[en] The temporal evolution of the plasma potential, V_p, in a pulsed dc magnetron plasma has been determined using the emissive probe technique. The discharge was operated in the 'asymmetric bi-polar' mode, in which the discharge voltage changes polarity during part of the pulse cycle. The probe measurements, with a time-resolution of 20 ns or better, were made along a line above the racetrack, normal to the plane of the cathode target, for a fixed frequency (100 kHz), duty cycle (50%), argon pressure (0.74 Pa) and discharge power (583 W). At all the measured positions, V_p was found to respond to the large and rapid changes in the cathode voltage, V_d, during the different phases of the pulse cycle, with V_p always more positive than V_d. At a typical substrate position (>80 mm from the target), V_p remains a few volts above the most positive surface in the discharge at all times. In the 'on' phase of the pulse, the measurements show a significant axial electric field is generated in the plasma, with the plasma potential dropping by a total of about 30 V over a distance of 70 mm, from the bulk plasma to a position close to the beginning of the cathode fall. This is consistent with measurements made in the dc magnetron. During the stable 'reverse' phase of the discharge, for distances greater than 18 mm from the target, the axial electric field is found to collapse, with V_p elevated uniformly to about 3 V above V_d. Between the target and this field-free region an ion sheath forms, and the current flowing to the target is still an ion current in this 'reverse' period. During the initial 200 ns of the voltage 'overshoot' phase (between 'on' and 'reverse' phases), V_d reached a potential of +290 V; however, close to the target, V_p was found to attain a much higher value, namely +378 V. Along the line of measurement, the axial electric field reverses in direction in this phase, and an electron current of up to 9 A flows to the target. The spatial and temporal measurements of V_p presented here confirm a simple picture of the evolution of V_p, predicted from previously made time-resolved mass spectroscopic measurements of the ionic component in the pulsed magnetron. This paper describes the development and characteristics of the emissive probe technique for such fast measurements, together with implications for the form of the measured transient potential profiles on the operation of the magnetron discharge. In particular, it addresses the charged particle drifts and the potential for sputtering of the walls and the anode by ion impact

[en] Emissive and Langmuir probe techniques have been used to obtain two-dimensional (2D) spatial maps of the plasma potential V_p, electric field E, and ion trajectories in a pulsed bipolar magnetron discharge. The magnetron was pulsed at a frequency of 100 kHz, with a 50% duty cycle and operated at an argon pressure of 0.74 Pa. The pulse wave form was characterized by three distinct phases: the 'overshoot', 'reverse', and 'on' phases. In the 'on' phase of the pulse, when the cathode voltage is driven to -670 V, the 2D spatial distribution of V_p has a similar form to that in dc magnetron, with significant axial and radial electric fields in the bulk plasma, accelerating ions to the sheath edge above the cathode racetrack region. During the 'overshoot' phase (duration 200 ns), V_p is raised to values greater than +330 V, more than 100 V above the cathode potential, with E pointing away from the target. In the 'reverse' phase V_p has a value of +45 V at all measured positions, 2 V more positive than the target potential. In this phase there is no electric field present in the plasma. In the bulk of the plasma, the results from Langmuir probe and the emissive probe are in good agreement, however, in one particular region of the plasma outside the radius of the cathode, the emissive probe measurements are consistently more positive (up to 45 V in the 'on' time). This discrepancy is discussed in terms of the different frequency response of the probes and their perturbation of the plasma. A simple circuit model of the plasma-probe system has been proposed to explain our results. A brief discussion of the effect of the changing plasma potential distribution on the operation of the magnetron is given

[en] Using an emissive probe, the temporal evolution of plasma potential V_p in front of an electrically isolated substrate in an asymmetric pulsed dc magnetron has been determined. The discharge pulsing frequency was 100 kHz, with a 50% duty cycle. Through a scheme of externally biasing the emissive probe, it was found that the time response of the probe could be improved greatly, and a resolution of 20 ns was achieved. This good response revealed that V_p is highly modulated by the transient cathode potential, following it closely and varying from a value just above the ground potential in the pulse 'on' phase, up to a value of +277 V during the positive overshoot in the 'reverse' pulse phase. During the whole pulse cycle, V_p was found to remain above the most positive surface in the discharge. The results confirm our previous prediction for V_p, based on energy-resolved mass spectrometry [Bradley et al., Plasma Sources Sci. Technol. 11, 165 (2002)], which indicated that ions must be created at high positive plasma potentials. However, measurements here show that the substrate floating potential V_f is also strongly modulated and the difference V_p-V_f, which determines the ion bombarding energy, always remains below 40 V during steady phases of the discharge throughout the pulse cycle

[en] A circular planar probe of diameter 4 mm has been used to determine the ratio of the electron E x B drift speed V_d to the thermal speed V_th in the bulk plasma of a magnetron discharge. It is found that when the probe is orientated into the E x B drift, with its surface parallel to the local B-field, significantly higher electron saturation currents are detected than when the probe is orientated away from the drift (i.e. 180 degrees from this orientation). Using the electron and ion saturation current measurements and a simple model of the anisotropic plasma-probe system, which assumes both Hall and collisional cross-field transport, V_d/V_th has been determined at different positions in the bulk plasma. The maximum in V_d/V_th is found to be about 0.14 (corresponding to V_d=1.10⁵ ms^-1) and the measured distribution of drift current agrees well with data that was found in a previous study, in which V_d was determined from the knowledge of E and B, namely V_d=E x B/B², and with V_th determined from electron temperature measurements, i.e. V_th=(2kT_e/m_e)^1/2. (copyright 2004 WILEY-VCH Verlag GmbH and Co. KGaA, Weinheim) (orig.)

[en] Field-enhanced metal-induced solid phase crystallization (FE-MISPC) of amorphous silicon is scaled down to nanoscale dimensions by using a sharp conductive tip in atomic force microscopy (AFM) as one of the electrodes. The room temperature process is driven by the electrical current of the order of 100 pA between the tip and the bottom nickel electrode. This results in energy transfer rates of 30-50 nJ s^-1. Amplitude of the current is limited by a MOSFET transistor to avoid electrical discharge from parasitic parallel capacitance. Limiting the current amplitude and control of the transferred energy (∼100 nJ) enables formation of silicon crystals with dimensions smaller than 100 nm in the amorphous film. Formation of the nanocrystals is localized by the AFM tip position. The presence of nanocrystals is detected by current-sensing AFM and independently corroborated by micro-Raman spectroscopy. The nanocrystal formation is discussed based on a model considering microscopic electrical contact, thermodynamics of crystallization and silicide formation.

[en] Plasma composition near the substrate was investigated in a high power impulse magnetron sputtering (HIPIMS) discharge using Langmuir probe analysis, mass spectroscopy, and atomic absorption spectroscopy. The HIPIMS discharge was operated in nonreactive Ar atmosphere at a pressure of 2.66 Pa and the magnetron cathode was furnished with Ti target. Plasma density, metal ion-to-neutral ratio, and gas ion-to-metal ion ratio were studied as a function of discharge current. At peak discharge current densities of ∼1 A cm^-2, the results show that a dense plasma (n_e∼10¹⁸ m^-3) expanded from the target toward the substrate and lasted more than 330 μs after the supplied power was turned off. The shape of the time-averaged ion energy distribution function of sputtered material exhibited a transition from Thompson to Maxwellian distribution, indicating efficient energy transfer in the discharge. The metal content in the plasma monotonically increased with discharge current and the metal ion-to-neutral ratio reached approximately 1:1 in the postdischarge plasma at peak current density of 5 A cm^-2

[en] Time-resolved measurements of the electron temperature T_e and density n_e at the centerline of a bipolar pulsed dc magnetron argon discharge were obtained using a triple probe. Two electron temperature spikes at the pulse transients were observed and are interpreted as being due to the presence of energetic electrons generated during these periods. During the off time the observed rapid decay of T_e and gradual decay of n_e are shown to be a consequence of enhanced plasma retention due to the magnetized electrons. The rapid rise in n_e during the on time was observed to reach a maximum, coinciding with a minimum in T_e (with T_e decaying rapidly), probably due to enhanced ionization by the energetic electrons. Throughout the rest of the pulse period T_e increased slightly whereas n_e decreased due to global collisional heating of electrons with an additional energetic electron group formed during the on time. The results also show that the electron temperature and plasma density increase with decreasing duty cycle. The plasma density increased linearly with the total energy input per pulse E and increases with pressure. The electron temperature decreases towards the higher pressures and was found to be approximately independent of E. The calculated ion power flux density to a floating substrate (averaged over one pulse cycle and being proportional to the ion-to-atom arrival ratio) was found to be higher by a factor between 2 and 4 than during dc at the same discharge conditions. The power flux was also found to increase linearly with time-averaged power with the steepest rise at the lowest duty cycle. Decreasing the duty cycle and increasing the time-averaged power will lead to the rise in the ion-to-atom arrival ratio and generally improve the quality of the deposited thin films. Finally, these results show the triple probe to be a reliable and efficient method to measure the temporal evolution of the plasma parameters in the pulsed magnetron

[en] HIPIMS (High Power Impulse Magnetron Sputtering) discharge is a new PVD technology for the deposition of high-quality thin films. The deposition flux contains a high degree of metal ionization and nitrogen dissociation. The microstructure of HIPIMS-deposited nitride films is denser compared to conventional sputter technologies. However, the mechanisms acting on the microstructure, texture and properties have not been discussed in detail so far. In this study, the growth of TiN by HIPIMS of Ti in mixed Ar and N₂ atmosphere has been investigated. Varying degrees of metal ionization and nitrogen dissociation were produced by increasing the peak discharge current (I_d) from 5 to 30 A. The average power was maintained constant by adjusting the frequency. Mass spectrometry measurements of the deposition flux revealed a high content of ionized film-forming species, such as Ti¹⁺, Ti²⁺ and atomic nitrogen N¹⁺. Ti¹⁺ ions with energies up to 50 eV were detected during the pulse with reducing energy in the pulse-off times. Langmuir probe measurements showed that the peak plasma density during the pulse was 3 x 10¹⁶ m^-3. Plasma density, and ion flux ratios of N¹⁺: N₂¹⁺ and Ti¹⁺: Ti⁰ increased linearly with peak current. The ratios exceeded 1 at 30 A. TiN films deposited by HIPIMS were analyzed by X-ray diffraction, and transmission electron microscopy. At high I_d, N¹⁺: N₂¹⁺ > 1 and Ti¹⁺: Ti⁰ > 1 were produced; a strong 002 texture was present and column boundaries in the films were atomically tight. As I_d reduced and N¹⁺: N₂¹⁺ and Ti¹⁺: Ti⁰ dropped below 1, the film texture switched to strong 111 with a dense structure. At very low I_d, porosity between columns developed. The effects of the significant activation of the deposition flux observed in the HIPIMS discharge on the film texture, microstructure, morphology and properties are discussed.

[en] In this study, Raman spectroscopy is used as a tool to determine the light-trapping capability of textured ZnO front electrodes implemented in microcrystalline silicon (μc-Si:H) solar cells. Microcrystalline silicon films deposited on superstrates of various roughnesses are characterized by Raman micro-spectroscopy at excitation wavelengths of 442 nm, 514 nm, 633 nm, and 785 nm, respectively. The way to measure quantitatively and with a high level of reproducibility the Raman intensity is described in details. By varying the superstrate texture and with it the light trapping in the μc-Si:H absorber layer, we find significant differences in the absolute Raman intensity measured in the near infrared wavelength region (where light trapping is relevant). A good agreement between the absolute Raman intensity and the external quantum efficiency of the μc-Si:H solar cells is obtained, demonstrating the validity of the introduced method. Applications to thin-film solar cells, in general, and other optoelectronic devices are discussed.