9.3 Thin film deposition: Issues

Even though thin film deposition of epitaxial metal oxides has been carried out since the 1960s and significant progress has been made on all fronts, yet there are some fundamental issues which are yet to be dealt with. To reduce the deposition temperature for depositing epitaxial films is the foremost challenge for most of the processing techniques, barring some chemical processes. Other aspects involve production of stable phases with precise composition and correct oxygen content, good structure and morphology of the films, high degree of desired orientation, and issues regarding the nucleation and growth of the films. Choice of device compatible and cheap substrates and deposition over larger area substrates are the main bottlenecks in the commercial success. The interdependence of various parameters, for example, processing temperature and epitaxy, poses serious questions about the choice of optimum conditions. In this section, it has been attempted to briefly elucidate these problems and the ongoing efforts to handle them.

Most of the metal oxides require high temperature for epitaxial growth, usually growth temperatures being 0.7-0.8Tm which is very high when compared to epitaxial growth temperatures of metals and semiconductors which fall usually between 0.2-0.5 T_m. These temperatures are very high when considering the processing compatibility with semiconductor circuit processing. The effect of deposition temperature is direct on the epitaxy, although it also depends upon the type of processes. Usually if the deposition temperature is lowered, then epitaxy deteriorates. But there are a few reports of deposition at lower temperatures by MOCVD processes²⁶ or evaporation processes²⁶ maintaining the desired orientation of the films. It has been shown for ion beam sputtered PZT films that bombardment of the growing film by low energy ions in ion beam sputtering and use of active oxygen species helps to reduce the crystallization temperature of PZT to some extent. But it is still not possible to deposit films at lower temperatures (below 500°C), with desired epitaxy, in processes like laser ablation or sputtering.

9.3.2 Phase Stability and Stoichiometry Control

Because of the thin film growth processes in general being non-equilibrium in nature, it becomes very important to know about the phase evolution under a given set of process parameters and how a desired phase can be obtained if they are altered accordingly. This problem arises predominantly in terms of deposition temperature and oxygen activity. Most of the perovskite oxides are not only complex oxides, but they also have some specific oxygen level for desired properties e.g. superconducting oxides like YBCO. These complexities put enormous restriction on the processing conditions to produce the correct phase. The work of Feenstra et al.³² on the effect of oxygen partial pressure and temperature on the stability of YBCO, gives a very good idea about operating windows for different processes.

Chemical complexity of these oxides also poses problems regarding the correct stoichiometry of the films. This is particularly difficult in sputtering because of the different sputter yields of different elements which cannot be controlled independently when using a compound target and due to resputtering effects on the growing film. This problem has been tackled by using multi-elemental targets.¹⁰ Also each element has different collision cross section with the plasma species because of different atomic masses which leads to a complexity in the transport processes during sputtering. Laser ablation has emerged as one of the most promising processes to address this problem with considerable success by using single compound oxide target and it has been demonstrated successfully that films can be deposited without any loss of stoichiometry.³³

9.3.3 Substrate Effects

The choice of substrate is a very important factor in the growing epitaxial oxide films. It has been a matter of great concern in growing superconducting and ferroelectric oxide thin films. SrTiO₃ (100) and MgO (100) have proven to be useful substrates for growing highly epitaxial films of YBCO. But moisture sensitiveness of MgO and high dielectric constant of SrTiO₃ makes them unattractive. They are also expensive. Numerous other substrates have been tried such as LaAlO₃, sapphire, NdGaO₃ etc. but the search is still on for the best which is inexpensive, should not react with the film at the film deposition temperature and should be able to give high quality epitaxial films with excellent properties.¹⁶ Ferroelectric thin films such as PZT have been grown epitaxially on platinised Si wafers (Pt/SiO₂/Si) but films on Pt give very poor fatigue characteristics which is very important for ferroelectric memory devices. This problem has forced researchers to look for other alternative electrodes such as LSCO, YBCO etc. but they have not yet fulfilled the promise because of some compromise required in other properties. These problems have been, to some extent, alleviated by the use of buffer layers. Buffer Layers act as intermediate layers between the substrate and film and minimise the various deleterious effects on epitaxy from factors such as mismatch in lattice parameters and coefficient of thermal expansion and interdiffusion of elements, and sometimes promote crystallisation of the film.³⁴ Various materials have been used for this purpose such as cerium oxide (CeO₂) for YBCO³⁵, and strontium ruthenium oxide (SrRuO₃) for PZT³⁶ and resulted in remarkable improvement in the device performance.

9.3.4 Structure and Morphology

The structure of the epitaxial oxide thin films such as YBCO has been characterised by X-ray diffraction extensively and it has been found that the best films grown have rocking curve width of the order of 0.15-20°.⁹ This degree of orientation is acceptable but when compared to substrate, it shows that there is a lot of scope for further improvement. The better the epitaxy of the film, better is the surface roughness. STM studies reveal that films grown by layer by layer growth mechanism are smoother than the films grown by 3-D island growth mechanism.³⁷ Stoichiometry control has also an important role to play in controlling the film structure. It was shown that by controlling the chemical composition correctly, films can be grown by layer by layer growth mode and this could also lead to smoother surface.³⁸ The effect of atomic oxygen on the growth and surface roughness of YBCO films grown on SrTiO₃ (100) by MBE process, has also been studied.³⁹ It was concluded that in the presence of atomic oxygen and with careful control of process parameters (low pO₂ and low deposition rate), it is possible to grow epitaxial films by 2-D growth mechanism and achieve atomically flat surfaces. So there are challenges to produce epitaxial oxide films with smoother surfaces and perfect orientation which require careful control of processing conditions.