(Received March 10, 1998) (Accepted in revised form August 3, 1998)
Introduction
Titanium alloys have excellent properties such as high specific strength and good corrosion resistance. However, they are extremely reactive to ceramics at high temperatures, resulting in a chemical reaction affected-surface [1]. The interstitial elements (e.g., C, N, O, H) from the ceramic have a great tendency to enter into the titanium alloys and cause the deterioration of mechanical properties [2]. Economos and Kingery [3] found that titanium melt could penetrate along the grain boundaries of oxides, such as Al2O3, ZrO2, and MgO, at 1800°C, leading to the alternation of these oxides. Ruh [4] indicated that the
titanium could react with zirconia and up to 10 at% of Zr and O elements from ZrO2were retained in
the solid solution of Ti, while the zirconia was transformed into oxygen deficient zirconia. Saha [5] also revealed that ZrO22x anda-Ti(O) were formed after Ti reacted with ZrO2. However, the interfacial
reactions between zirconia and titanium has not been fully elucidated to date.
In this study, the microstructure of the interface between zirconia and titanium has been investigated using an analytical transmission electron microscope with a dedicated energy dispersive spectrometer (TEM/EDS), and the mechanism of the interface reactions is also discussed.
Experimental Procedures
A commercially pure titanium powder (CP-Ti, 200 mesh, 99.8%) and several calcia partially stabilized ZrO2plates (;9 mm thick, 5 mol %CaO-ZrO2) were loaded in a zirconia crucible (5C–ZrO2), and then
put in an electric resistance furnace. The chamber was evacuated to 1024torr and then backfilled with argon to atmospheric pressure. This cycle of evacuation and purge with argon was repeated at least twice. It took 30 minutes to raise the temperature from 100 to 1600°C and 5 minutes from 1600 to 1750°C, and then held at 1750°C for 7 minutes. During cooling, the temperature was lowered to 1600°C at the cooling rate of 30°C/min, to 1000°C at 50°C/min. Next, the specimen was cooled down to room temperature in furnace. The interface of zirconia and titanium was observed using an analytical TEM (JOEL JEM 2010) as well as a scanning electron microscope (JXA 6400). The cross-sectional TEM specimens perpendicular to the interface of zirconia and titanium were prepared by standard procedures of cutting, grinding, polishing, and ion milling. The quantitative composition analyses, with an error less than6 5%, were carried out based on the principle of Cliff-Lorimer [6] by an energy dispersive spectrometer (EDS; Model ISIS 300) attached to the TEM.
Results and Discussions (a) SEM Analyses
Figure 1 displays an SEM micrograph showing the interface between zirconia and titanium after reaction at 1750°C/7 min. The unaffected ZrO2with a porous structure is located at the most left end,
while the unaffected titanium is in the right end. At high temperatures, the liquid titanium infiltrated into zirconia through the open pores forming a thick chemical reaction zone (indicated as Zone C) at the interface. A layer of the oxygen deficient zirconia (ZrO22x) (indicated as Zone B) is to the left of this
chemical reaction layer (Zone C). A change in the color of calcia partially stabilized zirconia (from light yellow to gray black) after reaction indicated oxygen deficiency caused by the extensive reduction of
Figure 2. A high resolution TEM micrograph of the ZrO2/TiO2interface.
Figure 1. An SEM micrograph showing the interface between zirconia and titanium after reaction at 1750°C/7 min. Zone A: unaffected ZrO2; Zone B: ZrO22xlayer; Zone C: chemical reaction layer; Zone D:a-case; Zone E: unaffected Ti.
INTERFACIAL REACTION
ZrO2[7]. To the right side of the chemical reaction zone (Zone C) there exists a needle-likea-phase
(indicated as Zone D). In addition, there were many micro-pores along the grain boundaries of this a-phase. This implies that zirconia lost oxygen and oxygen was dissolved in titanium, leading to the formation ofa-titanium and small oxygen pores.
(b) TEM Observations
Formation of TiO2. TiO2was observed in the chemical reaction zone (Zone C in Fig. 1). At high
temperatures, oxygen was leached from ZrO2and dissolved in titanium melt. TiO2could be formed if
the reaction between zirconia and titanium was extensive. Figure 2 presents a high resolution image of the ZrO22x/TiO2interface with (220) plane of ZrO22xbeing slightly misaligned with [101] plane of
TiO2. The phase diagram of ZrO2-TiO2[8] indicates that a two-phase region of ZrTiO4and TiO2in the
part of rich TiO2. But ZrTiO4was not found in this study.
Formation of a-Zr(O). Figure 3(a) displays a bright field image of a-Zr(O) and ZrO22x(Zone B or
Zone C in Fig. 1). Sphericala-Zr(O) particles were randomly distributed in ZrO22x. Inset in the lower
right corner of Fig. 3(a) is the selected area diffraction pattern (SADP) of a-Zr, while ZrO22x
(designated as primary ZrO22x) was identified to be a cubic phase. The EDS in Fig. 3(b) confirmed that
the spherical particles werea-Zr with a limited solubility of oxygen (note that Cu peaks in the EDS Figure 3. (a) A bright field image ofa-Zr(O) and ZrO22x; (b) an EDS ofa-Zr(O); (c) an EDS of ZrO22x.
spectrum were caused by the contamination of the copper grip during ion milling.) Figure 3(c) depicts an EDS result of the cubic ZrO22x, showing that the cubic ZrO22xdissolved 12.79 at% Ca. It implies
that Ca remained in ZrO22x and the ratio of CaO/ZrO2 after the precipitation of a-Zr(O) from the
primary ZrO22xbecame larger than that in the original 5C-ZrO2.
Zr-O phase diagram [9] displayed a two phase region of a-Zr(O) and c-ZrO22xin a wide region
(32.5 at%O-63.5 at%O) above the eutectoid temperature ('1525°C). Since x was around 0.3 in ZrO22x
in this study, the equilibrium Zr/O ratio in ZrO22x decreased with decreasing temperature. ZrO22x
could transform to the one with a higher oxygen content by the ex-solution ofa-Zr(O) during cooling, and sphericala-Zr(O) simultaneously precipitated in the primary ZrO22x.
Formation of secondary ZrO22xcrystallites. Nano-cystalline particles in Fig. 4(a) were identified to be
cubic zirconia from the diffraction rings in Fig. 4(b). The O/Zr ratio of these cubic zirconia crystallites was determined to be close to 1.9 by the Cliff-Lorimer method. Note that the diffraction spots in Fig. 4(b) were caused by the untransformed primary ZrO22x.
The transformation of zirconia in this study can be described as follows: ZrO2 was reduced to
primary ZrO22xby titanium at high temperatures. Spherical proeutectoida-Zr(O) would precipitate,
while the O/Zr ratio of the primary ZrO22xincreased. As the temperature decreased, secondary ZrO22x
crystallites with the O/Zr ratio being close to 1.9 were crystallized.
Oxygen deficient zirconia was stabilized during the interfacial reactions. The lowest temperature for cubic oxygen deficient zirconia was 1525°C and, at low temperatures, the monoclinic oxygen deficient zirconia existed in a very limited range [10]. However, the cubic ZrO22xwas found in present study.
There might be two reasons for the stabilization of ZrO22x: One is that its particle size was smaller than
the critical size for the t3 m transformation, while the other is that the CaO content increased in the primary ZrO22x.
Precipitation of CaZrO3. CaZrO3precipitated from the solid solution of primary ZrO22xas the ratio of
CaO/ZrO2increased due to the ex-solution ofa-Zr(O) (Zone C in Fig. 1). Figure 5(a) displays a TEM
micrograph showing several oval CaZrO3embedded ina-Zr(O). The EDS result in Fig. 5(b) confirms
the existence of CaZrO3. Figure 5(c) depicts the EDS result of a-Zr(O), indicating that a-Zr(O)
contained only Zr and O elements. The content of Ca in ZrO22xincreased aftera-Zr(O) was excluded
from ZrO22x. At the eutectoid temperature of CaO–ZrO2(1140°C) [11], the solid solution of ZrO22x
with more than 17 mol% CaO could decompose into tetragonal zirconia solid solution and two ordered phase (w1: CaZr4O9, andw2: Ca6Zr19O44). However, the rates of decomposition to ordered phases are
Figure 4. (a) A bright field image of secondary ZrO22xcrystallites; (b) an SADP showing diffraction rings caused by the
secondary ZrO22xcrystallites in (a).
INTERFACIAL REACTION
very slow at relatively low temperatures (,1400°C) [12]. Therefore, only CaZrO3precipitated, leaving
a-Zr(O) to remain in the matrix.
Summary
Interfacial reactions between zirconia and titanium were proceeded by the infiltration of liquid titanium through the open pores of zirconia. During reaction, zirconia was reduced to oxygen-deficient zirconia (ZrO22x) and resulted in the liberation of oxygen. The dissolution of oxygen in titanium gave rise to
the needle-likea-phase as well as TiO2in the reaction zone. The content of oxygen in ZrO22xincreased
because of the ex-solution of proeutectoid spherical particlesa-Zr(O). On cooling, secondary ZrO22x
crystallites with O/Zr' 1.9 were found. Furthermore, the Ca content increased in the solid solution of ZrO22xthat could induce the precipitation of CaZrO3.
References 1. R. L Saha and K. T. Jacob, Def. Sci. J. 36, 121 (1986).
2. M. J. Donachie Jr., Titanium: A Technical Guide, p. 40, ASM International, Metals Park, OH (1988). 3. G. Economos and W. D. Kingery, J. Am. Ceram. Soc. 36, 403 (1953).
4. R. Ruh, J. Am. Ceram. Soc. 46, 301 (1963).
5. R. L. Saha, T. K. Nandy, R. D. K. Misra, and K. T. Jacob, Metal. Trans. 21B, 559 (1990). 6. G. Cliff and G. W. Lorimer, J. Microsc. 103, 203 (1975).
7. G. M. Ingo, J. Am. Ceram. Soc. 74, 381 (1991).
8. R. S. Roth, T. Negas, and L. P. Cook, Phase Diagrams for Ceramists, p. 169, American Ceramics Society, Columbus, OH (1981).
9. H. Baker, ASM Handbook, Alloy Phase Diagrams, vol. 3, p. 2.326 (1992). 10. R. J. Ackermann, S. P. Garg, and E. G. Rauh, J. Am. Ceram. Soc. 60, 341 (1977). 11. J. R. Hellmann, and V. S. Stubican, Mater. Res. Bull. 17, 459 (1982).
12. J. R. Hellmann and V. S. Stubican, J. Am. Ceram. Soc. 66, 260 (1983).
INTERFACIAL REACTION
