Abstract:
Tungsten carbide (WC) is a widely used material for cutting tools due to its high mechanical hardness and friction toughness. When exposed to high temperature (e.g. due to friction), oxidation occurs and transforms the stable phase of tungsten carbide to a powdery phase. To avoid oxidation and enhance the tool lifetime, protective layers are deposited on top of WC substrates consisting of various oxides or nitrides, either as single layers or as multilayer structures. For physical vapor deposition (PVD) processes (e.g. magnetron sputtering), a bias potential is applied to the substrate to achieve the surface energy necessary for the growth process. When growing insulating layers (such as most of the metal oxides), the bias potential is reduced during growth, leading to poorer crystallinity as the layer thickness increases. Therefore, the development of a semi conductive oxide with advantageous mechanical properties for coating applications is of great interest. In this thesis, aluminum oxide (Al 2 O 3 ) was deposited on tungsten carbide substrates using dual magnetron sputtering (DSM) with argon as the sputtering gas. The deposition conditions during the growth process provide many different parameters that influence the crystal growth. A combination of argon and oxygen pressure during sputtering, together with the bias potential and substrate temperature was identified as necessary to achieve an electrically conductive aluminum oxide phase. The goal of this thesis was to analyze the conductive aluminum oxide phase with respect to the structural and electrical features. In addition to the enhanced quality of wear resistant coatings for cutting applications, conductive aluminum oxide is an interesting new member of the transparent conductive oxide (TCO) family. Strong efforts are made to find a replacement for the most widely used material indium tin oxide (ITO), since it relies on expensive and rare materials. A conductive aluminum oxide only requires easily available materials and is therefore an excellent candidate to play an important role in future TCO developments. Samples analyzed in this thesis were provided by Walter AG in Tübingen. The samples consist of WC substrates (for cutting applications) and wear resistive coatings of aluminum oxide were deposited using DMS. A subset of the samples was coated with an additional Ti 0.33 Al 0.67 N layer using high power impulse magnetron sputtering (HIPIMS). The structure of the conductive phase was identified as a disordered cubic structure, containing aluminum, oxygen and argon up to mole fractions of 2.5 at%. To make a distinction to insulating disordered cubic aluminum oxide phases known as γ aluminum oxide, the conductive phase is called pseudo-γ aluminum oxide. Due to its unique structure (known as disordered spinel) of γ aluminum oxide, it is possible to incorporate rather large amounts of noble gas atoms into the structure. But only when argon is incorporated in the growth using DMS, the conductive phase is formed. The Ar atoms seem to stabilize the pseudo-γ phase. To get a grasp on the role of argon in the formation of the conductive phase, the structural feasen to cover a wide range of samples properties such as film thickness, deposition bias, but also different hardness values. Quantitative energy-dispersive x-ray spectroscopy (EDX) in a scanning electron microscope (SEM) was carried out to trace the Ar mole fraction and to establish a correlation thereof to the sample hardness. By precisely determining the Ar mole fraction in plan-view as well as cross-section in SEM EDX point scans, the argon mole fraction was successfully correlated to the sample hardness, showing that the Ar atoms are essential to the forming of the stable structure. In addition, a shift in the lattice spacing towards a larger unit cell was observed when more Ar was incorporated into the pseudo-γ aluminum oxide phase. By performing SEM EDX linescans in cross-section, the Ar distribution was found to be nonuniform. The highest Ar amount was always found at the interface to the WC substrate, but different content profiles were found for the rest of the sample, ranging from a linear decay throughout the film to a second, smaller maximum in the Ar content. TEM investigation revealed that the differences in the Ar mole fraction are in fact linked to structures features of the pseudo-γ aluminum oxide phase. For both samples under investigation in the TEM, a multilayer structure was found at the interface to the substrate with high argon mole fractions up to 10 at%. For the rest of the sample, different structures were found for the sample with higher and lower hardness. The high hardness sample showed columnar grains with a constant Ar mole fraction at ≈ 2 at%, whereas the low hardness sample showed much smaller grains with decreasing Ar mole fraction towards the top of the film. In terms of electrical characterization, two different contacting methods were successfully applied, using DC sputtered Au contacts and Ti 0.33 Al 0.67 N contacts deposited using HIPIMS. Such contacts showed low ohmic electrical resistance values at room temperature and semiconducting behavior in temperature dependent measurements. The same samples were also used to determine the ratio of the thermopower of pseudo-γ aluminum oxide to its heat conductivity, revealing a positive sign of the thermopower. Temperature dependent measurements of the electrical resistance were also performed on samples with Ti 0.33 Al 0.67 N on top of pseudo-γ aluminum oxide. All samples showed curved lines in Arrhenius plots of the electrical conductivity as is expected for a polycrystalline semiconductor. By carefully choosing the right model to explain the curvature of the reference samples, it is possible to gain insight on the structural features that are essential for the applied model, such as grain structure and homogeneity of the films. For further use of pseudo-γ aluminum oxide as a transparent conductive oxide, the possibility to transfer the structure to different substrates was successfully investigated. Films were deposited on WC substrate as a reference and on electrically insulating SiO 2 , leading to the same pseudo-γ aluminum oxide phase in both cases. All applied methods of characterization such as x-ray diffraction (XRD), SEM and SEM EDX showed that the pseudo-γ aluminum oxide layers are identical and do not depend on the substrate material. Future work still needs to be done on the optimization of the contacting procedure of the pseudo-γ films. Especially the deposition of structured contacts on samples with insulating SiO 2 substrates is important to determine electrical resistivity and absolute thermopower values that can lead to a deeper understanding of the pseudo-γ aluminum oxide phase. Due to its cheap materials, pseudo-γ aluminum oxide is an excellent candidate for future use as a transparent conductive oxide. tures as well as the electrical properties were investigated in detail in this thesis. Investigations were performed both in plan-view and in cross-section. The samples were chosen to cover a wide range of samples properties such as film thickness, deposition bias, but also
different hardness values. Quantitative energy-dispersive x-ray spectroscopy (EDX) in a scan-
ning electron microscope (SEM) was carried out to trace the Ar mole fraction and to establish
a correlation thereof to the sample hardness.
By precisely determining the Ar mole fraction in plan-view as well as cross-section in SEM
EDX point scans, the argon mole fraction was successfully correlated to the sample hardness,
showing that the Ar atoms are essential to the forming of the stable structure. In addition, a shift
in the lattice spacing towards a larger unit cell was observed when more Ar was incorporated
into the pseudo-γ aluminum oxide phase.
By performing SEM EDX linescans in cross-section, the Ar distribution was found to be non-
uniform. The highest Ar amount was always found at the interface to the WC substrate, but
different content profiles were found for the rest of the sample, ranging from a linear decay
throughout the film to a second, smaller maximum in the Ar content.
TEM investigation revealed that the differences in the Ar mole fraction are in fact linked to
structures features of the pseudo-γ aluminum oxide phase. For both samples under investiga-
tion in the TEM, a multilayer structure was found at the interface to the substrate with high
argon mole fractions up to 10 at%. For the rest of the sample, different structures were found
for the sample with higher and lower hardness. The high hardness sample showed columnar
grains with a constant Ar mole fraction at ≈ 2 at%, whereas the low hardness sample showed
much smaller grains with decreasing Ar mole fraction towards the top of the film.
In terms of electrical characterization, two different contacting methods were successfully ap-
plied, using DC sputtered Au contacts and Ti 0.33 Al 0.67 N contacts deposited using HIPIMS. Such
contacts showed low ohmic electrical resistance values at room temperature and semiconduct-
ing behavior in temperature dependent measurements. The same samples were also used to
determine the ratio of the thermopower of pseudo-γ aluminum oxide to its heat conductivity,
revealing a positive sign of the thermopower.
Temperature dependent measurements of the electrical resistance were also performed on sam-
ples with Ti 0.33 Al 0.67 N on top of pseudo-γ aluminum oxide. All samples showed curved lines in
Arrhenius plots of the electrical conductivity as is expected for a polycrystalline semiconductor.
By carefully choosing the right model to explain the curvature of the reference samples, it is
possible to gain insight on the structural features that are essential for the applied model, such
as grain structure and homogeneity of the films.
For further use of pseudo-γ aluminum oxide as a transparent conductive oxide, the possibility
to transfer the structure to different substrates was successfully investigated. Films were de-
posited on WC substrate as a reference and on electrically insulating SiO 2 , leading to the same
pseudo-γ aluminum oxide phase in both cases. All applied methods of characterization such as
x-ray diffraction (XRD), SEM and SEM EDX showed that the pseudo-γ aluminum oxide layers
are identical and do not depend on the substrate material.
Future work still needs to be done on the optimization of the contacting procedure of the
pseudo-γ films. Especially the deposition of structured contacts on samples with insulating SiO 2
substrates is important to determine electrical resistivity and absolute thermopower values that
can lead to a deeper understanding of the pseudo-γ aluminum oxide phase. Due to its cheap
materials, pseudo-γ aluminum oxide is an excellent candidate for future use as a transparent
conductive oxide.