Microstructure and Interface Characteristics of 17-4PH/YSZ Components after Co-Sintering and Hydrothermal Corrosion

Combining stainless steel with zirconia components by powder technological shaping routes for manufacturing of multifunctional parts is an advantageous and promising one-step method making expensive and time-consuming additional joining steps redundant. However, several requirements for co-shaping and co-sintering of the very different compound partners have to be met. The microstructural and chemical constitution of the interface between both materials plays an important role for the mechanical properties, durability and corrosion resistance of the manufactured parts. In the present study, different shaping techniques for co-shaping of stainless steel and zirconia are introduced. The microstructure and the interphase properties of metal/ceramic hybrid parts have been investigated for samples made by tape casting, subsequent lamination and co-sintering. Nevertheless, the results of this study are valid for components made by other hybrid shaping processes as well. The interfaces were characterized by TEM, FESEM, EDX, and X-ray diffraction. Furthermore, the hydrothermal stability of the material compound was investigated.


Introduction
Combining different ceramics, glasses or metals in one component for attaining multifunctional properties have been reported in several publications for at least one decade. The authors used powder technological co-shaping routes for property combinations like ductility and wear resistance [1,2], electrical conductivity and insulation [3,4], dense and porous components [5], different coloring [6,7] or magnetic and non-magnetic properties [8]. Co-shaping thereby allows for a combination of different materials in only one processing step without any additional time-consuming joining steps. However, powder technological shaping routes always require debinding and sintering steps for removal of organic additives or binders and for attaining a complete densification of the sintered part with final properties. For multimaterial approaches the co-sintering of the components is even more demanding. In this case both materials which shall be combined durably in one part must fulfill a number of requirements: 1.
The coefficient of thermal extension (CTE) must be comparable over the whole range from sintering temperature down to ambient temperature. Without fulfilling this requirement the co-sintered materials compound definitely fails during cooling down from sintering temperature or, at latest during thermal cycling in the application of the part.
Whereas the first three requirements are material inherent properties which have to be taken into account already for choosing the fitting compound partners, the last point-the sintering behavior can be adjusted by choosing powders with suited particle size distributions and by equalizing the solid content of both components in the green state.
Beside the above mentioned requirements mixing up the material classes for multicomponent parts, e.g., combining ceramics with metals by powder technological routes has a further challenge in store. The chemical bonding mechanisms in ceramics are ionic or covalent bonding, whereas metallic bonding is characteristic for metals. For that reason the question arises which bonding can be achieved by co-sintering of both materials forming one material compound. The development work in this article pursues this questions for the material combination stainless steel and zirconia. This material combination is worth of investigation due to its excellent property combinations like hardness and ductility, electrical conductivity and insulation, white or black color and metallic gloss and the consequent applications for micro surgical instruments, heating elements, design parts or metal supported membranes. Several co-shaping technologies have been successfully investigated so far for this material combination, e.g., 2-component injection molding [9], Fused Filament Fabrication [10], Thermoplastic 3D Printing [11] or tape casting [12]. The last mentioned method had been chosen for describing the interphase formation in this article due to the simplicity of the shaping route and the large interphase area which can be produced with a relatively low amount of both materials.
Co-sintering generally describes the common heat treatment of the materials in a composite powder compact or in a green composite part. However, the step before is co-debinding for complete removal of any organic additives necessary for the shaping process in a powder technological route, which can have a significant effect on the corrosion behavior of the composite. The particle packing density and the interface formation between the material partners control the properties of the composite after the heat treatment. For a homogeneous structure and a high sinter density, the theoretical green density is crucial. In the case of spherical particles, they occupy an orthorhombic structure and thus achieve a maximum density of 62.5%. Multimodal particle distributions can obtain up to 97.5% [13]. For the formation of interfaces, [14] considers the wettability during the metal-ceramic composite formation. This is influenced by the presence of atomic oxygen, the predominant surface orientation of the material particles and their electrical as well as ionic conductivity. These influences become especially clear when a partial melting of the metal occurs at the maximum sintering temperature [14]. Regardless of whether a couple of particles or a layered composite is considered, local chemical processes, the bonding mechanisms and crystallographic orientation play a very important role in the formation of the interface. In addition to the wettability, the microstructure of the interface, such as the occurrence of vacancies, the oxygen loss of oxide ceramic, as well as dislocations or phase formation have a great influence on composite formation [15]. Simulations for co-sintering have been carried out only for composites of the same materials until now [16]. The mechanisms in metal-ceramic composites have to be extensively researched in order to understand the phase transformations and the effects on the corrosion behavior.
Dourandish [17] has intensively dealt with composite formation and co-sintering of stainless steels such as 420, 17-4PH and 316L [17]. In his work, he focused on the sintering process itself and the material properties as the main influencing factors during co-sintering. The coefficient of thermal expansion (CTE) and the shrinkage differences must be as close as possible to each other for defect-free composite production over the whole temperature range. The resulting interface phase is also described as a reaction layer, which compensates residual stresses in the composite. For a composite with ZrO 2 , 17-4PH is best suited for co-sintering due to its thermal properties, but no phase formation could be detected in the interface. With other steels, a phase is formed in the interface, and the difference in thermal properties can be compensated by the use of nanoscale ZrO 2 [18].
A material approach is very similar, but many factors have to be redesigned. The co-shaping, which also significantly influences the co-sintering, can be seen as simplified in Dourandish's work, because pressed composites are used without large amounts of additive components. This also reduces the influence of debinding or is no longer relevant as an influencing factor in the process. The use of 17-4PH appears to be advantageous in connection with zirconia. For this purpose, the connecting mechanism is investigated more closely and the existence of an interfacial phase will be pursued. For the production of a defect-free composite material, sintering is the greatest obstacle [18][19][20][21][22]. The residual stresses [18][19][20][21][22] play a major role in the production of functional composite. The CTE is material-dependent and thus cannot be influenced. As already described, the shrinkage can be positively modified not only by the ceramic but also by the metallic particles. The shrinking behavior of the metallic component in a layered composite can be adapted to the ceramic by selective mixing of the flat-shaped and spattered particles, so that a co-sintering of stainless steel and YSZ is possible [12].
For the co-sintering of 17-4PH and 316L components, which are joined by injection molding, Dutra observes the sintering behavior with varying parameters. The parallel simulation shows good accordance with the obtained results. It can be used for one-dimensional kinetic interfacial processes of completely dense, metallic materials, neglecting grain boundaries and vacancies [23,24]. A derivation of the probable chemical reactions during co-sintering can be made using the free enthalpies of formation of ceramics and metals published by Ondracek [25]. These primarily comprise elementary metals, but no alloys or steels. Thermodynamic data have already been used for the assessment of the interactions at the interface between the ceramic and selected alloying elements of the steel [26,27] and the verifiability of new phase formation [28,29]. For the analysis of the ceramic-sided phenomena, the experimental consideration is a fundamental task of this work.
For the production of composites made of TRIP steels and Mg-PSZ, extrusion molding is used by Aneziris et al. [30] to produce honeycomb structures with increased energy absorption properties. In addition to the superplastic properties of the steel, the stress-induced phase transformation from tetragonal to the monoclinic ZrO 2 phase plays an important role in the transformation-enhancing properties of the composite. A strength increase of the extruded honeycombs can be demonstrated [30]. The resulting transformation toughening mechanism can be illustrated by in-situ measurements of the compressive stresses and simultaneous observation of the phase transformation through EBSD and thus a correlation between local deformation and solidification effects can be made [31,32].
Other effects that affect the phase transformation of zirconia are due to the interaction with metallic ions [33]. In this case, Cr ions, but also Fe ions, can influence the transformation of monoclinic to tetragonal ZrO 2 [34][35][36][37]. Based on this, there are also studies of the oxidation behavior of ceramic matrix composites (CMC) in which the ZrO 2 also has a great influence on the corrosion behavior of the steel [38].
The corrosion behavior of materials is primarily described in relation to gaseous and liquid media. Ceramics are particularly inert and suitable for applications in contact with highly corrosive media such as hydrofluoric acid [39]. As a corrosion test for ZrO 2 ceramics as a dental prosthesis, Oblak et al. [40] uses exposure in an autoclave. The transformation of the tetragonal to the monoclinic phase lowers the strength of the oxide ceramic [40]. When milling Y-TZP from particle sizes in the submicron range to nanoscale powder, the sintered body is situated in the autoclave at 134 • C for 192 h with steam, but without corrosive phenomena. Although the grain refinement results in a reduced bending strength, it radically reduces the ability to transform from monoclinic to tetragonal [41]. Thermally sprayed Y-TZP coatings show good corrosion resistance in 0.5 molar sodium chloride solution (NaCl) as well as very good protection against sulfuric acid (H 2 SO 4 ) starting from a layer thickness of 5 µm [42]. The effect of composite formation was observed in the present study with regard to the hydrothermal corrosion behavior and the phase transformation. For this purpose, individual tapes were compared with the layers in the composite and evaluated.

Materials
The ceramic component used was TZ-3Y-E, a 3 mol% yttria-stabilized ZrO 2 (YSZ) from Tosoh. The starting powder was agglomerated in a wide particle size range and had a primary grain size of 70 nm. With the addition of 0.25 wt.-% alumina in the commercial powder, the sintering temperature was lowered. The metallic partner of the layered composite or laminate respectively was an austenitic Cr-Ni steel, which is used for medical applications. The starting powder of the 17-4PH steel (1.4542, X5CrNiCuNb16-4) from lad consisted of spherical particles with an average particle size of 10-12 µm prepared by gas atomization. The chemical composition of the steel is given in Table 1. Due to the adjustment of the sintering temperature of the YSZ, both materials can be sintered together at a temperature of 1350 • C. Figure 1 shows the shrinkage profiles of both materials. regard to the hydrothermal corrosion behavior and the phase transformation. For this purpose, individual tapes were compared with the layers in the composite and evaluated.

Materials
The ceramic component used was TZ-3Y-E, a 3 mol% yttria-stabilized ZrO2 (YSZ) from Tosoh. The starting powder was agglomerated in a wide particle size range and had a primary grain size of ~70 nm. With the addition of 0.25 wt.-% alumina in the commercial powder, the sintering temperature was lowered. The metallic partner of the layered composite or laminate respectively was an austenitic Cr-Ni steel, which is used for medical applications. The starting powder of the 17-4PH steel (1.4542, X5CrNiCuNb16-4) from lad consisted of spherical particles with an average particle size of 10-12 μm prepared by gas atomization. The chemical composition of the steel is given in Table 1. Due to the adjustment of the sintering temperature of the YSZ, both materials can be sintered together at a temperature of 1350 °C. Figure 1 shows the shrinkage profiles of both materials.  For the shrinkage measurement, a dilatometer curve of the tape laminates from the respective material was recorded. The thermogravimetric analyzes were carried out by means of the device STA 409 (Netzsch-Gerätebau GmbH, Selb, Germany). The diagram shows that the laminates had a constant shrinkage profile up to ~900 °C. The CTEs in the temperature range from room temperature to 1000 °C were matched to each other according to the material selection; YSZ had a CTE of 8-13 * 10 −6 /K and 17-4PH had one of 10-12 * 10 −6 /K. The shrinkage of both laminates starts with a temperature difference of 50 K and ends with a shrinkage difference of ~8% at 1400 °C.

Suspension Preparation and Tape Casting
The production of laminates of the same type for the dilatometer measurements, layered composites of a steel and YSZ tape as well as single tapes was achieved by tape casting. To prepare the ceramic suspension, the dispersion medium (distilled water), the dispersant (Dolapix CE 64, Zschimmer & Schwarz, Lahnstein, Germany), the defoamer (Foamaster F111, BASF, Ludwigshafen, For the shrinkage measurement, a dilatometer curve of the tape laminates from the respective material was recorded. The thermogravimetric analyzes were carried out by means of the device STA 409 (Netzsch-Gerätebau GmbH, Selb, Germany). The diagram shows that the laminates had a constant shrinkage profile up to~900 • C. The CTEs in the temperature range from room temperature to 1000 • C were matched to each other according to the material selection; YSZ had a CTE of 8-13 × 10 −6 /K and 17-4PH had one of 10-12 × 10 −6 /K. The shrinkage of both laminates starts with a temperature difference of 50 K and ends with a shrinkage difference of~8% at 1400 • C.

Suspension Preparation and Tape Casting
The production of laminates of the same type for the dilatometer measurements, layered composites of a steel and YSZ tape as well as single tapes was achieved by tape casting. To prepare the ceramic suspension, the dispersion medium (distilled water), the dispersant (Dolapix CE 64, Zschimmer & Schwarz, Lahnstein, Germany), the defoamer (Foamaster F111, BASF, Ludwigshafen, Germany), the wetting agent (Glydol N109, Zschimmer & Schwarz) and the ceramic powder were put in a ZrO 2 milling bowl with ZrO 2 milling balls (50 wt.% of powder, d = 10 mm). Subsequently, the suspension was homogenized at 300 rpm for 30 min in a planetary ball mill (PKM, pulverisette 5, Fritsch). After the first homogenization step, the plasticizer (glycerin, Carl Roth GmbH, Karlsruhe, Germany) and the PVA binder (13% Mowiol 20-98, Ter Hell GmbH, Hamburg, Germany) were added and mixed in the PKM for further 10 min at 150 rpm. The complete content of the milling bowl was transferred into a polyethylene bottle and mixed for a further 20 h on a roller mill (Multi Mix 10 Balik, Maschinenbau Wien, Vienna). After completion of the homogenization steps, the suspension was passed through a strainer to separate the milling balls and de-aired. For this purpose, the ceramic suspension was exposed to a vacuum of 100 mbar for 60 min under constant stirring. The metallic suspension was prepared by homogenizing all additives with a propeller stirrer for 10 min. Subsequently, the metal powder was added and mixed for a further 15 min. The suspension composition of YSZ and 17-4PH is shown in Table 2. The data refer to 100 g of powder. The metal tapes were produced using a discontinuous tape casting unit from Netzsch GmbH, Germany. The ceramic suspensions were cast on an antistatic carrier tape (polyester tape, Optimont, Brossard, QC, Canada) with a film casting unit from Erichsen (K Control Coater, model 624, K202) on which thin films in DIN A4 size can be produced. The coating on the film casting unit was carried out with a 40-µm roller blade. For ceramic tapes with a casting height of 170 µm, the tape casting unit could be used.

Characterization after Sintering and Hydrothermal Corrosion
The debinding took place in air up to 600 • C for 2 h (Nabertherm, Multitherm N 60/A). For the sintering at 5 K/min to 1350 • C and 2 h holding time, an atmosphere with the composition 80% Ar/20% H 2 was set in the furnace (MUT, high-temperature tungsten oven OHV 250/300-1900V).
Steam sterilization is a common method in medical technology to clean surgical instruments from any microorganisms. For this purpose, the 17-4PH/YSZ laminates were deposited in steam in accordance with DIN 58946. They must be exposed in an autoclave at 3 bar and at 134 • C for at least five hours. To test the durability of the composite samples, five specimens were deposited for 5 and 24 h, respectively. These were embedded on a Teflon holder and examined for macroscopic and microscopic abnormalities after completion of the exposure. Particular attention was paid to the YSZ layer, which is the less resistant component under hydrothermal aging.
The X-ray diffractometer (XRD) D8 Advance (Bruker AXS, Billerica, MA, USA) was used with Cu K-α radiation for phase analysis. Figure 2 shows the XRD analysis of the 17-4PH starting powder and the sintered tape. The suspension composition and the heat treatment have no major effect on the phase composition of the steel. In order to detect interfacial areas on sintered laminates, a calotte cut was performed on the ceramic side of the laminate and the XRD IT 3003TT (Seifert, Germany) was used in grazing incidence. The microstructure was examined in the sintered state for single and composite tapes and after hydrothermal aging. The laminates and single components were used to produce grinding patterns of the cross-sections of the tapes. Thereto, the test specimens were embedded in epoxy resin, ground and mechanically polished. The Epiphot 300 incident light microscope from Nikon GmbH (Düsseldorf, Germany) was used to examine the microstructure of the cross-sections at up to 500 times magnification. The cross-sections and surfaces were sputtered with carbon for microstructural characterization with a field-emission scanning electron microscope (FESEM) Ultra 55 (Carl Zeiss AG, Oberkochen, Germany) in backscattered electron (BSE) contrast. The transmission electron microscope (TEM) Zeiss Libra 200 was used for the deeper interface analysis. TEM samples were prepared by in-situ lift-out preparation using a FIB-SEM combination of the type Zeiss NVision 40. Energy-dispersive X-ray spectroscopy (EDX) was used in FESEM and TEM.

Heat Treatment of 17-4PH/YSZ Laminates
The difference in sinter shrinkage leaves stresses at the interface. Due to the higher shrinkage of the YSZ, the samples bend without any loading. In order to produce the lowest possible stress, laminates with thin YSZ layers (≤100 μm) were produced with metal substrate layers (~1000 μm). After the heat treatment, a thin and dense YSZ layer was visible which was connected to an also dense metal layer. As shown in the micrograph of Figure 3, the YSZ layer (dark grey) was on the top of the metal substrate (light grey), which was about 10-20 times thicker, and had an oxide-free edge zone and an oxide-rich inner region (filled dark grey pores), which is due to the binder removal in air. Debinding in inert or reducing atmospheres results in porous YSZ layers, which should be avoided under application-specific aspects. The black region around the composite sample is epoxy filler as investment material. The microstructure was examined in the sintered state for single and composite tapes and after hydrothermal aging. The laminates and single components were used to produce grinding patterns of the cross-sections of the tapes. Thereto, the test specimens were embedded in epoxy resin, ground and mechanically polished. The Epiphot 300 incident light microscope from Nikon GmbH (Düsseldorf, Germany) was used to examine the microstructure of the cross-sections at up to 500 times magnification. The cross-sections and surfaces were sputtered with carbon for microstructural characterization with a field-emission scanning electron microscope (FESEM) Ultra 55 (Carl Zeiss AG, Oberkochen, Germany) in backscattered electron (BSE) contrast. The transmission electron microscope (TEM) Zeiss Libra 200 was used for the deeper interface analysis. TEM samples were prepared by in-situ lift-out preparation using a FIB-SEM combination of the type Zeiss NVision 40. Energy-dispersive X-ray spectroscopy (EDX) was used in FESEM and TEM.

Heat Treatment of 17-4PH/YSZ Laminates
The difference in sinter shrinkage leaves stresses at the interface. Due to the higher shrinkage of the YSZ, the samples bend without any loading. In order to produce the lowest possible stress, laminates with thin YSZ layers (≤100 µm) were produced with metal substrate layers (~1000 µm). After the heat treatment, a thin and dense YSZ layer was visible which was connected to an also dense metal layer. As shown in the micrograph of Figure 3, the YSZ layer (dark grey) was on the top of the metal substrate (light grey), which was about 10-20 times thicker, and had an oxide-free edge zone and an oxide-rich inner region (filled dark grey pores), which is due to the binder removal in air. Debinding in inert or reducing atmospheres results in porous YSZ layers, which should be avoided under application-specific aspects. The black region around the composite sample is epoxy filler as investment material. Up to 600 °C, the steel had a low oxidation rate of 0.25 ma.-%, which exponentially increased above approx. 800 °C. The resulting oxides from the debinding and the oxides which already existed before on the powder surfaces of the steel form the inner oxides after sintering. Directly at the interface, oxide formation was also found. The edge area was reduced before the initial shrinkage in the hydrogen-rich sintering and could thus be completely compacted.

Interface of 17-4PH/YSZ Laminate
The interface of the 17-4PH/YSZ laminate was characterized by a continuous connection of the ceramic and metallic layers, which allows a form-fitting composite in which the fine YSZ particles completely fill the intermediate spaces of the coarse steel particles. There were no critical cracks or delaminations between the two layers. In Figure 4, an interfacial phase can be identified at the interface of the thin YSZ-layer and the metal substrate, which represents an additional material bond of the laminate. This occured directly at the transition between metallic and ceramic layers and is clearly identifiable in the BSE contrast by the dark gray coloration. By means of EDX, the resulting interface phase was characterized element-specifically. In addition to the main elements of the two material components Fe and Zr, oxygen, chromium, manganese, silicon, nickel, copper, and carbon could also be detected by the qualitative analysis (the secondary elements are listed downward according to the mass fraction in the interface phase). A quantitative evaluation as well as the stoichiometric composition of the phases is not based on the Up to 600 • C, the steel had a low oxidation rate of 0.25 ma.-%, which exponentially increased above approx. 800 • C. The resulting oxides from the debinding and the oxides which already existed before on the powder surfaces of the steel form the inner oxides after sintering. Directly at the interface, oxide formation was also found. The edge area was reduced before the initial shrinkage in the hydrogen-rich sintering and could thus be completely compacted.

Interface of 17-4PH/YSZ Laminate
The interface of the 17-4PH/YSZ laminate was characterized by a continuous connection of the ceramic and metallic layers, which allows a form-fitting composite in which the fine YSZ particles completely fill the intermediate spaces of the coarse steel particles. There were no critical cracks or delaminations between the two layers. In Figure 4, an interfacial phase can be identified at the interface of the thin YSZ-layer and the metal substrate, which represents an additional material bond of the laminate. This occured directly at the transition between metallic and ceramic layers and is clearly identifiable in the BSE contrast by the dark gray coloration. Up to 600 °C, the steel had a low oxidation rate of 0.25 ma.-%, which exponentially increased above approx. 800 °C. The resulting oxides from the debinding and the oxides which already existed before on the powder surfaces of the steel form the inner oxides after sintering. Directly at the interface, oxide formation was also found. The edge area was reduced before the initial shrinkage in the hydrogen-rich sintering and could thus be completely compacted.

Interface of 17-4PH/YSZ Laminate
The interface of the 17-4PH/YSZ laminate was characterized by a continuous connection of the ceramic and metallic layers, which allows a form-fitting composite in which the fine YSZ particles completely fill the intermediate spaces of the coarse steel particles. There were no critical cracks or delaminations between the two layers. In Figure 4, an interfacial phase can be identified at the interface of the thin YSZ-layer and the metal substrate, which represents an additional material bond of the laminate. This occured directly at the transition between metallic and ceramic layers and is clearly identifiable in the BSE contrast by the dark gray coloration. By means of EDX, the resulting interface phase was characterized element-specifically. In addition to the main elements of the two material components Fe and Zr, oxygen, chromium, manganese, silicon, nickel, copper, and carbon could also be detected by the qualitative analysis (the secondary elements are listed downward according to the mass fraction in the interface phase). A quantitative evaluation as well as the stoichiometric composition of the phases is not based on the By means of EDX, the resulting interface phase was characterized element-specifically. In addition to the main elements of the two material components Fe and Zr, oxygen, chromium, manganese, silicon, nickel, copper, and carbon could also be detected by the qualitative analysis (the secondary elements are listed downward according to the mass fraction in the interface phase). A quantitative evaluation as well as the stoichiometric composition of the phases is not based on the EDX spectra, since the inaccuracy of the method does not allow any reliable statement about this, mainly due to the activation volume of the X-ray. The carbon content can be attributed to the previous sputtering of the sample surface, because of which the detection of carbon in the composite is only conditionally possible. It should also be assumed that the phases are oxides. These consisted mainly of the alloying elements, possibly of elements of the YSZ layer and the remaining oxygen from the debinding atmosphere. It can therefore be assumed that similar phases were formed as in the steel interior. EDX mappings, which were prepared in TEM and are shown in Figure 5, confirm the formation of chromium-and silicon-rich oxides as well as pure copper precipitations at the interface.
Ceramics 2020, 3 FOR PEER REVIEW 8 EDX spectra, since the inaccuracy of the method does not allow any reliable statement about this, mainly due to the activation volume of the X-ray. The carbon content can be attributed to the previous sputtering of the sample surface, because of which the detection of carbon in the composite is only conditionally possible. It should also be assumed that the phases are oxides. These consisted mainly of the alloying elements, possibly of elements of the YSZ layer and the remaining oxygen from the debinding atmosphere. It can therefore be assumed that similar phases were formed as in the steel interior. EDX mappings, which were prepared in TEM and are shown in Figure 5, confirm the formation of chromium-and silicon-rich oxides as well as pure copper precipitations at the interface. The formed interface phase (1) manly consisted of chromium, oxygen and copper, while the steel substrate (2) showed a composition of iron, chromium, nickel and copper. For the distribution of the formed phase over the whole interface, no reliable statement can be made due to the electron microscopy results for the 17-4PH/YSZ laminate. When using destructive methods, in order to be able to perform a surface analysis of the interface, the laminate always fails in the ceramic layer and never at the interface.

Properties of YSZ Near the Interface
The YSZ layer showed a homogeneous structure in the FESEM images over the entire layer thickness up to the interface. In additional TEM images, a difference can be observed between the grains next to the interface and those which are still located in the interior layer. Figure 6 shows a vertically laminated interface, from which the steel substrate is to the left and the ceramic layer is to the right. The formed interface phase (1) manly consisted of chromium, oxygen and copper, while the steel substrate (2) showed a composition of iron, chromium, nickel and copper. For the distribution of the formed phase over the whole interface, no reliable statement can be made due to the electron microscopy results for the 17-4PH/YSZ laminate. When using destructive methods, in order to be able to perform a surface analysis of the interface, the laminate always fails in the ceramic layer and never at the interface.

Properties of YSZ Near the Interface
The YSZ layer showed a homogeneous structure in the FESEM images over the entire layer thickness up to the interface. In additional TEM images, a difference can be observed between the grains next to the interface and those which are still located in the interior layer. Figure 6 shows a vertically laminated interface, from which the steel substrate is to the left and the ceramic layer is to the right. Note the different appearance of the ceramic grains in a region of 1 μm next to the interface (dashed black line). Grain size a greater distance from the interface seems to be smaller, and grains appear in more different grey values, which hints at stronger differences in orientation. A comparison of elemental constitution close to the interface and a greater distance away (Figure 7) shows no difference in this first measurement. For composite formation, these results indicate that, due to the shrinkage-induced stresses during the sintering, no monoclinic phase transformation of the zirconia took place.

Hydrothermal Corrosion of Single Tapes and Laminates
On composite and single tapes, FESEM images ( Figure 8) were taken after 24 h at 134 °C under hydrothermal corrosion conditions. Due to the steam, a hydrothermal attack on the surface is visible on the ceramic layers. However, it is not possible to determine to what extent this superficial ablation affects the dense layer and whether monoclinic phase transformation took place. Note the different appearance of the ceramic grains in a region of 1 µm next to the interface (dashed black line). Grain size a greater distance from the interface seems to be smaller, and grains appear in more different grey values, which hints at stronger differences in orientation. A comparison of elemental constitution close to the interface and a greater distance away (Figure 7) shows no difference in this first measurement. Note the different appearance of the ceramic grains in a region of 1 μm next to the interface (dashed black line). Grain size a greater distance from the interface seems to be smaller, and grains appear in more different grey values, which hints at stronger differences in orientation. A comparison of elemental constitution close to the interface and a greater distance away (Figure 7) shows no difference in this first measurement. For composite formation, these results indicate that, due to the shrinkage-induced stresses during the sintering, no monoclinic phase transformation of the zirconia took place.

Hydrothermal Corrosion of Single Tapes and Laminates
On composite and single tapes, FESEM images ( Figure 8) were taken after 24 h at 134 °C under hydrothermal corrosion conditions. Due to the steam, a hydrothermal attack on the surface is visible on the ceramic layers. However, it is not possible to determine to what extent this superficial ablation affects the dense layer and whether monoclinic phase transformation took place. For composite formation, these results indicate that, due to the shrinkage-induced stresses during the sintering, no monoclinic phase transformation of the zirconia took place.

Hydrothermal Corrosion of Single Tapes and Laminates
On composite and single tapes, FESEM images ( Figure 8) were taken after 24 h at 134 • C under hydrothermal corrosion conditions. Due to the steam, a hydrothermal attack on the surface is visible on the ceramic layers. However, it is not possible to determine to what extent this superficial ablation affects the dense layer and whether monoclinic phase transformation took place. No corrosion removal is seen on the steel surfaces; only deposits of impurities after the process of exposure are visible. A remarkable result after the aging in steam can be noted at the YSZ layer of the composite. The diffraction pattern shown in Figure 9 has been measured with the grazing incidence of the X-ray beam. In order to be able to measure near the interface, a calotte cut is made, which penetrates into the steel substrate. The measurement takes place in the ceramic layer at the interface.  No corrosion removal is seen on the steel surfaces; only deposits of impurities after the process of exposure are visible. A remarkable result after the aging in steam can be noted at the YSZ layer of the composite. The diffraction pattern shown in Figure 9 has been measured with the grazing incidence of the X-ray beam. In order to be able to measure near the interface, a calotte cut is made, which penetrates into the steel substrate. The measurement takes place in the ceramic layer at the interface. No corrosion removal is seen on the steel surfaces; only deposits of impurities after the process of exposure are visible. A remarkable result after the aging in steam can be noted at the YSZ layer of the composite. The diffraction pattern shown in Figure 9 has been measured with the grazing incidence of the X-ray beam. In order to be able to measure near the interface, a calotte cut is made, which penetrates into the steel substrate. The measurement takes place in the ceramic layer at the interface.  There was no monoclinic phase of the YSZ layer to be found in the region next to the interface. The strongest corrosive attack by hydrothermal aging for an YSZ was not detectable through XRD. The strongest monoclinic reflections, which are at 28.2 • and 31.5 • , were not present and no phase transformation could be observed. In the case of identical tests with non-corroded laminates, which may have a monoclinic phase transformation due to the shrinkage difference during sintering, the same diffraction pattern could be seen.
Based on the main reflex of the Fe-Cr steel, typical at about 44 • , the actual proximity to the interface can be illustrated. The directly underlying steel layer is embraced by the activation volume and duly appears as a peak in the diffraction pattern. It can be assumed that there are alloying elements, which influence the YSZ layer to a higher hydrothermal stability. In the ceramic layer, these elements cannot be detected by EDX.

Conclusions
In addition to the formation of an interface phase, it could be demonstrated that the available methods were unable to detect a tetragonal to monoclinic conversion of the YSZ layer. After the sintering of the 17-4PH/YSZ laminate at 1350 • C in an 80% Ar/20% H 2 atmosphere with a shrinkage difference of the two materials of~8%, no monoclinic phase could be detected through EDX and XRD. This refers to the fact that there are no significant stresses during the production and heat treatment of the composite. After hydrothermal corrosion for 24 h at 134 • C, no phase transformation with the described characterization methods could be observed either. The YSZ layer shows an improved hydrothermal stability. One possible explanation for this behavior is the influence of the alloying elements of the steel, which must be proven in continuing work.
Furthermore, the reduction of the oxide formation in the interior of the steel layer should be effected with a suitable sintering regime. At the same time, sufficient hydrogen from the atmosphere must penetrate the steel layer and reduce the oxidized 17-4PH particles before the shrinkage begins and the oxides are included.