Controlling surface statistical properties using bias voltage: Atomic force microscopy and stochastic analysis

To perform a quantitative study on surfaces roughness, analytical and numerical treatments of simple growth models propose, quite generally, the height fluctuations have a selfsimilar character and their average correlations exhibit a dynamic scaling form 1,2,3,4,5,6 . In these models, roughness of a surface is a smooth function of the sample size and growth time (or thickness) of films. In addition, other statistical quantities such as the average frequency of positive slope level crossing, the probability density function (PDF), as well as drift and diffusion coefficients provide further complete analysis on roughness of a surface. Very recently, it has been shown that, by using these statistical variables in the Langevin equation, regeneration of rough surfaces with the same statistical properties of a nanoscopic imaging is possible 7,8 .

In practice, one of the effective ways to modify roughness of surfaces is applying a negative bias voltage during deposition of thin films 10 , while their sample size and thickness are constant. In bias sputtering, electric fields near the substrate are modified to vary the flux and energy of incident charged species. This is achieved by applying either a negative DC or RF bias to the substrate. Due to charge exchange processes in the anode dark space, very few discharge ions strike the substrate with full bias voltage. Rather a broad low energy distribution of ions bombard the growing films.

Generally, bias sputtering modifies film properties such as surface morphology, resistivity, stress, density, adhesion, and so on through roughness improvement of the surface, elimination of interfacial voids and subsurface porosity, creation of a finer and more isotropic grain morphology, and the elimination of columnar grains 10 .

In this work, the effect of bias voltage on the statistical properties of a surface, i.e., the roughness exponent, the level crossing, the probability density function, as well as the drift and diffusion coefficients has been studied. In this regard, we have analyzed the surface of Co(3 nm)/NiO(30 nm)/Si(100) structure (as a base structure in the magnetic multilayers, e.g., spin valves operated using giant magnetoresistance (GMR) effect 11,12 ) fabricated by bias sputtering method at different bias voltages. The behavior of statistical characterizations obtained by nanostructural analysis have been also compared with behavior of sheet resistance measurement of the films deposited at the different bias voltages, as a macroscopic analysis.

The substrates used for this experiment were n-type Si(100) wafers with resistivity of about 5-8 Ω-cm and the dimension of 5×11 mm 2 . After a standard RCA cleaning procedure and a short time dip in a diluted HF solution, the wafers were loaded into a vacuum chamber. The chamber was evacuated to a base pressure of about 4×10 -7 Torr. To deposit nickel oxide thin film, first high purity NiO powder was pressed and baked over night at 1400 • C in an atmospheric oven yielded a green solid disk suitable for thermal evaporation. Before each NiO deposition, a pre-evaporation was done for about 5 minutes. Then a 30 nm thick NiO layer was deposited on the Si substrate with applied power of about 350 watts resulted in a deposition rate of 0.03 nm/s at a pressure of 2×10 -6 Torr. After that, without breaking the vacuum, a thin Co layer of 3 nm was deposited on the NiO surface by using DC sputtering technique. During the deposition, a dynamic flow of ultrahigh purity Ar gas with pressure of 70 mTorr was used for sputtering discharge. The discharge power to grow Co layers was considered around 40 watts that resulted in a deposition rate of about 0.01 nm/s. The thickness of the deposited films was measured by styles technique, and controlled in-situ by a quartz crystal oscillator located near the substrate. The distance between the target (50 mm in diameter) and substrate was 70 mm. Before each deposition, a pre-sputtering was also performed for about 10 minutes. The deposition of Co layers was done at various negative bias voltages ranging from zero to -80 V at the same sputtering conditions. The schematic details about the way of exerting the bias voltage to the Si substrate can be found in 13 .

In order to analyze the deposited samples, we have used atomic force microscopy (AFM) on contact mode to study the surface topography of the Co layer. The surface topography of the films was investigated using Park Scientific Instruments (model Autoprobe CP). The images were collected in a constant force mode and digitized into 256 × 256 pixels with scanning frequency of 0.6 Hz. The cantilever of 0.05 N m -1 spring constant with a commercial standard pyramidal Si 3 N 4 tip with an aspect ratio of about

…(Full text truncated)…