VSEBINA – CONTENTS
PREGLEDNI ^LANKI – REVIEW ARTICLES
Environmental catalysis from nano- to macro-scale
Okoljska kataliza od nano- do makrovelikosti
A. M. T. Silva. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
IZVIRNI ZNANSTVENI ^LANKI – ORIGINAL SCIENTIFIC ARTICLES
Pressureless sintering and characterization of B4C, TiC and TiB2-particle-reinforced TiAl3-matrix composites
Sintranje in karakterizacija kompozitov na osnovi TiAl3, oja~anih z delci B4C, TiC in TiB2
V. Kevorkijan1, S. D. [kapin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Magnetic characteristics of isothermally aged Cr-Ni-Mo-based alloys with different δ-ferrite contents
Magnetne lastnosti izotermno `arjenih zlitin Cr-Ni-Mo z razli~no vsebnostjo δ-ferita
B. [u{tar{i~, B. Podmilj{ak, P. McGuiness, J. V. Tuma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
The oxidation of duplex stainless steel at moderately elevated temperatures
Oksidacija dupleksnega nerjavnega jekla pri zmerno povi{anih temperaturah
^. Donik, A. Kocijan, I. Paulin, M. Jenko . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Optimization of weld bead geometry in TIG welding process using Grey relation analysis and Taguchi method
Optimizacija geometrije TIG-varkov z Greyjevo analizo in Taguchijevo metodo
U. Esme, M. Bayramoglu, Y. Kazancoglu, S. Ozgun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Primary etalonnage of negative gauge pressures using pressure balances at the Czech metrology institute
Primarne kalibracijske metode za negativni relativni tlak s tla~nimi tehtnicami na ^e{kem in{titutu za metrologijo
J. Tesa , Z. Krají~ek, D. Pra`ák, F. Stank. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
The geological record as an indicator of the mudstones thermal characteristics in the temperature range of decarbonatisation
Geolo{ki zapis kot pokazatelj termi~nih lastnosti laporovcev v temperaturnem obmo~ju dekarbonatizacije
@. Poga~nik, J. Pav{i~, A. Meden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
STROKOVNI ^LANKI – PROFESSIONAL ARTICLES
Kerami~ni kompoziti na osnovi silicijevega nitrida
Ceramic composites based on silicon nitride
A. Maglica, K. Krnel, M. Ambro`i~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Investigations of micro-alloyed cast steels
Raziskave mikrolegiranih jeklenih litin
B. Chokkalingam, S. S. Mohamed Nazirudeen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
17. MEDNARODNA KONFERENCA O MATERIALIH IN TEHNOLOGIJAH, 16. – 18. november, 2009, Portoro`, Slovenija
17th INTERNATIONAL CONFERENCE ON MATERIALS AND TECHNOLOGY, 16–18 November, 2009, Portoro`, Slovenia . . . . . 175
ISSN 1580-2949
UDK 669+666+678+53 MTAEC9, 43(3)111–176(2009)
MATER.
TEHNOL.
LETNIK
VOLUME 43
[TEV.
NO. 3
STR.
P. 111–176
LJUBLJANA
SLOVENIJA
MAY–JUNE
2009

A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
ENVIRONMENTAL CATALYSIS FROM NANO- TO
MACRO-SCALE
OKOLJSKA KATALIZA OD NANO- DO MAKROVELIKOSTI
Adrián Manuel Tavares da Silva
Laboratório de Catálise e Materiais (LCM), Laboratório Associado LSRE/LCM, Departamento de Engenharia Química, Faculdade de
Engenharia, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal
adrian fe.up.pt
Prejem rokopisa – received: 2008-09-25; sprejem za objavo – accepted for publication: 2009-03-31
In today’s highly sensitive ecosystem, environmental protection is a great challenge to be faced by industry, with regulations
becoming more restrictive in the future. Advanced oxidation processes (AOPs) – or technologies (AOTs) – are promising
end-line solutions for waste water treatment, involving liquid-phase oxidation of pollutants by means of highly reactive species,
such as hydroxyl radicals. In general, installation costs of AOPs are comparable with the ones associated to biological
treatments; however, operating costs of AOPs are higher due to the use of costly chemicals. Development of highly active
heterogeneous catalysts is a key approach to scale-up cost-effective environmental technologies, minimizing the needs in
chemicals and increasing the level of treatment obtained. This paper is an overview of some different steps involved in the field
of environmental catalysis engineering to develop suitable materials from a nano-scale point to a macro-scale application.
Keywords: environmental protection; heterogeneous catalysts; advanced oxidation processes (AOPs);
V dana{njem ob~utljivem ekosistemu je varovanje okolja pomemben izziv, ki ga sre~uje industrija, {e posebej zaradi predpisov,
ki bodo v prihodnosti {e bolj omejujo~i. Napredni oksidacijski procesi (AOP) in tehnologije (AOT) so obetajo~e kon~ne re{itve
za obdelavo odpadnih voda in temeljijo na oksidaciji ne~isto~ z visoko reaktivnimi snovmi, kot so hidroksilni radikali.V
splo{nem so stro{ki AOP-naprav podobni tistim za biolo{ko obdelavo, vendar so obratovalni stro{ki AOP vi{ji zaradi uporabe
dragih kemikalij. Razvoj visokoaktivne heterogene katalize je klju~ni pogoj za napredek cenovno u~inkovite okoljske
tehnologije, ker minimizira porabo kemikalij in pove~a kakovost obdelave. Ta ~lanek je pregled razli~nih stopenj na podro~ja
in`enirstva okoljske katalize, ki ima za cilj razviti primerne materiale z nano- do makroobmo~jem uporabe.
Klju~ne besede: varovanje okolja, heterogena kataliza, napredni oksidacijski procesi (AOP)
1 INTRODUCTION
The treatment of effluents is one of the major social,
economical and political problems of our days. As con-
sequence, many researchers have focused their scientific
interests in environmental issues. Scientific journals with
high impact factors, as well as traditional and new
congresses include now environmental themes. This was
a crucial step in order to alert industry stakeholders to
the importance of reducing the industrial impact on the
environment, assuring a good quality of life. The
objective of the so-called green chemistry is mainly
directed to emerging technologies by designing
environmentally benign processes 1. However, operation
issues, deep research, long time consuming and high
economical investment are some shortcomings found at
industrial level when applying this concept to processes
already installed in traditional industries. Consequently,
the treatment of pollutants by remedial end-line techno-
logies is still important.
Conventional biological processes have associated
significant disadvantages, such as inhibition of
microorganisms by certain pollutants, impossibility to
treat high concentrations of organic compounds and
production of sludge, which is another source of
pollution. AOPs are alternatives in this domain. In brief,
the Fenton process consists in the chemical oxidation of
pollutants by means of hydrogen peroxide (H2O2) and
iron acting as catalyst, based in the Fenton reaction
known for over 100 years 2; ozonation is based in the
strong oxidant properties of ozone; photocatalysis uses
the photonic activation of the catalyst by light irradiation
producing reactive electron-holes; and in catalytic wet
oxidation (CWO) the pollutant molecules are oxidized
with pure oxygen or air at elevated temperatures
(130–250 °C) and pressures (5–50 bar). Other AOPs
such as sonolysis, which consists on the ultrasonically
induced acoustic cavitation, or the electrochemical
oxidation, which produces large amounts of hydroxyl
radicals directly in water, are at an earlier stage of
development. Each technology requires a different
catalytic material as well as the optimization of their
catalytic properties. Different metal oxides (e.g., Zn, Mn,
Fe, Co, Ni, Cr, and Bi oxides) and noble metals (e.g.,
Ru, Pt, Pd, Rh and Ir) supported typically on CeO2,
Al2O3, ZrO2, TiO2 or carbon materials have been used in
CWO studies 3,4 and few of them applied in ozonation 5.
TiO2-based materials are the preferred photocatalysts 6
and iron ions are the base of the Fenton reaction 7,8.
Actually, the sustainable technological implemen-
tation of AOPs as new solutions at the macro-scale
depends on the understanding of fundamental catalytic
processes taking place at the nano-scale. Development of
active, selective and economic catalysts, which permit to
Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121 113
UDK 54:504 ISSN 1580-2949
Review article/Pregledni ~lanek MTAEC9, 43(3)113(2009)
increase the process efficiency and decrease the con-
sumption of chemicals, is one of the main challenges 9.
The present paper does not intend to give an extensive
survey on the published work regarding catalytic
materials and reactor configurations that has been used
in AOPs, since this information can be found in many
reviews of the Fenton process 8,10, ozonation 5, photo-
catalysis 10–15 and CWO 3,4,10,16–23, but rather to give an
introductory background on the catalysis engineering
steps and tools to develop heterogeneous catalysts, with
the final aim of applying these materials in industrial-
scale reactors. In contrast with CWO, the ozonation and
the Fenton processes are relatively simple and have been
widely installed for the treatment of effluents 5,7. Photo-
catalysis differs from the other AOPs since a light-
irradiation is used and, as consequence, the main
concepts involved are also different when. Therefore, in
this paper, particular emphasis is given to CWO and
photocatalysis as two examples of AOPs with different
complexity.
Environmental catalysis engineering can be con-
sidered as an interdisciplinary integration of fundamental
and applied knowledge on materials, chemistry, physics
and engineering. The different levels of catalyst
development are strongly linked with basic concepts of
all these areas and they cannot be isolated if the aim is to
produce sustainable development of catalytic reactors. In
a catalytic point of view, at the nano-scale the reactivity
of the active sites is influenced by the surrounding
environment at a supramolecular level. This environment
drives the local adsorption/desorption of reactants and
products; for instance, local pH around the active site
can be different from the bulk pH depending on the
characteristics of the catalytic cavities. For this reason,
adequate tuning of catalysts at the nano-scale scale level
allows enhancement of the catalytic properties. The
nano-scale differs from the molecular level since the
later refers to the direct interaction between the active
site and these molecules. The nano-sized catalytic
materials should be assembled at the micro-scale with
the aim of obtaining optimal spatial distribution and
catalyst composition in a final macro-structured catalyst
(e.g., pellets, membranes, monoliths, foams) 24. Optimi-
zation of catalytic properties and reactor design involves
testing and modelling at different levels (reaction and
kinetic studies under different operating conditions,
mass/heat transfer and hydrodynamic investigations,
simulation and experimental work in pilot-units and final
scale-up to industrial level). In order to provide an
integrative point of view, the multidisciplinary strategy
for development of catalysts was divided in seven main
stages, as shown in Figure 1: catalyst synthesis; catalyst
screening and operating conditions; kinetic studies and
reaction network; catalyst moulding; pilot-plant unit and
process design; industrial-scale application. The length
associated to each step is only indicative and although
they are sequentially presented and described in the next
sections, in practice there are overlaps between the
different steps when optimizing catalytic materials/pro-
cesses.
2 CATALYST SYNTHESIS
In general, the chemical processes can be classified
according to the number and type of phases (liquid, gas
and solid) or to the nature of the catalyst (heterogeneous
or homogeneous). Gas-liquid reactions with solid
catalysts are one of the largest chemical commercial
applications, including oxidations, since the use of
homogeneous catalysts requires its removal from the
wastewater with a posterior recovery step. The concept
of heterogeneous catalysts is normally associated to
solids or mixtures of solids which accelerate chemical
fluid reactions without changes in the solids. During
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
114 Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121
Figure 1: Schematic diagram representing different levels of catalyst
development and application. Figures at the right side from bottom to
top: (i) HRTEM micrograph of nanosized ceria particles (≈ 3 to 8
nm); (ii) representation of a functionalized carbon surface; (iii)
HRTEM micrograph of Pt nanoparticles (dark spots ≈ 2–4 nm)
supported on Ce-Ti-O; (iv) scheme of a particle of catalyst in a liquid
media; (v) example of a shaped catalyst (6.4 × 6.4) mm; (vi) WO
Zimpro diagram.
Slika 1: Shemati~ni grafikon, ki predstavlja razli~ne stopnje razvoja
in uporabe katalizatorjev. Slike od desne spodaj navzgor: (i) HRTEM-
mikroposnetek nanoziranih delcev cerijevega oksida (≈ 3 – 8 nm); (ii)
funkcionalizirana povr{ina ogljika; (iii) HRTEM-posnetek nanodelcev
Pt (temne to~ke ≈ 2 – 4 nm) na Ce-Ti-O-podlagi; (IV) shema delca
katalizatorja v teko~em mediju; (V) primer oblikovanega katalizatorja
(6.4 × 6.4) nm; (VI) WO Zimpro-diagram
many years (essentially until 1970’s), development and
preparation of heterogeneous catalysts was considered
more as alchemy than science, consuming many efforts,
money and time without guaranty of the final results 25.
Nowadays, the relation between catalysis and other
emerging sciences, such as analytical or solid state
chemistry, makes possible to reduce the wide number of
variables in preparing heterogeneous catalysts to a series
of elementary steps. A heterogeneous catalyst is charac-
terized mainly by the relative amounts of different
components (active species, physical and/or chemical
promoters, and supports), shape and size, pore volume
and distribution, and surface area. The nature of the
active species is always the most important factor.
Various research projects worldwide investigate possible
applications of active and stable heterogeneous catalysts
in AOPs for the oxidation of various pollutants.
The first steps in heterogeneous catalytic ozonation
date back to 1970s 26, but only now interesting perspec-
tives are reappearing 5. In the Fenton process the for-
mation of sludge, a source of pollution and the reason of
catalyst loss, is the main shortcoming of using iron in the
typical homogeneous form. Iron-based heterogeneous
catalysts, or solid sources of iron, such as minerals, are
some alternatives 8. Attention has been also paid to new
heterogeneous catalysts for CWO and photocatalysis.
Catalysts able to oxidize typical refractory end products
of biological processes have been well accepted.
According to the most common preparation proce-
dures, the catalysts can be classified as 25: bulk catalysts
or supports where the catalyst is generated as a new
solid phase (e.g., precipitation, sol-gel) and impregnated
catalysts where the active phase is introduced or fixed on
a pre-existing solid (e.g., impregnation, ion exchange,
adsorption, deposition-precipitation and chemical- or
physical-vapour deposition). The formation and precipi-
tation of a crystalline solid occurs by supersaturation
with physical (e.g., variations in temperature) or
chemical (e.g., addition of bases or acids) perturbations
of a solution, formation of stable small particles by
nucleation and growth by agglomeration of those
particles. The sol-gel method consists on the transfor-
mation of a solution into a hydrated solid precursor
(hydrogel). This method has gained importance due to
the better control of different properties such as texture
and homogeneity. Impregnation is based in the contact
between the solid support and a certain volume of
solution, which contains the precursor of the active
phase. The method is called wet impregnation when an
excess of solution is used and incipient wetness
impregnation if the volume is equal or slightly less than
the pore volume of the support. The ion exchange
method is based on the replacement of ions on the
surface of a support by electrostatic interactions.
Adsorption consists in the controlled attraction of a
precursor contained in an aqueous solution by charged
sites on the support. In the deposition-precipitation
method, slurries are formed and the precipitation occurs
in interaction with a support surface by addition of an
alkali solution. The prepared solids are separated from
the mother liquor with filtration, decantation, centrifu-
gation, etc., usually washed with distilled water (or
specific organic solvents) for complete removal of
impurities and submitted to thermal treatments (e.g.,
drying and calcination-heating). After calcination,
modification of the nature and/or structure of the phases
and stabilization of mechanical properties occur.
Thorough chemical, textural and structural analysis
of catalytic materials is possible nowadays due to several
available techniques. Electron microscopy constitutes a
family of very powerful catalyst characterization techni-
ques. Scanning electron microscopy (SEM), trans-
mission electron microscopy (TEM) or high resolution
TEM (HRTEM) are analytical techniques used to iden-
tify structural constitution of heterogeneous catalysts,
including shape, atomic structure and dispersion of
active nanoparticles supported on high surface area
materials (e.g., Figure1-iii showing Pt-nanoparticles
supported on Ce-Ti-O), which are key information in
catalysis. When coupled with other techniques, such as
energy-dispersive X-ray spectroscopy (EDXS) or
selected-area electron-diffraction (SAED), catalyst
composition data can be obtained in diverse phases of
the sample, since it is possible to analyze selected areas
at the nano-size scale. New devices have been coupled
with microscopes in order to increase the analytical
potentialities. For instance, the high-angle annular
dark-field detector (HAADF) and the field emission gun
(FEG) are appropriate to identify some metals (e.g.,
gold) at the nano-scale in catalytic materials. X-ray
diffraction (XRD) is a common technique used to
identify the structure of materials as well as to determine
crystallite sizes with the Scherrer equation. X-ray
photoelectron spectroscopy (XPS) is frequently used to
analyze oxidation states. Textural characteristics (BET
surface area, micropore, mesopore and macropore
constitution) are determined from the nitrogen adsorp-
tion isotherm obtained at 77 K and using specific
calculation methods such as t-method, Stoeckli and
Dubinin equations. Temperature-programmed desorption
(TPD) has gained special relevance in the chemical
analysis of surface oxygen groups on carbon materials,
since CO and CO2 are formed and released when
applying heating ramps 27. By means of the
deconvolution of the TPD spectra, it is possible to
determine the amounts of the diverse surface groups
(carboxylic acids; carboxylic anhydrides; lactones;
phenols; carbonyl or quinones). This makes possible to
finely tune the surface chemistry of functionalized
carbon materials (representative image on Figure 1-ii)
which provides a proper control of chemical reactivity.
The referred techniques are only few examples among
many other important methods that are used to
characterize heterogeneous catalysts.
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121 115
3 CATALYST SCREENING AND OPERATING
CONDITIONS
In AOPs the appropriate catalyst will be the one that
operating in amenable reaction conditions shows the
highest level of pollutant degradation to carbon dioxide
and water or to less toxic/refractory compounds. Many
parameters should be evaluated and several materials
and compositions must be tested in order to increase the
probability of identifying a suitable candidate for the
process. Catalyst deactivation must be investigated as
well, being normally due to sintering, poisoning of active
sites, dissolution of active components to the liquid
phase (leaching) or fouling of the catalyst surface as a
result of deposition of reaction intermediates. Exami-
nation of fresh and used catalysts allows to identify
possible changes on the catalyst during the reaction,
essential for understanding the catalyst deactivation.
Therefore, an initial idea of catalyst activity/stability
may be obtained in batch reactor experiments by
recycling the catalyst for a small number of cycles.
Experiments that are more definitive may require
operation with a continuous reactor over hundreds or
possibly thousands of hours of operation. The level of
pollutants degradation to be achieved with a certain
technology is defined by process economic conside-
rations, where the cost of the catalyst replacement is
considered as part of the global economical analysis.
Novel methods have been developed for monitoring in
situ in which way the metal sites interact chemically
with the surrounding molecules, interconnecting
classical studies with advanced analysis of the chemical
events taking place at the atomic/molecular level. Sum
Frequency Generation (SFG) is one example 28.
The activity of catalysts depends on the operating
conditions used. Catalyst loading, solution pH and sub-
strate concentration/nature are variables that should be
considered among all the different catalytic-AOPs. Other
parameters are specific for each process. Of importance
is the oxygen partial pressure and the temperature in
CWO, the light intensity in photocatalysis and the ratio
Fe:H2O2 in the Fenton process. The ozonation process is
strongly dependent on the media pH. At low pH the
direct and selective reaction of molecular ozone with
pollutants occurs, while at high pH ozone is mainly
decomposed in hydroxyl radicals which in turn attack the
pollutants 5. In CWO, the increase of temperature leads
to a consequent increase in the oxidation rate. Regarding
possible effects of oxygen partial pressure and catalyst
loading, a hypothetical behaviour of the total organic
carbon (TOC) conversion in a generalized reaction
system at constant temperature is represented in Figure
2. Without catalyst and increasing the oxygen partial
pressure the conversion obtained will be small if the
pollutant is of difficult oxidation by non-catalytic WO. If
the experiment is performed without oxygen but in the
presence of a catalyst, some conversion can be achieved
depending on the nature of the catalytic material.
Increasing both oxygen and catalyst loading, a plateau of
maximum efficiency can be observed in a region where
changes in the correspondent values do not affect the
global conversion. Moreover, for higher oxygen partial
pressures decrease in conversion is represented as well,
in order to account for the possible occurrence of
over-oxidation by the catalyst, as referred in the lite-
rature 29. At a constant oxygen partial pressure, decrease
in conversion with higher amounts of catalyst represents
the eventual acceleration of the termination step
mechanism by the catalyst 30.
Regarding photocatalysis, it is quite well established
that the reaction rate is proportional to the radiant flux
and to the mass of catalyst in slurry conditions, up to a
certain loading for which all the particles are fully
illuminated (usually between 0.6–2.5 g/L). For higher
mass of catalyst, scattering effects can decrease the
efficiency of the process 12. In contrast to CWO, in
photocatalysis heating is not required since the process is
based in a photonic activation of the catalyst. In fact,
increase of temperature can have a negative effect in the
activity, since the adsorption of pollutants decreases and
tends to decelerate the reaction. Thus, photocatalytic
reactors operate at normal temperature and pressure.
This advantage together with the possibility of using
solar energy makes this technology very promissing.
Development of immobilized photoactive materials in
the visible-light region (> 400 nm) is the main challenge.
4 KINETIC STUDIES AND REACTION
NETWORK
Multiphase reactions often consist of multiple reac-
tion pathways with series of parallel reactions networks.
Identification of intermediates is usually an essential
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
116 Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121
Figure 2: Conversion of total organic carbon as a function of catalyst
loading and oxygen partial pressure (PO2), in CWO.
Slika 2: Popolna organska konverzija na CWO kot funkcija obreme-
nitve katalizatorja in parcialnega pritiska kisika (PO2)
issue, normally requiring both experimental analysis and
postulates from the available information. In AOPs the
strong oxidation media leads to the formation of several
intermediates. The difficulty in the development of a
reaction network increases in complex reactions when
many compounds are involved, which is of special
importance when using industrial effluents. For this
reason, besides Langmuir-Hinshelwood expressions,
kinetic models based in lumped parameters grouping
pollutants with similar reactivity have been widely used
for AOPs 16,31,32. Moreover, in general, the mechanisms
are not well understood but it is quite consensual that
free radicals participate actively. For this reason, the
normal strategy when studying AOPs is to oxidize as
much as possible the pollutants and look to the organic
pollutants obtained at the end of the experiments, which
are usually more resistant to oxidation than those
degraded along the process.
The kinetic parameters are generally obtained in
slurry operation batch-experiments for AOPs. Figure 3
shows the typical concentration profiles of a general
gas-liquid-solid system with external mass transfer and
intraparticle diffusion resistances for the oxidant gas (O)
and the pollutant (P). The local temperature is assumed
to be constant throughout the different phases and the
liquid backmixed. For the reactants the steps can be
described as: (A) transport of O from the gas bulk to the
gas-liquid interface; (B) transport of O from the
gas-liquid interface to the liquid bulk; (C) transport of O
and P from the liquid bulk to the catalyst surface; (D)
intraparticle diffusion of the reactants into the catalyst
pores; (E) adsorption of the reactants on the active
catalyst sites; (F) surface reactions of O and P to yield
reaction products; and (G) desorption of the products
into the liquid bulk. Therefore, before the reaction at the
active catalyst sites, different transport and reaction
processes take place. In order to model the intrinsic
reaction kinetics, the laboratory reactor should be
selected and operated with the aim of conducting the
experiments under conditions where interparticle and
intraparticle concentration and temperature gradients are
absent. With mechanically agitated reactors, it is
possible to eliminate or reduce the magnitude of these
resistances in a gas-liquid-solid catalyzed system and,
therefore, it becomes easier to obtain intrinsic reaction
rates directly or derived from the experimental data.
Continuous fixed-bed reactors have been also used for
this purpose in CWO 33; however, elimination of these
gradients may be difficult.
For slightly soluble gas oxidants in water (e.g.,
oxygen), the transport of the oxidant in a gas phase
constituted by two or more components (e.g., oxygen
and nitrogen) is faster than in the liquid. The gas phase
resistance can be then neglected, being the liquid film
resistance the controlling mass transfer. Different
experimental procedures can be considered in order to
eliminate the concentration gradients on the system; (1)
For the gas-liquid mass transfer resistance the preferred
approaches are to improve the gas-liquid contacting or to
increase the concentration of reactant in the gas phase.
The first case can be obtained by means of some reactor
specifications such as impeller rotational speed, design,
size, location or gas sparger design, vessel baffles and
other internals; (2) Concerning the liquid-solid mass
transfer resistance, the elimination of this phenomenon is
associated with the amount of catalyst, catalyst particle
size, impeller diameter and location, baffles and other
internals. The increase of the impeller rotation speed
decreases the liquid-solid resistance, having, however,
little effect once the speed attains the minimum value
required for complete and uniform solids suspension; (3)
The internal diffusion resistance can be reduced with the
amount of catalyst as well as with the particle diameter
since, consequently, the effectiveness factor increases
34–35.
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121 117
Figure 3: Schematic representation of typical concentration profiles for the gas (CO) and the liquid pollutant (CP) in a gas(G)-liquid(L)-solid(S)
catalytic system with and without(*) external mass transfer and intraparticle diffusion resistances (i = interface).
Slika 3: Shemati~na predstavitev tipi~nih koncentracijskih profilov za plinski (CO) in teko~i one~i{~evalec (CP) v teko~em (L) in v trdnem
katalitskem sistemu z zunanjim prenosom mase in brez njega(*) ter difuzijska odpornost med delci (1 – mejna povr{ina)
CWO experiments are conducted in stainless steel (or
titanium) high-pressure autoclaves equipped with
thermocouples and mechanically driven impellers for
proper control of temperature and agitation velocity,
respectively. In contrast, typical batch photocatalytic
reactors consist of a glass external cylindrical reaction
vessel, where is placed the reaction mixture, with
magnetic agitation, and an inner double-wall compart-
ment immersed inside the liquid housing the lamp and
acting as a cooling water jacket. This compartment is
made of glass or quartz depending on the type of
irradiation that is studied (UV or UV/VIS). Therefore,
batch reactors in slurry conditions are used to perform
kinetic studies in both CWO and photocatalysis.
However, the reactors are quite different in what respect
configuration and construction materials.
5 CATALYST MOULDING
The optimal catalyst composition defined by slurry
screening tests can differs when the experiments are
performed with shaped catalysts. In this context, the
different steps of catalyst development should be always
strongly connected. The solid catalysts are typically used
as a powder (slurry), or as irregularly shaped granule,
table, sphere, or extrudate. Usually, the catalysts used in
slurry have average particle diameters between 50 µm to
450 µm and the larger pellet catalysts used in fluidized
and fixed-bed processes between 0.8 mm to 10.0 mm (an
image of a shaped catalyst is shown in Figure1-v).
Structured catalysts, such as monoliths, membranes,
foams and fibre catalysts are new attractive alternatives
24. Shaped catalysts can be produced by traditional
processes such as press and extraction, or using new
technologies as powder injection moulding (PIM). Some
studies should be done to optimize the heating rate,
sintering temperature, time and atmosphere composition.
In particular for photocatalysis, the use of typical pellets
is not feasible since the semiconductor catalyst has to be
fully irradiated. For this reason, most of the known
systems use fine powdered catalyst dispersions/suspen-
sions, requiring a final step of filtration. As solution,
photocatalytic materials can be immobilized onto glass
or other supports as a film. Since these coatings are still
not mechanically robust for the traditional and complex
waste waters treatments, they are actual motive of
intensive research 13.
6 PILOT PLANT UNIT & PROCESS DESIGN
AOPs can be implemented in different three-phase
reactors. Most used configurations are mechanically
agitated reactors in slurry operation, bubble columns and
packed-bed reactors. The complexity in modelling
multiphase reactors with solid catalysts lies mainly in the
fluid flow patterns, which depend on the reactor cate-
gory. Prior to the intrinsic reaction that occurs at the
catalyst surface, the reactants have to diffuse from the
gas and liquid phase to the active catalyst sites. In this
context, the scale-up of a reaction system also involves
evaluation of heat and mass transfer as well as gas-liquid
hydrodynamics. Several models have been developed
using various parameters such as interface mass transfer
coefficients, intraparticle effective diffusion coefficients,
axial dispersion coefficient, flow regimes, liquid
hold-up, pressure drop and catalyst wetting 23–24. These
parameters depend on the hydrodynamic regime of the
reactor and the physical properties of the phases (gas,
liquid and solid). Interface mass transfer coefficients or
external diffusion involve the resistances that are
inherent to the reactor for transport from gas to liquid,
liquid to solid, gas to solid and from dynamic to static
liquid. The intraparticle diffusion or internal diffusion of
reactants depends on the particle shape. The flow regime
describes the flow pattern of liquid and gas throughout
the reactor. The liquid hold-up consists of the inter-
particle and intraparticle fraction, of the total reactor
volume, occupied by the liquid phase. The pressure drop
depends on the liquid hold-up as the latter determines the
effective void fraction available for the gas phase. The
catalyst wetting deals with the average fraction of the
external particle surface covered by the liquid. Many
studies have been conducted at laboratorial scale to
consolidate the knowledge in AOPs, but the bridge
between laboratory and commercial scale is still a
difficult task due to this high number of variables
involved in the design of multiphase catalytic reactors.
Other specific issues should be also considered. For
instance, in CWO experiments, it was found that heavy
polymers are formed when some pollutants, such as
phenol, are oxidized in slurry reactors. This phenomenon
can be attributed to the high liquid-to-solid ratio in the
slurry reactor, once these polymers can be formed
through oxidative coupling. Afterwards, the polymers
can be irreversibly adsorbed on the catalyst surface and
progressively block the active sites. In this context,
reactors with low liquid-to-catalyst volumetric ratio,
such as packed-beds operating in trickle-bed regime,
may be better options 36,37. Moreover, catalyst stability
should be addressed in continuous reaction application,
taking as base the stability results obtained in batch
operation. In fact, for a fresh catalyst, it is not unusual to
show an initially high level of activity during short time
testing. Nevertheless, sometimes, this is followed by a
gradual or rapid decline with extended exposure, which
may continue until some measure of activity reaches
zero, or achieves a steady-state level caused by the
deactivation of the catalyst.
Most of the studies performed in continuous ope-
ration deal with small-scale reactors. Relatively few
studies have been published in the field of CWO in
large-scale packed-bed systems, where mass transfer
limitations are involved 38. Therefore, experimental data,
description of mass- and heat-transfer processes, reactor
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
118 Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121
simulation (e.g., using Computational Fluid Dynamics –
CFD) and many other studies concerning continuous
operation on CWO are object of research 39,40 being
tested laboratorial pilot-scale reactors with lengths
between 15–350 cm and diameters between 0.1 cm and
10 cm 36,38,41,42,43.
Photocatalytic reactors differ in several aspects of
design when compared with typical reactors. Selection of
radiation sources (including solar light), respective
design according to the radiation implemented, materials
of construction and the need of other devices (such as
reflectors, windows and cleaning setup) should be
considered in their design 13. Typical photocatalytic
reactors consist of inner walled glasses housing the
lamps in contact with the waste waters contained in an
external reaction vessel. Other configurations consist of
external or distributed light sources. Most of the reactors
use powders as catalysts (slurry annular, open upflow,
integrated flow membrane with filtration system, swirl
flow, taylor vortex and turbulent slurry reactors) or, less
common, immobilized photocatalysts in solid matrix
(falling film, fiber optic cable, multiple tube, packed bed,
rotating disk and spiral glass tube reactors). Examples of
solar reactors are the parabolic trough reactor, the
parabolic collecting reactor, the double skin sheet
reactor, the tubular reactor and the solar pond 13.
The deposition of photoactive films on the internal
surfaces of the photoreactors is gaining more importance
due to the increased control on the preparation of such
materials. The liquid-phase processing methods that
have been used specially to synthesize thin films, usually
nanocrystalline, are (co-)precipitation and sol–gel
methods while among the gas phase methods sputtering,
spray pyrolysis deposition, chemical and physical vapour
deposition are used. More investigation is needed
regarding nano-adherence stability of crystalline films
when applied over a rigid inert surface and explicit
application of nano-films to photocatalytic reactors.
Mechanical characterization of the films can be carried
out by depth-sensing nanoindentation using Atomic
Force Microscopes (AFM).
7 INDUSTRIAL-SCALE APPLICATION
Examples of commercial WO processes are schema-
tized in Table 1 19–20. As it can be observed, the WO
technology has earned some place in the treatment of
different waste waters. In particular, Zimpro (from
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121 119
Table 1: Schematization of WO commercial processes
Tabela 1: Shematika komercialnih WO-procesov
Process N° Plants Waste examples Reactor T/°C P/bar Catalyst
Zimpro 200
20
50
sewage sludge
spent activated carbon
industrial
bubble column
(20 min – 4 h)
250–270
280–325
150–320
85–120
200
10–220
None (or
homogeneous
copper)
Kenox n.s. pulp & paper, metal
finishing,petrochemical, food
processing
recirculation reactor 204–260* 31–83 None (or
impregnated in
the static
mixer)
Wetox n.s. n.s. stirred tanks 207–247 40 None
VerTech n.s. sewage sludge below-ground
deep shaft
< 275* < 110* None
Oxyjet n.s. pharmaceutical and phenolic
chemical wastes
jet mixer and tubular
reactor (+ jet reactor)
< 300 n.s. None (or in a
jet reactor)
WPO n.s. n.s. n.s. (H2O2 as oxidant) 90–130 1–5 Fe
2+
Ciba-Geiby 3 pharmaceutical, pulp &
paper, petrochemical,
hazardous waste landfill
n.s. 300 n.s. Cu2+
Loprox n.s. Paper mill waste streams,
municipal sewage sludge
Single or multi-stage
bubble column
120–200 50–20 Fe2+
Orcan n.s. Textile n.s. 120 3 Fe2+
ITECP** n.s. n.s. n.s. n.s. n.s. Bromide,
nitrate and
manganese ions
Athos n.s. municipal and industrial
sludge
n.s. 240 44 Cu2+ (heteroge-
neous optional)
Osaka gas n.s. coal gasifier
coke ovens
cyanide
sewage sludge
slurry bubble
column
250 70 ZrO2 or TiO2
with noble or
base metals
NS–LC n.s. phenol, glucose, acetic acid,
etc
monolith 220 40 Pt–Pd/
TiO2–ZrO2
Kurita n.s. spent resins, ammonia n.s. < 170 n.s. Supported Pt
n.s.: not specified; *variable along the reactor; **IT Enviroscience catalytic process.
USFilter, U.S.) is well established at an industrial level
(WO diagram shown in Figure 1-vi). Treatment of
sludge has been one of the main applications of WO and
temperature ranges from 90 °C to 300 °C and pressures
up to 220-bar have been used. Different reactor confi-
gurations have been developed, such as bubble column,
tubular, stirred tanks, monolith and recirculation
reactors. The commercial processes are essentially based
in non-catalytic WO (five systems) or in homogeneous
catalysts (six processes). The use of non-catalytic
systems limits the application of the technology,
increases considerably the operating conditions, and,
therefore, the capital costs. Only few processes (three)
involving heterogeneous catalysts, which are mainly
based in expensive materials (Pt and Pd) are commer-
cially developed. Regarding the use of photocatalysis for
waste waters treatment, several companies are including
this technology in their commercial portfolios. Commer-
cial worldwide interest is clear, but many photocatalytic
systems are still under development. Some cases of
success are the solar plant in Almeria, Spain, and the
systems commercialized by Purifics Environmental
Technologies Inc., Canada 13,15. Effective immobilization
of catalysts that can operates under both artificial and
solar light seems to be the main challenges for sustain-
able development of industrial-scale photoreactors.
Therefore, the development of active and stable hetero-
geneous catalysts for industrial-scale application became
still a key factor for enhancing the efficiency of AOPs.
8 FINAL REMARKS
Several criteria should be considered when selecting
a given process for the treatment of waste waters. The
development of catalytic materials is based on a
methodical sequence of scientific and technological
approaches to achieve a successful actual application.
Identification of active phases by screening reactivity
tests combined with the analysis of catalyst stability are
the first steps in order to proceed with further kinetices
and reaction network studies, which are particular of
each reactional system, and consequently, of each type
of waste water. In this domain, powerful analytical tools
should be used especially for catalyst characterization
and deep understanding of the catalytic reaction events.
The integration of different AOPs have to be considered
as a common strategy in laboratorial studies since better
solutions can be obtained in an efficient-economic basis
regarding both capital and operating costs when
scaling-up these technologies. The catalyst powders
should be transformed in suitable and mechanical robust
forms for mass/heat transfer investigation and, even-
tually, hydrodynamic studies. In this task valuable com-
putational tools can be used to simulate the catalytic
processes, such as CFD, but kinetic data can be still only
obtained with experimental work. Final testing in real
conditions is achieved after the scale of these
technologies up to an industrial level.
Acknowledgments
The author acknowledges the Fundação para a
Ciência e Tecnologia (FCT), Portugal, for the financial
support by the project POCI/N010/2006.
9 REFERENCES
1 G. Centi, S. Perathoner, Catal. Today, 77 (2003), 287–297
2 H. J. H. Fenton, J. Chem. Soc., 65 (1894), 899–910
3 Y. I. Matatov-Meytal, M. Sheintuch, Ind. Eng. Chem. Res., 37
(1998), 309-326
4 F. Stüber, J. Font, A. Fortuny, C. Bengoa, A. Eftaxias, A. Fabregat,
Top. Catal., 33 (2005), 3–50
5 L. F. Liotta, M. Gruttadauria, G. Di Carlo, G. Perrini, V. Librando, J.
Hazard. Mater., in press
6 X. Chen, S. M. Mao, Chem. Rev., 107 (2007), 2891–2959
7 P. Bautista, A. F. Mohedano, J. A. Casas, J. A. Zazo, J. J. Rodriguez,
J. Chem. Technol. Biotechnol. 83 (2008), 1323–1338
8 J. Pignatello, E. Oliveros, A. MacKay; Crit. Rev. Env. Sci. Technol.,
36 (2006), 1–84
9 C. Comminellis, A. Kapalka, S. Malato, S.A. Parsons, I. Poulios, D.
Mantzavinos, J. Chem. Technol. Biot., 83 (2008), 769–776
10 S. Parsons, Advanced Oxidation Processes for Water and Waste-
water Treatment, IWA Publishing, Cornwall, UK, 2004
11 I. K. Konstantinou, T. A. Albanis, Appl. Catal. B-Environ. 49
(2004), 1–14
12 J.-M. Herrmann, Top. Catal. 34 (2005), 49–65
13 H. de Lasa, B. Serrano, M. Salaices (Editors). Photocatalytic Reac-
tion Engineering, Springer Science+Business Media, USA, 2005
14 M. Pera-Titus, V. García-Molina, M. A. Baños, J. Fiménez, S. Esplu-
gas, Appl. Catal. B-Environ., 47 (2004), 219–256
15 A. Mills, S-K. Lee, J. Photoch. Photobio. A., 152 (2002), 233–247
16 J. Levec, A. Pintar, Catal. Today, 124 (2007), 172–184
17 V. S. Mishra, V. V. Mahajani, J. B. Joshi, Ind. Eng. Chem. Res., 34
(1995), 2–48
18 S. Imamura, Ind. Eng. Chem. Res., 38 (1999), 1743–1753
19 F. Luck, Catal. Today, 53 (1999), 81–91
20 S. T. Kolaczkowski, P. Plucinski, F. J. Beltran, F. J. Rivas, D. B.
McLurgh, Chem. Eng. J., 73 (1999), 143–160
21 S. K. Bhargava, J. Tardio, J. Prasad, K. Föger, D. B. Akolekar, S. C.
Grocott, Ind. Eng. Chem. Res., 45 (2006), 1221–1258
22 A. Cybulski, Ind. Eng. Chem. Res., 46 (2007), 4007–4033
23 F. Larachi, Top. Catal., 33 (2005), 109–134
24 G. Centi, S. Perathoner, Cattech, 7 (2003), 78–89.
25 M. Campanati, G. Fornasari, A. Vaccari, Catal. Today, 77 (2003),
299–314
26 C. G. Hewes, R. R. Davison, AIChE Symp. Ser. 69 (1972), 71
27 J. L. Figueiredo, M. F. R. Pereira, M. M. A. Freitas, J. J. M. Órfão,
Carbon, 37 (1999), 1379–1389
28 G. Rupprechter, C. Weilach, Nanotoday, 2 (2007), 20–29
29 Z. P.G. Masende, B.F.M. Kuster, K. J. Ptasinski, F. J. J. G. Janssen,
J. H. Y. Katima, J. C. Schouten, Catal. Today, 79–80 (2003),
357–370
30 S. S. Lin, D. J. Chang, C. Wang, C. C. Chen, Water Res., 37 (2003),
793–800
31 A. M. T. Silva, R.M. Quinta-Ferreira, J. Levec, Ind. Eng. Chem.
Res., 42 (2003), 5099–5108
32 S. H. Lin, C. D. Kiang, Ozone-Sci. Eng., 27 (2005), 225–232
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
120 Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121
33 A. Eftaxias, Catalytic wet air oxidation of phenol in a trickle bed
reactor: kinetics and reactor modelling. PhD Thesis, Rovira I Virgili
University, Tarragona, Spain (2002)
34 P. L. Mills, P.A. Ramachandran, R.V. Chaudhari, Reviews in chemi-
cal engineering, Eds. N. R. Amundson, D. Luss, University of
Houston, Freud Publishing House LTD, London, England (1992)
35 H. S. Fogler, Elements of chemical reaction engineering, Ed. G. Os-
bourne, Prentice-Hall, London, UK (1986)
36 J. Guo, M. Al-Dahhan, Chem. Eng. Sci., 60 (2004), 735–746
37 A. Alejandre, F. Medina, X. Rodriguez, P. Salagre, Y. Cesteros, J.E.
Sueiras, Appl. Catal. B-Environ., 30 (2001), 195–207
38 A. Singh, K. K. Pant, K. D. P. Nigam, Chem. Eng. J., 103 (2004),
51–57
39 D. Nemec, J. Levec, Chem. Eng. Sci., 60 (2005), 6958–6970
40 R. J. G. Lopes, A. M. T. Silva, R. M. Quinta-Ferreira, Ind. Eng.
Chem. Res., 46 (2007), 8380–8387
41 A. Fortuny, C. Bengoa, J. Font, F. Castells, A. Fabregat, J. Hazard.
Mater., B64 (1999), 181–193
42 A. Pintar, G. Ber~i~, J. Levec, Chem. Eng. Sci., 52 (1997),
4143–4153
43 A. Quintanilla, J. A. Casas, J. A. Zazo, A. F. Mohedano, J. Rodrí-
guez, Appl. Catal. B-Environ., 62 (2006), 115–120
A. M. T. SILVA: ENVIRONMENTAL CATALYSIS FROM NANO- TO MACRO-SCALE
Materiali in tehnologije / Materials and technology 43 (2009) 3, 113–121 121

V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
PRESSURELESS SINTERING AND
CHARACTERIZATION OF B4C, TiC AND
TiB2-PARTICLE-REINFORCED TiAl3-MATRIX
COMPOSITES
SINTRANJE IN KARAKTERIZACIJA KOMPOZITOV NA OSNOVI
TiAl3, OJA^ANIH Z DELCI B4C, TiC IN TiB2
Varu`an Kevorkijan1, Sre~o Davor [kapin2
1Independent Researching plc, Betnavska cesta 6, 2000 Maribor, Slovenia
2Jo`ef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia
varuzan.kevorkijansiol.si
Prejem rokopisa – received: 2009-02-10; sprejem za objavo – accepted for publication: 2009-03-21
Dense (>95 % of theoretical density) TiAl3-based composites reinforced with the volume fractions of B4C or TiC 10–50 % and
10–30 % of TiB2 particles were successfully prepared by combining pressureless reaction synthesis and reaction sintering. The
first step was formation of TiAl3 from the elemental powders by reaction between solid Ti and liquid Al (at 800 °C/1h and 1250
°C/3h in a static atmosphere of Ar with the volume fraction of H2 4 %). The reaction product obtained with a nominal
composition of 88% TiAl3-10%Al-2 % Ti, was milled and subsequently blended with the ceramic powders. The green compacts
made from the blended powder mixture were reaction sintered at 1300 °C for 2 h in an Ar + 4 % H2-rich environment using a
vacuum furnace. The morphology of the TiAl3 powders and the microstructure of the as-sintered composites were studied by
scanning electron microscopy and X-ray diffraction analysis. The mechanical properties were measured by tensile tests
performed at room temperature. In the case of reactive systems (TiAl3-B4C and TiAl3-TiC), densification proceeds via
formation of various secondary bonding phases, whilst in non-reactive system (TiAl3-TiB2) the process seems to proceed by
infiltration of molten aluminium into pores. In this respect, densification of TiAl3-TiB2 samples with more than 30 % of ceramic
particles was incomplete. Regarding the room temperature tensile properties of the composites, superior improvement was
observed in TiAl3-B4C and TiAl3-TiC samples with the volume fraction of ceramic particles 50 %. However, in TiAl3-TiB2
samples the improvement was significantly lower due to the lack of chemical affinity (wetting) between ceramic reinforcement
and TiAl3 matrix.
Key words: aluminium titanates, reaction synthesis, intermetallic matrix composites (imcs), ceramic particulate reinforcement,
reaction sintering, metallographic characterization
Z reakcijskim sintranjem smo pripravili goste (>95 % teoreti~ne gostote) kompozitne materiale na osnovi spojine TiAl3, oja~ane
z anorganskimi delci s prostorninskimi dele`i B4C in TiC 10–50 % ter 10–30% TiB2. V prvi fazi smo sintetizirali spojino TiAl3
iz prahov elementov Ti in Al z reakcijo med trdnim Ti in teko~im Al pri 800 °C/1h in kasnej{im segrevanjem pri 1250 °C/3h v
za{~itni atmosferi Ar s 4 H2. Tako pripravljen vzorec, ki je vseboval 88 % TiAl3, 10 % Al in 2 % Ti smo zmleli ter mu prime{ali
kerami~ne delce v dolo~enem razmerju in dobro homogenizirali. Te izhodne me{anice smo enoosno stisnili in sintrali v pri 1300
°C 2 uri v za{~itni atmosferi Ar/H2. Morfologijo izhodnih delcev TiAl3 in mikrostrukturo pripravljenih vzorcev smo analizirali z
elektronsko vrsti~no mikroskopijo in energijsko disperzijsko analizo ter rentgensko pra{kovno difrakcijo. Mehanske lastnosti
smo dolo~ali z nateznim preizkusom pri sobni temperaturi. Med sintezo vzorcev na osnovi sistemov TiAl3-B4C in TiAl3-TiC je
zgo{~evanje potekalo s tvorbo vmesnih faz, v sistemu TiAl3 – TiB2 pa z infiltracijo teko~ega Al v pore in po mejah med zrni. V
vzorcih TiAl3 –TiB2 z ve~ kot 30 % TiB2 je bilo tak{no zgo{~evanje nepopolno, zato so vzorci porozni. Glede na mehanske
lastnosti pripravljenih kompozitnih materialov pri sobni temepraturi ka`ejo najve~je izbolj{anje le-teh materiali s
prostorninskim dele`em 50 % B4C in TiC. V sistemu TiAl3-TiB2 so bile izmerjene lastnosti nekoliko slab{e zaradi odsotnosti
kemijske reakcije med anorganskimi delci in TiAl3 matrico.
Klju~ne besede: aluminij titanati, reakcijska sinteza, kompoziti z intermetalno matrico, diskontinuirna oja~itev s kerami~nimi
delci, reakcijsko sintranje, metalografske preiskave
1 INTRODUCTION
Titanium aluminide alloys based on the equilibrium
solid phases of the Ti-Al system and their composites
reinforced with ceramic particles have been considered
and are being developed for high-temperature appli-
cations in the automotive, aerospace, and power-gene-
ration industries. Such intermetallics based-matrix
composites (IMCs) reinforced with ceramic particles
generally achieved specific strength and specific stiff-
ness, creep strength, toughness, and high-temperature
strength resistance remarkably better than their as
non-reinforced intermetallic counterparts 1,2. Members of
this class of advanced materials are of interest as poten-
tial high-temperature materials due to their high melting
point, low density, good oxidation resistance, and high
modulus of elasticity. Ultimately, bringing these
attractive intermetallic materials into use largely depends
upon the availability of practical processing routes.
Titanium aluminide IMCs reinforced with various
ceramic particles (B4C, TiC, TiB2) can be effectively
produced by powder processing, starting from commer-
cial titanium aluminide powders blended with ceramic
particulates 3. Although this method offers near-net
shapes with very little post-fabrication machining
required, it is at the same time very expensive for
Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128 123
UDK 669.71:669.018.9:621.762.5 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)123(2009)
commercial application due to the high cost of titanium
aluminide powders. Essentially, it is of great significance
to develop an effective technique for preparation of
titanium aluminide powders starting from cost-effective
raw materials (e.g. by the exothermic reaction of liquid
aluminium with solid titanium) and their conventional
densification to fully dense alloys and composites 4.
Since such reactive synthesis and sintering aims at the
formation of one or more aluminides, a prior knowledge
of the formation of specific aluminide(s) and their
microstructural development is particularly important for
further commercialization of the process 5.
Among all the equilibrium solid phases of the Ti-Al
system, only TiAl and Ti3Al have been extensively used
for alloys production and as a matrix in composites. This
is because the Ti rich compounds Ti3Al and TiAl exist
over a wide range of composition while TiAl3 exists as a
very narrow range of compounds (an intermediate phase
that appear as a vertical line in the binary phase diagram
at some stoichiometric ratio of mole fractions of the
elements) and any change in the definite composition of
the mole fraction of elements (e.g. by oxidation) leads to
its decomposition 6.
In the early stages of the development of titanium
aluminide alloys and composites, the single step elemen-
tal powder metallurgical route has gained particular
attention because near-net shape titanium aluminide
alloy products could be fabricated by the consolidation
and forming of blended Ti and Al elemental powders,
followed by a subsequent reactive sintering process 7–11.
However, due to the large difference between the partial
diffusion coefficients of Ti and Al, the synthesis of
titanium aluminide alloys via reactive sintering follows
the mechanism that Al atoms move faster into the Ti
lattice then vice versa, thus leading during reactive sin-
tering to the formation of Kirkendal diffusion pores 12–14.
Hot isostatic pressing (HIP) and refining of the Ti and Al
reactants through cold deformation, such as extrusion
and rolling of the Ti and Al powder mixture billet prior
to reactive sintering, have been reported to be effective
in eliminating the porosity of reactively sintered TiAl
and Ti3Al alloys. However, their high cost and low
productivity make these techniques unsuitable for
commercial use.
In the current attempt to optimize the cost and quality
of IMCs, a two stage composite fabrication process was
considered. The first step in composite production was
the cost-effective synthesis of titanium aluminide pow-
ders from liquid aluminium and solid titanium, while the
second step was the reactive sintering of the cold
compacted as-received titanium aluminide powders
blended with ceramic particles (B4C, TiB2 and TiC) to
composite samples. The main expectation was that poro-
sity due to Kirkendal diffusion could be reduced below 5
% by refining the microstructure with ceramic particu-
late reinforcement and by selecting from the titanium
aluminides the most reactive constituents, such as TiAl3.
TiAl3 was found to be attractive for at least two
reasons: (i) it is the main product of the reaction between
liquid aluminium and solid titanium, which is one of the
most economical ways of production of titanium
aluminides, and (ii) it is also a reactive, non-stable phase
(Daltonides) which could react with ceramic reinforce-
ment, enabling formation of fully dense composite
structures.
Hence, in the present paper we report the fabrication
and characterization of titanium aluminide-based IMCs
made by a pressureless two-step procedure: (i) the reac-
tive synthesis of titanium aluminide powders with
nominal composition 25Ti : 75Al by reactive synthesis
from liquid aluminium and titanium powder in combi-
nation with milling and, (ii) reactive sintering of cold
compacted 25Ti : 75Al powders blended with ceramic
particles (B4C, TiB2 and TiC) to fully dense composite
samples.
2 EXPERIMENTAL
TiAl3 powder was synthesized by high temperature
reaction synthesis from the elemental powders. As a first
step, a homogeneous powder mixture of titanium (purity
99.5 % and average particle size 30 µm) and aluminium
(purity 99.7 % and particle size 100 µm) with a compo-
sition corresponding to stoichiometric TiAl3, 25 % mole
fractions of Ti and 75 % Al, was prepared in a planetary
mill. After room temperature uniaxial and cold isostatic
compaction of the mixtures, the green pellets obtained
were heated in a vacuum furnace (800 °C/1h and 1250
°C/3h), in a static atmosphere of Ar + 4 % H2. This was
followed by cooling the pellets to room temperature and
subsequent milling of the reaction products in an attrition
mill. The milling time was 4 h, and the grinding
media-to-powder ratio was 2:1.
Finally, composites were formulated by blending the
synthesized TiAl3 powder with commercial ceramic
powders (B4C, TiC and TiB2) in appropriate amounts to
create IMC-based matrices with the volume fractions
(10, 20, 30, 40 and 50) % of B4C, TiC or TiB2 disconti-
nuous reinforcement.
The powder blends were thoroughly mixed in a
planetary mill and subsequently cold compacted. In all
cases, reaction sintering of the compacts was conducted
at 1300 °C for 2 h in an Ar + 4 % H2-rich environment
using a vacuum furnace.
The as-synthesized composite samples were cut,
machined and polished in accordance with standard pro-
cedures.
Microstructural characterization of titanium alumi-
nide powders, as well as sintered composites, was
performed by optical and scanning electron microscopy
(OM and SEM with EDS), whereas X-ray diffraction
(XRD) measurements were applied to the samples to
identify the phases and their crystal structures.
The specimens for OM observation were electro-
lytically polished in a solution of 95 % CH3COOH and
V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
124 Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128
5 % HClO4, and then etched in a solution of 5 % HNO3,
15 % HF, and 80% H2O. The main grain sizes were
measured by the linear intercept method.
The specimens for XRD were abraded with SiC
paper and were then subjected to diffraction using CuKa
radiation.
Quantitative determination of the volume percentage
of titanium aluminide and secondary phases involved in
the matrix, ceramic particles and retained porosity was
performed by analysing optical and scanning electron
micrographs of polished composite bars using the point
counting method and image analysis and processing
software.
Composite density measurements were carried out in
accordance with Archimedes’ principle, applying di-
stilled water as the immersion fluid.
The initial density of the green compacts was
calculated from the mass and geometry of the samples.
The tensile properties (tensile strength, 0.2 % tensile
yield strength and elongation) of the composite speci-
mens were determined in accordance with the ASTM
test method, E8M-96. The tensile tests were conducted
on cylindrical tension-test specimens 3.5 mm in diameter
and 16 mm gauge length using an automated servo-
hydraulic tensile testing machine with a crosshead speed
of 0.254 mm/60 s.
3 RESULTS AND DISCUSSION
3.1 Morphology and chemical composition of labora-
tory prepared titanium aluminide powders
The morphology of laboratory prepared TiAl3
powders, obtained by milling the pelletized products of
reactive synthesis, is presented in Figure 1. As evident,
the TiAl3 powder obtained is non-agglomerated, with
well shaped individual particles of particle size mostly
below 10 µm.
The morphology of the raw titanium powder, Figure
2, is very similar with the morphology of synthesized
TiAl3. The main difference is in the average particle size,
which in the case of the as-received titanium powder is
about three times larger.
The typical phase composition in synthesized TiAl3
is reported in Table 1.
Table 1: Phase composition in laboratory prepared TiAl3 powder
Tabela 1: Fazna sestava laboratorijsko pripravljenega TiAl3 prahu
Phase Volume fraction (f/%)
TiAl3 87 ± 5
Al 10 ± 1
Ti 2 ± 0.2
Other phases 1 ± 0.1
3.2 Microstructural development in IMCs
Generally, the microstructure of IMCs consists of an
intermetallic matrix (based on an ordered intermetallic
compound or a multiphase combination of intermetallic
compounds), the ceramic particulate reinforcement and
the interfacial region with secondary phases formed
during reactive sintering by chemical reactions pro-
ceeding between the ceramic reinforcement, TiAl3 and/or
non-reacted Al and Ti. As minor phases, solidified Al
inclusions and individual titanium particles are also
involved in the composite microstructure.
3.2.1 B4C-Ti3Al samples
Reactive sintering of B4C-TiAl and B4C-Ti3Al
composite samples resulted in specimens with densities
higher than 95 % of T. D. As in densification, the boron
carbide particulate reinforcement reacted with the
intermetallic matrix forming numerous secondary phases
(TiB2, TiC, Al3BC), well detectable in the microstructure
of sintered samples – Figure 3.
In the B4C-Ti3Al and B4C-TiAl samples intensive
chemical reactivity between the composite constituents
was observed during high temperature densification due
to the fact that boron carbide reacts with both molten
aluminium and solid titanium 15,16.
V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128 125
Figure 2: SEM micrograph of as-received commercial Ti powder
Slika 2: SEM-posnetek komercialnega Ti-prahu
Figure 1: SEM micrograph of TiAl3 after milling
Slika 1: SEM-posnetek TiAl3-delcev po mletju
Boron carbide reacts at high temperature (1300 °C)
with aluminium, forming mainly Al3BC/Al4C3 and some
AlB2 (Reaction 1):
5B4C(s) + 19Al(l) = 2Al3BC(s) + Al4C3(s) + 9AlB2 (1)
However, boron carbide also reacts with solid tita-
nium particles producing secondary TiC and TiB2:
B4C(s) + 3Ti(s) = 2TiB2(s) + TiC(s) (2)
Moreover, from thermodynamic point of view, boron
carbide could also react with TiAl3 causing its decom-
position. The possible overall stoichiometric reaction
between boron carbide and particulate TiAl3 is the
following:
64B4C(s) + 57TiAl3 = 38TiB2(s) + 19TiC(s) +
+ 18Al3BC(s) + 9Al4C3(s) +81AlB2(s) (3)
3.2.2 TiB2-TiAl3 samples
TiB2 is stable at 1300 °C in contact with both TiAl3
and aluminium. Because of that, dense composite sam-
ples, Figure 4, are obtained only if a sufficient amount
of elemental aluminium is present in the system as a
liquid phase. The experimental findings suggest that the
necessary amount of aluminium is about one third of the
volume fraction of ceramic reinforcement. Sintering of
TiAl3-TiB2 composites with a larger amount of ceramic
reinforcement (>30 %) and without proper addition of
elemental aluminium resulted in samples with a huge
fraction of Kirkendall diffusion pores, as it is shown in
Figure 5.
3.2.3 TiC-Ti3Al samples
In this system, the best results of pressureless
reactive sintering were experimentally obtained. Almost
all the sintered samples, including these with a high
amount (50 %) of ceramic reinforcement, had a retained
porosity below 5 %, among which many were fully
dense.
As in the case of B4C, TiC may also react with TiAl3
and elemental aluminium. The products of these reac-
tions are various Ti-Al-C phases (e.g. Ti3AlC, Ti3AlC2
and Ti2AlC) and Al4C3. The presence of these phases
was confirmed by SEM-EDS at the matrix-reinforcement
interface of the fabricated composites, Figures 6, 7.
3.3 Mechanical properties
The results of room temperature tensile tests on the
composite samples are listed in Table 2. As a result of
matrix reinforcement, significant improvements in
Young’s modulus, tensile strength and ultimate tensile
V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
126 Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128
Figure 3: SEM micrograph of sintered TiAl3-40 % B4C composite
sample with various phases identified by XRD: (1) Al, Al4C3, Al3BC
(2) AlTi, AlB2, (3) TiB2 and (4) B4C
Slika 3: SEM posnetek sintranega TiAl3-40 % B4C-vzorca kompozita
z razli~nimi fazami, identificiranimi z XRD: (1) Al, Al4C3, Al3BC (2)
AlTi, AlB2, (3) TiB2 in (4) B4C
Figure 5: Scanning electron micrograph of pressureless sintered
TiAl3-50 % TiB2 composite with a significant number of Kirkendall
diffusion pores: (1) Al3Ti and (2) TiB2
Slika 5: SEM-posnetek TiAl3-50 % TiB2 kompozita sintranega pri
atmosferskem tlaku s {tevilnimi Kirkendallovimi porami: (1) Al3Ti in
(2) TiB2
Figure 4: Microstructure of a pressureless sintered TiAl3-30 %TiB2
-10 % Al composite. The microstructural constituents are: (1) Al3Ti,
(2) TiB2 (3) Al and (4) traces of Al2O3
Slika 4: Mikrostruktura TiAl3-30 % TiB2 -10 % Al-kompozita, sintra-
nega pri atmosferskem tlaku. Mikrostrukturni gradniki so: (1) Al3Ti,
(2) TiB2 (3) Al in (4) sledovi Al2O3
strength of the fabricated composites were observed. At
the same time, the highest improvement of tensile
properties caused by reinforcement was observed in
samples reinforced with TiC and B4C and less in samples
reinforced with TiB2. Comparing the mechanical
properties of composite samples with various volume
fractions of ceramic particles in the matrix, it was found
that Young’s modulus, tensile strength and ultimate
tensile strength increase while elongation decreases with
an increasing fraction of ceramic reinforcement.
4 CONCLUSION
A TiAl3 sintering mixture (87 % TiAl3, 10 % Al and
1 % Ti) for composite fabrication was routinely reaction
synthesized from the elements – (solid titanium particles
and molten aluminium) with an almost 100 % reaction
yield. Reaction synthesis was performed at moderate
temperature (1300 °C) for 2 h under a flowing atmo-
sphere of Ar + 4 % H2. The reaction product was milled
in a conventional planetary mill, resulting in particulates
with an average particle size of 10 µm.
Dense (>95 % of theoretical density) TiAl3-based
composites reinforced with 10–50 % of B4C or TiC and
10–30 % of TiB2 particles were successfully prepared by
combining pressureless reaction synthesis and reaction
sintering.
Irrespective of the ceramic reinforcement applied,
dense composite samples reinforced with 10 % to 30 %
of ceramic particulates were successfully obtained,
revealing the significant industrial potential of this
fabrication method. The retained porosity in sintered
samples was less than 5 %.
The highest chemical affinity between ceramic
reinforcement and matrix was observed in the TiAl3-TiC
and TiAl3-B4C systems, while TiAl3-TiB2 was chemi-
cally inert. The sintering of TiAl3-TiC and TiAl3-B4C
samples proceeds via formation of various secondary
phases involved in the process of densification. Accord-
ingly, TiAl3-TiC and TiAl3-B4C samples with as much as
50 % of ceramic reinforcement were successfully pre-
ssurelessly sintered. On the other hand, sintering in
TiAl3-TiB2 was achieved by the liquid phase (molten
aluminium). Hence, in TiAl3-TiB2 samples with more
than 30 % of ceramic particulates, under the applied
sintering conditions Kirkendall diffusion pores were not
reduced below 10 %.
Regarding the room temperature tensile properties of
the composite samples, the highest improvement was
obtained in TiAl3-TiC and TiAl3-B4C samples with the
highest amount (50 %) of ceramic reinforcement, while
V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128 127
Figure 7: Microstructure developed in TiAl3-40 % TiC: (1) TiAl3, (2)
Ti-Al-C and (3) Al with up to 5 % of Ti
Slika 7: Mikrostruktura TiAl3-40 % TiC-kompozita: (1) TiAl3, (2)
Ti-Al-C in (3) Al z do 5 % Ti
Figure 6: SEM micrograph of a TiAl3-30 % TiC composite: (1) TiAl3
matrix, (2) Ti-Al-C phase and (3) Al
Slika 6: SEM-posnetek TiAl3-30 % TiC kompozita: (1) TiAl3 matrica,
(2) Ti-Al-C faza in (3) Al
Table 2: Average room temperature tensile properties of various
laboratory prepared TiAl3-B4C, TiC, TiB2 composite samples
Tabela 2: Povpre~ne vrednosti mehanskih lastnosti (dobljene pri
sobni temperaturi) kompozitov na osnovi TiAl3, oja~anih z B4C, TiC,
TiB2
Volume
fraction of
ceramic
reinforce-
ment in
TiAl3 matrix
ϕ/%
Retained
porosity in
sintered
samples
%
E/GPa
Tensile
strength
σ/MPa
0.2 % tensile
yield
strength
σ/MPa
Elongation
in 50 mm
∆l/%
10-B4C 1.4±0.1 187±20 332±33 274±27 0.8±0.08
20-B4C 2.2±0.2 207±22 352±35 298±30 0.6 ±0.06
30-B4C 3.1±0.3 228±23 377±38 324±32 0.5±0.05
40-B4C 3.9±0.4 241±24 394±40 346±35 0.3±0.03
50-B4C 4.6±0.5 249±25 423±42 368±37 0.2±0.02
10-TiC 1.1±0.1 218±22 374±37 329±33 0.5±0.05
20-TiC 1.7±0.2 241±24 409±41 366±37 0.4±0.04
30-TiC 1.9±0.2 270±27 453±45 400±40 0.3±0.03
40-TiC 2.1±0.2 297±30 489±49 431±43 0.3±0.03
50-TiC 2.4±0.2 321±31 513±51 467±47 0.2±0.02
10-TiB2 3,9±0.4 149±15 298±30 209±21 0.7±0.07
20-TiB2 4,6±0.5 167±17 318±32 229±23 0.5±0.05
30-TiB2 5,0±0.5 188±19 331±33 248±25 0.4±0.04
in TiAl3-TiB2 samples the improvement was rather mo-
derate. Generally, the improvement of tensile strength,
tensile yield strength and modulus was found to correlate
with amount of ceramic reinforcement in the matrix,
irrespective of the sort of reinforcement. However, quite
the opposite behaviour was found regarding elongation,
where the introduction of ceramic particles into the
intermetallic matrix led in all specimens to a significant
reduction of ductility and toughness.
Acknowledgment
This work was supported by funding from the Public
Agency for Research and Development of the Republic
of Slovenia, as well as the Impol Aluminium Company
from Slovenska Bistrica, Slovenia, under contract No.
1000-07-219308
5 REFERENCES
1 Gamma Titanium Aluminide 1999, Y. W. Kim, D. M. Dimiduk, M.
H. Loretto, eds., TMS, Warrendale, PA, 1999, 3
2 S. Djanarthany, J. C. Viala, J. Bouix: Mater. Chem. Phys., 72 (2001),
301
3 C. M. Ward-Close, R. Minor, P. J. Doorbar: Intermetallics, 4 (1996),
217
4 E. K: Y. Fu, R. D. Rawlings, H. B. McShane, J. Mater.Sci., 36
(2001), 5542
5 M. Sujata, S. Bhabgava, S. Sangal, J. Mater. Sci. Lett., 16 (1997),
1175
6 G. Sauthoff, Basic of thermodynamics and phase transitions in
complex intermetallics, edited by E. B. Ferre, World Scientific,
London, 2008, 147
7 G. X. Wang, M. Dahms, JOM, 5 (1993), 52
8 M. Dahms, F. Schmeizer, J. Seeger, B. Wildhagen, Mater. Sci.
Technol., 8 (1992), 359
9 S. Kazuhisa, F. Tsutomu, M. S. Kim, K. Masaki, Powder Metall., 41
(1994), 146
10 J. B. Yang, K. W. Teoh, W. S. Hwang, Mater. Sci. Technol., 13
(1997), 695
11 F. Wenbin, H. Lianxi, H. Wenxiong, W. Erde, L. Xiaoqing, Mater.
Sci. Eng., A 403 (2005), 186
12 G. X. Wang, M. Dahms, Powder Metall., 24 (1992), 219
13 F. J. J. Van Loo, G. D. Rieck, Acta Metall., 21 (1973), 61
14 M. Dahms, Mater. Sci. Eng. A, 110 (1989), 5
15 G. Arslan, F. Kara, S. Turan, J. Eur. Ceram. Soc., 23 (2003), 1243
16 J. Jung, S. Kang, J. Am. Ceram. Soc., 87 (2004), 47
V. KEVORKIJAN ET AL.: PRESSURELESS SINTERING AND CHARACTERIZATION OF ...
128 Materiali in tehnologije / Materials and technology 43 (2009) 3, 123–128
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
MAGNETIC CHARACTERISTICS OF ISOTHERMALLY
AGED Cr-Ni-Mo-BASED ALLOYS WITH DIFFERENT
δ-FERRITE CONTENTS
MAGNETNE LASTNOSTI IZOTERMNO @ARJENIH ZLITIN
Cr-Ni-Mo Z RAZLI^NO VSEBNOSTJO δ-FERITA
Borivoj [u{tar{i~1, Benjamin Podmilj{ak2, Paul McGuiness2, Jelena V. Tuma1
1 Institute of Metals and Technology, Lepi pot 11, 10000 Ljubljana, Slovenia
2 Jo`ef Stefan Institute, Jamova cesta 39, Ljubljana, Slovenia
borivoj.sustarsicimt.si
Prejem rokopisa – received: 2009-02-27; sprejem za objavo – accepted for publication: 2009-03-04
We have investigated the influence of δ-ferrite content in Cr-Ni-Mo stainless-steel cast alloys on the magnetic characteristics.
Samples of cast alloys with mass fractions 9-11 % of Ni, 18-21 % of Cr and 1.8-2.5 % of Mo, with three characteristically
different δ-ferrite contents, were prepared with careful selection of appropriate amounts of alloying elements and a controlled
solidification procedure. The samples were then aged in the operating temperature region (290 – 350 °C) for different periods of
time (up to two years). The δ-ferrite content was determined with three different methods: i.e., empirically on the basis of
chemical composition, magnetic-induction-based and metallographically. The magnetic-induction-based determination of
 -ferrite content showed that it does not change with ageing temperature and time. It confirms that only the internal ferrite
structure is changed during ageing because of the spinodal decomposition.
The absolute magnetic properties were determined with a hysteresisgraph and a vibrating-sampling magnetometer (VSM). The
results showed that the magnetic properties depend on the chemical composition (δ-ferrite content), the ageing temperature and
the time. The chemical composition has the biggest influence, but the influence of the ageing temperature and time is
insignificant and the scatter of the results is relatively large. The determination of the absolute magnetic properties is a
destructive method and the mechanical preparation of the samples can influence the magnetic properties. Therefore, the method
is not appropriate for the in-situ observation of the kinetics of spinodal decomposition. Measurements with the VSM are more
appropriate than with the hysteresisgraph because much smaller samples with any geometry can be used.
Keywords: Cr-Ni-Mo based stainless steel, isothermal ageing, magnetic properties, influence of δ-ferrite
Raziskovali smo vpliv vsebnosti delta ferita na magnetne lastnosti Cr-Ni-Mo nerjavnih jeklenih litin. Izdelali smo tri razli~ne
litine v obmo~ju sestav z masnimi dele`i med 9 in 11 % Ni, 18 in 21 % Cr in od 1,8 do 2,5 %Mo. Skrbno smo izbrali primerno
vsebnost posameznih legirnih elementov in kontrolirali pogoje strjevanja, da smo dobili litine z razli~no, karakteristi~no
vsebnostjo δ-ferita (s pribli`no 4, 15 in 30 prostorninskimi dele`i). Vzorce litin smo razli~no dolgo (do dveh let) izotermno
`arili v temperaturnem obmo~ju (med 290 in 350 °C), ki je karakteristi~no za obratovanje uparjalnikov nuklearnih elektrarn.
Vsebnost δ-ferita razli~no staranih vzorcev smo dolo~evali na tri razli~ne na~ine: empiri~no na osnovi kemijske analize, na
osnovi merjenja magnetne indukcije in metalografsko. Meritve vsebnosti δ-ferita na metalografskih vzorcih so pokazale, da se
le-ta ne spreminja s ~asom in temperaturo izotermnega `arjenja. To potrjuje predpostavko, da se med izotermnim `arjenjem ne
spreminja vsebnost δ-ferita temve~ le njegova notranja struktura zaradi spinodalnega razpada.
V vzorcih preiskovanih litin smo dolo~ili tudi absolutne magnetne lastnosti z merilnikom magnetne histereze in vibracijskim
magnetometrom (VSM). Rezultati meritev so pokazali, da so magnetne lastnosti odvisne tako od sestave (vsebnosti δ-ferita) kot
tudi temperature in ~asa izotermnega `arjenja. Najve~ji vpliv ima vsebnost δ-ferita. Vpliv temperature in ~asa izotermnega
`arjenja je manj{i z opaznim raztrosom merjenih vrednosti. Dolo~evanje magnetnih lastnosti je poru{itvena metoda, saj moramo
vzorec materiala izrezati iz elementa ali konstrukcije, ki jo preiskujemo. Mehanska izdelava preizku{anca lahko vpliva na
magnetne lastnosti. Zato ta metoda ni najprimernej{a za opazovanje sprememb in-situ med izotermnim `arjenjem oziroma
spinodalnim razpadom. Magnetne meritve z VSM so primernej{e, ker zahtevajo precej manj{e vzorce poljubne oblike.
Klju~ne besede: nerjavna jekla na osnovi Cr-Ni-Mo, izotermno `arjenje, magnetne lastnosti, vpliv vsebnosti δ-ferita
1 INTRODUCTION
High-alloyed Cr-Ni-based stainless-steel cast alloys
are frequently used in thermoelectric installations such
as conventional and nuclear power plants (NPPs)1,2.
Many years of exploitation of mechanical equipment in
these objects have shown that the toughness of these
alloys decreases with the operating time and tempera-
ture3-7. These alloys have a characteristic two-phase
microstructure consisting of austenite and δ-ferrite8–10.
The δ-ferrite content depends on the chemical compo-
sition of the alloy and on metallurgical factors: i.e.,
manufacturing technology and the exploitation condi-
tions. Therefore, in alloys with a chemical composition
the allowed ranges of content of alloying elements
completely different microstructures can form. The
result of this could be different mechanical properties of
a material and it different behavior during its exploita-
tion. Actually, our investigations have already confirmed
that the Charpy impact energy and the resistance to
stable crack growth of such alloys depend strongly on
the δ-ferrite content11,12.
Extensive investigations in the past indicate for the
change of properties of these alloys the spinodal decom-
position of  -ferrite is responsible. Investigations also
showed that the austenite phase does not play a
Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135 129
UDK 669.14.018.8:621.785.3 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)129(2009)
significant part in this process. Spinodal decomposition
in these types of alloys one can understand as it is
schematically shown in Figure 1, and it is clear that the
final result of spinodal decomposition is coherent,
two-phase structure with different Cr-to-Ni ratios, even
in a relatively low-temperature range (300 – 350 °C).
This temperature range is similar to the operating-
temperature range of vital parts of, for example, the
coolant system6. The formation of two phases with the
same crystal structure but different lattice parameters
causes high internal elastic stresses, resulting in
significant hardness increase and toughness decrease.
Besides the microstructural investigations, mecha-
nical testing and analyses of isothermally aged
materials11,12 the aim of our work was also to find out if
microstructural and related mechanical changes can be
monitored and connected to a change of magnetic
properties. One of the most useful non-destructive
methods for the determination of δ-ferrite content is
based on magnetic induction, because in two-phase
systems (for example, ferrite-austenite) ferrite is the
ferromagnetic phase. The amount of this phase is almost
linearly proportional to the measured magnetic induction
and therefore the method became a standard in-situ
checking procedure for thermoelectric installations9,14–15,
especially when new reparative welds or parts are
introduced during service. However, publications in the
context of changes of the absolute magnetic properties
(the hysteresis loops) of these types of alloys during
isothermal annealing could not be found. Therefore,
some investigations and analyses were performed
relating to the change of the magnetic properties of
thermally degraded material.
2 EXPERIMENTAL
The samples of a Cr-Ni-Mo-based cast alloy,
CF-8M-type (ASTM A351), were prepared with a
standard casting procedure. Twenty-kg batches of melts
of the selected chemical composition were prepared in
an inductive melting furnace and then cast into metal
moulds under a controlled casting and solidification
conditions. The solidification rate was controlled with
the selection of an appropriate type of mould and
melt/mould preheating. Cooled, cast ingots of each
chemical composition (see Table 1), designated as alloys
A, B and C, were cut and ground into parallel slices,
approx. 10 mm thick. The actual chemical compositions
of the prepared alloys were determined with optical and
ion-coupled-plasma atomic emission spectroscopy (OES
and ICP-AES).
On the basis of well-known empirical correlations
(equation (1) and (2)) the Cr and Ni equivalents, as well
as the δ-ferrite content were calculated16. Such a
calculation shows that the alloy CF-8M inside the
allowed chemical composition can practically contain
from 0 to 90 volume fractions of δ-ferrite.
CR
w Cr
w Ni
eq
eq
=
( )
( )
=
=
w Cr W Si w Mo w Nb
w Ni w C
( ) . ( ) . ( ) ( ) .
( ) ( ) .
+ + + −
+ +
15 14 4 99
30 05w Mn w n( ) ( ( ) . ) .+ − +26 0 02 2 77
(1)
F CR CR CR= − + − +68 768 157 909 133171 4718492 3. . . . (2)
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
130 Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135
Figure 1: Schematic presentation of: a) quasibinary Fe-Cr-Ni-Mo
phase diagram and b) redistribution of Fe and Cr atoms during
spinodal decomposition13
Slika 1: Shemati~na predstavitev: a) kvazibinarni diagram
Fe-Cr-Ni-Mo in b) prerazporeditev atomov Fe in Cr med spinodalnim
razpadom13
Table 1: Nominal and actual chemical composition of the prepared alloys
Tabela 1: Nazivna in dejanske kemi~ne sestave pripravljenih zlitin
Alloy
designation
Chemical composition in mass fractions (w/%) Cr and Ni equivalent
C Si Mn P S Ni Cr Mo Creq Nieq
CF-8M <0.08 <2.0 <1.5 <0.04 <0.04 9−12 18−21 2-3 15.8−23.2 11.3−17.4
A 0.06 0.43 1.59 0.03 0.01 11.9 18.0 1.84 16.2 16.8
B 0.07 0.67 1.04 0.03 0.01 11.0 21.7 2.03 20.6 15.9
C 0.06 1.68 0.67 0.03 0.01 9.0 20.8 2.46 21.8 13.4
At the characteristic ingot positions the samples for
metallographic investigations were cut and the average
 -ferrite content of each slice was determined magne-
tically and metallographically. Typical microstructures
visible under a light microscope (LM) are presented in
Figures 2 a, b and c.
A determination of the δ-ferrite content based on the
magnetic induction (Figure 3 a) was performed with two
instruments: i.e., an older one regularly used at NPPs and
a new, digital one, the Feritscope MP-30 Fischer GmbH,
Germany (Figure 3 b), with the appropriate calibration
samples. The measured values obtained with different
methods/instruments are shown in Table 2. It is clear
that the differences in the measured values are relatively
large. But the range of measured values obtained with
one measuring method seems acceptable. The magne-
tically obtained values strongly depend on the sample
surface and its shape, and therefore only the flat,
polished surfaces of metallographic samples were used.
Slices of prepared alloys were then put into the
laboratory batch furnaces and artificially aged (isother-
mally annealed) at elevated temperatures for different
periods of time. The selected temperatures of ageing
were 290, 320 and 350 °C and the time of 24 h, 1 and 6
months, and finally one and two years. After each ageing
temperature/time cycle samples for microstructure cha-
racterization and mechanical testing were prepared. The
average δ-ferrite content was again determined with an
induction-based method on the polished metallographic
samples.
Table 2: Initial δ-ferrite content, determined with different methods/
instruments
Tabela 2: Za~etna vsebnost δ-ferita, dolo~ena na razli~ne na~ine
Alloy
designation
δ-ferrite content in volume fractions (ϕ/%)
Calculated* Metallogra-phically
Ferritmesser
(NPP)
Ferritescope
MP-30
A 2.1 1.2−6.3 1.5−2.5 3.5−4.7
B 14.9 14.0−18.3 10.0−12.0 14.6−19.2
C 38.8 28.5−34.0 26.0−28.0 27.9−32.9
* equations (1) and (2)
For the determination of the absolute magnetic
properties with a hysteresisgraph (permagraph RE3,
Magnet Physic Dr. Steingroever) cylinders dimensions
of ( 10 × 12) mm were used. For measurements with a
vibrating-sampling magnetometer (VSM) small plates
with dimensions of approx. (1.5 × 1.5 × 1.5) mm were
used. The VSM has become a widely used instrument for
determining the magnetic properties of a large variety of
materials in the subzero and elevated temperature
regions18. The investigated material was placed into a
uniform magnetic field and mechanically vibrated, re-
sulting in some magnetic flux change. The induced
voltage in the pick-up coils is proportional to the mag-
netic moment of the sample.
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135 131
Figure 3: Determination of δ-ferrite content based on the magnetic
induction: a) schematic presentation and b) modern digital ferrite-
meter (Ferritescope MP-30, Fischer GmbH, Germany)17
Slika 3: Dolo~evanje vsebnosti δ-ferita na osnovi magnetne indukcije:
a) shemati~ni prikaz in b) moderen digitalni feritmeter (Ferritescope
MP-30, Fischer GmbH, Nem~ija)17
Figure 2: Micrographs of as-cast microstructures of investigated steels: a) alloy A with approx. 18Cr-12Ni, b) alloy B with 22Cr-11Ni and c)
alloy C with approx 21Cr-9Ni, light-grey areas represent austenite, dark-grey areas marked with arrows are islands of δ-ferrite; LM, original
magnification 200-times, selective etching of δ-ferrite (KOH : K3 Fe(CN)6 : H2O=0.25 : 0.25 : 0.50).
Slika 2: Posnetki mikrostrukture preiskovanih jekel: a) zlitina A z 18 % Cr in 12 % Ni, b) zlitina B s pribli`no 22 % Cr in 11 % Ni in c) zlitina C
s pribli`no 21 % Cr in 9 % Ni, svetla podro~ja predstavljajo austenit, temna podro~ja ozna~ena s pu{~icami pa so oto~ki δ-ferita; opti~ni
mikroskop, originalna pove~ava 200-krat, selektivno jedkanje na δ-ferit (KOH : K3 Fe(CN)6 : H2O=0,25 : 0,25 : 0,50).
3 RESULTS AND DISCUSSION
Figure 4 shows the results of the  -ferrite content
determined with Ferritescope, depending on the tem-
perature and the time of isothermal ageing of the
prepared alloys. It is clear that the δ-ferrite content is
virtually independent of the ageing time and tempera-
ture. A slight trend for a decrease in the δ-ferrite content
can be noticed, but only for alloy C at all temperatures
and for alloy B at the highest temperature (350 °C). But
it is in the range of the scatter of the measured results.
Therefore, we can conclude that the induction-based
method is not an appropriate method for monitoring of
the kinetics of thermal degradation of the selected
material. It seems an appropriate result and conclusion
because during spinodal decomposition the changes
occur at the nano-level, connected with small fluc-
tuations of chemical composition and the cell
parameters. However, the Ferritscope detects the average
magnetic induction of much larger local regions (approx.
1 mm2 and 1 mm3, respectively), made up of ferro-
magnetic (ferrite) and nonmagnetic (austenite) grains.
The method can, therefore, only serve as a control for the
initial content of the chemical composition (δ-ferrite
content), as well as an in-situ checking procedure for
new welds and the replaced structural elements of
thermoelectric installations during maintenance. As an
example, Figure 5 shows the change in δ-ferrite content
across the but-weld of thin plates made of 316 L
stainless steel at two different locations.
The hysteresis loop is the basic magnetic characte-
ristic for ferromagnetic materials. From such a loop it is
possible to read the remanence Br, the coercivity Hc, the
saturation magnetization Bs and the permeability µ.
These properties are composition and structure sensitive;
therefore, one can expect that it would be possible to
monitor the kinetics of the spinodal decomposition.
The investigated material can be treated as a
composite material with a different volume content of
magnetic phase (δ-ferrite). Correspondingly, it is
expected that the hysteresis curves of the investigated
alloys change size and shape with an increased content
of  -ferrite. Figure 6 shows the hysteresis loops of the
original non-aged state of all three prepared alloys. As
expected, alloy C, with the largest content of magnetic
phase, has the narrowest and most upright hysteresis
curve, with higher Br, Bs and lower Hc in comparison
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
132 Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135
Figure 6: Hysteresis loops determined by Remagraph RE3, sample
dimensions (φ 10 × 12) mm, original non-aged state of material
Slika 6: Histerezne zanke, ugotovljene z napravo Remagraph RE3,
dimenzije vzorcev (φ 10 × 12) mm, originalno ne`arjeno stanje
materiala
Figure 4: Average δ-ferrite content vs. time and temperature for
isothermal annealing of all the prepared alloys, determined with the
Feritscope MP-30 (Fischer GmbH, Germany)
Slika 4: Povpre~na vsebnost δ-ferita v zlitinah A, B in C v odvisnosti
od ~asa in temperature izotermnega `arjenja, dolo~ena s feritmetrom
(Feritscope MP-30, Fischer GmbH, Nem~ija)
Figure 5: Change of δ-ferrite content across but-welded joint of thin
plates made of 316 L stainless steel at two different measuring
locations
Slika 5: Sprememba vsebnosti δ-ferita vzdol` zvarnega spoja dveh
tankih plo{~ iz nerjavnega jekla vrste 316 L, merjena na dveh razli~nih
lokacijah
with alloy B, which has a lower content of ferromagnetic
phase. Alloy A has a very low content of ferromagnetic
phase and therefore shows an almost linear relation for
the applied field vs. the induction (paramagnetic).
Figure 7 shows the hysteresis loops for alloys aged at
350 °C for 720 h. Its shapes and sizes are similar, and the
differences in the magnetic properties compared to the
original non-aged alloys are negligible. Therefore, the
selected instrument is not the most appropriate for
monitoring the spinodal decomposition of the selected
alloys.
The hysteresis loops were then determined with the
VSM. This method enables the use of smaller samples
without any limits in terms of geometry and surface
quality. The magnetic properties of the prepared mate-
rials were measured at 20 °C and 300 °C. The influence
of the internal stresses had to be taken into account and
any cutting/mechanical machining must be avoided or
carefully carried out. Figure 8 shows the room-tempe-
rature hysteresis loops of alloys aged at 350 °C for two
years. The differences in the hysteresis curves are
evident and they are similar to those obtained with the
permagraph.
Figure 9 shows the room-temperature hysteresis
loops of alloy C aged at 350 °C for different periods of
time. It is clear that with increased ageing time the
saturation magnetization decreases, but the changes are
relatively small (approx. 27–29 emu/g). Also, the
changes in the coercivity and the remanent induction are
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135 133
Figure 9: Typical hysteresis curves obtained with the VSM for alloy
C, aged at 350 °C for different periods of time
Slika 9: Tipi~ne histerezne zanke, ugotovljene z VSM pri zlitini C,
razli~no dolgo ~asa izotermno `arjeni pri 350 °C
Figure 7: Hysteresis loops determined by Remagraph RE3, sample
dimensions (φ 10 × 12) mm, aged 720 h at 350 °C
Slika 7: Histerezne zanke, ugotovljene z napravo Remagraph RE3,
dimenzije vzorcev (φ 10 × 12) mm, izotermno `arjeno 720 h pri 350
°C
Figure 10: Average room-temperature remanent induction R vs.
ageing time obtained for cast alloys with two different δ-ferrite
content, aged at (290, 320 and 350) °C
Slika 10: Povpre~na remanentna indukcija R dolo~ena pri 20 °C v
odvisnosti od ~asa izotermnega `arjenja zlitin z razli~no vsebnostjo
δ-ferita, `arjeno pri (290, 320 in 350) °C
Figure 8: Typical hysteresis loops determined with the VSM,
obtained for alloys A, B and C and aged 2 years at 350 °C
Slika 8: Tipi~ne histerezne zanke, ugotovljene z VSM na zlitinah A, B
in C, izotermno `arjene dve leti pri 350 °C
very small and are close to the limits of the sensitivity of
the VSM. The remanent induction (R) and the coercivity
(Hci) were read from the measured hysteresis curves and
put into the diagrams in order to show the intrinsic
properties vs. the ageing conditions for the individual
alloys.
Figure 10 shows the room-temperature R of the
alloys A, B and C, aged up to two years at (290, 320 and
350) °C. Its change for non-magnetic alloy A is almost
negligible, but it increases significantly for alloy C,
which has the highest content of ferromagnetic phase.
The alloy B shows an almost negligible change, similar
to alloy A.
Figure 11 shows the room-temperature coercivity of
alloys A, B and C aged up to two years at selected
temperatures. In this case the alloys show an increase of
coercivity with ageing time. Surprisingly, alloy C at 320
°C does not show this increase.
Inconsistencies are noticed in the obtained results and
shown in Figures 10 and 11. However, it is obvious that
the change in the remanent induction and the coercivity
are connected with structural changes. Therefore, we can
also conclude that the kinetics of the spinodal
decomposition can be monitored from the change in the
absolute magnetic properties, but probably a more
suitable instrument, specially designed for measurements
of low coercivity soft magnetic materials, should be
used.
4 CONCLUSIONS
The magnetic-induction-based determination of the
δ-ferrite content showed that it does not change
significantly with ageing temperature and time. It
confirms that only the internal ferrite structure is
changed, during ageing, because of the spinodal decom-
position. Therefore, it can be concluded that magnetic
induction is not appropriate method to observe spinodal
decomposition in SS cast alloys. It can only serve as
information or control for either the initial materials’
content of  -ferrite or as checking procedure during the
maintenance of thermo-energetic objects.
The kinetics of spinodal decomposition can be
monitored by a change in the absolute magnetic pro-
perties. However, these methods are destructive, as well
as stress and temperature sensitive. The scatter of the
results is relatively large and, therefore, sometimes they
are inconsistent. The changes in the magnetic properties
are also relatively small and so an appropriate measuring
instrument has to be chosen.
5 REFERENCES
1 K. H. Herter et al: Leak-before-break behaviour of piping compa-
rison of testing with calculation, VGB PowerTech, 11 (2001), 84–92
2 L. Skanberg: PWSCC observations Ringhals 3 and 4 safe ends,
SMiRT-16 Conference, Swedish Nuclear Power Inspectorate,
August 2001
3 A. Trautwein, W. Gysel: Influence of long-time aging of CF8 and
CF8M Cast steel at temperatures between 300 and 500 °C on impact
toughness and structural properties, ASTM Symposium on Stainless
Steel and Related Alloys, ASTM Special Technical Publication, 756
(1980), 165–189
4 J. Kwon, J-c. Park, J. Lee, W. Lee, Y-w Park: An investigation of the
degradation characteristics for casting stainless steel, CF8M, under
high temperatures, Nuclear Engineering and Design, 198 (2000),
227–240
5 Rodney Fordham: The PWR Nuclear Power Station Risk Problem,
http://www.cabinet-office.gov.uk, July 2001, 1–6
6 C. E. Jaske: Process Equipment Fitness for Service Assessments
Using API RP 579, Clarion Technical Conferences, Houston, Texas,
November 2001, http://www.clarion.org/
7 S. R. Doctor: Overview of V. C. Summer hot leg loop a nozzle to
pipe dissimilar metal weld, SMiRT-16 Conference, U. S. Department
of Energy Pacific Northwest National Laboratory, August 2001
8 A. Mitra et al: Evaluation of magnetic phases in cold worked and
weldments of AISI 304 stainless steel, 15-th World Conference of
Non Destructive Testing, Rome, Italy, Oct. 2000
9 M. Vasudevan et al.: Delta ferrite prediction in stainless steel welds
using neural network analysis and comparison with other prediction
methods, Journal Materials Processing Technology, 142 (2003),
20–28
10 J. Janovec, B. [u{tar{i~, J. Medved, M. Jenko: Phases in austenitic
stainless steels, Mater. Tehnol., 37 (2003) 6, 307-312
11 B. [u{tar{i~, N. Gubeljak, J.V. Tuma: Effect of delta ferrite content
on the mechanical properties of cast stainless steel. Rev. Int. Ing.
Syst. Prod. Mechan., Metz, France, 10 (2004), 48–54.
12 J. V. Tuma, B. [u{tar{i~, F. Vodopivec: The effect of ageing tempe-
rature and time on the mechanical properties of Fe-NiCrMo alloys
with different contents of delta ferrite, Nucl. Eng. Des., 238 (2008) 7,
1511–1517
13 C. Pokor, C. Pages, B. Marini: Thermal Ageing of cast duplex stain-
less steels of PWR; material assessment on the removed cast elbows,
presentation at CEA-EDF-IMT meeting in the frame of CEA-IMT
bilateral collaboration, Saclay, France, January 2008
14 Determination of Ferrite Content in Austenitic Stainless Steel
Castings, International standard ISO/FDIS 13520:2002(E)
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
134 Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135
Figure 11: Average room-temperature coercivity vs. ageing time
obtained for the investigated cast alloys, aged at (290, 320 and 350)
°C
Slika 11: Povpre~na koercitivnost, izmerjena pri 20 °C, v odvisnosti
od ~asa izotermnega `arjenja zlitin z razli~no vsebnostjo δ-ferita,
`arjeno pri (290, 320 in 350) °C.
15 Standard practice for steel casting, austenitic alloy, estimating ferrite
content thereof, ASTM A800/A800M-01, 2003 16 ISO/FDIS
13520; Determination of ferrite content in austenitic stainless
castings, 2002
16 International standard EN ISO 8249:2000E; Welding-determination
of ferrite number (FN) in austenitic and duplex ferritic Cr-Ni
stainless steel weld materials, Brussels, 2000
17 Measurement of the ferrite content in austenitic and duplex steel,
feritscope MP 30, Helmut Fischer GmbH&Co, Germany,
commercial catalogue, http://www.helmut-fischer.de
18 B. C. Dodril, B. J. Kelly: Measurement with VSM permanent magnet
materials, Lake Shore Cryotronics, Inc., 1999, 1–6
B. [U[TAR[I^ ET AL.: MAGNETIC CHARACTERISTICS OF ISOTHERMALLY AGED Cr-Ni-Mo-BASED ALLOYS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 129–135 135

^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
THE OXIDATION OF DUPLEX STAINLESS STEEL AT
MODERATELY ELEVATED TEMPERATURES
OKSIDACIJA DUPLEKSNEGA NERJAVNEGA JEKLA PRI
ZMERNO POVI[ANIH TEMPERATURAH
^rtomir Donik, Aleksandra Kocijan, Irena Paulin, Monika Jenko
Institute of Metals and Technology, Lepi pot 11, SI-1000 Ljubljana, Slovenia
crtomir.donikimt.si
Prejem rokopisa – received: 2009-03-17; sprejem za objavo – accepted for publication: 2009-05-04
The surface oxidation of DSS 2205 duplex stainless steel was studied by X-ray photoelectron spectroscopy (XPS). Two
different techniques were used to produce thin oxide layers on polished and sputter-cleaned duplex stainless-steel samples.
These samples were exposed to a 10–5 mbar pressure of pure oxygen inside the vacuum chamber and oxidized with an oxygen
plasma from room temperature up to 300 °C. The experiments were made with the alloy after a controlled oxidation with
oxygen atoms created in an inductively coupled plasma. The experiments were performed in the temperature interval from room
temperature up to 300 °C. The compositions of the modified oxidized surfaces were determined from the XPS survey scans, and
the chemistry of selected elements from the higher-energy-resolution scans of the appropriate peaks. Various Fe/Cr oxidized
layers and various oxide thicknesses were observed and correlated with the temperature. It was found that all the techniques
produced oxide layers with various traces of metallic components and with the maximum concentration of chromium oxide and
iron oxide in layers close to the oxide-layer–bulk-metal interface.
Key words: duplex stainless steel, oxidation, plasma oxidation, XPS, stainless steel
Z rentgensko fotoelektronsko spektorskopijo (XPS) smo raziskovali povr{insko oksidacijo dupleksnega nerjavnega jekla DSS
2205. Z `eljo, da bi dobili tanke oksidne plasti na atomsko ~istih povr{inah, smo jih oksidirali na dva na~ina. Vzorci so bili
izpostavljeni ali atmosferi ~istega kisika v pripravljalni komori pri tlaku 10–5 mbar in nato analizirani z XPS ali oksidirani v
kisikovi plazmi pri temperaturah od 22 °C do 300 °C. Vse tanke plasti smo preu~ili z XPS- tehniko za dolo~itev sestave
oksidnih plasti na povr{ini. Posebej smo nato pri ve~ji energijski lo~ljivosti posneli {e elemente za natan~no dolo~itev
spreminjanja skozi oksidno plast, t. i. XPS-globinski profil. Ugotovili smo, da so nastale pri razli~nih pogojih razli~ne oksidne
plasti Fe/Cr razli~nih debelin v povezavi s temperaturo in na~inom oksidacije. Opazili pa smo tudi, da je skoraj pri vseh
temperaturah oksidacije ostal tudi dele` neoksidiranih kovinskih elementov na povr{ini.
Klju~ne beside: dupleksno nerjavno jeklo, oksidacija, plazemska oksidacija, XPS, nerjavno jeklo
1 INTRODUCTION
Stainless steel is one of the most widely used
materials and has a large variety of applications.
Therefore, it is exposed to a wide range of conditions.
Since it is an alloy, various oxides can be formed on its
surface, and they can change the mechanical, chemical
and physical properties of the material. Steel’s corrosion
resistance originates in the Cr-rich oxide layer that acts
as a barrier against ion diffusion between the alloy and
the ambient phase.1–4 Custom steel grades are designed
for specific applications by optimizing their properties
using specific alloy compositions. The study was on
mass fractions 22 % chromium, 5–6 % nickel, 3 %
molybdenum, 2 % manganese, nitrogen-alloyed duplex
stainless steel (DSS 2205, also known as W.Nr. 1.4462),
with high general, localized and stress corrosion
resistance properties in addition to high strength and
excellent impact toughness. Duplex stainless steels
contain two equilibrium phases, i.e., ferrite and austenite.
This type of stainless steel is increasingly used because
of its several superior properties. It also has better
corrosion resistance and erosion-fatigue properties as
well as a lower thermal expansion coefficient and higher
thermal conductivity than plain austenitic stainless steel.
Its yield strength is about twice as high as that of
austenitic stainless steels and this allows the designer to
save material and makes the alloy more cost competitive
in comparison to the 316L or 317L stainless steels. Alloy
2205 is particularly suitable for applications in the
temperature interval from –50 °C to 300 °C. Tempera-
tures outside this interval may also be considered, but
there are some restrictions, particularly for welded
structures.5,6
The oxidation and corrosion resistance of stainless
steel has already been the subject of many studies. The
corrosion resistance of stainless steel is known to be
based on the Cr2O3 chromium oxide at the surface that is
considered to act as a protective layer against corrosion
due to its low diffusion constants for oxygen and metal
ions.7,8 The oxide layer formed on stainless-steel surfaces
is usually not uniform in terms of the depth. Double or
even triple layers can be formed on the surface,
depending on the alloy composition, on the oxidizing
conditions (oxidizing atmosphere, time, and tempera-
ture), and also on the different duplex stainless-steel iron
phases.9,10 Oxide films may theoretically consist of
various iron and chromium oxides as well as their
Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142 137
UDK 669.14.018.8:620.191 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)137(2009)
mixtures. Several authors have studied the formation of
thin oxide layers on stainless steels, being oxidized in air
or oxygen at various pressures and various temperatures.
The oxidation in air showed that at most temperatures a
duplex oxide layer was formed; the outer layer, α-Fe2O3,
was formed on the top of the inner oxide layer, an
iron-chromium oxide. Under reduced oxygen partial
pressures a fine chromium-rich oxide grows through the
iron oxide layer that was formed initially. At very low
oxygen pressures (below 10–3 Pa) and temperatures
above 350 °C, chromium predominates throughout the
oxide layer 11–15. At low temperatures, on the other hand,
i.e., up to 400 °C, only iron oxide is formed on the
surface of stainless steels. This is not true, however, for
all the FeCr alloys. Hultquist et al. 16, e.g., reported that a
significant fraction of oxidized Cr was detected by the
X-ray photoelectron spectroscopy (XPS) of FeCr alloys
that were exposed to air at temperatures between 25 °C
and 180 °C.
The aim of the present study was to examine the
initial phases of oxide growth on the 2205 duplex
stainless steel as a function of temperatures up to 300 °C.
The oxide layers were produced by the controlled
exposure of polished duplex stainless-steel samples to
oxygen atoms. X-ray photoelectron spectroscopy (XPS)
was applied to measure the depth distributions of the
oxide films formed on the surface by the sputter depth
profiling. To our knowledge, this is the first report on the
depth distribution of oxide-layer compositions on DSS
2205 for the conditions specified in this paper. The
results of an ongoing study of the oxidation in air will be
described in a subsequent paper.
2 EXPERIMENTAL
Duplex 2205 stainless steel (DSS) was obtained from
the Acroni, d. o. o., steel plant, which certified the alloy
as 2205 duplex stainless steel with the composition as
given in Table 1.2
The samples were mounted in resin, ground with
emery papers up to No. 4000, and then polished with
1-µm diamond paste. Metallographic samples of
approximately (8 × 8 × 1) mm were thus prepared. These
samples were then mounted onto sample holders for the
XPS examinations, using UHV-compatible double-sided
sticky tape.
The oxidation was achieved in an experimental
reactor for the passivation of metals. The reactor was a
glass tube with a length of 40 cm and a diameter of 4 cm.
It was evacuated from one side with a two–stage rotary
pump with a pumping rate of 16 m3h–1. The ultimate
pressure in the system was about 4 Pa and the residual
atmosphere consisted mainly of water vapour. The other
end of the reactor was connected to a source of neutral
oxygen atoms. The source was a low-pressure oxygen
plasma created by an inductively coupled RF generator
operating at a frequency of 27.12 MHz and with a
variable output power. The applied experimental
procedure is described in detail elsewhere.17–22 In our
case the RF power was fixed at about 250 W. The
plasma was characterized with a double Langmuir probe
and a Fibre Optics Catalytic Probe (FOCP).17,21,23,24 The
Langmuir probe was placed in the centre of the plasma.
At the applied pressure of 75 Pa the electron temperature
was about 4 eV, while the electron density was about 7 ·
1015 m–3. The FOCP was mounted in the experimental
reactor for the passivation of metals close to the
samples.24 At a pressure of 75 Pa the O-atom density
reached its peak value of about 1 · 1022 m–3. The neutral
gas kinetics temperature in the passivation reactor was
close to room temperature.
The samples were mounted on a heatable holder,
heated to the appropriate temperature and kept in the
passivation reactor for 100 s. In this period they received
a dose of neutral oxygen atoms of about 1.5 · 1024 m2.
The XPS depth profiles of the oxide layers were
measured with a VG Microlab 310F AES/SEM/XPS. Mg
Kα radiation at 1253.6 eV with an anode power of 200 W
(the anode voltage was 12.5 kV and the emission current
16 mA) was applied in all the XPS measurements. A
3-keV, 1-µA Ar+ ion beam scanned over an (8 × 8) mm
area was applied for the sputter depth profiling with an
incidence angle of 60° (measured with respect to the
sample normal). The approximate sputtering rate under
the conditions applied here was 1 nm/min for SiO2. This
was not inconsistent with some calibration measure-
ments performed on metallic and oxide-type samples as
well as with some reference data for the sputtering rates
of Fe and Cr and their oxides.13,25–29
The spectra were acquired using the Avantage 3.41v
software supplied by the manufacturer. CasaXPS soft-
ware30 was used to process the data. The XPS spectra
were fitted to determine the different chemical states’
contributions to the total peak intensity.
3 RESULTS AND DISCUSSION
Figure 1 shows high-resolution XPS scans of the Cr
2p3/2 and Fe 2p3/2 transitions with the corresponding
metallic and oxide components before sputtering the
DSS 2205 that was prepared at 22 °C. In the case of Cr
2p3/2 four peaks were used for the fitting, i.e., the
metallic peak and the three oxide peaks, CrO3, Cr2O3 and
Cr(OH)3. The binding-energy differences of the compo-
^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
138 Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142
Table 2: Chemical composition of the duplex stainless steel in mole fractions x/%
Cr Ni Mn Si P S C Mo N Fe
23.96 4.82 1.48 0.85 0.05 0.00 0.13 1.83 0.67 66.21
nents were fixed at the values obtained from references
7,13,31,32 and the full widths at half maximum (FWHM) for
the three oxide peaks were fixed to be the same. The
FWHM applied for the metallic component was taken
from the Cr peak, being measured after the surface
oxides were removed. The expected presence of the
three oxides was based on previous work. The Fe 2p3/2
was fitted with five peaks in a similar way.6,26,33–39
Figure 2 shows the Cr 2p3/2 and 2p1/2 transitions after
various sputtering times for the sample prepared at 22
°C. At each depth the 2p3/2 peak was fitted as discussed
above (see Figure 1). The total Cr oxide concentration
was approximated by adding the areas of the three
oxide-peak components. The intensities of the oxide
components were reduced by sputtering and eventually
only the metallic component remained. The Fe data were
processed similarly.
Figure 3 shows an XPS depth profile of the oxide
layer after the oxidation of the DSS 2205 with an
exposure to 200 L (Langmuir) of oxygen in the prepa-
ration chamber. The stoichiometry of the oxide was, in
the best case, an approximation because of the ion-
sputtering-induced effects, such as the reduction of the
oxidation state.
XPS depth profiles were also measured with the DSS
2205 after much longer exposures to oxygen, i.e., up to
8000 L (Figure 4), without obtaining substantially
different results. This suggested that an increase in the
pressure and/or different atmospheric conditions, e.g.,
humidity, were needed to produce a thicker oxide layer.
The profiles in Figures 3 and 4 showed that these
oxide layers were ill-defined with slowly changing
concentrations of oxidized and unoxidized metals from
the surface of the oxide layer towards the interior of the
sample. Also, large fractions of unoxidized metals were
present at the very surface of the oxide layers. These
concentrations might be partially overestimated since the
estimations of oxide-layer thicknesses obtained from the
profiles in Figures 3 and 4 were 2–3 nm, i.e., of the
order of the probing depth of the applied XPS in the
^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142 139
Figure 4: XPS depth profile of DSS at 8000 L exposure in the
preparation chamber
Slika 4: XPS; globinski profil DSS pri izpostavljenosti 8000 L v
pripravljalni komori
Figure 2: Set of high-resolution XPS spectra of Cr 2p transitions used
for constructing the Cr depth profiles in Figure 3. The bottom
spectrum is from the surface before sputtering, and the spectra above
it belong to spots of increased depth from the original surface. The
spectra have been offset in intensity for clarity.
Slika 2: Visoko lo~ljivi XPS-spektri Cr 2p vrha, ki smo jih uporabili
za izdelavo globinskih profilov v XPS sliki 3. Spodnji spekter je na
povr{ini vzorca pred jedkanjem, zgornji spekter pa je po seriji jedkanj.
Spektri so zamaknjeni po intenziteti za bolj{i pregled.
Figure 3: XPS depth profile of DSS at 200 L exposure in the pre-
paration chamber
Slika 3: XPS; globinski profil DSS pri izpostavljenosti 200 L v pri-
pravljalni komori
Figure 1: Examples of fitted Cr 2p3/2, Fe 2p3/2 XPS spectra
Slika 1: XPS-spektra Cr 2p3/2, Fe 2p3/2 z oksidnimi komponentami
analysis, so there existed some possibility that unoxi-
dized metals from the bulk influenced the measurements.
Even in those poorly defined oxide layers, two-oxide
structures could already be detected: a higher concen-
tration of iron oxide at, and close to, the surface, with
chromium oxide being the predominant oxide at greater
depths, in contrast to the layer of the bulk interface.
The stoichiometry of the oxide was, in the best case,
an approximation due to ion-sputtering-induced effects,
such as the reduction of the oxidation states.16,27,28,40
However, valid comparisons could be made between the
relative oxidation degrees at various temperatures. The
formation of an oxide film with an inner region,
essentially consisting of chromium oxide, and an outer
region, mainly consisting of iron oxide, was also
observed in the samples that were heated up to 300 °C.
A. P. Greeff et al.41 found that the formation of Fe2O3 on
the FeCrMo steel was, in the interval between room
temperature and 300 °C, more rapid than that of the
initially formed Cr2O3, and it was eventually formed on
the top of the Cr-oxide. This is in agreement with our
results. The concentration of the unoxidized metallic
species in the topmost part of the oxide layer was much
lower than the concentration of their oxidized
counterparts, while there was virtually no unoxidized
chromium or iron in the oxide layer. Figure 6 shows a
different metal-oxide distribution for the sample that was
oxidized at a higher temperature of 150 °C (423 K). The
concentration of iron oxides close to the surface was, in
this case, much higher than that in the sample oxidized at
22 °C (Figure 5), being about mole fraction x = 85 %
compared to the previous value of about mole fraction of
x = 70 %, and it showed the segregation of iron oxides at
the surface. On the other hand, the concentration of
chromium oxides was about x = 10 %, compared to x =
25 %, showing the reduced surface segregation of Cr
(oxides) at 150 °C . The thickness of the oxide layer was
also somewhat greater than that formed with plasma
oxidation at room temperature and very much thicker
than that formed with just pure oxygen. In both cases,
the amount of chromium oxides increased at the
interface between the surface oxides and the underlying
steel. Figure 7 shows the XPS depth profile of DSS
oxidized at 300 °C (573 K). The further increase of iron
oxides and the further reduction of chromium oxides in
the oxide film were quite dramatic compared to the
depth profiles of the samples being oxidized at lower
temperatures, as well as the increased thickness of the
oxide layer itself. Other effects observed in the sputter
depth profiles at 22 °C and 150 °C (Figures 5 and 6)
remained unchanged, i.e., the increased concentration of
chromium oxides at the interface of the oxide layer and
steel substrate. A comparison of Figures 5, 6 and 7
makes it clear that the total oxide layer was thicker, and
the ratio of the iron oxides to chromium oxides increased
with increasing oxidation temperature (up to 300 °C).
However, it was observed that the amount of chromium
oxides increased at the interface of the surface oxide
layers and substrate to approximately x ≈ 35 % in all
cases. It is known, as a general rule, that iron oxide is the
predominant surface oxide in stainless steels at
temperatures up to 400 °C,9 and this research showed the
actual temperature dependence of this predominant
species within the mentioned temperature range. The
composition and oxidation behaviour of the various
grains was known to be very similar, and the XPS
analysis was made on a 1 mm × 2 mm area that was
much greater than the grain size.
^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
140 Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142
Figure 6: XPS depth profile of DSS oxidized at 150 °C
Slika 6: XPS; globinski profil DSS po oksidaciji v plazmi pri 150 °C
Figure 7: XPS depth profile of DSS oxidized at 300 oC
Slika 7: XPS; globinski profil DSS po oksidaciji v plazmi pri 300 °C
Figure 5: XPS depth profile of DSS oxidized at 22 °C
Slika 5: XPS; globinski profil DSS po oksidaciji v plazmi pri 22 °C
An enrichment with Mo throughout the oxide layer
was observed at all temperatures. The oxide layer of
DSS had a duplex structure of Fe oxides, predominantly
Fe2O3, on the outer side and Cr oxides, predominantly
Cr2O3, on the inner side of layer, with an enrichment of
Mo underneath the Cr oxide layer. This enrichment of
Mo might be explained by the assumption that Mo
oxidizes less readily than Fe or Cr. The Gibbs free
energies for the formation of Fe2O3, Cr2O3 and MoO3 at
700 K (427 °C) are – 979 J/mol, – 1209 J/mol and – 799
J/mol, respectively, and they correspond to – 744,8
J/mol, – 807,5 J/mol and – 535,5 J/mol O2. From the
viewpoint of thermodynamics, the oxidation of Mo is the
least favourable. Furthermore, according to Mathieu and
Landolt 5, the presence of Mo at the oxide/metal
interface could influence the oxide growth in two ways:
Mo could reduce the activity of Fe at the phase boundary
or it could form a diffusion barrier for Fe and Cr ions
between the oxide and the metal.
In summary, the oxidation of DSS 2205 with atomic
oxygen revealed two important factors that influenced
the growth and composition of the oxide layers that were
formed at various temperatures, i.e., the segregation of
the alloying elements to the surface and the influence of
atomic oxygen on the oxidation process.
4 CONCLUSIONS
Two different types of oxide layers were obtained.
The first one was very thin, of the order of 2–5 nm,
obtained with exposure to oxygen at 10–5 mbar,
independent of the total exposure time, even when the
exposure times differed by two orders of magnitude. The
second one was of the order of a few tenths of nm,
obtained by exposure to plasma oxidation, also at higher
temperatures. A more detailed analysis showed that
despite the great similarities between the samples that
were oxidized either with oxygen or with plasma there
were also differences in certain numerical parameters
(e.g., the ratios of unoxidized versus oxidized iron, or
metal in general) that allowed us to denote the plasma-
oxidized sample as a more "thoroughly" oxidized. A
common characteristic of all the examined layers was a
double-oxide stratification, with the regions closer to the
surface that contained higher concentrations of iron
oxide, and those more towards the interior, exhibiting
higher concentrations of chromium oxide. The compo-
sitions of the oxide films formed on the DSS 2205 by
exposure to plasma oxygen atoms are presented from
room temperature to 300 °C. Different oxide layers were
formed on the surface, depending on the temperature
during the atomic oxidation. Small increments of tempe-
rature above room temperature resulted in increased
amounts of Fe oxides and reduced amounts of Cr oxides
on the surface, while almost pure Fe oxides were formed
at 300 °C. There was an enrichment of chromium oxides
at the interface between the surface oxide and the bulk
material at all the temperatures. Due to this Cr enrich-
ment at the interface, the upper part of the bulk material
consisted of the so-called chromium-depleted layer. This
study also provided important results about the
behaviour of the DSS 2205 that was treated with oxygen
atoms from relatively low temperatures up to just 300
°C. To examine this advanced exposure a chamber with
the possibility of increasing the exposure pressure,
measured in mbars, could be used. Another possibility
for further studies is to use ARXPS for more accurate
depth profiling of the thinnest oxide layers.
5 REFERENCES
1 G. Betz, G. K. Wehner, L. Toth, Journal of Applied Physics 45
(1974), 5312–5316
2 ASM specialty handbook: Stainless Steels, 1996, 32–34
3 M. Veljkovic, J. Gozzi, Journal of Pressure Vessel Technology-
Transactions of the ASME 2007, 129, 155–161 10.1115/1.2389034
4 S. Ohkido, Y. Ishikawa, T. Yoshimura, POSAP analysis of the oxide
alloy interface in stainless-steel, 1994, Chapter, 261–265
5 H. J. Mathieu, D. Landolt, Corrosion Science 26 (1986), 547–559
6 C. Palacio, H. J. Mathieu, V. Stambouli, D. Landolt, Surface Science
295 (1993), 251–262
7 C. M. Abreu, M. J. Cristobal, X. R. Novoa, G. Pena, M. C. Perez,
Surface and Interface Analysis 2008, 40, 294–298 10.1002/sia.2796
8 H. Asteman, K. Segerdahl, J. E. Svensson, L. G. Johansson, M.
Halvarsson, J. E. Tang, p. trans tech, In: Oxidation of stainless steel
in H2O/O-2 environments – Role of chromium evaporation, 2004,
Chapter, 775–782
9 A. Vesel, A. Drenik, M. Mozetic, A. Zalar, M. Balat-Pichelin, M.
Bele, In: AES investigation of the stainless steel surface oxidized in
plasma, 2007, Chapter, 228–231
10 A. Vesel, M. Mozetic, A. Drenik, N. Hauptman, M. Balat-Pichelin,
Applied Surface Science 255 (2008), 1759–1765 10.1016/j.apsusc.
2008.06.017
11 Y. Ishikawa, K. Odaka, Reduction of outgassing from stainless
surfaces by surface oxidation, 1990, Chapter, 1995–1997
12 Y. Ishikawa, T. Yoshimura, Journal of Vacuum Science & Techno-
logy a-Vacuum Surfaces and Films 13 (1995), 1847–1852
13 A. Kocijan, C. Donik, M. Jenko, Corrosion Science 49 (2007),
2083-2098 10.1016/j.corsci.2006.11.001
14 S. Shibagaki, A. Koga, Y. Shirakawa, H. Onishi, H. Yokokawa, J.
Tanaka, Thin Solid Films 303 (1997), 101–106
15 G. J. Stokkers, A. Vansilfhout, G. A. Bootsma, T. Fransen, P. J.
Gellings, Corrosion Science 23 (1983), 195–204
16 G. Hultquist, M. Seo, N. Sato, Oxidation of Metals 25 (1986),
363–372
17 U. Cvelbar, M. Mozetic, A. Ricard, Characterization of oxygen
plasma with a fiber optic catalytic probe and determination of
recombination coefficients, ed. by Editor, City, 2005, Chapter,
834–837
18 U. Cvelbar, M. Mozetic, Journal of Physics D-Applied Physics 2007,
40, 2300–2303 10.1088/0022–3727/40/8/s09
19 U. Cvelbar, K. Ostrikov, A. Drenik, M. Mozetic, Applied Physics
Letters 92 (2008), 133505 10.1063/1.2905265
20 A. Drenik, U. Cvelbar, K. Ostrikov, M. Mozetic, Journal of Physics
D-Applied Physics 41 (2008), 115201 10.1088/0022-3727/41/11/
115201
21 M. Mozetic, A. Zalar, U. Cvelbar, D. Babic, AES characterization of
thin oxide films growing on Al foil during oxygen plasma treatment,
2004, Chapter, 986–988
22 M. Mozetic, U. Cvelbar, A. Vesel, N. Krstulovic, S. Milosevic, Ieee
Transactions on Plasma Science 36 (2008), 868–869 10.1109/tps.
2008.925383
^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142 141
23 U. Cvelbar, M. Mozetic, I. Poberaj, D. Babib, A. Ricard, Characteri-
zation of hydrogen plasma with a fiber optics catalytic probe, 2005,
12–16
24 M. Mozetic, Characterization of reactive plasmas with catalytic
probes, 2007, Chapter, 4837–4842
25 J. R. Lince, S. V. Didziulis, D. K. Shuh, T. D. Durbin, J. A. Yarmoff,
Surface Science 277 (1992), 43–63
26 D. Mandrino, M. Godec, M. Torkar, M. Jenko, Surface and Interface
Analysis 40 (2008), 285–289 10.1002/sia.2718
27 N. S. McIntyre, F. W. Stanchell, Journal of Vacuum Science &
Technology 16 (1979), 798–802
28 A. Zalar, Thin Solid Films 124 (1985), 223–230
29 A. Kocijan, I. Milosev, D. K. Merl, B. Pihlar, Journal of Applied
Electrochemistry 34 (2004), 517–524
30 N. Fairley, http://www.casaxps.com/
31 A. Maetaki, M. Yamamoto, H. Matsumoto, K. Kishi, Surface
Science 445 (2000), 80–88
32 A. Kocijan, I. Milosev, B. Pihlar, Journal of Materials Science-
Materials in Medicine 15 (2004), 643–650
33 Dj. Mandrino, M. Jenko A study of oxide layers on electrical steels
prepared with different thermal treatments, 2001, Chapter, 157–161
34 Dj. Mandrino, M. Godec, P. Skraba, B. Sustarsic, M. Jenko, AES,
XPS and EDS analyses of an iron-based magnetic powder and an
SMC material, 2004, Chapter, 912–916
35 Dj. Mandrino, M. Lamut, M. Godec, M. Torkar, M. Jenko, Surface
and Interface Analysis 39 (2007), 438–444 10.1002/sia.2534
36 J. Baltrusaitis, D. M. Cwiertny, V. H. Grassian, Physical Chemistry
Chemical Physics 9 (2007), 5542–5554 10.1039/b709167b
37 C. R. Brundle, T. J. Chuang, K. Wandelt, Surface Science 68 (1977),
459–468
38 A. P. Grosvenor, B. A. Kobe, M. C. Biesinger, N. S. McIntyre,
Surface and Interface Analysis 36 (2004), 1564–1574 10.1002/
sia.1984
39 E. Paparazzo, Journal of Electron Spectroscopy and Related
Phenomena 154 (2006), 38–40 10.1016/j.elspec.2006.09.004
40 C. Leygraf, G. Hultquist, Surface Science 61 (1976), 69–84
41 A. P. Greeff, C. W. Louw, H. C. Swart, Corrosion Science 42
(2000), 1725–1740
^. DONIK ET AL.: THE OXIDATION OF DUPLEX STAINLESS STEEL ...
142 Materiali in tehnologije / Materials and technology 43 (2009) 3, 137–142
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG
WELDING PROCESS USING GREY RELATION
ANALYSIS AND TAGUCHI METHOD
OPTIMIZACIJA GEOMETRIJE TIG-VARKOV Z GREYJEVO
ANALIZO IN TAGUCHIJEVO METODO
Ugur Esme1, Melih Bayramoglu2, Yugut Kazancoglu3, Sueda Ozgun4
1Mersin University Tarsus Technical Education Faculty Department of Mechanical Education, 33140 Mersin-Tarsus, Turkey
2Cukurova University Engineering and Architecture Faculty, Department of Mechanical Engineering, 01330 Balcali-Adana, Turkey
3Izmir University of Economics, Sakarya cad. 156, Balcova-Izmir, Turkey
4Mersin University Vocational School of Gulnar, 33400 Gulnar-Mersin, Turkey
uesme2003hotmail.com
Prejem rokopisa – received: 2009-01-26; sprejem za objavo – accepted for publication: 2009-02-17
This study investigated the multi-response optimization of tungsten inert gas welding (TIG) welding process for an optimal
parametric combination to yield favorable bead geometry of welded joints using the Grey relational analysis and Taguchi
method. Sixteen experimental runs based on an orthogonal array of Taguchi method were performed to derive objective
functions to be optimized within experimental domain. The objective functions have been selected in relation to parameters of
TIG welding bead geometry; bead width, bead height, penetration, area of penetration as well as width of heat affected zone and
tensile load. The Taguchi approach followed by Grey relational analysis to solve the multi-response optimization problem. The
significance of the factors on overall quality characteristics of the weldment has also been evaluated quantitatively by the
analysis of variance method (ANOVA). Optimal results have been verified through additional experiments. This shows
application feasibility of the Grey relation analysis in combination with Taguchi technique for continuous improvement in
product quality in manufacturing industry.
Keywords: TIG welding; Grey relation analysis; Tensile test
Predstavljeni so rezultati optimizacije odgovorov pri varjenju v inertni atmosferi z volframovo elektrodo (TIG) za dosego
optimalne geometrije varka z uporabo analize po Greyju in Taguchijevi metodi. [estnajst eksperimentalnih varkov je bilo
nanesenih v ortogonalni razporeditvi po Taguchijevi metodi, da bi dolo~ili parametre, ki jih je treba eksprimentalno optimirati.
Ti parametri, izbrani v povezavi s parametri TIG-varjenja, so: geometrija varka, {irina varka, vi{ina varka, penetracija, povr{ina
penetracije, {irina toplotne zone in natezna obremenitev. Taguchijevemu pribli`ku je sledila Greyjeva racionalna analiza zaradi
optimizacije ve~ odgovorov pri problemu optimizacije. Pomen dejavnikov za karakteristike splo{ne kakovosti zvarov je
kvantitativno ocenjen z analizo po meodi variance (ANOVA). Optimalne rezultate smo verificirali z dodatnimi poskusi. To ka`e
na mo`nost uporabe Greyjeve analize v kombinaciji s Taguchijevo tehniko za stalno rast kakovosti v predelovalni industriji.
Klju~ne besede: TIG-varjenje, Greyjeva analiza, natezni preizkus
1 INTRODUCTION
In today’s manufacturing world, quality is of vital
importance. Quality can be defined as the degree of
customer’s satisfaction as provided by the procured
product. The product quality depends on the desired
requirements gained in the product that suits its
functional requirements in various areas of application.
In the field of welding, weld quality mainly depends
on the mechanical properties of the weld metal and heat
affected zone (HAZ), which in turn is influenced by
metallurgical characteristics and chemical compositions
of the weld.
Moreover, these mechanical-metallurgical features of
the weldment depend on the weld bead geometry, which
are directly related to welding process parameters. In
other words, weld quality depends on welding process
parameters 1.
Tungsten inert gas (TIG) welding is a multi-objective
and multi-factor metal fabrication technique. Several
process parameters interact in a complex manner
resulting direct or indirect influence on weld bead
geometry, mechanical properties and metallurgical
features of the weldment as well as on the weld
chemistry. Basically, TIG weld quality is strongly
characterized by the weld bead geometry as shown in
Figure 1.
Bead geometry variables; heat affected zone, bead
width, bead height, penetration and area of penetration
are greatly influenced by welding process parameters;
welding speed, welding current, shielding gas flow rate
and gap distance and also it plays an important role in
Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149 143
UDK 621.791.05 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)143(2009)
Figure 1: Weld bead geometry
Slika 1: Geometrija varka
determining the mechanical properties of the weld such
as Tensile load.
It is necessary to find an optimal process condition
capable of producing desired weld quality. However, this
optimization should be performed in such a way that all
the objectives should fulfill simultaneously. Such an
optimization technique is called multi-response optimi-
zation.
Literature shows that work has been explored on
various aspects of modeling, simulation and process
optimization in TIG welding. In this study, detailed
analysis has been made to establish relationships
between welding parameters and weld bead geometry
and weld quality leading to an optimal process.
Saurav Datta and et al. 1 developed a grey-based
Taguchi method for multi-response optimization of bead
geometry in submerged arc bead-on-plate welding
process. Jackson and Shrubsall 2 performed optimization,
neural networks and regression analysis in submerged
arc welding process. Xie Yan-Min and et al. 3 used grey
relational analysis for optimizing the square hole
flanging process parameters with considerations of the
multiple response characteristics. Lin and Lin 4 studied
on the use of the grey-fuzzy logic based on orthogonal
array for optimizing the electrical discharge machining
process with multi-response characteristics. Murugan et
al. 5, developed mathematical models using a five-level
factorial technique to predict the weld bead geometry for
depositing 316L stainless steel onto structural steel
IS2062. Gunaraj and Murugan 6 determined the main and
interaction effects of process control variables on
important bead geometry parameters including bead
volume quantitatively and represented the results
graphically in submerged arc welding process.
The Taguchi method is very popular for solving
optimization problems in the field of production
engineering, Yang et al. 7, Rowlands et al. 8. The method
utilizes a well-balanced experimental design (allows a
limited number of experimental runs) called orthogonal
array design, and signal-to-noise ratio (S/N ratio), which
serve the objective function to be optimized (maximized)
within experimental domain. However, traditional Tagu-
chi method cannot solve multi-objective optimization
problem. To overcome this, the Taguchi method coupled
with Grey relational analysis has a wide area of
application in manufacturing processes. This approach
can solve multi-response optimization problem simulta-
neously 1,9.
Planning the experiments through the Taguchi
orthogonal array has been used quite successfully in
process optimization by Chen and Chen 10, Fung and
Kang 11, Tang et al. 12, Vijian and Arunachalam 13, Yang 14
as well as Zhang et al. 15. Therefore, this study applied a
Taguchi L16(44) orthogonal array to plan the experiments
on TIG welding process. Four controlling factors includ-
ing welding speed (V), welding current (I), shielding gas
flow rate (F) and gap distance (G) with four levels for
each factor were selected. The Grey relational analysis is
then applied to examine how the welding process factors
influence the bead geometry; bead width (BW), bead
height (BH), penetration (P), area of penetration (AP)
and heat affected zone (HAZ), as well as tensile load
(TL). An optimal parameter combination was then
obtained. Through analyzing the Grey relational grade
matrix, the most influential factors for individual quality
targets of TIG welding process can be identified.
Additionally, the analysis of variance (ANOVA) was
also utilized to examine the most significant factors for
the bead geometry in TIG welding process.
2 GREY RELATIONAL ANALYSIS
2.1 Data Preprocessing
In Grey relational analysis, experimental data i.e.,
measured features of quality characteristics are first
normalized ranging from zero to one. This process is
known as Grey relational generation. Next, based on
normalized experimental data, Grey relational coefficient
is calculated to represent the correlation between the
desired and actual experimental data. Then overall Grey
relational grade is determined by averaging the Grey
relational coefficient corresponding to selected responses
1. The overall performance characteristic of the multiple
response process depends on the calculated Grey
relational grade. This approach converts a multiple
response process optimization problem into a single
response optimization situation with the objective
function is overall Grey relational grade. The optimal
parametric combination is then evaluated which would
result highest Grey relational grade. The optimal factor
setting for maximizing overall Grey relational grade can
be performed by Taguchi method 1.
In Grey relational generation, the normalized HAZ,
BW and BH corresponding to the smaller-the-better (SB)
criterion which can be expressed as:
x k
y k y k
y k y ki
i i
i i
( )
max ( ) ( )
max ( ) min ( )
=
−
−
(1)
TL, P and AP should follow the larger-the-better
(LB) criterion, which can be expressed as:
x k
y k y k
y k y ki
i i
i i
( )
( ) min ( )
max ( ) min ( )
=
−
−
(2)
where xi(k) is the value after the Grey relational gene-
ration, min yi(k) is the smallest value of yi(k) for the kth
response, and max yi(k) is the largest value of yi(k) for
the kth response. An ideal sequence is x0(k) (k = 1, 2, 3,
..., 16) for the responses. The definition of Grey rela-
tional grade in the course of Grey relational analysis is
to reveal the degree of relation between the 16
sequences  x0(k) and xi(k), i = 1, 2, 3, ...,16. The Grey
relational coefficient ξi(k) can be calculated as:
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
144 Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149
ξ
∆ ψ∆
∆ ψ∆i i
k
k
( )
( )
min max
max
=
−
+0
(3)
where ∆0i = IIx0(k) – xi(k)II difference of the absolute
value x0(k) and xi(k); ψ is the distinguishing coefficient
0 ≤ ψ ≤ 1; ∆min = ∀jmin ∈ i∀kmin IIx0(k) – xi(k)II = the
smallest value of ∆0i; and ∆max = ∀jmax ∈ i∀kmax IIx0(k)
– xi(k)II = largest value of ∆0i. After averaging the Grey
relational coefficients, the Grey relational grade γi can
be computed as:
γ ξi i
k
n
n
k=
=
∑1
1
( ) (4)
where n is the number of process responses. The higher
value of Grey relational grade corresponds to intense
relational degree between the reference sequence x0(k)
and the given sequence xi(k). The reference sequence
x0(k) represents the best process sequence; therefore,
higher Grey relational grade means that the corres-
ponding parameter combination is closer to the optimal.
The mean response for the Grey relational grade with its
grand mean and the main effect plot of Grey relational
grade are very important because optimal process
condition can be evaluated from this plot 1.
3 EXPERIMENTAL PROCEDURE AND TEST
RESULTS
3.1 Experimental Details
The experiments were conducted according to the
design matrix at random order to avoid systematic errors
infiltrating the system. Weld beads were laid on the joint
to join 1.2 mm AISI 304 thin stainless steel plate with
the dimensions of (25 × 240) mm. The chemical compo-
sition of the workpiece material is given in Table 1.
Table 1: Composition (%) of AISI 304 steel workpiece
Tabela 1: Sestava (%) vzorca jekla AISI 304
C Mn P S Si Cr Mo Ni Cu
0.08 2.00 0.04 0.030 1.00 19 0.20 10.5 0.02
Fronius Variostar WTU 305 type arc welding
machine was used in the experiments. The specimens
were joined using a single pass butt welding with AWS
A 5.12-80 EW Th-2 thoriated tungsten electrode and
pure argon (99.99 %) as shielding gas.
The experimental set up was designed and con-
structed to control the linear movement of the torch
along the weld pad center line. The welded joints were
sectioned to produce specimens for examining the
quality parameters (BW, BH, P and AP) of weld bead
shape in the welded specimens. These specimens were
prepared by the usual metallurgical polishing methods
and etched with Marble’s etching reagent (CuSO4 + HCl
+ H2SO4). Macrographs were then taken for each cross
section using stereo microscope with 50X lens. In macro
examinations of the specimens, Motic stereo microscope
with image capture device mounted on top of the lens
section of the microscope was used.
The weld bead profile was outlined by using
Image-pro Plus 4.5 and NIH ImageJ software. The
spatial calibrations were made on the macrographs
before the measurement. The line drawings of the bead
profiles were then used to take measurements on UW,
UH, P, AP and HAZ.
Tensile load values were recorded from tensile
testing of the specimens prepared in accordance with the
EN 895 Standard. Tensile test specimens were taken
from the weld bead according to the transverse tensile
test method. Tensile test specimens were prepared in
such a way that the weld zones were centered in the gage
length. At the same time, heat affected zone was placed
in the gage length perpendicular to the weld.
3.2 Process Parameters and Test Results
In full factorial design, the number of experimental
runs exponentially increases as the number of factors as
well as their level increases. This results huge experi-
mentation cost and considerable time 1. So, in order to
compromise these two adverse factors and to search the
optimal process condition through a limited number of
experimental runs Taguchi’s L16(44) orthogonal array
consisting of 16 sets of data has been selected to
optimize the multiple performance characteristics of TIG
welding bead geometry.
Experiments have been conducted with the process
parameters given in Table 2, to obtain butt welding on
AISI 304 1.2 mm thickness stainless steel plate with 25
mm × 240 mm dimensions by TIG welding process.
Table 2: Process parameters and their limits
Tabela 2: Parametri procesov in njihove limite
Parameters Notation Unit Levels of factors
1 2 3 4
Travel speed V mm/s 1.06* 1.99 2.31 3.55
Current I A 40* 55 70 85
Gas flow rate F L/min 8* 10 12 14
Gap distance G mm 1.5* 2 2.5 3
*Initial factor settings
Table 3 shows the selected design matrix based on
Taguchi L16(44) orthogonal array consisting of 16 sets of
coded conditions and the experimental results for the
responses of TL, HAZ, BW, BH, P and AP. All these
data have been utilized for analysis and evaluation of
optimal parameter combination required to achieve
desired quality weld in terms of bead geometry within
the experimental domain.
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149 145
4 PARAMETRIC OPTIMIZATION OF TIG
WELDING PROCESS
4.1 Evaluation of Optimal Process Condition
First, by using Eqs. (1) and (2), experimental data
have been normalized to obtain Grey relational gene-
ration. The normalized data and ∆0i for each of the
responses of bead geometry as well as TL and HAZ have
been furnished in Table 4 and Table 5 respectively. For
TL, P and AP larger-the-better (LB) and for HAZ, BW,
BH smaller-the-better (SB) criterion has been selected.
Table 4: Grey relational generation of each performance characte-
ristics
Tabela 4: Generiranje Greyjeve odvisnosti za vsako zna~ilno perfo-
manco
Run no
TL P AP HAZ BW BH
Larger-the-better Smaller-the-better
Ideal
sequence 1 1 1 1 1 1
1 0.620 0.509 0.523 0.530 0.470 0.544
2 0.746 0.730 0.706 0.338 0.287 0.312
3 0.797 0.593 0.300 0.161 0.130 0.131
4 0.718 1.000 1.000 0.000 0.000 0.000
5 0.662 0.417 0.378 0.787 0.568 0.908
6 0.803 0.586 0.505 0.603 0.388 0.886
7 0.846 0.785 0.568 0.512 0.387 0.663
8 0.958 0.911 0.735 0.344 0.173 0.552
9 0.577 0.286 0.295 0.855 0.642 0.918
10 0.676 0.529 0.391 0.714 0.561 0.852
11 0.930 0.741 0.550 0.550 0.337 0.848
12 1.000 0.862 0.641 0.437 0.236 0.679
13 0.000 0.000 0.000 1.000 1.000 1.000
14 0.211 0.314 0.123 0.870 0.847 0.997
15 0.291 0.365 0.225 0.702 0.684 0.947
16 0.338 0.593 0.300 0.590 0.599 0.986
Table 5: Evaluation of ∆0i for each of the responses
Tabela 5: Ocena ∆0i za vsak odgovor
Run no TL P AP HAZ BW BH
Ideal
sequence 1 1 1 1 1 1
1 0.380 0.491 0.477 0.470 0.530 0.456
2 0.254 0.270 0.294 0.662 0.713 0.688
3 0.203 0.407 0.700 0.839 0.870 0.869
4 0.282 0.000 0.000 1.000 1.000 1.000
5 0.338 0.583 0.622 0.213 0.432 0.092
6 0.197 0.414 0.495 0.397 0.612 0.114
7 0.154 0.215 0.432 0.488 0.613 0.337
8 0.042 0.089 0.265 0.656 0.827 0.448
9 0.423 0.714 0.705 0.145 0.358 0.082
10 0.324 0.471 0.609 0.286 0.439 0.148
11 0.070 0.259 0.450 0.450 0.663 0.152
12 0.000 0.138 0.359 0.563 0.764 0.321
13 1.000 1.000 1.000 0.000 0.000 0.000
14 0.789 0.686 0.877 0.130 0.153 0.003
15 0.709 0.635 0.775 0.298 0.316 0.053
16 0.662 0.407 0.700 0.410 0.401 0.014
Table 6 shows the calculated Grey relational coeffi-
cients (with ψ = 0 of each performance characteristic
using Eq. (3).
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
146 Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149
Table 3: Orthogonal array L16(44) of the experimental runs and results
Tabela 3: Ortogonalna porazdelitev L16 (44) eksperimentalnih varkov in rezultatov
Run no
Process parameters Experimental results
V I F G
TL
(N)
HAZ
(mm)
BW
(mm)
BH
(mm)
P
(mm)
AP
(mm2)
1 1 1 1 1 11429 5.255 7.383 0.185 1.098 7.809
2 1 2 2 2 11870 6.204 8.391 0.230 1.190 8.899
3 1 3 3 3 12047 7.077 9.253 0.265 1.258 9.848
4 1 4 4 4 11772 7.874 9.972 0.291 1.302 10.654
5 2 1 2 3 11576 3.985 6.840 0.114 1.060 6.948
6 2 2 1 4 12066 4.895 7.835 0.118 1.130 7.704
7 2 3 4 1 12213 5.344 7.840 0.162 1.213 8.078
8 2 4 3 2 12606 6.172 9.017 0.183 1.265 9.072
9 3 1 3 4 11282 3.649 6.432 0.112 1.006 6.453
10 3 2 4 3 11625 4.343 6.880 0.125 1.107 7.022
11 3 3 1 2 12508 5.157 8.115 0.126 1.195 7.970
12 3 4 2 1 12753 5.714 8.670 0.159 1.245 8.514
13 4 1 4 2 9270 2.931 4.459 0.096 0.887 4.695
14 4 2 3 1 10006 3.573 5.304 0.097 1.017 5.429
15 4 3 2 4 10281 4.405 6.201 0.107 1.039 6.033
16 4 4 1 3 10448 4.955 6.670 0.099 1.133 6.481
Table 6: Grey relational coefficient of each performance characte-
ristics (ψ = 0)
Tabela 6: Koeficient Greyjeve relacije za karakteristike vsake per-
formance(ψ = 0)
Run no TL P AP HAZ BW BH
Ideal
sequence 1 1 1 1 1 1
1 0.568 0.504 0.512 0.515 0.485 0.523
2 0.664 0.649 0.629 0.430 0.412 0.421
3 0.711 0.551 0.417 0.373 0.365 0.365
4 0.640 1.000 1.000 0.333 0.333 0.333
5 0.597 0.462 0.446 0.701 0.537 0.845
6 0.717 0.547 0.503 0.557 0.449 0.815
7 0.765 0.699 0.536 0.506 0.449 0.597
8 0.922 0.848 0.653 0.433 0.377 0.528
9 0.542 0.412 0.415 0.775 0.583 0.859
10 0.607 0.515 0.451 0.636 0.532 0.772
11 0.877 0.659 0.526 0.526 0.430 0.767
12 1.000 0.784 0.582 0.470 0.396 0.609
13 0.333 0.333 0.333 1.000 1.000 1.000
14 0.388 0.422 0.363 0.794 0.765 0.993
15 0.413 0.441 0.392 0.626 0.613 0.904
16 0.430 0.551 0.417 0.550 0.555 0.974
The Grey relational coefficients, given in Table 7,
for each response have been accumulated by using Eq.
(4) to evaluate Grey relational grade, which is the overall
representative of all the features of weld quality. Thus,
the multi-criteria optimization problem has been
transformed into a single equivalent objective function
optimization problem using the combination of Taguchi
approach and Grey relational analyses. Higher is the
value of Grey relational grade, the corresponding factor
combination is said to be close to the optimal 1.
Table 7: Grey relational grade
Tabela 7: Razred Greyjeve relacije
Run no Grey relationalgrade Rank
1 0.5180 15
2 0.5342 14
3 0.4637 16
4 0.6066 6
5 0.5977 8
6 0.5980 7
7 0.5922 10
8 0.6268 4
9 0.5975 9
10 0.5855 11
11 0.6307 3
12 0.6402 2
13 0.6667 1
14 0.6208 5
15 0.5650 13
16 0.5794 12
Table 8 shows the S/N ratio based on the larger-
the-better criterion for overall Grey relational grade
calculated by using Eq. (5).
S N
n y ii
n
/ lg= −
⎡
⎣
⎢
⎤
⎦
⎥
=
∑10 1 12
1
(5)
where n is the number of measurements, and yi is the
measured characteristic value.
Table 8: S/N ratio for overall Grey relational grade
Tabela 8: Razmerje S/N za splo{no Grayjevo odvisnost
Run no S/N
1 –5.71
2 –5.45
3 –6.68
4 –4.34
5 –4.47
6 –4.47
7 –4.55
8 –4.06
9 –4.47
10 –4.65
11 –4.00
12 –3.87
13 –3.52
14 –4.14
15 –4.96
16 –4.75
Table 9: Response table for the mean Grey relational grade
Tabela 9: Tabela odgovorov za povpre~je odvisnosti Greyjevega
razreda
Factors
Grey relational grade
Level 1 Level 2 Level 3 Level 4 Max-Min
V 0.53 0.60 0.61 0.61 0.08
I 0.59 0.58 0.56 0.61 0.05
F 0.58 0.58 0.58 0.61 0.03
G 0.59 0.61 0.56 0.59 0.05
Total mean Grey relational grade is 0.59.
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149 147
Figure 2: S/N ratio plot for the overall Grey relational grade
Slika 2: S/N odvisnost za splo{no Greyjevo relacijo
Graphical representation of S/N ratio for overall Grey
relational grade is shown in Figure 2. The dashed line is
the value of the total mean of the S/N ratio.
As indicated in Figure 2, the optimal condition for
1.2 mm AISI 304 stainless steel becomes V3I4F4G2.
Table 9 shows the mean Grey relational grade ratio for
each level of the process parameters.
4.2 Analysis of Variance (ANOVA)
The purpose of the analysis of variance (ANOVA) is
to investigate which welding parameters significantly
affect the performance characteristic. This is accom-
plished by separating the total variability of the grey
relational grades, which is measured by the sum of the
squared deviations from the total mean of the grey
relational grade, into contributions by each welding
parameters and the error. Thus;
SST = SSF + SSe (6)
where
SS j
j
p
T m= γ γ( )−
=
∑
1
2 (7)
and
SST Total sum of squared deviations about the mean
γj Mean response for jth experiment
γm Grand mean of the response
p Number of experiments in the orthogonal array
SSF Sum of squared deviations due to each factor
SSe Sum of squared deviations due to error
In addition, the F test was used to determine which
welding parameters have a significant effect on the
performance characteristic. Usually, the change of the
welding parameter has a significant effect on the
performance characteristic when the F value is large.
ANOVA for overall Grey relational grade is shown in
Table 10.
Table 10: ANOVA results
Tabela 10: ANOVA-rezultati
Parameter Degree ofFreedom
Sum of
Square
Mean
Square F
Contribu-
tion (%)
V 3 0.0180 0.006109 4.42 52.41
I 3 0.00529 0.001765 1.28 15.40
F 3 0.00312 0.001042 0.75 9.09
G 3 0.00691 0.002305 1.67 20.12
Error 3 0.00102 0.000346 - 2.97
Total 15 0.03434 100
Result of the ANOVA indicates that the welding
speed (52.41 % contribution) is the most effective
parameter on the responses under the multi criteria
optimization (higher tensile load, penetration, area of
penetration and lower heat affected zone, bead width,
bead height). The percent contributions of other para-
meters are gap distance (20.12 %), current (15.40 %) and
shielding gas flow rate (9.09 %).
4.3 Confirmation Test
After evaluating the optimal parameter settings, the
next step is to predict and verify the enhancement of
quality characteristics using the optimal parametric
combination. The estimated Grey relational grade  γ
using the optimal level of the design parameters can be
calculated as:
  ( )γ = γ γ γm m+ −
=
∑ i
i
o
1
(8)
where γm is the total mean Grey relational grade, γ i is
the mean Grey relational grade at the optimal level, and
o is the number of the main design parameters that
affect the quality characteristics. Table 11 shows the
comparison of the predicted bead geometry parameters
as well as TL and HAZ with that of actual by using the
optimal TIG welding conditions; good agreement
between the actual and predicted results have been
observed (improvement in overall Grey relational
grade).
Table 11: Results of confirmation test
Tabela 11: Rezultati potrditvenega poskusa
Initial factor
settings
Optimal process condition
Prediction Experiment
Factor levels V1I1F1G1 V3I4F4G2 V3I4F4G2
TL 1165 1285
P 1.098 1.226
AP 7.809 8.518
HAZ 5.255 5.747
BW 7.383 8.569
BH 0.185 0.152
S/N ratio of overall
Grey relational grade −5.71 −3.14 −1.60
Overall Grey
relational grade 0.52 0.69 0.83
Improvement in Grey relational grade is 0.31
In Taguchi method, the only performance feature is
the overall Grey relational grade; and the aim should be
to search a parameter setting that can achieve highest
overall Grey relational grade. The Grey relational grade
is the representative of all individual performance
characteristics. In this study, objective functions have
been selected in relation to parameters of bead geometry
as well as TL and HAZ; and all the responses have been
given equal weight age. The results showed that using
optimal parameter setting (V3I4F4G2) caused higher P,
AP and TL.
5 CONCLUSION
Taguchi method is a very effective tool for process
optimization under limited number of experimental runs.
Essential requirements for all types of welding are
deeper penetration with higher tensile load and lower
heat affected zone, bead height and bead width for
reducing weld metal consumption. This study has
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
148 Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149
concentrated on the application of Taguchi method
coupled with Grey relation analysis for solving multi
criteria optimization problem in the field of tungsten
inert gas welding process. Experimental results have
shown that the tensile load, heat affected zone and
penetration, area of penetration, heat affected zone, bead
width and bead height of the weld bead in the TIG
welding of stainless steel are greatly improved by using
Grey relation analysis in combination with Taguchi
method.
6 REFERENCES
1 S. Datta, A. Bandyopadhyay, P. K. Pal, Int J Adv Manuf Technol.,
39 (2008), 1136–1143
2 C. E. Jackson, A. E. Shrubsall, Weld J., 32 (1953), 172–178
3 Y. M. Xie et al., Journal of Zhejiang University Science A, 8 (2007),
805–811
4 J. L. Lin, C. L. Lin, J Mater Process Technol., 160 (2005), 9–14
5 N. Murugan, R. S. Paramar, S. K. Sud, J Mater Process Technol., 37
(1993), 767–780
6 V. Gunaraj, N. Murugan, Weld J., (2000), 286–294
7 W. H. Yang, Y. S. Tarng, J Mater Process Technol., 84 (1998),
122–129
8 H. Rowlands, J. Antony, G. Knowles, The TQM Magazine, 12
(2000), 78–83
9 S-H. Lim, C-M. Lee, W.J. Chung, International Journal of Precision
Engineering and Manufacturing, 7 (2006), 18–23
10 D. C. Chen, C. F. Chen, J. Mater Process Technol., 190 (2007),
130–137
11 C. P. Fung, P.C. Kang, J. Mater. Process Technol., 170 (2005),
602–610
12 S. H. Tang, V. J. Tan, S.M. Sapuan, S. Sulaiman, N. Ismail, R.
Samin, J Mater Process Technol., 182 (2007), 418–426
13 P. Vijian, V.P. Arunachalam, J Mater Process Technol., 180 (2006),
161–166
14 L. J. Yang, J Mater Process Technol., 185 (2007), 113–119
15 J. Z. Zhang, J. C. Chen, E. D. Kirby, J Mater Process Technol., 184
(2007), 233–239
U. ESME ET AL.: OPTIMIZATION OF WELD BEAD GEOMETRY IN TIG WELDING PROCESS ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 143–149 149

J. TESA : PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
PRIMARY ETALONNAGE OF NEGATIVE GAUGE
PRESSURES USING PRESSURE BALANCES AT THE
CZECH METROLOGY INSTITUTE
PRIMARNE KALIBRACIJSKE METODE ZA NEGATIVNI
RELATIVNI TLAK S TLA^NIMI TEHTNICAMI NA ^E[KEM
IN[TITUTU ZA METROLOGIJO
Ji í Tesa , Zdenk Krají~ek, Dominik Pra`ák, Franti{ek Stank
Czech Metrology Institute, Okru`ní 31, 638 00 Brno, Czech Republic
dprazak cmi.cz
Prejem rokopisa – received: 2008-04-14; sprejem za objavo – accepted for publication: 2008-08-23
This paper summarizes the methods of using the piston pressure balance principle for the etalonnage of negative gauge
pressures, i.e., the classical piston pressure balance with an inversed piston cylinder or working under an evacuated bell jar,
digital non-rotating piston pressure balances and an absolute pressure divider. It describes the principles of their use (focusing
on their utilization at the Czech Metrology Institute), the methods of traceability and analyzes their achievable uncertainties.
Keywords: negative gauge pressure, pressure balance, traceability
V ~lanku so predstavljene razli~ne metode z batnimi tla~nimi tehtnicami za umerjanje merilnikov negativnega tlaka, tj. klasi~ne
batne tla~ne tehtnice z obrnjenim sklopom bat-valj ali z evakuirano zvonasto posodo, digitalne batne tla~ne tehtnice z nevrte~im
se batom in z delilnikom absolutnega tlaka. Opisan je princip dela s poudarkom na njihovi rabi na ^e{kem in{titutu za
metrologijo, na~ini sledljivosti in analiza dosegljivih merilnih negotovosti.
Klju~ne besede: merilnik negativnega pritiska, tla~na tehtnica, sledljivost
1 INTRODUCTION
The etalonnage of negative gauge pressures is a very
much neglected branch of primary metrology (although
very important in practice). This peripheral position is
documented by the fact that no Key Comparison has
ever been performed in this range. The only such
comparisons known to the authors were organized at the
national level, e.g. 1,2. Simultaneously, the use of piston
pressure balances as the primary standards of negative
gauge pressure leads to very interesting constructions.
The Laboratory of Pressure Metrology of the Czech
Metrology Institute intends to ensure this pressure range
via piston pressure balances.
In this paper the values of the negative gauge
pressure Png will be positive and all the uncertainties will
be unexpanded (k = 1).
2 CLASSICAL PRESSURE BALANCES
Pressure balances can also be utilized for the primary
etalonnage of negative low pressures 3. There are three
basic set-ups enabling this. Firstly, an inverse piston-
cylinder design; secondly, the utilization of an eva-
cuated bell jar; and thirdly, an absolute pressure
balance with a reference pressure monitor, enabling
the calculation of the negative gauge pressure from their
pressure difference (not discussed further in this article).
The first solution is based on a piston cylinder that is
vertically reversed (Figure 1). It is held in its floating
position by the atmospheric pressure Pa acting upwards,
balancing the residual pressure P acting downwards on
the piston together with the force F that is the sum of the
gravitational forces acting on the piston and the loaded
masses. So, the generated negative gauge pressure is
given by the following equation:
P P P
M g
A p T
ng = − =
−
⎛
⎝
⎜
⎞
⎠
⎟
a
a
M
ef
1
ρ
ρ
( , )
(1)
where:
Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156 151
UDK 006.91:543.5 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)151(2009)
Figure 1: The inverse piston-cylinder design
Slika 1: Prikaz delovanja obrnjenega sklopa bat-valj
Aef (p,T) is the effective area of the piston cylinder that is
generally dependent on the acting pressure p and its
temperature T,
M is the total mass of the piston and the masses loaded
on it,
 M is the average density of the piston and the masses,
 a is the density of the ambient air,
g is the local acceleration due to gravity.
In practice, the piston cylinder is mounted on tubing,
enabling it to be calibrated in the gauge mode in the
normal position and then turned down and utilized for
negative gauge pressures. This solution, however, brings
about certain restrictions. For the range up to 100 kPa
generated by a piston cylinder with a nominal effective
area of 10 cm2 (widely used in this range of gauge
pressures) a total mass of 10 kg is needed, which
introduces design complications. This is solved by
inverse piston cylinders with a nominal effective area of
1 cm2, which, on the other hand, reduce the sensitivity.
The second solution is based on a classical
piston-cylinder design working under a bell jar (Figure
2). This is normally used for working in the absolute
mode when the space under the bell jar is evacuated (and
the residual pressure is measured with a suitable vacuum
meter). In this case, however, the ambient atmospheric
pressure is applied under the piston, while under the bell
jar such an absolute pressure is set in order to balance
(together with the pressure defined by the piston
cylinder) the atmospheric pressure. So the generated
negative gauge pressure is given as:
P P P
M g
A p T
ng = − =
−
⎛
⎝
⎜
⎞
⎠
⎟
a
g
M
ef
1
ρ
ρ
( , )
(2)
where  g is the density of the gas pressure medium
under the bell.
The second solution is very useful because it is
possible to utilize a piston manometer in a reversed
set-up of its absolute mode, only with the possibility to
employ piston cylinders with large effective areas.
However, there are also certain disadvantages. The space
beneath the vacuum bell jar has a large volume that
requires a very skilled operator. Another task is to
accurately determine the density of the gas under the bell
jar. Furthermore, for measuring the low negative gauge
pressures, it is necessary to evacuate the bell jar first in
order to reach an efficient sealing level, and only then is
it possible to return to low negative gauge pressures.
The laboratory utilizes the classical DHI PG7601
piston gauge 4,5 with a piston cylinder 10 cm2 in the
mode described above. It is also equipped with an
automated mass-handling system and automated
pressure regulators. This system was developed for work
in the absolute mode, but the laboratory also utilizes it in
the negative gauge mode with very satisfactory results.
The simple free-deformation piston cylinders are
utilized in this mode. Dadson and Sutton’s model of an
effective area of a piston cylinder also depends on the
pressure distribution along the gap and the viscosity and
density of the pressure medium used as functions of the
pressure. It would therefore be necessary to evaluate the
effective area not only for various working modes
(absolute, gauge, negative gauge) but also for various
generated pressure points. Numerical and experimental
studies have proven that these changes never exceed
0.0002 %, which is negligible.
The calculation of the uncertainty of the negative
gauge pressure generated by the above-mentioned piston
cylinder is based on equation (2). The effective area
value and its uncertainty uAef were obtained from
geometrical measurements using Dadson and Sutton’s
model 3 and were confirmed by the hydrostatic cross-
floating method, both in absolute and gauge modes. The
J. TESA: PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
152 Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156
Figure 3: The influence of the uncertainties of all the input quantities
on the total uncertainty of the negative gauge pressure generated by a
classical pressure balance with an evacuated bell jar
Slika 3: Vpliv negotovosti vseh vstopnih veli~in na skupno merilno
negotovost negativnega pritiska, ki ga generiramo s klasi~no batno
tla~no tehtnico z evakuirano zvonasto posodo
Figure 2: Principle of utilizing a bell to measure negative gauge
pressures
Slika 2: Prikaz rabe zvonaste posode za merjenje negativnega pritiska
pressure-deformation coefficient of the piston cylinder
can be omitted in this pressure range, but the influence
of its thermal expansion must be taken into account. The
second most important contribution to the total uncer-
tainty is the uncertainty of the mass uM. The influence of
the uncertainty of the density of the gas under the
vacuum bell jar u g has been a source of relatively high
uncertainty and so pure nitrogen has been introduced as
the pressure medium. Its density values are tabulated in 6
with a precision 0.01 %. The influence of the head-
pressure uncertainty uPh depends on the difference bet-
ween the reference levels, which is usually very small,
but an exaggerated value of 25 cm was used here. The
correlations between uρg and the uncertainty of the
acceleration due to gravity ug that also appear in the
calculation of uPh were neglected. The remaining sources
of uncertainty are the uncertainties of the mass density
uρM and the sensitivity us. The graph in Figure 3 shows
the influence of the uncertainties of all the input
quantities on the total uncertainty of the generated
negative gauge pressure uPng and compares it with the
internationally accepted measurement capability of the
Czech Republic uCMC = 0.5 Pa + 0.0006 % of the
measured value in the range from 5 kPa to 100 kPa 7.
3 DIGITAL PRESSURE BALANCES
The measurement of small pressure differences and
low absolute pressures using a classical piston mano-
meter with a rotating piston has two basic limitations.
First, it is only possible to measure pressures that are
high enough to balance the mass of the piston (a few
kilopascals). Second, the periodic pressure fluctuations
caused by rotation of the piston that have amplitudes of
the order of tenths of a pascal are an important source of
uncertainty in low-pressure measurements. A solution
based on the connection of a non-rotating piston to an
electronic dynamometer can measure even very small
changes of the state of equilibrium of the piston. It also
offers the possibility to use a larger effective area of the
piston cylinder.
The laboratory uses an improved, commercially
available Furness Controls FRS 4 HR standard able to
work in differential, gauge and (reversing the outputs)
also negative gauge modes 8. The nominal area of the
piston is 100 cm2, and the width of the gap is approxi-
mately 30 µm. The electronic balance can be calibrated
using an external mass before every measurement. The
friction between the piston and the cylinder is prevented
by a lever mechanism (Figure 4), allowing axial move-
ment of the piston without any contact with the surface
of the cylinder.
The flow of dry air enters the upper part of the
cylinder during the measurement. This air then flows
through the gap and exits through the outputs in the base.
The system was highly sensitive to small instabilities of
atmospheric pressure. It was therefore decided to use
mass-flow controllers in the input and output instead of
the flow-controlling unit in order to insulate the
instrument’s ports from the ambient and to stabilize and
control the gas flow (i.e., the pressure difference).
This instrument has made it possible to ensure the
etalonnage of the negative gauge pressure in the range
from 1 Pa to 3.2 kPa. The internationally accepted
measurement capability of the Czech Republic uCMC =
0.01 Pa + 0.004 % of the measured value 7.
Figure 5 introduces the DHI FPG 8601 manometer
9–12. It is based on a non-rotating tungsten-carbide piston
10 cm2 placed in a tungsten-carbide cylinder, where the
gap has a 1-to-6 µm conical profile. The piston is centred
by means of a lubricating gas with a pressure that is
about 40 kPa higher than the reference pressure and is
balanced by a mass comparator (electronic balance). The
pressure and the gas flow are regulated by a control unit
equipped with two flowmeters (for coarse and fine
adjustments), and also a PC. The attainable uncertainty
J. TESA: PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156 153
Figure 5: Digital non-rotating piston manometer DHI FPG 8601
Slika 5: Digitalni nerotirajo~i batni manometer DHI FPG 8601
Figure 4: Digital non-rotating piston manometer FRS 4 HR
Slika 4: Digitalni nerotirajo~i batni manometer FRS 4 HR
in the negative gauge mode in the range from 1 Pa to 15
kPa is uPng = 0.01 Pa + 0.0014 % of the measured value.
This instrument was evaluated using primary
methods and compared with a classical pressure balance,
both in gauge (evaluating number En up to 0.4) and in
absolute (En up to 0.3) modes. Both the digital non-
rotating piston pressure balances were also compared
mutually in the gauge mode (En up to 0.6).
4 ABSOLUTE PRESSURE DIVIDER
The principle of this standard is the same as in the
classical pressure divider 13. It is based on three con-
centric pistons (A, B and C). The effective area of the
piston B is 101-times larger than that of the pistons A
and C (Figure 6). The pressure generated by an
oil-piston manometer is connected to the base of piston
C. Due to the ratio of the effective areas of the piston
cylinders of the divider, the increase in the pressure
above piston B is nominally 100-times lower than the
increase in the hydraulic pressure under piston C. There
is a vacuum pump connected via an insulating valve to
the lower chamber beneath piston B. The device under
test is also connected to the lower chamber.
The DH-Budenberg 1600 absolute pressure divider
can be used to generate absolute, gauge, negative gauge
and differential (at line pressures other than atmospheric)
pressures. The work in the negative gauge mode is as
follows. First, the spaces under and over the central
piston B are opened to the atmosphere. Such masses
must be put on the oil-piston manometer in order to
generate an appropriate pressure with a magnitude of
approximately 1 MPa, acting on the base of piston C and
compensating for the gravity of the rotating piston
system. After reaching an equilibrium state of the
oil-piston standard and the divider, the valve connecting
the lower chamber to the atmosphere can be closed.
Now, an additional mass must be placed onto the oil-
piston manometer to generate an increase in the hydrau-
lic pressure 100-times higher than is the demanded
negative gauge pressure value. This will increase the
pressure upwards acting on the base of piston C, i.e., on
the piston system of the divider, and raise this system in
the upper stroke. Then the lower chamber is evacuated
until the demanded negative gauge pressure is reached.
Consequently, the vacuum pump is disconnected by the
valve and the hydraulic pressure is trimmed in order to
ensure the middle floating position of the piston system
of the divider.
Since the effective area of piston B is nominally
101-times larger than that of pistons A and C, the ratio of
the hydraulic and the differential pressure (dividing
ratio) is nominally:
R
p
p
A A
AD
o
ng
B C
C
= =
−
= − =
∆ 101 1
1
100 (3)
where:
 po is the change in hydraulic pressure
AC is the effective area of piston C
AB is the effective area of piston B
The hydraulic pressure at the time of the initial
equilibrium is defined as:
p
m c
A p T
h gmo
a0
o o o
o=
−⎛
⎝
⎜
⎞
⎠
⎟ +
+
1
0
ρ
ρ
σ
ρ
( , )
(4)
where:
m is the total applied mass by initial equilibrium
m is the mean density of the piston and loaded masses
Ao is the effective area of the hydraulic pressure piston
cylinder
To0 is the temperature of the hydraulic pressure piston
cylinder during initial equilibrium
a0 is the density of ambient atmosphere during initial
equilibrium
c is the piston circumference at its exit from the oil
 is the surface-tension coefficient of the oil
o is the density of the oil
h is the height of the piston base above the reference
level
The change in the hydraulic pressure (caused by the
additional mass mp) can be written as:
∆
∆
p
m m g c
A p T p
m
o
p
m m=
+ −⎛
⎝
⎜
⎞
⎠
⎟ +
+
−
−⎛( )
( , )
1 1
0
ρ
ρ
σ
ρ
ρ
a
o o o o
a0
⎝
⎜
⎞
⎠
⎟ +g c
A p T
σ
o o o( , )0
(5)
where a is the density of the ambient atmosphere and
To is the temperature upon generating the differential
pressure.
The generated negative gauge pressure is:
J. TESA : PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
154 Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156
Figure 6: Principle of the absolute pressure divider
Slika 6: Prikaz delovanja absolutnega delilnika tlaka
p
R
m m g c
A p T p
mp
m
ng
D
a
o o o o
a0
=
+ −
⎛
⎝
⎜
⎞
⎠
⎟ +
+
−
−
1
1 1
0
( )
( , )
ρ
ρ
σ
ρ
∆
ρ
σ
m
g c
A p T
⎛
⎝
⎜
⎞
⎠
⎟ +
⎡
⎣
⎢
⎢
⎢
⎢
⎤
⎦
⎥
⎥
⎥
⎥o o o
( , )0
(6)
The following notation that simplifies the previous
equation is used for the purpose of analysing the un-
certainty:
[ ]P
R
P Cng
D
o= −
1
(7)
The graph in Figure 7 shows the influence of the
uncertainties of all the input quantities on the total
uncertainty of the generated negative gauge pressure
uPng. The major influences come from the uncertainty of
the dividing ratio uRD and the component uRo, followed
by the sensitivity us and the component uC. The influence
of the head-pressure uncertainty uPh depends on the
difference between the reference levels (again, an
exaggerated value of 25 cm was used). The correlations
between uPo, uC and uPh were neglected. The resultant
uncertainty (for negative gauge pressures higher than
1 kPa) can be approximated by uPng = 0.15 Pa + 0.0027
% of the measured value.
5 TRACEABILITY OF THE ABSOLUTE
PRESSURE DIVIDER
The traceability of the pressure divider is composed
of two items. First, it is the traceability of the hydraulic
pressure balance, which is trivial. Second, it is the
traceability of the dividing ratio that was performed with
the national standard in the gauge pressure (up to
450 kPa, see Figure 8). The dividing ratio is, in general,
dependent on the pressure, but this can be neglected in
the negative gauge mode. The obtained value RD =
(99.997 ± 0.005) is in agreement with the value obtained
three years before RD = (99.998 ± 0.005).
J. TESA: PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156 155
Figure 7: Influence of the uncertainties of all the input quantities on
the total uncertainty of the negative gauge pressure generated by the
absolute pressure divider
Slika 7: Vpliv negotovosti vseh vstopnih veli~in na skupno merilno
negotovost negativnega pritiska, ki ga generiramo z absolutnim
delilnikom tlaka
Figure 9: Results of the traceability of the dividing ratio in the
absolute mode for the years 2007 (diamonds) and 2004 (squares)
Slika 9: Rezultati sledljivosti razmerja delitve za absolutni tlak v letu
2007 (rombi) in 2004 (kvadrati)
Figure 8: Results of the traceability of the dividing ratio in the gauge
mode for the years 2007 (diamonds) and 2004 (squares)
Slika 8: Rezultati sledljivosti razmerja delitve za relativni tlak v letu
2007 (rombi) in 2004 (kvadrati)
Figure 10: Uncertainties attainable via different methods
Slika 10: Primerjava merilne negotovosti razli~nih metod
In order to be sure that the dividing ratio is not
influenced by the working mode its traceability in the
absolute mode was also performed. The above-men-
tioned DHI FPG8601 digital non-rotating piston mano-
meter served as a standard for this (Figure 9). The
measured pressure points were (5, 10 and 15) kPa. In this
case the result was RD = (99.997 ± 0.007), and three
years before it was also RD = (99.998 ± 0.007).
Furthermore, the comparison in the negative gauge
mode between the absolute divider and the classical
piston manometer, the digital non-rotating piston mano-
meter with flow centring and the digital non-rotating
piston manometer with lever centring was performed.
The instruments were separated by a zero indicator
(MKS 1 torr Baratron). The comparisons with the digital
non-rotating piston-pressure balances gave worse results
at the lower ends of their ranges, where the noise is
higher. The comparison with the classical piston-
pressure balances also gave worse results at the lower
end of its range (problems with the free rotation time of
the piston), but also at the higher end (problems with the
regulation of the pressure). However, the evaluation
number was lower than one in each case.
6 CONCLUSION
Utilization of the classical piston gauge operating
under a bell jar for the etalonnage of negative gauge
pressures is a useful and cost-effective solution. Figure
10 shows a comparison of the attainable uncertainties of
all of the mentioned instruments. The Czech Metrology
Institute seems to be the only case in the CMC tables
covering the negative gauge pressure range using such a
principle. The problems it brings have been solved to a
satisfactory degree. On the other hand, the use of the
absolute divider for this range is much faster, and with
only an insignificant (for typical calibrations) decrease in
the uncertainty. Moreover, it enables a comparison of
classical and digital pressure balances in the negative
gauge range.
7 REFERENCES
1 M. Rantanen, S. Semenoja, Intercomparison in gauge pressure range
from –95 kPa to +100 kPa, MIKES Publication J1/2006,
Mittatekniikan keskus, Espoo 2006, ISBN 952-5610-11-X, ISSN
1235–2704
2 S. Zuñiga-Gonzalez, P. Olvera-Arana, J.C. Torres-Guzman, F.J.
Flores-Martinez, Pneumatic gauge pressure proficiency test in the
range from –70 kPa to 0 kPa for Mexican accredited laboratories, in:
Abstracts of IMEKO TC3 & TC16 & TC22 International
Conference, Merida, Mexico 2007, 154–155
3 R.S. Dadson, S.L. Lewis, G.N. Peggs, The pressure balance – theory
and practice, Her Majesty’s Stationery Office, London 1982
4 M.H. Orhan, Y. Calkin, J. Tesar, Z., Krajicek, Pneumatic gauge
pressure comparison measurements between the UME (Turkey) and
the CMI (Czech Republic) – EUROMET project No. 537,
Metrologia 38 (2001), 173–179
5 J. Tesar, J. Jaeger, Z. Krajicek, W. Schultz, Pressure comparison
measurement between CMI and PTB in the range 0.07 MPa to 0.4
MPa, Metrologia 36 (1999), 647-650
6 IUPAC – International thermodynamic tables of the fluid state – Vol.
6. – Nitrogen. Pergamon Press, Oxford 1979
7 Available from World Wide Web: www.bipm.org
8 C. G. Rendle, A large area piston gauge for differential and gauge
pressure from zero to 3,2 kPa, Metrologia 30 (1994), 611–614
9 A. Ooiwa, Novel Nonrotational Piston Gauge with Weight Balance
Mechanism for the Measurement of Small Differential Pressures,
Metrologia 30 (1994), 607–610
10 J. Tesar, D. Prazak, The limitations for using the vacuum standards
based on piston-cylinder technique, Vacuum 67 (2002), 311–316
11 J. Tesar, P. Repa, D. Prazak, Z. Krajicek, L. Peksa, The new method
of traceability of a force balanced piston gauge used as primary
vacuum standard, Vacuum 76 (2004), 491–499
12 J. Tesa, D. Pra`ák, F. Stank, P. epa, L. Peksa, Ensuring primary
realization of pressure unit in the vacuum range without typically
utilized static expansion system, Vacuum 81 (2007), 785–787
13 J. Tesa, D. Pra`ák, The methods of CMI traceability for the stan-
dards of the differential pressure at high line pressures. in: Middle
East measurement and instrumentation, Bahrain, 2004, 313–328
J. TESA: PRIMARY ETALONNAGE OF NEGATIVE GAUGE PRESSURES ...
156 Materiali in tehnologije / Materials and technology 43 (2009) 3, 151–156
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
THE GEOLOGICAL RECORD AS AN INDICATOR OF THE
MUDSTONES THERMAL CHARACTERISTICS IN THE
TEMPERATURE RANGE OF DECARBONATISATION
GEOLO[KI ZAPIS KOT POKAZATELJ TERMI^NIH LASTNOSTI
LAPOROVCEV V TEMPERATURNEM OBMO^JU
DEKARBONATIZACIJE
@eljko Poga~nik1, Jernej Pav{i~2, Anton Meden3
1Salonit Anhovo, Building Materials, Joint-Stock Co., Vojkova 1, 5210 Deskle, Slovenia
2Chair of paleontology and strathigraphy, University of Ljubljana, Privoz 11, 1000 Ljubljana, Slovenia
3Faculty of Chemistry and Chemical Technology, University of Ljubljana, A{ker~eva cesta 5, 1000 Ljubljana, Slovenia
zeljko.pogacnik salonit.si
Prejem rokopisa – received: 2008-10-03; sprejem za objavo – accepted for publication: 2008-12-06
The aim of this work is to find a connection between the marlstones geological record, i.e., its origin, and the thermal properties
in the range of the decarbonatisation processes. As the geological record we took the recognized alteration or modification of
the chemical and mineralogical difference, the amount of heavy metals and the fossilized nanoplankton material. For this
research we collected and analyzed mudstones from four different lithological horizons of the subaquatic gravitational
avalanche – turbidite, Paleocene flyschoidic rocks from the Anhovo area.
The remains of the nanoplankton fossilized skeletons are the main bearer of the amount of CaCO3 in this kind of sedimentary
rock, besides a small quantity of carbonate fragments, mostly from older sedimentary rocks. We used optical microscopy, XRF,
XRD and TGA-DTA analytical methods and granulometric analysis. The clay mineralogy was obtained from the acid residual
XRD and DTG-DTA (temperature range to 1000 °C) analyses.
The research results demonstrated that we can expose four different groups of mudstones with similar thermal characteristics in
the temperature range of the decarbonatisation. The chemical alternation degrees – the CIA of the rocks showed a connection
between the clay mineralogy and the temperature course of the thermal decomposition of the mudstone carbonate. The quantity
of nanoplankton skeletons followed a similar trend. The fundamental digressions in the heavy metals were not recognized, but
the Ba quantity showed a good correlation with the samples that had a larger number of skeletons. We related this behavioural
particularity with the biological activity of the environment. Both the CIA factor and the correlation of the skeleton number with
the Ba content demonstrated different sedimentary environments and subsequently the influence of the geological record on the
thermal decomposition nature of the rock in the range of decarbonatisation.
Key words: chemistry and mineralogy, cocolithophycae, mudstone thermal properties
V prispevku `elimo poudariti povezavo med zgodovino (geolo{kega zapisa) nastanka laporovca litotipa mudstone in njegovimi
termi~nimi lastnostmi pri procesu dekarbonatizacije. Kot geolo{ki zapis pojmujemo prepoznavno spremembo v kamnini, ki se
ka`e v njeni kemijski in mineralo{ki sestavi, vsebnosti te`kih kovin in {tevilu fosilnih skeletov kalcitnega nanoplanktona. V ta
namen so bili analizirani vzorci laporovcev iz {tirih litostratigrafskih ~lenov treh ciklotem (zaporedje kamnin, ki so nastale kot
posledica delovanja podvodnih gravitacijskih plazov) paleocenske starosti na obmo~ju Anhovega.
Dele` CaCO3 v laporovcih so ve~inoma skeleti mikrofosilov–kokolitoficej in karbonatni drobir. Vzorci so bili analizirani z
opti~no mikroskopijo, XRF, XRD, TGA-DTA in granulometrijsko analizo. Za dolo~itev mineralov glin so bili vzorci
izpostavljeni kislinskemu razklopu s klorovodikovo kislino. Netopni ostanek je bil ponovno analiziran z XRD- in
DTG-DTA-metodo v temperaturnem obmo~ju do 1000 °C.
Na osnovi rezultatov lahko med analiziranimi vzorci poudarimo {tiri skupine, ki imajo podobne termi~ne lastnosti v obmo~ju
dekarbonatizacije. Stopnja kemijsko spremenjene kamnine-CIA ka`e, da so koli~ina mineralov glin in njihove spremembe
povezane s temperaturnim potekom termi~nega razpada laporovca tipa mudstone. Podobno usmeritev je opaziti tudi pri {tevilu
ohranjenih skeletov kokolitoficej. Bistvenih odmikov med vzorci glede vsebnosti te`kih kovin ni opaziti, razlikujejo pa se v
vsebnosti Ba, ki posredno z ohranjenimi skeleti kokolitoficej ka`e na biolo{ko aktivnost sedimentacijskega okolja. Tako faktor
CIA kot {tevilo kokolitoficej ka`eta razli~no sedimentacijsko okolje kamnine litotipa mudstone oz. da ima geolo{ki zapis vpliv
na lastnosti kamnine v temperaturnem obmo~ju dekarbonatizacije.
Klju~ne besede: kemizem in mineralogija, kokolitoficeje, termi~ne lastnosti laporovcev
1 INTRODUCTION
The mineral raw material used to produce portland
cement clinker (OPC) in the Anhovo area is bound to
subaquatic gravitational avalanches 1, which set off at
unstable shores in the late Paleocen. The raw material
consists of marly breccias, marly limestone and lime-
stones, coarse-grained, medium-grained and fine-grained
lime sandstones, marlaceous limestones and marls, and a
combination of shales, carbonate and silicyclastic silt-
stones and slatestones. The mudstones are carbonate
sedimentary rocks with a silty base and a weight
proportion of grains (either of organic or inorganic
origin) not exceeding 10 %. The grains are 30 µm to 2
mm in diameter. They are typical of a calm and
deep-water environment. In the pelagic environment the
producers of carbonate material are mainly plankton
organisms.
Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163 157
UDK 552.54:54:55 ISSN 1580-2949
Original scientific article/Izvirni znanstveni ~lanek MTAEC9, 43(3)157(2009)
When a marly grain is thermally decomposed, a
sharp edge is produced between a new CaO and a yet
unreacted carbon core, showing that the grain
decarbonatisation process is created first on the inside
and then it spreads outside. The marls carry SiO2. Due to
the genetic character of the rock, it is mostly equally
spread and forms the basis as clay minerals in between
carbonated particles. All the clay minerals lose bound
water between the layers in the temperature range
between 100 °C and 200 °C. First off, the minerals of the
kaolonite groups fall apart between 450 °C (or 500 °C
for the illite group) and 600 °C; they lose chemically
bound OH in the process of complete decomposition. In
the temperature range between 600 °C and 700 °C, the
minerals of the montmorilite group lose the OH group
and this corresponds to the process area of decarbona-
tisation.
This article aims to emphasize the influence of the
geological record of mudstone creation on its thermal
characteristics in the process area of decarbonatisation,
which is in the temperature interval between 650 °C and
950 °C. This adheres to the recognized rock change in
the geological record, reflecting a direct or indirect
sedimentary environment in which the rock has been
created. The changes occur in the mineralogy and the
preserved number of fossil remnants, indirectly
indicating the biological and physico-chemical activities
of the sedimentary environment.
2 MATERIAL SELECTION AND ANALYTICAL
METHODS
The sampled minerals belong to the group of marly
olistolithes of basal breccias (cyclothems Rode` and
Perunk) and the marl of the upper cyclothem part
Podbrdo. Due to the specific genesis3 and rock position4
that involve the sites of mineral raw materials,
significant differences between the individual eroded and
pre-sedimented marls in flyschoidic cyclothems need to
be determined (Figure 1).
The following samples have been analyzed: Sample
13 and Sample 14 (eroded marlstones from marly
calcareous breccia of cyclothem Rode`), marlstone
Sample 9 (from the upper part of cyclothem Rode`),
Sample 18 (marlstone of Podbrdo cyclothem), Sample
20, Sample 21 and Sample 23 (eroded marlstones from
marly calcareous breccia of cyclothem Perunk). To
determine the significant differences between the marl-
stones, the XRF and XRD analyses were carried out,
along with the thermal methods (TG-DTG-DTA), the
analysis of strong acid digestion, granulometric analysis
of the insoluble residue and the petrographic micro-
scopic analyses.
Part of the samples was crushed in a laboratory mill,
and the rest was used to carry out the micro-
paleonthological analysis. We used a scraping tool to
obtain some pulverized parts from the undamaged
sample of the rock. Those parts were dissipated in the
water, dried up and covered in Canadian balm. That
preparation was then evenly distributed on the micro-
scopic glass and covered with a square coverslip, the a
side of 22 mm. We used transmitted light by applying
cross nicols and regularly transmitted light magnified
1000 times; with a microscopic analysis we evaluated
(per unit of surface area) the quantity of nanoplankton
skeletons.
To determine the main oxides (SiO2, Al2O3, Fe2O3,
CaO, MgO, K2O, Na2O, SO3 and calculated CaCO3 ,
assuming that the whole CaO is bound to carbonates)
and certain metals (V, Ba, Ni, Cr, Ti in Mn), we used
XRF analyses. The voltage and current of the x-ray tube
of the ARL 8480 instrument were 30 kV and 80 mA,
whereas the recording time was set to 40 s. The samples
were examined by applying XRD analyses and the
TG-DTG-TGA methods (defining CaCO3 from the
emitted CO2).
The samples for the XRD analysis were recorded on
a PANalytical X’Pert PRO MPD diffractometer by
applying the reflection technique and the CuK 
radiation in the 2  range from 3° to 70° in steps of
0.034° 2 , an integration time of 400 s, a 20-mm mask
and a divergent and an antiscatter slit, set to 20 mm of
the sample irradiation. Soller slits of 0.02 radians were
used on both sides and the whole detector range was
applied. The sample was rotated around the vertical axis.
The Rietveld analysis of the diffraction pattern was
performed with the TOPAS 3.0 programme and with the
use of models containing 9 to 12 of the crystalline
phases. In all cases a very good final match of the
calculated and measured pattern with Rwp at around 6 %
was obtained.
While analysing the samples with thermal methods,
well-homogenized measured parts were heated in cylin-
drical pots (1 mm to 3 mm in height and 4 mm in
diameter) made of inert materials (platinous and Al2O3
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
158 Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163
Figure 1: The classification of the analyzed samples in mega cyclo-
thems
Slika 1: Razporeditev analiziranih vzorcev v megaciklotemah
pots) by applying a temperature scan of 10 °C/min at
room temperature and up to 1400 °C. We used the STD
2960 Simultaneus DTA-TGA analyser to carry out this
procedure. When the signal analysis was made, we
elaborated on the temperature parameters, which are
presented in Table 1, see below. The software tool Uni-
versal Analysis for Windows 95/98/NT, edition 2.5H,
was used to define the parameters.
Table 1: Temperature parameters5,6,7
Tabela 1: Temperaturni parametri 5,6,7
Para-
meter Explanation
T2
The temperature at which the reaction ends,
corresponds to Toffset (identified with 1. derivative
of the TG-curve)
Tonset
The temperature which triggers the decomposition
reaction – crystallization − is the lowest
temperature which triggers the process   T1
Tmax
The temperature at which the process triggered
reached its peak
To divide the clay component basis in marlstones, we
carried out an acid decomposition of the samples. In a
container with 450 mL of distilled water there was a
100-gram sample, to which we gradually added diluted
hydrochloric acid (37 % concentrated HCl acid, mixed
with distilled water in a 1:1 ratio) with a total capacity of
200 mL. The insoluble residue was decanted with
destilled water to reach pH  4. Then the material was
dried in a drier at 37 °C for 18 h. The samples obtained
were passed through a sieve of 45 µm and the residue
was analysed by applying the XRD-method and thermal
analyses. On the basis of the results obtained, we
evaluated the mineralogy of the insoluble residue.
3 RESULTS OF THE ANALYSIS
The microscopic analysis showed that Sample 23
(followed by Sample 21 and Sample 14 – Table 2)
contains the largest number of fossil skeletons of the
nanoplankton organisms.
Table 2: The number of fossil units (the number on the diameter area
of 22 mm2) in the individual sample
Tabela 2: [tevilo fosilnih enot ({t. na povr{ini premera 22 mm2) v
posameznem vzorcu
SAMPLE Number of units
Sample 9 162
Sample 13 142
Sample 14 370
Sample 18 120
Sample 20 165
Sample 21 468
Sample 23 880
In Sample 23, besides many other types of calcite
nanoplankton, we also detected the following types
(Figure 2): Chiasmolithus bidens (Bramlette et Sulli-
van), Ellipsolithus macellus (Bramlette et Sullivan) ter
Fasciculithus tympaniformis (Hay et Mohler).
According to the proportion of the calcium carbo-
nate, determined on the basis of the emitted CO2 at the
decarbonatisation process (Table 3), we concluded that
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163 159
Figure 2: Calcite nanoplankton from the Sample 23: a) Chiasmolithus bidens, b) Ellipsolithus macellus in c) Fasciculithus tympaniformis.
Pictures a) and c) are recorded under crossed nicols, b) is recorded by applying regular light.
Slika 2: Kalcitni nanoplankton iz vzorca 23: a) Chiasmolithus bidens, b) Ellipsolithus macellus in c) Fasciculithus tympani-
formis. Sliki a) in c) sta posneti pod navzkri`nimi nikoli, b) pa pri navadni svetlobi.
Figure 3: DTA curves of the Samples 21, Sample 23 and Sample 14
Slika 3: DTA krivulja vzorcev Vz. 21, Vz. 23 in Vz. 14
Sample 14 contains the largest amount of carbonate,
whereas Sample 21 and Sample 23 contain the smallest
amount.
Figure 3 shows concave curves of the DTA samples
23, 21 and 14. The number of DTA signals shows at
least two endothermic reactions in the range of the
thermal decomposition of carbonate, whereas the curve
of Sample 14 has an extremely low amplitude and a
shifted Tmax.
The TG-DTG-DTA analysis of the insoluble residue
in the decarbonatisation range shows strong endotermic
peaks; they show characteristic temperatures at the
beginning of the reaction and maximum temperatures at
the very end of the reaction of the thermal carbonate
decomposition (see Table 4).
The results of the XRD analysis (Table 5) show the
presence of carbonate in the insoluble residue.
In all the primary samples the calcite is dominant,
and flexures of different clay minerals at low angles can
be clearly seen. Besides calcite, the following were also
defined in the primary sample: quartz (SiO2), muscovite
(KAl2(AlSi3O10)(OH)2, vermiculite (Mg4(AlSi3O10)
(OH)2(H2O)4, kaolinite (Al2(Si2O5)(OH)4, kronstedtite
(Fe3(FeSi)O5(OH)4, cristobalite (SiO2), iIlite
(K(Al4Si2O9(OH)3)), dolomite (CaMg(CO3)2, ankerite
(CaFe(CO3)2, aragonite CaCO3 and protoenstatite
(MgSiO3). In the insoluble residue the calcite was not
identified in samples 9 and 21. The proportion of false
analysis is 3 % absolute with calcite, 1 % with musco-
vite, kaolinite and illite and 0.5 % with other compo-
nents. Total proportion of the unidentified crystalline
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
160 Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163
Table 3: The results of the XRF- and TG-DTG-DTA analyses of the primary sample
Tabela 3: Rezultati XRF- in TG-DTG-DTA analize osnovnega vzorca
SAMPLE
proportion CO2 CaCO3
TG Ti SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O CaCO3 V Ba Ni Cr Mn
TG-DTG-DTA XRF
(mass proportion, w/%) w/(µg/g)
Smp. 9 23.65 53.75 0.15 19.60 5.16 2.40 37.65 1.37 1.11 0.22 67.21 56 550 53 54 568
Smp. 13 25.29 57.48 0.14 18.96 4.73 2.28 38.18 1.49 1.02 0.21 68.16 49 454 52 53 499
Smp. 14 26.73 60.75 0.16 21.80 5.52 2.80 35.33 1.67 1.25 0.25 63.07 61 457 60 57 533
Smp. 18 25.75 58.52 0.15 20.74 4.83 2.25 37.48 1.40 1.00 0.26 66.90 54 324 50 57 381
Smp. 20 23.70 53.86 0.15 21.05 4.85 2.28 37.34 1.38 1.00 0.26 66.65 52 377 51 58 382
Smp. 21 19.14 43.50 0.20 26.14 6.71 3.09 32.39 1.42 1.32 0.18 57.81 66 199 65 66 463
Table 5: XRD and TG-DTG-DTA determination of certain mass proportions of minerals (%) in the primary sample (BULK) and its insoluble
residue (<45 µm)
Tabela 5: XRD-in TG-DTG-DTA-dolo~itev nekaterih masnih dele`ev mineralov (%) v osnovnem vzorcu (BULK) in njegovem netopnem
ostanku (<45 µm)
SAMPLE
Kaolinite
Al2(Si2O5)(OH)4
Illite
K(Al4Si2O9)(OH)3
Muscovite
KAl2(AlSi3O10)(OH)2
Illite/
muscovite
Albite
NaAlSi3O8
Calcite
CaCO3
Calcite
CaCO3
BULK 45 µm< BULK BULK <45 µm <45 µm BULK <45 µm
Smp. 9 3.30 9.32 3.81 9.26 35.38 8.84 75.53 -
Smp.13 2.97 2.87 3.94 9.04 13.63 7.25 73.13 57.84
Smp.14 3.04 5.72 4.91 9.46 36.31 7.96 72.73 14.46
Smp.18 2.52 1.03 3.58 9.06 8.78 4.49 74.60 74.75
Smp. 20 2.69 2.04 3.70 9.05 10.19 4.38 74.07 71.22
Smp. 21 2.81 7.13 5.47 11.23 45.67 8.02 69.71 -
Smp. 23 2.81 5.52 6.20 8.84 35.97 5.72 71.06 23.84
Table 4: TG-DTG-DTA analysis of the original sample and the
insoluble residue (the residue on a sieve of 45 µm) in the the tempera-
ture interval of decarbonatisation
Tabela 4: TG-DTG-DTA-analiza originalnega vzorca in netopnega
ostanka (ostanek na situ 45 µm) v temperaturnem intervalu dekarbona-
tizacije
SAMPLE
Analysis of the original sample
Tonset Tmax
Tmax
-Tonset
T2 CO2 CaCO3
°C w/%
Smp. 9 717.13 778.62 61.49 812.60 23.65 53.75
Smp.13 715.62 779.58 63.96 812.70 25.29 57.48
Smp.14 757.94 835.10 77.16 911.00 26.73 60.75
Smp.18 716.18 786.78 70.60 825.30 25.75 58.52
Smp. 20 713.88 777.31 63.43 809.50 23.70 53.86
Smp. 21 709.51 769.41 59.90 798.40 19.14 43.50
Smp. 23 706.43 770.20 63.77 801.60 18.68 42.45
Analysis of the insoluble sample residue on the sieve of 45 µm
Smp. 9 - - - - -
Smp.13 682.91 767.27 84.36 798.90 23.13 52.57
Smp.14 728.26 795.22 66.96 824.90 10.61 24.11
Smp.18 695.72 777.98 82.26 809.76 27.47 65.18
Smp. 20 682.91 767.27 84.36 798.90 23.13 62.43
Smp. 21 - - - - -
Smp. 23 636.58 751.11 114.53 837.60 10.52 23.91
phases is evaluated to less than 5 %; the amorphous
phase is probably less than 10 %.
4 DISCUSSION
The number of determined nanoplankton skeletons
(Table 1) can be compared to the determined carbonate
proportion in the individual sample (Tables 2 and 3). The
difference can occur because of the group of carbonate
minerals, parts of them, not bound to the then existing
living environment of the nanoplankton organisms.
This is confirmed by the results of the diffraction and
thermal analyses. The data show that the samples diffe-
rentiate by the process type of decarbonatisation (double
endothermic peaks; Figure 3), in the Tonset temperature
and in the time to reach the maximum temperature of the
carbonate decomposition (Tmax, Table 4). In the sifted
residue of the acid-treated sample (the size of grains
below 45 µm) there is an unreacted carbonate (TG-DTG
and XRD analyses), showing the presence of relatively
older carbonate rocks that were eroded off from the older
beds in the sedimentation process (of subaquatic avalan-
che) or brought in from the different sedimentary
environment (most likely diagenetically already altered –
alumosilicate cement).
The interdependence of the mudstone rock samples
from the Paleocenic flyschoidic beds on the level of the
temperature parameters (Tonset, Tmax in Tmax–Tonset), the
geological record indicating when the marlstones were
created (CIA*,8, K2O/Al2O3, K2O+Na2O, K2O, CaCO3,
Cr/Ni, Cr/V, Mn, Ti/Al, K/(Fe+Mg), Ba and the number
of skeletons) and the ratio F** are summarized in the
Table 6.
Through a statistical overview of the results it can be
concluded that the CIA factor and the number of
preserved cocolithophycae skeletons have a great
influence on the start-up reaction temperature Tonset and
the maximum temperature of the reaction process Tmax.
From the geological record it can also be concluded that
K-alumosilicates and clay minerals are prevalent in the
samples (as shown by the results of the diffraction
analysis – Table 5). The influence is shown in the
Tmax–Tonset temperature interval. Diagenesis runs were
made to disintegrate the acid feldspars (orthoclase) into
K-feldspars and further into clay minerals. In our case,
the diffraction analysis did not identify the orthoclase;
the albite, however, was identified. It appears together
like the mineral joint with orthoclase in medium and acid
magmatic rocks.
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163 161
Table 6: Correlation table of the temperature parameters, the geological record of the marlstones and the F ratio
Tabela 6: Korelacijska tabela temperaturnih parametrov, geolo{kega zapisa laporovcev in razmerja F
F
T
on
se
t
T
m
ax
T
m
ax
–T
on
se
t
C
IA
w
(K
2O
)/
w
(A
l 2
O
3)
w
(K
2O
+
N
a 2
O
)
w
(C
aC
O
3)
{t
.
sk
el
et
ov
w
(B
a)
w
(K
2O
)
w
(C
r)
/w
(V
)
w
(C
r)
/w
(N
i)
w
(K
)/
w
(F
e
+
M
g)
w
(T
i)
/w
(A
l)
w
(S
iO
2)
/w
(A
l 2
O
3)
F 1
Tonset 0.892 1
Tmax 0.932 0.990 1
Tmax–Tonset 0.946 0.853 0.917 1
CIA 0.770 0.871 0.898 0.876 1
w(K2O)/w(Al2O3) 0.836 0.828 0.803 0.644 0.507 1
w(K2O + Na2O) 0.617 0.433 0.509 0.669 0.336 0.317 1
w(CaCO3) 0.859 0.648 0.676 0.682 0.341 0.867 0.616 1
The No.of skeletons 0.765 0.919 0.879 0.665 0.731 0.895 0.149 0.592 1
w(Ba) 0.479 0.461 0.416 0.238 0.054 0.846 0.131 0.710 0.675 1
w(K2O) -0.289 0.009 -0.011-0.066 0.377 -0.402-0.469-0.719 0.042 -0.518 1
w(Cr)/w(V) -0.385-0.653-0.596-0.367-0.570-0.505-0.029-0.178-0.718-0.331-0.315 1
w(Cr)/w(Ni) -0.075-0.366-0.312-0.122-0.524-0.188 0.584 0.271 -0.553-0.060-0.763 0.558 1
w(K)/w(Fe + Mg) -0.582-0.266-0.317-0.429-0.056-0.471-0.514-0.780-0.117-0.310 0.795 -0.400-0.546 1
w(Ti)/w(Al) 0.073 -0.192-0.143 0.014 -0.349-0.095 0.623 0.352 -0.444-0.123-0.748 0.449 0.948 -0.646 1
w(SiO2)/w(Al2O3) 0.269 -0.029 0.041 0.238 -0.157 0.029 0.735 0.474 -0.303-0.054-0.729 0.550 0.885 -0.791 0.925 1
*
CIA
w
w w w w
=
+ + +
⋅(Al O )
(Al O ) (CaO) (Na O) (K O)
2 3
2 3 2 2
100; proportion of CaO* is a calculated proportion of CaO on the basis of the CaO difference,
set by XRF-analysis and CaO, defined with the decarbonatisation process applying DTG-DTA-analysis.
**
F
w
w
= (CaO )
(CaO )
t
f
, CaOf is a calcium oxide, which is created through decarbonatisation process and the proportion of CaOt is bound to an absolute
quantity of Ca-oxide, defined in the sample on the basis of XRF-analysis.
The proportion of heavy metals in the samples
(Table 3) is shown by the influence of the disintegrated
ultramaphic complex 9 (correlation between Cr/Ni and
Cr/V with SiO2/Al2O3 proportion, Table 6) and the
terigen insertion of individual minerals in the source of
the clay component in the marlstones. Titanium is an
extremely conservative trace element 10 and as such it is
resistent to diagenetic factors, while the Al proportion
has a dominant role in indicating the clay minerals. The
correlation between Ti/Al and SiO2/Al2O3 (Table 6)
shows that Ti entered the sedimentary environment
through the river input or through earlier, eroded
sedimentary rocks.
The hypothesis about the vulcanoclastic source of
sediments is not acceptable, as this theory is disproved
by a negative correlation 10 between the Ti/Al and the
K/(Fe+Mg). The proportion of Mn11 in the samples also
shows the terigen import of the sediment or presedi-
mentation. The negative correlation between the Ba and
K2O shows that Ba is not bound to the clay component
12, although Ba2+ and K2+ can interchange in the crystal
structure. So the source of Ba is to be ascribed to a
general carbonate cycle, which corresponds to the ratio
of CaCO3 proportion and the number of nanoplankton
skeletons in individual samples (Table 2).
5 CONCLUSION
The thermal analysis results show that marlstones of
the mudstone lythotype react differently during the
decarbonatization process (Figure 4) in the temperature
range between 650 °C and 950 °C. Four groups can be
distinguished among the curves. The two pairs of curves
of the Sample 23 and Sample 21 are prominent among
them, having two endothermic peaks. Sample 18 and
Sample 13 are also worth mentioning, with the second
endothermic peak being less emphasized.
The oldest marlstones – olistholites (Sample 13 and
Sample 14) – taken from a marly calcareous breccia
comper with Sample 9, from the upper part of the Rode`
cyclothem, have the highest F ratio among all the
samples, considering their litostratigraphic sequence.
They differ from one another in the number of nano-
plankton skeletons and in the proportion of chemical
change CIA factor. We can therefore conclude that the
rocks were formed in two different sedimentary
environments.
The marlstones of the Podbrdo cyclothem (Sample
18) differ from the previously mentioned samples in the
quantity of the biogenic carbonate (number of skeletons)
and Ba proportion, therefore confirming the biological
activity of the environment. They also differ in the CIA
factor, which is lower in the sample richer in fossils. In
Sample 18 carbonate was identified in the insoluble
residue; we therefore assume that in the sedimentation
period different sources of carbonate material were
present.
The marlstone samples (Sample 21, Sample 23 and
Sample 20), taken in the youngest horizon of the marly
calcareous breccia of Perunk cyclothem, differ from one
another by the F ratio. Among the marlstones of this
lythological horizon the F value is larger in the samples
which hold a higher proportion of Ba. The number of
skeletons and the content of Ba show that there are at
least two different sedimentary environments to be
distinguished in this marlstone group.
The CIA proportion (Figure 5) confirms that the
number of feldspars and clay minerals, and the
accordingly bound diagenetic changes, are related to the
temperature process of the thermal decomposition of the
carbonate rock. This is also evident from the correlation
factor between the CIA and the start-up reaction tempe-
rature Tonset, as well as the maximum decarbonatization
temperature Tmax. A similar trend can be seen in the
number of preserved cocolithophycae skeletons. Despite
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
162 Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163
Figure 5: The ratio between the number of skeletons, Ba proportion
(µg/g), start-up reaction temperature Tonset (°C) and the F factor
Slika 5: Razmerje med {tevilom skeletov, dele`em Ba (µg/g), tempe-
raturo za~etka reakcije Tonset (°C) in faktorjem F
Figure 4: Comparison of marlstone DTA curves of mudstone type
Slika 4: Primerjava DTA-krivulj laporovcev tipa mudstone
the approximately even number of skeletons the curves
of Sample 23 and Sample 13 are shaped differently. The
difference can be seen in the CIA ratio, as the latter is 13
% higher in Sample 23 than in Sample 13.
The proportion of heavy metals indicates the source
of fine-grained material13 (most likely the area of today’s
Carnic Alps), whereas the number of preserved skeletons
and the Ba content indicate the change of the sedi-
mentary environment in which the rock was created. The
proportion of carbonate in the insoluble residue still
remains unknown at this stage, although it does allow for
the hypothesis of the material in the sample being of
secondary origin (either presedimented or eroded from
older cyclothems).
The highest CaO proportion, bound in carbonate, and
the largest amount of cocolithophycae can be found in
Sample 14. This is followed by Sample 18, having the
largest carbonate proportion in the insoluble residue;
here the ratio between Ba and the number of skeletons
allows for the hypothesis of the larger carbonate propor-
tion being bound to fossil pelagic organisms, i.e. it is of
biogenic origin.
The marlstone reflects the chemical and mineralogic
features influencing the temperature reaction process in
the carbonate decomposition range. The results of the
analysis show that a sole definition of the rock lythotype
does not clearly establish the temperature characteristic
of the sedimentary rock 2 in the decarbonatisation area.
Consequently, the geological record has an important
influence on the temperature range of the decarbonati-
sation of the mudstone rock in the Paleocenic sedi-
mentary environment in the close-by area of Anhovo.
6 REFERENCES
1 Skaberne, D. Sedimentological investigation of flysch from the
Anhovo area. Diploma degree (in Slovene), Katedra za geologijo in
paleontologijo, NTF, Ljubljana, 1973, 285
2 Dunham, R. J. Classifications of carbonate rocks according to
depositional texture. V: Clasification of Carbonate Rocks, a Sim-
posium. Edited by Ham, E.W. Tulsa, (1962), 108–122
3 Skaberne, D. Observation Point N° II/9, The Paleocene megaturbi-
dites – Anhovo. V Evolution of the karstic carbonate platform –
Excursions Guidebook, 5. June – 6. June. Trieste : Universitá degli
studi di Trieste, Instituto di geologia e paleontologia, 1987, 37–45
4 Poga~nik, @., Car, M. Prostorska umestitev fli{a v okolici Rode`a-
integrirani raziskovalni pristop. In: Horvat, Aleksander (ur.). 18.
posvetovanje slovenskih geologov, (Geolo{ki zbornik, 19). Ljublja-
na: Univerza v Ljubljani, Naravoslovnotehni{ka fakulteta, Oddelek
za geologijo, 2007, 92–96
5 Mackenzie, R. C. The differential thermal investigation of clays.
London: Mineralogical society (clay minerals group), 1957, 456 str.
6 Ramachandran, V. S., Phil, D. Applications of differential thermal
analysis in cement chemistry. Chemical pulishing company, 1969,
308
7 Yen, F. S., Lo, H. S., Wen, H. L., Yang, R. J.   – to  – phase trans-
formation subsystem induced by  – Al2O3 – seeding in boehmite –
derived nano – sized alumina powders. Journal of Crystal Growth,
(2003), 249, 283–293
8 Nesbitt, H. W., Young, G. M. Early Proterozoic climates and plate
motions inferred from major element chemistry of lutites. Nature,
(1982) 299, 715–717
9 Barbara B., Mclennan, S. M., Hanson, G. N. Geochemistry and
provenance of the Middle Ordovician Austin Glen Member
(Normanskill Formation) and the Taconian Orogeny in New
England. Sedimentology 45 (1998), 635–655
10 Meyers, S. R., Sageman, B. B., Hinnov, L. A. Integrated quantitative
stratigraphy of the Cenomanian–Tturonian Bridge Creek limestone
member using evolutive harmonic analysis and stratigraphic
modeling. Journal of Sedimentary Research, 71 (2001), 4, 628–644
11 Veizer, J., Demovi~, R. Environmental and climatic controlled
fractionation of elements in the mesozoic carbonate sequences of the
western Carpathians. 1. Journal of Sedimentary Research, 43 (1973)
1, 258–271
12 Dinelli, E, Lucchini, F., Mordenti, A., Paganelli, L. Geochemistry of
Oligocene-Miocene sandstones of the northen Apennines (Italy) and
evolution of chemical features in relations to provenance changes.
Sedimentary Geology, 127 (1999), 193–207
13 Venturini, S., Tunis, G. La composizione dei conglomerati cenozoici
del Friuli: dati prliminari. In: Studi Geologici Camereti, Studi
preliminari all’ acqusizione dati del profilo CROP 1-1A La Spezia-
Alpi orientali, 1992, 285–295
@. POGA^NIK ET AL.: THE GEOLOGICAL RECORD AS AN INDICATOR OF THE MUDSTONES ...
Materiali in tehnologije / Materials and technology 43 (2009) 3, 157–163 163

A. MAGLICA ET AL.: KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA NITRIDA
KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA
NITRIDA
CERAMIC COMPOSITES BASED ON SILICON NITRIDE
Aljo{a Maglica, Kristoffer Krnel, Milan Ambro`i~
Odsek za in`enirsko keramiko, Institut "Jo`ef Stefan", Jamova 39, Ljubljana
aljosa.maglica ijs.si
Prejem rokopisa – received: 2008-10-07; sprejem za objavo – accepted for publication: 2009-04-08
Keramika na osnovi silicijevega nitrida (Si3N4) ima odli~ne mehanske lastnosti in korozijsko obstojnost, zato se uporablja za
razli~ne namene, tudi pri visokih temperaturah, kjer so drugi materiali manj primerni. V tem prispevku je opisana uporaba
nekaterih kompozitov Si3N4. Z dvema tak{nima kompozitoma (sialoni in silicijev nitrid s prevodno manj{insko kerami~no
komponento) smo se ukvarjali tudi v na{i raziskovalni skupini.
Klju~ne besede: Si3N4, SiAlON, kerami~ni kompoziti
Ceramics based on silicon nitride (Si3N4) have excellent mechanical properties and corrosion resistance, thus, they are used for
various applications, even at high temperatures where other materials are less appropriate. In this paper the use of some Si3N4
composites is presented. Two of such composites (sialons and silicon nitride with conducting minor ceramic component) were
studied in our research group.
Key words: Si3N4, SiAlON, ceramic composites
1 UVOD
Silicijev nitrid (Si3N4) spada med strukturne kera-
mi~ne materiale z najbolj{imi mehanskimi lastnostmi,
kot ~ista snov ali kot sestavina kompozitov. V naravni
obliki so ga na{li le v obliki minerala nierit v meteoritih,
sicer pa ga izdelujejo z direktno reakcijo med silicijem in
du{ikom pri visoki temperaturi.1 ^isti Si3N4 uporabljajo
zaradi njegove trdote, obstojnosti pri visokih tempera-
turah in dobre obrabne obstojnosti za trde prevleke na
orodjih, posebno za obdelavo sive litine, medtem ko je
za obdelavo jekel primernej{i titanov nitrid. Silicijev
nitrid v trdih prevlekah orodij, kjer so potrebne velike
obdelovalne hitrosti, tekmuje z drugimi kerami~nimi
materiali: korundom (Al2O3), kompozitom Al2O3/TiC,
polikristalnim diamantom in kubi~nim borovim nitridom
(c-BN). Kerami~ne trde prevleke se na orodja nana{ajo
najve~krat z naparevanjem, ~eprav so pomembni tudi
drugi na~ini, npr. laserski difuzijski postopki. Si3N4 je
primeren tudi za krogli~ne le`aje, saj je v primerjavi s
kovinami tr{i in bolj tog. Zato se manj deformira, tako da
manj{a trenje med le`aji in njihovim nosilcem ter obrabo
pri velikih hitrostih. Tak{ne le`aje najdemo v vozilih,
turbinah in celo mlinih na veter. V mikroelektroniki se
Si3N4 uporablja za elektri~noizolacijske plasti.2
Obstajajo tri kristalografske modifikacije silicijevega
nitrida: α, β in γ.1 Medtem ko sta prvi dve obi~ajni,
lahko sintetiziramo γ-strukturo Si3N4 le pri visokih tem-
peraturah in tlakih; ta struktura ima zelo visoko trdoto 35
GPa (za primerjavo: zna~ilne trdote aluminijevega
oksida Al2O3, ki velja za zelo trd material, so okrog 15
GPa). α- in β-Si3N4 imata heksagonalno kristalno struk-
turo, γ-Si3N4 pa je kubi~en. Kristalna mre`a α- in
β-Si3N4 je iz tetraedrov SiN4, tako da je Si v te`i{~u
vsakega tetraedra, 4 du{ikovi atomi pa v ogli{~ih.3 V
enem ogli{~u se stikajo trije tetraedri, zato imamo
nazadnje pravo razmerje {tevila silicijevih in du{ikovih
atomov, 3 : 4 (slika 1).
Razlika med strukturama je v simetriji porazdelitve
tetraedrov: manj simetri~na  -modifikacija ima pribli`no
dvakrat ve~jo osnovno celico (osnovna celica je pravilna
3-strana prizma). Robovi prizme so: a = b = 0,7765 nm,
c = 0,5675 nm za α-Si3N4; a = b = 0,7709 nm, c = 0,2908
nm za β-Si3N4 (nm = 10–9 m). Gostota Si3N4 je 3,44
g/cm3, Youngov modul okrog 300 GPa, upogibna trdnost
vsaj 500 MPa, tali{~e pa 1900 °C. Silicijev nitrid dobimo
kot siv prah brez vonja (proizvajalci: UBE Industries
Ltd., Japonska; Merck, Nem~ija itd.), in sicer v α- in
β-modifikaciji (slika 2).
Sintranje silicijevega nitrida do visokih gostot je
mogo~e le ob upo{tevanju naslednjih pogojev, kot so:
uporaba zunanjega pritiska, povi{anje temperature sin-
Materiali in tehnologije / Materials and technology 43 (2009) 3, 165–169 165
UDK 54:666 ISSN 1580-2949
Strokovni ~lanek/Professional article MTAEC9, 43(4)165(2009)
Slika 1: Kristalna struktura α- in β-Si3N41
Figure 1: Crystal structure of α- and β-Si3N41
tranja in tlaka du{ika, uporaba dodatkov za sintranje.
Treba je tudi zagotoviti ravnote`ne parne tlake vseh
komponent z uporabo zasipa iz prahu Si3N4. Navadno ob
upo{tevanju omenjenih pogojev ne dobimo dobrih
rezultatov oz. je priprava tovrstne keramike predraga,
zato je treba Si3N4 dodati ustrezne snovi, ki pri sintranju
tvorijo teko~o fazo. ^e teko~a faza dobro omaka Si3N4 in
je ta v njej dobro topen, lahko zgo{~evanje opi{emo s
klasi~no teorijo sintranja, kot jo je predlagal Kingery:
prerazporeditev, raztapljanje-difuzija-precipitacija in
koalescenca.
Mehanske in druge lastnosti silicijevega nitrida lahko
izbolj{amo tudi z drugimi dodatki, npr. z ogljikovimi
vlakni, nanodelci SiC (silicijev karbid) itd. Si3N4 pa
dodajajo tudi kot manj{insko komponento drugim
kerami~nim materialom, npr. Al2O3 ali ZrO2, da bi
izbolj{ali njihove lastnosti, kot je obrabna obstojnost. Z
dodatki pa lahko spremenimo tudi funkcionalne lastnosti
Si3N4. Slika 3 prikazuje kompozite silicijevega nitrida s
cirkonijevim nitridom Si3N4/ZrN kot enega od razli~nih
prevodnih kerami~nih materialov. Prevodno komponento
smo za preskus dodajali tudi sialonom. Zgleda za to sta
1) ZrN, ki nastane med reakcijskim sintranjem zaradi
dodanega cirkonijevega oksida ZrO2 v za~etno me{anico
prahov, in 2) titanov nitrid TiN, ki nastane z reakcijo
med silicijevim nitridom in dodanim titanovim oksidom
TiO2.
SEM-posnetka polirane povr{ine sintranih kompo-
zitnih materialov na osnovi Si3N4 z dispergiranimi delci
ZrN ali TiN sta prikazani na sliki 4. S slike je razvidno,
da imata oba kompozita homogeno mikrostrukturo z
visoko relativno gostoto (> 96 %), ki daje dobre mehan-
ske lastnosti. Mikrostrukturi obeh materialov ka`eta
podolgovata zrna β-Si3N4, med katerimi so razporejena
svetla zrna ZrN (slika 4 a) ali TiN (slika 4 b).
A. MAGLICA ET AL.: KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA NITRIDA
166 Materiali in tehnologije / Materials and technology 43 (2009) 3, 165–169
Slika 3: Vzor~ne pal~ke za preskuse trdnosti: svetle pal~ke so surovci
Si3N4 z dodanim prahom ZrO2, temne pal~ke pa so sintrani vzorci;
~isti Si3N4 je svetlo siv, ti vzorci pa so temnej{i zaradi prisotnosti
ZrN, ki je nastal pri kemijski reakciji med sintranjem; izdelano na IJS.
Figure 3: Test bars for measuring flexural strength: bright bars are
Si3N4 green-bodies with dispersed ZrO2 powder while the dark ones
are sintered samples; pure Si3N4 is bright grey, but these samples are
darker due to the presence of ZrN which was formed during sintering;
fabricated at the JSI.
Slika 2: Prah β-Si3N4 na grafitnem lepilnem traku; vrsti~ni elektron-
ski mikroskop na poljsko jakost (FEG-SEM), pove~ava 30 000-kratno,
Kemijski in{titut
Figure 2: β-Si3N4 powder on graphite sticking tape, field emission
scanning electron microscope (FEG-SEM), magnification 30000.
National Institute of Chemistry.
Slika 4: SEM-posnetek polirane povr{ine sintranega kompozita: a)
Si3N4/ZrN in b) Si3N4/TIN, pove~ava 3000-kratna. IJS
Figure 4: SEM picture of polished surface of sintered composite: a)
Si3N4/ZrN in b) Si3N4/TIN, magnification 3000. IJS
Si3N4 je tudi surovina za kompozitno SiAlON
(sialonsko) keramiko. Le-ta je trdna raztopina Si3N4 in
Al2O3 z raznimi dodatki za sintranje, kot je itrijev oksid
Y2O3. Sialoni nastanejo z delno zamenjavo silicija in
du{ika v kristalni mre`i Si3N4 z aluminijem in kisikom
(odtod njihovo ime: Si = silicij, Al = aluminij, O = kisik,
N = du{ik).4 Pomembni modifikaciji sialona sta α in β,
osnovani na α- in β-Si3N4, zato je morfologija kristalnih
zrn sialona α in β podobna morfologiji zrn Si3N4.
Okvirne vrednosti najpomembnej{ih lastnosti sialonov
so prikazane v tabeli 1. α-sialon ima v primerjavi z
β-sialonom odli~no trdoto, vendar slab{o trdnost,
`ilavost, toplotno prevodnost in oksidacijsko odpornost,
medtem ko ima β-oblika podolgovata zrna z dobro
lomno `ilavostjo in toplotno prevodnostjo. S primerno
kombinacijo obeh faz je mo`no dose~i optimalne
mehanske lastnosti.5
Tabela 1: Fizikalne lastnosti sialonov2
Table 1: Physical properties of sialons2
Mehanske lastnosti pri sobni temperaturi
Upogibna trdnost, MPa 800
Youngov modul, GPa 290
Gostota, g/cm3 3,3
Trdota, GPa 17
Lomna `ilavost, MN m–3/2 6,2
Termi~ne in elektri~ne lastnosti
Specifi~na toplota, J/(kg K) 710
Elektri~na upornost, Ω cm 1012
Koeficient termi~nega raztezka, 10–6 K–1 3,3
Toplotna prevodnost, W/(m K) 18
β-SiAlON-e opi{emo s formulo Si6–xAlxOxN8–x (0 = x
= 4,2).6 Keramika z nizko vrednostjo x ima dobro trdnost
pri visoki temperaturi in je odporna proti lezenju in
oksidaciji, keramika z visoko vrednostjo x pa ima dobre
lastnosti pri ni`jih temperaturah, visoko trdnost, `ilavost
in abrazijsko obstojnost. α-SiAlON opi{emo s formulo
MexSi12–(m+n)Al(m+n)OnN16–n.7,8 Me ozna~uje dodatno ko-
vino (Me = metal, npr. Li, Mg, Ca itd.), njen dele` x pa
je pri znanih parametrih m in n dolo~en z valenco kovine
Me. Iz formul za oba sialona je razvidno, da je osnovna
celica modifikacije α pribli`no dvakrat ve~ja kot za
modifikacijo β, tako kot pri silicijevem nitridu.
Sialonska keramika je v zadnjem ~asu dobila velik
pomen in se uvrstila med komercialne materiale visoko-
tehnolo{kih kerami~nih produktov. Zaradi dobrih
visokotemperaturnih mehanskih in termi~nih lastnosti je
v zadnjem desetletju do`ivela nagel razvoj in vse ve~je
povpra{evanje. Pomembna prednost te keramike je
predvsem v enostavnej{i pripravi in doseganju bolj{ih
mehanskih lastnosti v primerjavi s ~istim Si3N4. V
metalurgiji izrabljamo obstojnost sialonov v stiku s sta-
ljenim bakrom, aluminijem in njunimi zlitinami. V stroj-
ni{tvu izrabljamo njene dobre tribolo{ke lastnosti
(lastnosti pri agresivnem trenju z drugimi materiali) in
odpornost proti abraziji. Iz sialonov se izdelujejo orodja
za vle~enje cevi in `ic, krogelni in drsni le`aji, rezalna
orodja za obdelavo super zlitin in sive litine pri visokih
hitrostih rezanja, hkrati pa je zaradi visoke `ilavosti
rezalno orodje iz sialonov mo`no uporabljati tudi za
rezkanje.9
2 EKSPERIMENTALNO DELO
Na Odseku za in`enirsko keramiko K6 IJS smo
raziskovali pripravo materiala za prototipno kerami~no
~epno `arilno sve~ko iz Si3N4 z dispergiranimi prevod-
nimi delci. Z ta namen lahko uporabimo tudi sialonsko
matrico, sialoni pa so zaradi dobrih mehanskih lastnosti
zanimivi tudi kot strukturni materiali.
2.1 Prototipna kerami~na ~epna `arilna sve~ka
Kot zanimiv primer zdru`itve odli~nih mehanskih
lastnosti Si3N4 in pravih elektri~nih lastnosti dodatkov za
dosego primerne funkcionalnosti omenimo izdelavo
kerami~ne ~epne `arilne sve~ke za dizelske motorje. Ti
motorji nimajo klasi~nih v`ignih sve~k, vseeno pa ta
vozila potrebujejo `arilno sve~ko, ki omogo~a zagon in
bolj{e delovanje hladnega motorja. V preteklosti so se v
ta namen uporabljale kovinske `arilne sve~ke, njihovi
slabosti pa sta dolg odzivni ~as in omejena trajnost
uporabe. Zaradi vi{jih ekolo{kih zahtev naj bi nove
`arilne sve~ke hitreje dosegale temperaturo, potrebno za
v`ig goriva. Doba trajanja take sve~ke naj bi bila
primerljiva s trajnostno dobo dizelskega motorja. Zato so
razli~na podjetja po svetu pred okoli 20 leti za~ela
razvijati kerami~no ~epno `arilno sve~ko (K^@S).10
Tak{na sve~ka naj bi izpolnjevala naslednje zahteve:
temperaturo 1000 °C naj bi dosegla v manj kot dveh
sekundah z za~etnim tokom manj{im od 20 A, delovna
temperatura sve~ke naj bi bila okrog 1300 °C, kar bi
zagotavljala tok manj{i od 8 A in napetost okrog 8 V.
Silicijev nitrid (Si3N4) je med najprimernej{imi materiali
za elektri~no neprevodno kerami~no komponento, kot
prevodno komponento pa najpogosteje uporabljajo
volframov karbid (WC) in molibdenov disilicid
(MoSi2).11 Na svetu je ve~ tiso~ patentov v zvezi z `aril-
no sve~ko in od tega ve~ sto, ki se nana{ajo na kera-
mi~no `arilno sve~ko. Po {tevilu prijavljenih patentov
prednja~ita japonski NGK in podjetje Bosch, ki je prve
patente s tega podro~ja prijavilo `e pred petnajstimi leti.
Tudi na IJS posku{amo sodelavci na keramiki zaradi
podpore doma~i industriji razviti prototip kerami~ne
sve~ke na osnovi silicijevega nitrida in molibdenovega
disilicida.
Si3N4/MoSi2-kompoziti so bili pripravljeni iz nasled-
njih kupljenih prahov: Si3N4, MoSi2, Al2O3 in Y2O3.
Tabela 2 prikazuje njihove masne dele`e. Dodatka Y2O3
in Al2O3 izbolj{ata sintranje Si3N4 zaradi nastanka pre-
hodne teko~e faze med sintranjem.
Pravilna izbira vrste in koli~ine oksidnih dodatkov ter
pogojev sintranja, to je temperature, tlaka in atmosfere,
mo~no vpliva na kon~no mikrostrukturo, lastnosti in
A. MAGLICA ET AL.: KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA NITRIDA
Materiali in tehnologije / Materials and technology 43 (2009) 3, 165–169 167
obstojnost produkta. ^eprav ima kerami~ni prah Si3N4
strukturo α, prevladuje po sintranju β-Si3N4, ki ima
ustreznej{e lastnosti kot modifikacija α (posebno bist-
veno druga~no toplotno prevodnost). V tabeli 3 sta
navedena upogibna trdnost in elektri~na prevodnost
kompozitov.
Tabela 2: Sestava kompozita Si3N4/MoSi2, izdelanega na IJS: masni
dele`i sestavin so podani v odstotkih
Table 2: Chemical composition of Si3N4/MoSi2 composite, fabricated
at the JSI: weight portions are given in percentages
Oznaka Si3N4 MoSi2 Y2O3 Al2O3
1 61,2 32,0 3,9 2,9
2 49,9 44,6 3,1 2,4
3 44,7 50,3 2,8 2,2
Tabela 3: Primerjava upogibne trdnosti σup in elektri~ne prevodnosti
σel sintranih kompozitov glede na vsebnost MoSi2
Table 3: Comparison of the bend strength σb and electrical con-
ductivity σel of sintered composites in dependence on a portion of
MoSi2
Oznaka σup/MPa σel/(S/m)
1 480 8,5 × 10–9
2 350 7,15
3 270 306
Po pri~akovanju je specifi~na elektri~na prevodnost
vzorcev mo~no odvisna od koli~ine prevodne kompo-
nente MoSi2, saj je sam Si3N4 elektri~ni izolator. Sestave
nekako med tistima z oznakama 2 in 3 imajo primerno
elektri~no prevodnost, ki naj bi se pri kerami~nih grelcih
in ~epnih sve~kah gibala v obmo~ju 1–100 S/m. Vendar
pa upogibna trdnost materiala hitro pada z vsebnostjo
MoSi2, sve~ke pa morajo imeti zelo visoko trdnost, da
vzdr`ijo termi~ni {ok med delovanjem. Vrednosti trd-
nosti, podane v tabeli 3, so prenizke. Zato je treba iskati
re{itve, npr. spreminjati {e druge parametre poleg dele`a
MoSi2. Ena od mo`nosti bi bila naslednja. Podatki za
izmerjeno elektri~no prevodnost v tabeli 3 so za dolo-
~eno povpre~no velikost delcev MoSi2 v prahu (≈2 µm).
Z manj{o velikostjo delcev MoSi2, ki bi jo dosegli s
temeljitim mletjem, se da pri enakih masnih dele`ih
MoSi2 pove~ati prevodnost, ne da bi s tem poslab{ali
upogibno trdnost kompozita. Kljub omenjenim te`avam
nam je na Odseku za in`enirsko keramiko uspelo izdelati
prototipno ~epno sve~ko (slika 5).
Bralec lahko najde nekaj informacij o komercialnih
~epnih sve~kah, izdelanih v svetu, na spletnih straneh.12
2.2 Priprava SiAlON-ov
Priprava sialonske keramike je v primerjavi s
silicijevim nitridom v splo{nem enostavnej{a zaradi
raznovrstne me{anice prahov, kot so Si3N4, AlN, Al2O3
in Y2O3, ki olaj{ajo sintranje. Le-to je omejeno s tempe-
raturo kemijskega razpada silicijevega nitrida, 1900 °C.
Na IJS smo izdelali sialonsko keramiko s sintranjem pri
temperaturi 1800 °C za 2 uri v pretoku du{ika in pri
navadnem tlaku, in sicer v zasipu iz me{anice prahov
Si3N4 in BN v masnem razmerju 1 : 1. Glede na dose-
`eno gostoto (97 % teoreti~ne gostote) in nastalo mikro-
strukturo se je najbolje zgostila tista sialonska keramika,
pri kateri smo dosegli zadovoljive vrednosti upogibne
trdnosti (750 MPa).
Zanimiva mo`nost je tudi priprava mnogoplastnih
kompozitov, kjer imajo posamezne plasti razli~ne dele`e
α- in β-sialona. Z ulivanjem α- in β-suspenzij sialona
(prostorninski dele` suhe snovi 40 %) v mav~ne modele
smo pripravili valjaste in ravne kompozite. Ulite vzorce
smo nato sintrali pri 1800 °C 2 h v du{ikovi atmosferi.
Tako smo pripravili relativno goste mnogoplastne kom-
pozite. Za dosego njihovih ustreznih mehanskih lastnosti
je treba upo{tevati ve~ faktorjev, na primer reolo{ke
lastnosti suspenzij prahov, optimalne parametre sintra-
nja, razli~ne temperaturne razteznostne koeficiente
posameznih plasti v odvisnosti od njihove sestave itd. Na
primer, zaradi razli~nega temperaturnega raztezanja
posameznih plasti pri ohlajanju do sobne temperature po
sintranju lahko pride do velikih termi~nih preostalih
napetosti v materialu, ki povzro~ijo nastanek razpok in
celo razplastitev. Podobna te`ava se lahko pojavi tudi pri
su{enju suspenzij v kalupu zaradi razli~nih hitrosti su{e-
nja posameznih plasti. Ravne mnogoplastne kompozite
smo ulivali v pravokoten model plast za plastjo, pri
~emer smo pred ulivanjem nove plasti nekoliko po~akali,
da se je vrhnja plast delno posu{ila. Izmerili smo upo-
gibne trdnosti teh kompozitov, ki jih `elimo optimizirati.
3 RAZPRAVA
Izkazalo se je, da se pri sintranju tako Si3N4 kot
sialon zgostita do relativno visoke gostote. Pomembni
sta predvsem vrsta in koli~ina dodatkov za sintranje, pa
tudi tehnika sintranja. Sialonska keramika se v primer-
javi s Si3N4 pri sicer enakih pogojih sintranja bolj zgosti
zaradi dodanega AlN (npr. 2,2 %).
A. MAGLICA ET AL.: KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA NITRIDA
168 Materiali in tehnologije / Materials and technology 43 (2009) 3, 165–169
Slika 5: Delovanje prototipne kerami~ne `arilne sve~ke, izdelane na
na{em odseku
Figure 5: Functioning of the prototype ceramic glow plug fabricated
in our department
Pri del~nih kompozitih na osnovi Si3N4 z dispergi-
ranimi prevodnimi delci so mehanske in elektri~ne
lastnosti odvisne od dodanega dele`a prevodne kompo-
nente, predvsem od tega, ali ta dele` prese`e perkola-
cijski prag prevodnosti. V splo{nem velja, da z ve~jim
dele`em prevodne komponente pove~amo elektri~no
prevodnost kon~nega materiala, vendar pa s tem poslab-
{amo mehanske lastnosti, zato moramo najti kompromis
med temi lastnostmi. Pregled znanstvenih publikacij s
podro~ja elektri~no prevodnih kerami~nih kompozitov
na osnovi silicijevega nitrida je pokazal, da lahko
ustrezno elektri~no prevodnost materiala dose`emo z
majhnim dodatkom prevodne komponente v za~etno
me{anico prahu, vendar pa morajo ti delci tvoriti elek-
tri~no prevodno mre`o okrog ve~jih zrn Si3N413. Izdelava
tak{nih kompozitov zahteva uporabo granularnega prahu
Si3N4, ki mora biti oble~en s prevodnimi nanodelci, in
uporabo druge tehnike sintranja, ki dopu{~a sintranje pri
ni`jih temperaturah ob povi{anem tlaku. Tak{na izbira
oz. priprava izhodnih surovin in tehnike sintranja
omogo~a, da je material elektri~no prevoden `e pri
volumenskem dele`u prevodne faze 5 %. S tem
dose`emo ustrezne elektri~ne in mehanske lastnosti za
izdelavo kerami~nega grelca ali kerami~ne ~epne `arilne
sve~ke za dizelske motorje.14,15
Iz raziskav mnogoplastnih sialonskih kompozitov je
razvidno, da lahko z upo{tevanjem klju~nih parametrov
pri pripravi suspenzij in optimizaciji pogojev sintranja
dose`emo goste kompozite z dobrimi mehanskimi
lastnostmi. V literaturi najdemo veliko eksperimentalnih
in teoreti~nih raziskav tudi na drugih mnogoplastnih
kompozitih, npr. iz Al2O3 in ZrO2. Bile so usmerjene
tako v primerne metode za njihovo izdelavo kot v
optimizacijo efektivne upogibne trdnosti in lomne
`ilavosti.16–22 Optimizacijo upogibne trdnosti je mogo~e
dose~i s sestavo, razporeditvijo in debelino posameznih
plasti. Te raziskave so lahko vodilo tudi za optimizacijo
mnogoplastnih sialonov. V prihodnosti utegnejo kompo-
zitom na osnovi sialonske keramike zaradi dobrih
mehanskih in fizikalnih lastnost nameniti ve~jo pozor-
nost na podro~ju in`enirske keramike.
4 SKLEP
Na podro~ju kompozitne keramike Si3N4 smo se
ukvarjali predvsem s prevodno keramiko z disper-
giranimi MoSi2-delci in s sialoni. Obe vrsti materiala je
treba {e optimizirati: 1) pri keramiki Si3N4/MoSi2
moramo izbolj{ati prevodnost, ne da bi bistveno po-
slab{ali mehanske lastnosti; 2) pri mnogoplastnih
kompozitih na osnovi α- in β-sialonov moramo odpraviti
nekatere te`ave, kot je razplastitev. Dosedanje raziskave
pa so spodbudne.
5 LITERATURA
1 P. F. Becher, M. J. Hoffmann, Influence of additives on anisotropic
grain growth in silicon nitride ceramics, Material Science and
Engineering, A422 (2006), 85–91
2 Spletni naslov: http://en.wikipedia.org/wiki/silicon_nitride
3 Frank L. Riley, Silicon nitride and related materials, Journal of the
American Ceramic Society, 83 (2000) 2, 245–65
4 Calis N., Kara F., Mandal H., Functionally graded SiAlON ceramics,
Journal of the European Ceramic Society, 24 (2004), 3387–3393
5 Ekstrom T., Nygren A, SiAlON Ceramics, Journal of American
Ceramic Society, 75 (1992) 2, 259–276
6 L. K. L. Falk, Imaging and microanalysis of liquid phase sintered
silicon-based ceramic microstructures, Journal of Materials Science,
39 (2004), 6655–6673
7 Pernilla Pettersson, Zhijian Shen, Mats Johnsson, Thermal shock
properties of β-sialon ceramics, Journal of the European Ceramic
Society, 22 (2002), 1357–1365
8 Izhevskiy V. A., Genova L. A., Aldinger F., Progress in SiAlON
ceramics, Journal of the European Ceramic Society, 20 (2000),
2275-2295
9 Erhan Ayas, Alpagut Kara, Hasan Mandal, Production of α-β
SiAlON-TiN/TiCN composites by gas pressure sintering, Sil. Ind.
Special Issue, 69, Turkey, 2005
10 H Kawamura, Thermos construction made of Si3N4 in the heat
isulated engine, Key engineering materials, 89 (1994), 713–718
11 K. Yamada, N. Kamiya, High temperature mechanical properties of
Si3N4-MoSi2 and Si3N4-SiC composites with network structures of
second phases, Mat. Sci. Eng. A-Struct. Mat., Properties, micro-
structure and Processing, 261 (1999), 270–277
12 Spletni naslov:
http://www.le-mark.cn/E-pages/product/glow%20plug.asp
13 E. Ayas, A. Kara, F. Kara, A novel approach for preparing electri-
cally conductive α/β SiAlON-TiN composites by spark plasma
sintering, J. Ceram. Soc. Jap., 116 (2008) 7, 812–814
14 S. Kawano, J. Takahashi, S. Shimada, Highly electroconductive
TiN/Si3N4 composite ceramics fabricated by spark plasma sintering
of Si3N4 particles with a nano-sized TiN coating, J. Mater. Chem., 12
(2002), 361–365
15 L. Gao, J. Li, T. Kusunose, K. Niihara, Preparation and properties of
TiN-Si3N4 composites, J. Eu. Ceram. Soc., 24 (2004), 381–386
16 J. Requena, R. Moreno, J. S. Moya, Alumina and alumina/ zirconia
multilayer composites obtained by slip casting, J. Am. Ceram. Soc.,
72 (1989) 8, 1511–13
17 P. Sarkar, X. Huang, P. S. Nicholson, Structural ceramic micro-
laminates by electrophoretic deposition, J. Am. Ceram. Soc., 75
(1992) 12, 2907–09
18 C. J. Russo, M. P. Harmer, H. M. Chan, G. A. Miller, Design of a
laminated ceramic composite for improved strength and toughness, J.
Am. Ceram. Soc., 75 (1992) 12, 3396–400
19 D. D. Barnett-Ritcey, P. S. Nicholson, Failure prediction maps for a
model Al2O3⏐c-ZrO2/Al2O3⏐Al2O3 brittle polycrystalline trilayer
composite, J. Am. Ceram. Soc., 86 (2003) 1, 121–28
20 C.-H. Hsueh, Modeling of elastic deformation of multilayers due to
residual stresses and external bending, J. Appl. Phys., 91 (2002) 12,
9652–56
21 Ambro`i~ M., Kosma~ T., Optimization of the bend strength of
flat-layered alumina-zirconia composites, J. Amer. Ceram. Soc. 90
(2007) 5, 1545–1550
22 Berani~ Klop~i~ S., Ambro`i~ M., Kosma~ T., Novak S., Tunneling
cracks in Al2O3/Al2O3-ZrO2 layered composites, J. Eur. Ceram. Soc.
27 (2007), 1333–1337
A. MAGLICA ET AL.: KERAMI^NI KOMPOZITI NA OSNOVI SILICIJEVEGA NITRIDA
Materiali in tehnologije / Materials and technology 43 (2009) 3, 165–169 169

B. CHOKKALINGAM, S. S. MOHAMED NAZIRUDEEN: INVESTIGATIONS OF MICRO-ALLOYED CAST STEELS
INVESTIGATIONS OF MICRO-ALLOYED CAST STEELS
RAZISKAVE MIKROLEGIRANIH JEKLENIH LITIN
Bommannan Chokkalingam, S. S. Mohamed Nazirudeen
Department of Metallurgical Engineering, PSG College of Technology, Peelamedu, Coimbatore-641 004. India
bchokkalingam gmail.com
Prejem rokopisa – received: 2008-11-06; sprejem za objavo – accepted for publication: 2009-03-23
This paper presents a review of the effect of micro-alloying with vanadium, niobium, titanium and zirconium in carbon and
low-alloy cast steels of different compositions. The elements were added in different percentages and combinations to the base
cast steels. Also, the effect of heat treatment on the experimental steels is presented. The micro-alloying improved the
mechanical properties of the as-cast and heat-treated base steels significantly.
Key words: cast steel, chemical composition, micro-alloying, mechanical properties, heat treatment
Predstavljeni so rezultati raziskav o vplivu mikrolegiranja z vanadijem, niobijem, titanom in cirkonijem na mehanske lastnosti
in trdoto ogljikove in malolegirane jeklene litine z razli~no sestavo. Mikrolegirni elementi so bili dodani osnovnemu jeklu
individualno in v razli~nih kombinacijah. Mikrolegiranje je mo~no vplivalo na lastnosti tega jekla v litem stanju in po toplotni
obdelavi.
Klju~ne besede: jeklene litine, kemi~na sestava, mikrolegiranje, toplotna obdelava, mehanske lastnosti
1 INTRODUCTION
Micro-alloyed steels are low, medium carbon steels
or low-alloy steels with the addition of elements such as
niobium, titanium, vanadium and zirconium. These
additions increase the strength of the steel, individually,
and also in different combinations. Initially, micro-
alloyed wrought steels were produced with micro-
alloying and with thermo-mechanical processing (heat
treatment) for higher strength as well as toughness and
good weldability. Similar improvements are required for
steels in the as-cast condition, especially carbon and
low-alloy steels that are used for many applications.
Moreover, these castings are frequently used in extreme
conditions of tensile and fatigue stressing, impact
loading, creep, abrasion and corrosion. Generally, cast
steels have the major advantage of wider ranges of
composition, higher casting temperatures and fewer
impurities over wrought steels. To meet the present
critical applications of components, cast steels with
higher strengths and toughnesses are used. In this paper,
the effect of the micro-alloying elements on the pro-
perties of the cast steels is investigated.
2 TYPES OF ALLOY STEELS
Steels containing one or more alloying elements,
other than carbon, are designed to enhance one or more
of the steels’ properties. Specifically, alloy steels are
superior to plain carbon steels due to the effect of the
alloying elements. All alloy steels can be classed into
three groups with respect to the total content of alloying
elements: low-alloy steels1,2 with up to 5 %, medium-
alloy steels1 with 5 % to 10 % and high-alloy steels with
more than 10 %.
In micro-alloy steel, minor amounts of vanadium,
niobium and/or titanium, etc., are added2,3. The addition
of these individual elements is generally less than 0.10 %
and the total micro-alloying is less than 0.20 %.
3 BASE STEEL AND MICRO-ALLOYING
The chemical compositions of the base steels used
for the micro-alloying experiments are presented in
Table 1. The micro-alloying was achieved with the
addition of vanadium, niobium, titanium and zirconium
to the base steels. The contents of the micro-alloying
elements are given in Table 2.
4 MECHANICAL PROPERTIES
The maximum yield and the ultimate tensile
strengths3–7 of the experimental steels, the base (UMA)
and the micro-alloyed (MA), are shown in Figure 1 and
Figure 2. After micro-alloying (MA), the yield and
tensile strengths3 were increased by 110 % and 100 %,
respectively, when compared to the un-micro-alloyed
steels. The experimental steels3 were quenched and
tempered at various temperatures. A low carbon content
was maintained in all the experimental steels to achieve
sufficient ductility. The minimum yield strength and the
ultimate tensile strength were obtained for 0.003 V and
0.001 Ti after tempering at 700 °C. In this steel the
highest yield strength and the ultimate tensile strength
were obtained with combined tempering at temperatures
of 700 °C and 400 °C.
Materiali in tehnologije / Materials and technology 43 (2009) 3, 171–174 171
UDK 669.15:669.14.018.298 ISSN 1580-2949
Professional article/Strokovni ~lanek MTAEC9, 43(3)171(2009)
The mechanical properties for the micro-alloyed steel
5 and the corresponding base steel were tested after nor-
malizing (920 °C), quenching (920 °C) and tempering
(580 °C) as well as homogenization (1100 °C) treat-
ments4. Due to the micro-alloying, the yield strength was
increased by 118 % and the tensile strength by 65 %,
respectively, when compared to the un-micro-alloyed
steels. The minimum and maximum yield and tensile
strengths5 were obtained in steel 5 with the addition of
0.09 vanadium after the normalizing and homogenizing
treatments.
The mechanical properties5,6 were tested in the
as-cast condition without any heat treatment. After the
micro-alloying the yield and tensile strengths were
increased by 55 % and 26 % with respect to the
un-micro-alloyed steels. The minimum yield and the
tensile strengths were obtained in the un-micro-alloyed
steel 9 without any heat treatment. The addition of
B. CHOKKALINGAM, S. S. MOHAMED NAZIRUDEEN: INVESTIGATIONS OF MICRO-ALLOYED CAST STEELS
172 Materiali in tehnologije / Materials and technology 43 (2009) 3, 171–174
Table 1: Summary of chemical compositions of the base cast steels
Tabela 1: Pregled kemi~ne sestave osnovnih in mikrolegiranih jekel
Ref
No.
Experimental
Steel
Elements in wt %
C Si Mn P S Cr Mo Ni Al
3
Steel 1 (UMA)
Steel 1 (MA)
Steel 2 (MA)
Steel 3 (MA)
Steel 4 (MA)
0.23
0.23
0.18
0.16
0.15
0.36
0.36
0.30
0.32
0.34
0.88
0.88
0.85
0.83
0.82
0.04
0.04
0.03
0.02
0.04
0.04
0.04
0.03
0.03
0.03
0.39
0.39
0.40
0.40
0.40
0.10
0.10
0.10
0.10
0.10
–
–
–
–
–
0.004
0.004
0.012
0.015
0.056
4
Steel 5 (UMA)
Steel 5 (MA)
Steel 6(MA)
0.21
0.21
0.20
0.39
0.39
0.52
1.13
1.13
1.24
–
–
–
–
–
–
0.11
0.11
0.41
0.18
0.18
0.29
0.91
0.91
0.76
0.035
0.035
0.007
5,6
Steel 7(UMA)
Steel 8(MA)
Steel 9(UMA)
Steel 9(MA)
Steel 10(MA)
0.14
0.15
0.14
0.14
0.14
0.35
0.31
0.30
0.30
0.32
1.56
1.41
1.54
1.54
1.55
0.010
0.010
0.009
0.009
0.009
0.015
0.015
0.02
0.02
0.016
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.03
0.02
0.03
0.03
0.02
7
Steel 11(UMA)
Steel 12(MA)
Steel 13(UMA)
Steel 13(MA)
Steel 14(MA)
Steel 15(MA)
0.06
0.06
0.06
0.06
0.07
0.07
0.14
0.14
0.09
0.09
0.09
0.09
1.60
1.80
1.85
1.85
1.85
1.90
0.03
0.03
0.01
0.01
0.03
0.01
0.07
0.08
0.10
0.10
0.07
0.09
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.05
0.05
0.01
0.01
0.04
0.03
Table 2: Summary of the micro-alloying elements
Tabela 2: Pregled vsebnosti mikrolegiranih elementov
Reference Experimental Steel
Elements in wt %
Vanadium Niobium Titanium Zirconium
Ref. 3
Steel 1(UMA) – – – –
Steel 1 (MA) 0.003 – 0.001 –
Steel 2(MA) 0.11 – 0.01 –
Steel 3(MA) 0.11 0.036 0.01 –
Steel 4(MA) 0.11 0.036 0.04 –
Ref. 4
Steel 5 (UMA) – – – –
Steel 5(MA) 0.09 – – –
Steel 6(MA) 0.09 0.035 – –
Ref. 5, 6
Steel 7(UMA) – – – –
Steel 8(MA) 0.096 – – –
Steel 9(UMA) – – – –
Steel 9(MA) 0.092 0.047 – –
Steel 10(MA) 0.10 0.043 0.023 –
Ref. 7
Steel 11(UMA) – – – –
Steel 12(MA) – – – 0.005
Steel 13(UMA) – – – –
Steel 13(MA) – – – 0.01
Steel 14(MA) – – – 0.02
Steel 15(MA) – – – 0.10
vanadium individually increased both the yield strength
and ultimate tensile strength up to 100 MPa. After
micro-alloying, the steel 9, with 0.092 vanadium and
0.047 niobium, had its yield strength and ultimate tensile
strength increased by 160 MPa and by 125 MPa.
Zirconium was also added as a micro-alloying
element to the low-carbon base steel7. Four trials were
made by varying the content of zirconium in the base
steel and the mechanical properties were tested after
quenching, tempering and homogenizing. After the
addition of zirconium, the yield and tensile strengths
were increased by 120 % and 50 % with respect to the
un-micro-alloyed steel. The minimum values for the
yield and tensile strengths were found for the micro-
alloyed steel 13 after homogenization. The tempering
treatment improved both the ultimate tensile strength and
the yield strength in steel 13 with 0.01 zirconium more
than the quenching and the homogenizing.
The elongation and reduction of area obtained for the
experimental steels un-micro-alloyed and micro-alloyed
are shown in Figure 3 and Figure 4, respectively3,4,7.
The elongation and the reduction of the area after
micro-alloying3 for the steel 11 are lower by 27 % and
28 %, respectively.
The minimum and maximum elongation and reduc-
tion of area3 were found for the steel 1 after tempering at
400 °C and 700 °C, respectively.
For the steel 5, after micro-alloying the elongation
and the reduction of area were decreased by 44 % and 45
%, respectively. The minimum elongation and reduction
of area for the steel 5 were found after homogenization
and normalization and the benefit of micro-alloying was
achieved in steel 5 only after quenching and tempering at
580 °C.
After micro-alloying5,6 in steel 9, the elongation was
decreased by 36 %. The minimum elongation was found
for the micro-alloyed steel 9 and maximum for the
un-micro-alloyed steel 9.
The elongation and the reduction of area due to
micro-alloying7 in steel 13 were decreased by 40 % and
70 %, respectively, with respect to the un-micro-alloyed
steel 11. The minimum elongation and reduction of area
were found after quenching and the maximum after
tempering at 450 °C.
The hardness3,5,6,7 of the experimental steels is shown
in Figure 5. After micro-alloying, the hardness was
increased by 146 % in steel 1, 27 % in steel 9 and 130 %
in steel 13, respectively.
B. CHOKKALINGAM, S. S. MOHAMED NAZIRUDEEN: INVESTIGATIONS OF MICRO-ALLOYED CAST STEELS
Materiali in tehnologije / Materials and technology 43 (2009) 3, 171–174 173
Figure 2: Ultimate tensile strength
Slika 2: Raztr`na trdnost
Figure 4: Percentage reduction in area
Slika 4: Skr~ek v odstotkih
Figure 1: Yield strength
Slika 1: Meja plasti~nosti
Figure 6: Impact toughness
Slika 6: Udarna `ilavost
Figure 3: Percentage elongation
Slika 3: Raztezek v odstotkih
Figure 5: Hardness
Slika 5: Trdota
The minimum and maximum hardness3 were
obtained in steel 1 with 0.003 V and 0.001 Ti after
hardening and tempering at 500 °C. The minimum
hardness was found for the un-micro-alloyed steel 9 and
the maximum for the micro-alloyed steel 9 without heat
treatment5,6. The maximum hardness7 was found for the
micro-alloyed quenched steel 13 with 0.005 of zirconium
and the minimum for the un-micro-alloyed steel 13.
Generally, the hardness was higher after quenching
and tempering treatments than in the as-cast condition.
The toughness5,6,7 obtained with the micro-alloyed
and the un-micro-alloyed for the experimental steel 9
and steel 13 is shown in Figure 6. The toughness5,6,7 due
to micro alloying7 was decreased by 90% in the micro-
alloyed steel 9 without heat treatment and increased by
110% in the micro-alloyed tempered steel 13 with 0.01
zirconium.
5 MICRO-ALLOYING MECHANISMS
A fine grain size3 was achieved in micro-alloyed steel
1 compared with that of the un-micro-alloyed steel 1.
Generally, grain refinement was found in all the
micro-alloyed steels. In micro-alloyed steels the titanium
precipitates as TiN and the vanadium as VC/V(CN). Nb
precipitates from the matrix as NbC/Nb (CN) inde-
pendently, or on the existing TiN particles. The grain
refinement and precipitation hardening induce the higher
strength and hardness of the micro-alloyed steels.
In steel 5, micro-alloyed with vanadium,4 the
strengthening is achieved with a fine precipitation of VC
in the ferrite matrix and also with the fine interlamellar
spacing of pearlite. The vanadium addition in the as-cast
condition affects the grain size of the ferrite also.
The increase in yield and ultimate tensile strengths
and hardness5,6 in the micro-alloyed steel 9 is due to the
formation of a fine scale of carbonitride precipitates. The
micro-alloy carbonitrides precipitate on the advancing
alpha/gamma interface. The micro-alloying of the
niobium precipitates in the matrix or at the austenite
grain boundaries, whereas the vanadium precipitates as
V(C,N) as an interphase or random precipitation.
Zirconium is a strong nitride-7 and oxide-forming
element. In the zirconium micro-alloyed steels the ZrC
precipitates form at higher temperatures and the ZrN
precipitates at lower temperatures. The precipitation of
ZrN and the heat treatment improves the mechanical
properties of the micro-alloyed steel 13 very much with
respect to the addition of zirconium.
6 CONCLUSION
I In this article the effects of micro-alloying additions
of vanadium, niobium, titanium and zirconium in
low-carbon base steels of different compositions, as
cast and after heat treatment are presented and
discussed.
The chemical composition of the base steel used and
the individual and combined additions of micro-
alloying elements used for the experimental work are
given.
The minimum and maximum values of the mecha-
nical properties obtained in the micro-alloyed cast
steels in the quoted references are summarized and
discussed.
II Generally, the ultimate tensile strength, yield strength
and hardness are greater with micro-alloyed cast
steels, while the ductility, measured as elongation
and the reduction of area, is smaller. Also, the notch
toughness is lower after micro-alloying.
III In steels micro-alloyed with vanadium, niobium and
titanium, the tensile strength was lower in the
presence of titanium. The vanadium addition
individually increased the yield strength and the
tensile strength up to 100 MPa, whereas vanadium
and niobium additions increased the yield strength
and ultimate tensile strength by 128 MPa and 90
MPa, respectively.
IV The hardness was higher after micro-alloying the
zirconium, especially after tempering. An increased
hardness and toughness were obtained with
0.005–0.01 % of zirconium.
7 REFERENCES
1 T. V. Rajan, C. P. Sharma, Ashok Sharma , Heat treatment principles
and techniques, 1st ed., Prentice Hall of India private limited, New
delhi (2006), ISBN-81-203-0716-X, 53
2 V. Raghavan, Physical metallurgy principles and practice, 2nd ed.,
Prentice Hall of India private limited, New delhi 2006, 106, 193,
ISBN-81-203-3012-9
3 B. D. Jana, A. K. Chakrabarti., K. K. Ray, Study of cast micro-
alloyed steels, Materials science and technology, (1993), January,
80–86
4 Barbara Kalandyk, Hubert Matysiak, Jan Glownia, Microstructure
strength relationship in microalloyed cast steels, Reviews on
advanced material science, 1 (2004) 8, 44–48
5 H. Najafi, J. Rassizadehghani, A. Halvaaee, Mechanical properties of
as cast microalloyed steels containing V, Nb and Ti, Material science
and technology, 23 (2007) 6, 699–705
6 J. Rassizadehghani, H. Najafi, M. Emamy, Eslami Saeen,
Mechanical properties of V, Nb and Ti bearing as cast micro alloyed
steels, Journal of material science and technology, 23 (2007) 6,
779–784
7 H. A. Akbarzadeh, M. Tamizifar, Mirdamadi Sh, A. Abdolhossini,
Mechanical properties and micro structures of Zr – microalloyed cast
steel, ISIJ international, 45 (2005), 1201–1204
B. CHOKKALINGAM, S. S. MOHAMED NAZIRUDEEN: INVESTIGATIONS OF MICRO-ALLOYED CAST STEELS
174 Materiali in tehnologije / Materials and technology 43 (2009) 3, 171–174