak through the paper, which
becomes significant when porous paper is measured and causes misleading results for the
roughness value.
42
Fig. 11. Topography of a coated fine paper sheet. The size of the large image is 2.5 x 2.5 mm
and smaller one is 0.5 x 0.5 mm.
In the next experiment, five laboratory-made sheets of paper were measured using LCI.
The roughness parameters were calculated and the results were compared with a
reference provided by the Technical Research Centre of Finland (VTT). The reference
contains surface topografies and roughness parameters measured with a commercial
Altisurf 500 optical profilometer. The size of the measurement beam on the surface of the
sample is 1 m and it gives 10 nm depth resolution. The aim of this experiment was to
verify the results achieved by the device described in section 4.1. Each sample had
undergone a slightly different calendering method and filler properties described in Table
3 (VTT Processes 2003).
Table 3. Samples.
Sample Filler content Calendering
A 0% uncalendered
B 0% medium calendering
C 0% strong calendering
D 15% uncalendered
E 30% uncalendered
Fig. 12 and 13 give two examples of topografies of these samples measured using LCI.
The sizes of the images are 500 x 500 pixels with 5 m transversal step increments,
which yields 2.5 x 2.5 mm images. They are selected to give an overview of the filler
distribution in the paper and the effect of calendering on the paper surface. As these
figures demonstrate, calendering gives a smoother surface, thereby reducing the
topographic visibility of its fiber distribution.
43
Fig. 12. Topography of sample D.
Fig. 13. Topography of sample C.
In addition to parameters R
a
and R
q
, several other roughness parameters were analyzed in
this experiment. They are presented and explained in detail in the list of terms, symbols,
and abbreviations (Gademawla et al. 2002, Veeco Instruments Inc. 2002).
The results are shown in Table 4. If we look at the table and compare the values given
by LCI and the Altisurf 500, we find that all the parameters give a different value. Some
x-position [mm]
y
-
p
o
s
i
t
i
o
n

[
m
m
]
0 0.5 1.0 1.5 2.0 2.5
0
0.5
1.0
1.5
2.0
2.5
x-position [mm]
y
-
p
o
s
i
t
i
o
n

[
m
m
]
0 0.5 1.0 1.5 2.0 2.5
0
0.5
1.0
1.5
2.0
2.5
44
of them differ slightly and some of them are completely different. One reason for the
differences can be the difference in the measured area. In the case of LCI, it was 2.5 x 2.5
mm, and in the case of the Altisurf 500, it was 4 x 4 mm. Such an area was too large for
our system.
Table 4. Roughness values for LCI and an Altisurf 500. The values are given in m.
Sample
A B C D E
LCI
Ra 7.83 5.76 5.78 6.85 14.90
Rq 9.92 7.44 7.41 9.32 22.00
Rp 27.20 57.20 23.90 92.40 123.00
Rv 58.30 74.10 75.70 64.30 99.80
Rt 85.50 131.00 99.70 157.00 222.00
Rsk -0.79 -1.04 -0.96 -0.26 1.12
Rku 3.99 5.47 4.69 8.65 6.78
Rz 80.80 88.80 78.90 144.00 201.00
Altisurf
500

Ra 5.48 4.30 4.40 7.27 15.44
Rq 7.36 5.86 6.46 10.72 24.44
Rp 27.65 45.16 180.60 161.00 170.70
Rv 64.37 51.76 44.42 73.27 90.40
Rt 92.02 96.92 225.00 234.20 261.10
Rsk -1.69 -1.56 0.94 0.57 2.26
Rku 7.27 7.31 38.02 13.65 11.72
Rz 84.91 86.91 186.60 212.30 248.50
Table 5 gives us the linear correlation coefficients for all the roughness parameters. It can
be seen that the parameters R
a
and R
q
have the highest correlation and all the others
correlate more or less badly. The differences in correlation can be explained by the
differences in the focusing geometry of the measurement devices.
Table 5. Correlation coefficients for roughness values measured by LCI and an Altisurf
500.
Roughness parameter
Ra Rq Rp Rv Rt Rsk Rku Rz
Correlation
Coefficient
0.97 0.98 0.45 0.48 0.65 0.76 -0.17 0.79
Thus, due to the Altisurfs smaller focus spot, which is superior in the detection of large
narrow peaks, the Altisurf tends to produce a larger Rp value than LCI. On the other
hand, the ability of LCI to detect deep pits is better than that of Altisurf. This can be
explained by the assumed large NA of the Altisurfs focusing optics that causes that some
of the deep pits to be invisible to such a system, and in most cases it gives a smaller value
45
for R
v
. It is very likely that this focusing geometry lies at the root of the other observed
non-correlations, as well. As an example, take the parameter Rsk; if this value is positive,
there are more peaks than pits on the paper surface, if the value is negative, there are
more pits. Consequently, because the Altisurf is better at detecting high peaks than deep
pits, the Rsk value is in most cases positive compared with the value measured using
LCI. The same reason causes a larger value for R
ku
with the Altisurf, which detects high
narrow peaks on the surface.
But, if we compare the results sample by sample by taking into account all the
roughness parameters, we get interesting results. In table 6, the linear correlation
coefficients are calculated for each sample.
Table 6. Correlation coefficient of all the roughness parameters measured using LCI and
an Altisurf 500.
Sample
A B C D E
Correlation
Coefficient
1.00 0.98 0.73 0.99 0.99
The above table reveals that although the correlation of single roughness parameters
between the samples is poor, excluding R
a
and R
q
, the relationship of the parameters from
the same sample measured using LCI and the Altisurf 500 shows nice correlation. The
value for sample C stays unknown.
5.2 Surface quality evaluation of paper using LCI and a DOG
Micro surface roughness and related gloss have been observed to be important parameters
in paper science (MacGregor 1994). In this section, two methods, low coherence
interferometry and diffractive optical element based glossmetry (DOG), were utilized to
give joint information on both surface roughness and gloss that helps papermakers in
their research and development of optimal paper surface quality, which is crucial for
optimal ink absorption in the printing process.
Micro surface roughness of paper has an important role in the gloss of paper.
Unfortunately, commercial glossmeters do not provide information on the local gloss of
paper. In the next experiment, LCI was employed to assess the average surface roughness
of three different paper samples with the aid of the recorded topography maps.
Furthermore, local and average gloss were measured with a diffractive optical element
based glossmeter (DOG), which has previously been applied to local gloss detection of
plastic products (Myller 2003, Silvennoinen 2004). More information on the DOG setup
used in this experiment can be found in Paper III.
The samples of this study present rather different grades of paper. The first one,
conventional commercial copy paper, was investigated. It is calendered and contains, in
addition to pulp, usually about 10-20% filler material, which is typically PCC. The
second sample, supercalendered (SC) paper, was not coated and depending on the
calendering process and filler contents, it has different applications, such as paper for
46
offset printing. Typically the filler used in SC paper is caoline. The third sample,
supercalendered fine paper, was coated and different coating pigments such as PCC are
typically used. This paper product usually provides a glossy surface and good printing
quality. The supercalendering process tends to smooth the surface and assists in the
formation of mirror-like facets. As a rule of thumb, the smoother the paper surface, the
better the printing quality, which can be distinguished, e.g. in the reproduction of images.
The samples were marked with needle holes so that the same area was measured using
LCI and a DOG. The 3 x 3 mm areas were measured in 15 m transversal step increments
with both devices. In addition to LCI and a DOG, gloss was also measured with a
commercial Zehntner ZGM 1020 Glossmeter with two angles of incidence: 20 and 60.
All the gloss measurements were performed at the University of Joensuu.
Fig. 14 gives an example of one of the topographies measured. It shows the
topography of a Xerox copy paper sample. The corresponding gloss map can be found in
Paper III, Fig. 6. A direct comparison of topography and gloss maps was not conducted in
this study. Surprisingly, the roughest paper gives the lowest variation in gloss. This effect
is clearly seen if we compare the topographies and gloss maps of fine and Xerox copy
paper samples. This effect is invisible if only the average values of roughness and gloss
are used, and it leads to a conclusion that local variation in surface properties is an
important factor when surface quality is evaluated. All the topographies and gloss maps
can be found in Paper III. The gloss and roughness values, with corresponding standard
deviations calculated for these samples, are shown in Table 7. Holes were excluded in the
calculation of roughness values. Fig. 15 shows the normalized gloss measured with a
DOG and a Zehntner 1020 with two geometries as a function of roughness R
a
.
Fig. 14. Topography of copy paper
0
1
2
3
0
1
2
3
x-position [mm]
y-position [mm]
z
-
p
o
s
i
t
i
o
n

[

m
]
47
Table 7. Averages and standard deviations of gloss measure with a DOG (G) and a
Zehnter 1020 (GU Gloss unit), and roughness measured using LCI (m).
Paper
sample
G
G
GU20
GU20
GU60
GU60
R
a

Ra

Fine 1.03 0.04 4.80 0.72 32.80 2.27 1.78 1.03
SC 1.08 0.03 2.56 0.15 13.10 0.86 5.40 1.95
Xerox 0.87 0.01 1.50 0.01 3.30 0.11 7.38 3.15

Fig. 15. Normalized gloss G, GU 20, and GU 60 as a function of roughness R
a
.
The Zehntner values GU20 and GU60 give nice negative linear correlation to
roughness; the correlation coefficients are -0.9994 and -0.9997, respectively. But for the
DOG value G, the correlation coefficient is only -0.61. It can be seen in Table 7 that SC
and Fine papers have large variations in local gloss measured with the DOG. It is evident
that the surface roughness of paper is not the only dominant factor affecting gloss, and it
cant be used alone to predict gloss variation on the surface. Variations in the refractive
index due to the fillers and surface coatings can cause large differences in local gloss. As
can be seen in Fig. 15, G seems to be less sensitive to roughness changes than GU20 and
GU60. Thats probably caused by the differences in measurement geometries of the
Zehntner and the DOG. The Zehntner gives more global information about gloss because
it measures an area of ~1.5cm
2
compared with the DOGs 0.09cm
2
. Islands of filler or
uneven coating can induce strong variations in the refractive index, which has a strong
effect on local gloss without any significant effect on roughness.
0
0.2
0.4
0.6
0.8
1
1.2
1 2 3 4 5 6 7 8
Ra [m]
N
o
r
m
a
l
i
z
e
d

g
l
o
s
s

[
a
.

u
.
]
G
GU 20
GU 60
48
5.3 Filler content evaluation of paper
Fillers are added to paper mainly to improve printability, but their inexpensiveness also
endows other benefits. Thus, when machine-made paper started to be sold by weight
rather than the number of sheets, as previously, cheap fillers came in handy. Other
reasons for using fillers in printing paper include improving their opacity, fairness, ability
to absorb printing ink, as well as surface smoothness and pleasantness, i.e., how the paper
feels in ones hands. There are also drawbacks; fillers lower the strength and gluing
properties of paper, for example (Aaltio 1969).
Measuring the filler content of paper relies on the fact that the scattering properties of
pulp and paper change with the addition of filler. The filler actually fills in spaces
between fibers, increasing the refractive index of the fibers surroundings and decreasing
the mismatch between the refractive indices of the different scattering components. The
refractive index of fiber is assumed to be 1.55, which is the index for cellulose, and the
index of air is 1. Secondly, increasing the filler content makes the filler particles
aggregate, which makes the fillers less effective in scattering light (Leskel 1998).
Consequently, if the scattering coefficient of the investigated sample changes due to
changes in the filler content, the slope, i.e. the reflectivity versus depth slope of the OCT
signal will show a corresponding change. In a highly scattering media like paper, the
scattering coefficient

>> the absorption coefficient, and the scattering properties can be
evaluated using Lambert-Beers law, which gives the exponential decay of light as a
function of depth. When using this formala as a basis for measuring the filler content of
paper, one must choose the linear part of the signal where the detected light is assumed to
be single-scattered. The maximum depth of determining the filler content with this
approach depends on the optical properties of the paper. This method doesnt give us an
absolute value for the filler content, and only relative changes can be measured.
To prove this assumption, three series of laboratory-made paper samples were
measured. These samples were made of unground pine pulp with a filler content of 0%,
15% and 30%, respectively. Precipitated calcium carbonate (PCC) was used as the filler
material, while Percol was employed as a retention chemical. In this context, filler
content is defined by the contribution it makes to the dry weight of the sheet, not to its
total weight (Saarela 2003).
Two measurement sequences were performed for each sample. First, 500 a-scans were
recorded in 2 m transversal step increments in the x-direction. A second sequence was
performed by moving the sample 0.5 mm in the y-direction and again recording 500 a-
scans in 2 m transversal step increments in the x-direction. The fringe signal envelopes
were averaged for each measurement sequence, and we got two average slopes for every
sample. Then, depicted in Fig. 16, the least squares (LSQ) line was fitted into the linear
part of the envelope. Table 8 shows the slopes of these LSQ lines with averages and
corresponding standard deviations calculated for each sample. Fig. 17 shows the
normalized average values of the slopes as a function of filler content with standard
deviation bars. The dependence of the slope is clearly indicated in Fig. 17; the slope of
the LSQ line decreases with increasing filler content. A filler content increase from 0% to
30% produced a 22% decrease in the LSQ slope. Although the sensitivity of the
measurement remains unknown until slopes with smaller differences in filler content can
49
be determined, the experiments do corroborate one fact: the effect of the filler is fairly
obvious, indicating that the refractive index mismatches have indeed decreased, resulting
in smaller LSQ slope values.
Fig. 16. Example of slope determination using the LSQ line fitted into the averaged output of
the logarithmic envelope output of the amplifier.
Table 8. Slopes of the LSQ lines.
0% filler
Sample no. (a.u.)
Slope (a.u.) 15% filler
Sample no. (a.u.)
Slope (a.u.) 30% filler
Sample no. (a.u.)
Slope (a.u.)
-8.09 -7.08 -5.74 0/8
-8.05
15/3
-7.03
30/7
-5.74
-8.24 -6.45 -5.73 0/10
-8.17
15/5
-6.47
30/8
-5.77
-8.20 -6.16 -6.47 0/11
-8.21
15/6
-6.21
30/9
-6.57
-7.23 -6.43 -6.62 0/12
-7.27
15/8
-6.43
30/10
-6.74
Average -7.93 -6.53 -6.17
Standard deviation 0.43 0.34 0.46
Normalized average 1.00 0.82 0.78
Normalized std 0.05 0.04 0.06
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
0 50 100 150
Optical depth [m]
A
m
p
l
i
t
u
d
e

[
V
]
50
Fig. 17. Normalized average value of slopes of the LSQ lines with standard deviation as a
function of filler content.
5.4 2D and 3D structural imaging of paper
As stated earlier, paper consists of a stochastic network of fibers, but since these fibers
are much longer than the thickness of the paper sheet, the network can be treated as
planar and almost two-dimensional. This two-dimensional structure can be used to
determine a number of parameters, and indirectly it even reveals some of the papers
three-dimensional characteristics. However, to improve paper quality, it is important to
know its three-dimensional porous structure (Niskanen 1998). Hitherto, research on the
structure of paper has suffered from an absence of nondestructive, fast, and cost-effective
measurement techniques for micro- and macrostructure imaging. Traditional methods are
based on ultrasonic, X-ray, magnetic resonance imaging (MRI), scanning electron
microscopy (SEM), and conventional light microscopy techniques (Hggvist 1999, Rutar
et al. 2001, Samuelsen et al. 2001). A problem in many of these techniques arises from
the measurement event itself; either the measurement is very slow and labor-intensive, or
it is expensive to carry out. Typically, 2D imaging with, e.g. SEM or light microscopy
requires a substantial amount of preparation before the analysis. Such tasks include
staining the samples with methyl blue and embedding them in epoxy. Moreover, the
sample can be viewed only after drying (Niskanen et al. 2002).
In recent years, the availability of reasonably cost-effective lasers and other light
sources have launched the development of optical ranging and imaging methods like low
coherence interferometry (LCI), optical coherence tomography (OCT), optical coherence
microscopy (OCM), and confocal laser scanning microscopy (CLSM). Optical techniques
have been used for structural imaging in medicine for years, but industrial applications
are few, because the complex propagation of light inside a scattering medium, such as
Filler 0%
Filler 15%
Filler 30%
0.5
0.6
0.7
0.8
0.9
1.0
1.1
0 10 20 30
Filler content [%]
N
o
r
m
.

s
l
o
p
e

[
a
.
u
.
]
51
paper, results in coherent multiple-scattering processes that degrade resolution and image
contrast (Bestemyanov et al. 2004, Zakharov et al. 2002, Zimnyakov et al. 2002).
This section introduces the use of OCT for 2D and 3D structural characterization and
imaging of paper. The first experimental measurement results of using OCT for structural
imaging of paper are introduced in Paper I. A 120 m thick paper sample was measured
to give a view of the OCT a-scan of paper. Before these results, only Fercher et al.
introduced few OCT a-scans of a sheet of paper in 2000. The reference value for
thickness was measured using a Lorenzen & Wettre Micrometer 51. Fig. 18 shows four a-
scans from a 120 m thick paper sample. The transversal step increment between the a-
scans was 20 m.
Fig. 18. A-scans of paper.
The scaling from optical to geometrical depth was performed by taking the first a-scan as
a reference. This scaling coefficient was then used for the other scans. The arrows in Fig.
18 show the front and back surfaces of the paper sample. It was assumed that the
structure doesnt change much in this range, but as seen in Fig. 18, the position of the
assumed reflection from the back surface changes. The result of this experiment
suggested that the structure of paper is visible to OCT, but in the following experiments,
this statement was proven to be wrong if no refractive index matching agents are used.
So, the result was too optimistic and promising. It is possible that the measured data
shows the real back surface, but if the measurement would have been taken from the
-50 0 50 100 150 200
0
0.05
0.1
Depth [um]
V
o
l
t
a
g
e

[
V
]
-50 0 50 100 150 200
0
0.05
0.1
Depth [um]
V
o
l
t
a
g
e

[
V
]
-50 0 50 100 150 200
0
0.05
0.1
Depth [um]
V
o
l
t
a
g
e

[
V
]
-50 0 50 100 150 200
0
0.05
0.1
Depth [um]
V
o
l
t
a
g
e

[
V
]
52
other transversal position, the results would have been completely different. One possible
cause for these peaks is random noise caused by speckle modulation of the OCT signal,
and they can probably be made to disappear by averaging a sequence of different A-scans
obtained inside a small area in the zone of interest.
A paper sheet is a 3D network of fibers, fines such as fiber fragments, and various
kinds of additives. Since this kind of network is optically very complex, it was best to
begin with a controlled simulation network, with fibers as the only scatterers. In the next
experiment, a 3D wood fiber network was constructed to simulate the structure of paper.
A bleached pine pulp solution was spread on top of a smooth glass plate and dried at
room temperature for about 24 hours to obtain a completely dry sample. To keep the
amount of data as small as possible, a fairly low sampling frequency was used, and the
data was decimated before 3D transformation. Fig. 19 presents a 3D OCT image of this
simulation network.
Fig. 19. 3D image of a paper simulation network. The size of the image is 2 x 2 x 0.075* mm
(*optical).
53
The size of the image is 2 x 2 x 0.075* mm (*optical), and it was constructed using 200
OCT slice images with 10 m transversal step increments. In addition, the image was
slightly smoothed using 3D Gaussian smoothing. A detailed data analysis and 3D image
formation was described in section 4.4. The image reveals that 3D imaging of a wood
fiber network is possible even with a conventional OCT system without any preparation
of the sample. This is because the diameter of the fibers (30 m - 40 m) is well above
the axial and transversal resolution of the system, and only fines like fiber fragments are
present in addition to pure fibers. However, problematic with this kind of visualization
and processing of 3D images using a regular PC is that the process requires a large
amount of time for processing, even after the slight decimation of the data. Moreover, it
has to be pointed out that this network is still quite different from a real paper structure,
which contains more densely packed fibers and variable additives. But, it can be stated
that the first layers of the paper surface are visible to OCT if the plain fibers are the only
scatterers. However, thats not a very common case in real commercial grades of paper.
Most paper surfaces are calendered and coated, which changes the structure presented to
one that is more complex for OCT imaging.
In the next experiments, typical copy paper was measured using OCT. Several
research groups have corroborated the fact that OCT is incapable of measuring the
structure of paper, such as copy paper, beneath the surface through many layers of fibers.
It is partly true if no preparation is performed before the measurement. This type of paper
contains about 20-25% filler and its densely packed fibers and fines produce very strong
light scattering. Nonetheless, there is a way to make the structure of the entire sheet
visible through measurements performed only on one side.
Several so-called clearing agents, i.e. refractive index matching agents, can be added
to paper to enhance its optical properties. One of the most applicable is benzyl alcohol,
which fills the pores within the paper structure effectively and sufficiently fast.
Furthermore, its refractive index, 1.54, is close to the refractive index of cellulose, 1.55,
which is the main construction material of paper fibers. The effect of various refractive
index maching liquids on paper is discussed in Paper V. In this paper, the aim was to find
the best possible liquid for this purpose and to demostrate experimentally that, as in
medical applications, the addition of an appropriate refractive index matching liquid
reduces the effective scattering coefficient and improves the probability of recording
photons carrying important information about the inner structure of paper using OCT
(Zimnyakov et al. 2002).
The light transmittance measurements introduced in detail in Paper V suggested that
the sorption and evaporation characteristics of ethanol and glycerol were impractical for
OCT measurements. Minimizing the refractive index mismatch to reduce coherent
multiple scattering by the refractive index matching liquid doesnt completely depend on
the optical properties of the liquid used, but also on its sorption and evaporation
characteristics when added to paper. Ethanol evaporated too quickly and both glycerols
sorpted too slowly and didnt give stable conditions for measurement, mainly because of
sorption in the fibers, which caused structural changes in the paper. Only 1-pentanol and
benzyl alcohol had suitable sorption properties, but only benzyl alcohol offered
sufficiently good imaging depth. The simulations done by Kirillin et al. also supported
this result (Kirillin et al. 2006).
54
The effect of refractive index matching is demonstrated in Fig. 20, where enhancement
of the imaging depth was tested with dry paper and then by wetting the paper sample with
1-pentanol and benzyl alcohol. The reconstructed cross-section of dry copy paper
demonstrates a case where the imaging depth of OCT is not adequate. As the scanning
depth of the OCT system was limited, the reconstructed cross-section profile of dry paper
seems to be cut away (a). In reality, however, the signal will fade away in the same way
as in the sample representing 1-pentanol (b), where the bottom surface of the paper is
unclear. Although paper wetted with 1-pentanol transmits much more light than dry
paper, the imaging depth is still insufficient. A proper image of the sheets bottom surface
can only be obtained with benzyl alcohol (c). Reflectance from the top and bottom
surface of a paper sheet wet with benzyl alcohol was equal. Light scattering and
absorption, however, attenuates the reflected intensity from the bottom surface.
Fig. 20. Reconstructed 2D cross-sections of the paper samples: (a) dry paper, paper wet with
(b) 1-pentanol and (c) benzyl alcohol. The pixel size of the images is 15 m and 0.39 m in the
transversal and depth directions, respectively. The black arrow points to the scanning depth
limit of the measurement system used. The white dashed line depicts an estimation of the rear
border of the paper.
The addition of the refractive index matching liquid yields a higher probability of
recording photons carrying important information about the inner structure of paper and
increases the imaging depth of OCT and enables the detection of the top and bottom
surface of the paper sheets. It should be noted that the imaging quality of the inner
structure of paper depends on the resolution provided by the OCT system. The method
presented also allows obtaining information about the three-dimensional structure of
paper and the evaporation of the refractive index matching liquid used.
In the next experiment, the copy paper sample was exposed to benzyl alcohol, which
filled the pores and pits of the surface and penetrated through the capillaries into the
pores and cavities of the sheet, without absorbing or diffusing into the fibers during the
55
measuring period. A droplet of benzyl alcohol was placed on both sides of the sample
sheet, which was measured right after this liquid exposure. Fig. 21 presents a 3D image
of this sample from the front side, while Fig. 22 shows an image from the back side.
Fig. 21. 3D image of a copy paper sample (front). The size of the image is 0.75 x 0.75 x 0.16*
mm (*optical).
Fig. 22. 3D image of a copy paper sample (back). The size of the image is 0.75 x 0.75 x 0.16*
mm (*optical).
56
The sizes of the images are 0.75 x 0.75 x 0.16* mm (*optical), and they were constructed
using 50 OCT slice images with 15 m transversal step increments. However, the sample
was measured only from the front side and then flipped relative to the z-axis in a
visualization program to show the back side. In addition, the images were slightly
smoothed using 3D Gaussian smoothing. A detailed data analysis and 3D image
formation were described in section 4.4.
It is obvious that single fibers are not visible on the surface, due to moderate filler
content, soft calendering, and the use of benzyl alcohol to fill surface pores and pits.
Nonetheless, the back side of the sample can be made visible without physically flipping
the sample. A Lorenzen & Wettre 51 standard thickness measurement device gave us a
thickness value of 102 m for the sample. Using OCT, the average optical thickness of
the sample was 158.8 m, corresponding to a physical thickness of 102.8 m, assuming a
porosity of 50%.
The experiments reported in this section suggest that optical coherence tomography
has the potential to become a new key method in paper characterization and evaluation. It
is a viable alternative to conventional methods and has the capacity to produce additional
information related to paper quality. In addition to surface characterization, the internal
structure of paper can be imaged optically in a non-contact manner. However, being a
complex process, optical imaging of the internal structure of paper poses a great
challenge for research and development. The results presented here are promising, and it
is suggested that OCT can be applied to paper structure imaging. For example, the 2D
and 3D images of pulp and paper presented in Fig. 19-22 can be greatly enhanced by
using a light source with a broader spectral bandwidth together with signal or image
processing algorithms, such as the deconvolution method and Kalman filtering (Hast et
al. 2004, Izatt et al. 2002). As stated earlier, imaging paper necessitates the use of an
agent with a certain set of characteristics. Of particular interest here is the viscosity,
refractive index, and chemical composition of the solution, because it has to fill the pores
and pits of the paper sheet without penetrating the fiber itself.


6 Discussion
The need to analyze and characterize paper is continually increasing in the modern paper
industry. Current measurement methods tend to be either slow, labor-intensive, expensive
or invasive. The aim of this work was to explore the capabilites of low coherence
interferometry and optical coherence tomography in paper measurements. The original
idea was to provide a new method for on-line measurement of paper thickness, but it was
noticed at a very early stage that it was unrealistic to realize this idea with these methods.
It is a fact that unprepared paper presents such a challenge for optical imaging that
neither LCI nor OCT can give the desired result if the complete sheet must be imaged
from one side. There are several optical profilometers commercially available, but it is
surprising that in the modern paper industry, old-fashioned methods are used to evaluate
surface quality. Nevertheless, some of the methods are non-scientific, they all yield
slightly different values for roughness, and they typically require a number of separate
devices to cover all types of paper.
The combination of a profilometer and an imaging device makes the device
construction used in this thesis project attractive. Commercial optical profilometers are
accurate devices for giving a nice topography of the sample and for calculating surface
roughness parameters. This is where OCT comes in, by offering the capability to
reconstruct 2D and 3D images of papers internal structure. As stated earlier, this is not a
straightforward process, and some preparation must be carried out before the
measurement.
At the beginning of this project, a multipurpose OCT device was constructed. The aim
was to keep the construction as flexible as possible to enable easy changes of its
components. High speed and optimization of the measurement sequence of the device
were not the goal of this thesis project, and the basic construction has been the same
during all the measurements. The main idea was to introduce a new method for
characterizing the structure of paper.
Several paper surfaces were measured to give an overview of the capabilities of the
LCI technique for surface characterization. The analysis of the results includes a
roughness evaluation and a comparison of the results with those obtained with a
commercial optical profilometer. A combined analysis of topography, roughness, gloss
58
maps, and average gloss provides more information for the paper researcher about the
surface parameters and how they affect printability, for example.
2D and 3D imaging of paper poses a great challenge for any imaging device, and
especially for OCT, which has been used in medicine for years. Surprisingly, however,
there are very few industrial applications. Of course the optical complexity of paper is
extremely high, which has slowed development of the OCT technique for that particular
application. But, as the experiments done in this thesis project indicate, there is a way to
make the paper structure visible to OCT.
6.1 Measurements
The first studies focused on surface analysis of paper. The surface topografies and
corresponding roughness values of three different grades of paper presented in Paper VI
give a nice overview of the capabilities of LCI in measuring several grades of paper.
Grades of paper from rough base paper to smooth photo printing papers can be measured
using the same setup. Compared with conventional air-flow methods, where typically two
devices are needed to cover all types of paper, LCI can offer a cost-effective alternative,
and in addition has less error sources present in the measurement event.
The next set of samples was measured to compare roughness values measured using
LCI with roughness values measured with a commercial profilometer. The linear
correlation coefficient was acceptable only with roughness parameters R
a
and R
q
. All the
other parameters correlated more or less badly. The noncorrelation was assumed to be
caused by differences in the focusing geometry of the devices and differences in the
measured area. This correlation coefficient was calculated by comparing the roughness
values parameter by parameter. The interesting point was that when all the parameters of
a single sample from A to D measured using LCI and the Altisurf were compared, high
correlation was found except in sample C. This revealed that even though the single
parameters didnt correlate well, the relativity of the parameters inside single samples
remained constant.
The surface roughness of paper has an important role in the gloss of paper. If joint
information on both surface roughness and gloss could be measured at the same time and
from the same areas of paper, it would help papermakers in their research and
development of optimal paper surface quality, which is crucial for optimal ink absorption
in the printing process. Three different paper samples were measured using two
glossmeters and LCI. The results show linear correlation between the roughness and gloss
value measured with a commercial glossmeter, which was an expected result. The
roughness and averaged gloss measured with a DOG didnt give such a correlation as
LCI and the commercial glossmeter, but the measurements revealed that the surface
roughness of paper was not the only dominant factor affecting gloss, and it cant be used
alone to predict gloss variation on the surface. Variations in the refractive index due to
fillers and surface coatings can cause large differences in local gloss. A more detailed
comparison of topography, local roughness, gloss, and local gloss should be made to
come up with a better combination of these techniques and correlation of the results.
59
Filler content evaluation of paper using OCT was discussed briefly in this thesis. The
correspondence of the slope of the least squares (LSQ) fitted line to the averaged
logarithmic fringe signal envelope was clear. A change from 0% to 30% in filler content
decreased the slope of the LSQ-fitted line by 22%. The correspondence was not perfectly
linear, but it is hard to speculate the sensitivity or linearity of the measurement with such
a limited number of samples. The effect is the well known optical clearing of tissues in
medicine to enhance the imaging depth of OCT in living tissue, but it can also be adapted
to determine the filler content of paper. The refractive index of PCC, n=1.59, is close to
the refractive index of cellulose, n=1.55. The filler actually fills in the spaces between
fibers, increasing the refractive index of the fibers surroundings and decreasing the
mismatch between the refractive indices of the different scattering components that
caused the slope, i.e. the reflectivity versus depth slope of the OCT signal, to show a
corresponding change. Secondly, increasing the filler content causes the filler particles to
aggregate, which makes the fillers less effective in scattering light (Leskel 1998), which
partly effects the slope of the signal. The method is not capable of detecting the absolute
value of the filler content without any calibration, but it is a useful tool for monitoring the
filler content in a manufacturing process.
2D and 3D imaging of paper pose a great challenge for OCT. It has been stated in this
thesis that imaging of unprepared commercial paper using OCT is impossible. The idea
of measuring paper was introduced in section 5.4, where the simplest form of so-called
paper, i.e. a network of fibers, was constructed. The result was promising and showed
that a wood fiber network is visible to OCT. But, when the experiments were continued
with commercial paper, it was immediately clear that only the surface or the very first
layers are visible in the image. Multiple scattering distorted the images completely, and a-
scans showed a signal far beyond where the papers back surface would have been. It was
a clear effect of a far more complex structure than the measured simulation network.
The idea of optical clearing of tissues in medicine was adapted also to paper
measurements. In the case of paper it is better to use the term refractive index matching.
It was found that certain alcohols had appropriate optical properties and they were inert
enough to prevent damage to paper. The best agent was found to be benzyl alcohol, which
was then used in the imaging of copy paper. However, commercial copy paper was not a
very convenient grade of paper for the experiments because it is almost impossible to
know its exact constituents. On the other hand, there is a practically unlimited number of
samples available for measurements and this grade of paper is widely used, which makes
its research interesting. The best possible sample would be a grade of paper that contains
more or less plain fibers. But such a grade is far from real paper in every day use, and it is
not interesting in paper research. But for OCT research, it gives a completely new
application area. The 2D and 3D OCT images cant be compared with the results of any
other group. There is practically no OCT research on paper other than the research
presented in this thesis. Medical applications are more attractive and the main focus of
research seems to stay there. But, the main point was to prove that paper can be treated so
that it can be imaged using OCT. The ability to resolve the structural components of
paper remains unkown until a more close comparison with a technique that gives
information from the exact same position on the paper can be performed. Preparation of
the sample makes that a challenging problem, and for that reason it was not conducted
during this research.
60
6.2 Future research
Typically, a paper sheet is around 100 m thick, but the thickness depends highly on the
grade of paper. It is obvious that it is not possible to reconstruct highly detailed 2D slice
or 3D images of paper with a 15 m spatial resolution. A resolution in the range of 1-2
m would highly enhance the capabilities of OCT in paper imaging. Micrometer-scale
resolution can be achieved, e.g. by using the Ti:Sapphire femtosecond laser currently
available in the Optoelectronics and Measurement Techniques Laboratory. The drawback
of such a laser is the need for a stable and clean operating environment. Its only for
laboratory purposes, but the results that could possibly be achieved by using such a high-
resolution imaging system in paper imaging would greatly enhance the results presented
in this thesis.
Optimization of the speed of the imaging device was not the main point in the research
of this thesis project. More resources will be focused on that in the future. By using the
FD-OCT (Fourier Domain Optical Coherence Tomography) scheme, the imaging
sequence can be greatly speeded up to 30-50 2D slice images per second without any
moving parts in the system (Endo et al. 2005, Leitgeb et al. 2004, Yatagai et al. 2005).
Transversal resolution is limited by the focusing optics. There is a trade-off between
resolution and imaging depth. The smaller the focused spot in the sample, the smaller is
the imaging depth if the transversal resolution is assumed to stay within the accepted
range. Fortunately a typical paper sheet is rather thin compared with tissue imaging in
medicine, but still the setup could be enhanced by using a dynamic focusing scheme,
which ensures that the transversal resolution is constant in the depth direction.
The surface and near-surface structure of paper present challenging problems, which
could be imaged using OCT. If the spatial resolution is enhanced to the micrometer level,
the coating color and printing layers could possibly be separated from the base paper. The
spatial distribution of the coating color, and spread and penetration of printing ink in
paper are important factors determining paper quality and printability. These layers are
assumed to be visible to OCT without any preparation of the paper.
Signal and image processing and post-processing were discussed very briefly in this
thesis. Techniques like iterative deconvolution can be used to enhance the resolution of
the system without any changes to the imaging system itself (Paes et al. 2005). Even
without the knowledge of the phase of the fringe signal, the iterative deconvolution
algorithm has been applied to enhance the resolution of OCT in paper imaging (Alarousu
et al. 2005, Hast et al. 2004). The Kalman filtering method can be applied instead of
analog processing to signal filtering and envelope detection (Alarousu et al. 2003). The
Kalman filter is a noise-immune and stable way to condition a digital signal. The next
step is to apply the Kalman filtering method to real-time signal processing and compare
the results with an analog processing method. In an ideal case the Kalman filter should be
noise-free and give a better signal-to-noise ratio compared with analog processing.


7 Summary
In this thesis project, Low Coherence Interferometry and Optical Coherence Tomography
were applied to paper measurements. The developed LCI/OCT system was first used to
measure the surface structure of paper. In the first experiments, different grades of paper
were selected to give a maximum variation in surface structure. The results show that
grades of paper from a rough base paper to highly coated photo printing paper can be
measured and roughness parameters calculated. The roughnesses R
a
ranged from 0.8 m
to 6.9 m.
In the second experiment, a set of five samples was measured to compare the
roughness values measured using LCI with roughness values measured with a
commercial Altisurf 500 profilometer. The linear correlation coefficient was acceptable
only with roughness parameters R
a
and R
q
: 0.97 and 0.98, respectively. All the other
parameters correlated more or less badly. This correlation coefficient was calculated by
comparing the roughness values parameter by parameter. But, when all the parameters of
a single sample from A to D measured using LCI and the Altisurf were compared, linear
correlation coefficients ranging from 0.98 to 1 were found, except in sample C, where the
corresponding value was 0.73.
Next, the surface quality of paper was evaluated using LCI, a DOG and a commercial
glossmeter. The results show linear correlation between roughness and the gloss value
measured with the commercial glossmeter. The linear correlation coefficients between the
roughness R
a
and Gloss GU20/ GU60 were -0.99/-0.99. The roughness R
a
and averaged
gloss measured with the DOG didnt give a linear correlation coefficient (-0.61) like that
of gloss measured with the commercial glossmeter.
In the next experiment, the filler content of paper was determined using OCT. The
correspondence of the slope of the LSQ-fitted line to the averaged logarithmic fringe
signal envelope and filler content was clear. A change from 0% to 30% filler content
decreased the slope of the LSQ-fitted line by 22%.
The last experiments focused on 2D and 3D imaging of paper. First a simple self-made
fiber network was constructed and imaged using OCT. The visibility of fibers was clear.
In the next experiments, commercial copy paper was used as a sample. Several refractive
index matching agents were tested by light transmittance and OCT measurements to find
the best possible agent for enhancing the imaging depth in paper. Benzyl alcohol was
62
found to have the best possible combination of optical, evaporation, and sorption
characteristics. Lastly, copy paper was exposed to benzyl alcohol and measured using
OCT. 3D visualizations of the sample were presented. The reference value of the
thickness of the paper was measured with a Lorenzen & Wettre 51 standard thickness
measurement device, which gave a value of 102 m. The thickness calculated from the
3D image obtained with OCT was 102.8 m when the porosity of the paper was assumed
to have a value of 50%.
The combination of an optical profilometer and a tomography imaging device
introduced in this thesis makes the construction attractive. The results show that the
surfaces of various grades of paper can be evaluated with a single device. When
combined with a device that measures the local gloss of the surface, two of the most
important parameters characterizing the surface properties of paper can be measured in
parallel to evaluate the surface quality, e.g. for printing. In addition to the surface, the
structure beneath the surface is visible to the device. If only the very first layers are
visible without any paper preparation, the whole sheet can be imaged with a spatial
resolution defined by the spectral characteristics and focusing optics of the light source
by using refractive index matching agents to fill the pores inside the paper. The results
introduced in this thesis suggest that these methods and this device construction could be
a viable alternative to conventional methods in paper research and even in on-line quality
control.


References
Aaltio E (1969) Paperin rakenne ja ominaisuudet. In: Ryti N (ed), Paperin valmistus, Suomen
Paperi-insinrien Yhdistyksen oppi- ja ksikirja III, Chap E, Frenckell printing works,
Helsinki, Finland, (In Finnish).
Akcay C, Parrein P & Rolland J P (2002) Estimation of longitudinal resolution in optical coherence
imaging. Appl. Opt. 41(25): 5256-5262.
Alarousu E, Gurov I, Hast J, Myllyl R, Prykri T & Zakharov A (2003) Optical coherence
tomography evaluation of internal random structure of wood fiber tissue. Proc. of SPIE 5132:
149-160.
Alarousu E, Gurov I, Hast J, Myllyl R, Prykri T & Zakharov A (2003) Optical coherence
tomography evaluating the random tissues based on dynamical processing the stochastic low-
coherence interference fringes. Proc. of SPIE 5140: 33-42.
Alarousu E, Gurov I, Hast J, Myllyl R & Zakharov A (2003) Optical coherence tomography of
multilayer tissue based on the dynamical stochastic fringe processing. Proc. of SPIE 5149: 13-
20.
Alarousu E, Okkonen M & Myllyl R (2005) Iterative Restoration of Optical Coherence
Tomography Images of Paper Structure. Proc. of 6
th
Japan-Finland Joint Symposium on Optics
in Engineering: 61-62.
Aschan P-J (1969) Pllystyksen teoriaa. In: Ryti N (ed), Paperin valmistus, Suomen Paperi-
insinrien Yhdistyksen oppi- ja ksikirja III, Chap O, Frenckell printing works, Helsinki,
Finland, (In Finnish).
Ballif J, Giannotti R, Chavanne P, Wlti R & Salath R P (1997) Rapid and scalable scans at 21
m/s in optical low-coherence reflectometry. Opt. Lett. 22: 757-759.
Baumgartner A, Hitzenberger C K, Sattmann H & Drexler W (1998) Signal and Resolution
Enhancements in Dual Beam Optical Coherence Tomography of the Human Eye. J. Biomed.
Opt. 3(1): 45-54.
Baumgartner A, Mller B, Hitzenberger C K, Drexler W & Fercher A F (1997) Measurement of the
Posterior Structures of the Human Eye In Vivo by Partial Coherence Interferometry using
Diffractive Optics. Proc. SPIE 2981: 85-93.
Bestenyamov K P, Gordienko V M, Konovalov A N & Podshivalov A A (2004) Studying of mass
transfer processes and defectoscopy of paper porosity using femtosecond OCT. Proc. SPIE
5475: 56-65.
Bland M-C & Bennet J M (2000) Effect of local microroughness on the gloss uniformity of
printed paper surfaces. Appl. Opt. 39(16): 2719-2726.
64
Bhushan B, Wyant J C & Koliopoulos C (1985) Measurement of surface topography of magnetic
tapes by Mirau interferometry. Appl. Opt. 24(10): 1489-1497.
Borch J (2002) Optical and appearance properties. In: Borch J, Lyne M B & Mark R E, Habeger C
C Jr (eds) Handbook of Physical Testing of Paper, Vol. 2, Second Edition, Revised and
Expanded, Chap. 4, Marcel Dekker Inc., New York, USA.
Brezinski M E & Fujimoto J G (1999) Optical Coherence Tomography: High-Resolution Imaging
in Nontransparent Tissue. IEEE J. Sel. Top. Quant. Electron. 5(4): 1185-1192.
Bristow J A (1986) The Paper Surface in Relation to the network. In: Bristow J A & Kolseth P
(eds), International Fiber Science and Technology Series 8, Paper Structure and Properties,
Chap. 8, Marcel Dekker, Inc., New York, USA.
Carlsson J, Hellentin P, Malmqvist L, Persson A, Persson W & Wahlstrm C-G (1995) Time-
resolved studies of light propagation in paper. Appl. Opt. 34(9): 1528-1535.
Carlsson J, Persson W, Hellentin P & Malmqvist L (1995) The propagation of light in paper.
TAPPI PRESS, Atlanta: 83.
Chinga G & Helle T (2002) Structure characterization of pigment coating layer on paper by
scanning electron microscopy and image analysis. Nordic Pulp Paper Res. J. 17(3):307-312.
Clements J C, Zvyagin A V, Silva K K M B D, Wanner T, Sampson D D & Cowling W A (2004)
Optical coherence tomography as a novel tool for non-destructive measurement of the hull
thickness of lupin seeds. Plant breeding 123: 266-270.
Danielson B L & Whittenberg C D (1987) Guided-wave reflectometry with micrometer resolution.
Appl. Opt. 26: 2836-2842.
Drexler W (2004) Ultrahigh-resolution optical coherence tomography. J. Biomed. Opt. 9(1): 47-74.
Drexler W, Hitzenberger C K, Baumgartner A, Findl O, Sattmann H & Fercher A F (1998)
Investigation of Dispersion Effects in Ocular Media by Multiple Wavelength Partial Coherence
Interferometry. Exp. Eye Res. 66: 25-33.
Endo T, Yasuno Y, Makita S, Itoh M & Yatagai T (2005) Profilometry with line-field Fourier-
domain interferometry. Opt. Expr. 13(3): 695-701.
Ettl P, Bohn G, Horneber C, Andretzky P, Konzog M, Knauer M, Pavlicek P & Husler G (1999)
Modifications of broad band interferometry for high accuracy 3D-measurements in medicine
and industry. Proc. ODIMAP II Topical Meeting on Optoelectronic Distance/Displacement
Measurements and Applications 36-41.
Fercher A F (1996) Optical coherence tomography. J. Biomed. Opt. 1: 157-173.
Fercher A F & Hitzenberger C K (1999) Optical Coherence Tomography in Medicine. In: Asakura
(ed) International Trends in Optics and Photonics ICO IV. Part. VII, Springer-Verlag, Berlin
Heidelberg, Germany.
Fercher A F, Hitzenberger C K, Moreno-Barriuso E, Sticker M, Leitgeb R, Sattmann H (2000)
Optical Coherence Tomography Technique for Thermal Light Sources. Proc. of SPIE 4160: 63-
68.
Fercher A F & Roth E (1986) Opthalmic laser interferometry. Proc. SPIE 658: 48-51.
Fujimoto J G (2002) Optical Coherence Tomography: Introduction. In: Bouma B E & Tearney G J
(eds), Handbook of Optical Coherence Tomography, Chap. 1, Marcel Dekker, Inc., New York,
USA.
Fujimoto J G (2003) Optical coherence tomography for ultrahigh resolution in vivo imaging. Nat.
Biotechnol. 21: 1361-1367.
Gadelmawla E S, Koura M M, Maksoud T M A, Elewa I M & Soliman H H (2002) Roughness
parameters. J. Mater. Process. Tech. 123: 133-145.
Gilgen H H, Novk R P & Salathe R P (1989) Submillimeter Optical Reflectometry. J. Lightwave
Technol. 7(8): 1225-1233.
65
Hahn L D (2002) The paper and board testing laboratory. In: Borch J, Lyne M B & Mark R E,
Habeger C C Jr (eds) Handbook of Physical Testing of Paper, Vol. 2, Second Edition, Revised
and Expanded, Chap. 1, Marcel Dekker Inc., New York, USA.
Hast J, Gurov I, Alarousu E, Zakharov A & Myllyl R (2004) Enhancing the OCT images by the
low coherence fringe envelope deconvolution method. Proc. SPIE 5486: 180-186.
Hee M R (2002) Optical Coherence Tomography: Theory. In: Bouma B E & Tearney G J (eds),
Handbook of Optical Coherence Tomography. Chap. 2, Marcel Dekker, Inc., New York, USA.
Hee M R, Izatt J A, Swanson E A, Huang D, Schuman J S, Lin C P, Puliafito C A & Fujimoto J G
(1995) Optical Coherence Tomography for Ophthalmic Imaging: New technique delivers
micron-scale resolution. IEEE Engineering in Medicine and Biology Magazine 14(1): 67-76.
Heikkurinen A (1999) Single fiber properties. In: Levlin J-E & Sderhjelm L (eds), Papermaking
science and technology 17, Pulp and Paper Testing, Chap. 2, Gummerus Printing, Jyvskyl,
Finland.
Hiltunen E (1999) Papermaking properties of pulp. In: Levlin J-E & Sderhjelm L (eds),
Papermaking science and technology 17, Pulp and Paper Testing, Chap. 3, Gummerus Printing,
Jyvskyl, Finland.
Hirsimki O (1972) Paperimassan flokkisuuden mittaaminen fokusoidun ultranen avulla. Thesis
for the Diploma Engineer Degree, University of Oulu (In Finnish).
Hitzenberger C K, Baumgartner A, Drexler W & Fercher A F (1999) Dispersion Effects in Partial
Coherence Interferometry: Implications for Intraocular Ranging. J. Biomed. Opt. 4(1): 144-151.
Hitzenberger C K, Baumgartner A & Fercher A F (1998) Dispersion induced multiple signal peak
splitting in partial coherence interferometry. Opt. Comm. 154: 179-185.
Hitzenberger C K, Drexler W, Baumgartner A & Fercher A F (1997) Dispersion effect in partial
coherence interferometry. Proc. SPIE 2981: 29-36.
Hitzenberger C K & Fercher A F (2002) Alternative OCT techniques. In: Bouma B E & Tearney G
J (eds), Handbook of Optical Coherence Tomography. Chap. 13, Marcel Dekker, Inc., New
York, USA.
Holmstad R, Antoine C, Nygrd P, Helle T (2002) Quantification of the three-dimensional paper
structure - Methods and potential. Proc. of 88th Annual Meeting Technical Section A: 9-12.
Huang D, Swanson E A, Lin C P, Schuman J S, Stinson W G, Chang W, Hee M R, Flotte T,
Gregory K, Puliafito C A & Fujimoto J G (1991) Optical Coherence Tomography. Science 254:
1178-1181.
Hggkvist M (1999) Porous structure in Paper studied by NMR. Licenciate thesis, Kungl. Tekniska
Hgskola, Royal Ins. of Tech., Stockholm.
Izatt J A, Rollins A M, Ung-Arunyawee R & Yazdanfar S (2002) System Integration and
Signal/Image Processing. In: Bouma B E & Tearney G J (eds), Handbook of Optical Coherence
Tomography. Chap. 6, Marcel Dekker, Inc., New York, USA.
Kajanto I, Laamanen J & Kainulainen M (1998) Paper bulk and surface. In: Niskanen K (ed),
Papermaking science and technology 16, Paper Physics, Chap. 3, Gummerus Printing,
Jyvskyl, Finland.
Karppinen T, Kassamakov I, Haeggstrm E & Stor-Pellinen J (2004) Mesuring paper wetting
processes with laser transmission. Meas. Sci. Tech. 15: 1223-1229.
Karppinen A, Kilpel A, Karras M, Tornberg J & Myllyl R (1995) Papermaking furnish properties
estimated by time-resolved spectroscopy. J. Pulp Paper Sci. 21(5): 151-154.
Kirillin M Yu, Priezzhev A V, Hast J & Myllyl R (2006) Monte Carlo simulation of optical
clearing of paper in optical coherence tomography. J. Quant. Electron. 36(2): 174-180.
Kolseth P & De Ruvo A (1986) The Cell Wall Components of Wood Pulp Fibers. In: Bristow J A
& Kolseth P (eds), International Fiber Science and Technology Series 8, Paper Structure and
Properties, Chap. 1, Marcel Dekker, Inc., New York, USA.
66
Krehut L (2003) Integrated Electronic System for Optical Coherence Tomography. Thesis for the
Diploma Engineer Degree, University of Oulu.
Krogerus B (1999) Fillers and pigments. In: Neimo L (ed), Papermaking science and technology 4,
Papermaking Chemistry, Chap. 6, Gummerus Printing, Jyvskyl, Finland.
Kubelka P (1948) New contributions to the optics of intensely light-scattering materials. Part 1. J.
Opt. Soc. Am. 38: 448-457.
Kubelka P (1954) New contributions to the optics of intensely light-scattering materials. Part 2. J.
Opt. Soc. Am. 44: 330-335.
Kubelka P & Munk F (1931) A contribution to the optics of colorant layers. Z. Tech. Phys.
12(11a): 593-601, (In German).
Kubota T, Nara M & Yoshino T (1987) Interferometer for measuring displacement and distance.
Opt. Lett. 12: 310-312.
Kulkarni M D, Thomas C W & Izatt J A (1997) Image enhancement in optical coherence
tomography using deconvolution. Electron. Lett. 33(16): 1365-1367.
Laubscher M (2004) Innovative experimental concepts for optical coherence tomography. Doctoral
thesis. Swiss Federal Institute of Technology.
Lee H-K, Joyce M K & Fleming P D (2003) Influence of Pigment Particle Size and Packing
Volume on Printability of Glossy Inkjet Paper Coatings. Proc. of IS&T NIP19: 613-618.
Lee H-K, Joyce M K & Fleming P D (2005) Influence of Pigment Particle size and Pigment Ratio
on Printability of Glossy Ink Jet Paper Coatings. J. Imaging Sci. Tehcn. 49(1): 54-60.
Leitgeb R A, Drexler W, Unterhuber A, Hermann B, Bajraszewski T, Le T, Stingl A & Fercher A F
(2004) Ultrahigh resolution Fourier domain optical coherence tomography. Opt. Expr. 12(10):
2156-2165.
Leskel M (1993) A model for the optical properties of paper part 1. The Theory. (1993) Paperi ja
puu Paper and Timber 75: 683-688.
Leskel M (1998) Optical properties. In: Niskanen K (ed), Papermaking science and technology 16,
Paper Physics, Chap. 4, Gummerus Printing, Jyvskyl, Finland.
Levlin J-E (1999) Aim of pulp and paper testing. In: Levlin J-E & Sderhjelm L (eds),
Papermaking science and technology 17, Pulp and Paper Testing, Chap. 1, Gummerus Printing,
Jyvskyl, Finland.
MacGregor M A, Johansson P & Beland M C (1994) Measurements of small-scale gloss
variation in printed paper. Proc. Intl. Printing and Graphics Arts Conf.: 22-43.
Masters B R (1999) Early development of optical low-coherence reflectometry and some recent
biomedical applications. J. Biomed. Opt. 4: 236-247.
M-real Digital imaging (2004) Paper School - An education on paper, Bergisch Gladbach,
Germany.
Myller K, Peiponen K-E & Silvennoinen R (2003) Two-dimensional map of gloss of plastic
measured by diffractive-element-based glossmeter. Opt. Eng. 42: 3194-3197.
Nesbakk T, Mrseburg K & Helle T (2001) Relationship between fibre properties and cross-
sectional paper characteristics of mechanical pulp handsheets. The 3
rd
biennial Gullichsen
colloquium: 63-72.
Niskanen K, Kajanto I & Pakarinen P (1998) Paper structure. In: Niskanen K (ed), Papermaking
science and technology 16, Paper Physics, Chap. 1, Gummerus Printing, Jyvskyl, Finland.
Niskanen K & Rajatora H (2002) Statistical Geometry of Paper Cross-Sections. J. Pulp Paper Sci.
28(7): 228-233.
Paes S, Ruy S Y, Na J, Choi E S & Lee B H (2005) Application of Iterative Deconvolution
Methods for Optical Coherent Imaging. Proc. of SPIE 5690: 486-495.
Pan Y, Lankenau E, Welzel J, Birngruber R & Engelhardt R (1996) Optical Coherence-Gated
imaging of Biological Tissues. IEEE J. Sel. Top. Quant. Electron. 2(4): 1029-1034.
67
Pan Y, Birngruber R, Rosperich J & Engelhardt R (1995) Low-Coherence optical tomography in
turbid tissue: theoretical analysis. Appl. Opt. 34(28): 6564-6574.
Pauler N (1986) Opacity and Reflectivity of Multilayer Structures. In: Bristow J A & Kolseth P
(eds), International Fiber Science and Technology Series 8, Paper Structure and Properties,
Chap. 10, Marcel Dekker, Inc., New York, USA.
Rutar V & Letnar M (2001) Sorption Abilities of Paper and Board from Recycled Fibers. Project
document of Cost action E11.
Saarela J & Myllyl R (2001) Change in time of flight due to paper porosity measured with a tof
lidar. Proceedings of 4
th
Finnish Optics Days, 20-21 April, Tampere, Finland.
Saarela J, Trmnen M & Myllyl R (2001) Changes in time of flight in thermomechanical pulp
(TMP) measured with a streak camera. Proceedings of SPIE 4242: 156-163.
Saarela J, Trmanen M & Myllyl R (2003) Measuring pulp consistency and fines content with
streak camera. Meas. Sci. Technol. 14: 1801:1806.
Sampson D D (2004) Trends and prospect for optical coherence tomography. Proc. SPIE 5502: 51-
58.
Samuelsen E J, Gregersen O W, Houen P J, Helle T, Raven C & Snigirev A (2001) Three-
dimensional imaging of paper by use of synchrotron X-ray microtomography. J. Pulp Paper Sci.
27(2): 50-53.
Scallan A M & Borch J (1972) An interpretation of paper reflectance based upon morphology I.
Initial considerations. Tappi 55(4): 583-588.
Schmitt J M (1999) Optical Coherence Tomography (OCT): A Review. IEEE J. Sel. Top. Quant.
Electron. 5: 1205-1215.
Schmitt J M, Xiang S H and Yung K M (1999) Speckle in Optical Coherence Tomography. J.
Biomed. Opt. 4(1): 95-105.
Silvennoinen R, Myller K, Peiponen K-E, Pkknen E J & Salmi J (2004) Diffractive optical
sensor for gloss differences of injection molded plastics product. Sensors and Actuators A 112:
74-79.
Takada K, Yokohama I, Chida K & Noda J (1987) New measurement system for fault location in
optical waveguide devices based on an interferometric technique. Appl. Opt. 26: 1603-1606.
Tearney G J, Brezinski M E, Bouma B E, Boppart S A, Pitris C, Southern J F & Fujimoto J G
(1997) In vivo endoscopic optical biopsy with optical coherence tomography. Science 276:
2097-2099.
Tomlins P H & Wang R K (2005) Theory, developments and applications of optical coherence
tomography. J. Phys. D: Appl. Phys. 38: 2519-2535.
Vaarasalo J (1999) Optical properties of paper. In: Levlin J-E & Sderhjelm L (eds), Papermaking
science and technology 17, Pulp and Paper Testing, Chap. 8, Gummerus Printing, Jyvskyl,
Finland.
Vabre L, Dubois A, Beaurepaire E & Boccara C (2000) Optical coherence microscopy for high
resolution biological imaging. Proc. SPIE 4160: 24-30.
Van Der Reyden D (1992) Recent Scientific Research in Paper Conservation. J. Am. Inst. Conserv.
31(1): 117-138.
Veeco Instruments inc. (2002) Surface Measurement Parameters for Wyko

Optical Profilers.
Applications notes AN505, Tucson, USA.
Vorburger T V & Teaque E C (1981) Optical techniques for on-line measurement of surface
topography. Precis. Eng. 3(2): 61-83.
VTT Automation & Helsinki University of Technology (2005) Knowpap e-Learning
Environment for Papermaking and Automation (http://www.knowpap.com).
VTT Processes (2003) Research report reference number PRO5659903.
Wang R K (1999) Resolution improved optical coherence-gated tomography for imaging through
biological tissues. J. Modern Optics. 46(13): 1905-1912.
68
Wang R K & Tuchin V V (2004) Optical Coherence Tomography Light Scattering and Imaging
Enhancement. In: Tuchin V V (ed), Biomedical Diagnostics, Environmental and Material
Science, Volume 2, Part IV, Optical Coherence Tomography, Chap. 13, Kluwer Academic
Publishers, Boston, USA.
Wang C S, Cheng W H, Hwang C J, Burns W K & Moeller R P (1982) High-power low-
divergence superradiance diode. Appl. Phys. Lett. 41: 587-589.
Whitehouse D J (1994) Handbook of Surface Metrology. Chap. 2, Institute of Physics Publishing,
London, UK.
Yatagai T & Yasuno Y (2005) Ultra-fast Fourier-domain optical coherence tomography for
biomedical applications. Proc. of 6
th
Japan-Finland Joint Symposium on Optics in Engineering:
3-4.
Yoshimura T, Kida K & Masazumi N (1995) Development of an image-processing system for a
low coherence interferometer. Opt. Commun. 117: 207-212.
Youngquist R C, Carr S & Davies D E N (1987) Optical coherence domain reflectometry: a new
optical evaluation technique. Opt. Lett. 12: 158-160.
Yung K M, Lee S L & Schmitt J M (1999) Phase-domain processing of optical coherence
tomography images. J. Biomed. Opt. 4(1): 125-136.
Zakharov P V & Zimnyakov A A (2002) Wavelet Analysis of Fluctuations of Laser Radiation
Scattered from Interphase Boundaries in Porous Media. Tech. Phys. Lett. 28(12): 1015-1017.
Zhang Y, Sato M & Tanno N (2001) Numerical investigations of optimal syntesis of several low
coherence sources for resolution improvement. Opt. Comm. 192: 183-192.
Zimnyakov D A & Tuchin V V (2002) Optical tomography of tissues. Kvantovaya Elektron.
32(10): 849-867. (J. Quant. Electron. 32(10): 849-867.)
Zimnyakov D A, Zakharov P V Trifonov V A & Chanilov O I (2001) Dynamic Light Scattering
Study of the Interface Evolution in Porous Media. JETP Lett. 74(4): 216-221.


Original papers
I Alarousu E, Myllyl R, Gurov I & Hast J (2001) Optical Coherence Tomography in
Scattering Material for Industrial Applications. Proc. of SPIE 4595: 223-230.
II Alarousu E, Krehut L, Myllyl R & Hast J (2004) Optical coherence tomography
device for paper characterization. Proc. of SPIE 5475: 48-55.
III Peiponen K E, Alarousu E, Juuti M, Silvennoinen R, Oksman A, Myllyl R &
Prykri T (2006) Diffractive optical element based glossmeter and low coherence
interferometer in assessment of local surface quality of paper. Opt. Eng. 45, 043601.
IV Alarousu E, Prykri T, Hast J & Myllyl R (2004) Low Coherence Interferometry in
Paper Surface and Near-surface Characterization. Proc. of ODIMAP IV: 36-41.
V Fabritius T, Alarousu E, Prykri T, Hast J & Myllyl R (2006) Characterization of
Optically Cleared Paper by Optical Coherence Tomography. J. Quant. Electron.
36(2): 181-187. (DOI: 10.1070/QE2006v036n02ABEH013121)
VI Alarousu E, Krehut L, Prykri T & Myllyl R (2005) Study on the use of optical
coherence tomography in measurements of paper properties. Meas. Sci. Technol. 16:
1131-1137. (www.iop.org/journals/mst)
Original publications are not included in the electronic version of the dissertation.


ACTA UNI VE RS I TATI S OUL UE NS I S
Distributed by
OULU UNIVERSITY LIBRARY
P.O. Box 7500, FI-90014
University of Oulu, Finland
Book orders:
OULU UNIVERSITY PRESS
P.O. Box 8200, FI-90014
University of Oulu, Finland
S ERI E S C TE CHNI CA
240. Hmlinen, Matti (2006) Singleband UWB systems. Analysis and measurements
of coexistence with selected existing radio systems
241. Virtanen, Jani (2006) Enhancing the compatibility of surgical robots with magnetic
resonance imaging
242. Lumijrvi, Jouko (2006) Optimization of critical flow velocity in cantilevered fluid-
conveying pipes, with a subsequent non-linear analysis
243. Stoor, Tuomas (2006) Air in pulp and papermaking processes
244. Gyrgy, Zsuzsanna (2006) Glycoside production by in vitro Rhodiola rosea
cultures
245. zer-Kemppainen, zlem (2006) Alternative housing environments for the
elderly in the information society. The Finnish experience
246. Laurinen, Perttu (2006) A top-down approach for creating and implementing data
mining solutions
247. Jortama, Timo (2006) A self-assessment based method for post-completion audits
in paper production line investment projects
248. Remes, Janne (2006) The development of laser chemical vapor deposition and
focused ion beam methods for prototype integrated circuit modification
249. Kinnunen, Matti (2006) Comparison of optical coherence tomography, the pulsed
photoacousti c techni que, and the ti me-of -f l i ght techni que i n gl ucose
measurements in vitro
250. Iskanius, Pivi (2006) An agile supply chain for a project-oriented steel product
network
251. Rantanen, Rami (2006) Modelling and control of cooking degree in conventional
and modified continuous pulping processes
252. Koskiaho, Jari (2006) Retention performance and hydraulic design of constructed
wetlands treating runoff waters from arable land
253. Koskinen, Miika (2006) Automatic assessment of functional suppression of the
central nervous system due to propofol anestheti c i nfusi on. From EEG
phenomena to a quantitative index
254. Heino, Jyrki (2006) Harjavallan Suurteollisuuspuisto teollisen ekosysteemin
esimerkkin kehitettess hiiliterksen ympristmynteisyytt
255. Gebus, Sbastien (2006) Knowledge-based decision support systems for
production optimization and quality improvement in the electronics industry
C256etukansi.fm Page 2 Tuesday, November 14, 2006 1:26 PM
A
B
C
D
E
F
G
UNI VERSI TY OF OULU P. O. Box 7500 FI - 90014 UNI VERSI TY OF OULU FI NLAND
A C T A U N I V E R S I T A T I S O U L U E N S I S
S E R I E S E D I T O R S
SCIENTIAE RERUM NATURALIUM
HUMANIORA
TECHNICA
MEDICA
SCIENTIAE RERUM SOCIALIUM
SCRIPTAACADEMICA
OECONOMICA
EDITOR IN CHIEF
EDITORIAL SECRETARY
Professor Mikko Siponen
Professor Harri Mantila
Professor Juha Kostamovaara
Professor Olli Vuolteenaho
Senior Assistant Timo Latomaa
Communications Officer Elna Stjerna
Senior Lecturer Seppo Eriksson
Professor Olli Vuolteenaho
Publication Editor Kirsti Nurkkala
ISBN 951-42-8213-2 (Paperback)
ISBN 951-42-8214-0 (PDF)
ISSN 0355-3213 (Print)
ISSN 1796-2226 (Online)
UNI V E RS I TAT I S OUL UE NS I S
ACTA
C
TECHNICA
OULU 2006
C 256
Erkki Alarousu
LOW COHERENCE
INTERFEROMETRY AND
OPTICAL COHERENCE
TOMOGRAPHY IN PAPER
MEASUREMENTS
FACULTY OF TECHNOLOGY,
DEPARTMENT OF ELECTRICAL AND INFORMATION ENGINEERING,
INFOTECH OULU,
UNIVERSITY OF OULU
C

2
5
6
A
C
T
A

E
r
k
k
i

A
l
a
r
o
u
s
u
C256etukansi.fm Page 1 Tuesday, November 14, 2006 1:26 PM