ICMCTF2019 Session B7-TuA: Plasma Diagnostics and Growth Processes

This work reports the study of the growth mechanisms involved in TiO_x thin film deposition by magnetron sputtering. An Ar-O₂ plasma chemistry obtained by dual magnetron setup operating in DC mode is used. Growth from both metallic and TiO_1.8 targets are compared. Isotopic ¹⁸O₂ is used as reactive gas for all different configurations. The aim is to differentiate oxygen coming from ceramic targets and oxygen coming from the gas phase. Indeed, using ion beam analysis techniques such as Rutherford Backscattering Spectroscopy it is possible to precisely analyze Ti, ¹⁶O and ¹⁸O content of the samples. Other investigation techniques such as AFM, SEM … are also used to fully characterize deposited coatings. In order to study the effect of energetic ions bombardment of samples during deposition, 70° and normal incidence samples are compared. In addition to sample’s characterization, Langmuir probe and energy flux probe are used to obtain plasma phase properties.

In a second part, simulations tools are used to predict both discharge and coating’s properties. In order to simulate the complete plasma process, three different software are used. Each one is handling a defined step of the process (i) neutral particle motion, (ii) charged particle motion and (iii) film growth. We will discuss the comparison between simulation predictions and experimental investigations. Those two approaches allow us to achieve a better understanding on the growth of oxide layers by reactive magnetron sputtering and how plasma parameters influence coating properties.

Reactive high power impulse magnetron sputtering (R-HiPIMS) offers a great opportunity for high quality coating production thus understanding the processes accompanying deposition is of great importance. The hysteresis curve in R-HiPIMS generally exhibits a narrower shape compared to dcMS, or it can even be entirely suppressed, which is beneficial for high-rate deposition of stoichiometric compound films. The main reason of the hysteresis suppression is not yet completely understood. A recently developed effective branching fraction method is utilized to determine absolute ground state number densities of sputtered titanium species from the optical-emission signal. We report on evolutions of titanium atom and ion ground state densities in R-HiPIMS discharges in oxygen, nitrogen and acetylene gases for constant mean power and pulse duration, when varying the repetition frequency. A fast feedback system is employed to allow working in the transition region of the hysteresis curve in a well-controlled manner. The ionization fraction of sputtered species increases with the partial pressure of the reactive gas. The increased ionization of titanium is attributed to the combination of the following effects: a longer residual time of sputtered species in the target vicinity; a higher maximal discharge current attained at the end of the pulse; lower amount of sputtered species due to the target poisoning which may positively affect electron distribution function. It is furthermore found that the hysteresis curve shape changes when varying the repetition frequency at the same mean power. The difference is more pronounced for R-HiPIMS with higher sputtered species ionization fraction. The experimental results are compared to the results obtained by a reactive ionization region model (R-IRM). The absolute ground state number densities of Ti atoms and Ti ions measured at the target vicinity are also substituted into the Berg model modified to include ion back attraction, and a rather good match between the measurements and simulation results for different experimental conditions is found.

Structure zone models give an overview of the microstructure as a function of the deposition conditions. It has been shown by our research team that these overviews can be interpreted in a quantitative way by studying the ratio between the diffusivity (D) and the deposition flux (F)[1-3]. The deposition flux can easily be derived from the deposition rate. To calculate the diffusivity the available energy per deposited atom (EPA) needs to be quantified which becomes possible by measuring the total energy flux with a passive calorimetric probe. In this paper, this approach is applied to understand the phase formation during reactive sputtering of copper in an argon/mixture. The influence of the total pressure, and the discharge current was investigated. Within the experimental range hardly any changes in the EPA could be noticed, illustrating that the phase formation in the case of copper oxide thin films is solely defined by the oxygen partial pressure in the system. In the case of pure tenorite (CuO) thin films deposited at relative high oxygen partial pressures, the EPA could by increased, and it is shown that this leads to less crystalline films. The origin of this behavior is further investigated by energy-resolved mass spectrometry.

[1] Review paper : S. Mahieu, D. Depla, Journal of Physics D: Applied Physics 42 (2009) 053002

[2] Critical review paper : D. Depla, B. Braeckman, Thin Solid Films, (2016) 90-93

[3] J. Xia, W. Liang, Q. Miao, D. Depla, Applied Surface Science 439 (2018) 545–551

Transparent and conductive materials are important for a large number of applications including: touch screens, solar cells, smart windows and light emitting diodes. Oxides doped with metals, generically known as transparent conductive oxides (TCO) are successfully used nowadays, with indium tin oxide (ITO) as the best material. However, the high demand for large area applications, conflicts with the reduced abundance of indium. This motivation sustains an intensive research on alternative materials with aluminum doped zinc oxide (AZO) as one of the most promising choices. There are several methods used to deposit AZO. Among them, magnetron plasma sputtering is successfully used for ITO on large area substrates and it is also investigated as a viable cost effective method for AZO. However, the resistivity of AZO thin films is about 5 to 10 time higher than ITO, with promising values only for limited areas on the substrate. One of the main reasons is the electronegativity of oxygen that forms negative ions, resulting in a growth mechanism assisted by energetic ions with a spatial distribution correlated with the erosions tracks. In this context, a proper deposition process requires spatially-resolved plasma diagnostics in direct correlation with spatially-resolved thin film properties. This work reviews the status in plasma diagnostics during TCO growth with special emphasis on AZO. New results by mass spectrometry, optical emission and probes are also presented, both for disk and rotatable cathodes operated in DC, RF and MF using oxide (ZnO/Al₂O₃, 2, 2.5 and 3% Al) and metallic targets (Zn, 2 and 3% Al). Pressure, target to substrate distance and discharge power have been investigated as discharge parameters with the aim of obtaining a resistivity below 10^-3 W cm over the whole length of the sample (50 mm for disk cathode and 200 mm for rotatable cathode). The resistivity, transmittance and film thickness were measured with a spatial resolution of 2 mm. XPS, XRD, TOF-SIMS and SEM were used for surface characterization. It is shown that the AZO resistivity can change with two orders of magnitude over 10 mm span on the substrate, a behavior that can be correlated with plasma parameters and the growth mechanism.

The preparation of ultra-thin film is crucial for the development of cutting-edge technologies in the field of microelectronics, optoelectronics, nanotechnology or catalysts. Furthermore, covering 3-D objects in a nanometer scale requires a high degree of uniformity of deposited thin films preserving high aspect ratio of complex shape objects. Atomic layer deposition (ALD) has proven to be almost an indispensable deposition technique for conformal deposition of mostly metal or oxide thin films. Plasma-enhanced ALD (PE-ALD) process brings energy for surface reactions between precursors and reactants through electrons, ions, radicals or excited particles. Many studies have shown that the PE-ALD process is able to operate at significantly lower substrate temperatures. In this work, a microwave surfatron plasma source as alternative to CCP, ICP or other remote plasma sources was used for activation of reactants during the PE-ALD process. We carried out measurements with the Langmuir probe to obtain spatial map of electron temperature, plasma density and potentials in the ALD chamber designed for deposition on wafers with diameter 100 mm during typical deposition conditions of TiO₂, Al₂O₃ and TiN thin films. The plasma parameters were investigated for different working gas pressures and mixtures. Obtained results proved that radial homogeneity of the plasma density could be improved if mass flow rate of working gas is reduced. On the other hand, for higher pressure of the working gas the plasma density rapidly decreased in axial distance from the surfatron nozzle outlet. Results clearly demonstrated that spatial inhomogeneity of the plasma parameters correlates with thin film properties. The possibilities of deployment of the multi-nozzle surfatron system for achieving a sufficient level of homogenization of the plasma parameters was suggested. Effect of substrate temperature and microwave power delivered into the surfatron on deposited thin film properties was studied too. The Raman spectroscopy proved that substrate temperature above 200°C led to anatase phase formation of TiO₂ thin films whilst lower temperatures produced amorphous thin films. Further, gradually increased microwave power resulted in the rise of thin film thickness measured by the spectroscopic ellipsometry for constant number of ALD cycles. Impact of the microwave power on the Growth per Cycle (GPC) parameter was studied by the optical emission spectroscopy detecting evolution of oxygen or nitrogen atoms and radicals.

The authors would like to thank for the financial support from Technology Agency of the Czech Republic under project no. TF03000025.