Surface & Coatings Technology 201 (2007) 9541 – 9548

www.elsevier.com/locate/surfcoat

Protective coatings for induction casting of titanium

Federico Smeacetto ⁎, Milena Salvo, Monica Ferraris
Materials Science and Chemical Engineering Department, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy
Received 8 November 2006; accepted in revised form 6 April 2007
Available online 18 April 2007

Abstract

Commercially pure titanium is among the most biocompatible metals available for dental uses: it combines good mechanical properties with
high corrosion resistance in simulated body fluids. However, titanium casting shows some problems mainly related to its high melting temperature
and chemical reactivity with oxygen at T N 600 °C; as a consequence, molten titanium reacts with crucibles and mould components during casting.
In this research, new Y2O3-based slurry was used to produce yttria coated mullitic crucibles and wax pattern coatings for induction melting of
titanium; furthermore, a detailed comparison among 4 different investment materials (one silica-based, one ZrSiO4-based and two MgO-based)
was performed. The titanium surface after casting was investigated with optical and electron microscopy (SEM) and energy dispersion
spectroscopy (EDS). A reaction zone was observed on Ti when cast in an uncoated crucible, while a clean, pure Ti surface was obtained by using
yttria coated crucibles and yttria coated wax patterns before investing with silica-based and zircon-based investment.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Titanium; Slurry; Yttrium oxide; Scanning electron microscopy

1. Introduction The whole casting process (“Investment casting”) has the

advantage of reproducing complex shapes at relatively low cost,
Titanium is a very promising material especially in dentistry, together with a good scalability from a single item to a large
because of its high corrosion resistance, low density, good number of identical items; however, certain defects in the final
mechanical properties and reasonable price, compared to those of products may result from the investment casting process such as
gold alloys [1]. porosity, shrinkage and contamination due to mould–casting
There are two melting methods currently used for titanium: interactions [7].
argon/arc melting with a non-consumable tungsten electrode
and high frequency induction melting [2,3]. Table 1
Several difficulties still arise in titanium casting, since it is Description of wax pattern used in this work
very reactive at high temperatures, easily oxidised, and reactive
Crown-shaped Investment Coating on wax Mold temperature
with refractory crucibles and mould components. wax patterns material pattern (before casting) (°C)
Oxides formation during the casting process increase titanium
A1 SiO2-based – 200
hardness and brittleness; other impurities may result in inhomo- (Ticoat)
geneous microstructure, increasing the susceptibility to corrosion A2 Zircon-based – 600
or decreasing the titanium biocompatibility. The formation of (Shera)
hard, brittle reaction layers lead to a contaminated casting surface B1 SiO2-based Y2O3/colloidal silica 200
(Ticoat)
[4], which can show increased microhardness [5] and surface
B2 Zircon based Y2O3/colloidal silica 600
roughness, not compatible with dental applications [6]. (Shera)
C1 MgO-based – 600
(Titavest)
⁎ Corresponding author. C2 MgO-based – 600
(Gilvest)
E-mail address: [email protected] (F. Smeacetto).

0257-8972/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.surfcoat.2007.04.027
9542 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548

Fig. 1. Wax pattern coating and investment casting process.

Silica-based phosphate-bonded investments have been expansion coefficients and mechanical properties less matched
commonly used for investment casting of titanium and base to those of Ti than the silica-based ones.
metal alloys in dentistry. Zirconia, because of its refractoriness, and yttria, because of its
The reactivity between titanium and silica-based phosphate- refractoriness and chemical inertness, has been extensively
bonded investment materials could be reduced by using MgO- studied as wax pattern [10] and ceramic crucible coating materials
based investment materials, where silica is replaced by MgO [9]. The refractory crucibles coated by applying Y2O3-colloidal
and ZrO2, less reactive than silica towards molten Ti. The silica slurries seem to be the most effective, stable and the less
problem with MgO-based investments is that they have thermal expensive ones. In order to avoid the premature gelation of these

Fig. 2. XRD of refractory crucible.

 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548 9543

slurries and the consequent formation of poor and not-uniform were performed by: optical microscopy, SEM and EDS analysis,
coatings, slurry containing sources of hydroxyl ions was proposed micro hardness. The reaction layer and the micro hardness of the
[9]. Besides, these hydroxyl containing slurries have to be titanium castings were investigated. Furthermore, an analysis on
stabilised with an alkali and their pH must be higher than 10.2 the dimensional accuracy of the titanium castings was conducted in
after 6 days after mixing. order to evaluate the precision of the titanium specimen.
The objective of this work consists in the design and testing of
a new Y2O3-colloidal silica slurry which could be used both to 2. Experimental
protect the mullitic refractory crucibles and to coat the wax
patterns (with the aim of modifying the surface of the mould in 2.1. Crucible coating
direct contact with molten titanium), in order to avoid the titanium
reactivity during casting. This new slurry is gelation stable, can be Commercially pure titanium was cast in a centrifugal
used for a long period without pH stabilisation and tailoring. controlled atmosphere induction machine (Neutrodyn Easyti,
In order to investigate the protection effectiveness of wax Manfredi S.p.A., Italy).
pattern coating, produced with the new Y2O3 slurry, a detailed The commercially pure titanium (cp Ti) was initially melted in
study of titanium castings obtained from 4 different investment zirconia coated mullitic refractory crucible, labelled as C15
materials (one silica-based, one ZrSiO4 based and two MgO-based) (Ceramique et Refractaries, France), for comparison purposes. A

Fig. 3. SEM image of the cross section of the interface between the crucible and the as deposited protective coating. a. EDS mapping of protective coating.
9544 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548

slurry made of yttrium oxide powder (≤10 μm, Stanford the yttria coating was practically transferred to the internal
Materials, USA) and colloidal silica with a specific surface area surface of the mould. The whole process is shown in Fig. 1.
of 230 m2/g and a 25 (wt.%) silica content was prepared in water The patterns C (C1 and C2) were invested with two
solution (at different yttria/silica weight ratios) and applied by commercial MgO-based investment materials (C1: Titavest
dipping on the C15 crucible inner surface. The as prepared CB Morita, Japan and C2: Gilvest Giulini, Germany).
crucible, after drying for 12 h at 25 °C, was heat treated The melting chamber was alternatively air evacuated and
(T N 900 °C) in a conventional oven, in order to obtain a dense filled with argon in order to prevent contamination from air.
ceramic protective layer on it; the cross-section of the coated A 40 second melting time was employed to melt the titanium
crucible was examined by optical and scanning electron (18 g) in the crucible; then, the liquid metal was transferred by a
microscopy before and after the contact with molten titanium. centrifugal force inside the mould (at a mold temperature related
to the used investment) (Table 1).
2.2. Investment materials and wax pattern coating After the casting procedures, the obtained titanium samples
were mechanically removed from the investment materials.
Three series of crown-shaped wax patterns were prepared the The Ti samples obtained from wax patterns A, B and C were
patterns A (A1 and A2) were directly invested with two prepared for metallographic observations of the cross sections,
commercial SiO2-based and ZrSiO4-based, phosphate-bonded in order to study the contamination layer depth on the titanium
investment materials (A1: Ticoat powder. (Manfredi) mixed surface. The samples were embedded in an epoxy resin, grinded
with special liquid (Manfredi), F.lli Manfredi S.p.A, Italy;. A2: using SiC papers from 320 to 4000 grit, cleaned in ultrasonic
SheraTitan-easy mixed with SHERALIQUID, (Shera Werkstoff bath for 5 min and etched with a 2% HF solution. The cross-
Technologie, Germany). The patterns B (B1 and B2) where sections were analysed by an optical microscope and with a
dipped in a slurry prepared by mixing yttrium oxide with SEM microscope equipped with energy dispersion spectrometry
colloidal silica (as described in Section 2.1) and subsequently system (EDS) and SEM JEOL JXA 8600 equipped with EPMA
dried in air for 2 h, before being invested with the same WDS Noran, controlled by Software “Vantage”, in order to
investment materials used for patterns of A series. (Table 1). detect the presence of elements, new phases or contamination
After the removal of the wax (during the burn-out procedure) zones formed during casting operations.

Fig. 4. Titanium slag after 18 g. melting in yttria coated crucible. b. Titanium slag after 18 g. melting in zirconia coated crucible. c. Cross section of yttria coated
crucible after 18 g titanium melting. d. Cross section of zirconia coated crucible after 18 g titanium melting.
 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548 9545

Fig. 5. Cross section and EDS analysis of titanium casting deriving from A1 wax pattern.

Micro hardness tests (Vickers) were performed on the molten titanium and the crucible (Fig. 4b). the titanium slag
titanium samples to evaluate the degree of contamination of the from yttria coated crucible is grey and ductile as the pure
titanium casting surfaces. titanium (Fig. 4a).
After the melting, the yttria coated crucible, is still 350 μ thick
3. Results and discussion and no degradation or reaction zone are observed (Fig. 4c); the
zirconia coated crucible shows a wide reaction zone (Fig. 4d)
3.1. Ceramic coating for crucibles; characterization and together with degradation in the coating thickness.
testing

The X-Rays diffraction pattern of the refractory crucible,

C15, revealed the presence of quartz, mullite and a residual
glassy phase, as reported in Fig. 2.
Many studies have been carried out on the reactivity of
titanium with a variety of refractory materials (borides,
carbides, nitrides and oxides): oxides are considered the most
suitable, taking in account the free energy data and solubility
effects of elements in liquid titanium. Yttrium oxide resulted to
be the most stable oxide [8–10].
For this reason, an yttrium oxide/colloidal silica slurry was
prepared and applied on the C15 crucible inner surface, heated
at temperature above 900 °C to obtain a dense ceramic
protective layer on it; the cross-section of the coated crucible
was examined by scanning electron microscopy: Fig. 3 shows
the cross section of the interface between the crucible and the as
deposited protective coating; the interface is continuous,
without pores and cracks. The average thickness of the
protective coating is 300–350 μm.
The EDS mapping on the protective coating (Fig. 3a) shows the
presence of yttrium oxide together with a small amount of silica
coming from the colloidal silica used for the slurry preparation.
In order to compare the protective effect of the yttria based
coating toward the molten titanium, an 18 g. titanium-ingot was
melted both in a commercial zirconia coated crucible and in the
yttria-coated crucible. The comparison between the titanium
slags and the cross sections of the two crucibles are reported in
Fig. 4a–d. It can be observed the different colour of the two
titanium slags; the titanium slag obtained with the zirconia
coated crucible is very brittle and its surface changes colour Fig. 6. a. Alpha-case layer (30 μ) on the titanium casting deriving from A1 wax
from blue to yellow, while the part in contact with the coated pattern. b. Alpha-case layer (30 μ) on the titanium casting deriving from A2 wax
crucible is black; this indicates a severe reaction between pattern.
9546 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548

Fig. 7. Cross section and EDS analysis of titanium casting coming from B1 wax patterns. a. Optical micrograph cross section of castings coming from wax pattern B1.

EDS measurements conducted on some titanium slags collected The hardness of this oxygen rich layer, known as alpha-case,
at the bottom of the yttria coated crucible revealed only the presence shows increased microhardness. This contaminated zone is
of titanium, confirming that no reaction and consequent contam- undesirable because reduces the surface quality of the castings
ination of titanium with the Y2O3 coated crucible took place. [11].
Furthermore, the efficiency of the Y2O3 based coating to prevent Fig. 6a, shows the alpha-case layer (30 μm) on the titanium
titanium oxidation and crucible damaging was demonstrated by casting deriving from A1 wax pattern; the same layer (30 μm)
using the same coated crucible up to five melting runs of 18 g each. was observed on the titanium casting deriving from A2 wax
pattern (Fig. 6b). The edge interface is characterized by three
3.2. Titanium castings analysis layers with a distinct appearance in the Fig. 6a; a shiny outer
layer (1), a second layer (2) with a globular structure and a third
The same Y2O3 based coating was proven to be effective layer (3) with a dendritic structure.
also for coating the wax patterns.
A macroscopic difference between titanium castings was
noticed between samples deriving from wax patterns of series A
and B; the former (A) showed black surfaces, while the latter
(B) were grey. The colour on the titanium castings surface also
reflected the degree of reaction between titanium and
investment during casting; a black surface indicates a more
severe reaction with the mould.
Fig. 5 shows a cross section of titanium casting deriving
from A1 wax pattern: EDS on the indicated area which was in
direct contact with the silica-based investment (Ticoat) showed
the presence of oxygen and silicon which indicates the
formation of a contamination layer during casting (thickness
about 50 μm). The oxygen liberated from the oxides diffuses
into the surface of the titanium casting to form a hardened layer. Fig. 8. Cross section of titanium casting coming from B2 wax patterns.
 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548 9547

Fig. 7 shows a SEM cross section of titanium casting coming

from B1 wax patterns; in this case EDS on the edge of the specimen
revealed only the presence of titanium; the alpha-case is absent (as
detected in the same zone by optical microscope, see Fig. 7a).
Yttrium oxide could not be reduced by titanium and played the role
of a barrier between the cast titanium and the investment, as
predicted by Ellingham diagrams, reported in Okabe et al [2], since
yttria is much more chemically stable than titanium oxide. On the
contrary, silica-based and ZrSiO4-based investments exhibited a
high reactivity with molten titanium determining an extensive
contamination (alpha-case layer) on the titanium castings.
There is also an important difference in surface roughness
between samples of Figs. 5 and 7; in the former samples a high
reactivity between titanium and investment occurred and, as a Fig. 10. Optical micrograph cross section of castings coming from wax pattern
consequence, a rougher surface was formed [12], while in the C2.
latter case, where there was the protective yttria based coating, i.e.
no direct contact between titanium and silica based investment,
the contamination was avoided and a smoother titanium surface Finally, the cross sections of all titanium castings were
was obtained. However, no formations of undesirable internal submitted to Vickers indentation. The results of Vickers
porosity or incomplete casting were observed in the titanium hardness (VH) changes in titanium castings from outer to
castings [13]. inner surface are shown in Fig. 11. It can be observed that the
A cross section of titanium casting coming from B2 wax best results were obtained with samples of series B1 and B2
pattern (Fig. 8) was also submitted to EDS analysis. EDS did were no alpha-case was detected; the behaviour of samples of
not reveal the presence of alpha-case layer and the yttria series C1 and C2 is similar to A1 and A2 ones, and the alpha
protective layer contains only Si, Y, and O. No Y was detected case layer is between 15 and 30 μm. The alpha-case layer
on the titanium casting surface layer and yttria coating was not resulted from contamination with oxygen, which increased the
detached from the walls of mold by liquid titanium during hardness. The Vickers hardness results confirmed the optical
casting. It can be concluded that no diffusion of elements microscopy observations conducted on the titanium castings in
occurred during casting; furthermore, the yttria protective layer order to reveal the presence of alpha case layer; the best results
was easily removed from titanium surface by ultrasonic were obtained with B2 series (zircon based investment), and B1
cleaning. (silica based investment). Concerning the use of MgO-based
Figs. 9 and 10 show the optical micrographs of the casting investments, the reactivity with titanium is reduced, but not
cross sections coming from wax pattern C1, and C2 (Table 1). In completely, and an alpha-case layer (average thickness 20 μm)
both these cases, the alpha case layer is about 15 μm. MgO- is present.
based investment materials showed a lower reactivity with The analysis on the dimensional accuracy of the titanium
molten titanium than silica-based and ZrSiO4-based invest- castings was conducted comparing the section width of the
ments as predicted by the standard free energy of formation for titanium castings obtained from the crown shaped wax patterns
MgO compared with titanium oxide [2]. No contamination of with a standard stainless steel reference. The dimensional
titanium castings were only obtained by using yttria coated accuracy is expressed as a percentage [14,15]. The patterns of
invested materials discussed in this paper (B1 and B2, in
Table 1).

Fig. 11. Vickers indentation on the cross sections of all titanium castings.
Fig. 9. Optical micrograph cross section of castings coming from wax pattern (a): series A1, Ticoat; (b): series A2, Shera; (c) series C1 and C2, Titavest and
C1. Gilvest; (d) series B1 (dipped); (e) series B2 (dipped).
9548 F. Smeacetto et al. / Surface & Coatings Technology 201 (2007) 9541–9548

series B1 and B2 gave a good fit with the stainless steel for its financial support and Manfredi S.p.A for the melting and
reference sample, while the pattern of series C1 and C2 gave an casting experiments and to Dr. Davide Giordano for his precious
undersized value. experimental preliminary work and to Dr. S. Spriano for her
helpful advice.
4. Conclusions
References
The proposed yttrium oxide based ceramic coating was
demonstrated to be effective in protecting titanium castings [1] M. Taira, J.B. Moser, E.H. Greener, Dent. Mater. 5 (1989) 45.
from contamination and oxidation. [2] T. Okabe, C. Ohkubo, I. Watanabe, O. Okuno, Y. Takada, JOM 50 (1988)
24.
New yttria based slurry has been successfully used to [3] H. Shimizu, T. Habu, Y. Takada, K. Watanabe, O. Okuno, T. Okabe,
produce yttria coated mullitic crucibles and as wax pattern Biomaterials 23 (2002) 2275.
coatings in a centrifugal controlled atmosphere induction [4] A.M. Ferenczi, B. Demri, M. Moritz, D. Muster, Biomaterials 19 (1998)
machine for titanium casting. 1513.
High quality titanium melting was conducted in the new [5] I. Watanabe, J.H. Watkins, H. Nakajima, M. Atsuta, T. Okabe, J. Dent. Res. 3
(1997) 773.
yttria coated crucibles. Furthermore, the protection effective- [6] X.P. Luo, T.W. Guo, Y.G. Ou, Q. Liu, Dent Mater. 18 (2002) 512.
ness of wax pattern coating, produced with the new Y2O3 slurry, [7] R.C. Atwood, P.D. Lee, R.V. Curtis, Dent. Mater. 21 (2005) 178.
was studied on 2 different investment materials (one silica- [8] R.L. Saha, T.K. Nandy, R.D.K. Misra, K.T. Jacob, Metall. Trans. B:
based, one ZrSiO4 based) and two MgO-based for comparison Process:Metalurggy 21B (1990) 559.
purpose: a contamination layer of about 30 μm on the titanium [9] Horton RA, Method of casting a reactive metal against a surface formed
from an improved slurry containing yttria. US Patent n°4947927, 1990.
casting deriving from silica-based and zircon-based invest- [10] M. Koike, Z. Cai, H. Fujii, M. Brezner, T. Okabe, Biomaterials 24 (2003)
ments, while a contamination layer of about 15 μm was 4541.
observed with MgO-based investments. [11] Si-Young Sung, Young-Jig Kim, Mater. Sci. Eng. A 405 (2005) 173.
No alpha-case was detected in titanium castings obtained by [12] T. Papadopoulos, S. Zinelis, M. Vardavoulias, J. Mater,. Sci. 34 (1999)
3639.
using yttria coated wax patterns with zircon or silica based
[13] D. Eliopoulos, S. Zinelis, T. Papadopoulos, J. Prosthet. Dent. 92 (4) (2004)
investments. 377.
[14] C.C. Hung, G.-L. Hou, C.C. Tsai, C.C. Huang, Dent. Mater. 20 (2004) 846.
Acknowledgements [15] D. Low, Dent. Mater. 15 (1999) 185.

The authors are grateful to PROVIN project and Consortium

for Research and Permanent Education Politecnico di Torino,