Bonding of TRIP-Steel/Al2O3-(3Y)-TZP Composites and (3Y)-TZP Ceramic by a Spark Plasma Sintering (SPS) Apparatus

A combination of the high damage tolerance of TRIP-steel and the extremely low thermal conductivity of partially stabilized zirconia (PSZ) can provide controlled thermal-mechanical properties to sandwich-shaped composite specimens comprising these materials. Sintering the (TRIP-steel-PSZ)/PSZ sandwich in a single step is very difficult due to differences in the sintering temperature and densification kinetics of the composite and the ceramic powders. In the present study, we successfully applied a two-step approach involving separate SPS consolidation of pure (3Y)-TZP and composites containing 20 vol % TRIP-steel, 40 vol % Al2O3 and 40 vol % (3Y)-TZP ceramic phase, and subsequent diffusion joining of both sintered components in an SPS apparatus. The microstructure and properties of the sintered and bonded specimens were characterized. No defects at the interface between the TZP and the composite after joining in the 1050–1150 °C temperature range were observed. Only limited grain growth occurred during joining, while crystallite size, hardness, shear strength and the fraction of the monoclinic phase in the TZP ceramic virtually did not change. The slight increase of the TZP layer’s fracture toughness with the joining temperature was attributed to the effect of grain size on transformation toughening.


Introduction
Below 1170˝C, zirconia transforms from the tetragonal phase into a monoclinic structure, accompanied by a volume expansion of 3%-5%. In partially stabilized ZrO 2 (by Y 2 O 3 , for instance), the tetragonal phase is metastable and displays stress-induced martensitic transformation into the monoclinic structure [1]. In the Nature manuscript named "Ceramic steel?" in 1975, the authors called partially stabilized zirconia "a ceramic analogue of steel" [2]. The toughness of zirconia realized upon such stress-induced transformation was analogously related to that observed in TRIP (transformation-induced plasticity)-steels [2,3]. TRIP-steels exhibit the phase transformation of metastable austenite into α'-martensite during plastic deformation. This phase transformation is called the TRIP-effect and depends on the chemical composition of the steel, temperature and strain rate.

Powder Consolidation in SPS
In Figure 1a, the microstructure of the SPS-consolidated (3Y)-TZP specimen is shown. The measured relative density of the (3Y)-TZP samples was 98.4%. Various properties of the (3Y)-TZP ceramic after SPS are shown in Table 1. The microstructure of the sintered composite is shown in Figure 1b. Bright TRIP-steel grains can be seen throughout the ceramic matrix.
Mean crystallite size of (3Y)-TZP 60 nm Mean crystallite size of the (3Y)-TZP phase in the composite layer 70 nm Fraction of the monoclinic phase in (3Y)-TZP <1 vol % Hardness of (3Y)-TZP 13.2 GPa Hardness of the composite 10.4 GPa Indentation fracture toughness ( ) of (3Y)-TZP 5.3 MPam 0.5  Non-destructive analysis of the joint using ultrasonic waves showed no detachments. A graphical representation of the ultrasonic test results can be seen in Figure 3. Each area unit was  SPS-sintered (3Y)-TZP specimens were self-joined at 1150˝C for 120 min. The HR-SEM image of a slightly etched surface ( Figure 2) shows the sub-micron equiaxed zirconia grain structure. The bonding area, marked by a dash-dot red line, is continuous, with no boundary/seam being obtained.     Non-destructive analysis of the joint using ultrasonic waves showed no detachments. A graphical representation of the ultrasonic test results can be seen in Figure 3. Each area unit was  Non-destructive analysis of the joint using ultrasonic waves showed no detachments. A graphical representation of the ultrasonic test results can be seen in Figure 3. Each area unit was scanned and the reflection results were translated into a colorimetric/numeric scale, where good bonding is indicated by lower numbers (blue color on the scale) and poor bonding is indicated by higher numbers (red color on the scale). Most of the bonding area, except for some alterations at the sample periphery, showed excellent bonding quality. scanned and the reflection results were translated into a colorimetric/numeric scale, where good bonding is indicated by lower numbers (blue color on the scale) and poor bonding is indicated by higher numbers (red color on the scale). Most of the bonding area, except for some alterations at the sample periphery, showed excellent bonding quality. Ultrasonic wave reflections were translated into a colorimetric/numeric/scale reflecting bonding quality (0/blue: high bonding quality, 10/red: low bonding quality).

Composite/(3Y)-TZP Joining
The main problem with the single-step sintering of steel-ceramic/ceramic sandwich structures derives from the different sintering temperatures and kinetics of the metal and ceramic components. Due to the fact that carbon from the graphite die diffuses into the steel, leading to a reduction in the melting point, the sintering temperature of the steel/ceramic composite is limited to 1150 °C. However, significantly higher temperatures (about 1350-1450 °C) are necessary to achieve fully dense (3Y)-TZP. To overcome this problem, solid-state bonding of the sintered composite with a ceramic content of 80 vol % to the sintered (3Y)-TZP part was applied (Figure 4a). No cracks or voids were observed at the bonding interface ( Figure 4b). A typical microstructure of the (3Y)-TZP layer after joining for 120 min at 1150 °C is shown in Figure 5. Ultrasonic wave reflections were translated into a colorimetric/numeric/scale reflecting bonding quality (0/blue: high bonding quality, 10/red: low bonding quality).

Composite/(3Y)-TZP Joining
The main problem with the single-step sintering of steel-ceramic/ceramic sandwich structures derives from the different sintering temperatures and kinetics of the metal and ceramic components. Due to the fact that carbon from the graphite die diffuses into the steel, leading to a reduction in the melting point, the sintering temperature of the steel/ceramic composite is limited to 1150˝C. However, significantly higher temperatures (about 1350-1450˝C) are necessary to achieve fully dense (3Y)-TZP. To overcome this problem, solid-state bonding of the sintered composite with a ceramic content of 80 vol % to the sintered (3Y)-TZP part was applied (Figure 4a). No cracks or voids were observed at the bonding interface ( Figure 4b).
Materials 2016, 9, 558 4 of 11 scanned and the reflection results were translated into a colorimetric/numeric scale, where good bonding is indicated by lower numbers (blue color on the scale) and poor bonding is indicated by higher numbers (red color on the scale). Most of the bonding area, except for some alterations at the sample periphery, showed excellent bonding quality. Ultrasonic wave reflections were translated into a colorimetric/numeric/scale reflecting bonding quality (0/blue: high bonding quality, 10/red: low bonding quality).

Composite/(3Y)-TZP Joining
The main problem with the single-step sintering of steel-ceramic/ceramic sandwich structures derives from the different sintering temperatures and kinetics of the metal and ceramic components. Due to the fact that carbon from the graphite die diffuses into the steel, leading to a reduction in the melting point, the sintering temperature of the steel/ceramic composite is limited to 1150 °C. However, significantly higher temperatures (about 1350-1450 °C) are necessary to achieve fully dense (3Y)-TZP. To overcome this problem, solid-state bonding of the sintered composite with a ceramic content of 80 vol % to the sintered (3Y)-TZP part was applied (Figure 4a). No cracks or voids were observed at the bonding interface ( Figure 4b). A typical microstructure of the (3Y)-TZP layer after joining for 120 min at 1150 °C is shown in Figure 5. A typical microstructure of the (3Y)-TZP layer after joining for 120 min at 1150˝C is shown in Figure 5. Only limited grain growth (from the initial 208 nm to 226 nm after joining at 1150 °C) occurred during joining, while the crystallite size (as revealed by XRD analysis) and the fraction of the monoclinic phase in the TZP ceramic virtually did not change. This means that the martensitic phase transformation from the tetragonal to the monocline structure in the pure TZP did not occur during the joining process. The representative XRD patterns assessed before and after the joining of (3Y)-TZP are shown in Figure 6. The results are in agreement with data reported by Ruiz et al. [39], where significant grain growth and change of phase composition appeared only after isothermal heat treatment at temperatures above 1550 °C.  Only limited grain growth (from the initial 208 nm to 226 nm after joining at 1150˝C) occurred during joining, while the crystallite size (as revealed by XRD analysis) and the fraction of the monoclinic phase in the TZP ceramic virtually did not change. This means that the martensitic phase transformation from the tetragonal to the monocline structure in the pure TZP did not occur during the joining process. The representative XRD patterns assessed before and after the joining of (3Y)-TZP are shown in Figure 6. The results are in agreement with data reported by Ruiz et al. [39], where significant grain growth and change of phase composition appeared only after isothermal heat treatment at temperatures above 1550˝C. Only limited grain growth (from the initial 208 nm to 226 nm after joining at 1150 °C) occurred during joining, while the crystallite size (as revealed by XRD analysis) and the fraction of the monoclinic phase in the TZP ceramic virtually did not change. This means that the martensitic phase transformation from the tetragonal to the monocline structure in the pure TZP did not occur during the joining process. The representative XRD patterns assessed before and after the joining of (3Y)-TZP are shown in Figure 6. The results are in agreement with data reported by Ruiz et al. [39], where significant grain growth and change of phase composition appeared only after isothermal heat treatment at temperatures above 1550 °C. The hardness values (Figure 7a) and the shear strengths for broken SPS-bonded samples ( Figure  8) likewise did not change during joining, although the fracture toughness of the TZP slightly The hardness values (Figure 7a) and the shear strengths for broken SPS-bonded samples (Figure 8) likewise did not change during joining, although the fracture toughness of the TZP slightly increased with the increased joining temperature (Figure 7b). This phenomenon may be attributed to the effect of grain size on transformation toughening [40]. Above a critical grain size, tetragonal grains would spontaneously transform into the monocline structure [39][40][41][42]. However, the temperature for critical grain growth was not achieved during bonding. increased with the increased joining temperature (Figure 7b). This phenomenon may be attributed to the effect of grain size on transformation toughening [40]. Above a critical grain size, tetragonal grains would spontaneously transform into the monocline structure [39][40][41][42]. However, the temperature for critical grain growth was not achieved during bonding.

Materials
The composite (20 vol % TRIP-steel, 40 vol % Al2O3 and 40 vol % (3Y)-TZP ceramic phase) was synthesized from high alloy austenitic CrMnNi-TRIP-steel (Table 2), alumina and yttria-stabilized zirconia powders. Median particle size of the steel powder was about 17 μm. Alumina (Pengda, Munich, Germany) and (3Y)-TZP ceramic powder (3 mole % yttria-stabilized zirconia, TOSOH, Tokyo, Japan) had median particle sizes of 0.49 μm and 40 nm, respectively. For composite fabrication, the powders were mixed in a planetary ball mill (Pulverisette 5, Fritsch GmbH, Idar-Oberstein, Germany) under high-energy ball milling conditions for 2 h. The ball material was hardened Cr-steel, the ball diameter was 25 mm, the powder to ball ratio was 1:10 and the rotating speed was 180 rpm.  increased with the increased joining temperature (Figure 7b). This phenomenon may be attributed to the effect of grain size on transformation toughening [40]. Above a critical grain size, tetragonal grains would spontaneously transform into the monocline structure [39][40][41][42]. However, the temperature for critical grain growth was not achieved during bonding.

Processing
Specimens were fabricated in a two-step process. First, TRIP-steel/Al 2 O 3 -(3Y)-TZP composite and pure (3Y)-TZP specimens were sintered separately. The powders were consolidated as cylindrical bodies with 20 mm diameter using the SPS technique in FCT-HP D 25/2-2 apparatus. Pure (3Y)-TZP and composite samples had a final height of approximately 3 mm.
The (3Y)-TZP powder was sintered at 1400˝C under uniaxial pressure of 60 MPa for holding time of 5 min. The heating and cooling rates were 100 K/min. The (3Y)-TZP specimens were self-joined in the SPS apparatus at 1150˝C with a holding time of 120 min under an argon atmosphere (10´2 torr) and a uniaxial pressure of 16 MPa. The composite was consolidated at 1150˝C for a holding time of 10 min under a pressure of 60 MPa.
The mating surfaces of the samples before joining were prepared by conventional metallographic technique on a 1 µm diamond paste (Struers Nap B) stage, cleaned in acetone and dried in air. The samples were placed in a graphite die with 20 mm inner and 40 mm outer diameters and inserted into the SPS apparatus (FCT-HP D5/1 System, Rauenstein, Germany) for joining. Composite and PSZ specimens were SPS-joined at 1050, 1100 and 1150˝C, respectively, with holding time of 120 min under an argon atmosphere (10´2 torr) and a uniaxial pressure of 16 MPa. The temperature was measured by a pyrometer focused on the upper graphite punch. Pulse-mode DC current (pulse 5 ms and pause 2 ms) was used throughout the joining experiments. The cooling rate after bonding was about 12 K/min.

Characterization
Determination of the density of the samples was achieved using the Archimedes method in distilled water. Quantitative phase analysis was carried out by XRD measurement (Cu-Kα radiation). Phase compositions were estimated using the Rietveld method.
Microstructure was characterized by scanning electron microscopy (SEM). Sample surfaces were ground and polished to a 1 µm diamond finish followed by vibratory polishing for 24 h. After polishing, the prepared ceramic samples were thermally etched at 950˝C for 1 h. Subsequently, grain size and size distribution were determined by the linear intercept method on more than three SEM micrographs, such that over 900 intercepted grains were considered for each condition.
The quality of the (3Y)-TZP/(3Y)-TZP joined region was tested by ultrasonic measurement using a 15 MHz pulse/receiver probe with 3.157 mm diameter (V260, Panametrics, Houston, TX, USA). Joined specimens were divided into squares with an area of about 9 mm 2 (3 mmˆ3 mm) each. The specimens were scanned with the transducer and the resulting reflections were documented according to the known sound speed in the tested material and specimen thickness. Reflections from the bonding area and those from the bottom of the tested specimen were collected, classified qualitatively and translated according to a colorimetric scale (blue to red) and numeric scale (ranging from 0 to 10, indicating well and poor bonding quality, respectively).
The maximal shear forces needed for fracture of the composite/(3Y)-TZP bonded specimens were determined using an LRX Plus apparatus (Lloyd Instruments, Fareham Hants, UK). The shear test specimen and tools are shown in Figure 9. Test specimen dimensions were 5ˆ5ˆ6 mm 3 (widthˆlengthˆheight). The test tools consisted of the static block, into which the specimen was mounted, and the moving block. When mounted into the testing apparatus prior to testing, the static block was placed on the static portion of the apparatus, while its dynamic portion was adjusted to the moving block of the tool. Four specimens were examined for each processing parameter tested.
were determined using an LRX Plus apparatus (Lloyd Instruments, Fareham Hants, UK). The shear test specimen and tools are shown in Figure 9. Test specimen dimensions were 5 × 5 × 6 mm 3 (width × length × height). The test tools consisted of the static block, into which the specimen was mounted, and the moving block. When mounted into the testing apparatus prior to testing, the static block was placed on the static portion of the apparatus, while its dynamic portion was adjusted to the moving block of the tool. Four specimens were examined for each processing parameter tested. The fracture toughness of the (3Y)-TZP specimens after sintering and joining was investigated by the indentation method. The applied indentation load was 98.07 N. Crack lengths were measured using an optical microscope. The fracture toughness (MPam 0.5 ) was calculated from the equation of Niihara et al. [43] for Palmqvist cracks: where is the Vickers hardness, is Young's modulus, represents the half length of the indentation diagonal, is the mean Palmqvist crack length and is a pseudo-constant. The value of depends on the ratio between Young's modulus and the uniaxial yield stress ( / ), as well as Poisson's ratio ; is reported in literature to be 2.7-3 for most ceramic materials [44]. Assuming that 2.7 and Vickers hardness equals 0.4636 • / , Equation (1) can be written as: It should be noted that Equation (2) is applicable for Palmqvist cracks only. At low crack-to-indent ratios, the dominantly formed crack geometry are Palmqvist cracks, typically showing a ratio ≲ 1.5 . At higher loads, crack geometry changes to halfpenny-shaped cracks (radial-median cracks) showing a ratio ≳ 2.5. Within the range 1.5 ≲ / ≲ 2.5, both crack systems can occur. Accordingly to the literature, Y-TZP ceramics preferentially crack in the Palmqvist rather than in the halfpenny mode [45,46].
Hardness was calculated from crack-free Vickers indentations generated at a lower applied indentation load of 4.90 N.

Conclusions
A two-stage fabrication approach was employed to obtain a sandwich-shaped specimen consisting of a high-alloy TRIP-steel/Al2O3-(3Y)-TZP composite and a (3Y)-TZP layer. The composite and ceramic specimens were separately consolidated and joined using SPS.


No evidence of cracks or voids was observed at the composite/TZP interface.  Limited grain growth from the initial 208 nm to 226 nm after joining at 1150 °C occurred during joining. The crystallite size and the fraction of the monoclinic phase in the Y-TZP ceramic Figure 9. Shear test tools and specimen. (a) Test specimen, dimensions 5ˆ5ˆ6 mm 3 ; (b) test tools and specimen; (c) test tools and specimen before and after mounting, full views; (d) test tools and specimen before and after mounting, cross-section views; (e) testing apparatus with mounted tools.
The fracture toughness of the (3Y)-TZP specimens after sintering and joining was investigated by the indentation method. The applied indentation load P was 98.07 N. Crack lengths were measured using an optical microscope. The fracture toughness K Ic (MPam 0.5 ) was calculated from the equation of Niihara et al. [43] for Palmqvist cracks: where H V is the Vickers hardness, E is Young's modulus, a represents the half length of the indentation diagonal, l is the mean Palmqvist crack length and ϕ is a pseudo-constant. The value of ϕ depends on the ratio between Young's modulus and the uniaxial yield stress (E{σ Y ), as well as Poisson's ratio ν; ϕ is reported in literature to be 2.7-3 for most ceramic materials [44]. Assuming that ϕ " 2.7 and Vickers hardness equals H V " 0.4636¨P{a 2 , Equation (1) can be written as: It should be noted that Equation (2) is applicable for Palmqvist cracks only. At low crack-to-indent ratios, the dominantly formed crack geometry are Palmqvist cracks, typically showing a ratio l a À 1.5. At higher loads, crack geometry changes to halfpenny-shaped cracks (radial-median cracks) showing a ratio l a Á 2.5. Within the range 1.5 À l{a À 2.5, both crack systems can occur. Accordingly to the literature, Y-TZP ceramics preferentially crack in the Palmqvist rather than in the halfpenny mode [45,46].
Hardness was calculated from crack-free Vickers indentations generated at a lower applied indentation load of 4.90 N.

Conclusions
A two-stage fabrication approach was employed to obtain a sandwich-shaped specimen consisting of a high-alloy TRIP-steel/Al 2 O 3 -(3Y)-TZP composite and a (3Y)-TZP layer. The composite and ceramic specimens were separately consolidated and joined using SPS.

‚
No evidence of cracks or voids was observed at the composite/TZP interface.
‚ Limited grain growth from the initial 208 nm to 226 nm after joining at 1150˝C occurred during joining. The crystallite size and the fraction of the monoclinic phase in the Y-TZP ceramic virtually did not change.

‚
The hardness values and the shear force for broken SPS-bonded samples did not change during joining. Slightly increased fracture toughness of the TZP-layer with increased joining temperatures was attributed to the effect of grain size on transformation toughening.
‚ SPS was proven to be an effective technique for sintering and solid-state joining of ceramics and metal/ceramic composites.