Formation of omega and Ti2Ni phases in a residual Ti-Ni-Ti multilayer between stainless steel and zirconia after bonding at 1173K

This article was downloaded by: [National Chiao Tung University 國立交通大學] On: 21 July 2015, At: 02:13

Publisher: Taylor & Francis

Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place, London, SW1P 1WG

Click for updates

Philosophical Magazine Letters

http://www.tandfonline.com/loi/tphl20

Formation of omega and Ti

2

Ni phases in

a residual Ti–Ni–Ti multilayer between

stainless steel and zirconia after

bonding at 1173 K

a

Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan

Published online: 02 Sep 2014.

To cite this article: Shen-Hung Wei & Chien-Cheng Lin (2014) Formation of omega and Ti2Ni phases

in a residual Ti–Ni–Ti multilayer between stainless steel and zirconia after bonding at 1173 K, Philosophical Magazine Letters, 94:10, 609-619, DOI: 10.1080/09500839.2014.951705

To link to this article: http://dx.doi.org/10.1080/09500839.2014.951705

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &

and-conditions

Formation of omega and Ti

Ni phases in a residual Ti–Ni–Ti

multilayer between stainless steel and zirconia after bonding at

1173 K

Shen-Hung Wei and Chien-Cheng Lin*

Department of Materials Science and Engineering, National Chiao Tung University, Hsinchu 300, Taiwan

(Received 9 December 2013; accepted 3 July 2014)

Stainless steel (316L) and 8 mol pct Y2O3-stabilized zirconia (8Y-ZrO2) were

bonded using a Ti–Ni–Ti multilayer at 1173 K (900 °C) for 1 h. Cross-sectional transmission electron microscopy specimens were prepared by an innovative focused ion beam plus lift-out technique. In addition to acicular α-Ti, the dissolution of Fe, Cr, and Ni diffusing outwards from 316L into β-Ti led to the precipitation of the omega (ω) phase with different variants in the residual Ti foil between 316L and Ni. The ω-phase was not found in the residual Ti foil between Ni and 8Y-ZrO2, while Ti2Ni precipitates were

precip-itated in someα-Ti grains owing to the exclusion of Ni from the β-Ti. Keywords: bonding; omega phase; interfaces; titanium; stainless steels; transmission electron microscopy

1. Introduction

The joining of stainless steel 316L and 8Y-ZrO2(8 mol pct Y2O3-stabilized ZrO2) has

attracted increasing attention in planar solid oxide fuel cell applications because it pro-vides excellent thermal stability and electrical conductivity at high temperatures [1,2]. The coefficients of thermal expansion of 316L and 8Y-ZrO2 are very different

(19.5 × 10−6/ °C for 316L and 10.5 × 10−6/ °C for FSZ) [3,4]. It is therefore necessary to find a specific multilayer, such as a Ti–Ni–Ti trilayer, that is ductile and suitably reactive to both stainless steel 316L and 8Y-ZrO2 to enhance the bonding strength and

reliability. The Ti layers in Ti–Ni–Ti act as reactive materials to promote chemical bonding, while the Ni layer acts as a buffer layer to absorb strain energies arising from the mismatch in the thermal expansion coefficients.

In commercially pure titanium, the low-temperatureα-phase is stable up to approxi-mately 890 °C. Increasing the O, Al, and/or N contents increases theα → β transition temperature. Moreover, the addition of these elements has a significant effect on the mechanical properties of Ti (and Ti alloys) [5]. Metastableβ-Ti alloys are such Ti alloys in which the martensiticβ → α transformation is suppressed by sufficient concentration of β stabilizing elements (e.g. V, Cr, Zr, Mo, and Fe) [6–10]. The metastable β-phase can transform to an athermal ω-phase upon quenching from the β-phase field, or to an

*Corresponding author. Email:[email protected]

© 2014 Taylor & Francis

Philosophical Magazine Letters, 2014

Vol. 94, No. 10, 609–619, http://dx.doi.org/10.1080/09500839.2014.951705

isothermal ω-phase on ageing at a low temperature. Formation of the ω-phase is generally considered undesirable since it can lead to severe embrittlement of the alloys. Banerjee and Mukhopadhyay reported that the ω-phase resulted from a metastable β-phase in the Ti, Zr, and Hf alloys with the required levels of β-stabilizing elements [11]. Furthermore, Williams et al. reported that the addition of Al, O, Sn, or Zr to metastable β-phase titanium alloys (e.g., Ti–V and Ti–Mo alloys) reduced the volume fraction, the upper temperature limit of formation, and the time of stability of the ω-phase [12].

As mentioned above, athermal and isothermal ω-phases are obtained by quenching and ageing, respectively. The latter occurs through a diffusion process, and the former is formed through the lattice collapse mechanism [11]. The formation of the athermal ω-phase can be explained in terms of a typical TTT diagram [13]. The actual time and temperature values not only depend on the cooling rate but also on the typical alloy additions to the titanium alloy.

The ω-phase is formed from the metastable β-phase by collapsing a pair of neighbouring (1 1 1) planes to the intermediate position, leaving the next (1 1 1) plane unaltered, collapsing the next pair, and repeating this process. This process produces a structure of hexagonal symmetry (called the ideal omega) through a complete double-plane collapse, as described by Silcock [14]. Devaraj et al. showed direct evidence of the collapse of neighbouring (2 2 2)βplanes using high-resolution transmission electron

microscopy (TEM) [15]. However, the double-plane collapse in some alloys or at the initial stage of this transformation may not be always complete, leading to a trigonal structure (trigonalω) [16].

Mixedω- and β-phases induced by the microsegregation of additions of X are often mistaken for a single β-phase in Ti–X alloys. These two phases can be distinguished under direct observation using TEM. In addition, few studies have been conducted on theω-phase at the joining interface between the metal and the ceramic because it is dif-ficult to prepare cross-sectional TEM specimens by conventional cutting and grinding techniques.

In this study, stainless steel (316L) and zirconia (8Y-ZrO2) were bonded using a

Ti–Ni–Ti multilayer, and the interfacial microstructures between the base materials and the Ti foils were investigated using analytical electron microscopy. It was supposed that the interfacial reactions took place between base materials and interlayers through atomic interdiffusion mechanisms. The elements, such as Fe, Cr, and Ni, diffusing from 316L to Ti foils are considered as β-Ti phase stabilizers, while O evolving from 8Y-ZrO2is aα-Ti phase stabilizer. The objective of this study was to explore the effect

of atomic interdiffusion on the phase transformation in the two residual Ti foils between stainless steel (or ZrO2) and Ni. The formation of the ω-phase in the residual Ti foil

was elucidated from a crystallographic perspective with the aid of a model.

2. Experimental procedures

Hot-pressed 8 mol pct Y2O3fully stabilized ZrO2(8Y-ZrO2) (2 mm thick, mainly 91.35

ZrO2 + HfO2, 8.65 Y2O3 in mol pct) and stainless steel 316L sheets (5 mm thick,

mainly 60.8 Fe, 19.3 Cr, 14.2 Ni, 1.7 Mo, 2.0 Mn, and 2 Si in at pct) were isothermally joined with a Ti–Ni–Ti multilayer under 0.1 MPa at 1173 K (900 °C) for 1 h in a tube furnace. The Ni foil (100μm thick and 99.5 pct pure) and the Ti foils (50 μm thick and

99.5 pct pure) were employed as a core layer and as outer layers, respectively. To prevent the sandwiched samples from being oxidized, they were annealed in an Ar-protective atmosphere. The heating and cooling rates were 5 K/min and 3 K/min, respectively.

Cross-sectional SEM samples were cut and ground along the direction perpendicular to the interface, followed by grinding and polishing. Cross-sectional TEM specimens were prepared by an innovative focused ion beam (FIB) plus lift-out technique (Model Nova 200) [17]. The FIB milling was performed with a Ga+ ion beam at 30 keV. Prior to the FIB milling, a Pt layer (1μm thick) was deposited by ion beam chemical deposi-tion using C9H16Pt as the precursor gas. The Pt layer served as a marker and prevented

the outer surface of the sample from being directly exposed to the Ga+ ion beam implantation during subsequent ion milling operations. After rough milling (7–1 nA), polishing (0.5–0.1 nA), and final polishing (50–10 pA), a thin foil (17 × 2 × 0.05 μm) was cut off, and then a micromanipulator was used to transfer the foil from the sample to a TEM grid for subsequent TEM analyses.

Cross-sectional samples prepared as described above were characterized using ana-lytical SEM (scanning electron microscopy, Model JSM 6500F) and TEM (Model JEM 2000 FX), each with an attached energy-dispersive spectroscope (EDS, Model ISIS 300). The Cliff-Lorimer standardless technique was performed on the TEM/EDS, which was equipped with an ultra-thin Be window so that elements lighter than Na, such as C, N, and O, could be detected accurately [18]. The crystal structures of various phases were obtained from the analyses of selected area diffraction patterns (SADPs) using computer-simulation crystallography software.

3. Results and discussion

Figure 1a shows a BEI (backscattered electron image) of a cross-section of the 316L/ 8Y-ZrO2joint with a Ti/Ni/Ti multilayer after bonding at 1173 K (900 °C) for 1 h.

Sev-eral reaction zones occurred at the stainless steel 316L/Ti, Ti/Ni, and Ti/8Y-ZrO2

inter-faces. Residual Ti and Ni foils were also observed because of incomplete reactions at these interfaces. The two residual Ti foils, which were mainly composed of α-Ti and β-Ti, exhibited very different microstructures, whereas the residual Ni foil was a single phase with Ti in solid solution. As shown in Figure 1b, the residual Ti foil between 316L and Ni consisted of needle-like α-Ti (98.8 ± 0.4 Ti, 0.5 ± 0.1 Fe, 0.4 ± 0.3 Ni, and 0.3 ± 0.1 Cr in at pct) and scattered Ti2Ni (65.4 ± 1.9 Ti, 28.8 ± 2.1 Ni, 3.2 ± 0.8

Fe, and 2.6 ± 0.6 Cr in at pct) along with retainedβ-Ti. At the place, where large Ti2Ni

grains were present, the neighbouring area was depleted of Ni, facilitating the formation of the needle-like α-Ti. The retained β-Ti, consisting of 93.4 ± 1.1 Ti, 3.1 ± 0.5 Fe, 1.8 ± 0.7 Cr, and 1.7 ± 0.8 Ni in at pct, was consistent with the fact that the dissolution of Fe, Cr, and Ni gave rise to the stabilization ofβ-Ti [19,20]. Figure1c shows that the residual Ti foil between Ni and 8Y-ZrO2 contained uniformly dispersed Ti2Ni

precipi-tates (bright) in theα-Ti phase (dark). In addition, β-Ti could undergo a eutectoid trans-formation of β-Ti → α-Ti + Ti2Ni (not shown). The precipitation of Ti2Ni consumed

almost all of the Ni diffusing from theβ-Ti grains during cooling, leading to the local transformation ofβ-Ti → α-Ti in the matrix. The retained β-Ti consisted of 87.8 ± 1.9 Ti, 11.3 ± 1.8 Ni, and 0.9 ± 0.3 O in at pct. A sufficient Ni solute content can suppress the phase transformation of β → α-Ti even though O is a strong α stabilizer. It is well

Philosophical Magazine Letters 611

known that ZrO2tends to be reduced to oxygen-deficient zirconia (ZrO2−x) in a

reduc-ing environment. Because Ti has a high affinity to O, it can act as a reducreduc-ing agent of ZrO2. When Ti is in contact with ZrO2, oxygen can diffuse out of ZrO2 and be

dis-solved into Ti, leading to the formation of ZrO2−x. Oxygen ions are readily dissolved

and occupy the interstitial sites in Ti up to high concentrations. Some titanium subox-ides (e.g. TiO, TiO2, and Ti2O3) are likely formed as the amount of dissolved oxygen

increases in Ti. In this study, ZrO2was partially reduced by Ti to form ZrO2−x, while

oxygen was dissolved into the residual Ti foil to form the intermediate compound TiO (48.2 ± 1.7 Ti, 47.8 ± 2.8 O, and 4.0 ± 1.1 Ni in at pct). No Zr solute was detected in the residual Ti foil between Ni and 8Y-ZrO2, indicating that the TiO sublayer at the Ti/

8Y-ZrO2interface could retard Zr diffusion from ZrO2into the residual Ti foil.

Figure 2a displays a bright-field image (BFI) of the needle-like α-Ti in the β-Ti

matrix within the residual Ti foil between 316L and Ni after bonding at 1173 K (900 °C) for 1 h. At a higher magnification, Figure 2b shows various fine precipitates of the ω-phase in the β-Ti matrix. The diffusion of Fe, Cr, and Ni from the steel resulted in a composition shift in the Ti foil to a metastable β-Ti region. It was such a composition that made the formation of theω-phase possible [21]. These nanometricω precipitates (approximately 30 nm) with two different variants (designated as ω1 and

ω2, respectively) were orientated with a true angle of approximately 109.5°. Silcock

reported that there are four crystallographic orientations of the ω-phase with respect to the β-phase [14]. However, only two orientations can be observed along the 1 ½ 1 0_b-Ti zone axis, and the remaining two can be visible in TEM with the incident electron

Figure 1. Backscattered electron images of the 316L/8Y-ZrO2 jointed by a Ti–Ni–Ti trilayer at

1173 K (900 °C) for 1 h: (a) all the reaction layers at a low magnification, (b) residual Ti between the 316L and Ni, and (c) residual Ti between Ni and ZrO2.

beam along the 1 1 0½ _b-Ti zone axis. In this study, the samples were cooled at a moder-ate rmoder-ate (about 3 K/min) in furnace, leading to the simultaneous transformations of β → β + α and β → β + ω so that α, β, and ω coexisted in the residual Ti foil between 316L and Ni. The ω-phase was formed preferentially in the solute (Fe, Cr, etc.)-depleted regions of the β matrix in the Ti alloyed with transition metals [22–24],

Figure 2. (a) A BFI showing the needle-likeα in the β-matrix in the residual Ti foil between the 316L and the Ni after bonding at 1173 K (900 °C) for 1 h, (b) a BFI of theβ + ω two-phase structure, (c) and (d) CDFIs showing the different variants ofω precipitates using the 1 0 1 0ð Þ_x1 and ð1 0 1 0Þ_x2 diffraction spots, respectively, (e) and (f ) SADPs of the β-Ti and ω-phase and their schematic diagrams (O, β-Ti; ●, ω1; Δ, ω2) with zone axes of 1 ½ 1 0_b-Ti// 1 ½ 2 1 0_x(A:

1 0 1 0

ð Þ_x1, A′: 0 0 0 1ð Þ_x

1; B: ð1 0 1 0Þx2, B′: 0 0 0 1ð Þx2), and (g) the EDS spectra of α-, β-, and

ω phases.

Philosophical Magazine Letters 613

indicating that some β-stabilizing elements were expelled from the ω-phase during transformation. In alloys, which contain a sufficient concentration of β-stabilizers to suppress athermalω-phase formation, the metastable β-phase can be subjected to a spin-odal decomposition into a solute-rich β-matrix and a corresponding solute-lean β′-phase [22]. The solute-lean β′-phase is thermodynamically favourable for the formation of the isothermal ω-phase. As shown in Figure 2c, the central dark-field image (CDFI) using

the 1 0 ð 1 0Þω1 diffraction spot (g vector) revealed theω1precipitates with their

longitu-dinal direction oriented from the lower left to the upper right. As shown in Figure 2d, another CDFI formed by the ð1 0 1 0Þ_x2 diffraction spot revealed the ω2 precipitates

with their longitudinal direction oriented from the lower right to the upper left. The occurrence of the ω-phase in the residual Ti foil could be mainly attributed to the fol-lowing two factors. First, the dissolution of sufficient β-stabilizers (Fe, Cr, and Ni) retained the metastable β-Ti and made the ω-phase formation thermodynamically favourable. Second, a moderate cooling rate (3 K/min) created favourable kinetic condi-tions for theω-phase formation. Moffat and Larbalestier reported the effect of the cool-ing rate on the β-phase decomposition in Ti–Nb alloys, indicating that the formation of α″ martensite was favoured by a rapid cooling rate, but a relatively slow cooling rate favoured the formation of theω-phase [25]. Figure2e displays the SAD patterns of the β + ω two-phase region. In addition to the strong bcc reflections of β-Ti, weak sharp spots corresponded to ω1and ω2, respectively. Previous studies indicated that the

crys-tallineω-phase formation was limited to less than a 10 at pct solute (Fe, Cr, and/or Ni) content in β-Ti [26], while a solute content of more than 10 at pct led to retained β-Ti with diffuseω scattering [26,27]. In this study, approximately 9.6 at pct of Fe, Cr, and Ni was dissolved in β-Ti at high temperatures, resulting in the direct transformation of β → β + ω during cooling. The SADPs were schematically redrawn in Figure 2f, with the diffraction spots being indexed, indicating that the orientation relations between the

bcc β-Ti matrix and the hcp ω-phase were recognized as follows:

z¼ 1 1 0½ _b-Ti // 1 ½ 2 1 0_x and 1 1 1ð Þ_b-Ti // 0 0 0 1ð Þ_x. It was noted that the ω1 and ω2

were precipitated in the 1 1 1f gβ habit planes (forming an octahedron). When the

inci-dent electron beam was parallel to the 1 ½ 1 0β-Ti, the view angle between the two

varia-tions of the ω-phase was 109.5 (the intersecting angle between TA
! and TB
!), corresponding to the dihedral angle of the octahedron. Furthermore, a mirror plane

0 0 2

ð Þ_b-Ti of ω1andω2was oriented parallel to the incident electron beam, so that the

diffraction spots ofω1andω2were symmetric with respect to the reciprocal lattice

vec-tor r*ð0; 0; 2Þ_b-Ti. The rotation of the diffraction pattern of ω1about r

*_{ð0; 0; 2Þ} b-Ti

by 180° corresponded to that of ω2, and vice versa. Figure 2g displays the EDS spectra

of α-, β-, and ω- phases in the residual Ti foil between 316L and Ni after bonding at 1173 K (900 °C) for 1 h. The needle-likeα-Ti consisted of 98.3 ± 0.6 Ti, 0.8 ± 0.2 Cr, 0.5 ± 0.2 Fe, and 0.4 ± 0.1 Ni in at pct based on TEM/EDS analyses, indicating that the Fe, Cr, and Ni atoms (i.e., β-stabilizers) were hardly dissolved in α-Ti. In contrast, the β-Ti matrix contained approximately 9.6 at pct β-Ti stabilizers, such as 6.0 ± 1.2 Fe, 1.9 ± 0.6 Cr, and 1.7 ± 0.5 Ni in at pct so that the β-Ti was retained at room tem-perature. The fine precipitates of the ω-phase consisted of 94.3 ± 0.8 Ti, 3.3 ± 0.6 Fe, 1.3 ± 0.4 Ni, and 1.1 ± 0.2 Cr in at pct within the residual Ti foil between 316L and Ni. Artefacts of additional Cu and Ga signals were induced by the copper grid and FIB operation, respectively, so that Cu and Ga were excluded from the quantitative analyses for a better comparison between the TEM/EDS and SEM/EDS results.

Figure 3. Formation of the crystalline ω-phase from a bcc β-Ti lattice: (a) the bcc β-Ti, (b) the 1 1 1

ð Þ projection of β-Ti, (c) a side view of the crystal structure in (b), showing the ABCABC sequence of the 2 2 2ð Þ_b planes, and (d) the crystallographic relation of theω-phase and the bcc β-Ti. For clarity, both crystal structures are overlapped and cut away, where aω = 1 0 ½ 1b-Ti,

b_ω= ½1 1 0_b-Ti, and c_ω = 1=2 1 1 1½ _b-Ti. The dashed arrows indicate the same atoms in the two cutaway halves.

Philosophical Magazine Letters 615

The formation mechanism of the ω-phase could be described from the crystallo-graphic viewpoint in correlation with the SADPs. Figure3a shows the bcc unit cell of β-Ti with an atom at each corner (depicted by grey spheres) and an atom at its body centre (depicted by red and blue spheres at various layers). Figure3b shows the 1 1 1ð Þ projection of the β-Ti crystal structure, displaying the most closely packed planes in sequence with the black atoms located at the topmost layer. Figure 3c illustrates the β-Ti crystal structure in the edge-on view of 1 1 1ð Þβ along the 2 ½ 1 1β-Tidirection with

seven layers, indicating an ABCABC… stacking sequence of 2 2 2f g_β-Ti. The diffraction spot of 2 2 2ð Þ_βwas located at the triple distance of the diffraction spot 0 0 0 1ð Þ_ω1 away from the direct spot, as shown in Figure 2e, because the distances of A to B, B to C, and C to A were equal to pffiffiffi3a=6. Furthermore, Figure 3d shows the crystallographic relationship of the hcp ω-phase (solid line) and the bcc β-Ti (dashed line) with a cut-away view for clarity. While the stacking sequence of β-Ti (solid circles) was AB-CABC, the crystal structure of the ω-phase was based on a stacking sequence of AB′ AB′ (hollow circles), where B′ was a collapsing pair of (2 2 2)β-Tiplanes. As a result,

the orientation relations between the ω-phase (solid line) and β-Ti (dash line) were 1 1 1

ð Þ_b-Ti== 0 0 0 1ð Þ_x and ½1 1 0_b-Ti // 1 ½ 2 1 0_x, as indicated by the SADPs in Figure2e.

Figure 4. (a) A bright-field image showing the coexistence of α-Ti and Ti2Ni in the residual Ti

foil between Ni and 8Y-ZrO2after bonding at 1173 K (900 °C) for 1 h, (b) the SADP of Ti2Ni

with z¼ 0 1 1½ , and (c) the EDS of Ti2Ni.

Figure4a displays the BFI ofα-Ti and Ti2Ni in the residual Ti foil between Ni and

8Y-ZrO2 after bonding at 1173 K (900 °C) for 1 h. Figure 4b shows an SADP of the

Ti2Ni phase along the 0 1 ½ 1 zone axis, corresponding to an fcc structure with a lattice

constant a = 1.17 nm (Ti2Ni, a = 1.131 nm in JCPDS # 720442). Figure4c shows the

EDS of the Ti2 Ni phase, indicating that the composition of the Ti2Ni was 65.1 ± 3.7

Ti, 33.2 ± 3.7 Ni, and 1.7 ± 0.4 O in at pct. Additionally, the composition of the α-Ti was 93.9 ± 0.8 Ti, 4.7 ± 0.8 O, and 1.4 ± 0.2 Ni in at pct, indicating that the precipita-tion of Ti2Ni led to the Ni depletion in Ti during cooling to room temperature. Artefacts

of the Cu and Ga peaks were caused by the copper ring and FIB, respectively, as men-tioned above. No ω-phase was found in the residual Ti foil between Ni and 8Y-ZrO2

because a significant amount of oxygen diffusing out of 8Y-ZrO2was dissolved in this

residual Ti foil [12].

The identified phases and their EDS results in the two residual Ti foils are summa-rized in Table 1. The compositions ofα, β, ω, and Ti2N were measured at several

dif-ferent points because a concentration gradient had formed along the direction perpendicular to the interface. Table1 shows the mean compositions and standard devi-ations of these phases. The residual Ti foil abutting 316L consisted of α, β, ω, and Ti2Ni, while the other Ti foil lacked the ω-phase. It is well known that the bonding

between base materials and interlayers results from interdiffusion across various inter-faces. Consequently, the residual Ti foil neighbouring the 316L contained Fe, Cr, and Ni, and the other Ti foil contained O and Ni. Amongst these elements, Fe, Cr, and Ni

Table 1. Compositions of phases formed at the residual Ti foils after bonding at 1173 K (900 °C) for 1 h. Interlayer Phase Composition (at pct) Notes Ti Fe Cr Ni O Residual Ti between 316L and Ni α 98.8 ± 0.4 0.5 ± 0.1 0.3 ± 0.1 0.4 ± 0.3 – SEM/EDS result β 93.4 ± 1.1 3.1 ± 0.5 1.8 ± 0.7 1.7 ± 0.8 – SEM/EDS result α 98.3 ± 0.6 0.5 ± 0.2 0.8 ± 0.2 0.4 ± 0.1 – TEM/EDS resulta β 90.4 ± 1.6 6.0 ± 1.2 1.9 ± 0.6 1.7 ± 0.5 – TEM/EDS resulta ω 94.3 ± 0.8 3.3 ± 0.6 1.1 ± 0.2 1.3 ± 0.4 – TEM/EDS resulta Ti2Ni 65.4 ± 1.9 3.2 ± 0.8 2.6 ± 0.6 28.8 ± 2.1 – SEM/EDS result Residual Ti between Ni and 8Y-ZrO2 α 93.9 ± 0.8 – – 1.4 ± 0.2 4.7 ± 0.8 TEM/EDS resulta β 87.8 ± 1.9 – – 11.3 ± 1.8 0.9 ± 0.3 SEM/EDS result Ti2Ni 65.1 ± 3.7 – – 33.2 ± 3.7 1.7 ± 0.4 TEM/EDS resulta TiO 48.2 ± 1.7 – – 4.0 ± 1.1 47.8 ± 2.8 SEM/EDS result a

Artefact signals for Cu and Ga were removed for a better comparison between TEM and SEM EDS results.

Philosophical Magazine Letters 617

are considered as β-stabilizers, and O is considered an α-stabilizer. The β-phase matrix was found in the residual Ti foil between ZrO2 and Ni, because it had an extended

dissolution of Ni up to 11.3 at pct. Oxygen atoms diffusing from 8Y-ZrO2 to the

residual Ti foil could further retard the formation of the ω-phase and result in the α+β two-phase structure. In contrast, the formation of theω- and α-phases resulted from the metastable β-Ti with approximately 9.6 at pct of Fe + Cr + Ni in the residual Ti foil between 316L and Ni. Laheurte et al. showed the critical solute concentrations (βc) of

several elements (including Fe, Cr, and Ni) required to retain 100% of the metastable β-Ti on quenching from the β-field [28]. For the metastable β-Ti alloys, the martensitic β → α is suppressed and an associated ω phase can be formed on quenching and/or ageing. Comparing the α-phases of the two residual Ti foils, one had an insignificant amount of Fe, Cr, and Ni (β-stabilizers), while the other contained a significant amount of dissolved O (α-stabilizer, ~4.7 at pct). It was also noted that the ω-phase contained less β-stabilizers (Fe + Cr + Ni ~ 5.7 at pct) than the β-phase (Fe + Cr + Ni ~ 9.6 at pct) did.

4. Conclusions

The ω-phase was observed in the residual Ti foil in a stainless steel and zirconia joint using Ti/Ni/Ti as an interlayer after annealing at 1173 K/1 h. The diffusion of Fe, Cr, and Ni from the steel to the Ti foils resulted in metastableβ-Ti, so that the ω-phase for-mation might be thermodynamically favourable. It was observed that ω- and β-Ti phases existed along with acicularα-Ti in the residual Ti between the 316L and the Ni, and the metastable-retained β-Ti contained approximately 9.6 at pct β-stabilizers (Fe, Cr, and Ni). Different variants ofω precipitates were observed in the β-matrix. The ori-entation relationship between theω-phase and the β-Ti phase was identified as follows: z¼ 1 1 0½ _b-Ti // 1 ½ 2 1 0_x and 0 0 0 1ð Þ_x== 1 1 1ð Þ_b-Ti. In contrast, no ω precipitate was found in the other residual Ti foil between Ni and 8Y-ZrO2 even though the β-Ti (Ni,

O) was retained at room temperature. This implied that the solid solution of oxygen could effectively suppress the formation of the ω-phase. The transformation of β-Ti → α-Ti was triggered following the precipitation of Ti2Ni during cooling. Whether

the formation of the ω-phase at the residual interlayer affected the bonding strength should be subjected to further investigation.

Funding

This work was supported by the National Science Council, Taiwan [grant number NSC98-2221-E-009-039-MY2].

References

[1] I. Antepara, I. Villarreal, L.M. Rodríguez-Martínez, N. Lecanda, U. Castro and Laresgoiti: J. Power Sources 151 (2005) p.103.

[2] W.Z. Zhu and S.C. Deevi, Mater. Sci. Eng. A 348 (2003) p.227.

[3] W.B. Hanson, K.I. Ironside and J.A. Fernie, Acta Mater. 48 (2000) p.4673. [4] I. Yasuda and M. Hishinuma, Electrochemistry 68 (2000) p.526.

[5] G. Lütjering and J.C. Williams, Titanium, 2nd ed., Springer-Verlag, Berlin, 2007, p.175.

[6] R. Banerjee, P.C. Collins, D. Bhattacharyya, S. Banerjee and H.L. Fraser, Acta Mater. 51 (2003) p.3277.

[7] W. Sinkler and D.E. Luzzi, Acta Metall. Mater. 42 (1994) p.1249. [8] B.A. Hatt and J.A. Roberts, Acta Metall. 8 (1960) p.575. [9] W.F. Ho, J. Med. Biol. Eng. 28 (2007) p.47.

[10] D. De Fontaine, N.E. Paton and J.C. Williams, Acta Metall. 19 (1971) p.1153.

[11] S. Banerjee and P. Mukhopadhyay, Phase Transformation: Examples from Titanium and Zirconium Alloys, Pergamon, Oxford, 2007, p.473.

[12] J.C. Williams, B.S. Hickman and D.H. Leslie, Metall. Trans. 2 (1971) p.477. [13] F.L. Harmon and A.R. Troiano, Am. Soc. Metals 53 (1961) p.43.

[14] J.M. Silcock, Acta Metall. 6 (1958) p.481.

[15] A. Devaraj, S. Nag, R. Srinivasan, R.E.A. Williams, S. Banerjee, R. Banerjee and H.L. Fraser, Acta Mater. 60 (2012) p.596.

[16] R. Pynn, J. Phys. F: Met. Phys. 8 (1978) p.1.

[17] L.A. Giannuzzi and F.A. Stevie, Micron 30 (1999) p.197. [18] G. Cliff and G.W. Lorimer, J. Microsc. 103 (1975) p.203.

[19] M. Ghosh, K. Bhanumurthy, G.B. Kale, J. Krishnan and S. Chatterjee, J. Nucl. Mater. 322 (2003) p.235.

[20] F.C. Campbell, ASM Handbook, Elements of Metallurgy and Engineering Alloys, ASM Inter-national, Materials Park, OH, 2008, p.527.

[21] S. Nag, Influence of beta instabilities on the early stages of nucleation and growth of alpha in beta titanium alloys, PhD thesis, The Ohio University, 2008.

[22] Z. Fan and A.P. Miodownik, J. Mater. Sci. 29 (1994) p.6403. [23] B.S. Hickman, J. Mater. Sci. 4 (1969) p.554.

[24] A.A. Popov, Phys. Met. Metall. 76 (1993) p.524.

[25] D.L. Moffat and D.C. Larbalestier, Metall. Trans. A 19 (1988) p.1687. [26] W. Sinkler and D.E. Luzzi, Acta. Metall. Mater. 42 (1994) p.1249. [27] S.L. Sass, J. Less-Common Met. 28 (1972) p.157.

[28] P. Laheurte, A. Eberhardt and M.J. Philippe, Mater. Sci. Eng. A 396 (2005) p.223.

Philosophical Magazine Letters 619