VSEBINA – CONTENTS
PREGLEDNI ^LANKI – REVIEW ARTICLES
Experimental study of some masonry-wall coursework material types under horizontal loads and their comparison
Eksperimentalna raziskava zgradbe nekaterih zidarskih zidov – vodoravna obremenitev in primerjava uporabljenih materialov
M. Kamanli, M. S. Donduren, M. T. Cogurcu, M. Altin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
IZVIRNI ZNANSTVENI ^LANKI – ORIGINAL SCIENTIFIC ARTICLES
Synthesis of aluminium foams by the powder-metallurgy process: compacting of precursors
Sinteza aluminijevih pen po postopku metalurgije prahov: stiskanje prekurzorjev
I. Paulin, B. [u{tar{i~, V. Kevorkijan, S. D. [kapin, M. Jenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
A new method for determining the remaining lifetime of coated gas-turbine blades
Nova metoda za izra~un preostale trajnostne dobe lopatic plinskih turbin
L. B. Getsov, P. G. Krukovski, N .V. Mozaiskaja, A. I. Rybnikov, K. A. Tadlja . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Reduction of ultra-fine tungsten powder with tungsten (VI)-oxide in a vertical tube reactor
Redukcija ultrafinih prahov volframovega(VI) oksida v reaktorju z vertikalno cevjo
@. Kamberovi}, D. Filipovi}, K. Rai}, M. Tasi}, Z. An|i}, M. Gavrilovski . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Oxygen diffusion in the non-evaporable getter St 707 during heat treatment
Difuzija kisika v getru St 707 med toplotno obdelavo
S. Avdiaj, B. [etina - Bati~, J. [etina, B. Erjavec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
The modeling of auger spectra
Modeliranje augerjevih spektrov
B. Poniku, I. Beli~, M. Jenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Modelling of the directional solidification of a leaded red brass flange
Modeliranje usmerjenega strjevanja prirobnice iz rde~e svin~eve medenine
V. Grozdani} . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Characterization of the inclusions in spring steel using light microscopy and scanning electron microscopy
Karakterizacija vklju~kov v vzmetnih jeklih s svetlobno in vrsti~no elektronsko mikroskopijo
A. Bytyqi, N. Puk{i~, M. Jenko, M. Godec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Characterization of the carbides in a Ni-Ti shape-memory alloy wire
Karakterizacija karbidov v @ici zlitine s spominom Ni-Ti
M. Godec, A. Kocijan, M. Jenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Fracture toughness of the vacuum-heat-treated spring steel 51CrV4
Lomna `ilavost vakuumsko toplotno obdelanega vzmetnega jekla 51CrV4
B. Sen~i~, V. Leskov{ek . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Similarity criteria and effect of lubricant inertia at cold rolling
Merila podobnosti in vpliv vztrajnosti maziva pri hladnem valjanju
D. ]ur~ija, F. Vodopivec, I. Mamuzi} . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
IN MEMORIAM
Oskar Kürner 1925–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
ISSN 1580-2949
UDK 669+666+678+53 MTAEC9, 45(1)1–81(2011)
MATER.
TEHNOL.
LETNIK
VOLUME 45
[TEV.
NO. 1
STR.
P. 1–81
LJUBLJANA
SLOVENIJA
JAN.–FEB.
2011

M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
EXPERIMENTAL STUDY OF SOME MASONRY-WALL
COURSEWORK MATERIAL TYPES UNDER
HORIZONTAL LOADS AND THEIR COMPARISON
EKSPERIMENTALNA RAZISKAVA ZGRADBE NEKATERIH
ZIDARSKIH ZIDOV – VODORAVNA OBREMENITEV IN
PRIMERJAVA UPORABLJENIH MATERIALOV
Mehmet Kamanli1, Mahmut Sami Donduren1, Mustafa Tolga Cogurcu1, Mustafa Altin2
1Department of Civil Engineering, Engineering and Architecture Faculty, Selcuk University, 42060 Konya, Turkey
2Technical Science College, Selcuk University, 42060 Konya, Turkey
sdonduren@selcuk.edu.tr
Prejem rokopisa – received: 2010-06-11; sprejem za objavo – accepted for publication: 2011-01-11
In this study, the collapsing loads and shear stress of wall samples made of full blend bricks have been examined under
horizontal loads within their planes using various bricking types and the results have been compared. To this end, samples were
produced using the plain-bricking, lock-bricking and cross-bricking methods. In the performed experimental study, the standard
sliding-resistance test technique for the masonry-wall samples was used as recommended in ASTM 1391-81.The behaviours
observed with the experimental samples and the cracks that occurred at the end of the experiment were examined and the 
(shear stress) graphs and the horizontal load-displacement graphs were drawn and compared with the experiment results
obtained with numerical modelling using the ANSYS program. When these three bricking types were compared, the biggest
collapse load occurred for the cross bricking with 240 kN. The biggest shear stress was also found in the cross bricking, with a
value of 55 000 kN/mm2.
Keywords: materials, full blend brick, experimental, lateral load performance, shear stress
Raziskani sta bili ru{ilna obremenitev in stri`na obremenitev zidov iz `gane opeke pri vodoravni obremenitvi v ravnini
postavitve opek in primerjani so rezultati. Uporabljeni so vzorci, izdelani z enostavno, kro`no in zaprto postavitvijo opek.
Uporabljen je preizkus standardne drsne upornosti vzorcev zidov, priporo~en v ASTM 1391-81. Vedenje vzorcev in razpoke,
nastale na koncu preizkusa, in odvisnosti stri`na napetost () ter vodoravna obremenitev – premik so bile dolo~ene in rezultati
primerjani z rezultati numeri~nega modeliranja z uporabo programa ANSYS. Pri primerjavi treh vrst razporeditve opek je bila
najve~ja ru{ilna obremenitev 240 kN pri kri`ni postavitvi. Pri tej postavitvi opeke je bila izmerjena tudi najve~ja stri`na sila
55 000 kN/mm2.
Klju~ne besede: materiali, `gana opeka, preizkusi, bo~na trdnost, stri`na napetost
1 INTRODUCTION
Human beings have constructed buildings, bridges,
and canals in order to maintain their lives using the
existing possibilities and the technology of the day since
the earliest times. The oldest building types in the history
of humanity are the masonry buildings which are still
being used today. Although reinforced concrete and steel
constructions are the most common building types of
today, masonry buildings are still being constructed in
various countries of the world 1.
Especially in rural areas, people prefer masonry
buildings for economic reasons, they are easy to con-
struct from the local materials and they can construct
buildings in places that are difficult to get at because of
geographic conditions. These buildings which are con-
structed without complying with the existing regulations,
getting engineering services and made of heavy and
rusty materials are rather weak with respect to horizontal
loads 2.
The carrier walls of a masonry building exposed to
horizontal loads have two deficiencies. The first one is
their failure that is not planar and the cracks are seen
here all along the bed joint lines. The second one is
planar failure and generally characterized by a diagonal
shrinking crack. If the failure which is not planar can be
avoided, then the structural resistance is radically
affected by the planar behaviour of the cut wall 3.
The vertical carrier elements of masonry buildings
are the walls. The horizontal loads may affect these
vertical carrier walls due to earthquake, storm, or other
reasons besides the vertical loads. In the case that the
vertical and horizontal loads have an affect together, a
two-axis loading condition occurs on the walls. It is of
significant importance to know the behaviour of the wall,
which is a composite material under two-axis resistance
in terms of providing the loads overlaid on the building
to carry securely 2.
The buildings should have the capacity of flexing and
absorbing energy when they are exposed to an
earthquake impact. This is related to the ductility of the
building. As masonry buildings are rigid and friable, they
have no capacity of flexing and absorbing energy like
flexible buildings. Eventually, they are exposed to major
horizontal forces as their capacity to decrease the energy
Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11 3
UDK 624.04:693:539.4 ISSN 1580-2949
Review article/Pregledni ~lanek MTAEC9, 45(1)3(2011)
exposed during the earthquake is low. Furthermore, they
demonstrate unwanted behaviours in term of engineering
due to the fact that the materials used with masonry
buildings are friable and cause sudden cracks and breaks
when they go beyond their loading limits. Because of
these reasons, masonry buildings cannot be considered
as buildings that are resistant to an earthquake. Never-
theless, masonry buildings should not be completely per-
ceived as negative buildings in terms of an earthquake.
Just like all other buildings, masonry buildings can be
strong and secure provided that they are built in com-
pliance with the standards, regulations and engineering
studies are applied. Masonry houses such as earthen,
brick and stone structures are composed of building
blocks with weak inter-binding action which have low
tension capacity 19.
Before starting the improvement and strengthening
activities for a masonry building, the earthquake beha-
viour of the masonry building exposed to an earthquake
and their collapsing mechanism should be well known.
The earthquake behaviours of masonry building are
diverse although they have similarities in comparison
with reinforced concrete buildings. It can be said that the
most impressive difference is the break of the masonry
wall in the out of plane direction. The cutting forces and
moments are constituted with masonry walls under
horizontal loads. Therewith, the in-plane break of the
masonry wall occurs by axial pressure/tensile forces
formed by the moment and/or the slant inert tension
effects the created moment 2.
Several studies have been carried out related to the
behaviours of masonry buildings. A multi-surface plastic
has been used so as to define the pressure, cut and
resisting behaviours of the filling plaster using an inte-
resting approach developed by Laurence 6,7 and modelled
by an interface element. Gamboratta and coworkers have
generated a model evaluating the constructive parity of
the interface in terms of two internal variables repre-
senting the friction surface and the filling plaster damage
using a similar approach 5. This model has been broken
under the stretching resistance and shown a decreasing
hardness and friction distribution. This feature is used
with the definition of hysteretic planar behaviours of
masonry building walls. These that have experienced
earthquakes have damaged constructions significantly
and caused the deaths of thousands of people 20.
The above-mentioned numerical procedures are quite
effective and take many facts which are seen with the
behaviours of masonry building. However, their appli-
cation is figuratively very difficult. In addition to these, a
few more simple methods are available. For instance, the
homogenization application or anisotropy elasto-plastic
foundation equalities 3, which are some of the appro-
aches 4,8 based upon the crack modelling of masonry
buildings, are much simpler methods.
The horizontal loads occurring during earthquakes
reveal strong planar and non-planar forces on these
walls. The behaviour and the damages to these con-
structions under seismic forces are relatively big 7. The
walls in the direction of the shearing force have a big
role in increasing the seismic force durability 15. The
majority of the existing buildings in this region have not
been designed to withstand earthquakes. Most of them
are typically unreinforced masonry, low-rise buildings
and have been exiguously designed 16. They showed poor
performance during earthquakes and most of the damage
and casualties resulted from these structures. However,
in terms of earthquake engineering, significant lessons
were learned from the surveys of damaged masonry
buildings after earthquakes 17. Damage and losses arising
from landslides can include cracks in the masonry,
damage to the electricity and water supply, subsidence or
at worst the complete collapse of buildings 18.
Masonry building constructions constitute a signi-
ficant part of the construction inheritance in the world.
Actually, the structural walls of these buildings have
been designed so as to resist the force of gravity. The
horizontal loads incurred by earthquakes reveal strong
planar and non-planar forces with these walls. The
damage that occurs in these buildings under seismic
forces are relatively big 2. Figure 1 shows some pictures
of masonry buildings that were damaged after the
Sultandagi-Cay Earthquake, 2002.
The masonry buildings are built with different rows
and bricking forms positioning the bricks end-to-end and
side by side in a net way. In this study, the lock bricking,
cross bricking and plain bricking, which are some of the
most commonly used bricking types in the world, are
compared. The plain bricking is formed by plain rows
which are piled over and over. The second row is started
with a half brick and the vertical sutures are ensured not
to match up with each other (Figure 2).
The lock bricking is formed by placing the rows over
each other by sliding a quarter brick. In order to provide
a quarter-brick sliding, the second row is completed
flatways and two pieces of three quarter brick are
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
4 Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11
Figure 1: Some damaged masonry buildings14 (Sultandagi-Cay Earth-
quake, 2002)
Slika 1: Nekaj zgradb, po{kodovanih v potresu Sultanagi-Cayju, 2002
14
positioned flatways at the end. It is used to brick the
plain and curved walls with a one brick thickness
(Figure 3).
The cross bricking is made using lock and plain rows.
The vertical sutures of the plain rows are skidded by half
a brick from the vertical sutures of the plain row right
under. Hence, the cross images are displayed on the
surface of the wall (Figure 4). It is used for bricking in
one-brick thickness or more thickness walls 5.
In this study a simple numerical modelling is re-
commended with an experimental approach in order to
research the shear behaviours of the horizontal loads of
walls made of full blend bricks within their planes. The
wall samples prepared for this purpose were tested to see
how they are forced to collapse and the shear stress of
the wall samples in the non-linear zones and the change
of the shear rigid of the wall sections were determined.
The numerical modelling was performed using the
ANSYS program with the finite elements taking into
account the local mechanics parameters of the brick and
plaster. The results that were obtained at the end of the
modelling and the results obtained from the experiments
were compared.
2 THE FEATURES OF THE USED MATERIALS
2.1 The features of the full blend brick
Full blend bricks were used in the experiments
(Figure 5). The average pressure resistance of the bricks
is 12 MPa; the elasticity module is 3 000 MPa and the
tensile strength is 0.9 MPa. While calculating the
average pressure resistance, the gross area of the brick is
the area where the pressure force affects and a very neat
distribution is ensured, neglecting the pressure resistance
values of the samples from the ones too high and too
low.
2.2. Features of the plaster
The same type plasters whose volumetric compo-
sition rates have been used in the production of the wall
samples were employed. The cement, sand, water
volumetric composition rate of this plaster has been
produced as 5 : 15 : 3 ratio. The calculated pressure
resistance of the used plaster is 2.26 MPa and the tensile
strength at bending is 1.64 MPa. The plaster thickness
was 3 cm.
3 EXPERIMENTAL STUDY
3.1 Features and production of experimental samples
A total of 9 pieces of samples with 600 mm × 600
mm × 200 mm sizes have been produced, with 3 from
each of the various bricking types. The vertical and
horizontal suture thickness of the samples was about 10
mm. Although the calculation methods in the ASTM
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11 5
Figure 5: Full Blend Brick
Slika 5: @gana opeka
Figure 4: Cross Knitting
Slika 4: Kri`na postavitev opek
Figure 3: Lock Knitting
Slika 3: Zaporna razporeditev opek
Figure 2: Plain Knitting
Slika 2: Enostavna postavitev opek
1391-81 are used 13, the sample size has been selected as
600 mm instead of 1200 mm, which is the sample size
recommended in this standard taking into account the
sizes of the bricks and the laboratory conditions (Figure
6).
All the samples were prepared in the laboratory
under the same conditions. The full blend bricks were
used in the formation of all the samples. The prepared
plaster to brick the samples is also used in the same
proportion. The wall samples were produced using the
very first prepared plaster, according to the bricking rules
(Figure 7). After completion of the wall bricking
process, the plaster mortar was formed using the same
composition rates and the facing process was started;
after realizing the watering process for 7 days, they were
left to dry for 3 days. The samples whose watering and
drying process were completed have been made ready
for the experiment by painting them using whitewash
prepared in the laboratory.
3.2 Experimental technique and program
In this study, carried out for the purpose of finding
the sliding resistance of the wall samples produced with
full blend bricks using various bricking type, the force
controlled experiment technique was used, trying to
maintain the force increment rate. The displacements on
the samples during the experiment were measured using
ten units of Linear Variable Displacement Transducers.
A vertical pressure force was applied to the wall samples
whose horizontal sutures are positioned to make an angle
of 45° by loading, and the breaking loads, breaking
shapes, crack relieves and the displacement readings
with the increasing load stages were recorded.
3.3 Experiment setup and loading
Two pieces of steel heads were made to provide the
samples to stand at 45° angle vertically. A surface made
of plaster of Paris was prepared inside the lower head
standing in its guide and the wall sample was positioned
to the head in its plumb line (Figure 8). Also preparing
the surface made of plaster of Paris in the upper head,
the upper head was placed on the sample. Keeping the
sample in its plane during the loading was ensured by the
metal arms pulled up the front and back of the sample, as
illustrated in Figure 8. In order to decrease the frictional
effect between these metal arms and the sample, the
slipperiness of it was increased, putting oil to the
interfaces.
The vertical load on the experimental samples was
applied by means of a 500 kN capacity manual hydraulic
jack. A steel plate was placed on the steel upper head in
order to adjust the level of the height and a hydraulic
jack was placed right on it. The value of the applied load
was measured using a 500 kN capacity load cell.
3.4 Collection of the experimental data and measuring
system
The load measuring was carried out by using load
cells and the displacements by LVDT displacement
meters in all the experiments. The values that these
devices read were instantly transferred to the computer
by means of a data-logger system called CoDA and
taken under records. All the values read from the
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
6 Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11
Figure 6: Bricking Types
Slika 6: Vrste kri`ne postavitve opek
Figure 8: Loading form of the experimental samples
Slika 8: Na~in obremenitve preizku{ancev
Figure 7: Production of the samples
Slika 7: Izdelava preizku{ancev
channels are recorded to computers immediately and
also all the readings taken from the requested channels
can be monitored in graphics. All the print outs of the
taken readings are in the form to be read by the
"EXCEL" program. The data-collecting system used for
the evaluation of the readings taken from the load cell
and the LVDTs and the computer set up are shown in
Figure 9.
The displacement readings with all the experiments
were carried out using ten units of LVDT (Linear
Variable Displacement Transducer). Two units of
displacement transducers were placed on each surface of
the front and back of the wall samples each, two
displacement transducers on the loading axis and two
units of displacement transducers in the vertical direction
to the loading axis. Furthermore, two units of
displacement transducers were placed on the front side
of the sample in order to control the movement out of
plane. The layout diagram and the illustration of the
displacement transducers are given in Figure 10. Here,
the direction of the arrows shows the approach direction
to the panel consisting of the displacement transducers’
measuring points.
The data read by the displacement transducers
numbered T9 and T10 during the experiments are
constantly controlled and the movement of the sample
out of the plane is provided to be within the certain
limits.
3.5 Experimental work
A total of 9 experiments – 3 from each bricking type
– have been carried out in order to compare some of the
bricking types in terms of horizontal load-bearing
capacity and cutting force. The pictures of the samples
before and after the experiment are given in Figure 11.
The loading was accomplished with approximately 10
kN load steps and the displacement measuring was taken
right after each step.
3.6 Modelling and numerical interpretation
For the finite-element modelling (FEM) of the
masonry prisms, SOLID45 elements in the ANSYS
element library were used for the 3-D modelling of the
solid structures. The element is defined by eight nodes
having three degrees of freedom at each node:
translations in the nodal x, y, and z directions. The
element has plasticity, creep, swelling, stress stiffening,
large deflection, and large strain capabilities. The
element is defined by eight nodes and the orthotropic
material properties. With using this element, the analysis
cannot capture the bending behaviour with a single layer
of elements. In the loading system of these studies there
are no bending effects on the masonry infill and the
SOLID 45 elements are suitable for the analysis 9.
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11 7
Figure 11: The pictures of the plain-, lock- and cross-bricking sam-
ples during the experiment
Slika 11: Posnetki preizku{ancev z enostavno, zaporno in kri`no raz-
poreditvijo opek med preizkusom
Figure 10: The layout diagram of the displacement transducers
Slika 10: Razporeditev transduktorjev premikov
Figure 9: Data-collecting system
Slika 9: Sistem za zbiranje podatkov
The masonry units and mortar layers are modelled
independently and different material properties are
obtained. In the elaborate FE model developed in this
study, masonry units are linked to the mortar units by a
series of nonlinear springs, which is also available in the
ANSYS library. COMBIN39 is a unidirectional element
with nonlinear generalized force-deflection capability
that can be used in every analysis. The longitudinal
option is a uniaxial tension-compression element with up
to three degrees of freedom at each node: translations in
the nodal x, y, and z directions. No bending or torsion is
considered. The element has a large displacement
capability for which there can be two or three degrees of
freedom at each node.10 The element is defined by two
(preferably coincident) node points and a generalized
force-deflection curve. The springs are introduced to
handle the tensile and shear stress failure in the mortar
joints, Figure 12.
The material model for the masonry panel was
assumed to be orthotropic parallel and normal to the bed
joints.11 The material stress versus strain relationship is
represented in Figure 13.
The incremental full Newton-Raphson iterative
solution procedure was used in order to account for both
large deformation effects and the material plasticity. In
order to capture the complete load-deflection behaviour
including the post peak response, the top node of the
masonry prism was subjected to a vertical downward
displacement.12
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
8 Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11
Figure 16: Von-Misses Stress distribution for specimen type 2
Slika 16: Von-Misesova razporeditev za vzorec 2
Figure 13: Stress versus strain relationship
Slika 13: Odvisnost napetost – deformacija
Figure 12: Mortar-brick joint interface modelling
Slika 12: Modeliranje povezave med opeko in malto
Figure 15: Principal Stress distribution for specimen type 1
Slika 15: Razporeditev glavnih napetosti za vzorec tipa 1
Figure 14: Von-Misses Stress distribution for specimen type 1
Slika 14: Von-Misesova razporeditev za vzorec 1
The Von-Misses stress distribution and principal
stress graphics for the loaded masonry panels are repre-
sented in Figure 14 through 19. The applied maximum
displacement level is chosen for comparison purposes.
(Type 1: Plain Knitting, Type 2: Cross Knitting, Type 3:
Lock Knitting).
3.7 Behaviours of the experimental samples and the
experimental results
The first crack in the sample reached to the biggest
collapsing load with the plain bricking samples occurred
at around 40 kN. The first cracks started from the top-left
side of the sample and continued downwards. The
maximum collapsing load was monitored to be 60 kN
and a break on the top-right of the sample is seen.
The first crack in the sample carried the biggest
collapsing load with the lock-bricking samples occurred
at 100 kN. This crack is formed on the right-bottom side
of the sample. The loading was continued and the sample
started not to carry the load at 160 kN and broke at the
right-top side.
The first crack in the sample where the maximum
collapsing load is obtained with the cross-bricking
samples occurred on the right-top side of the sample at
170 kN. The collapsing load was determined to be 240
kN. The sample was broken from the right-top side.
The horizontal load-displacement graphics of the
plain-, lock- and cross-bricking samples is given in
Figure 20.
3.7.1 The comparisons among the bricking types
a) Comparison in terms of horizontal load-displacement
The samples reaching to the biggest collapsing load
out of the plain-, lock- and cross-bricking samples were
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11 9
Figure 19: Principal Stress distribution for specimen type 3
Slika 19: Razporeditev glavnih napetosti za vzorec tipa 3
Figure 18: Von-Misses Stress distribution for specimen type 3
Slika 18: Von-Misesova razporeditev za vzorec 3
Figure 17: Principal Stress distribution for specimen type 2
Slika 17: Razporeditev glavnih napetosti za vzorec tipa 2
Figure 20: The horizontal load-displacement graphics of the plain-,
lock- and cross-knitting samples
Slika 20: Odvisnost vodoravna obremenitev – premik za enostavno,
zaporno in kri`no razporeditvijo opek
compared in terms of the horizontal load-displacement.
The comparisons of the horizontal load-displacement
graphics obtained from the biggest collapsing load in 3
bricking types are given in Figure 21.
b) Comparison of the shear resistance-displacement
While the shear force is determined from the plain-,
lock- and cross-bricking samples, the displacement
values have been dictated finding out the average of the
T1-T3 LVDT reading values and the T2-T4 LVDT
values. The sliding resistance was obtained by taking the
Cos45 value of the load and dividing it by the area. The
shear resistance-displacement graphics obtained in this
way are given in Figure 22.
c) Comparison related to the Numerical Modelling
The magnitude of this displacement was sufficiently
greater than that observed in the actual test, after which a
load-deflection plateau was attained indicating that the
contribution of the infill almost entirely diminished and
that no appreciable increase in the load resistance
occurred. The three different masonry configurations are
modelled and loaded in time steps, and the applied
vertical load versus displacement values are obtained.
The comparison of the analytical and experimental data
is represented in Figure 23, 24 and 25.
4 DISCUSSION AND CONCLUSION
The total of 9 units of samples made up of three
different bricking types to be used in the formation of a
masonry building were compared in terms of the
horizontal load and shear force using a standard sliding
resistance experiment technique for the masonry-wall
samples recommended in ASTM 1391-81, and the
following outcomes have been obtained. The analytical
results obtained by the solution with the finite-element
method together with the results obtained as a result of
the experiments were evaluated. The results are as
follows:
The biggest collapsing load occurred with the cross
bricking as per the horizontal loads. The collapsing load
with this type of bricking is 240 kN. The collapsing load
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
10 Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11
Figure 21: The comparisons of the horizontal load-displacement
graphics
Slika 21: Primerjava najve~je ru{ilne sile za 3 vrste postavitve opek
Figure 25: Load versus displacement comparison for specimen type 3
Slika 25: Primerjava breme – premik za vzorec tipa 3
Figure 23: Load versus displacement comparison for specimen type 1
Slika 23: Primerjava breme – premik za vzorec tipa 1
Figure 22: The shear resistance-displacement graphics
Slika 22: Odvisnosti stri`na sila – premik
Figure 24: Load versus displacement comparison for specimen type 2
Slika 24: Primerjava breme – premik za vzorec tipa 2
with the lock bricking is 160 kN, and the plain bricking
is found to be 60 kN.
It was determined that the strength of the shear stress
formed in cross knitting is 2.25 times higher than lock
knitting and 3.6 times more than plain knitting.
When horizontal load displacement graphs are taken
into consideration, it was found that the energy-absorp-
tion capacity is 3 410 kN mm in cross knitting, 2 420 kN
mm in lock knitting and 625 kN mm in plain knitting. It
was concluded that cross knitting consumed 1.4 times
more energy than lock knitting, whereas it consumed 5.8
times more energy than plain knitting.
According to the horizontal load-displacement graph,
the displacement ductility of the cross-knitting wall was
19 mm, that of the lock-knitting wall was 23 mm and
that of the plain-knitting wall was 14 mm.
The values obtained analytically by using the finite-
elements method (ANSYS) were in accordance with the
ones obtained as a result of the experiments.
5 CONCLUSIONS
When the results of both the empirical and analytical
studies on the samples formed by using different knitting
types were evaluated, a masonry construction with cross
knitting is considered to be the most suitable knitting
type in terms of collapse load, shear stress, and energy-
absorption capacity. Although the collapse load of lock
knitting is low, its displacement ductility is higher than
other knitting types. It was observed that plain knitting
has the lowest values among the other knitting types.
These results show that different knitting types used in
the masonry constructions of rural regions have an
important role in the protection of the construction
against earthquakes. Using a cross-knitting type in
masonry constructions in rural regions that are especially
at risk of earthquakes will make the construction more
resilient to earthquake damage.
6 REFERENCES
1 E. Yokel, S. G. Fattal, Failure hypothesis for masonry shear walls
Journal of the Structural Division, ASCE, 102 (1976) ST3, 515–532
2 R. Kanit, E. Atimtay, et al., The Experimental behavior of the
masonry walls which are loaded out of plane, YDGA The Increase in
the earthquake security of the Masonry Buildings Workshop, ODTÜ,
17th February, 2005, Ankara
3 J. Lopez, S. Oller, E. Onate, J. Lubliner, Homogenous constitution
model for masonry Int. J. Numer Meth Eng., 41 (1999), 1651–71
4 J. Lee, G. Pande, J. Midleton, B. Kralj, Numerical modelling of brick
masonry panels subject to lateral loadings, Comput. Struct., 61
(1996) 4, 735–45
5 L. Gambarotta S. Lagomarsino, Damage models for the seismic re-
sponse of brick masonry walls. Part I: the mortar joint model and its
applications. Earthquake Eng Struct Dynamics, 26 (1997), 423–39
6 P. Laurence, J. Rots, Multisurface interface model for analysis of
masonry structures. J. Eng. Mech., 123 (1997) 7, 660–8
7 P. B. Laurence, Computational strategles for masonry structures. Ph.
D. Thesis, The Netherlands :Technical Universitiy Delft, Delfth
University Pres: ISBN 90-407-1221-2 (1996)
8 R. Luciano, E. Sacco, Homogenization technique and damage model
for old masonry model Int. Solid Struct., 34 (1997) 24, 3191–208
9 A. Gabor, A. Bennani, E. Jacquelin, F. Lebon, Modelling aappro-
aches of the in-plane shear behaviour of unreinforced and FRP
strengthened masonry panels. Composite Structures, (2005),
277–288
10 D. Abruzzese, L. Miccoli, J. Yuan, Mechanical behavior of leaning
masonry, Huzhu Pagado Journal of Cultural Heritage, (2009),
480–486
11 M. Fathy, J. Planas, M. Sancho, A numarical study of masonry cra-
caks, Engineering Failure Analysis, (2008), 675–689
12 V. Mallardo, R. Malvezzi, E. Milani, G. Milani, Seismic vulnera-
bility of historical masonry buildings: A case study in Ferrara,
Engineering Structures, (2007), 2223–2241
13 ASTM 1391-81 Standard test method for diagonal tension (shear) in
masonry Assem-Blages
14 M. Y. Kaltakci, M. H. Arslan, H. H. Korkmaz, M. Kamanli, Failure
of superstructures after the February 03, 2002 Sultandagi-Cay
Earthquake, Journal of Engineering and Applied Sciences, 2 (2007)
10, 1569–1575
15 A. Gabor, E .Ferrier, E. Jacquelin, P. Hamelin, Modelling approaches
of the in-plane shear behaviour of unreinforced and FRP strength-
ened masonry panels, Composite structures, (2005), 277–288
16 K. Korkmaz, Seismic safety assessment of unreinforced masonry
low-rise buildings in Pakistan and its neighbourhood, Nat. Hazards
Earth Syst. Sci., 9 (2009), 1021–1031
17 H. Ramazi, H. S. Jigheh, The Bam (Iran) earthquake of December
26, 2003: From an engineering and seismological point of view, J.
Asian Earth Sci., 27 (2006), 576–584
18 A. Blöchl B. Braun, Economic assessment of landslide risks in the
Swabian Alb, Germany – research framework and first results of
homeowners and experts surveys, Natural Hazards and Earth System
Sciences, 5 (2005), 389–396
19 H. H. Korkmaz, S. Z. Korkmaz, M. S. Donduren, Earthquake hazard
and damage on traditional rural structures in Turkey, Natural Hazards
and Earth System Sciences, 10 (2010), 605–622
20 M. Kamanli, F. S. Balik, The behaviour of roof gable walls under the
effect of earthquake load, Natural Hazards and Earth System Scien-
ces, 10 (2010), 1–13
M. KAMANLI et al.: EXPERIMENTAL STUDY OF SOME MASONRY-WALL COURSEWORK ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 3–11 11

I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
SYNTHESIS OF ALUMINIUM FOAMS BY THE
POWDER-METALLURGY PROCESS: COMPACTING OF
PRECURSORS
SINTEZA ALUMINIJEVIH PEN PO POSTOPKU METALURGIJE
PRAHOV: STISKANJE PREKURZORJEV
Irena Paulin1,2, Borivoj [u{tar{i~2, Varu`an Kevorkijan3, Sre~o D. [kapin4,
Monika Jenko2
1TALUM, d. d., Kidri~evo, Tovarni{ka cesta 10, SI-2325 Kidri~evo, Slovenia
2Institute of Metals and Technology, Lepi pot 11, SI-1000 Ljubljana, Slovenia
3Independent Researcher, Betnavska cesta 6, SI-2000 Maribor, Slovenia
4Jozef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia
irena.paulin@imt.si
Prejem rokopisa – received: 2010-12-20; sprejem za objavo – accepted for publication: 2011-01-19
Aluminium foams, produced by the powder-metallurgy route, have a good potential for use in weight-sensitive structural parts.
The goal of this study was to evaluate the properties and to optimize the preparation of precursors by a powder-compacting
process. Various compacting pressures, from 200 MPa to 900 MPa, were used in the double-axial powder-compacting process
for two different aluminium powders: pure aluminium and an AlSi12 alloy with the addition of 1 % of TiH2 as a foaming agent.
The green density of the precursors and the distribution of the foaming agent were examined. The powder particles was also
characterised. The results of the effective preparation of precursors are shown as the effectiveness of the foaming of the
precursors.
The relation between the powder characteristics, the aluminium-alloy properties and the preparation of precursors was studied
by SEM/EDS analysis, powder-metallurgy standard testing of metallic powders, granulometry, etc. Different parameters were
used for the precursor preparation and foaming. The foaming temperature varied between 680 °C and 770 °C, and the foaming
time was from 6 min to 13 min. The relation between the properties and the applied production parameters was studied in detail
and is described in this paper.
Keywords: aluminium foam, Al, AlSi12 alloy, powder-metallurgy, powder compacting, sintering
Aluminijeve pene, narejene po postopku metalurgije prahov, imajo velik potencial v uporabi lahkih konstrukcij. Cilj raziskave je
ugotoviti lastnosti in parametre za optimizacijo priprave prekurzorja po postopku hladnega stiskanja. Za pripravo prekurzorjev
smo uporabili obojestransko stiskanje s tlaki od 200 MPa do 900 MPa. Stiskali smo me{anico prahov tehni~no ~istega aluminija
99,7 % in zlitine AlSi12, v obeh primerih z dodatkom 1 % TiH2 kot penila. Dolo~ili smo zelene gostote in porazdelitev penila v
prekurzorjih. Naredili smo tudi karakterizacijo vseh treh uporabljenih prahov kot vstopnega materiala. Rezultat uspe{ne priprave
prekurzorjev se ka`e v uspe{nosti penjenja materiala.
Povezavo med lastnostmi prahov, lastnostmi aluminijevih zlitin in priprave prekurzorjev smo raziskovali s SEM/EDS-analizami,
s preizku{anjem prahov s standardnimi metodami, uveljavljenimi v metalurgiji prahov, granulometrijo itd. Uporabljeni so bili
razli~ni parametri za pripravo in penjenje prekurzorjev. Temperature penjenja so bile med 680 °C in 770 °C, ~as penjenja pa
med 6 min in 13 min. Povezava med lastnostmi in uporabljenimi parametri za pripravo aluminijevih pen je bila raziskana in je
podrobno opisana v tem ~lanku.
Klju~ne besede: aluminijeve pene, aluminij, zlitina AlSi12, metalurgija prahov, stiskanje prahov, sintranje
1 INTRODUCTION
Aluminium foams are metallic materials with various
physical and chemical properties that can be used for
several purposes1,2. Production methods can be classified
into four groups: powder-metallurgy (PM), molten-metal
foaming, metallic deposition, and sputter deposition3.
Each production method gives its own characteristic
range of densities, cell sizes and shapes. In this study a
method that is adapted to produce complex shapes of
metallic foams is discussed, i.e., the powder-metallurgy
process.
PM is one of the possible techniques to produce
metallic foams. This production process is not as widely
used as the less-expensive, molten-metal foaming pro-
cess, but it also has advantages 4,5. Its general application
in the metal-foaming industry is not very extensive due
to the relatively high costs of the input materials. The
principle of PM is simple and the process consists of
three stages: 1) mixing the metallic powder and the
powder of the foaming agent, usually TiH2, ZrH2, etc; 2)
compacting the powder mixture and; 3) sintering as the
final stage of the process. All three steps are important
for the quality of the final production and the properties
of the aluminium-foam products.
The first step of the PM process is mixing of the
metallic powder and the powder of the foaming agent in
the proportions needed for the final properties of the
product. The higher is the fraction of foaming agent, the
greater and/or more numerous are the pores. The addi-
tion of the powder ceramic compounds (SiC, CaO) is
desired for the stabilization of the foams, but this is not
essential 6–8. The important task of mixing is to achieve a
Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19 13
UDK 66.017:669.715:621.746 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)13(2011)
homogeneous distribution of foaming-agent particles and
of the ceramic-powder particles if they are added.
The next important step of the foaming process is the
compacting of the powder mixture. The density of the
compacted material, the so-called green or theoretical
density, is a significant property of precursors and it
must be very high (up to 99 %). The shape and size of
the metallic-powder particles plays an important role in
achieving the highest possible green density of the pre-
cursors. However, depending on those properties of the
powders, various angles of the repose and bulk densities
are achieved. The green density of the precursors also
depends on the possible plastic deformation of the
metallic powder particles. The lower is the porosity of
the precursors, the more of the liberated gas during the
sintering is captured in the matrix material.
Sintering is the final stage of the PM production
process. During the sintering process at a temperature
that is above the temperature of the melting point, and
which is specific for each alloy and depends on the
foaming agent used, hydrogen is released from the
hydride and forms pores in the material. A larger fraction
of liberated gases is captured in the matrix material if the
molten metal has a higher viscosity 9–11.
The sintering time varies from 3 min up to 10 min,
depending on the alloy used for foaming, the foaming
agent and the size of the sintered piece. In the decom-
position of the TiH2 used as a foaming agent, the Ti
remains and hardens the matrix material while the
hydrogen captured in the matrix forms foams with
closed-cell structures.
In this study we have focused on the second stage of
the PM process. The production of precursors by com-
pacting powder mixtures can be performed in a variety
of ways, e.g., by uni-axial, double-axial or isostatic
pressing, extrusion, rolling, etc., and all the methods can
be hot or cold. Furthermore, the compacting process can
be performed in an inert atmosphere, in air or in vacuum
12. The most economical way is double-axial pressing in
air, but the most efficient one is high-temperature
extrusion.
2 EXPERIMENTAL WORK
The PM foaming process was applied in our research
work. Air-atomized Al powder with a purity of 99.7 %
and a D50 of 98 μm, air-atomized AlSi12 alloy powder
with D50 of 42 μm, and TiH2 powder with a purity of 98
% and a D50 of 12.6 μm were used as the starting mate-
rials (the purity was specified by manufacturer). The
metallic powders were prepared as Ecka granules (Non
Ferrum Kranj, d. o. o.), while the TiH2 powder (325
mesh, 98 %) was purchased from the Sigma-Aldrich
Company. In all cases, the metallic powder was mixed
with mass fraction 1 % of TiH2 powder in a turbular
mixer (TURBULA WAB Type T2C, 50 Hz, 180 W) for
an hour. The characteristics of the powders are shown in
Table 1. The size and the distribution of the powders
were determined by laser granulometry (Alcatel CILAS
HR 850-B, isopropanol as analysis medium).
The bulk density, tapped density and free-flowability
of the powders were determined. The free-flowability
and the bulk density of the powders were determined by
the Hall flowmeter funnel that contains a standard flow-
meter funnel for metallic powders 13. A dry test specimen
must be carefully loaded into the flowmeter funnel and
permitted to run into the density cup through the dis-
charge orifice. In order to determine the free-flowability
of the powders it is necessary to measure the time
needed for 50 g of powder to pass through the orifice.
When the bulk density is determined the powder must
completely fill the density cup till it starts to overflow,
then the powder is levelled with the top of the cup using
a spatula with a blade held perpendicular to the top of
cup. Afterwards, the powder is weighed. For a deter-
mination of the powder’s tapped density a standard
mechanical device is used that enables tapping of the
graduated cylinder containing powder at a rate of 100
taps per minute 13. The analyzed powder must be
weighed (50 g) and the volume of the powder is read
after the tapping.
Powder mixtures were compacted cold, using a
double-axial compaction, in a 24-mm-diameter, lubri-
cated, tool-steel die with pressures in the range from 200
MPa to 900 MPa to achieve different green densities 14.
The samples were compacted in INSTRON 1255 equip-
ment with an INSTRON 8800 computer system and the
Bluehill2 program. Different pressures were used for
both metallic powder mixtures and the green densities
were calculated by assuming that the density of the base
metal is 2.7 kg/dm3 when aluminium is used, and 2.65
kg/dm3 for the AlSi12 alloy. The samples were later
sintered at a pre-determined temperature in the air. After
various times of sintering, samples were taken out of the
retort furnace and quenched into water to solidify the
foam structures.
The sintered samples were characterized with a light
microscope (LM, Nikon microphot – FXA) and analyzed
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
14 Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19
Table 1: Characteristics of the used powders
Preglednica 1: Lastnosti uporabljenih prahov
Powders Manufacturer Mean particle size d/μm Purity w/% Method of manufacturing
Al Non ferrum, Kranj 106 99.7 Air atomization
AlSi12 Non ferrum, Kranj 47 compositionw(Si) = 11.29 % Air atomization
TiH2 Sigma-Aldrich 14 98 Ball milling
with the AnalySIS PRO 3.1 program. The size and the
distribution of the pores were determined by a standard
metallographic method for determining the size and the
distribution of the grains in the microstructures 15.
Metallographic samples of the precursors and the foam
samples were prepared using a standard metallographic
procedure, examined with scanning electron microscope
(FE-SEM JEOL JSM-6500F) and analyzed by energy-
dispersive electron spectroscopy (EDS, INCA X-SGHT
LN2 detector, INCA ENERGY 450 analyzing program)
for the elemental analyses. SEM images of the powders
were also made.
3 RESULTS AND DISCUSSION
Metallic powders produced in a gas atomizer, in our
case in air, have an oblong, quasi-spherical shape with
varying size distributions. Figure 1 shows SEM images
of the used powders; (a) Al 99.7 %, (b) AlSi12 and (c)
TiH2. The size and he characteristics of the powders are
shown in Table 1. The aluminium powder particles
(Figure 1 (a)) have a mean size of 106 μm, while the
AlSi12 powder particles (Figure 1 (b)) have a mean size
of 47 μm, which is half the size of the Al powder
particles. However, the morphology of the powders is
typical for gas-atomized powders from a liquid metallic
melt, and it can be described as a bulk prolonged,
semi-spherical shape. On the other hand, the morphology
of the TiH2 powder particles is quite different from that
of the metallic powders (Figure 1 (c)). The TiH2 powder
was prepared by the ball milling of titanium in a
hydrogen atmosphere under pressure 16–18 and it exhibits
an irregular, sharp, polygonal morphology. The mean
particle size was 14 μm, significantly smaller than that of
the matrix-material powder particles 19.
Table 2: Influence of compacting pressures on the green densities of
precursors
Preglednica 2: Vpliv pritiska stiskanja na zeleno gostoto prekurzorjev
Green density of precursor
Compacting
pressure
p/MPa
w(Al + 1 %
TiH2)/%
w(AlSi12 + 1 %
TiH2)/%
200 94.8 75.6
300 97.9 82.4
400 98.1 86.1
500 98.5 89.9
700 98.8 93.2
750 / 94.8
800 / 95.8
900 / 96.2
The surface of all the powder particles was oxidized.
The thickness of the oxide layer depended on the affinity
of the basic material for oxygen, which was very high for
both the aluminium and TiH2 20–24. This oxide layer was
examined in our previous work 21,23 and found to be one
of the influential parameters for the formation of metallic
foams by the powder-metallurgy process. The oxide
layer on TiH2 powders caused a delayed gas evolution,
while the oxide layer on the aluminium powder particles
increased the viscosity of the melt, which had a favour-
able influence on the foaming process 11,25.
The free flowability of the Al powder particles was
75 s per 50 g of powder, while the AlSi12 and TiH2
particles did not flow freely, since they did not have good
flowing properties. The bulk density (cumulative volume
fraction as a function of the particle size of powder/Al)
of the Al powder was 1.23 kg/dm3, and 1.20 kg/dm3 for
the AlSi12 powder. The TiH2 powder had a bulk density
of 1.34 kg/dm3. Both powders, the Al and of the AlSi12,
had a tapped density of 1.43 kg/dm3, while the TiH2
powder had a tapped density of 1.61 kg/dm3. There was
no major difference in the densities of both metallic
powders, but there was a difference in the flow properties
of the two. Though produced in the same way, with simi-
lar tapped densities, the two metallic powders had diffe-
rent final green densities after the compacting process.
The characteristics of the TiH2 powder did not influence
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19 15
Figure 2: Green density of the compacted precursors
Slika 2: Zelene gostote stisnjenih prekurzorjev
Figure 1: SEM images of as-received powders at the same magni-
fication; (a) Al powder of 99.7 % purity, (b) powder of the AlSi12
alloy, and (c) TiH2 powder
Slika 1: SEM-slike prahov, narejene pri enaki pove~avi; (a) Al prah s
~istoto 99,7 %, (b) prah zlitine AlSi12 in (c) prah TiH2
the properties of the precursors, since a relatively small
amount of TiH2 powder was used in the mixture, but its
properties did have an influence on the distribution of the
TiH2 powder in the metallic powder matrix. In any case,
it was preferable to have a homogenous distribution of
TiH2 particles in the precursor material.
After one hour of mixing the powder mixtures were
compacted with various pressures. The influence of the
compaction pressures on the green densities of the pre-
cursors is presented in Table 2 and in Figure 2. The
precursors were examined by SEM and analyzed by
EDS. The aluminium precursor that was compacted by
the double-axial process with a pressure of 400 MPa is
shown in Figure 3 (a). The porosity of the precursor was
approximately 2 %. TiH2, as the foaming agent, was
homogenously distributed in the matrix material. On the
other hand, the powder of the AlSi12 alloy was also
double-axially compacted, with pressures in the range
from 200 MPa to 900 MPa. With a pressure of 400 MPa
the porosity of the precursors was approximately 14 %.
When the compacting pressure was raised up to 900
MPa, the porosity was reduced to only 4 %. Different
theoretical densities of the precursors depended on the
different qualities of matrix powders, e.g., on the size
and the shape of the particles, the bulk and the tapped
density. Though the matrix powders had practically the
same shape, but different sizes, the compressibility of the
powders differed. Compressibility also depended on the
bulk and tapped densities of the powders. Double-axial
compacting of both, the Al and the AlSi12 powders, at
room temperature did not enable us to achieve the same
green density through using higher pressures than for the
alumosilicates. The limitation of the used compacting
pressures was also related to the lubricated tool-steel die
that started to stick due to the too high forces in com-
pacting aluminium, while the AlSi12 alloy powder was
harder and had a higher compacting resistance. In other
words, a larger green density could be achieved with
materials that had a better plastic deformability. Though
high compacting pressures were applied, the porosity,
especially in the AlSi12 precursors, still remained
relatively high.
The EDS analysis determined the distribution of Si
and Ti in the matrix material. The Si in the AlSi12
matrix was distributed uniformly throughout the whole
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
16 Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19
Figure 4: Cross-section of pure Al foams; (a) cold pressed with 200
MPa, sintered at 750 °C for 7 min, (b) cold pressed with 400 MPa,
sintered at 750 °C for 6 min 30 s.
Slika 4: Prerez Al-pene iz tehni~no ~istega aluminija; (a) hladno obo-
jestransko stisnjeno z 200 MPa, sintrano pri 750 °C, 7 min, (b) hladno
obojestransko stisnjeno s 400 MPa, sintrano pri 750 °C, 6 min 30 s.
Figure 3: Precursors; double-axially compacted with pressure of 400
MPa, (a) Al precursor of 99.7 % pure Al, determined density was
approximately 98.1 %, (b) AlSi12 precursor, determined density was
approximately 86.1 %. BE image – bright spots are titanium particles
and dark spots the porosity.
Slika 3: Prekurzorji, obojestransko stisnjeni s 400 MPa, (a) Al pre-
kurzor iz Al prahu ~istote 99,7 % z zeleno gostoto 98,1 %, (b) AlSi12
prekurzor z zeleno gostoto 86,1 %. BE slike – svetli predeli so titanovi
del~ki, temni predeli so poroznost.
Figure 5: Histogram of pore-size distribution of pure Al foams
Slika 5: Histogram velikosti in porazdelitve por v Al-penah
matrix material, which proved that the AlSi12 powder
was manufactured from the AlSi12 aluminium alloy and
was not a mixture of Al and Si powder in the demanded
ratio. The titanium was randomly but uniformly distribu-
ted as individual particles that could be seen in Figure 3
(b) as bright phases in the BE image. On the other hand,
only titanium particles were detected in the aluminium
matrix, and they were similarly distributed as individual
particles in the matrix material, as in the sample of the
AlSi12 alloy (Figure 3 (a)). The distribution of TiH2,
with the captured liberated gases around the particles of
foaming agent, enabled a homogeneous distribution of
pores in the aluminium foams.
The precursors were heat treated in a retort furnace at
temperatures that were slightly higher than the melting
temperature of the matrix material. All the samples were
sintered at temperatures between 680 °C and 770 °C, for
various sintering times. The temperatures of the retort in
the furnace were controlled with CrNi thermocouple.
After sintering at a pre-determined temperature and for a
chosen time, the samples were taken out of the retort and
quenched in air and/or water. The densities of the sam-
ples were then determined with the Archimedes’ method.
The samples of pure aluminium that were compacted
at lower pressures had fewer pores, and those were not
spherical, while the samples that were pressed at 400
MPa or higher pressures contained more pores that were
semi-spherical. Cross-sections of the obtained alumi-
nium foams are shown in Figure 4.
Light microscopy was employed to determined the
size and the distribution of the pores. A diagram of the
pore-size distribution in our samples is presented in
Figure 5. The same comparison was also made with
samples of the AlSi12 alloy that were pressed with
various forces and sintered at various temperatures and
for various times. The cross-sections of the foams are
shown in Figure 6 and a diagram of the pore size distri-
bution is shown in Figure 7.
The Al precursors were foamed at 750 °C from 6 min
to 10 min. A histogram of the Al foampore size distri-
bution (Figure 5) gave evidence that the majority of the
pore sizes varied from 1 mm to 3 mm, giving an average
pore size of 2.6 mm. Only a smaller amount of pores
larger than 5 mm was found. On the other hand, the
AlSi12 foams were sintered at 770 °C and for times from
6 min to 13 min. A histogram of the pore-size distri-
bution is presented in Figure 7. The distribution was
very scattered and there was a larger amount of pores
with sizes of more than 10 mm.
The difference between the size and the distribution
of pores in both foamed materials was due to the green
densities of the precursor materials. The compacted
material with the higher green density was more easily
and uniformly foamed. Therefore, the foaming effective-
ness of the AlSi12 precursors was relatively low due to
their low green densities (only up to 96 % in our case).
The pores in this material were larger due to the
prolonged foaming times that were needed to foam the
material. This process causes further coarsening of the
pores (the large pores were growing at the expense of the
small ones).
The density of the foamed material, determined by
the Archimedes’ method, was from 0.58 kg/dm3 to 1.2
kg/dm3 for pure Al foams and from 0.9 kg/dm3 to 1.6
kg/dm3 for the AlSi12 foams. The reason for the
difference in the densities was in foamability, which
depends on the green density of the precursors.
Another difference in the foaming of both aluminium
foams was caused by the Si. The AlSi12 alloy is a
so-called casting alloy, and according to the Al-Si phase
diagram, it is lying near the eutectic point, thus its
melting-cooling properties are different from those of
pure aluminium. The area between the liquidus and
solidus lines is much larger for pure aluminium than for
an alloy of the near-to-eutectic composition. Therefore,
the temperature range between the liquidus and solidus
state in which the pores can be formed is higher with
pure aluminium.
4 CONCLUSIONS
It is very important for better foaming results to
achieve a green density of at least 98 %. This can be
easily achieved with pure aluminium powder. On the
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19 17
Figure 7: Histogram of pore-size distribution of the AlSi12 foam
Slika 7: Histogram velikosti in porazdelitve por v penah zlitine
AlSi12
Figure 6: Cross-section of AlSi12 foam, cold pressed with 900 MPa,
sintered at 770 °C for 9 min
Slika 6: Prerez pene zlitine AlSi12, hladno obojestransko stisnjeno z
900 MPa, sintrano pri 770 °C, 9 min
other hand, it is practically impossible to achieve green
densities higher than 96 % with AlSi12 powder using
only a double-axial compression at pressures even up to
900 MPa at room temperature. For industrial applica-
tions it is also important to use as simple industrial
machines as possible or to use techniques that are
already used in the PM industry for the preparation of
different types of complex products. Higher green
densities can be achieved by increasing the compacting
pressure and/or pre-heating the precursors to increase
their compressibility 26,27.
The difference in size of our powder particles does no
play such an important role in the difference in the green
densities of both precursors, Al and AlSi12, though the
bulk density is slightly higher with larger aluminium
powder particles. The tapped density is approximately
the same for both powders. The most important para-
meter for achieving higher green densities is the com-
pressibility of powder particles, which depends on the
properties of the metallic alloy.
A very important parameter in the foaming process is
the foaming temperature. In our study, a temperature of
750 °C was chosen as the most suitable temperature for
pure aluminium and 770 °C for the AlSi12 alloy. The
results of our study and from the references confirm that
the foaming temperature differs from the melting
temperature of the matrix material and it cannot be
correlated directly from the relation with the melting
point of the material. However, while there is no direct
correlation between the foaming and the melting tempe-
rature, so the foaming temperature should be determined
for each matrix material separately, depending on the
applied foaming agent. Also, the foaming time has to be
chosen in a similar way. It should be determined separa-
tely for each alloy and for each foaming temperature.
The most efficient aluminium alloys used for foam-
ing by the PM route are the alloys with higher contents
of aluminium and technically pure Al powder. Therefore,
so-called aluminium wrought alloys that have a lower
content of impurities and of alloying elements (usually
less than w = 1 %) are more often used for foaming than
the casting alloys that have higher amounts of alloying
elements. The reason for the selection of wrought alloys
is also due to the fact that according to the Al-Si phase
diagram the alloys closer to the aluminium corner of the
diagram have a higher temperature interval between the
liquidus and the solidus temperature in which pores can
be formed. In any case, the viscosity of wrought alloys is
higher than that of casting alloys, and therefore pores are
formed due to the capture of liberated gases from the
foaming agents.
Acknowledgements
This work has been supported by the "Physics and
Chemistry of Porous aluminium for Al panels capable of
highly efficiently energy absorption" project L2-2410
(D), funded by the Slovenian Research Agency. The
views expressed in this publication reflect only the views
of the authors and the SR Agency is not liable for any
use that may be made of the information contained
therein.
5 REFERENCES
1 J. Banhart, Manufacture, characterisation and application of cellular
metals and metal foams, Progress in Materials Science, 46 (2001),
559–632
2 F. Simancik, J. Jerz, J. Kovacik, P. Minar, Aluminium foam – A new
light-weight structural material, Kovove Materialy-Metallic
Materials, 35 (1997), 265–277
3 G. J. Davies, S. Zhen, Review Metallic foams: their production,
properties and applications, Journal of Materials Science, 18 (1983),
1899–1911
4 C. J. Yu, H. H. Eifert, J. Banhart, J. Baumeister, Metal foaming by a
powder metallurgy method: production, properties and applications,
Materials Research Innovations, 2 (1998), 181–188
5 J. Banhart, H. W. Seeliger, Aluminium foam sandwich panels:
Manufacture, Metallurgy and Applications, 10 (2008), 793–802
6 M. Haesche, J. Weise, F. Garcia-Moreno, J. Banhart, Influence of
particle additions on the foaming behaviour of AlSi11/TiH2
composites made by semi-solid processing, Materials Science and
Engineering a-Structural Materials Properties Microstructure and
Processing, 480 (2008), 283–288
7 M. Guden, S. Yuksel, SiC-particulate aluminum composite foams
produced from powder compacts: foaming and compression
behavior, Journal of Materials Science, 41 (2006), 4075–4084
8 S. Esmaeelzadeh, A. Simchi, Foamability and compressive properties
of AlSi7-3 % SiC-0.5 % TiH2 powder compact, Materials Letters, 62
(2008), 1561–1564
9 R. E. Raj, B. S. S. Daniel, Manufacturing challenges in obtaining
tailor-made closed-cell structures in metallic foams, Springer
London Ltd, (2008), 605–612
10 R. Surace, L. A. C. De Filippis, A. D. Ludovico, G. Boghetich,
Experimental analysis of the effect of control factors on aluminium
foam produced by powder metallutgy, Proc. Estonian Acad. Sci.
Eng., 13 (2007), 156–167
11 Z.-L. Song, L.-Q. Ma, Z.-J. Wu, H. De-Ping, Effects of viscosity on
cellular structure of foamed aluminum in foaming process, Journal of
Materials Science, 35 (2000), 15–20
12 C. Jimenez, F. Garcia-Moreno, M. Mukherjee, O. Goerke, J. Banhart,
Improvement of aluminium foaming by powder consolidation under
vacuum, Scripta Materialia, 61 (2009), 552–555
13 M. p. i. federation, Standard test methodes for metal powders and
powder metallurgy products, Metal Powder Industries Federation,
Princeton, New Jersey 08540 USA, 1985–1986
14 B. [u{tar{i~, L. Kosec, T. [pan, M. Jelenko, M. Torkar, An instru-
mented cell to analyse the behaviour of metal powders during cold
uniaxial die compaction, Mater. Tehnol., 35 (2001), 351
15 M. Nosko, F. Simancik, R. Florek, Reproducibility of aluminum
foam properties: Effect of precursor distribution on the structural
anisotropy and the collapse stress and its dispersion, Materials
Science and Engineering a-Structural Materials Properties Micro-
structure and Processing, 527 (2010), 5900–5908
16 H. Zhang, E. H. Kisi, Formation of titanium hydride at room tem-
perature by ball milling, Journal of Physics-Condensed Matter, 9
(1997), L185–L190
17 R. S. Vennila, A. Durygin, M. Merlini, Z. Wang, S. K. Saxena, Phase
stability of TiH2 under high pressure and temperatures, International
Journal of Hydrogen Energy, 33 (2008), 6667–6671
18 J. W. Zhao, H. Ding, X. F. Tian, W. J. Zhao, H. L. Hou, Thermo-
dynamic Calculation on the Formation of Titanium Hydride, Chinese
Journal of Chemical Physics, 21 (2008), 569–574
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
18 Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19
19 A. Ibrahim, C. Korner, R.F. Singer, The effect of TiH2 particle size
on the morphology of Al-foam produced by PM process, Advanced
engineering materials, 10 (2008), 845–848
20 I. Paulin, D. Mandrino, C. Donik, M. Jenko, AES and XPS charac-
terization of titanium hydride powder, Mater. Tehnol., 44 (2010) 2,
73–76
21 C. Donik, A. Kocijan, J. T. Grant, M. Jenko, A. Drenik, B. Pihlar,
XPS study of duplex stainless steel oxidized by oxygen atoms,
Corrosion Science, 51 (2009), 827–832
22 C. Donik, A. Kocijan, D. Mandrino, I. Paulin, M. Jenko, B. Pihlar,
Initial oxidation of duplex stainless steel, Applied Surface Science,
255 (2009), 7056–7061
23 D. Mandrino, ^. Donik, M. Jenko, AES of thin oxide layers on a
duplex stainless steel surface, Surface and Interface Analysis, 42
(2010), 762–765
24 B. Matijasevic, J. Banhart, Improvement of aluminium foam techno-
logy by tailoring of blowing agent, Scripta Materialia, 54 (2006),
503–508
25 S. Asavavisithchai, A. R. Kennedy, The effect of compaction method
on the expansion and stability of aluminium foams, Advanced
engineering materials, 8 (2006), 810–815
26 M. Shiomi, S. Imagama, K. Osakada, R. Matsumoto, Fabrication of
aluminium foams from powder by hot extrusion and foaming,
Journal of Materials Processing Technology, 210 (2010), 1203–1208
27 K. Mori, H. Nishikawa, Cold repeated forming of compact for alumi-
nium foam, Journal of Materials Processing Technology, 210 (2010),
1580–1586
I. PAULIN et al.: SYNTHESIS OF ALUMINIUM FOAMS BY THE POWDER-METALLURGY PROCESS ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 13–19 19

L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
A NEW METHOD FOR DETERMINING THE
REMAINING LIFETIME OF COATED GAS-TURBINE
BLADES
NOVA METODA ZA IZRA^UN PREOSTALE TRAJNOSTNE DOBE
LOPATIC PLINSKIH TURBIN
Leonid Borisovich Getsov1, P. G. Krukovski2, N .V. Mozaiskaja1,
Aleksander I. Rybnikov1, K. A. Tadlja2
1NPO ZKTI Russian, 2ITTPh NAN Ukraina
guetsov@yahoo.com
Prejem rokopisa – received: 2010-02-01; sprejem za objavo – accepted for publication: 2011-01-03
In the course of the exploitation of gas-turbine units (GTE) their turbine blades accumulate defects. These defects result from
both the static stress caused by the centrifugal forces and by the thermo-cyclic stress. Since gas turbines usually operate under
variable conditions, the time of exploitation cannot be taken as a reliable parameter for a determination of the remaining
lifetime. Unfortunately, in many cases, at the moment when a GTE is stopped for maintenance, the thermal conditions of the
previous exploitation of the blades are unknown.
In this presentation a new method of determining the remaining lifetime for coated blades is considered. A long-term
strength-decay evaluation is carried out on the basis of the equivalent exploitation temperature Teq, the known exploitation time
 and the static stresses .
Key words: Larson-Miller dependence, turbine blade, remaining lifetime, coating, diffusion, equivalent temperature
Med uporabo plinskih turbin (GTE) se v lopaticah kopi~ijo napake. Te so posledica stati~nih napetosti zaradi centrifugalnih in
termo-cikli~nih napetosti. Ker plinske turbine obratujejo v spremenljivih razmerah, ~asa eksploatacije ni mogo~e upo{tevati kot
zanesljivega parametra za dolo~itev preostale trajnostne dobe. Na nesre~o v mnogih primerih, ko je GTE ustavljena zaradi
vzdr`evanja, niso poznane termi~ne razmere prej{nje uporabe lopatice.
V tem sestavku predstavljamo novo metodo za dolo~anje preostale trajnostne dobe za lopatice s prekritjem. Ocena dolgotrajnega
zmanj{anja trdnosti je izvr{ena na podlagi ekvivalentne temperature uporabe Teq, znanega ~asa uporabe  in stati~nih napetosti .
Klju~ne besede: Larson-Miller odvisnost, turbinska lopatica, preostala trajnostna doba, pokritje, difuzija, ekvivalentna
temperatura
1 MAIN THESES
Four different cases can be considered:
1. Uncooled blades, working in constant conditions (at a
constant temperature). Blades of this type (type 1) in
a stationary mode of operation have a constant
temperature and constant stress in each cross section;
2. Cooled blades in a stationary mode of operation (at a
constant gas temperature). In this case blades have a
specific distribution of stresses and temperatures in
each cross-section. For the calculation of the remain-
ing lifetime of these blades (type 2) the thermostatic
stress is usually omitted and only the static centrifu-
gal stress is used;
3. Uncooled blades in a variable mode of operation
(type 3) – for each mode of operation in each cross-
section they have a constant temperature and stress;
4. Cooled blades working under variable conditions
(type 4).
In recent works 1–6 the surface-layer processes in the
blades were investigated as a function of the long expo-
sures to high temperature. By means of solving the direct
and inverse diffusion problem, the authors have deve-
loped some models and methods of prediction for these
processes. On the basis of these results the temperature
distribution in the surface layer for the weak cross-sec-
tion can be found using an X-ray analysis and a quanti-
tative metallographic analysis of the blade, taken out
from the engine. These calculations are based on the
temperature dependences of the coating elements’
diffusion characteristics. In this way, for example, an
evaluation of the average operating temperatures of the
coating, type NiCoCrAlY with the initial concentration
of the c(Al) = 8.4 %, was performed 6 – Figure 1 (cooled
blades, IN738 alloy; worked for 26 400 h in the turbine).
2 DETERMINATION OF Teq

Due to the differences in the temperature dependence
for the diffusion coefficients and long-term strength, the
temperature TeqD (obtained from the surface-layer
elements redistribution) can be used for the remaining
lifetime determination only in the case of Type-1 blades.
In all the other cases a comparison of Teq and TeqD is
required (see Figure 2).
The method, described below, is based on the
following assumptions:
1. The remaining lifetime of the blade is determined by
the remaining lifetime of its bulk metal;
2. The temperature of the blade cross-section is suppo-
sed to be constant and it corresponds to the maximum
Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26 21
UDK 669.1:621.4:620.179 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)21(2011)
temperature value in the maximally loaded cross-
section;
3. The damages for different temperatures and different
stresses are summed according to a linear law and at
the moment of destruction the following relationship
is true:
(i /ti) = 1 (1)
where i is the exploitation time at the temperature Ti
and under the stress i; ti is the time of destruction under
the same conditions. It is necessary to mention that for a
traditional estimation, it is more appropriate to use
"0.87" instead of "1" in the right-hand part of Eq.1.
Let us suppose that we have the following distri-
bution of work-times for a gas-turbine unit: 1 : 2 : 3 : 4
: 5 etc. under different conditions with the correspon-
ding power N1, N2, N3, N4, N5 etc. and according to the
blades temperatures T1, T2, T3, T4, T5 etc. In this case the
time of exploitation can be found as  = ai.
Let us use the Larson-Miller dependence:
P() = T(C + lg t) (2)
where t is the time of destruction at the temperature T
and stress . The value of the parameter C (in the first
approximation C = 20) can be specified for the involved
temperature interval from the experimental long-term
strength of the blade material.
Using (2):
t1 = 10 [P(1)/T1] – C ; t2 = 10 [P(2)/T2] – C ;
t3 = 10 [P(3)/T3] – C ; t4 = 10 [P(4)/T4] – C ;
t5 = 10 [P(5)/T5] – C ... (3)
where i is determined by the rotation frequency in the
regime Ni, P(i) and C can be obtained from the refe-
rence data for the corresponding material.
Taking into account that the damage summation is
linear (1) and the relationship (3), we can determine the
value a. It corresponds to the given time distribution for
the different operation regimes 1 : 2 : 3 : 4 : 5 etc.
The value of the equivalent temperature Teq can be
obtained from the dependences:
( )( )T
P
P Ti i i
eq
eq
eq


  
=
−∑ ∑
2 3
2 3
, ( )
ln / exp , ( ) / )
P(eq) = Teq
 (C + lg p eq)


i i
i i
T
P CT( )−
=∑ 1 (4)
which parameters are values of the operating time 1 : 2
: 3 : 4 : 5 etc. by the various power settings cha-
racterized by the values Ti and i. Here, P(eq) – para-
meter of dependence of Larson-Miller. Thus, the
remaining lifetime of the blades 2–4 after the turbine
operation during the time  can be determined for the
stress eq multiplied by the long-term strength coeffi-
cient (obtained from the initial turbine project or taken
from the strength standards) as the destruction time at
the temperature Teq minus the turbine operating time.
If we suppose that the exploitation temperatures
changed according to the requirements’ specification (or
according to the recorded exploitation parameters), then
it is possible, as it is described above, to determine the
remaining lifetime. Thus, the calculated Teq is based on
the time ratio 1 : 2 : 3 : 4 : 5 etc., but in most cases
this ratio is unknown. For a determination of the real
(objective) ratio 1 : 2 : 3 : 4 : 5 etc. the calculated TeqD
can be used.
L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
22 Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26
Figure 1: Working temperature (°C) distribution around the cross-
section contour after exploitation for 26 400 h. The numbered points
on the plot correspond to the numbered positions on the blade. Curve
1 – calculated; curves 2 – error range of measurements.
Slika 1: Porazdelitev delovne temperature (°C) vzdol` oboda preseka
po 26 400 h eksploatacije. O{tevil~ene to~ke ozna~ujejo o{tevil~ene
to~ke na lopatici. Krivulja 1 – izra~unano, 2 – obmo~je merilne
napake.
Figure 2: The diagram of the change of temperature during operation
(a) and the temperature dependences of long durability and the factors
of diffusion (b)
Slika 2: Diagram spremembe temperature med operacijo (a) in tempe-
raturna odvisnost dolge trajnosti in difuzijski faktorji (b)
3 METHOD OF DETERMINING THE
WORKING-TIME RATIO FOR DIFFERENT
REGIMES
By solving the inverse diffusion problem for the
experimental Al distribution in the surface layer, the
value of the diffusion coefficient can be determined (for
the real temperature-change law during the turbine
exploitation). Then, if the diffusion temperature depen-
dence is known for the selected element, the value of
TeqD can be determined. This temperature (TeqD) will be
equivalent to the average temperature determined using
the Al diffusion parameters. It can be used as an
equivalent temperature for the determination of the
residual corrosion lifetime of the coating. But for the
determination of the blade’s remaining lifetime under
applied stress (centrifugal bending stress, or especially
thermo-cyclic stress), this temperature (TeqD) must be
corrected according to the aforesaid statements.
Let us suppose that the element distribution in the
diffusion layer does not depend on the consecution of the
temperature-time conditions of the process. Then, by
solving the direct diffusion problem, step by step (for
different initial conditions), at different temperatures Ti
and times of exploitation i, the distributions of the Al
and oxide layer thickness can be obtained (correspond-
ing to the exploitation times). By comparing this data
with the calculated TeqD for the different points of the
profile (Figure 1), the ratio of times 1 : 2 : 3 : 4 : 5
etc. can be obtained.
4 METHOD OF Teq
D DETERMINATION
Let us consider the processes in the solid coating
MeCrAlY (Figure 3).
The oxide Al2O3 is formed during the blade explo-
itation by a combination of Al and oxygen absorbed
from the gas environment and diffused trough the oxide
layer (x1 – x0) to the oxide-coating interface marked as
x1. The diffusion of Al from the coating occurs in two
directions:
• to the oxide-coating interface x1, where it reacts with
diffused oxygen;
• to the coating-bulk alloy interface x4, where it is
accumulated in the inter-diffusion zone and then,
depending on the Al concentration in the bulk metal,
diffuses either into the bulk alloy orin to the coating.
Due to the Al diffusion from the  +  two-phase
coating region, the depleted Al single-phase zones with a
reduced concentration of Al (-phase) are formed on
both sides of the coating – on the side of the oxide and
on the side of the bulk alloy (depleted zones I and II,
Figure 2). All Al, diffused from the coating, is diffused
from the  +  two-phase coating region due to the
disappearance (consumption) of the -phase. The Al
concentration curve in the MCrAlY coating has the
shape of a stepped line, and in the bulk alloy region – a
curve with the maximum in the inter-diffusion zone.
The process of the Al mass-exchange within the
calculation region x1 < x < x can be described with the
following diffusion equation:
∂
∂
=
∂
∂
∂
∂
⎡
⎣⎢
⎤
⎦⎥+
C
x
D
C
x
W
 eff
(5)
 	 x x x1 < < ∞ C C x= ( , )
W W x= ( , ) D D xeff eff= ( )
x x1 1= ( ) x x2 2= ( ) x x3 3= ( ) x x5 5= ( )
with the initial condition:
C x, C x
C x x
C x x
n
oc
( ) ( )0
00 4
0
4
= =
< <
≥
⎧
⎨
⎩
(6)
the boundary condition for infinity:
∂
∂
=∞
C x
x
( , )
0 (7)
and the boundary condition on the moving interface x1,
describing the diffusion flux (due to the concentration
gradient) of the Al from the coating to the left (Figure
3), forming the oxide:
J x D
C(x , )
xOK eff
( , )1
1
+ = −
∂
∂


(8)
The flux (8) conjointly with the flux C x dx /d( , )( )1 1 
(formed by moving the interface x1 to the right) produce
the total Al flux, which creates an oxide film with a
thickness of 
x. The kinetics of the oxide film growth
can be described with two parabolic equations:
Δx x x( ) ( ) ( )  = − =1 0
=
⋅ <
⋅ < ∞
⎧
⎨
⎩
k
k
OK
OK
*
**
  
   
	
 
	 
0
(9,10)
where kox*, kox** – are the coefficients of the oxide-
film-growth intensity (the so-called constants of
oxidation), * – boundary time point, before it the
kinetics obey the oxidation law (9), after – the oxidation
law (10). The curves (9) and (10) for intersect at the
time point * ; 0 – time point (negative value) of the
intersection for the curve (10) and the time axis . The
L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26 23
Figure 3: Typical Al concentration distribution in the coating layer
and the bulk alloy
Slika 3: Tipi~na porazdelitev koncentracije Al v plasti pokritja in v
zlitini
necessity of using two parabolic laws (9) and (10) for
the description of the Al-oxide growth can be explained
by the faster formation (growth) of the -oxide phase
before the time point *, and by the slower growth of the
Al2O3 -oxide phase for  > *. The coefficients kOK*,
kOK**, 0 and the value * in (9) and (10) are determined
from the experimental data of the oxide growth as a
function of time.
The following concentration conditions are taken for
the moving boundaries x2 and x3:
C x C x C y( , ) ( , )2 3− += = 
C x C x C y( , ) ( , )2 3+ − += =   (11)
Thus, by analogy with the boundary condition (8) for
the fixed boundary x4, the diffusion flux of the Al from
the coating to the left (which forms the diffusion zone 
y
= x5 – x4) and the diffusion of Al to the bulk alloy can be
written as:
J x D
C(x , )
xoc eff
( , )4
4
−
−= −
∂
∂


(12)
According to the physical model (Figure 3), the flux
of the Al (12) from the coating induces a new phase
formation in the diffusion zone having the thickness 
y
= x5 – x4, i.e.:
J x W x yoc ( , ) ( , ) ( )4 − = ⋅  Δ (13)
where W is the Al mass, which diffused from the coat-
ing to the diffusion zone x4 < x < x5 depending on C
– C and determined by the relationship
Δy x x k( ) ( )    	= − = ⋅5 4 D D (14)
The coefficient kD and the value D0 in (14) are deter-
mined from the experimental data of the diffusion-zone
increase with time. 
y = x5 – x4 – the diffusion-zone
width, which increases with time due to boundary x5 shift
to the right according to the parabolic law:
W W x= ( , ) =
=
⋅ − <
< >
⎧
⎨
⎩
k C C x x x
x x x x x
w
m( )
,
   

4 5
1 4 50
(15)
where kw is the coefficient of the Al precipitation in the
diffusion zone.
Since in the system the coating/oxide film/bulk alloy
mass balance should be fulfilled, according to (8) and
(12) the total flux of Al from the two-phase coating will
take the form:
J J x J xOK oc   ( ) ( , ) ( , )= ++ −1 4 (16)
According to the physical model, all the Al outgoing
from the coating, actually outgoes from the  + -two-
phase coating zone by consuming the -phase. Then the
movement of the boundaries x2 and x3, and the Al
concentration decrease in the  + -two-phase zone can
be described by the equation of the mass balance bet-
ween the Al diffusion fluxes through these boundaries
and for the diffusion fluxes due to the Al concentration
differences in the  + - and -phases (ΔC C C= −   ):
J J x J x   = + =− +( , ) ( , )2 3
= ⋅ + ⋅ + −Δ ΔC
x
C
x
x x
Cd
d
d
d
d
d
2 3
3 2  
 
( ) (17)
where J x D
C(x , )
x


( , )2
2
−
−= − eff
d
d
and J x D
C(x , )
x


( , )3
3
+
+= − eff
d
d
are the diffusion fluxes towards the oxide and bulk alloy
due to the Al concentration gradients to the left and to
the right of the boundaries x2 and x3, respectively.
The expression (17) after dividing by J takes the
form:
1
2 3= ⋅ + ⋅ +Δ ΔC
x
J C
x
J
d
d
d
d  
/ /
+ − = + +( ) ,x x
C
J g g g3 2 2 3 2 3
d
d
 

(18)
where g2 and g3 are the portions of the total Al mass,
which left the coating due to the boundaries x2 and x3
shifting correspondingly; g2,3 = (1 – g2 – g3) – the
portion of the total Al mass, which left the coating due
to the decrease of the Al -phase concentration in the
region x2 < x < x3. The values g2 and g3 directly affect
the speed of the shifting of the borders x2 and x3 and are
taken as functions of the concentration differences
ΔC C C= −   :
g k
C
C n
2 2 0= ⋅
Δ
g k
C
C n
3 3 0= ⋅
Δ
(19)
From the equations (15) and (16) it is possible to
obtain the laws of the borders x2 and x3 shifting and the
law of the decrease of concentration of Al C+ in the
two-phase zone:
d
d
x
J
k
C n
2 2
0 
=
d
d
x
J
k
C n
3 3
0 
=
d
d
C
J x x g g
 

= − − −( )( )3 2 2 31 (20)
the constants k2 and k3 in (19) can be obtained from the
equation (20) using the experimental data of the borders
x2 and x3 shifting dynamics and of the C   plateau
value in the region x2 < x < x3 for different times of the
tests or exploitation.
The total concentration of Al C   depends on the
-phase concentration C  ( ) in the coating as:
[ ]C C C C CAl y      ( ) ( ) (= ⋅ + −1 (21)
where C Al – concentration of Al in the -phase.
The diffusion coefficient Deff, the coefficient kw and
the power m in the mathematical model (5)–(21) can be
obtained from the experimental data by means of the
inverse-problem solution.
L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
24 Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26
The diffusion coefficient Deff in (5) is correct for the
whole solution region, excluding the sub-region x2 < x <
x3, where it was taken as a greater value due to the
absence of the Al-concentration space gradient.
Al accumulated in the diffusion zone partially
diffuses back into the coating, due to the absence of the
Al-concentration space gradient.
The Al accumulated in the diffusion zone partially
diffuses back into the coating, due to the gradient in the
concentration to the right of the border x4. The related
diffusion flux:
J x D
C(x , )
xmd eff
( , )4
4
+
+= −
∂
∂


returns back into the intra-diffusion zone and is added to
the main flux (13).
The program complex COLTAN created in the ITTPh
NAN Ukraina allows us to calculate the Al concentration
distribution in the coating and in the base blade material,
TeqD and time ratios 1 : 2 : 3 : 4 : 5 etc. in the cooled
blades for different operation modes. In the case of the
uncooled blades (type 3) in order to obtain the time
ratios the inverse problem must be solved for the blades
taken from the GTE for different times of operation 1 :
2 : 3 : 4 : 5 etc.
5 ALGORITHM FOR THE CALCULATION OF
THE RESIDUAL LIFE TIME
On the basis of the aforesaid the following several
steps are used for the determination of the residual
lifetime of the coated blade removed from the turbine
that has been running for the operation time :
1) determination of the parameters of diffusion for
the blade coating by means of experiments and
calculations,
2) experimental determination of the distribution of
aluminium in different points of the surface layer of the
coated blade taken out from the turbine,
3) calculation of the equivalent temperature TeqD
according to a special method based on the solution of
the reverse diffusion problem for experimental data on
the distribution of aluminium (see step 2),
4) calculation of the allocation of the time of GTE
operation in different modes 1 : 2 : 3 : 4 : 5 etc. = r1 :
r2 : r 3 : r 4 : r 5 etc. , corresponding to the power values
N1, N2, N3, N4, N5 etc. (operation time  = a ri, a I = ari.
according to the data obtained during the step 2,
5) calculation of the equivalent temperature Teq using
equation (4)
6) calculation of residual lifetime of the set of blades
(after the operation during the time ).
The algorithm of the solution of the problem men-
tioned in step 4 consists of a number of successive calcu-
lations of the direct diffusion problem (with different
initial conditions) for different temperatures Ti (within
the real range for the examined GTE) and operation
times i (in different modes) as a result of which we
obtain the relevant distributions of aluminium and the
width of the oxide layer, corresponding to the operation
times. Having these set of points for different conditions
we chose those in which:
• the calculated Al distribution coincides with its
experimental distribution,
• the equivalent temperature coincides with TeqD.
The accord between these data and those obtained
with the help of the calculations for the definition TeqD in
different points of the blade profile allows us to calculate
the ratio of times 1: 2 : 3 : 4 : 5 etc. (see Figure 1,
included as an example of the solution of the problem of
finding TeqD distributions) allows us to calculate the ratio
of times 1 : 2 : 3 : 4 : 5 etc.
In other words:
1. Knowing TeqD and the total operation time of the
examined blade, we are able to write the following
equation representing the sum of solutions of reverse
problems for different operating modes contributing to
the resulting distribution curve
F1(, Teq
D) =F1(i, Ti
D) (22)
2. Knowing TeqD, we obtain a number of possible
values I – TiD from the equation (22).
3. Knowing different values TeqjD for different blade
points we find the true values i – TiD from the equation
(22).
4. Having found the true values i – TiD (step 3), we
determine the value Teq using the formula (4).
The residual lifetime of the set of blades (after the
operating time ) is calculated as the time value before
fracture at the temperature Teq under the stress eq,
multiplied by the strength safety factor (obtained in the
course of the strength calculations for blades at the
design stage or by the safety factor chosen from the
strength standards) minus the elapsed operating time .
L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26 25
Figure 4: Al concentration distribution across the coating layer
[NiCoCrAlYRe (Sicoat 2464) coating after exposition at 950 °C]
Slika 4: Porazdelitev koncentracije Al na prerezu plasti pokritja
[NiCoCrAlYRe (Sicoat 2464) pokritje po `arjenju pri 950 °C]
As a basis for the above-mentioned calculations for
MeCrAlY coatings the experimental data obtained in the
present work can be used, namely:
• determination of the Al concentration distribution on
the basis of the coating-layer thickness using the
X-ray microanalysis method;
• determination of the volumetric share of the -phase
on the basis of the coating-layer thickness using
digital optical metallography (for example, with the
help of software made by "IstaVideoTest");
• measurement of the oxide film thickness on the
coating surface and the thicknesses of inner and outer
de-alloyed layers.
These experimental data must be obtained for not less
than three temperatures from the real GTE temperature
range and for several exposures that must produce mean-
ingful results for the sought parameters.
6 X-RAY AND METALLOGRAPHIC ANALYSIS
METHOD
The experimental data required for the above calcu-
lations for the MeCoCrAlY coatings include:
• a determination of the Al concentration distribution
across the coating layer using the X-ray microana-
lysis method;
• a determination of the -phase volume-fraction
change across the coating layer using Digital Optical
Metallography (for example with the software made
by "IstaVideoTest");
• measurements of the oxide film thickness on the
surface of the coating, and the inner and outer de-
alloyed layers thicknesses (Figure 3).
This experimental data should be obtained for at least
three different temperatures in the range of the real
gas-turbine operating conditions with several time
exposures that should bring meaningful results.
For the coatings of a different type, an element sen-
sible to the working-temperature history should be taken
as a diffusing element.
Some illustrative examples of the NiCoCrAlReY
coating study on the alloy Rene80 can be seen in Figure
4–5.
The samples were isothermally exposed for times up
to 20 000 h at the temperatures 900–980 °C. From
Figure 4 it can be seen that on the surface (up to 40 μm
depth) and at the interface coating-alloy (at 180–220 μm)
the formation of a de-alloyed layer is observed.
7 CONCLUSION
A new method of determining the remaining lifetime
of coated blades was developed for stationary conditions.
This method requires one blade’s removal from the
turbine after a long operating time, metallographic
analysis and calculations using the data obtained from
the laboratory experiments performed in advance.
The method is based on the Larson-Miller depen-
dence, the law of linear summation of damages, and the
assumption that the elements-distribution in the diffusion
layer of the coating does not depend on the sequence of
the temperature-time conditions.
8 REFERENCES
1 Getsov L. B., Rybnikov A. I., Krukovski P. G., Kartavova E. S.
De-alloying and fatigue of high temperature alloys used for gas
turbine blades. Materials at high temperature, 13 (1995) 2, 81–86
2 Getsov L. B., Rybnikov A. I., Krukovski P. G. Oxidizing modifi-
cation of surface composition of high-temperature alloys. Protection
of Metals. 4 (1995), 376–381
3 P. Krukovsky, V. Kolarik, K. Tadlya, A. Rybnikov, I. Kryukov, M.
Juez-Lorenzo, Theoretical and experimental approach for long term
modelling of oxidation and diffusion processes in MCrAlY coatings.
European Federation of Corrosion Publications, Number 34,
Proceedings of an EFC Workshop 2001, Edited by M. Schutze, W. J.
Quadakkers and J. R. Nicholls, ISSN P.1354-5116
4 Krukovsky P. G, Kartavova E. S., Getsov L. B. Prediction model of
high temperature diffusion and gas corrosion of gas turbine blade.
Thin Solid Films 270 (1995), 679–688
5 Getsov L. B. Rybnikov A. I. Krukovskiy P. G. et al. A procedure for
determination the change of material fatigue resistance characteris-
tics in long-term operation. Materials at high temperatures. 14
(1997) 1, 15–18
6 Rybnikov A. I., Krukovsky P., Kolarik V., Tadlya K., Mojaiskaia N.
V., Kryukov I. I., Juez-Lorenzo M. Lifetime prediction for gas
turbine blades MCrAlY coatings. Microscopy and Microanalysis,
V10, Suppl.
L. B. GETSOV et al.: A NEW METHOD FOR DETERMINING THE REMAINING LIFETIME ...
26 Materiali in tehnologije / Materials and technology 45 (2011) 1, 21–26
Figure 5: -phase volume fraction distribution across the coating
layer (different duration of exposure at 950 °C)
Slika 5: Porazdelitev volumskega dele`a -faze (razli~no trajanje `ar-
jenja pri 950 °C)
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
REDUCTION OF ULTRA-FINE TUNGSTEN POWDER
WITH TUNGSTEN (VI)-OXIDE IN A VERTICAL TUBE
REACTOR
REDUKCIJA ULTRAFINIH PRAHOV VOLFRAMOVEGA(VI)
OKSIDA V REAKTORJU Z VERTIKALNO CEVJO
@eljko Kamberovi}1, Du{ica Filipovi}2, Karlo Rai}1, Milo{ Tasi}3, Zoran An|i}3,
Milorad Gavrilovski3
1Faculty of Technology and Metallurgy, Karnegijeva 4, 11120 Belgrade, SRB-11000
2IMPOL Seval a. d., Sevojno, SRB-11000
3Scientific Research Center, Veliki Park 1, 31000 U`ice, SRB-11000
kamber@tmf.bg.ac.rs
Prejem rokopisa – received: 2010-09-14; sprejem za objavo – accepted for publication: 2010-11-09
The reduction of WO3 with hydrogen in a vertical tube reactor is a new approach to the reduction of tungsten oxide compared to
conventional procedures in a stationary layer. The advantage of this method is the intensive contact between the reducent and
the oxide particles, a higher degree and a shorter time of reduction of only a few seconds, compared to the several hours
processing time in a horizontal tube reactor.
The characterization of the WO3 powder and of the tungsten powder included a SEM microstructural examination and DT/TG
analyses. The SEM examination of the tungsten powder indicates the presence of small particles with a size of about 1 μm, also
within an agglomerated porous foam structure. Tungsten powder with a particle size of less than 1 μm is obtained at an optimal
temperature with a suitable flow of hydrogen, as a result of stretching and the cracking of large particles, first in the reaction
zone, and then, due to appropriate shock of the temperature, outside of this zone.
Using an appropriate mathematical model, the degree of reduction and the time required for the formation of tungsten particles
during the hydrogen reduction of WO3 in a vertical tube reactor were determined. Experimental studies have shown that the
degree of reduction of WO3 with hydrogen in a vertical tube reactor increases with the temperature and the flow of hydrogen.
Keywords: tungsten (VI)-oxide, reduction, vertical tube reactor, tungsten powder
Redukcija WO3 z vodikom v reaktorju z vertikalno cevjo je nov postopek v primerjavi s konvencionalnimi postopki v
stacionarni plasti. Prednost postopka je intenziven kontakt med reducentom in oksidnimi zrni, kar izbolj{a in skraj{a ~as
redukcije na nekaj sekund, medtem ko traja procesiranje v reaktorju z vodoravno celo nekaj ur.
Karakterizacija prahov WO3 je obsegala SEM-opazovanje in DT/TG-analize.
SEM-opazovanje volframovega prahu odkrije prisotnost majhnih delov z velikostjo 1 μm s strukturo iz penastih aglomeratov.
Volframov prah z velikostjo manj{o od 1μm dobimo pri optimalni temperaturi in ustreznem toku vodika kot rezultat
obremenitve in razpokanja ve~jih delcev, najprej v reakcijski zoni nato pa zaradi temperaturnega {oka zunaj te zone.
Razvit je bil ra~unalni{ki model za nastanek zrn volframa z redukcijo WO3 z vodikom v reaktorju z vertikalno cevjo. Rezultati
raziskave ka`ejo, da stopnja redukcije WO3 raste s temperaruro in s tokom vodika.
Klju~ne besede: volframov(VI) oksid, redukcija, reaktor z vertikalno cevjo, volframov prah
1 INTRODUCTION
The more intensive development of science and in-
dustry requires the application of new materials with
specific properties. The basis for most of these materials
is rare metals and their alloys. Due to their good mechan-
ical, electrical, electro-erosive, tribological and magnetic
properties, tungsten and its alloys are frequently used
materials in industry 1. Besides, tungsten has a high melt-
ing temperature, a high boiling point, a low evaporation
rate at high temperatures and a low thermal expansion
coefficient.
In industrial conditions, these materials, are often
submitted to a reduction process with a gas as the most
commonly used reducent, which easily penetrates into
the oxide pores and ensures a good contact with the
surface and in the interior of the reduced material 2.
Recently, a number of authors investigated the synthesis
of tungsten-based powders using different methods, such
as: the sol-gel method, chemical method co-precipi-
tation, processes of mechanical activation, and plasma
reduction 3–7. With the aim to obtain a pure tungsten
powder, the latest world research is oriented toward the
application of new reducing agents and the development
of new or a modification of existing technological
processes. The process of the reduction of tungsten
oxide, due to the low reaction temperature and the high
degree of purity of the obtained products, was the subject
of extensive research 3,6,7. The results indicated that the
properties of the synthesized tungsten powder depend,
above all, on the properties of the starting powders and
of the parameters of the technological process, mostly its
simplification. At the same time, the variation of the
parameters of the technological process, such as the
temperature and the flow of hydrogen, as well as the
processing in isothermal and non-isothermal conditions,
Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32 27
UDK 669.27 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)27(2011)
led to the conclusion that with the process it is possible
to obtain ultra fine, homogenous and loose tungsten
powders. With the aim to shorten the time of the process,
Lackner 8 performed the reduction of tungsten oxide in a
vertical tube for four temperatures at (800, 900, 1000 and
1100) °C. The starting powder tungsten(VI)-oxide had
an average particles size of 28 μm and tungsten powder
particles with a size of 17.3 μm were obtained by a
reduction at 1100 °C in 1.7 s.
A study of the mechanism and kinetics of hydrogen
reduction of tungsten oxide pointed out that the main
problem of this process is the formation of several
sub-oxides as intermediate products 9–12. The reduction
of WO3, occurs through the following stages of oxidation
8,13: WO3 (-oxide)  WO2,9 (-oxide)  WO2,72
(-oxide)  WO2 (-oxide)  metal, - or -tungsten.
According to Fouad 10, the reduction of WO3 with hy-
drogen occurs through three stages: WO3  WO2.72 
WO2  W, where the activation energy is in the range
from 100.48 kJ/mol to 133.98 kJ/mol. In accordance
with 10,13,14, the reduction of WO3 with hydrogen under
isothermal conditions occurs in two stages: WO3  WO2
 W, with the activation energy, according to Bustnes 15,
of the reduction of WO2 to W of 83kJ/mol. A kinetic
analysis of the reduction of tungsten oxide in a vertical
tube 8 indicates that the process occurs with an activation
energy of 126 kJ/mol.
Although the most common method of producing
tungsten powder is tungsten oxide reduction with hydro-
gen, the reaction conditions and the kinetics of other re-
ducing agents, such as the reduction of WO2 with carbon
in the form of graphite or soot 16,17, resulting in pure
tungsten, and the reduction of oxide tungsten with CO,
when the WC is the final products were also investi-
gated18.
As a part of this study the process of hydrogen reduc-
tion of WO3 powder in a vertical tube reactor was exam-
ined. It is a completely new approach to the reduction of
oxides of tungsten, compared to conventional procedures
in a horizontal tube reactor. The advantage of the proce-
dure is the intensive contact between the reducing agents
and the oxide particles, which results in a higher degree
of reduction and a shortening of the reduction time to
only a few seconds, while the process in a horizontal
tube lasts for several hours.
2 EXPERIMENTAL
For the synthesis of ultra-fine tungsten powder with
the reduction of tungsten oxide a vertical tube reactor
was used. In Figure 1, the scheme of the apparatus for
the reduction of WO3 by hydrogen in a vertical tube
reactor is shown. The apparatus consists of:
• a system of supply,
• gas-purifier system,
• units for heating,
• system to add a starting powder (vibration feeder),
• instruments for measurement and controlling the pro-
cess,
• quartz tubes, diameter of 22 mm and height h = 540
mm,
• cooling system.
The system for gas supply consists of two bottles,
one with compressed hydrogen and one with nitrogen.
The bottle with hydrogen is connected through the
rubber hose to the column containing CaCl2, which is
further connected through the rubber hose to a rinser
with sulphuric acid, both used to remove moisture from
the hydrogen. The flow of gas at the exit from the bottle
with hydrogen is measured with a flowmeter. The quartz
heating tube is used in an electric furnace. The process
parameters are the temperature and the flow of hydrogen.
The holding time of the WO3 powder particles in the
furnace reaction zone is calculated using an equation
based on values of the relevant parameters: tube length,
particle diameter, particle density, gas speed, gas density,
viscosity and gravity. The temperature is measured and
controlled with a thermoregulator, connected to the
furnace with a PtRh-Pt thermocouple. The quartz tube
(inner diameter,  = 22 mm and height, h = 540 mm) is
placed vertically through the furnace and it is attached to
the metal frame. The tube is closed from both sides with
rubber seals. The part of the quartz tube in the furnace is
cooled using a copper heat sink, of spiral form, through
which water circulates. The oxygen supply and the drain
occur from the bottom of the quartz tube. At the end the
hydrogen hose outlet has a rinser with water that serves
as a powder collector. During the process of reduction,
the bottom of the quartz tube is placed in a glass cup
filled with cold water and ice, and the Pyrex powder
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
28 Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32
Figure 1: Schematic review of the apparatus for tungsten trioxide
reduction in a vertical tube reactor. 1. Hydrogen bottle, 2. Nitrogen
bottle, 3. Rotameter, 4. Purificiation columns, 5. Thermoregulator, 6.
Thermopar, 7. Vibration supplier, 8. Radiator, 9. Electroresistant
furnace, 10. Quartz tube, 11. Glass supplement, 12. Ice bowl, 13.
Cleanser
Slika 1: 1. Jeklenka z vodikom, 2. Jeklenka z du{ikom, 3. Rotameter,
4. ^istilne kolone, 5. Termoregulator, 6. Termopar, 7. Vibrator, 8.
Radiator, 9. Elektrouporovna pe~, 10. Kremenova cev, 11. Stekleni
dodatek, 12. Posoda z ledom, 13. ^istilnik
collector is completely submerged to prevent the
re-oxidation of the reduced particles. This will achieve
additional stresses in the particles due to cracking
thermoshocks, which leads to their fracture. On the
quartz tube top a rubber seal is placed and it is drawn
into the glass tube with a rubber hose. A glass funnel is
placed to the rubber hose in which a certain amount of
powder is inserted. The inner diameter of the glass
funnel is 4.75 mm and outside is 7.75 mm. The powder
dosing is accomplished with a vibratory feeder. At the
end of this hose there is a rinser with water.
Synthetic powder of tungsten (VI)-oxide (Merck,
Dermstadt) with a particle size of less than 30 μm was
used as a base material. This material is hygroscopic and
prone to agglomeration. Before the reduction process,
WO3 powder is dried at a temperature of 100 °C. Occa-
sionally, the powder was taken out from the dryer and
mechanically ground. To remove the remaining agglom-
erates, after the completion of the drying process, the
powder is sifted through a sieve with 63 μm holes.
The WO3 powder’s hydrogen reduction was carried
out at six temperatures (700, 725, 750, 800, 850 and 900)
°C and four hydrogen flow rates (10, 20, 30 and 50)
dm3/h.
The characterization of the used WO3 powder and of
the obtained tungsten consisted of a SEM microstruc-
tural examination and a DT/TG analysis. The SEM anal-
ysis was carried out using a JEOL T20 JSM 5300 micro-
scope. For the DT/TG analysis, a NETZCH STA model
409EP with a maximum temperature of 1100 °C was
used. A thermocouple Pt-Pt-Rh S type (10 % Rh) and a
reference material -Al2O3 were used.
The degree of reduction was determined and, using
the appropriate mathematical model, the time required
for the formation of tungsten particles during the hydro-
gen reduction of tungsten in a vertical tube reactor was
deduced. Based on the obtained results the influence of
the temperature, the time and the hydrogen flow on the
reduction degree of tungsten trioxide were determined.
3 RESULTS AND DISCUSSION
In Figure 2, the initial powder of synthetic WO3, and
in Figure 3, the initial powder of synthetic WO3 after
drying, grinding and sifting are shown. The examination
of the initial powder of synthetic WO3 (Figure 2) indi-
cates large agglomerates with a size of 80 μm and parti-
cles smaller than 30 μm. After drying, grinding and sift-
ing, an irregular shape of particles with a size below 15
μm and large agglomerates of size below 40 μm consist-
ing of a great number of small particles connected to one
another are observed. In addition, small particles cover
the large as a patch and cause a spongy morphology.
Since the initial powder is very fine and hygroscopic, the
primary ’condition’ for agglomerate formation is ful-
filled. Namely, the agglomeration of the small particles
is due to their large surface, i.e., a high surface energy
and the interaction through Van der Waals forces that ex-
ist between them.
Figure 4 shows the resulting tungsten powder in the
form of a porous agglomerate of size less than 30 μm.
Also, particles of tungsten of size about 1 μm and
smaller, particles are present suggesting that the reduc-
tion process contributed to the fragmentation of the pow-
der particles.
Figure 5 indicates clearly the presence of a large, po-
rous foam structure agglomerate with a size of about 50
μm with smaller agglomerates in the background. All the
agglomerates consist of single particles smaller than 1
μm. The occurrence of the porous tungsten powder is ex-
plained by the fact that the reduction reaction was heter-
ogeneous and chemically controlled.
The reduction speed was determined from the speed
of the chemical reaction at the interface of the ox-
ide/product, which causes the porosity of tungsten pow-
der. Tungsten powder particles with a size of less than 1
μm are produced with stretching and cracking of the
large particles in the reduction zone. The hydrogen mole-
cules diffuse through the surface of the WO3 particles
and metal tungsten nucleation may occur at the begin-
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32 29
Figure 3: SEM of the initial WO3 powder after drying, grinding and
sifting
Slika 3: Za~etni prah WO3 po su{enju, drobljenju in sejanju
Figure 2: SEM of the initial WO3 powder
Slika 2: Posnetek za~etnega prahu WO3
ning of the process of reduction; however, tungsten nu-
clei begin to grow when most of the WO3 is changed into
WO2. The released oxygen surplus created during the
transformation of WO3 into WO2 diffuses through the
surface and reacting with hydrogen, creates steam, which
as reduction product becomes trapped in the WO3 lattice.
The steam and the growth of tungsten nuclei cause
stresses and later the cracking of larger oxide particles
into smaller fragments. Additional stresses occur during
the transition of reduced particles from the furnace reac-
tion zone, heated to a temperature of 900 °C in the un-
heated to the quartz tube zone and during cooling in the
collector iced water. The temperature shock causes the
rupture and further fragmentation of particles. The de-
scribed reduction mechanism suggests that the process
advances in two stages.
The degree of reduction was determined stoichio-
metrically from the weight difference of the starting
WO3 powder and the reduced tungsten powder, i.e., the
differences in the amounts of oxygen. Table I shows the
obtained experimental values of the degree of reduction.
The time when the WO3 particles fall through the
quartz tube of the vertical reactor during the reduction
process could not be precisely determined with a simple
measurement using a stopwatch. For this reason, it was
determined using the appropriate mathematical model
based on the physical laws of the movement of particles
through a fluid, i.e., on the application of Stock’s law to
the moving of one particle through the fluid. One of the
conditions of the application Stock’s law is that the parti-
cles diameter is not greater than 50 μm 19,20, a condition
that was fulfilled as the starting WO3 powder mean parti-
cles size was about 30 μm.
According to Stock’s law, the speed limit of small
spherical particles which fall out of the fluid is:
V
R g
t =
−2
9
2 ( ) 

s
(1)
where:
s – particle density,
 – fluid density,
 – dynamic viscosity,
g – acceleration due to gravity,
R – diameter of a spherical particle,
Vt – speed of depositing particles.
In this case, the time for a small spherical particle to
fall through the fluid which streams in the opposite di-
rection is:
30 Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
Table 1: Experimental values of the degree of reduction and the mathematically calculated values of the reduction time of the tungsten trioxide
particles
Temperature, T/°C 700 725 750 800 850 900
q/dm3/h 10 10 10 10 10 10
/% 24 29 34 44 53 59
t/s 3.04 3.1 3.16 3.29 3.42 3.53
q/dm3/h 20 20 20 20 20 20
/% 35 42 49 60 67 74
t/s 3.28 3.36 3.44 3.62 3.78 3.95
q/dm3/h 30 30 30 30 30 30
/% 45 53 60 70 78 84
t/s 3.55 3.65 3.76 3.99 4.21 4.44
q/dm3/h 50 50 50 50 50 50
/% 65 72 78 88 95 99
t/s 4.32 4.51 4.69 5.12 5.58 6.09
Figure 5: SEM of the obtained tungsten powder
Slika 5: SEM-posnetek pridobljenega volframovega prahu
Figure 4: SEM of the obtained tungsten powder
Slika 4: SEM-posnetek pridobljenega volframovega prahu
t
v v
=
−
1
t gas
(2)
where:
l – length of the tube reaction zone,
vt – speed of a spherical particle,
vgas – fluid rate.
Since the particle does not flow through an unlimited
fluid, but through a fluid in a dish, the influence of the
walls on the dish during the particles’ falling must be
taken into account. If a particle falls on the axis cylinder
dish, the actual fall time is shorter and it is 20:
t
t
r
R
r
R
* =
+ ⎛⎝
⎜ ⎞
⎠
⎟ + ⎛⎝
⎜ ⎞
⎠
⎟1
9
4
9
4
2 (3)
where:
R – dish diameter,
r – distance from the dish’s central axis.
In Table I the obtained experimental values for the
degree of reduction and the mathematically calculated
values of the reduction time for the tungsten trioxide
particles are shown.
The described mathematical model to determine the
time refers to a single particle. However, during the re-
duction of tungsten trioxide powder in a vertical tube re-
actor the particles fall in the form of small agglomerates.
The particle agglomerates mass is significantly greater
than the mass of one particle and for this reason, the fall-
ing speed of the agglomerated particles is higher than the
falling speed of single particles. For this reason, the ac-
tual reaction time with hydrogen is less than 1 s.
Based on the obtained results, shown in Table I, the
analysis of the influence of temperature, time and hydro-
gen flow on the degree of reduction of tungsten trioxide
was carried out. The influence of the temperature and
time on the degree of reduction of WO3 is shown in Fig-
ure 6. The reduction degree of tungsten trioxide in-
creases with the temperature and the highest values (re-
duction degree 99 %) were achieved at a temperature of
900 °C. In Figure 7 the dependence of the degree of re-
duction of tungsten trioxide on hydrogen flow is shown.
With an increasing of the gas flow, the reduction de-
gree increases and the greatest degree of reduction is
achieved with a hydrogen flow of 50 dm3/h. This phe-
nomenon is explained as a result of the increasing
gas-adsorption properties with its quantity increase per
unit of time and the increase of the diffusion coefficient,
which is greater with a smaller molecular weight of gas.
Bearing in mind that hydrogen has the smallest molecu-
lar mass of all known gas reducers, i.e., a high coeffi-
cient of diffusion, it is clear that the speed of penetration
of hydrogen into small, spherical particles of tungsten
trioxide is high, and therefore the degree of reduction of
the oxide is greater.
During the experiment the analyses of the DT and
DG (Figure 8) of obtained tungsten powder were also
carried out. The described mechanism of reduction
indicates that the process goes through two stages and it
is completely in accord with the results of the DT
analysis, where two exothermic peaks, at 365 °C and 469
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32 31
Figure 8: DT/TG analysis of tungsten powder
Slika 8: DT/TG-analiza volframovega prahu
Figure 6: Influence of temperature and time on the degree of
reduction for the WO3
Slika 6: Vpliv temperature in ~asa na stopnjo redukcije WO3
Figure 7: Influence of the flow of hydrogen on the degree of
reduction for the WO3
Slika 7: Vpliv toka vodika na stopnjo redukcije WO3
°C, respectively (Figure 9) were observed for the sample
W + WxOy.
The TG analysis shows a mass increase in the oxida-
tion of tungsten (250–500 °C) to a value of 22.61 %. The
increase in mass in this temperature zone for pure tung-
sten is 26.14 %, and this indicates that the examined
tungsten sample was not totally pure, i.e., that it con-
tained a significant content of oxides of tungsten. Based
on these results, the calculated degree of reduction is
86.49 %, and in agreement with the experimental results
obtained (88 %) with the appropriate stoichiometric
analysis.
4 CONCLUSION
With the reduction procedure in a vertical tube reac-
tor, tungsten powder with particles of size less than 1 μm
were obtained. Compared to conventional procedures
performed in a stationary layer the procedure is a new
approach to the reduction of oxides of tungsten. The ad-
vantage of this method is intensive contact between the
reducent and the oxide particles, which results in a
higher degree of reduction and a shortening of the reduc-
tion time to only a few seconds, when compared to the
process in a horizontal tube reactor of several hours that
results in a less-reduced powder.
Tungsten powder particles with a size of less than 1
μm were obtained for the optimum temperature condi-
tions and flow of hydrogen as result of the stretching and
cracking of large particles, first in the reaction zone and
then, due to an appropriate temperature shock, outside of
this zone. The obtained tungsten powder shows a similar
tendency to agglomerate as the basic tungsten trioxide
powder and the agglomerate formed is due to its large
area and the high surface energy and the effect of the at-
tractive forces.
The described mechanism of reduction indicates that
the process occurs in two stages and in accordance with
the results of the DT analysis, where two exothermic
peaks can be observed, at 365 °C and 469 °C, respec-
tively, for the sample W + WxOy. Based on the results of
the TG analysis a degree of reduction of 86.49 % is cal-
culated, which is in good agreement with the experimen-
tal results obtained (88 %) with the stoichiometric analy-
sis.
The results showed that the degree of hydrogen re-
duction of WO3 in a vertical tube reactor increases with
temperature and the highest values (99 %), achieved at
900 °C, of less than 1 s indicate that the process is very
fast and intensive. With an increasing of the gas flow, the
degree of reduction increases and the highest values were
achieved with a hydrogen flow of 50 dm3/h.
Further research will be aimed towards optimizing
the process of reduction in a vertical tube reactor with in-
tensification of the reactions and cost-reduction. Also, it
would be useful to achieve a separation of the large frac-
tions of powder at the exit of the quartz tube and their re-
direction in the process of reduction and obtain a circular
system and therefore a higher efficiency of the process.
5 REFERENCES
1 A. Hesham, E. Abdel-Hady, S. Sesnadri, Metallurgical and Materials
Transactions B, 41B (2010), 161–172
2 W. Tremel, H. Kleinke, V. Derstroff, C. Reisner, J. Alloys Compd.,
219 (1995), 471–475
3 A. Al Mohammad, Acta Physica Polonica A, 116 (2009), 240–244
4 M. C. Yang, J. Xu, Z. Q. Hu, Int. J. Refractory Metals Hard Mater.
22 (2004), 1
5 Z. Xiong, G. Shao, X. Shi, X. Duan, L. Yan, Int. J. Refractory Metals
Hard Mater. 26 (2008), 242
6 A. Al Mohammad, Phys. Status Solidi A 205 (2008), 2880
7 G. Da Silva G, C. Alves, V. Hajek, A. Da Silva, G. P., International
journal of powder metallurgy, 41 (2005) 4, 43–50
8 A. Lackner, A. Filzwieser, P. Paschen, Int. J. Refractory Metals Hard
Mater. 14 (1996), 383
9 L. Jiqiao, C. Shaoyi, Z. Zhiqiang, L. Haibo, H. Baiyun, Int. J. Re-
fractory Metals Hard Mater. 17 (1999), 233
10 N. E. Fouad, Journal of analytical and applied pyrolysis, 44 (1997),
13–28
11 N. Fouad, A. Nohman, M. Zaki, Thermochimica Acta 343 (2000),
139–143
12 S. Eroglu, T. Baykar, Journal of Materials Processing Technology
103 (2000), 288–292
13 P. Taskinen, M. H. Tikkanen, Scandinavian Journal of Metallurgy 6
(1997), 228–232
14 J. A. Bustnes, D. Sichen, S. Seetharaman, EPD Congres, The Mine-
rals, Metals & Materials Society, (1993), 581–600
15 J. A. Bustnes, D. Sichen, S. Seetharaman, Metall. Trans. B, 24B
(1993), 475
16 D. Venables, M. Brown, Thermochimica Acta 282/283 (1996),
251–264
17 D. Venables, M. Brown, Thermochimica Acta 282/283 (1996),
265–276
18 D. Venables, M. Brown, Thermochimica Acta 291 (1997), 131–140
19 D. R. Poirier, G. H. Geiger, Transport Phenomena in Materials, Pro-
cessing, TMS, Warrendale, Pennsylvania, USA (1994)
20 D. Filipovi}, K. Rai}, M. Kne`evi}, M. Tasi}, A. Vujovi}, Reduction
of tungsten oxide in vertical tube reactor. Mathematical model for
estimate time of reduction, 4th Balkan Conference on Metallurgy,
Proceedings, Zlatibor, Serbia, 2006, 311–316
32 Materiali in tehnologije / Materials and technology 45 (2011) 1, 27–32
@. KAMBEROVI] et al.: REDUCTION OF ULTRA-FINE TUNGSTEN POWDER ...
Figure 9: DT/TG analysis of sample W + WxOy
Slika 9: DT/TG-analiza vzorca W + WO3
S. AVDIAJ et al.: OXYGEN DIFFUSION IN THE NON-EVAPORABLE GETTER St 707 ...
OXYGEN DIFFUSION IN THE NON-EVAPORABLE
GETTER St 707 DURING HEAT TREATMENT
DIFUZIJA KISIKA V GETRU St 707 MED TOPLOTNO OBDELAVO
Sefer Avdiaj1,2, Barbara [etina - Bati~2, Janez [etina2, Bojan Erjavec2
1Lotri~, d. o. o, Selca 163, 4227 Selca, Slovenia
2Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia
sefer.avdiaj@imt.si
Prejem rokopisa – received: 2010-08-26; sprejem za objavo – accepted for publication: 2011-01-11
The oxygen diffusion and microstructure were investigated in specimens of commercial Non Evaporable Getter under the
trademark of St 707, which is an alloy of Zr(70)-V(24.6)-Fe(5.4) using EDS, AES and XRD. Specimens were in the form of
tablets compressed from St 707 powder of typical particle size from 50 μm to 100 μm. To obtain an observable amount of
oxygen in the bulk material, we loaded the getter specimen at 450 °C at oxygen partial pressure of 1.2 · 10–4 Pa for 1 h. Using
AES and EDS we investigated the distribution of oxygen in different phases of the alloy. From the concentration profile, we can
see the considerable diffusion length of the oxygen into the getter. This diffusion length is much larger in the Laves phase
compared to the zirconium phase. On the zirconium phase two different layers are formed, an oxide layer composed of ZrO2
followed by a diffusion layer, while on the Laves phase only a diffusion layer can be found. In this paper, we also determined
the phases present in our materials, as well as the changes of these phases after oxidation.
Keywords: diffusion, oxidation layer, getter, zirconium, Laves phase
Preu~evali smo mikrostrukturo in difuzijo kisika v neuparljivih getrih St 707(zlitina Zr(70)-V(24,6)-Fe(5,4)) z metodami EDS,
AES in XRD. Preizku{anci so bili v obliki stisnjenih tablet z zna~ilno velikostjo delcev med 50 μm in 100 μm. Pri temperaturi
450 °C smo preizku{ance izpostavili kisiku pri tlaku 1.2 · 10–4 Pa za 1 h. Z metodami AES in EDS smo opazili razli~no
pove~anje vsebnosti kisika v razli~nih fazah zlitine. Iz koncentracijskega profila je razvidna velika difuzijska dol`ina za kisik v
getrskem materialu. V Lavesovih fazah opazimo znatno dalj{o difuzijsko dol`ino kot v cirkonijevi fazi. Na povr{ini cirkonijeve
faze smo na{li dve razli~ni plasti. Na vrhu je oksidna plast ZrO2, ki ji sledi difuzijska plast. V Lavesoih fazah pa najdemo le
difuzijsko plast. Prav tako smo ugotovili fazno sestavo materiala in spremembe zaradi oksidacije.
Klju~ne besede: difuzija, plast oksidacije, geter, cirkonij, Lavesove faze
1 INTRODUCTION
Getters are solid materials capable of chemisorbing
gas molecules to their surface: getters are chemical
pumps. Non-evaporable getters are applied in vacuum
systems where a very low partial pressure of hydrogen is
required, e.g., in ultra-high-vacuum (UHV) chambers 8.
Because they exhibit selective pumping, they are also
used for inert-gas purification. In general, getters are
classified into two main groups: evaporated getters (EGs)
and non-evaporated getters (NEGs). During the handling
of NEGs under usual ambient conditions their surface
becomes covered with a passive layer consisting mainly
of chemisorbed oxygen and carbon. But a surface
phenomenon is not the only characteristic of NEGs. This
passive layer must be eliminated to start the gettering
action and the metal surface is less oxidized after O2
exposure to high temperatures than at low temperatures,
because the oxygen diffuses from the surface into the
bulk 7. The process of activation of the getter is described
in 1,5,11,12. Lately, getters have found widespread appli-
cation in various fields, such as ultra-high-vacuum
systems of particle accelerators, sealed vacuum devices
such as vacuum tubes, and field emission displays. In
this study (St 707) non-evaporable getter material has
been investigated. Techniques such as Energy-Dispersive
Spectroscopy (EDS), Auger Electron Spectroscopy
(AES) and X-Ray Diffraction (XRD) were used for the
material characterization.
2 EXPERIMENTAL
Our specimen is an alloy with mass fractions of
Zr(70)-V(24.6)-Fe(5.4) %. In the typical procedure,
specimens were prepared through grinding and
polishing, and then immersing in an ultrasonic bath with
alcohol for cleaning. Then the specimens were inserted
into the test chamber (CH2), Figure 1. Then we pumped
the gases out of the CH2 (through the turbomolecular
pump) until the pressure in the vacuum chamber reached
10–7 mbar. Afterwards, the specimen was heated for
activation, and simultaneously the temperature was
measured by a thermocouple attached to CH2. In the
experiment the specimen surface was exposed to the
given partial pressure of test gas at elevated temperature.
The sources of the test gases are gas cylinders. When
desired temperature was reached, the regulating valve
was opened and the gas (in our case pure oxygen) is
permitted to flow into the system. Opening of the
regulating valve was controled to reach the pressure of
1.2 · 10–4 Pa in the CH2, while, the temperature was 450
°C. The specimen was treated at this conditions, for a
time of one hour. During this treatment the diffusion of
Materiali in tehnologije / Materials and technology 45 (2011) 1, 33–37 33
UDK 533.5:543.428.2 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)33(2011)
O2 into the specimen started. The purity of the oxygen
admitted to the getter sample was better than 99.99 %.
The investigation of the diffusion depth of oxygen
was performed using SEM and AES with different
accelerating voltages and different magnifications. The
crystal structures of the obtained specimen were
characterized by powder X-ray diffraction (XRD). The
specimens were characterized in both the as-received and
oxidized states.
3 RESULTS
3.1 EDS analyses
The compositions of the constituents of the
as-received St 707 specimen and after their oxidation in
the vacuum system were investigated by EDS. The
results of the EDS investigations are summarized in
Table 1 and Figures 2a and 3a. The specimen in the
as-received state contained particles which consisted of
two different compositions of zirconium-rich phase and
Laves phase, as can be seen from Figure 2a. Areas rich
in zirconium are surrounded by those with which
vanadium is the major component. It was found that iron
is present only in the vanadium-rich areas 2. According
to Figure 2 and Figure 3, the bright areas in the
secondary electron images could be assigned to areas
rich in elemental zirconium, whereas the dark ones
indicate the dominance of the Laves phase.
When the getter is exposed to air, it absorbs oxygen
and the gases present in the atmosphere. As con-
sequence, a layer is formed at its surface which prevents
further adsorption. For further use of the getter it should
be activated, which is achieved through increasing the
temperature. For this reason, a small amount of oxygen
will be found in the "as-received" getter, Figures 2a and
3a. The presence of oxygen in the "as-received"
specimen is barely noticeable compared to the oxygen
presence in the oxidized specimen Figures 2b and 3b,
and Table 1. Also in the "as-received" specimen the
presence of oxygen is almost similar in all phases, the
zirconium-rich phase and the Laves phase.
From the EDS spectra as well as the EDS elemental
mapping images of the as-received getter material it is
obvious that the amount of oxygen is almost the same in
S. AVDIAJ et al.: OXYGEN DIFFUSION IN THE NON-EVAPORABLE GETTER St 707 ...
34 Materiali in tehnologije / Materials and technology 45 (2011) 1, 33–37
Figure 2: (a) Secondary-electron image of a polished surface of the
as-received St 707 specimen and element distribution images for O, V,
and Zr. (b) Secondary-electron image of a polished surface of the
oxidized specimen St 707 at temperature T = 450 °C, pressure P =
1.2 · 10–4 Pa, and time t = 1 h and the element distribution images for
O, V, and Zr
Slika 2: (a) Polirana povr{ina preizku{anca St 707 in porazdelitev ele-
mentov O, V in Zr. (b) Polirana povr{ina oksidiranega preizku{anca St
707 pri temperaturi T = 450 °C, tlaku P = 1.2 · 10–4 Pa za ~as t = 1 h
ter porazdelitev elementov O, V in Zr
Figure 1: Experimental setup of the vacuum system. SRG is Spinning
Rotor Gauge, CDG is Capacitance Diaphragm Gauge.
Slika 1: Eksperimentalna postavitev vakuumskega sistema. SRG je
viskoznostni vakuumski merilnik in CDG kapacitivni membranski
merilnik.
both phases, the zirconium-rich phase and the Laves
phase Figure 2a. On the other hand, through the same
techniques we observed a larger oxygen content in the
zirconium-rich phase compared to the Laves phase
(Figure 2b). Also from Figure 3 and Table 1 it is
evident that there is more oxygen in the zirconium-rich
phase than in the Laves phase. The spectrums 1, 2, 3 are
of the zirconium-rich phase and spectrums 4, 5, 6 are of
the Laves phase. Measured oxygen concentration in
spectrums 1, 2 and 3 of oxidized specimen is close to
60%. This happened due to the fact that close to 500 °C
the formation of ZrO2 bond occurres 2 and also the
solubility values indicate that metals of the 4th group (in
our case zirconium) have higher storage capacities for
oxygen than those of the 5th group (V), or, putting it
another way, that hcp crystalline structures have in
general higher storage capacities than the bcc ones 11.
3.2 AES analyses
The surface of the getter materials is particularly
difficult to analyze because of their high chemical
reactivity. The results obtained can, therefore, be strong-
ly influenced by the experimental set-up and the proce-
dures 3,4, as recently confirmed through our measurements.
The AES results are also compared with complementary
Materiali in tehnologije / Materials and technology 45 (2011) 1, 33–37 35
S. AVDIAJ et al.: OXYGEN DIFFUSION IN THE NON-EVAPORABLE GETTER St 707 ...
Figure 4: (a) Two-dimensional element distribution images for O, V,
and Zr by Auger Electron Spectroscopy on a polished surface of the
oxidized St 707 alloy, (b) AES analyses for "oxidized" specimen in
different phases, more oxygen is present in Zr (spectrum A1-59.50 %)
than in the Laves phase (spectrum A4-20.47 %)
Slika 4: (a) AES ploskovna porazdelitev elementov O, V in Zr na
polirani povr{ini oksidiranega St 707 (b) AES to~kovna analiza oksi-
diranega preizku{anca v razli~nih fazah, kjer je razvidno, da je v
Zr-fazi pove~ana koncentracija kisika glede na Lavesovo fazo.
Figure 3: EDS analyses: a) as-received specimen, the presence of
oxygen is almost the same in both phases b) EDS analyses for
"oxidized" specimen in different phases, oxygen is more present in the
zirconium than in the Laves phase
Slika 3: EDS-analize: a) za~etni material, kjer je vsebnost kisika v
obeh fazah podobna, b) pri oksidiranem preizku{ancu je vsebnost
kisika v zirkonski fazi ve~ja kot v Lavesovi.
Table 1: Composition of "as-received" specimen and "oxidized" spe-
cimen using EDS
Tabela 1: Sestava za~etnega in oksidiranega preizku{anca
O V Fe Zr Total
Spectrum "as-received " specimen
Spectrum 1 13.32 45.28 9.31 32.09 100
Spectrum 2 14.94 4.93 0.00 80.12 100
Spectrum 3 12.52 8.99 0.00 78.48 100
Spectrum 4 18.82 42.07 8.99 30.12 100
Spectrum "oxidized" specimen
Spectrum 1 59.75 4.47 0.00 35.78 100
Spectrum 2 60.99 0.45 0.00 38.56 100
Spectrum 3 60.48 1.81 0.00 37.70 100
Spectrum 4 19.29 40.69 9.07 30.95 100
Spectrum 5 16.46 41.41 9.85 32.28 100
Spectrum 6 0.00 51.88 10.86 37.25 100
EDS measurements. The results from the AES, Figure 4,
are similar to those obtained with EDS, Figure 3. Again,
we can observe that oxygen is more present in the
Zr-rich phase than in the Laves phase. From Figure 4 we
see that the presence of oxygen in atomic percent (in the
depth 41 nm) in the Zr (spectrum 1) is 59.50 %, and in
the Laves phase (spectrum 4) is 20.47 %. Also, from
mapping, a larger oxygen content in Zr compared to the
Laves phase (Figure 4a) was observed. In addition, a
larger amount of oxygen in the grain boundary and grain
particles is observable, because the grain boundaries, as
dislocations, are considered as short-circuits for atomic
diffusion in metals 11.
AES depth profiling measurements were made at
some points of the specimen in order to be able to
estimate the penetration depth of the oxygen within the
getter, concretely for the zirconium-rich phase, (Figure
5a) and the Laves phase (Figure 5b). This measurement
technique makes it possible to distinguish the oxidized
zone (the presence of ZrO2) and the diffused zone (the
zone where oxygen is inserted). Because of the forma-
tion of ZrO2 chemical bond in the zirconium-rich phase
there is the presence of the oxidized layer. This oxidized
layer is absent in the Laves phase. On the other hand, the
thickness of the diffused layers is higher for the Laves
phase compared to the zirconium-rich phase.
3.3 X-ray diffraction analysis
XRD analyses were carried out in order to determine
the phases present in the materials. We confirmed that
the as-received alloy is biphasic at room temperature
(300 K), consisting of hexagonal -Zr and the cubic
Laves phase Zr(V1-xFex)2 as shown in Figure 6a and
indicated in references 2,10. We can see in Figure 6b that
for the sample treated at 450 °C under oxygen partial
pressure of 1.2 · 10–4 Pa the intensity of the Laves phase
will decrease and, at the same time, a new phase of ZrO2
will emerge. This new phase comes as consequence of
ZrO2 bond formation at high temperatures. The ZrO2 has
a monoclinic structure. The as-received specimen
exhibits narrower XRD peaks, while with the oxidized
specimen a certain broadening of the peaks (mainly
Laves phase) due to lattice strains and distortions can be
observed.
4 DISCUSSION
The diffusion length during the time t at temperature
T is given by 13:
L D t t= ⋅( )
where L/cm is the thickness, D/(cm2 s–1) is the diffusion
coefficient and t/s is the time. The diffusion coefficient
(or diffusivity) in the metal at temperature T is:
D D t
E
RT
= ⋅ −⎛⎝
⎜ ⎞
⎠
⎟
0 ( ) exp
D0 is a constant, E is the activation energy for diffusion,
and R is the universal gas constant. In the literature 15
we have found following values for hexagonal -Zr
(mean of two data): D0 = 0.21 cm2/s and E = 184.1
kJmol–1K–1, and for cubic V: D0 = 0.013 cm2/s and E =
121.3 kJmol–1K–1. Using these values we get for Zr: D(T
= 723 K) = 1.3 · 10–14 cm2/s and for V: D(T = 723 K) =
2.6 · 10–11 cm2/s.
The corresponding diffusion lengths for 1 h treatment
at 450 °C are 70 nm for Zr and 3050 nm for V. We can
S. AVDIAJ et al.: OXYGEN DIFFUSION IN THE NON-EVAPORABLE GETTER St 707 ...
36 Materiali in tehnologije / Materials and technology 45 (2011) 1, 33–37
Figure 5: AES depth profiles for a) Zr phase and b) Laves phase at
450 °C for 60 min
Slika 5: AES globinski profil za a) Zr-fazo ter b) Lavesovo fazo.
Preizku{anec je bil oksidiran pri temperaturi 450 °C, 60 minut.
Figure 6: X-Ray diffraction spectra: a) as-received specimen, b) oxi-
dized specimen at 450 °C for 1 h
Slika 6: XRD-spektri: a) za~etni preizku{anec, b) oksidiran preizku{a-
nec pri temperaturi 450 °C in ~asu 1 h
see from Figure 5a that in the diffusion layer in Zr the
drop of oxygen concentration from 66 % to nearly zero
occurs in approximately 100 nm. Because of almost 50
times larger diffusion length in V we can not see in Fi-
gure 5b a considerable change of oxygen concentration
in Laves phase over the investigated depth of 600 nm.
The data for the diffusivity and the solubility limit
from the literature give us a global view about the
diffusivity and solubility limit in single-crystal pure
metals. However, the NEG materials under investigation
are alloys. The diffusivity and the solubility-limit values
that can be found in the literature are valid for single
crystals, so they are not very helpful to evaluate the
diffusivity and the solubility limit in the case of alloy. It
is known that the diffusivity and the solubility limit
increase in the presence of grain boundaries. From the
literature we find that metals of the 5th group (in our case
vanadium) have a higher diffusivity for oxygen than
elements of the 4th (in our case zirconium) group,
whereas the solubility limit values indicate that metals of
the 4th group (Zr) have higher storage capacities for
oxygen than those of the 5th group 11. This can be clearly
seen from the results of our measurements, which are
presented in Figure 5. There are two layers in the
zirconium-rich phase, the first one is the oxidized layer,
and the second is the diffused layer. On the other hand,
in the Laves phase there is only the diffused layer. The
appearance of the oxidized layer in the zirconium-rich
phase is the consequence of the affinity of Zr to form
chemical bonds with O2 at elevated temperatures. As a
consequence of the composition of ZrO2, the presence of
oxygen in the zirconium-rich phase in mole fractions (%)
can be twice that of the zirconium atoms at most, which
is evident in Figure 5a. By increasing the temperature or
time of oxidation, we will observe an increase in the
thicknes of the oxidized layer, whereas the structure of
diffused layer will remain similar. Thus we have a
saturation of the zirconium-rich phase with oxygen in the
oxidized layer. On the other hand, in the Laves phase the
oxidized layer practically does not exist. Only the
diffused layer is present there, which is in accordance
with theory 11. According to theory, the elements of the
5th group have a greater affinity for oxygen diffusion
compared to those of the 4th group, whereas the solubi-
lity in these elements is quite small.
5 CONCLUSION
According to the results reported in this paper, it can
be stated that:
– The XRD together with EDS and AES confirms that
the as-received alloy is biphasic.
– The hexagonal zirconium and cubic Laves phases
coexist in this material.
– The presence of oxygen in the oxidized specimens is
more observable in the zirconium than in the Laves
phase, whereas in the as-received specimen, the
presence of oxygen is almost the same in both
phases.
– XRD analysis of the sample oxidised at 450 °C
showed a decrease in the intensity of the Laves
phase reflections and the formation of ZrO2 was
indicated.
– With AES depth profiling it is possible to determine
the size of the zones affected by the oxygen. The
presence of the oxide layer composed of the ZrO2.
was confirmed on the Zr phase.
Acknowledgement
The presented work was partly financed by the
European Union, the European Social Fund. Operation
implemented in the framework of Operational Program
for Human Resources Development for the period
2007-2013, Priority axis1: Promoting entrepreneurships
and adaptability, Main type of activity 1.1.: Experts and
Researcher for competitive enterprises.
6 REFERENCES
1 Lafferty J., Foundations of Vacuum Science and Technology, John
Wiley & Sons, New York (1998)
2 Gunter M., et al., Microstructure and bulk reactivity of nonevapo-
rable gettter Zr57V36Fe7, J. Vac. Sci. Technol. A, 16 (1998) 6,
3526–3535
3 Prodromides A., Scheuerlein, Taborelli M., The characterization of
non-evaporable getters by Auger electron spectroscopy: analytical
potential and artefacts, Applied Surface Science, 191 (2002),
300–312
4 Prodromides A., Scheuerlein, Taborelli M., Lowering the activation
temperature of TiZrV non-evaporable getter films, Vacuum 60
(2001), 35–41
5 Jousten K., Handbook of Vacuum Technology, Wiley-WCH Verlag
GmbH & Co. KgaA, Weinheim (2008)
6 Malyshev O., Middleman K., Activation and measurements of
nonevaporable getter films, J.Vac. Sci. Technol. A, 27 (2009) 2,
321–327
7 Vedel I., Schlapbach L., Surface reactivity of ZrVFe getter alloys
exposed to H2O, CO and O2 at 300 K and 700 K, J. Vac. Sci.
Technol. A, 11(1993) 3, 539–542
8 Roupcova P., Schneeweiss O., Changes in phase composition of
Zr-Fe-V getter after hydriding and vacuum dehydriding cycles,
Journal of Alloys and Compounds, 492 (2010), 160–165
9 Knize R., Cecchi J., Theory of bulk gettering, J. Appl. Phys., 54
(1983) 6, 3183–3189
10 Boffito C., Ferrario B., Della Porta P., Rosai L., A nonevaporbale low
temperature activatable getter material, J. Vac. Sci. Technol., 18
(1981) 3, 1117–1120
11 A. Prodromides, Non-Evaporbale Getter Thin Film Coatings for
Vacuum Applications, Ecole Polytechnique Federale de Lausanne,
PhD thesis, (2002), 16–32
12 B. Ferrario, Chemical pumping in vacuum technology, Vacuum,
47(1996) 4, 363–370
13 C. Schuman, E. Lenarduzzi, S. Weber, M.J. Philippe, D. Petelot, P.
Bounie, Modelling the oxygen diffusion profile in TA3V and T40
during heat-treatment under air between 620 and 850 °C, Surface and
Coating Technology, 200 (2006), 4572–4578
14 Flinn B., Zhang C., Norton P., Oxygen diffusion along the [0001]
axis in Zr(0001), Physical review B, 27 (1993) 4, 16499–16505
15 J. D. Fast, Gases in Metals, The Macmillan Press Ltd. (1976), 127
S. AVDIAJ et al.: OXYGEN DIFFUSION IN THE NON-EVAPORABLE GETTER St 707 ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 33–37 37

B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
THE MODELING OF AUGER SPECTRA
MODELIRANJE AUGERJEVIH SPEKTROV
Besnik Poniku, Igor Beli~, Monika Jenko
Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia
besnik.poniku@imt.si
Prejem rokopisa – received: 2010-12-10; sprejem za objavo – accepted for publication: 2011-01-11
The presented work deals with the process of simulated AES spectra creation. The purpose of virtual spectra creation is to use
them as a testing ground for different background-removal and noise-reduction algorithms. Numerous methods for background
removal and noise reduction exist already, but there is uncertainty about their performance and the accuracy of the data
produced, since it is well known that the raw data is altered any time it undergoes any kind of processing conducted for different
purposes. This is due to the fact that the levels of background and noise are at unknown levels in the measured AES spectra. By
using simulated spectra the case is different. The levels of background and noise are added by the user, thus their levels are
known exactly prior to processing. For this reason, they make a valuable means for assessing the performance of different
algorithms intended for background removal and noise reduction.
The article describes the principles of simulation of AES spectra and shows the basic elements of the developed software.
Keywords: AES spectra, data processing, simulation, neural networks, VBA
Predstavljeno delo obravnava proces simulacije Augerjevih spektrov. Osnovni namen generiranja virtualnih spektrov je
vzpostavitev okolja, namenjenega preizku{anju razli~nih algoritmov za odstranjevanje spektralnega ozadja in {uma. Obstaja
veliko splo{nih metod za odstranjevanje ozadja in {uma, ki pa so bile narejene za druge namene, zato njihov vpliv na
interpretacijo sestavin spektra ni znan. Znano je, da vsaka metoda, ki odstranjuje ozadje in {um, neizogibno popa~i tudi osnovno
informacijo. Do sedaj je bilo nemogo~e zanesljivo oceniti napake, ki jih povzro~ajo posamezni algoritmi, ker so bili vedno
uporabljeni na merjenih spektrih, za katere ni to~no znano, kako so sestavljeni, prav tako tudi ne vemo, kak{no je ozadje in
koliko je zares {uma. Pri simuliranih spektrih pa v vsakem primeru posebej natan~no vemo, kako so sestavljeni, zato je ocena
delovanja metod za odstranjevanje ne`elenih delov spektra lahko zelo zanesljiva.
^lanek opisuje postopke simulacije Augerjevih spektrov in prikazuje osnovne gradnike razvitega programskega orodja.
Klju~ne besede: Augerjevi spektri, obdelava podatkov, simulacija, nevronski sistemi, VBA
1 INTRODUCTION
The main goal of our work was to design a system
which will be able to generate a series of AES (Auger
Electron Spectroscopy) spectra with one clear aim – to
provide an environment where numerous algorithms for
AES spectra analysis can be verified, compared and
tested. In order to achieve this goal, we have analyzed
the appearance of many measured spectra. The article
describes the way the AES spectra were analyzed and
finally how the different elements were brought together
to produce the artificial AES spectra.
1.1 A brief description of AES
Auger electron spectroscopy is a powerful analytical
technique for studying surfaces.1 AES is a very
surface-sensitive technique due to the fact that most of
the signal comes from the topmost atomic layers of the
surface of the sample.2 Together with X-ray photo-
electron spectroscopy (XPS), it has been extensively
used for determining the chemical compositions of thin
films of nanometer thicknesses in a large variety of
applications.3 This technique is based on the Auger
effect, a phenomenon discovered by both Pierre Auger
and Lise Meitner independently in the 1920s, even
though Pierre Auger is accredited with the discovery in
the scientific community.4 The Auger effect constitutes
the emission of an electron from an atom following the
filling of a vacancy in an inner electron shell.5 Even
though the discovery of the phenomenon was made in
the early 1920s, it was not until the 1950s that AES
became a practical characterization technique for
analyzing surface features.6
AES involves probing the surface with an electron
beam, creating excited states of atoms present in the
sample7 through core-shell ionization, and emitting
"Auger electrons" with characteristic energies,8 which
are used to determine the surface chemical composition.9
Afterwards, these electrons enter the analyzer, where
depending on the type of the analyzer used they get
deflected to degrees which depend on their kinetic
energies, and then electrons with different kinetic
energies enter the detector at different locations. With the
detector the number of electrons reaching specific
locations is measured. At the end of the process we
obtain the data plotted on the screen as the number of
electrons per second reaching the detector versus their
kinetic energy. A simplified diagram of this whole
process is represented in Figure 1.
After obtaining the spectrum the data needs to be
processed in order to derive from it the necessary
information. Different tools (programs) are used for
processing the AES spectra. During these operations
Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46 39
UDK 543.428.2:519.68 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)39(2011)
though, e.g., differentiation of the spectra for quanti-
fication, inevitably the raw data is being altered, often at
unknown levels to the user. It is important to have
information about the influence the program itself has on
the data during the processing. In this way the post-
processing features observed on the spectra that are not
representative of the material, but are due to the influ-
ence of the program, will be readily accounted for.
When the spectra are measured, the peaks and noise
are superimposed on a slowly varying background.11 For
different purposes, such as the automation of the
recognition of AES spectra, data processing like back-
ground subtraction and noise reduction is obligatory.
Alongside noise, the background is usually regarded as a
problem of data processing.12 In the process of back-
ground subtraction and noise reduction through different
tools, the spectra are being altered, often at unknown
levels to the experimenter.
This work is going to deal with the process of
modeling the AES spectra, with the intention of creating
simulated spectra to be used later for assessing the
performance of the background-removal and noise-
reduction algorithms. The advantage of using simulated
AES spectra lies in the fact that the levels of background
and noise are added by the user, and thus are known
prior to the spectra processing. Having prior knowledge
about the levels of these two elements of the spectra is a
clear advantage, because after processing them through
the algorithms for background removal and noise
reduction, by comparison the influence the program
itself has on the data during the processing can be
determined. By taking into account this fact and
correcting for it, the uncertainty in the outgoing data will
be reduced.
2 EXPERIMENTAL PART
The modeling of the AES spectra involved the careful
investigation of numerous real spectra measured from
metallic samples. Figure 2 represents some typical
examples of the measured spectra.
The spectra were taken using the Microlab 310-F
spectrometer at the IMT laboratory for surface analysis.
2.1 Approximation of the background
For the purpose of defining the general shape
properties of the background, the data were transferred
in a digital form from the standard .vms files of
CasaXPS into Excel. Figure 3 represents the digital data
already transferred into an Excel file, with the graph of
the spectra being drawn from the imported data.
There the number of points representative of the
background were taken, and using the graphing tools in
Excel the approximate background for a couple of
spectra were drawn. Figure 4 represents such a case.
Those background shapes for a couple of spectra, in
the normalized form, were plotted together in order to
investigate and possibly reach some conclusions about
their shapes. Figure 5 represents the manually defined
background shapes for a number of spectra.
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
40 Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46
Figure 3: Spectra in digital form transferred from a .vms file and the
corresponding graph in Excel
Slika 3: Digitalni zapis spektra, prenesenega z datoteke .vms, ter pri-
padajo~ graf v Excelu
Figure 2: The "Direct" AES spectra used for the determination of the
shape of the spectra
Slika 2: Nekaj primerov Augerjevih spektrov, ki so bili uporabljeni za
dolo~evanje posameznih sestavnih delov spektra
Figure 1: Detection process in Auger Electron Spectroscopy10
Slika 1: Proces detekcije pri Augerjevi elektronski spektroskopiji10
From Figure 5 it is obvious that the AES spectra
background is very much variant. Nevertheless, some
distinctive features can be observed there, like the
"low-energy part" – the descending part where the back-
ground shows a rapid decrease in the signal with the
kinetic energy increase; the "mid-energy part" – the
central part where the increase of the background is
observed, which is again followed by the decrease; and
the "high-energy part" – the raising part where the back-
ground is steadily increasing.
The manual background definition is a very time-
consuming process, as much as it is not very reliable.
2.2 Neural network system for background AES appro-
ximation
In order to produce a more reliable impression of the
background, where a larger amount of AES spectra can
be checked with regard to the shape of the background,
we have developed a method capable of approximating
the background shape. For this purpose, a feedforward
neural network was used.
For our purposes, we have used the neural networks
for data modeling or function approximation. Neural
networks are model-less approximators, meaning they
are capable of accomplishing the approximation tasks
regardless of any knowledge of the nature of the modeled
system. For classical approximation techniques, it is
often necessary to know the basic mathematical model of
the approximated problem.
The most important property of neural networks is
their ability to learn the model from the data presented.
This means that the user needs only to be sure that a
causal relation between the presented parameters really
exists. Almost no further knowledge of the system
behavior is needed. When the neural network builds the
model, the dependences among the parameters are
included in the model.
For implementing the neural networks we used the
Neuralyst (Cheshire Ltd.), which is a supplement to
Microsoft Excel and VBA (Visual Basic for Accesso-
ries), to be able to make the best use of Neuralyst.
The backgrounds for 63 measured AES spectra were
approximated using an error backpropagation neural
network13,14. Since the cumulative AES spectra count
values differ in magnitudes significantly, and we have
been interested in the comparison of the shape of the
spectra, all the samples were normalized prior to the
approximation by the neural network. Figure 6 repre-
sents the outcome of the neural network formation of the
background for all 63 measured spectra.
From Figure 6 we can see the same characteristics as
was the case with the method where the background was
determined manually (Figure 5).
2.3 Modeling the background shape
After obtaining a general visual impression for the
shape of the backgrounds of the spectra and investigating
them, we went to the next level of finding a represen-
tative function that would best fit that shape, and the
function of which would be used later to simulate the
background.
According to the characteristics observed in Figures
5 and 6 the ideal function for the AES spectra back-
ground was proposed to be:
f A
px
bx c BB
e
e= + +
⎛
⎝
⎜
⎞
⎠
⎟ +
1
(1)
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46 41
Figure 5: Manually defined background in normalized form
Slika 5: Ro~no dolo~eno ozadje v normirani obliki
Figure 6: The curves of AES spectra backgrounds for various
measured spectra as approximated by the neural network.
Slika 6: Krivulje ocenjenih ozadij za Augerjeve spektre. Ozadje je
bilo ocenjeno z nevronskim sistemom.
Figure 4: Manually defined background for one spectrum
Slika 4: Ro~no dolo~evanje ozadja spektra
Where:
fB: represents the background function
xe: the kinetic energy
A, B, p, b, c: parameters which define the shape of the
background
The construction of the background function is
shown in Figure 7.
The construction of the randomly shaped background
function starts with the random selection of the para-
meters of the linear part
f bx clin e= + (2)
where the points P1 and P2 (Figure 7) are randomly
selected. The point P1 is the zero crossing of the linear
part, and P2 is the value of the normalized count for the
kinetic energy of 1650 eV. The random generator
generates the values for P1 in the range [0, 1000], and
P2 in the range [60, 140]. The limits for both values
were defined experimentally, by observing the behavior
of real spectra.
From the known points P1 and P2 the coefficients b
and c of the linear part are defined.
y bx c= +e
b
P
P P
y
x x
=
−
2
2 1
; P y2 1650= ; P y1 0= (3)
c bP x= − 1
Then follows the definition of the non-linear part
f
pxnlin e
=
1
(4)
For this part the random generator generates the
value of the non-linear part for the kinetic energy of 50
eV. In figure 8 this point is denoted by P3. The random
generator selects the values from the interval [20, 500].
Again the border values were taken experimentally.
Finally the value for the parameter p is set.
P
P Px y
=
⋅
1
3 3
(5)
Parameters A and B are set in the scaling process,
where the minimal part of the background function is set
close to 0, and the maximal part close to 100.
By using equation 1 as a model function and having
the computer randomly generate the necessary
parameters within the previously set limiting values,
some simulated primary background shapes were
obtained. Figure 8 represents a number of the simulated
shapes of the primary background.
2.4 Completing the simulation
To complete the simulation we also had to gather
information about the characteristic peaks of elements.
For that purpose the characteristic elemental AES spectra
of Al, C, Co, Cu, Fe, Au, Ni, O, Si, Ag, Ti, and Va were
gathered from Compro10, a database easily accessible on
the net.
Generally, we suppose that the measured spectra are
constructed according to equation 6.
S f a s P Ni i i= + +( ) B (6)
where S represents the measured AES spectra, ƒi is
some non-linear function binding the standardized
spectra into the composed elements of the spectra, ai is
the amount of i-th element in the spectra, si the i-th pure
element represented in the AES spectra, PB is the
primary background, and N is the noise.
In the case of synthesis of the artificial AES spectra,
the non-linear dependence of the various standardized
spectra is assumed to be linear, thus producing equation
7.
S a s P Ni
i
M
i= + +
=
∑ ( )
1
B (7)
where the symbol M denotes the number of all the
spectra in the database.
For the purpose of creating the so-called standard
spectra database, the elementary spectra were analyzed.
Figure 9 shows the spectra of pure Fe.
From visually inspecting Figure 9 we can see that
the AES spectra consist of the three distinctive elements:
1. The primary background,
2. The peaks base,
3. The peaks.
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
42 Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46
Figure 7: The construction of the background function
Slika 7: Konstrukcija funkcije, ki opisuje ozadje spektra
Figure 8: The randomly chosen backgrounds for the simulated AES
spectra
Slika 8: Naklju~no izbrana ozadja za simulirane Augerjeve spektre
Figure 10 illustrates these distinctive parts. For the
purpose of extraction of the three spectra elements,
several steps were introduced as follows:
1. Two points P1 and P2 (Figure 10) were selected.
They are common to all three curves.
2. The line that represents the primary background
through P1 and P2 was constructed. We are aware that
in the measured spectra the primary background is
nowhere linear, but the use of the line is close enough
for the intended purpose.
3. On the part of the AES spectra between P1 and P2 the
set of points describing the base of the peaks was
selected (Figure 10). Typically, the number of selec-
ted points was up to 10.
The shape of the base function was approximated by
the four-layer neural network (configuration 1-4-8-1).
The approximated function was stored in a table
containing the Kinetic energy and count values for
the area between P1 and P2 (Figure 11).
4. Once the base function is known it is subtracted from
the original spectra, thus obtaining the shape of the
extracted peaks (Figure 12).
2.5 The virtual AES spectra generator
The system for the random generation of AES spectra
was developed under Microsoft Excel using Visual Basic
for Applications and consists of six distinctive elements:
1. The generator Console
2. The Files control center
3. The Primary background generator
4. The Noise generator
5. The standard elements peaks database
6. The standard elements peak base database.
2.5.1 The generator Console
The generator console is the central part of the AES
spectra generator. It provides the basic constituent data
regarding the generated AES spectra. The following are
data provided on the Console datasheet:
• The generator mode of operation. Two modes are
currently possible: "ONE AES" producing only one
randomly generated AES spectra without the saving
to the file; "SAVE TO FILE" option, which generates
random AES spectra and stores the data to the files.
• The number of elements provided in the standard
spectra database. Currently, the standard spectra
database contains 11 spectra of standard elements.
Additional spectra are very easily included in the
database.
• The number of actual AES spectra constituents.
Although the number of standard spectra is provided,
the process of random AES spectra generation can be
limited to the lower number of actual constituents.
The number is specified by the user.
• The graphical representation of the randomly gene-
rated spectra. The generated spectra are joined with
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46 43
Figure 12: The extracted peaks of Fe AES spectra
Slika 12: Izlo~eni vrhovi za Augerjev spekter Fe
Figure 9: The AES spectrum of Fe
Slika 9: Augerjev spekter Fe
Figure 10: The AES spectra elements
Slika 10: Elementi Augerjevega spektra
Figure 11: The approximated function of the base of the peaks for the
Fe AES spectrum
Slika 11: Aproksimirana funkcija podlage vrhov za Augerjev spekter Fe
the primary background as well as with the noise.
The result of the multiple random generation is
graphically presented. The additional information is
provided on the table where the amounts of the
constituents are given (Figure 13).
2.5.2 The Files control center
The Files control center defines the way in which the
randomly generated AES spectra are stored. The user
defines the name of the files where the spectra are to be
stored. The provided name is joined by the subsequent
number of the generated spectra and stored in the
directory, the name of which has already been provided
by the user. It is also necessary to specify the number of
generated AES spectra that are to be saved in the
separate file.
The data contained in each file are:
• The sum of all the peaks contained in the generated
AES spectra
• The sum of all the base elements that are part of each
standard spectra
• The randomly generated primary background
• The noise
• The AES spectra construction data
2.5.3 The Primary background generator
The primary background generator controls the
process of AES spectra primary background generation.
The user must define the intervals for the primary
background parameters. The AES spectra generator uses
the given intervals to randomly re-shape the basic
primary background equation. Figure 14 represents the
graphical user interface for the AES spectra primary
background generation.
2.5.4 The Noise generator
The noise generator provides the random noise that is
to be added to the AES spectra. The user can define the
upper limit for the AES spectra noise. The noise
generator then forms the noise according to the given
boundary. The generated noise is bipolar, providing that
the AES spectra fluctuations affect the spectra both as
the addition and subtraction to the signal (Figure 15). In
the literature, the noise level is understood to be in a
non-linear dependence with the kinetic energy. In the
first stage of the AES signal modeling we opted to use
the linear dependence of the noise amplitude over the
kinetic energies used.15
2.5.5 The standard-elements peaks database
The standard-elements peaks database contains the
data on the pure – standard – elements that are the basis
of the AES spectra generation. So far, 11 elements have
been added to the database. This database contains only
the peaks data (Figure 16). Other parts are stored
elsewhere.
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
44 Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46
Figure 15: The AES spectra noise
Slika 15: Dodan {um Augerjevemu spektru
Figure 13: The AES spectra generator main console
Slika 13: Osrednji del generatorja Augerjevih spektrov
Figure 14: The AES spectra primary background generation
Slika 14: Generiranje osnovnega ozadja Augerjevega spektra
2.5.6 The standard-elements peaks base database.
The standard-elements peaks base database is
formatted in exactly the same way as in the case of the
standard-elements peaks database.
3 RESULTS AND DISCUSSION
Comparing Figure 17 with Figure 2 it is obvious that
the randomly simulated AES spectra adequately resem-
ble the measured AES spectra. The variations of the
backgrounds are sufficient to produce a set of quite
different spectra that will allow a thorough analysis of
the background- and noise-removal algorithms.
The process of construction of the complete AES
spectra starts with the selection of the parameters that
control the process. The steps are the following:
1.) The number of constituent elements that are to be
joined in the spectra is selected by the user. In our
case the user can select 12 different elements. This
number can be easily changed by adding the new
spectra for additional elements.
2.) The random generator selects the constituent ele-
ments (from the number provided by the user in the
previous step) and sets the relative amplitude for each
selected element (parameter ai from equation 6).
3.) The primary background is created.
4.) One by one the spectra of the selected elements are
added to the primary background.
5.) The randomly generated noise is added.
6.) The normalized form of the simulated spectra is
scaled to receive the "realistic" count values.
7.) The created spectrum is stored in a separate file,
which will later represent the testing and validation
ground for different background- and noise-removal
techniques and algorithms. Each file contains
detailed data on how the concrete spectra were
constructed, so the background and the noise are
known in detail. Therefore, a very exact basis for the
further analysis is set.
Figure 17 represents two cases of simulated AES
spectra, with the corresponding data given alongside in
the table.
It is very important to note that the simulated AES
spectra do not necessarily have to represent the spectra
of real materials. The only condition that has to be
fulfilled is for them to resemble the real AES spectra,
i.e., to consist of the same constituent parts.
4 CONCLUSIONS
AES is a technique capable of deriving information
from the topmost atomic layers of the surface. It is able
to identify a very wide range of elements present at solid
surfaces, and sometimes it can be used to obtain
information about the chemical environment and
bonding of surface species.16 As knowledge about the
properties of surfaces and their composition is increasing
in importance, so is the need to get accurate information
about them through techniques such as AES. An
inevitable step after performing a measurement with this
technique is the data processing that follows in order to
harvest the desired information out of the spectra already
taken. Background and noise are two undesired elements
that are present alongside the characteristic peaks of
elements in measured AES spectra. The background
must be removed and the noise reduced to continue with
the proper treatment of the data, but being at unknown
levels in the raw data and knowing that any data
treatment inevitably alters them to some extent, there is
uncertainty as to the quality of the outgoing information.
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46 45
Figure 17: Two cases of simulated AES spectra with their correspon-
ding data
Slika 17: Primera simuliranega Augerjevega spektra
Figure 16: The standard-elements peaks database
Slika 16: Podatkovna baza standardnih spektrov
The highlight of this work is the ability to have prior
knowledge about the levels of background and noise
before the processing of the spectra by various means
takes place. Having this in mind, after the processing is
done the pre-processed and the post-processed data can
be compared to learn how much error the background-
subtraction and noise-reduction techniques are intro-
ducing into the results that we read. By correcting for
this factor the accuracy of the obtained information from
the background-removal and noise-reduction tools is
greatly improved.
It is of utmost importance to know that the presented
work serves only one purpose, i.e., to provide the envi-
ronment where various algorithms for noise reduction
and background removal can be tested and evaluated.
Such algorithms can afterwards be used on measured
spectra, but with a thorough understanding of the con-
sequences they have for the phase of data processing.
5 REFERENCES
1 F. A. Settle, (editor) Handbook of instrumental techniques for ana-
lytical chemistry, Prentice Hall, New Jersey, 1997
2 C. R. Brundle, C. A. Evans, Jr., S. Wilson, Encyclopedia of Materials
Characterization. Manning Publications & Reed Publishing, 1992
3 C. J. Powella, A. Jablonski, W. S. M. Werner, W. Smekal, Characte-
rization of thin films on the nanometer scale by Auger electron
spectroscopy and X-ray photoelectron spectroscopy, Applied Surface
Science, 239 (2005), 470–480
4 J. T. Grant, D. Briggs, Surface analysis by Auger and X-ray Photo-
electron Spectroscopy, Chichester, IM Publications 2003
5 IUPAC. Compendium of chemical terminology, 2nd ed. (the "Gold
Book"). Compiled by A. D. McNaught and A. Wilkinson. Blackwell
Scientific Publications, Oxford 1997
6 M. Thompson, M. D. Baker, A. Christie, J. F. Tyson, Auger Electron
Spectroscopy, Chichester: John Wiley & Sons, 1985
7 G. Gauglitz, T. Vo-Dinh (editors), Handbook of Spectroscopy.
WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 2003
8 D. Briggs and M. P. Seah (editors), Practical surface analysis, John
Willey & Sons, West Sussex 1990
9 J. A. Venables, Introduction to surface and thin film processes, Cam-
bridge University Press, 2000
10 Microlab 310-F, Operators Manual, V.G. Scientific, 1997
11 V. Mazet, C. Carteret, D. Brie, J. Idier, B. Humbert, Background
removal from spectra by designing and minimising a non-quadratic
cost function, Chemometrics and Intelligent Laboratory Systems, 76
(2005) 2, 121–133
12 J. T. Grant, Background subtraction techniques in surface analysis, J.
Vac. Sci. Technol. A, 2 (1984), 2
13 R. M. Golden, Mathematical Methods for Neural Network Analysis.
MIT Press, Cambridge MA 1996
14 I. Beli~, Neural networks and modelling in vacuum science, Vacuum.
80 (2006), 1107–1122
15 M. F. Koenig, J. T. Grant, Signal-to-noise measurements in X-ray
photoelectron spectroscopy, Surface and Interface Analysis, 7 (1985)
5, 217
16 J. T. Grant, Surface analysis with Auger electron spectroscopy, App-
lications of Surface Science, 13 (1982), 35–62
B. PONIKU et al.: THE MODELING OF AUGER SPECTRA
46 Materiali in tehnologije / Materials and technology 45 (2011) 1, 39–46
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
MODELLING OF THE DIRECTIONAL SOLIDIFICATION
OF A LEADED RED BRASS FLANGE
MODELIRANJE USMERJENEGA STRJEVANJA PRIROBNICE IZ
RDE^E SVIN^EVE MEDENINE
Vladimir Grozdani}
University of Zagreb, Metallurgical Faculty, Aleja narodnih heroja 3, 44103 Sisak, Croatia
Prejem rokopisa – received: 2010-04-21; sprejem za objavo – accepted for publication: 2010-11-02
A mathematical model and a numerical simulation of the directional solidification of a leaded red brass flange are presented.
The mathematical model is based on physically realistic assumptions and it is solved by the finite-difference method, implicit
alternating directions method. This method has great accuracy and is of second order with regard to the approximation of time
and space. The initial conditions are an analytical solution of the heat conduction equation in the case of the contact of two
semi-infinite media. The latent heat of fusion incorporated into the equation for the specific heat capacity of metal, and
temperature dependent thermophysical properties of all materials in the system mould–casting–core–chill, enables us to
accurately numerically represent the flange solidification, which is casting and a common example in foundry practice. The
simulation of the directional solidification is a modern and scientific way of pointing to casting points in which the defect
occurrence is possible, and how to prevent them using the chill.
Keywords: mathematical model, directional solidification, leaded red brass flange, graphite chill
Predstavljena sta matemati~ni model in numeri~na simulacija usmerjenega strjevanja prirobnice iz rde~e svin~eve medenine.
Podlaga matemati~nega modela so realne fizikalne predpostavke, model pa je razvit z metodo kon~nih razlik, implicitno metodo
z alternativno smerjo. Ta metoda je zelo natan~na in je drugega reda glede na pribli`ek ~asa in prostora. Za~etni pogoji so
analiti~ne re{itve ena~be za toplotno prevodnost za primer kontakta dveh semineskon~nih medijev. Latentna talilna toplota, ki je
uporabljena za specifi~no toplotno kapaciteto kovine in termofizikalne lastnosti vseh materialov v sistemu forma – litje – jedro –
kokila omogo~ajo, da se numeri~no modelira strjevanje prirobnice, torej litje, kar je splo{en primer dela v livarni. Simulacija
usmerjenega strjevanja je moderna metoda, da se znanstveno opredeli mesta ulitka, kjer lahko nastajajo napake in kako te
napake s kokilo prepre~iti.
Klju~ne besede: matemati~ni model, usmerjeno strjevanje, prirobnica, svin~eva rde~a medenina, grafitna kokila
1 INTRODUCTION
The directional solidification of castings of a given
geometry may be obtained using risers or chills. Risers
are reservoirs of molten metal and feed the casting in the
liquid state. In the case of classical risers, only 17 pct. of
the initial volume of the riser is available for the feeding
of the casting 1. In contrast to risers, chills better lead
away the heat and in this way speed up the solidification.
Usually, they are used when the placement of risers is
impossible. However, in the case of nonferrous metals
this is not a rule. In this way, for example, in the inve-
stigated system of a leaded red brass flange, directional
solidification is obtained using a graphite chill. With the
placement of an appropriate chill, the hot spot is shifted
in areas of the casting which are later machined. This is
desirable because the hot spot solidifies last, and in their
place the occurrence of casting defects (e.g., shrinkage
cavity, porosity etc.) is possible. The investigated system
is complex because it consists of a flange, a mould core
and a chill, and it is shown in Figure 1. For a given
system the mathematical model of the solidification is
formulated and investigated.
2 MATHEMATICAL MODEL
In the development of the mathematical model, the
following partial differential equation of heat flow
corresponding to the flange (Figure 1) 2 should be
solved:
∂
∂
∂
∂
∂
∂
∂
∂
T
t
a
T
r r
T
r
T
z
= + +
⎛
⎝
⎜
⎞
⎠
⎟
2
2
2
2
1
(1)
Since for the horizontal axis of the flange symmetry r
= 0, equation (1) should be modified according to L’
Hospital’s rule considering the equation:
Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53 47
UDK 621.74:519.68 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)47(2011)
Figure 1: Flange and graphite chill with the corresponding dimensions
Slika 1: Prirobnica in grafitna kokila z dimenzijami
∂
∂
∂
∂
∂
∂
T
t
a
T
r
T
z
= +
⎛
⎝
⎜
⎞
⎠
⎟2
2
2
2
2 (2)
Initial conditions
The mould temperature and the temperature of its
outer side are equal to Ts, whereas the temperature of the
metal is equal to the casting temperature TL. The initial
temperature at the mould/casting boundary interface can
be obtained by solving Fourier’s differential equation for
heat flow through the contact area of two semi-infinite
media:
T T
T T
K
K
a
a
i s
L s
K
m
m
k
= +
−
+1
(3)
The derivation of eq. (3) is given in Appendix 2.
Boundary conditions
The outer mould surface maintains a constant tem-
perature Ts. For the contacts mould/metal, metal/core,
mould/core, metal/chill, mould/chill and chill/core area
there are continuous heat flows for which the boundary
condition of the fourth type holds 3:
Km
∂
∂
T
n
m
= KK
∂
∂
T
n
K
(4)
Km
∂
∂
T
n
m
= Kj
∂
∂
T
n
j
(5)
Km
∂
∂
T
n
m
= Kh
∂
∂
T
n
h
(6)
Kk
∂
∂
T
n
K
= Kj
∂
∂
T
n
j
(7)
Kh
∂
∂
T
n
h
= Kk
∂
∂
T
n
k
(8)
Kh
∂
∂
T
n
h
= Kj
∂
∂
T
n
j
(9)
Thermophysical properties of the material
It has been assumed that the thermal properties of the
mould, metal, core and chill are temperature dependent 4,
which is as shown in Figures 2 to 9.
48 Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
Figure 2: Temperature-dependent thermal conductivity of the leaded
red brass flange
Slika 2: Odvisnost toplotne prevodnosti od temperature za prirobnico
iz svin~eve rde~e medenine
Figure 4: Temperature-dependent thermal conductivity of the graphite
chill
Slika 4: Odvisnost toplotne prevodnosti od temperature za grafitno
kokilo
Figure 3: Temperature-dependent specific heat capacity of the leaded
red brass flange
Slika 3: Odvisnost specifi~ne toplotne kapacitete od temperature za
prirobnico iz svin~eve rde~e medenine
Figure 5: Temperature-dependent specific heat capacity of the graphi-
te chill
Slika 5: Odvisnost specifi~ne toplotne kapacitete od temperature za
grafitno kokilo
3 IMPLICIT ALTERNATING DIRECTION
METHOD
Differential heat-flow equations (1) and (2) with the
corresponding initial and boundary conditions have been
numerically solved using the implicit alternating
direction method 5,6. The method utilises the division of
the time interval into two steps.
In the first half of the time interval the equation is
solved implicitly for z and explicitly for the r direction.
The procedure is reversed in the second half of the time
interval.
Consequently, for the differential equation (1) and the
first half of the time interval t/2 we have:
T T T
r
i,j
n
i,j
n
i,j
n
− +− +1 1
2
2
( )Δ
+
T T
r r
i,j
n
i,j
n
j
+ −−
⋅
1 1
2 Δ
+
+
T T T
z
i ,j
n
i,j
n
i ,j
n
− +− +1 1
2
2
( )Δ
=
1
2a
T T
t/i j n
i,j i,j
n
, ,
* −
Δ
(10)
Whereas for the second t/2 we obtain:
T T T
r
i,j
n
i,j
n
i,j
n
−
+
+
+− +1
1
1
1
2
2+1
( )Δ
+
T T
r r
i,j
n
i,j
n
j
+
+
−
+−
⋅
1
1
1
1
2 Δ
+
+
T T T
z
i ,j i,j i ,j− +− +1 1
2
2* * *
( )Δ
=
1
2a
T T
t/i j n
i,j i,j
, ,
*n +1 −
Δ
(11)
The numerical solution for the differential equation
(2) of the heat flow for the first t/2 is:
4
2 1
2
T T
r
i,
n
i,
n−
( )Δ
+
T T T
z
i , i, i ,− +− +1 1 1 1 1
2
2* * *
( )Δ
=
=
1
21
1 1
a
T T
t/i n
i, i,
n
, ,
* −
Δ
(12)
and for the second t/2:
4
2 1
1
2
T T
r
i,
n
i,
n+1 − +
( )Δ
+
T T T
z
i , i, i ,− +− +1 1 1 1 1
2
2* * *
( )Δ
=
=
1
21
1 1
a
T T
t/i n
i, i,
, ,
*n +1 −
Δ
(13)
The use of the implicit alternating direction method
results in a system of simultaneous linear algebraic
equations with the variables v1, v2,…, vn of tri-diagonal
form:
b1 v1 + c 1 v2 = d1
a2 v1 + b2 v2 + c2 v3 = d2
a3 v2 + b3 v3 + c4 v4 = d3
………………………
ai vi-1 + bi vi + ci vi+1 = di (14)
………………………
aN–1 vN–2 bN–1 vN–1 + cN–1 vN = dN–1
aNvN–1 + bN vN = dN
The special efficient algorithm for solving the
tri-diagonal system of equations is 7:
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53 49
Figure 9: Temperature-dependent temperature diffusivity of the core.
Slika 9: Odvisnost toplotne difuzivnosti od temperature za jedro
Figure 8: Temperature-dependent thermal conductivity of the core
Slika 8: Odvisnost toplotne prevodnosti od temperature za jedro
Figure 7: Temperature-dependent temperature diffusivity of the mould
Slika 7: Odvisnost toplotne difuzivnosti od temperature za formo
Figure 6: Temperature-dependent thermal conductivity of the mould
Slika 6: Odvisnost toplotne prevodnosti od temperature za formo
vN = N (15)
vi = i –
c vi i
i
+1

, i = n – 1, n – 2, ..., 1 (16)
where  and  are calculated from recursive formulas
1 = b1 (17)
1 = d1/1 (18)
i = bi –
a ci i
i
−
−
1
1
, i = 2,3,...,n (19)
i = –
d ai i i
i
− −

1
, i = 2,3,...,n (20)
The tri-diagonal coefficients that give the algorithm
of the flange solidification in a sand mould are presented
in Appendix 3. Based on the presented algorithm a
computer program was written in FORTRAN 77 and
solved on an Athlon AMD 64 computer.
4 FLOW DIAGRAM
A detailed flow diagram is shown in Figure 10a and
10b.
The main feature of the program is the use of two
temperature matrixes, namely T and T*. The first matrix
contains the temperatures at the start and the end of the
time step, and the second contains the temperature at the
end of first t/2. Initial values are assigned to the
program variables and constants. The program module
TYP provides for the initial values of the temperatures,
of particular net points as well as for the standardization
of all the points in the mould, the casting, the chill core
and their boundary interfaces. The module PRINT 1
prints out the initial temperature distribution. The system
of tri-diagonal equations is then solved firstly, row by
row (module ROWS), and then column by column
(module COLS). The results are periodically printed
over the whole geometry of the casting, the mould, the
chill and the core (module PRINT 1) or over the casting
geometry only (module PRINT 2) until the prescribed
time tmax.
5 DISCUSSION
Leaded red brass (alloy C83600) has the following
composition: 85 % Cu, 5 % Sn, 5 % Pb and 5 % Zn 8.
The alloy C83600 is used not only for flanges, but also
for valves, pipe fittings, water pumps’ meter housings
and impellers, small gears, high-quality plumbing goods,
statuary and plaques. These applications are possible
because of its corrosion resistance, machinability,
strength, bearing properties, colour and the excellent
castability of the alloy.
The simulation of the solidification of the leaded red
brass flange in a sand mould is carried out by the space
steps z = 15 mm and r = 7 mm and the time step t =
5 s to tmax = 600 s.
The casting temperature was 1180 °C and the initial
temperature of the mould, the core and the chill was 25
°C. The initial temperature at the sand/casting interface
was 1057 °C, on the casting/core interface it was 1013
°C, and at the casting/chill interface it was 577 °C. On
the basis of successive temperature print outs for particu-
lar net points, a solidification time of 367 s was obtained.
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
50 Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53
Figure 10: Flow diagram
Slika 10: Flow diagram
Using the isosolidus (854 °C) shift, as seen in Figure 11,
it can be concluded that the potential defect site is
obviously the point of the final solidification of the heat
centre, i.e. in the vicinity of the core.
The accuracy of the simulation is limited, depending
on the assumptions used in the mathematical model, the
method of the numerical analysis and the values of ther-
mophysical material properties used. Several assump-
tions were used in the elaboration of the mathematical
model. The most important are as follows: the complete
heat-transfer rate, a casting temperature equal to the
initial temperature of the metal in the mould, and a
mould/casting interface with perfect thermal contact.
The first assumption restrains the analysis to the
mould-casting-core-chill system with heat conduction
only, i.e., partial heat flows associated with the mould
and core moisture are not considered. The second
assumption is the simplification introduced to avoid a
complex consideration of the metal flow through the gate
system and the mould cavity and the matching heat
transfer. The assumption of a perfect thermal contact on
the interface is acceptable because there is only a partial
appearance of a gaseous clearance, and therefore in the
mathematical formulation the boundary condition of the
fourth kind is usually taken as valid. The partial
differential equation for the heat flow is solved using a
numerical method of finite difference – the implicit
alternating direction method, which was chosen because
of its high accuracy during the approximation of both
time and space. Insufficient knowledge about the ther-
mophysical properties of the material, especially at high
temperatures, has a strong influence on the simulation of
the solidification. It holds especially with respect to the
thermal properties of the mould and the core material,
which can be determined only experimentally. Moreover,
the values for the thermal properties at higher tempe-
ratures show a considerable a dissipation.
6 CONCLUSIONS
A numerical simulation of the directional solidifi-
cation of a leaded red brass flange was carried out on the
basis of a suitable mathematical model. The complex
model system is composed of four materials: the mould,
the core, the chill and the casting of comparatively
complicated geometry. The mathematical model of soli-
dification was developed assuming thermal conduction
as only heat flow in the system, which is considered as a
physically real assumption. A differential equation for
the heat flow suited to the flange’s geometry was
modified and numerically solved by the use of implicit
alternating direction method. The temperature depen-
dence of the thermophysical material’s properties was
taken into account. Based on the obtained algorithm a
computer program written in FORTRAN 77 for an
Athlon AMD 64 computer was used for the simulation of
the solidification. It was determined that the complete
solidification takes 367 seconds. The progress of the
solidification as well as the hot spot, i.e., the site of any
potential shrinkage cavity, can be determined by a shift
of isosolidus.
7 REFERENCES
1 V. Grozdani}, Metalurgija 25 (1986) 2, 51
2 J. H. Leinhard IV, J. H. Leinhard V, A heat transfer textbook, 3rd ed.,
Phlogiston Press, Cambridge, 2006
3 V. P. Isachenko, V. A. Osipova, A. S. Sukomel, Heat transfer, 3 rd
ed., Mir Publishers, Moscow 1980
4 R. D. Pehlke, A. Jeyarajan, H. Wada, Summary of thermal properties
for casting alloys and mold materials, University of Michigan, Ann
Arbor, 1982
5 J. Douglas, H. H. Rachford, Trans. Amer. Math. Soc., 82 (1956), 421
6 V. Grozdani}, Materiali in tehnologije 43 (2009) 5, 233
7 B. Carnahan, H. A. Luther, J. O. Wilkes, Applied numerical methods,
John Wiley, New York, 1969
8 AFS, Casting coppar-base alloys, American Foundrymen’s Society,
Des Plaines, 1984
9 M. R. Spiegel, Laplace transforms, McGraw-Hill, New York, 1965
10 M. Abramowitz, I. A. Stegun (eds.), Handbook of mathematical
functions with formulas, graphs and mathematical tables, National
Bureau of Standards, Washington, 1964
Appendix 1
Abbreviations used:
a –- temperature conductivity
ai, bi, ci, di – coefficients adjoining unknowns in a tri-dia-
gonal system of algebraic equations
cp – specific heat capacity at constant pressure
k – thermal conductivity
n – vertical direction
r – space coordinate
t – time
T – temperature
vi – unknown in system of simultaneous equations
z – space coordinate
Appendix 2
The following partial differential equation of heat
flow with the appropriate initial and boundary conditions
must be solved to derive the equation for the temperature
distribution at the mould/metal interface:
Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53 51
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
Figure 11: Progress of the isosolidus (854 °C) after (150, 240 and
300) s
Slika 11: Napredovanje izotr. (854 °C) po (150, 240 in 300) s
∂
∂
∂
∂
T
t
a
T
x
=
2
2 (21)
T(x,0) = Ts
T(0,t) = Ti (22)
|T(x,t)| < M
where M is a positive real constant.
The equation of heat flow should be solved for the
case of the contact of two semi-infinite media. The
partial differential equation (21) must be transformed
into a common differential equation by the Laplace
transform defined as 9:
{ }L T x t x s T x t tst( , ) ( , ) ( , )= = −
∞
∫Q e d
0
(23)
Equation (21) in the Laplace region has the from:
d
dx a
s
a
Ts
2
2
1 1Q
Q− = − (24)
The solution of (24) has the form:
Q( , ) ( ) exp( / ) ( ) exp( / )x s c s x s a c s x s a
T
s
s= + − +1 2 (25)
Because of temperature limitations, |T(x,t)| < M,
c1 = 0 for x, i.e.
Q( , ) exp( / )x s c x s a
T
s
s= − +2 (26)
For the boundary condition T(0,t) = Ti, i.e.
Q( , )0 2s c
T
s
T
s
s i= + = , c
s
T Ts2 1
1
= −( )
The final result in the Laplace region is:
Q( , ) exp( / )x s
T T
s
x s a
T
s
i s s=
−
− + (27)
By passing from the Laplace region into real space,
we have:
T x t T T
x
at
Ts s( , ) ( )= −
⎛
⎝
⎜ ⎞
⎠
⎟ +1 2
erfc (28)
respectively
( )
( )
T T
T T
x
at
i
s i
−
−
=
⎛
⎝
⎜ ⎞
⎠
⎟erf
2
(29)
where the error function is defined as 10:
erf ( e dt uu
t
) = −∫
2 2
0
(30)
The temperature gradient along the x-axis is:
∂
∂
T
x
T T
at
x / at
s i=
−
−

exp( )2 4 (31)
respectively
∂
∂
T
x
T T
atx
s i⎛
⎝
⎜ ⎞
⎠
⎟ =
−
=0 
(32)
Two semi-infinite media (mould and metal) are in
interfacial contact and the boundary condition of the
fourth kind is valid:
−
⎛
⎝
⎜ ⎞
⎠
⎟ = −
⎛
⎝
⎜ ⎞
⎠
⎟
= =
k
T
x
k
T
xk
k
x
m
m
x
∂
∂
∂
∂
0 0
(33)
By including the proper temperature gradients:
k
T T
a t
k
T T
a tk
i s
k
m
L i
m
−
=
−
 
(34)
Finally, the initial temperature distribution on the
boundary mould/metal surface is obtained:
T T
T T
K
K
a
a
i
L s
K
m
m
k
= +
−
+
s
1
(35)
Appendix 3
The constants, which in tri-diagonal coefficients, are:
p
a t
r1 22
=
Δ
Δ( )
p
a t
r rj
2 4
=
⋅
Δ
Δ
p3 = p1 – p2
p4 = p1 + p2
p
t k k
c r5 22
=
+Δ
Δ
( )
( )
A B
p
t k k
cr rj
6 4
=
+
⋅
Δ
Δ
( )A B
c
k
a
k
a
= +A
A
B
B
q
a t
z1 22
=
Δ
Δ( )
q
k t
c z2 2
= A
Δ
Δ( )
q
k t
c z3 2
= B
Δ
Δ( )
q
t k k
c z4 22
=
+Δ
Δ
( )
( )
A B
q
k t
c r5 2
= A
Δ
Δ( )
q
k t
c r6 2
= B
Δ
Δ( )
Tri-diagonal coefficients
1. Point (i,j) in the mould, metal core or chill
– first t/2:
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
52 Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53
ai = ci = –q1
bi = 1 + 2q1 (36)
d p T p T p Ti i j
n
i j
n
i j
n= + − +− +3 1 1 4 11 2, , ,( )
– second t/2:
aj = –p3
bj = 1 + 2p1
cj = –p4 (37)
d q T q T q Tj i j i j i j= + − +− +1 1 1 1 11 2,
*
,
*
,
*( )
2. Point (i,j) on the boundary surface parallel to the r
axis separating the material A (left) and B (right)
– first t/2:
a qi = − 2
b q qi = + +1 2 3
c qi = − 3 (38)
d p p T p T p p Ti i j
n
i j
n
i j
n= − + − + +− +( ) ( ) ( ), , ,5 6 1 5 5 6 11 2
– second t/2:
a p pj = − −( )5 6
b pj = +1 2 5
c p pj = − −( )5 6 (39)
d q T q q T q Tj i j i j i j= + − − +− +2 1 2 3 3 11,
*
,
*
,
*( )
3. Point (i,j) on the boundary surface parallel to the z
axis separating the material A (down) and B (up).
– first t/2:
a c qi i= = − 4
b qi = +1 2 4 (40)
d q q T p T q q Ti i j
n
i j
n
i j
n= − + − + +− +( ) ( ) ( ), , ,5 6 1 5 5 6 11 2
– second t/2:
a p qj = −6 5
b pj = +1 5
c q pj = − −( )5 6 (41)
d q T q T q Tj i j i j i j= + − +− +4 1 4 4 11 2,
*
,
*
,
*( )
4. Point (i,1) out of the boundary surface
– first t/2:
a c qi i= = − 1
b qi = +1 2 1 (42)
d p T p Ti i
n
i
n= − +( ) , ,1 4 41 1 1 2
– second t/2:
b pj = +1 4 1
c pj = 4 1 (43)
d q T q T q Tj i i i= + − +− +1 1 1 1 1 1 1 11 2,
*
,
*
,
*( )
5. Point (i,1) on the boundary surface, which sepa-
rates the material A (left) and B (right)
– first t/2:
a qi = − 2
b qi = +2 14
c qi = − 3 (44)
d p T p Ti i
n
i
n= − +( ) , ,1 4 45 1 5 2
– second t/2:
b pj = +4 15
c pj = −4 5 (45)
d q T q T q Tj i i i= + − +− +2 1 1 4 1 3 1 11 2,
*
,
*
,
*( )
V. GROZDANI]: MODELLING OF THE DIRECTIONAL SOLIDIFICATION OF A LEADED RED BRASS FLANGE
Materiali in tehnologije / Materials and technology 45 (2011) 1, 47–53 53

A. BYTYQI et al.: CHARACTERIZATION OF THE INCLUSIONS IN SPRING STEEL ...
CHARACTERIZATION OF THE INCLUSIONS IN SPRING
STEEL USING LIGHT MICROSCOPY AND SCANNING
ELECTRON MICROSCOPY
KARAKTERIZACIJA VKLJU^KOV V VZMETNIH JEKLIH S
SVETLOBNO IN VRSTI^NO ELEKTRONSKO MIKROSKOPIJO
Arsim Bytyqi, Nu{a Puk{i~, Monika Jenko, Matja` Godec
Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia
[tore Steel Company, d. o. o., @elezarska cesta 3, 3220 [tore, Slovenia
arsim.bytyqi@imt.si
Prejem rokopisa – received: 2010-04-08; sprejem za objavo – accepted for publication: 2011-01-11
The quality of high-grade steel depends mainly on the size, distribution and composition of the inclusions in the steel. A
good-quality product can be obtained by controlling the size, distribution and chemical composition of the inclusions. This
study was limited to non-metallic inclusions, mainly compounds of metallic elements and carbon, nitrogen, oxygen or sulfur.
The inclusions in spring steel were investigated with light microscopy and scanning electron microscopy. Generalized extreme
value theory was applied to the size distribution of the inclusions. The types of inclusions were evaluated and classified in the
majority as having a complex non-metallic composition. The experimental results show the possibility to define permissible
limits for non-metallic inclusions with a minimal negative influence on the steel’s properties.
Keywords: non-metallic inclusions, SEM, EDS, spring steel, GEV
Glavni dejavnik, od katerega je odvisna visoka kakovost jekla, je velikost, razporeditev in sestava nekovinskih vklju~kov. Z
nadzorom velikosti, porazdelitve in kemijske sestave vklju~kov lahko dobimo kakovosten proizvod. Preiskava je bila omejena
na nekovinske vklju~ke, ki so v glavnem spojine kovinskih elementov z ogljikom, du{ikom, kisikom ali `veplom. V prispevku
sta karakterizacija nekovinskih vklju~kov in ~istost vzmetnega jekla opredeljeni s preiskavami v svetlobnem mikroskopu in v
vrsti~nem elektronskem mikroskopu. Porazdelitev velikosti vklju~kov je bila opredeljena s posplo{eno teorijo ekstremnih
vrednosti. Iz rezultatov je bilo mogo~e oceniti naravo vklju~kov, in sicer je bila ve~ina vklju~kov kompleksnih, zra{~enih z
enofaznimi vklju~ki. Rezultati eksperimentalnih postopkov ka`ejo na mo`nost dolo~itve dopustne meje nekovinskih vklju~kov z
minimalnim negativnim vplivom na lastnosti jekla.
Klju~ne besede: nekovinski vklju~ki, SEM, EDS, vzmetno jeklo, posplo{ena teorija ekstremnih vrednosti
1 INTRODUCTION
Non-metallic inclusions are formed during the steel’s
production process and have a great impact on the
strength, plasticity, fracture toughness, fatigue strength
and other properties1. Producers can achieve better
control of the processing and improve the quality of pro-
ducts by a characterization of the individual inclusions.
The size and distribution of inclusions are particularly
important, because large macro-inclusions are the most
harmful for the mechanical properties of steels2. How-
ever, large inclusions are difficult to inspect because of
their low occurrence rate. Many experimental results
show that the largest inclusions are the most probable
fracture origins in a given volume of material3. There-
fore, it is important that steelmakers learn to better
control the inclusions’ characteristics. The characteri-
zation of non-metallic inclusions in steel in terms of
number, size, shape and chemical composition is thus an
essential requirement for metallurgical process techno-
logy. It is necessary to measure the size and spatial
distributions of the inclusions for a correlation with the
mechanical properties4. Different statistical models can
be used to characterize the size distribution of inclusions
and a lot of effort was expended in recent years to
predict the limits for inclusion size or mechanical
properties by extrapolating the data5.
However, despite the major advances in inclusion
control, there is still no rapid and accurate method for
determining the type, size and number of inclusions
present in a steel sample. The scope of possible changes
in the production and characterization processes requires
scientific studies.
2 EXPERIMENTAL
High-purity steel was produced in [tore Steel, d. o. o.,
in compliance with the EN 51CrV4 standard. The
chemical composition is given in Table 1. Five samples
with dimensions of 20 mm × 28 mm × 10 mm were cut
from the slab according to the ISO 4967 standard.
Samples were prepared with standard metallographic
techniques, then examined by light microscopy and by
scanning electron microscopy (SEM). The advantage of
optical microscopy is the ability to examine a large area
in a short time, but the information we are able to gather
about inclusions is limited. Imaging at higher magni-
fication and microchemical analyses were made using
Materiali in tehnologije / Materials and technology 45 (2011) 1, 55–59 55
UDK 669.14.018.252:620.18:543.428.2 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)55(2011)
SEM with an energy-dispersive x-ray spectroscopy
(EDS) capability.
The samples were examined with optical microscopy
in order to determine the quantity and size of the
inclusions. A total of 554 images of polished samples at
100x magnification with a total area of 150 mm2 were
acquired. Image processing and measurements of the
inclusion cross-section areas were made with the
analySIS software.
The SEM was used in the secondary-electrons
imaging and backscattered-electrons imaging modes to
analyze the inclusions; details of the composition of the
inclusions were investigated by EDS mapping and point
analyses.
The light microscopy and SEM analyses (JEOL
JSM-6500F) were performed at the Institute of Metals
and Technology in Ljubljana.
Table 1: Chemical composition of the spring steel sample No. 49115
in w/%
Tabela 1: Kemijska sestava vzorca vzmetnega jekla {t. 49115 v
masnih dele`ih, w/%
Element C Si Mn Cr V S Al
Composition w/% 0.52 0.35 0.96 0.93 0.12 0.007 0.010
3 RESULTS AND DISCUSSION
3.1 Light microscopy
Since thorough inspection is difficult and time-con-
suming, statistical methods have been developed for
predicting the characteristic maximum inclusion size in a
large volume of steel by extrapolating from data gathered
on small samples. In this study, Generalized Extreme
Value (GEV) theory was applied to the results and a
prediction for the characteristic inclusion size was calcu-
lated.
The basic concept of extreme value theory is that
when a number of data points following a basic distri-
bution are collected on multiple samples, the maxima
and minima of each of these sets also follow a certain
distribution. The distribution function was given by
Gumbel 6,7:
G z
z
a
( ) exp exp
( )
= −
− −⎛
⎝
⎜ ⎞
⎠
⎟ (1)
where is the probability that the maximum inclusion is
no larger than the size z, and a and  are the scale and
location parameters. It was first applied to the inclusion
sizes in steels by Murakami and coworkers. The
distribution was later generalized with the addition of
another parameter 	:
G z
z
a
( ) exp
( )
/
= − +
−⎛
⎝
⎜ ⎞
⎠
⎟⎡
⎣⎢
⎤
⎦⎥
⎛
⎝
⎜
⎞
⎠
⎟
−
1
1
	

	
(2)
which reduces to the Gumbel distribution (Eq. 1) for 	 =
0.
A standard inspection area S0 = 0.27 mm2 was de-
fined and 554 such areas were examined. The cross-
section area of the largest inclusion in each inspection
area was measured and a square root of the area z =# was
calculated. Values of the distribution parameters were
then calculated using the maximum-likelihood method.
The hypothesis 	 = 0 was not statistically supported and
thus abandoned. The parameter values of the generalized
extreme-value distribution are listed in Table 2.
Table 2: Estimated parameters with standard errors from the GEV
method for the experimental steel
Tabela 2: Ocene parametrov s standardnimi napakami za generalizi-
rano metodo ekstremnih vrednosti
Parameter Value Std. error

 1.96 0.06
 6.37 0.09
	 –0.0089 0.0018
To estimate the size of the maximum inclusion in the
volume of spring steel V, the return level T is defined as
follows:
T =V/V0 (3)
where V0 = h · S0 is the standard inspection volume, as
defined by Murakami et al., where h = area imax,∑ /N.
The estimate for the characteristic size of the maximum
inclusion is then calculated from [refs]:
z
a
Tmax
ln= − − − −⎛⎝
⎜ ⎞
⎠
⎟⎡
⎣⎢
⎤
⎦⎥
⎛
⎝
⎜
⎞
⎠
⎟
−

	
	
1 1
1
(4)
There exists an estimation upper limit of the inclu-
sion size when 	 < 0. The characteristic sizes of the
maximum inclusion in different volumes of spring steel
are shown in Figure 1. The estimation’s upper limit is
224 μm. The predicted result of the inclusion size can be
used in a database for a reference in the steel-making
process.
A. BYTYQI et al.: CHARACTERIZATION OF THE INCLUSIONS IN SPRING STEEL ...
56 Materiali in tehnologije / Materials and technology 45 (2011) 1, 55–59
Figure 1: Estimated characteristic sizes of the maximum inclusions in
different volumes. Confidence intervals were calculated using the
maximum-likelihood method for a 95% confidence level.
Slika 1: Ocenjene karakteristi~ne velikosti najve~jih vklju~kov za raz-
li~ne volumne jekla. Intervali zaupanja so bili izra~unani po metodi
najve~je verjetnosti za 95-odstotno stopnjo zaupanja.
3.2 SEM analysis
Scanning electron microscopy of the investigated
spring-steel samples revealed different types of inclu-
sions. Figure 2 shows a typical complex inclusion. Three
analyses were performed on this inclusion to reveal the
detailed composition, mainly of calcium sulfide and
aluminum oxide, Table 2.
Aluminum oxide inclusions are the result of aluminum
de-oxidation. The great advantage of steel de-oxidation
with aluminum and the use of calcium to modify the
form of alumina inclusions is that they reduce the
dissolved oxygen activity in liquid steel, resulting in a
final product with a large index of cleanliness and a
smaller tendency for porosity formation8.
Different inclusions were observed with scanning
electron microscopy in sample 2. This chemical
composition was similar, but the size was different.
According to the EDS analysis the content of calcium
sulfide (CaS) was greater than the content of aluminum
oxide (Al2O3). As can be seen in Figure 3, aluminum
oxide (Al2O3) in the inclusion center is surrounded by
calcium sulfide (CaS). The distribution of the elements
(Table 3) indicates that sulphur is bound in the man-
ganese sulphide (MnS).
The analysis of the non-metallic inclusions in the
steel 100 Cr6 indicated the presence of complex,
oxide-sulfide and sulphide inclusions with different
shape characteristics, while, a higher content of the
sulphide inclusions is related to the high content of
sulphur9.
A complex inclusion can be seen in Figure 4. The
detailed analysis of the inclusion supported the finding
and showed that the inclusion consisted of calcium
sulfide and aluminum oxide. The acquired SEM
A. BYTYQI et al.: CHARACTERIZATION OF THE INCLUSIONS IN SPRING STEEL ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 55–59 57
Table 2: Semi-quantitative chemical analysis of the inclusion in the studied spring steel in w/%
Tabela 2: Semikvantitativna kemijska analiza nekovinskega vklju~ka preiskovanega vzmetnega jekla v masnih dele`ih, w/%
Elements O Mg Al Si S K Ca Mn Fe
Spectrum 1 4.89 — 1.53 — 28.1 — 39.6 3.33 7.80
Spectrum 2 5.48 — 1.53 — 36.1 — 41.1 3.95 6.33
Spectrum 3 48.7 2.42 34.8 2.19 1.29 1.92 4.35 — 4.21
Table 3: Semi-quantitative chemical analysis of the inclusion in w/%
Tabela 3: Semikvantitativna kemijska analiza nekovinskega vklju~ka v masnih dele`ih, w/%
Elements O Mg Al S Ca Mn Fe
Spectrum 4.20 2.06 2.14 33.8 5.52 47.1 5.07
Figure 3: (a) SEM image of a complex, non-metallic inclusion and the corresponding elemental distribution for (b) O K
1, (c) Mn K
1, (d) Al
K
1, (e) S K
1, (f) Ca K
1.
Slika 3: (a) Nekovinski vklju~ek, posnet s sekundarnimi elektroni v vrsti~nem elektronskem mikroskopu in ploskovni posnetki elementov (b) O
K
1, (c) Mn K
1, (d) Al K
1, (e) S K
1, (f) Ca K
1
Figure 2: SEM image of a complex, non-metallic inclusion taken by
secondary electrons
Slika 2: Kompleksen nekovinski vklju~kek, posnet s sekundarnimi
elektroni v vrsti~nem elektronskem mikroskopu
examination revealed that the sulfur and calcium content
are lower in the light area of the inclusion than in the
dark area. Correspondingly, the oxygen and aluminum
contents are higher in the dark area compared with the
light area. This indicates that the inclusion is composed
of two different parts: the light part mainly consists of
aluminum oxide (Al2O3), while the dark areas mainly
consist of calcium sulfide (CaS) Table 4. The characte-
ristic shape of the inclusion is globular and the size is 4
μm. It is quoted in references that inclusions with an
irregular shape and sharp edges cause larger stress
concentrations around the inclusions than inclusions with
a smooth shape, even if there is not a significant
difference in the size10.
Manganese sulfide was found either as a singular
inclusion or as an outer shell on an oxide. Figure 5
shows an SEM image of a sulfide type of inclusion
consisting of mass fractions 48.2 % of manganese and
26.4 % of sulfur Table 5. These inclusions are narrow
and long because of their high ductility at the steel
A. BYTYQI et al.: CHARACTERIZATION OF THE INCLUSIONS IN SPRING STEEL ...
58 Materiali in tehnologije / Materials and technology 45 (2011) 1, 55–59
Figure 5: (a) SEM image of a complex, non-metallic inclusion and the corresponding elemental distribution for (b) S K
1, (c) Mn K
1, (d) Al
K
1
Slika 5: (a) Nekovinski vklju~ek, posnet s sekundarnimi elektroni v vrsti~nem elektronskem mikroskopu in ploskovni posnetki elementov (b) S
K
1, (c) Mn K
1, (d) Al K
1
Table 5: Semi-quantitative chemical analysis of the inclusion in Figure 4 in w/%
Tabela 5: Semikvantitativna kemijska analiza nekovinskega vklju~ka s slike 4 (w/%)
Elements V S Cr Mn Fe Total
Spectrum 3 0.48 26.4 1.51 48.2 22.5 100.0
Figure 4: (a) SEM image of a complex, non-metallic inclusion and the corresponding elemental distribution for (b) S K
1, (c) Ca K
1, (d) Mg
K
1, (e) Al K
1, (f) O K
1
Slika 4: (a) Nekovinskih vklju~ki posneti v vrsti~nem mikroskopu in ploskovni posnetki elementov (b) S K
1, (c) Ca K
1, (d) Mg K
1, (e) Al
K
1, (f) O K
1
Table 4: Semi-quantitative chemical analysis of the inclusion in the studied spring steel in w/%
Tabela 4: Semi-kvantitativna kemi~na analiza nekovinskega vklju~ka v masnih dele`ih, w/%
Elements Al S Ca Mn Fe Cr Cu
Spectrum 0.81 26.4 0.37 48.2 22.4 0.64 —-
rolling temperature, and then deform and are elongated
as the steel is rolled11. Because of their elongated shape,
MnS inclusions may affect the transverse toughness of
the steel12. Apart from the chemical composition, the size
of the inclusions is crucial for the steel’s properties. The
results show that the critical inclusion size in the spring
steel 51CrV4 was 0.14mm13.
4 CONCLUSION
The nonmetallic inclusions in spring steel were
investigated in the laboratory, and the inclusion size
distribution and compositions were determined using
optical microscopy according the international standard
ISO 4967 and with SEM combined with EDS,
respectively.
The more specific conclusions from this work are as
follows. The majority of inclusions found in steel contain
manganese sulfide (MnS), calcium sulfide (CaS) and
aluminum oxide (Al2O3) or a mixture of these. A
comparison between the inclusion compositions in the
spring steel showed that they are very similar in terms of
chemical composition. The SEV method can be
effectively used to estimate the size of the maximum
inclusion in different volumes of the spring steel.
However, a reasonable estimation can be derived from
the combined analysis of two methods, i.e., the
metallographic and statistically methods. The inclusion
characteristics are not sufficiently investigated for many
of the commercially available spring steels and should be
amended.
5 REFERENCES
1 Gao Hong-bin., Jia Yun-hai., Li Mei-ling., Yuan Liang-jing, WANG
Hai-zhou. Research on original position statistic distribution analysis
model of aluminum inclusion grain size in steel. Central Iron&Steel
Research, 2009
2 Lifeng Zhang., Brian G. Thomas. Inclusions in continuous casting of
steel. Mech. Engr. Buldg. (2003), 138–183
3 A. Ahmad, J. Purbolaksono, Z. Yahya., Estimating inclusion size in
WE43-T6 magnesium alloys based on Gumble extreme values.
Materials Science and Engineering. A 513–514 (2009), 319–324
4 Shehata M. T. Boyd J. D. Measurement of Spatial Distribution of
Inclusions. Inclusions and their Influence in Material Behavior, 4
(1998), 123–131
5 Atkinson H. V., Sh. G. Characterization of Inclusions in clean steels;
A review including the statistics of extremes methods. Progress in
Materials Science, 48 (2003) 5, 457–520
6 J. M. Zhang, J. F. Zhang, Z. G Yang, G. Y. Li, G. Yao, S. X. Li, W. J.
Hui, Y. Q. Wenig, Estimation of maximum inclusion size and fatigue
strength in high strength ADF1steel. 394 (2005), 126–131
7 C. W. Anderson, G. Shi, H. V. Atkinson, C. M. Sellars, J. R. Yates,
Interrelationship between statistical methods for estimating the size
of maximum inclusion in clean steels. Acta Materiala. 51 (2003),
2331–2343
8 Jose Carlos, S. Pires, Amauri Garcia. Modification of oxide
inclusions present in aluminum-killed low carbon steel by addition of
calcium. Metalurgia & Materialis., 57 (2004), 1–5
9 A. Gigovi~, M. Oru~, I. Vitez, B. Vuji~i~. Analyse and research of
nonmetallic inclusions for steel 100Cr6. Metalurgija, 48 (2009),
29–32
10 Pekko Juvonen. Effectc of non-metallic inclusions on Fatigue
properties of Calcium Treated Steels. Helsinki University of
Technology, PhD Thesis, (2004), 14
11 L. M. Cabalin, M. P. Mateo, J. J. Laserna. Large area mapping of
non-metallic inclusions in stainless steel by an automated system
based on laser ablation. Elsevier. Spectrochimica Acta Part B, 4
(2004), 567–575
12 Claude Lincourt, Madhavarao Krishnadev. Effect of inclusion
morphology on the fracture toughness of high strength steels. Mater.
Technol. 27 (2009), 1687–1698
13 Miha Kova~i~, Sandra Sen~i~. Critical Inclusion Size in Spring Steel
and genetic programming. Materials and Geoenvironment, 57
(2010), 17–23
A. BYTYQI et al.: CHARACTERIZATION OF THE INCLUSIONS IN SPRING STEEL ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 55–59 59

M. GODEC et al.: CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti SHAPE-MEMORY ALLOY WIRE
CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti
SHAPE-MEMORY ALLOY WIRE
KARAKTERIZACIJA KARBIDOV V @ICI ZLITINE S SPOMINOM
Ni-Ti
Matja` Godec, Aleksandra Kocijan, Monika Jenko
1Institute of Metals and Technology, Lepi pot 11, 1000 Ljubljana, Slovenia
matjaz.godec@imt.si
Prejem rokopisa – received: 2010-09-27; sprejem za objavo – accepted for publication: 2010-10-19
The microstructure of a commercially available Ni-Ti shape-memory alloy was investigated with Scanning Electron Microscopy
(SEM), Energy-Dispersive X-ray Spectroscopy (EDS) and Electron Backscatter Diffraction (EBSD) patterns. The material
investigated was a Ni-Ti shape-memory alloy with the chemical composition in the mass fraction: C 0.1 %, Ni 45 % and Ti 54.9
%. The microstructure of the alloy consisted of nanosized crystal grains of the Ni-Ti phase and particles of titanium carbides.
The majority of the particles were aligned in the longitudinal direction parallel to the wire’s axis. The problems occurring during
the EBSD analysis of the Ni-Ti phase are discussed and some orientational relationships between the carbides and the matrix are
suggested.
Key words: shape memory alloy, Ni-Ti, TiC, EBSD, EDS
Raziskali smo mikrostrukturo komercialne zlitine s spominom Ni-Ti z uporabo vrsti~ne elektronske mikroskopije (SEM),
energijsko disperzijske spektroskopije (EDS) in uklona sipanih elektronov (EBSD). Preiskovani material je bila zlitina s
spominom Ni-Ti s kemijsko sestavo v masnih dele`ih: C 0.1 %, Ni 45 % in Ti 54.9 %. Mikrostruktura zlitine je bila sestavljena
iz nanokristalnih zrn faze Ni-Ti in titanovih karbidov. Ve~ina karbidov je bila usmerjena vzdol`no z osjo `ice. Opisani so
problemi EBSD-analize faze Ni-Ti in predlagane orientacijske odvisnosti med karbidi in matrico.
Klju~ne besede: zlitina s spominom, Ni-Ti, TiC, EBSD, EDS
1 INTRODUCTION
Porous Ni-Ti shape-memory alloys are widely used
in numerous biomedical applications (orthodontics,
cardiovascular, orthopedics, urology, etc.) because of
their good biocompatibility, unique shape-memory pro-
perties, mechanical properties, superior damping capa-
bility, excellent corrosion resistance and wear resistance
1–6. They combine their special functional properties with
a high mechanical strength 7. These characteristics are
due to the martensitic transformation and its reversion,
which can be activated by thermal or mechanical loads 7.
The Ni-Ti alloy has similar mechanical characteristics to
natural biomaterials, e.g., a high recoverable strain and a
low elastic modulus 8,9, both of which are very similar to
bone. This makes the alloy an ideal biological engi-
neering material for orthopedic surgery and orthodontics
10. Despite these advantages, the high Ni content (50 %)
in the Ni-Ti shape-memory alloy is a major health
concern 11: the toxicity, carcinogenicity and allergic
hazards associated with Ni have been the subject of a
number of studies 12,13. One of the most effective and
obvious solutions is to improve the surface microstruc-
ture and modify the properties. Of course, one of the
basic requirements for any metallic implant is that it
should not exhibit any toxicity and that it should be
biocompatible.
The properties of a Ni-Ti shape-memory alloy
depend strongly on the exact chemical composition, the
processing history, and the level of impurities 7. For
example, contaminants such as oxygen and carbon can
dramatically affect the properties of the alloy; their
penetration tends to occur during the production and
processing of the alloy 7.
To obtain a good shape-memory effect it is crucial to
have a good chemical homogeneity of the material,
which then leads to single-phase Ni-Ti formation. The
usual process to produce Ni-Ti shape-memory alloys is
called Vacuum Induction Melting (VIM), using high-
density graphite crucibles to minimize the carbon
contamination of the melt. The carbon combines with the
titanium, which results in the precipitation of TiC
particles. The consequence is an alloy matrix that is
richer in nickel than the nominal composition, which
means a lowering the martensitic transformation tem-
peratures 3,14–16. The higher concentration of Ni in the
Ni-Ti matrix increases the risk of allergic and toxic
reactions 1 as well as reducing the shape-memory effect.
The precipitated titanium carbides have been reported to
have a certain orientational relationship with the matrix 16.
In the present study the microstructure of a
commercial Ni-Ti orthodontic wire produced by vacuum
induction melting was examined with Scanning Electron
Microscopy and Electron Backscatter Diffraction
(EBSD) patterns. The problems occurring during the
indexing of the EBSD patterns of the Ni-Ti phase are
described and some of the orientational relationships
Materiali in tehnologije / Materials and technology 45 (2011) 1, 61–65 61
UDK 669.24'29:543.428.2 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)61(2011)
between the carbide particles and the matrix are
suggested.
2 EXPERIMENTAL
The material used was a Ni-Ti shape-memory alloy
with the following chemical composition in mass
fractions: C 0.1 %, Ni 45 %, Ti 54.9 %. An orthodontic
archwire specimen ( 0.2 mm × 100 mm) was hot
mounted in conductive Bakelite, ground and polished.
The usual diamond polishings with 3 μm and 1 μm
particles were prolonged to 10 min each; these were
followed by colloidal silica oxide polishing for 5 min
and then cleaning in an ultrasonic bath. The specimen
was analyzed in a FE-SEM JEOL JSM 6500F
field-emission scanning electron microscope using
energy-dispersive X-ray spectroscopy (EDS), an INCA
X-SGHT LN2-type detector, INCA ENERGY 450
software, and an HKL Nordlys II EBSD camera using
Channel 5 software. The microstructure was revealed
with etching in 85 % H20, 5 % HF and 10 % HNO3. For
the EDS analyses a 15-kV accelerating voltage and a
probe current of 0.7 nA were used. The parameters
chosen represent a good compromise between the size of
the analyzing volume and the overvoltage needed to
detect the elements present. The EBSD was performed at
a 15-kV accelerating voltage and a 2.6 nA probe current.
The analysis was made using 40 reflectors and 4 × 4
binning. For the noise reduction, 10 frames were
used, and each frame was obtained at 20 ms. The
obtained crystallographic data were analyzed using the
Channel 5 software and the ICSD 2003 database.
3 RESULTS AND DISCUSSION
The thermomechanical behavior of shape-memory
alloys is obtained with the martensitic-austenitic phase
transition, which depends on the temperature and the
stress, and occurs as a result of the martensitic twinning
at low temperatures. The hysteresis loop of the analyzed
shape-memory alloy has a martensite start at 2 °C and a
martensite finish at –15 °C, and an austenite start at 18
°C and an austenite finish at 36 °C, in order to achieve
the phase transformation from martensite to austenite,
which is typical for applications at human-body tempe-
rature 17. The microstructure of the investigated alloy
consists of the Ni-Ti phase with a number of titanium
carbide particles (Figure 1). The high carbon content of
0.1 % is due to the dissolution of carbon from the
crucible graphite in the melt during the vacuum induc-
tion melting 16. The titanium carbides might lower the
concentration of nickel in the matrix; however, a bene-
ficial effect of increased hardness may occur, particularly
if the interface between the titanium carbide and the
Ni-Ti phase is semi-coherent 16.
Chemical etching of the Ni-Ti surface prior to the
surface characterization causes a rapid attack on the NiTi
matrix, while the titanium carbides were not affected.
Therefore, in the highly tilted mode in the Scanning
Electron Microscope the carbide particles, and especially
their shape, become clearly observable (Figure 2). The
longitudinal direction of the carbide is parallel to the
wire’s longitudinal direction. This specific situation may
result from two possible factors concerning the precipi-
tation of carbide particles. The first one is related to the
manufacturing of the shape-memory Ni-Ti alloy, which
is associated with the casting and the plastic deformation
of the wire at elevated temperatures. The carbides might
precipitate during the casting solidification 16 and later
during the plastic working with hot rolling and drawing
when the non-deformable particles are aligned in the
longitudinal direction of the wire. In this case, the
carbide particles have no orientational relationship with
the matrix because of the re-crystallization during or
after the wire drawing. On the other hand, the particles
M. GODEC et al.: CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti SHAPE-MEMORY ALLOY WIRE
62 Materiali in tehnologije / Materials and technology 45 (2011) 1, 61–65
Figure 2: Etched transversal cross-section of NiTi wire with visible
carbides in longitudinal direction. Micrograph obtained from
secondary electrons in 60° tilted mode.
Slika 2: Prikaz jedkanega pre~nega prereza NiTi-`ice z vidnimi
karbidi v vzdol`ni smeri. Slika je bila posneta s sekundarnimi
elektroni, vzorec nagnjen za 60°.
Figure 1: Polished transversal cross-section of NiTi wire. Black spots
represent titanium carbides in the NiTi matrix. Micrograph obtained in
scanning electron microscope using backscattered electrons.
Slika 1: Prikaz poliranega pre~nega prereza NiTi `ice. Temna
podro~ja so titanovi karbidi v NiTi matrici. Slika je bila posneta v
vrsti~nem elektronskem mikroskopu z odbitimi elektroni.
may have precipitated during the wire drawing, which
would impact on the specific growth direction of the
particles. The titanium carbides have a rod-like shape;
however, the transversal cross-section is not completely
circular; in some cases it is closer to a square with
rounded edges. It is most likely that the wire was
manufactured by continuous casting and drawing, so that
the carbides precipitate during the casting and have a
specific orientation due to the tension gradient in the
drawing direction. This can explain why the carbides are
not broken into small pieces, which is what usually
occurs with aluminum or silicon non-metallic inclusions
in steels during plastic deformation. Some of the carbide
particles are slightly thinner at their ends, which is
typical evidence that the precipitation and the drawing
took place at a high enough temperature to allow plastic
deformation also of the carbides.
The EBSD analysis was performed on highly
polished surfaces where the carbide particles formed just
a slight topography. Based on the available ICSD 2003
database 18,19, the EBSD patterns were solved almost
equally as TiC or Ti8C5. However, a slightly better MAD
(mean angular deviation) number was obtained for the
TiC phase. Zhang et al. reported the lattice of titanium
carbide particles as being cubic TiC 16. TiC is in a wider
region according to the phase diagram calculated using
the Thermo Calc program 20. TiC can have the B1 (or
NaCl-type) lattice with the titanium atoms situated in a
face-centered-cubic, closed-packed arrangement with the
octahedral interstitial sites occupied by carbon atoms 21.
The equilibrium phase diagram exhibits only one carbide
phase, TiC, which is characterized by a wide region of
composition (from TiC0.48 to TiC1.00) and melts con-
gruently at 3068 °C 22. The reason why both carbides
create almost identical patterns from the interaction of
electrons with the crystal structure is that in both lattices
the atoms are in a close-packed formation and the similar
ordering creates similar electron refraction. Because of
the very similar EBSD patterns of both the equilibrium
and non-equilibrium carbides, it is not possible to
exclude the existence of the Ti8C5 carbide. Nevertheless,
the shape of the carbide rods, with their slightly square
transversal shape, might be more indicative of the
equilibrium cubic TiC. The internal stress can have an
influence on the carbide growth at high temperature and
might change the bulky crystal shape. Some hexagonal
form, for instance an angle of 120°, is to be expected
also. Figure 3 shows the EBSD pattern of a typical TiC
particle. The orientation of the crystal is such that the
longitudinal direction of the carbide is parallel to the
wire.
EDS microchemical analyses were performed on
some carbide particles. However, an accurate EDS
carbon analysis is always a challenge. The system used
has no liquid-nitrogen trap and therefore some excess
carbon due to contamination is always observed.
Depending on the material analyzed, some carbon build
up under the electron beam is usually observed during
the EDS measurement in the range of mole fractions 10
% or 15 % in steels and Ni-Ti alloys, respectively. The
carbon in the vacuum analyzing chamber originates
mostly from the oil vapor in the diffusion pumps. Taking
into account this particular fact and from the EDS
measurements on titanium carbides, it is possible to
conclude that the carbides are closer to TiC than Ti5C8.
Unfortunately, using the WDS technique gives no impro-
vement in such conditions; on the contrary, it makes the
measurements even less reliable. The phenomenon of
carbon build up is even more intense due to the higher
current. Figure 4 shows EDS spectra from typical
carbides in the Ni-Ti shape-memory alloy and the cal-
culated chemical composition of the carbide. From the
EDS spectra the calculated atomic ratio of Ti to C is
approximately 1:1. There is, however, some excess
amount of carbon due to the vacuum problem mentioned
above. A certain amount of oxygen was also observed in
M. GODEC et al.: CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti SHAPE-MEMORY ALLOY WIRE
Materiali in tehnologije / Materials and technology 45 (2011) 1, 61–65 63
Figure 4: EDS spectrum of carbides shows slightly larger amount of
carbon that belongs to the TiC. The excess amount is due to carbon
build up under the electron beam.
Slika 4: EDS-spekter karbidov, ki prikazuje nekoliko povi{an dele`
ogljika glede na TiC. Prese`ek ogljika je zaradi nalaganje pod elek-
tronskim curkom.
Figure 3: EBSD pattern of typical TiC carbide in longitudinal direc-
tion of the wire. The pattern was solved as TiC with the orientation of
the crystal shown in the figure.
Slika 3: EBSD-slika tipi~nega TiC-karbida v vzdol`ni smeri `ice.
Elektronska uklonska slika prikazuje TiC s kristalno orientacijo,
prikazano na sliki.
the EDS spectra. It was reported previously that the
presence of carbon and oxygen can result in the
formation of different titanium carbides and oxides 23.
To prepare the Ni-Ti alloy for the EBSD investigation
was a demanding task. Two different approaches were
tested: polishing of the surface with argon ions and
polishing by using silica oxide nanoparticles. The latter
was more successful, however. It is believed that the
highly energetic ions may affect the lattice and induce
some internal stresses, which could be reflected in the
EBSD pattern, as observed during the EBSD mapping.
In these analyses the indexing and the band-contrast
image constantly deteriorated during the EBSD
mapping. The explanation for this may also be found in
the austenite-martensite transformation as well as the
martensite twining caused by the interaction of electrons
and the specimen being analyzed. Most likely this
problem is related to the contamination of the surface
during EBSD analysis as well as local oxidation of the
surface.
The orientational relationship of the titanium car-
bides (with the B1 crystal structure) and the Ni-Ti matrix
(with the B2 crystal structure) was established (Figure
5). It was found that the majority of the TiC particles
have the same orientation, and without any clear
relationship with the NiTi matrix lattice. However, it is
related to the longitudinal direction of the wire, as the
EBSD analysis showed that majority of titanium carbides
had the longitudinal direction [111]. The grain size of the
Ni-Ti phase was in the range of 200 nm.
4 CONCLUSIONS
The present study was conducted in order to cha-
racterize a commercially available Ni-Ti shape-memory
alloy. It was found that the microstructure of the
analyzed shape-memory alloy consisted of nanosized
crystal grains of the Ni-Ti phase with a number of
titanium carbides. The chemical analyses confirmed that
the excess amount of carbon originated from the
induction-melting process in the graphite crucible, which
decreased the shape-memory effects due to the formation
of TiC precipitates. The titanium carbides were rod
shaped, while the transversal cross-section shape was in
the form of a modified circle, although in some cases it
was close to a square with rounded edges. The majority
of the carbides were aligned parallel to the wire axis,
which can be associated with an orientation change
during the plastic deformation. It is also possible that the
particles grow faster in the longitudinal direction due to
the stresses produced by the wire drawing. The grain size
of the Ni-Ti phase was in the range of 200 nm, and the
majority of the carbides had a direction parallel to the
wire’s longitudinal direction [111].
Acknowledgements
The authors are grateful to Dr. Janez Rozman for pro-
viding the investigated samples.
5 REFERENCES
1 Neelakantan L, Swaminathan S, Spiegel M, Eggeler G, Hassel A W,
Selective surface oxidation and nitridation of NiTi shape memory
alloys by reduction annealing Corros. Sci., 51 (2009), 635–641
2 Kim J I, Miyazaki S, Effect of nano-scaled precipitates on shape
memory behavior of Ti-50.9at.%Ni alloy, Acta Materialia, 53
(2005), 4545–4554
3 Saburi T, in: Shape memory materials, Otsuka K, Wayman C M, eds,
Cambridge University Press, Cambridge, UK, 1998, p. 49
4 Otsuka K, Ren X, Recent developments in the research of shape
memory alloys Intermet,. 7 (1999), 511–528
5 Duerig T, Pelton A, Stöckel D, An overview of nitinol medical
applications, Mater. Sci. Eng. A, 273–275 (1999), 149–160
6 Chu C L, Chung C Y, Lin P H, Wang S D, Fabrication of porous NiTi
shape memory alloy for hard tissue implants by combustion
synthesis Mater. Sci. Eng. A, 366 (2004), 114–119
7 Mehrabi K, Bahmanpour H, Shokuhfar A, Kneissl A, Influence of
chemical composition and manufacturing conditions on properties of
NiTi shape memory alloys Mater. Sci. Eng. A, 481 (2008), 693–696
8 Gjunter V E, Superelastic shape memory implants in maxillofacial
surgery, traumatology, orthopaedics and neurosurgery, Tomsk Uni-
versity Publishing House (TUP), Tomsk, Russia, 1995
9 Helsen J A, Breme H J, Metals as biomaterials, Wiley, Baffins Lane,
Chichester, England, 1998
10 Gil F X, Planell J A, Manero J M, Relevant aspects in the clinical
applications of NiTi shape memory alloys J. Mater. Sci.: Mater.
Med.,7 (1996), 403–406
11 Chu C L, Guo C, Sheng X B, Dong Y S, Lin P H, Yeung K W K,
Chu P K, Microstructure, nickel suppression and mechanical
characteristics of electropolished and photoelectrocatalytically
oxidized biomedical nickel titanium shape memory alloy Acta
Biomaterialia, 5 (2009) 6, 2238–2245
12 Williams D F, Fundamental aspects of biocompatibility, in: Toxi-
cology of implanted metals, CRC Press, Boca Raton, FL, 1981, vol.
II
13 Takamura K, Hayashi K, Ishinishi N, Yamada T, Sugioka Y,
Evaluation of carcinogenicity and chronic toxicity associated with
M. GODEC et al.: CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti SHAPE-MEMORY ALLOY WIRE
64 Materiali in tehnologije / Materials and technology 45 (2011) 1, 61–65
Figure 5: Marked area of EBSD spot analysis with corresponding
EBSD pattern. Spot 1 shows the carbides, spots 2 to 7 show Ni-Ti
matrix with nanograins.
Slika 5: Ozna~eno mesto EBSD to~kovne analize s pripadajo~o EBSD
uklonsko sliko. Mesto 1 prikazuje karbid, mesta od 2 do 7 prikazujejo
Ni-Ti matrico z nanozrni.
orthopedic implants in mice J. Biomed. Mater. Res., 28 (1994),
583–589
14 Otubo J, Rigo O D, Neto C M, The Mei P R, effects of vacuum
induction melting and electron beam melting techniques on the
purity of NiTi shape memory alloys Mater. Sci. Eng. A, 438 (2006),
679–682
15 Tang W, Sundmann B, Sandström R, Qiu C, New modelling of the
B2 phase and its associated martensitic transformation in the Ti-Ni
system Acta Mater., 47 (1999), 3457–3468
16 Zhang Z H, Frenzel J, Somsen C, Pesicka J, Eggeler G, in: Progress
in Crystal Growth Research, Karas G V, ed., Nova Science Publi-
shers, New York, USA, 2005
17 https://www.goodfellow.com (2008–2010)
18 Zhou Y C, Wang X H, Sun Z M, Chen S Q, Electronic and structural
properties of the layered ternary carbide Ti3AlC2 J. Mater. Chem., 11
(2001), 2335–2339
19 Khayenko B V, Golub V I, Arbuzow M P, Kristallografiya 25 (1980),
112–118
20 Sundman B, Thermo-Calc Software Version S, Royal Institute of
Technology, Stockholm, 2008
21 Jeitschko W, Pottgen R, Hoffman R D, Structural chemistry of hard
materials, in: Handbook of Ceramic Hard Materials, Riedel R, ed,
Wiley-VCH, New York, 2003, p. 9
22 Jonsson S, Assessment of the Ti–C system Z. Metallkd., 87 (1996),
703–772
23 Mentz J, Frenzel J. Wagner M F -X, Neuking K, Eggeler GBuch-
kremer, H P, Stöver D, Powder metallurgical processing of NiTi
shape memory alloys with elevated transformation temperatures,
Mater. Sci. Eng. A, 491 (2008), 270–278
M. GODEC et al.: CHARACTERIZATION OF THE CARBIDES IN A Ni-Ti SHAPE-MEMORY ALLOY WIRE
Materiali in tehnologije / Materials and technology 45 (2011) 1, 61–65 65

B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
FRACTURE TOUGHNESS OF THE
VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
LOMNA @ILAVOST VAKUUMSKO TOPLOTNO OBDELANEGA
VZMETNEGA JEKLA 51CrV4
Bojan Sen~i~1, Vojteh Leskov{ek2
1[tore Steel, d. o. o., @elezarska cesta 3, SI-3220 [tore, Slovenia
2 Institute of metals and technology, Lepi pot 11, 1000 Ljubljana, Slovenia
bojan.sencic@
Prejem rokopisa – received: 2011-01-12; sprejem za objavo – accepted for publication: 2011-01-17
In the work the possibilities of the vacuum heat treatment of spring steel grade 51CrV4 are presented. Charpy-V notch (CVN)
impact-test values are widely used in toughness specifications for spring steels, even though the fracturing energy is not directly
related to the spring design. The plain-strain stress-intensity factor (KIc) at the onset of unstable crack growth can be related to
the spring design; however, KIc values are not used in the toughness specifications. This is surprising since to the designer KIc
values are more useful than CVN values, because the design calculations for springs from high-strength steels should also take
into account the strength and the toughness of materials to prevent rapid and brittle fracture. An investigation was conducted to
determine whether standardized fracture-toughness testing (ASTM E399-90), which is difficult to perform reliably for hard and
low ductility materials, could be replaced with a non-standard testing method using circumferentially notched and
fatigue-precracked tensile specimens. The results of this investigation have shown that using the proposed method it was
possible to draw, for the normally used range of working hardness, combined tempering diagrams (Rockwell-C hardness –
Fracture toughness KIc – Tempering temperature) for the vacuum-heat-treated spring steel grade 51CrV4. Fractographic and
metallographic analyses of the KIc-test specimens used shows in steel the presence of positive and negative segregations. It was
found that the width of the segregations bands and the distance between the positive and negative segregations influence
significantly the fracture toughness due to the presence of bainite in the negative segregations.
Keywords: spring steels, vacuum heat treatment, hardness, fracture toughness, microstructures
V ~lanku so predstavljene mo`nosti vakuumske toplotne obdelave vzmetnega jekla 51CrV4. Vrednosti Charpy-V udarne
`ilavost (CVN) so pogosto navedene v specifikacijah za vzmetna jekla, ~eprav energija loma ni neposredno povezana z
dimenzioniranjem vzmeti. Ravninsko deformacijski faktor intenzitete napetosti (KIc) na za~etku nestabilne razpoke lahko
uporabimo pri dimenzioniranju vzmeti, vendar pa vrednosti KIc ne najdemo v specifikacijah. To je presenetljivo, saj so za
konstruktorje vrednosti KIc bolj uporabne od vrednosti CVN za visokotrdnostne vzmeti, pri katerih je treba upo{tevati poleg
trdnosti tudi `ilavost materiala, da prepre~imo lom. Opravljena je bila preiskava, s katero smo `eleli ugotoviti, ali lahko
standardizirano preizku{anje lomne `ilavosti (ASTM E399-90), ki je te`ko izvedljivo za trde in krhke materiale z majhno
duktilnostjo, nadomestimo z nestandardnim postopkom preizku{anja lomne `ilavosti s cilindri~nim nateznim preizku{ancem z
zarezo po obodu in utrujenostjo razpoko v dnu zareze. Rezultati raziskave so pokazali, da s predlagano metodo lahko
konstruiramo diagram popu{~anja (trdota Rockwell-C – lomna `ilavost KIc – temperatura popu{~anja) za vakuumsko toplotno
obdelano vzmetno jeklo 51CrV4. S fraktografsko in metalografsko analizo KIc preizkusnih vzorcev smo ugotovili prisotnost
pozitivnih in negativnih izcej v jeklu. Ugotovljeno je bilo, da {irina izcej in razdalja med njimi pomembno vplivata na lomno
`ilavost, in sicer zaradi prisotnosti bainita v negativnih izcejah.
Klju~ne besede: vzmetno jeklo, vakuumska toplotna obdelava, trdota, lomna `ilavost, mikrostruktura
1 INTRODUCTION
A manufacturer of spring steels must provide a tech-
nical steel description in their production program.
Namely, the durability of the springs is limited by plastic
deformation, fatigue and fracturing. From this point of
view, the use of spring steel with the following properties
is recommended: high ductility and toughness at ope-
rating temperatures from –40 °C to +50 °C, good harde-
nability that provides the required mechanical properties,
even at the maximum dimensions.
Steels with a similar chemical composition may
behave differently due to various mechanical properties
as a consequence of the manufacturing route. The
description of the spring steel includes the chemical
composition and the basic mechanical and physical as
well as technological properties. For the manufacturers
of springs, the information relating to the heat treatment
of a specific spring steel is important. Assuming that the
chemical composition and initial microstructure of the
steel correspond to prescribed for steel grade 51CrV4
(DIN 17221 and DIN 17222), then the mechanical
properties for a specific application depend mainly on
the appropriately selected parameters of the heat treat-
ment.
Hypo-eutectoid steels, such as the spring steel
51CrV4, are usually heat treated conventionally in a
furnace with or without a protective atmosphere at a
temperature of 30 °C to 50 °C above the Ac3 point soaked
at the austenizing temperature to obtain "homogeneous
austenite", and oil cooled to the temperature of the
quench oil. The quenching is followed by a tempering at
a selected temperature. The conventional tempering
diagram for spring steel gives us information about the
Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73 67
UDK 669.14.018.252:539.42 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)67(2011)
mechanical properties, i.e., the tensile strength Rm, yield
strength Rp0.2, elongation A5 and necking Z as a function
of the tempering temperature in the range from 350 °C to
700 °C for a specific austenitizing temperature. In the
case of spring steels, a minimum impact toughness
Charpy-V measured by standard CVN-specimens, must
be given. Lately, more and more users also request
measured values of the fracture toughness KIc for the
springs after heat treatment, which with the known
ultimate tensile strength allows us to calculate the
fracture stress f and the critical defect size acr at the
applied stress.
The trend of the heat treatment of machine parts,
including springs, is in the direction of vacuum heat
treatment. Namely, cooling rates that can be achieved in
modern vacuum furnaces with cooling in a stream of N2
or in a mixture of He and N2 under a pressure up to 25
bar, already reach cooling speeds close to the cooling
rates in oil. One can expect that the obtained micro-
structure and mechanical properties of the springs after
vacuum quenching and tempering will be comparable to
those achieved with a conventional heat treatment. In
vacuum heat treatment, there is no risk of decarburi-
zation, no formation of the oxide layer and no residues
of quench oil on the surface. It offers then a significant
advantage when compared to conventional heat treat-
ment. The objective of our work was to create a specific
tempering diagram for the investigated steel 51CrV4
from which the relations between the Rockwell-C
hardness, the fracture toughness KIc, the austenizing and
the tempering temperature could be determined.
2 THEORETICAL PART
2.1 Hardness, ductility and toughness
Factors contributing to the isotropic mechanical
properties of steel are a high cleanliness, a low level of
residual elements, of non-metallic inclusions (NMI) and
primary carbides without segregations. This can be
achieved with an appropriate process route of the
continuous cast spring steel through the Clean Steel
Concept1 using the Continuous Soft Reduction (CSR)2
process.
The properties used to evaluate the heat treatment of
spring steel are the Rockwell-C or Brinell hardness. In
the hardened state the hardness may be an indication of
the austenitizing temperature, from which the spring
steel has been quenched. In the tempered state, the
hardness is important for users, although based on the
hardness, it is not possible to distinguish between the
springs that were quenched and tempered in various
ways. For example, the same hardness can be achieved
by varying the austenitizing and tempering temperature,
but the toughness can be different. From this point of
view, it has sense to introduce into the tempering dia-
gram an additional criterion such as toughness. Unfortu-
nately, for the determination of toughness there are no
generally accepted test methods. Consequently, the
literature often refers to the data determined by different
test methods. The comparison of these data may some-
times lead to confusing discussions.
The Charpy-V notch (CVN) impact test is widely
used in toughness specifications for spring steels, even
though the fracturing energy is not directly related to the
spring design. The user requirements for springs are in
the direction of the operation of springs at a higher
tensile strength (1800 MPa and more), which allows the
use of only a single parabolic leaf spring, resulting in a
lower weight of the vehicle. However, with such a high
strength, the toughness is low, and therefore it is
necessary to measure the toughness of such springs with
a more sensitive method than the Charpy-V impact
toughness. The plain-strain stress-intensity factor (KIc) at
the onset of unstable crack growth can be related to the
spring design; however, the test values are not used in the
toughness specifications.
Introducing a linear elastic fracture mechanic
(LEFM) for the evaluation of the heat treatment of
springs makes it necessary to draw a distinction between
the ductility and the toughness. In this context, the
ability of a material to resist the initiation and spread of
fracture at gross plastic deformation is ductility and the
correct meaning of the toughness is the ability to resist
the growth of an existing crack at a tensile load 3,4. The
ductility and toughness are, consequently, two different
material properties, even though both, unfortunately, are
denominated as toughness. The opposite of both pro-
perties is, however, the same, i.e., brittleness.
The most reliable measure of toughness is the
plain-strain fracture toughness (KIc)5. The minimum size
of the specimens depends on the yield stress and the
fracture toughness of the material, both of which are
required for the plain-strain deformation. A fatigue crack
of defined length is propagated from a mechanical notch
in the specimens, ensuring that the notch effect is at a
maximum and equal for all tests. The same value of
fracture toughness should be found for tests on
specimens of the same material with different geometries
and with a critical combination of crack size and shape
and fracture stress. Within certain limits, this is indeed
the case, and information about the fracture toughness
obtained under standard conditions can be used to
predict the failure for different combinations of stress
and crack size, and for different geometries6.
To determine the fracture toughness KIc, of spring
steels, standard CT (Compact Tension) and SENB
(Single Edge Notch Bend) specimens and non-standard
circumferentially notched and fatigue-precracked ten-
sile-test specimens (KIc-test specimen) can be used7.
2.2 Relationship between ultimate fracture stress and
defect size
In metal materials there are always defects present,
with their importance increasing with increasing strength
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
68 Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73
and reducing ductility and toughness. In springs the
two-dimensional defects are critical. The fracture
strength of brittle materials depends on the size and the
orientation of the defect. It usually decreases rapidly
with the increasing defect size. If LEFM can be applied
to calculate the material behaviour, the fracture strength
can be calculated with the following equation8–10(1):
 f
IC=
⋅ −
K
Y aπ
for: f < Ys (1)
KIc fracture toughness
a defect size
f fracture stress (tensile strength)
ys yield stress
Y correction factor (depending on the geometry: Y =
1.11 for surface cracks)
Equation (1) can be applied if fracture occurs prior to
reaching the yield stress. As spring steels are relatively
hard (brittle) materials, they can be compared to tool
steels with low ductility and toughness. For this reason
LEFM can be applied over a wide range of defect sizes.
According to the literature, equation (1) is suitable
for describing the fracture behaviour of tool steels for a
typical crack-nucleating defect size (e.g., carbide
diameter, NMI) above 10 μm to 20 μm. In the case of
conventional manufactured tool steels with a banded
microstructure, the width of the carbide agglomerates
has to be taken as the critical defect size and it has been
further noted that the fracture occurs along a path with a
high carbide density. The fracture toughness of con-
ventionally produced tool steels depends, therefore, on
the specimen orientation8. It is highest for a crack
propagation perpendicular to the rolling direction and it
is lowest for a crack propagation in the rolling direction.
If the fracture stress according to equation (1)
approaches the yield stress (e.g., if no larger crack-
nucleating particles are present) elastic-plastic methods
have to be applied. In this region the material behaviour
can be described with equation10 (2), if the ultimate
fracture strength is known.
 
 f U
Ic
U
= ⋅
−
⎛
⎝
⎜ ⎞
⎠
⎟
⋅ ⋅
⎡
⎣
⎢
⎢
⎢
⎢
⎤
⎦
⎥
⎥
⎥
⎥
⎧
⎨−
2
8
1
2
π
π
cos exp
K
Y
a
⎪
⎪
⎩
⎪
⎪
⎫
⎬
⎪
⎪
⎭
⎪
⎪
(2)
U final tensile fracture stress (tensile strength Rm)
For a large defect size equation (2) approximates the
results from equation (1). Deviations from equation (1)
are found for small defect sizes. Especially for very
small defect sizes, the influence of the defect size is
almost negligible. A decrease of the fracture strength is
observed if a critical defect size is exceeded.
3 EXPERIMENTAL
3.1 Hardness and fracture-toughness tests
The Rockwell-C hardness (HRc) was measured on
individual groups of KIc-test specimens using a Wilson
4JR hardness machine. Circumferentially notched and
fatigue-precracked tensile-test specimens11 with the
dimensions indicated in Figure 1 were used for the
investigation.
The advantage of the KIc-test specimens used here
over standardized CT specimens (ASTM E399-90) is in
the radial symmetry, which makes them particularly
suitable for studying the influence of the microstructure
of metallic materials on the fracture toughness. The
advantage of these specimens is related to heat transfer,
ensuring a completely uniform microstructure.
Due to the high notch sensitivity of hard and brittle
metallic materials, such as continuous casted spring steel
grade 51CrV4, it is very difficult – sometimes even
almost impossible – to create a fatigue crack in the test
specimen. However, with KIc-test specimens the fatigue
crack can be created with rotating-bending loading
before the final heat treatment11; the second advantage of
such test specimens is that plain-strain conditions can be
achieved using specimens with smaller dimensions than
those of conventional CT test specimens12.
For the linear elastic behaviour up to fracture of such
specimens13 the following equation is applied:
K
P
D
D
dIC
= − +
⎛
⎝
⎜ ⎞
⎠
⎟3 2 127 172/ . . (3)
where P is the load at failure, D is the outside diameter,
and d is the notched-section diameter of the test
specimen, i.e., the diameter of the ligament next to the
crack. Equation (3) is valid as long as the condition 0.5
< d/D < 0.8 is fulfilled.
Measurements of the fracture toughness were per-
formed at room temperature using an Instron 1255 ten-
sile-test machine. The cross-head speed was 1.0 mm/min
for standard tensile tests on specimens with a nominal
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73 69
Figure 1: Circumferentially notched and fatigue-precracked KIc-test
specimen. All dimensions are in mm.
Slika 1: Cilindri~ni natezni preizku{anec za merjenje lomne `ilavosti
z zarezo po obodu in utrujenostno razpoko v dnu zareze. Vse dimen-
zije so v milimetrih.
test length of 100 mm. In the tests two specially prepared
cardan fixed jaws, ensuring the axiality of the tensile
load, were used. During the tests the tensile-load/ dis-
placement relationship until failure was recorded. In all
cases this relationship was linear, and the validity of
equation (3) for the tests was confirmed.
3.2 Material, sampling and vacuum heat treatment
Commercial, continuous-cast spring steel grade
51CrV4, delivered as rolled bars of dimensions (20 ×100
× 4000) mm from two heats with similar chemical com-
positions (Table 1) were used. As can be seen from the
table there is no difference in the chemical composition
of the two heats. The rolled bars were delivered after hot
plastic deformation in the as-rolled condition with a
mixed martensitic-bainitic microstructure and a hardness
of HRc 42–45. Before the sampling, the bars were soft
annealed. The bars of each heat were selected at the
beginning, middle and end of the rolling.
KIc-test specimens in the form of circumferentially
notched and fatigue-precracked tensile-test specimens
were cut from soft annealed bars with HB2.5/187.5274–277
in the rolling direction with the fatigue crack at the notch
root in the transverse direction.
The specimens were heat treated in a horizontal
vacuum furnace with uniform high-pressure gas-quen-
ching using nitrogen (N2) at a pressure of 5 bar. After the
first preheat (650 °C) the specimens were heated (10
°C/min) to the austenitizing temperature of 870 °C,
soaked for 10 min, gas quenched to a temperature of 80
°C, and then single tempered for one hour at different
temperatures between 200 °C and 575 °C, as shown in
Figure 2. Sixteen KIc-test specimens were tested for each
tempering temperature.
4 RESULTS AND DISCUSSION
The average measured hardness and fracture-tough-
ness data are shown for the range of hardness (HRc
30–56 ; i.e. Rm = 1275–2161 MPa) in a so-called com-
bined tempering diagram (Rockwell-C hardness – Frac-
ture toughness KIc – Tempering temperature for austeniti-
zing temperature 870 °C) in Figure 2.
From the diagram it is clear that the highest hardness
of HRc 58 and the associated fracture toughness KIc i.e.,
18.8 MPa m1/2, are achieved in the as-quenched condition
after vacuum quenching from the austenitizing
temperature of 870 °C. In examining the course of
tempering, it is observed that the minimum scatter of
results within ±2 = 3.2 MPa m1/2 is in the as-quenched
state and after single tempering at 200 °C. At higher
tempering temperatures between 300 °C and 575 °C, the
scattering of the KIc results slightly increased up to ±2 =
9.4 MPa m1/2, while the scatter of the Rockwell-C
hardness is up to ±2 = 1.2 HRc in the whole range of
used tempering temperatures. Within each group of
KIc-specimens, which were quenched and tempered at the
same temperature, this can be attributed to the kinetics of
carbide precipitation during tempering at selected tempe-
ratures as well as to the heterogeneity of the investigated
steel. Since the differences in chemical composition
(Table 1) are minimal and the austenitizing temperature
is the same, it is clear that the fracture toughness KIc is a
very selective mechanical property with regard to the
tempering temperature. It should be noted that the
KIc-test specimens were taken from the middle of the bar,
and therefore the microstructures of the KIc-test
specimens with the lowest and highest fracture toughness
are comparable.
In the as-quenched condition the microstructure
consists of untempered martensite and bainite, Figure 3.
Grain boundaries of the prior austenite grains are not
pronounced, and therefore the determination of the
austenite grain size according to ASTM E112 was not
possible.
Strong positive (bright) and negative (dark) segre-
gations are visible due to the lower etching intensity of
the untempered martensite.
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
70 Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73
Table 1: Chemical composition of continuous-cast spring steel grade 51CrV4 in mass fractions, w/%
Tabela 1: Kemi~na sestava kontinuirno litega vzmetnega jekla 51CrV4 v masnih dele`ih, w/%
Steel grades
51CrMoV4 C Si Mn P Cr Mo V Ni AL Cu Ti Sn Ca N
Heat 1
Heat 2
0.49
0.50
0.30
0.33
0.93
0.94
0.008
0.004
0.008
0.97
1.00
0.08 0.1
0.11
0.14
0.006
0.009
0.17
0.20
0.018
0.019
0.011
0.010
0,0010
0.0015
0.013
Figure 2: Tempering diagram for two heats of continuous-cast spring
steel grade 51CrV4
Slika 2: Diagram popu{~anja za dve {ar`i kontinuirno litega vzmet-
nega jekla 51CrV4
The fractured surface of the KIc-test specimens from
the same heat with the highest and lowest fracture
toughness tempered at 457 °C examined in binocular at
16-times magnification, Figure 4.
From Figure 4, it is evident that on the fractured
surface of the KIc-test specimen the highest fracture
toughness is higher with the density of positive segre-
gations (black features), the size of segregation is smaller
and the distribution is more even in comparison to the
fractured surface of the KIc-test specimen with the lowest
fracture toughness with nearly the same hardness.
The microstructure of the same KIc-specimens just
below the fractured surface was examined in the optical
microscope, Figure 5.
The microstructure in the quenched and tempered
condition consists of tempered martensite and bainite
( 20 %). In the microstructure, non-metallic inclusions
of the sulphide type can be observed that are located in
positive segregations and oriented in the rolling
direction.
The microhardness HV0.025 on the positive (bright)
and negative (dark) segregations of the KIc-test speci-
mens with the highest fracture toughness is shown in
Figure 6. The microhardness in the positive segregations
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73 71
Figure 5: Typical microstructure of KIc-test specimens a, b) specimen
with the lowest fracture toughness (KIc = 75.7 MPa m1/2, HRc 43.8) in
transverse direction and c, d) in longitudinal direction and e, f)
specimen with the highest fracture toughness (KIc = 82.2 MPa m1/2,
HRc 43.2) in transverse direction and g, h) in longitudinal direction
Slika 5: Zna~ilna mikrostruktura KIc -preizku{anca: a, b) z najmanj{o
lomno `ilavostjo (KIc = 75,7 MPa m1/2 in HRc 43,8), pre~na smer c, d)
vzdol`na smer; e, f) z najve~jo lomno `ilavostjo (KIc = 82,2 MPa m1/2
in HRc 43,2), pre~na smer; g, h) vzdol`na smer
Figure 3: The microstructure of the KIc-test specimen in the
as-quenched condition: a, b) specimen with the lowest fracture
toughness (KIc = 16.2 MPa m1/2 and HRc 58.9) c, d) specimen with the
highest fracture toughness (KIc = 22.3 MPa m1/2 and HRc 58.1),
transverse direction, natal
Slika 3: Mikrostruktura KIc-preizku{anca v kaljenem stanju: a, b) z
najmanj{o lomno `ilavostjo (KIc= 16,2 MPa m1/2 in HRc 58,9) c, d) z
najve~jo lomno `ilavostjo(KIc = 22,3 MPa m1/2 in HRc 58,1), pre~na
smer
Figure 4: Fractured surfaces of KIc-test specimen, tempered at 475 °C:
a, b) specimen with the lowest fracture toughness (KIc = 75.7 MPa
m1/2, HRc 43.8) c, d) specimen with the highest fracture toughness (KIc
= 82.2 MPa m1/2, HRc 43.2)
Slika 4: Prelomna povr{ina KIcpreizku{ancev, popu{~enih na 475 °C:
a, b) z najmanj{o lomno `ilavostjo (KIc = 75,7 MPa m1/2 in HRc 43,8)
c, d) z najve~jo lomno `ilavostjo(KIc = 82,2 MPa m1/2 in HRc 43,2)
was 554–584 HV0.025, in negative segregations it was
458 HV0.025. From the difference of the microhardness
one can conclude that the microstructure in positive
segregations consists of tempered martensite and in
negative segregations with a lower microhardness of
martensite and bainite.
The microstructure of the KIc -test specimens with the
highest fracture toughness was examined in the SEM on
positive and negative segregations in the longitudinal
direction, Figure 7.
At higher magnification it can be seen that the
microstructure in the positive segregations consist of
tempered martensite and in negative of tempered
martensite with islands of bainite, Figure 8.
From the comparison of the microstructures of the
KIc-test specimens with the lowest and highest fracture
toughness from the same heat and nearly the same
hardness and fracture toughnesses, it can be concluded
that the density, the size and the distribution of segre-
gations have a significant influence on the fracture
toughness. The increase in the fracture toughness can be
ascribed to the presence of bainite in the negative segre-
gations.
As mentioned above, if the fracture stress according
to equation (1) approaches the yield stress (e.g., if no
larger crack-nucleating particles are present), elastic-
plastic methods have to be applied. In this region the
material behaviour can be described with equation (2), if
the ultimate fracture strength is known, Figure 9.
In the diagram the decrease of the fracture strength is
observed when the critical defect size acr is exceeded,
i.e., acr = 50 μm and 100 μm, respectively.
5 CONCLUSIONS
The investigated spring steel grade 51CrV4 was
successfully vacuum quenched in a horizontal vacuum
furnace with uniform high-pressure gas-quenching using
nitrogen (N2) at a pressure of 5 bar. The obtained
as-quenched Rockwell-C hardness was HRc 58.4 ± 0.8,
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
72 Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73
Figure 9: Fracture stress f calculation for the investigated spring
steel tempered at various temperatures, ultimate fracture strength (u =
Ruc) was calculated from the Rockwell-C hardness according to DIN
50150.
Slika 9: Lomna napetost f za preiskovano vzmetno jeklo, popu{~eno
pri razli~nih temperaturah; kon~na natezna napetost (u = Ruc) izra~u-
nana iz trdote Rockwell-C, DIN 50150
Figure 7: SEM Microstructure of the KIc-test specimens with the
highest fracture toughness: a, b) positive segregations and c, d)
negative segregations, longitudinal direction
Slika 7: SEM-mikrostruktura KIc-preizku{anca z najve~jo lomno
`ilavostjo: a, b) pozitivna izceja, c, d) negativna izceja
Figure 8: Bainite grain from Figure 7
Slika 8: Kristalno zrno bainita s slike 7
Figure 6: Microhardness measurements on positive (bright) a) and
negative (darker), b) segregations of the KIc -test specimens with the
highest fracture toughness
Slika 6: Mikrotrdota a) na pozitivnih (svetli pasovi) in b) na nega-
tivnih (temnej{i pasovi) izcejah KIc-preizku{anca z najve~jo lomno
`ilavostjo
i.e., high enough to obtain, after a single tempering, the
required hardnesses from HRc 30 to HRc 50.
The obtained microstructure consists of tempered
martensite and bainite ( 20 %). Such a microstructure
is allowed according to the criteria TB1402 (Scania
Standard STD512090 and STD4153) permitting that the
microstructure consists of volume fraction   80 % of
martensite in the middle of the spring. This condition
was met for the investigated spring steel. From the
results we can conclude that the investigated spring steel
51CrV4 with a thickness up to 20 mm is suitable for heat
treatment in a vacuum furnace with uniform high-
pressure gas-quenching using nitrogen (N2) at a pressure
of 5 bar or higher.
The fracture toughness of the steel was determined
with a non-standard circumferentially notched and fati-
gue-precracked tensile-test specimen (KIc-test specimen).
In all cases the load-displacement relationship was linear
and the validity of equation (3) for the tests was con-
firmed. Because of the radial symmetry of the heat
transfer, the microstructure of the metallic material along
the circumferential area is completely uniform; there-
fore, the scatter of the fracture toughness could be
ascribed to the heterogeneity of the investigated steel
only.
For three tempering temperatures and the associated
ultimate fracture strength calculated from the Rock-
well-C hardness according to DIN 50150, the critical
defect size acr was estimated, i.e., acr = 50–100 μm,
respectively.
The fractured surfaces of the KIc-test specimens show
positive segregations (black features) and confirmed with
metallographic analyses as well as with the microhard-
ness HV0.025 measurements. The microstructure of the
positive segregations consists of tempered martensite
with a microhardness HV0.025554–584. In the microstruc-
ture, non-metallic inclusions of the sulphide type can be
observed, which are located in the positive segregations
and oriented in the rolling direction. The microstructure
in the negative segregations with a lower microhardness
consists of low-alloyed tempered martensite and bainite.
In analyzing the fracture surfaces of the KIc-test
specimens of the same heat with the highest and lowest
fracture toughness, tempered at the same tempering tem-
perature, as well as from the examined microstructure,
which confirmed the presence of positive and negative
segregations, we found that specimens with a higher
density of positive segregations (black features), which
are smaller in size and evenly distributed over the
fractured surface, exhibit up to 27 % higher fracture
toughness. This finding, which is also very interesting
for the automotive industry, will be the subject of a
further investigation.
Acknowledgement
[tore Steel, d. o. o., @elezarska cesta 3, SI-3220
[tore, Slovenia, is thanked for financial support as well
as for the supply of test materials. Thanks also to Prof.
dr. Franc Vodopivec and dr. Matja` Godec for helpful
discussions.
6 REFERENCES
1 A. Sandberg, M. Nzotta, High performance matrix tool steels
produced via a clean steel concept, 7th Tooling Conference, Torino,
May 2–6, 2006
2 P. Hansson, New multi-purpose pre-hardened tool steels, 7th Tooling
Conference, Torino, May 2–6, 2006
3 H. Jesperson, Toughness of tool steels, Proceeding of the 5th
International Conference on Tooling, Leoben, Sept. 29th to October
1st, 1999, University of Leoben, p. 93
4 R. Ebner, H. Leitner, F. Jeglitch, D. Caliskanoglu, Methods of
property oriented tool steel design, Proceeding of the 5th
International Conference on Tooling, Leoben, Sept. 29th to October
1st, 1999, University of Leoben, p. 3
5 J. F. Knott, Fundamentals of fracture mechanic, Butterworth, 1973
6 M. Janssen, J. Zuidema, R. J. H. Wanhill: Fracture mechanics, 2-nd
Edition, Delft University Press, Delft, Netherlands, 2002,15
7 B. Ule, V. Leskov{ek, B. Tuma, Eng. Frac. Mech., 65 (2000), 559/72
8 H. F. Fischmeister, L. R. Olson: Cutting tool materials, Proc. Int.
Conf. on Cutting Tool Materials, September 15–17(1980), Ft.
Mitchell, Kentucky, p. 111
9 S. Kargagöz, H. F. Fischmeister: Steel Research, 58 (1987) 8, 353
10 P. Billgren: Hot isostatic pressing of high speed steels, Proc. First
Int. High Speed Steel Conference, March 26–29, Leoben, Avstrija,
(1990), 115
11 V. Leskovsek, B. Ule, B. Liscic, Journal of Materials Processing
Technology, 127 (2002) 3, 298–308
12 Wei Shen, Tingshi Zhao, Daxing Gao, Dunkang Liu, Poliang Li,
Xiaoyun Qiu: Engng. Fract. Mech., 16 (1982) 1, 69–92
13 H. F. Bueckner: ASTM-STP, 381 (1965), 82
B. SEN^I^, V. LESKOV[EK: FRACTURE TOUGHNESS OF THE VACUUM-HEAT-TREATED SPRING STEEL 51CrV4
Materiali in tehnologije / Materials and technology 45 (2011) 1, 67–73 73

D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
SIMILARITY CRITERIA AND EFFECT OF LUBRICANT
INERTIA AT COLD ROLLING
MERILA PODOBNOSTI IN VPLIV VZTRAJNOSTI MAZIVA PRI
HLADNEM VALJANJU
Du{an ]ur~ija1, Ilija Mamuzi}1, Franc Vodopivec2
1Croatian Metallurgical Society, Berislavi}eva 6, 10000 Zagreb, Croatia
2Institute of Metals and technology, Lepi pot 11, 1000 Ljubljana, Slovenia
plutonijanac21@net.hr
Prejem rokopisa – received: 2010-09-01; sprejem za objavo – accepted for publication: 2011-01-17
Modern rolling mills operate at speed requiring the consideration of inertial forces for the explanation of the behaviour of
lubricant layer in the enter gap of the metal deformation zone. For this calculation a similarity criterium and the improved
Mizun-Grudev equation were applied. Good results were obtained in comparison with the numerical Monte Carlo method. A
significant point of the calculation is the definition of the initial state in the point of singularity.
Key words: cold rolling, lubrification, lubricant inertia, gripping angle, Reynolds equation, Monte Carlo method
Moderne valjarne obratujejo pri hitrostih valjanja, ki zahtevajo upo{tevanje sile vztrajnosti za razlago vedenja plasti maziva na
vhodni re`i zone deformacije metala. Za izra~un sta bili uporabljeni merilo podobnosti in Mizum-Gridevova ena~ba. Dobljeni so
dobri rezultati v primerjavi z numeri~no metodo Monte Carlo. Pomembna to~ka izra~una je definicija za~etnega stanja v to~ki
singularnosti.
Klju~ne besede: hladno valjanje, mazanje, vztrajnost maziva, kot oprijema, Reynoldsova ena~ba, metoda Monte Carlo
1 INTRODUCTION
The lubricant achieves in the enter gap of the metal
deformation1–5 zone a wedge shape determined with the
geometry of the rolls and the rolled sheet, as shown in
Figure 1. The flow of lubricant in the rolling gap can be
described with the simplified Reynol’s differential 6–9
equations:
∂
∂
=
∂
∂
p
x
v
y
x

2
2 (1a)
∂
∂
=
p
y
0 (1b)
∂
∂
+
∂
∂
=
v
x
v
y
x y
2
0 (1c)
Z
v
x
y C x
x= −
∂
∂
+∫ d ( ) (1d)
According to the differential equation (1b), the
pressure in the lubricant layer is constant over the gap
height and that it changes along the layer length, only,
and the approximate analytical solution of equation (1a)
is:
v
dp
dx
y
C y Cx = + +
1
2
2
1 2
( ) (1e)
Assuming boundary conditions from Figure 1
v v y
v v y x
x
x R
= =
= =
0 0
( )
(1f)
the integration constants are:
C
v v
x
dp
dx
xRx
1
0 1
2
=
−
−
 

( )
( )
(1g)
C v2 0= (1h)
Including (1g) and (1h) in (1d) we obtain
[ ]v dp
dx
y x y
v v
x
y vx
Rx= − +
−⎡
⎣⎢
⎤
⎦⎥
+
1 2 0
0



( )
( )
(1i)
The lubricant consumption along the strip perimeter
is:
Q x v dy
dp
dx
x
v v
xx
x
Rx
( ) ( ) ( )
( )
= = − +
+⎡
⎣⎢
⎤
⎦⎥∫0
01
2



  (1j)
For x = 0 Q
v v R=
+⎛
⎝
⎜ ⎞
⎠
⎟0
2
 (1k)
Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80 75
UDK 621.771:621.892 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 45(1)75(2011)
Figure 1: Scheme of cold rolling with lubricant
Slika 1: Shema hladnega valjanja z mazivom
Equalizing (1j) and (1k) we obtain3–9:
( )d
d
p
x
v v
x
Q
x
R=
+
−
6 120



 
( ) ( )
(1l)
The equation was solved applying numerical
methods10,11 for dressing rolling including inertial forces.
The solution is:
( )d
d
p
x
v v
x
C
x
R=
+
+ +
6 0 1



 
( ) ( )
[ ]+ + −x
R x
v v x C


120
16 0
2 2
1
2
( )
( ) ( )R (2)
[ ]C k k v v v v kR R1
2
0 0 0 02 4
2 8 3= − + + + +( ) ( )  (3)
k vR
x
=120
(4)
The lubricant thickness in the gap range (–a, 0) is:
  
 
( ) cos sinx R
x
R
= + − − −⎛⎝
⎜ ⎞
⎠
⎟
⎧
⎨
⎩
⎫
⎬
⎭
0
2
1 (5a)
This equation can be developed in the power series:
  ( ) ...x x
R
x
R
x= − + − +0
2
2
31
2
1
2
(5b)
Neglecting inertia forces, analytical solutions can be
developped in the form:
[ ]    
 
0 01 01 01) ( ) /d = − − (6)
[ ] 

  




  	   

−
+
⎡
⎣⎢
⎤
⎦⎥
+ −
2
1 3
2
3 2
0
1 2 2 0
2 2
R x R
R/ / (7)
With: R/m – rolls diameter, v0 and vR/(m/s) – rolling
speeds, μ0/(Pa s) – dynamical viscosity of the lubricant,
/(kg/m3) – fluid (lubricant) density, v/(m2/s) – kinema-
tical viscosity of the lubricant, 
 (rad) – rolling grip
angle, A/m–1 – technological parameter, 0/m – lubricant
layer thickness in the initial section of metal deformation
zone, 10/m – lubricant layer thickness for 
→ 0, vx and
vy/(m/s) – speeds in Descartes coordinates x and y,
/(m2/N) – pieso coefficient of lubricant viscosity, 0∗ –
lubricant layer thickness in the singularity point 
∗,
p0/(N/m2) – rolling pressure. The parametric symbols are
explained in Table 1.
These equations can be used for calculations and
computer modelling of the behaviour of the lubricant
layer in the zone of plastic deformation1,10,11,12 of metals
with cold rolling.
The characteristics of thin sheets rolling process
allow to find analytical solutions for equation (1l), how-
ever, it is difficult to find a solution for equation (2). For
lubricated rolling of sheets, the analytical solutions are
acceptable for high gripping angles, while simulation
models are mostly used for continous rolling 12,13.
2 SIMILARITY CRITERIA
Similarity criteria are frequently used in fluid meca-
nics calculations. Basic data for calculation related to the
behaviour of the lubricant in the metal deformation zone
in Table 2 for the term M0 and for the transcendent
equation (7) were deduced for the following processing
conditions: A = 1.965512 · 106 m–1, v0 = 6 m/s, vR = 10
m/s, 
i = (m)±(1/3)
. The iso values are based on investi-
gations aimed to obtain a better equation than that of
Mizunov-Grudev given in Table 2 and analytically more
acceptable than the transcendent equation (7). The
D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
76 Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80
Table 1: Parametrical symbols for Figure 1 and equations (1) to (7)
Tabela 1: Parametri~ni simboli za sliko 1 in ena~be od (1) do (7)
01 (2R/128A2)1/3
0 (1/2)R(
)2

 (8/15RA)1/3
D Square determinant of equation (6) forthe members (
 : 0)
 –
 (2/R)0 – 
2
 ln–
 – ()1/2/(–
 + ()1/2
A (1 – exp(–p0))/60(v0 + vR)
Linear interaction 

  1.246(01/R)1/2 ; 0  0.7726 01;
0 = R(
)2/2
Mizun-Grudev
equation 
M
0 = 1/2A

(0)d Linearisation for the dressing rolling
Figure 2: Comparison of different methods for the calculation of the
lubricant layer thickness (R = 0.35 m, A = 898519 m–1. 1 – Mizunov-
Grudev equation, 2. o – improved Mizunov-Grudev equation (8), 3 –
method of linearisation of the equation (6), 4. ! – numerical Monte
Carlo solution.
Guide R = 0.35 m A = 1965512 m–1
Standard R = 0.35 m A = 800 000 m–1
Clone R = 0.35m A = 898 519 m–1
Slika 2: Primerjava razli~nih metod za izra~un debeline plasti maziva
(R = 0.35m, A = 898 519 m–1. 1 – Mizunov-Grudev ena~ba, 2. o –
izbolj{ana Mizunov-Grudev ena~ba (8), 3 – metoda of linearizacije
ena~be (6), 4. ! – numeri~na Monte Carlo re{itev
Vodilo R = 0.35 m A = 1 965 512 m–1
Standard R = 0.35 m A = 800 000 m–1
Klon R = 0.35m A = 898 519 m–1
authors have developed the following aproximation for
this equation:
4 10
2 00
2
0
1V
R
Aw w
−⎛
⎝
⎜ ⎞
⎠
⎟ + − =−


 
)
V M=
⎛
⎝
⎜
⎞
⎠
⎟ln
Re2 2
0



(8)
With: e – natural logarithm base.
Tests have shown that the error is for equation (8) in
comparison to equation (7) smaller than 1 % for gripping
angles 
 > 0.03 rad. Further, equation (8) preserves the
same law of iso values than the Mizunov-Grudev
equation. It follows, that by applying the criteria of iso
values it is possible by cold rolling to pass over from a
greater to a smaller gripping angle and to calculate the
lubricant layer thickness according to equation (8) using
the relation 
i = (m)±(1/3)
. The values in Table 2 cannot
be calculated using equation (1) developed to a poly-
nome of the third or higher order but only to the square
polynome according to equation (5b).
D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80 77
Table 2: Etalon (standard) for theoretical investigations of similarity criteria (m = R/Ri)
Tabela 2: Etalon (standard) za teoreti~no raziskavo meril podobnosti (m = R/Ri)
R/m
Rolling grip angle/rad
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
0.05 1.725186 1.385324 1.234872 1.155365 1.108717 1.079329 1.059809
0.1 1.510205 1.259770 1.153034 1.098537 1.067490 1.048406 1.035988
0.15 1.410961 1.203689 1.117577 1.074549 1.050460 1.035857 1.026462
0.2 1.350680 1.170418 1.096966 1.060832 1.040848 1.028848 1.021185
0.25 1.309130 1.147908 1.083234 1.051802 1.034578 1.024310 1.017789
0.30 1.833992 1.278294 1.131459 1.073322 1.045345 1.030127 1.021105 1.015400
0.35 1.774128 1.254262 1.118808 1.065776 1.040466 1.026783 1.018708 1.013621
0.40 1.725186 1.234872 1.108718 1.059809 1.036633 1.024169 1.016841 1.012238
0.45 1.684164 1.218813 1.100447 1.054955 1.033532 1.022062 1.015341 1.011130
0.50 1.649108 1.205241 1.093520 1.050917 1.030965 1.020324 1.014107 1.010220
0.55 1.618683 1.193583 1.087620 1.047498 1.028800 1.010886 1.013072 1.009458
0.60 1.591942 1.183434 1.082523 1.044561 1.026947 1.017615 1.012190 1.008810
- - - - - - - - -
0.8 1.510205 1.153034 1.067490 1.035988
Table 3: Comparison of different methods to the Monte-Carlo method for bridging the problematic area in Figure 3 for a = 0.02 to a = 0.035 rad
Tabela 3: Primerjava razli~nih metod z metodo Monte-Carlo za premostitve problemati~ne povr{ine na sliki 3 za a = 0.02 do a = 0.035 rad
Gripping angle / rad Equation (6) Method of linearisation Monte-Carlo Equation (8)

 = 0.02 19.909 · 10–6 19.949 · 10–6 19.916 · 10–6 -

 = 0.025 - 17.532 · 10–6 17.499 · 10–6 -

 = 0.03 - 15.536 · 10–6 15.509 · 10–6 15.471 · 10–6

 = 0.035 - 13.884 · 10–6 13.866 · 10–6 13.716 · 10–6
Second part

 = 0.04 - 12.304 · 10–6 12.501 · 10–6 12.341 · 10–6

 = 0.045 - 11.342 · 10–6 11.353 · 10–6 11.215 · 10–6

 = 0.05 - 10.352 · 10–6 10.387 · 10–6 10.271 · 10–6

 = 0.055 - - 9.560 · 10–6 9.467 · 10–6
Third part
Derivation method Monte-Carlo Equation (8)

 = 0.07 7.884 · 10–6 7.686 · 10–6 7.639 · 10–6

 = 0.08 6.950 · 10–6 6.784 · 10–6 6.754 · 10–6

 = 0.09 6.224 · 10–6 6.067 · 10–6 6.046 · 10–6
Figure 3: Modelled solutions for Figure 2 at 
 = 0.03 rad. Same
denotations as in Figure 2. The tool REVSURF allows the rotation
around the linearisatioin method (curve 3).
Slika 3: Modelirane re{itve za sliko 2 pri 
 = 0.03 rad. Iste ozna~be
kot na sliki 2. Orodje REVSURF omogo~a rotacijo okoli metode
linearizacije (krivulja 3).
In Figure 2 the comparison of results of the calcu-
lation of the lubricant layer thickness according to
different methods is shown. The figures show that the
linearisation method can be applied only for the dressing
rolling. For cold rolling the Mizunov-Grudev solution
differs from the numerical solution for aproximately 
 =
0.06 rad, while equation (8) differs for only 
 = 0.025
rad. For the range 
 = 0.02 to 
 = 0.03 rad on Figure 2
the criterium of similarity can be used and an approxi-
mate solution can be obtained with interpolation from
known values.
The criterium of similarity is based on equation (6)
and the selected transfer function was a hyperbola. The
thickness of the lubricant layer is calculated using the
linearisation method and applying the standard data in
Table 2 over the guider to the case on Figure 2. The
hyperbola is used to transmit the lubricant layer thick-
ness defined by the equation (8), as in this case, or using
the boundary conditions for a = 0.02 and a = 0.03, as
shown in Table 3. In this table the comparison is given
of the lubricant layer thickness calculated using the
similarity criteria and the Monte Carlo method for the
gripping angles indicated. The guider is defined for R =
0.35 and A = 8 · 105 m–1. It is clear from Table 3 that the
methods of linearisation according to equation (6), of
similarity criteria and of the improved Mizunov-Grudev
equation are complementar and represent an algorithm
for the thickness of the lubricant layer suitable for fast
practical application.
In Figure 3 the AutoCAD modelling of Figure 2
using different modelling methods is shown. In the initial
part of abscissa a good approximation is obtained
between the numerical and the linearisation method and
that in the border part of abscissa with 
 = 0.03 the
agreement between the corrected equation (9) and the nu-
merical Monte-Carlo calculation is acceptable. Straight
lines obtained with linear programming and derivation at
the point of singularity are shown in Figure 4. The linear
representation was selected to obtain the optimal transfer
of similarity criteria to the cylinder-clone for which the
solution of differential equation is obtained using the
known solutions for the guider and the standard on the
base of only one solution for the singular point (
, 0).
This approach is valid for gripping angles in the
range 0.0124 rad to 0.0404 rad with allowed angle
change for 
  0.00362 rad. In the second part of
Table 3 the transfer of similarity criteria from the
standard over the guider to the cylinder clone using the
straight lines 4, 5 and 6 in Figure 5, is shown. If the
strainght lines transferring the accuracy from the first to
the fourth quadrant become unreliable for the transfer of
similarity criteria, it is possible to use, as help, the
singular point derivation with direction of transfer of
similarity criteria to the first quadrant for which the
calculation of the lubricant layer thickness is performed.
The singular point transfers well the calculation of the
lubricant film, although the gripping angle is relatively
distant. It is usefull to remind that the method of deri-
vation in Table 3 offers the possibility of increasing the
mathematical accuracy in comparison to the Monte-
Carlo method. In this case, the simple analytical defi-
nition would approach a form approaching equation (8).
Further, data in Table 3 show that the improved equation
(8) can, for smooth surfaces of rolls and of rolled metal,
substitute the Monte Carlo method. In this case, the
transfer of similarity criteria for the standard and the
guider can be calculated using equation (8), while the
clone cylinder is defined mathematically with differential
equations connecting the longitudinal and transverse
roughness and with possible analytical solutions for the
point of singularity.
The solution of differential equations related to the
equation (1a), as for instance equation (2), can be
obtained through the solution for one point, as boundary
condition, in this case the solution for the singular point.
If the accuracy as in Table 1 is expected, the calculation
of the lubricant layer for smooth surfaces can not be
simplified further.
Using a complex method of statistical analysis of
data in Table 2 and for the rolls of diameter of R = 0.3 m
the following relation was developed for the lubricant
layer thickness:
D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
78 Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80
Figure 5: Relationship lubricant layer thickness – rolling gap geome-
trical shape
Slika 5: Odvisnost debeline plasti maziva od geometri~ne oblike va-
ljalni{ke re`e
Figure 4: The method of linearisation (1, 2, 3) through the points (0 :
10) and (a : 0) and of derivation in the point (a : 0) (4, 5, 6)
Standard (3, 4), guider (1, 5), Clone cylinder (2, 6)
Slika 4: Metoda linearizacije (1, 2, 3) skozi to~ke (0 : 10) in (a : 0)
in derivacije v to~ki (a : 0) (4, 5, 6).
Standard (3, 4), vodilo (1, 5), valj klon (2, 6)



 
 
 

0
0
1 2 1 2 2
M
T S B / C L /= + + +( ) ( )
/ /
Q = 
1 2/ W = ( / )  
0 0
2M T (9)
The results of the calculation are shown in Table 4.
Table 4: Some statistical data on the regression analysis
Tabela 4: Nekateri statisti~ni podatki o regresijski analizi
S 0.8519
B –0.2633
C 0.0346
L –0.0015
Rsq 0.99999
F 66829.2
d. f. 5
Q, W Abscissa and ordinate
3 EFFECTS OF LUBRICANT INERTIA
The effect of lubricant inertia depends on the speed
of the processing, the gap geometry and the rheological
characteristics of the lubricant. In this work only
investigations of the effect in the gap geometry is
discussed. It was deduced with a numerical solution of
the equation (1) applying the Monte-Carlo method with
A = 1 965 512 m–1, R = 0.2 m,  = 854 kg/m3, dressing
angle 0.02 rad and other data as in Table 2 and it is
shown in Figure 4.
The reliable analysis range lies both sides near of the
point 2 in Figure 5. On the arc 2"3 of the curve, by
increasing of the strip lubricant layer thickness, the
thickness of the lubricant layer in the initial section of
the metal deformation zone is increased and the effect of
inertia forces is decreased, while, on the arc 2"1 of the
curve the effect is opposite. For a selected processing,
the vertex at the point 2 can be deduced with a regression
analysis for determined processing parameters. In Table 5
data on the effect of inertia forces on 0 calculated from
equations (1a) and (2) using the Monte-Carlo method are
given. The calculation was similar as for the curves in
Figure 2 and the data in Table 2. For the arc 2"3 and
dressing processing the correction for the lubricant layer
thickness would be unnecessary, while, considering the
inertia forces it is significant for the arc 2"1. If the
working velocity is increased from 16 m/s, as in Table 5,
to 50 m/s for the angle 0,05 rad 0 = 13.854 · 10–6 m is
deduced, while the consideration of inertia forces gives
IN0 = 14,951 · 10–6 m.
Data in Table 5 show that the shape of the gap in the
area (–a, 0) prevails over the lubricant rheological cha-
Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80 79
D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
Figure 6: Change of pressure gradient for the gripping angler 
 = 0,02
rad and rolls diameter R = 0.2 m in dependence of the ratio (–a) / (–a)
in the range of – 0.001 m to – 0.000015 m and other parameters as for
Table 2
Slika 6: Sprememba gradienta pritiska za kot oprijema 
 = 0,02 rad in
premer valjev R = 0,2 m v odvisnosti od razmerja (–a)/(–a) v obmo~ju
od –0.001 m do – 0.000015 m in drugih pamaterov kot za tabelo 2
Table 5: Effect of inertia forces(IN0) for dressing rolling
Tabela 5: Vpliv sil vztrajnosti (IN0) za oblikovalno valjanje

/rad 0 0.011335 0.02 0.03 0.04 0.05
0 /m 15.863 · 10–6 12.255 · 10–6 9.416 · 10–6 7.244 · 10–6 5.797 · 10–6 4.695 · 10–6
IN0 /m 15.863 · 10–6 12.335 · 10–6 9.542 · 10–6 7.447 · 10–6 5.959 · 10–6 4.906 · 10–6
Table 6: Dependence of the lubricant layer thickness on the gripping angle and the lubricant layer thickness on the strip ahead the deformation
zone. Calculated using equations (1a) i (2).
Tabela 6: Odvisnost debeline plasti maziva od kota oprijema in debeline plasti maziva na traku pred zono deformacije. Izra~unano z uporabo
ena~b (1a) in (2).

/rad a/ m a/m a/m a/m a/m
0.001 0.002 0.003 0,004 0.005
Without inertia 12.613· 10–6 12.642· 10–6 12.648· 10–6 12.651· 10–6 12.652· 10–6

 = 0.01 12.793 · 10–6 12.853 · 10–6 12.884 · 10–6 12.903 · 10–6 12.912 · 10–6
Without inertia 9.390 · 10–6 9.407 · 10–6 9.412 · 10–6 9.414 · 10–6 9.415 · 10–6

 = 0.02 9.491 · 10–6 9.539 · 10–6 9.562 · 10–6 9.578 · 10–6 9.589 · 10–6
Without inertia 7.225 · 10–6 7.238 · 10–6 7.241 · 10–6 7.242 · 10–6 7.243 · 10–6

 = 0.03 7.286 · 10–6 7.322 · 10–6 7.340 · 10–6 7.352 · 10–6 7.361 · 10–6
Without inertia 5.786 · 10–6 5.793 · 10–6 5.795 · 10–6 5.796 · 10–6 5.796 · 10–6

 = 0.04 5.820 · 10–6 5.848 · 10–6 5.862 · 10–6 5.871 · 10–6 5.878 · 10–6
Without inertia 4.788 · 10–6 4.793 · 10–6 4.794 · 10–6 4.795 · 10–6 4.795 · 10–6

 = 0.05 4.809 · 10–6 4.830 · 10–6 4.841 · 10–6 4.848 · 10–6 4.854 · 10–6
racteristics and the kinematics of the processing because
the gripping angle approaches to zero. For this reason,
the inertia forces do not affect significantly the lubricant
behaviour by dressing processes, while these forces
shold be considered by cold rolling. In Table 6 the effect
of a (lubricant layer thickness on the sheet) on 0 (lubri-
cant layer thickness at entrance cross section) is shown,
as deduced using equation (2) and the Monte-Carlo
calculation with R = 0.2 m and other data, as for Table 5.
The value of 0 is increased for 1–2 %, while the value
of a increases for the sheet for five times for the rolling
velocity up to 16 m/s.
In Table 6 the effect of lubricant height on the sheet
on the lubricant sheet on the entering section of the
deformation zone is shown as function of lubricant
inertia and gripping angle. The inertia effect increases
significantly with the gripping angle and even faster with
the lubricant height on the sheet ahead the rolls.
On Figure 6 the change of pressure gradient ahead
the rolling gap is shown with the maximum for x =
–0.00025 m.
It is very difficult to obtain approximate analytical
solutions of the differential equation (2). For this reason,
it is analysed applying the numerical Monte-Carlo
integration. The analysis should be complemented13
considering surface roughness and with correlations with
dependences to dynamkcal viscosity and rolling pre-
ssure.
4 CONCLUSIONS
On the base of the results of the calculations pre-
sented the following conclusions are proposed:
– The criteria of similarity are transferred acceptably
with the solution of the equation (1a) from the
standard over the guider to the cylinder clone
applying the method of linearisation up to a gripping
angle determined by linear programming. For a
greater gripping angle the transfer of similarity
criteria can be achieved using straight lines obtained
with derivation in the point of singularity.
– The effect of inertia forces can be neglected for
lower dressing speed, while this effect becomes
significant for incresead speed of cold rolling.
– The use of the improved Mizunov-Grudev equation
(8) gives solutions in good agreement with the
numerical solution of equation (1) applying the
Monte-Carlo method. Also, it is more practical for
use than the transcendent equation (7).
– For a determined relation rolls diametere versus
gripping angle very similar results are obtained
using the Mizunov – Grudev relation and the trans-
cendent equation (7).
5 REFERENCES
1 D. ]ur~ija, I. Mamuzi}, Schemes of metal-working processes and the
related tribological equations of fluid mechanics, Mater. Tehnol., 43
(2009) 1, 23–30
2 R. Boman, Jean-Philippe Ponthot, Numerical simulation of lubri-
cated contact in rolling processes, J. Mater. Process. Techn., 125–126
(2002), 405–411
3 S. Thiruvarudchelvan, M. J. Tan, Recent developments in friction-
assisted sheet metal forming processes, J. Mater. Process. Techn.,
167 (2005) 2–3, 161–166
4 Dengyue Sun, Xun Chen, Fengshan Du, Guangming Zhu, Jiangang
Zhou, Dynamic thickness simulation of five-stand continuous cold
rolling, J. Mater. Process. Techn., 167 (2005) 2–3, 218–223
5 W. R. D. Wilson, Friction and lubrication bulk metal forming pro-
cess, J. Appl. Met. Work, 1 (1979), 7–19
6 D. ]ur~ija, I. Mamuzi}, Lubrication flow during the rolling of seam-
less tubes, Mater. Tehnol., 42 (2008) 2, 59–63
7 S. V. Mazur, Postanovka zada~i i zakonomernosti te~enija smazki v
o~age deformacii pri prokatke trub, Su~asni problemi metalurgii,
Plasti~na deformacija metaliv, 8 (2005), 447–452
8 O. P. Maksimenko, A. A. Semen~a, Issledovanie kontaktno-gidro-
dinami~eskoj smazki pri prokatke, Su~asni problemi metalurgii,
Plasti~na deformacija metaliv, 8 (2005), 99–103
9 W. Huang, D. B. Bogy, Three-dimensional direct simulation Monte
Carlo methods for slider air bearings, Phys. Fluids, 9 (1997) 6,
1764–1769
10 M. I. Sobol: Die Monte Carlo Metode, Ver. H. Deutsch, Frankfurth a.
Main, 1986
11 D. ]ur~ija, I. Mamuzi}, Mathematical Modelling of Metal Drawing
Process, Metalurgija, 44 (2005) 2, 113–118
12 M. R. Jensen, L. Olovsson, J. Danckert, Numerical model for the oil
pressure distribution in the hydromechanical deep drawing process,
J. Mater. Process. Techn., 103 (2000), 74–79
13 P. Heyer, J. Läuger, Corelation between friction and flow of
lubricating greases in a new tribometer device, Lubrication Science,
21 (2009) 7, 253–268
The authors are indebted to professor Marica Pe{i}
and professor Desanka Radunovi} of the Faculty of
Natural Sciences and Matematics of the University of
Belgrade for reviewing the mathematical treatment.
D. ]UR^IJA, I. MAMUZI]: SIMILARITY CRITERIA AND EFFECT OF LUBRICANT INERTIA ...
80 Materiali in tehnologije / Materials and technology 45 (2011) 1, 75–80
In memoriam
OSKAR KÜRNER
1925–2010
Konec septembra nas je zapustil Oskar Kürner, univ.
dipl. in`. metalurgije in vi{ji strokovni sodelavec v
Raziskovalnem oddelku @elezarne Jesenice.
Rodil se je 21. 10. 1925 v Ljubljani. Po diplomi leta
1956 se je zaposlil v @elezarni Jesenice kot asistent
martinarne in kasneje v metalur{kem oddelku tehni~ne
kontrole. V istem letu je bil imenovan za vi{jega stro-
kovnega sodelavca in leta 1983 za vodjo sektorja teh-
ni~ne kontrole, kjer je ostal do upokojitve leta 1989.
Njegovo strokovno delo je bilo pestro in inovativno.
Najprej je prevzel nalogo, da dopolni proizvodnjo
elektroplo~evine, tako da je bilo opu{~eno valjanje jekla
v obliki plo~evine, ki je bilo nadome{~eno z izdelavo v
obliki hladno valjanih trakov. Tehnolo{ko je dopolnil
izdelavo jekla tako, da se bile elektromagnetne karakte-
ristike jekla v skladu s takratnimi DIN-normami.
Postavil in opremil je eksperimentalno linijo v razis-
kovalnem oddelku, tako da je bilo mogo~e spreminjati
posamezne parametre za doseganje najbolj{ih elektro-
magnetnih karakteristik jekla.
Brez dvoma je s svojimi rezultati prispeval k tehno-
logiji procesov v hladni valjarni na Beli in so se rezultati
kot mozaik vgrajevali v temelje razvoja hladne valjarne
Bela in kasneje v ACRONI ter tudi v odlo~itve o konfi-
guraciji tehnolo{kih procesov v novi jeklarni.
Usmeril se je predvsem v razvoj elektroplo~evin, od
mehkega relejnega `eleza, siliciranih dinamoplo~evin do
poizkusov izdelave transformatorske plo~evine v skladu
z mednarodnimi standardi. Uspelo mu je pridobiti nove
merilne naprave. Za~el je ustvarjati interne predpise in z
njimi dosegati kvaliteto elektroplo~evin po mednarodnih
standardih. Z novimi merilnimi napravami so se oprav-
ljale meritve po metodi Epstein in z njimi je lahko
posredoval porabnikom elektroplo~evin podatke o
kvaliteti, ki so bili kompatibilni s svetovnimi normami za
te materiale.
Da bi lahko uresni~il nadaljnji razvoj elektroplo-
~evine v `elezarni, je projektiral in v okviru eksperi-
mentalne delavnice v oddelku tudi udejanil pripravo
laboratorija za te namene, saj je bilo mogo~e `arenje
plo~evin v razli~nih plinskih me{anicah in pri razli~ni
vsebnosti vlage. V `arilni pe~i je bilo mogo~e po
kontinuirnem postopku izvajati razli~no toplotno
obdelavo za doseganje optimalnih magnetnih lastnosti
plo~evine ter pripraviti tehnolo{ka navodila za indu-
strijsko proizvodnjo. Ugotavljati je bilo mogo~e tudi
vpliv oligoelementov na proces staranja jekla in vpliv na
elektromagnetne karakteristike.
In`enir Kürner je deloval tudi na naslednjih pod-
ro~jih:
– vodil je raziskovalni oddelek, ki ga je opremil za
izvajanje posebnih raziskovalnih in tehnolo{kih
posegov;
– ustanovil je oddelek za statisti~ne preiskave in
spremljanje tehnolo{kih parametrov v proizvodnji
jekla;
– uvedel je pripravo interne tehnolo{ke dokumentacije
za izdelavo in predelavo jekla v obliki regulativov
oziroma internih standardov; ti so obsegali tudi inter-
no kontrolo proizvodov v `elezarni.
– re{eval je teko~o proizvodno problematiko v `elezar-
ni in sodeloval pri razvoju novega jekla.
– v letu 1961 je za~el razvijati novo vrsto jekla za
rudarske verige in jekla za krogli~ne le`aje; rudarske
verige so bile izvozni izdelek tovarne.
Delo in`. Kürnerja je bilo zelo raznoliko in obse`no,
saj je bil vklju~en v re{evanje razli~nih proizvodnih
problemov, ki so se pojavljali v vsakdanji proizvodnji.
Bil je mehkega zna~aja, razumel je te`ave in stiske
sodelavcev, pod njegovim vodstvom pa so se le-ti tudi
strokovno izpopolnjevali.
Za svoje raziskovalno delo je prejel Pantzovo
priznanje za leto 1981, najvi{jo nagrado, ki mu jo je
podelila @elezarna Jesenice.
Veliko dogodkov je povezanih z delom pokojnega
Oskarja. Zato lahko trdimo, da je bil eden izmed za-
slu`nih stebrov, ki so nosili razvoj `elezarne na Jesenicah
in s tem ohranjali `elezarstvo na Gorenjskem. V razvoj
stroke je vlo`il veliko truda in znanja, ki se prena{a na
naslednike in je to dejansko kapital podjetja.
Nam, ki smo ga poznali, bo ostal v trajnem spominu.
Prof. dr. in`. Marin Gabrov{ek
Materiali in tehnologije / Materials and technology 45 (2011) 1, 81 81