Svetlana Radovanov

Noninvasive, real-time measurements of plasma parameters via optical emission spectroscopy

Plasma process control applications require acquisition of diagnostic data at a rate faster than the characteristic timescale of perturbations to the plasma. Diagnostics based on optical emission spectroscopy of intense emission lines permit rapid noninvasive measurements with low-resolution (∼1 nm), fiber-coupled spectrographs, which are included on many plasma process tools for semiconductor processing. Here the authors report on rapid analysis of Ar emissions with such a system to obtain electron temperatures, electron densities, and metastable densities in argon and argon/mixed-gas (Ar/N2, Ar/O2, Ar/H2) inductively coupled plasmas. Accuracy of the results (compared to measurements made by Langmuir probe and white-light absorption spectroscopy) are typically better than ±15% with a time resolution of 0.1 s, which is more than sufficient to capture the transient behavior of many processes, limited only by the time response of the spectrograph used.

Ion energy distributions in a pulsed plasma doping system

Discharge parameters in a pulsed dc plasma doping system have been studied using measurements of time-resolved ion energy distributions, relative ion density, plasma potential, and electron temperature in BF3 and Ar plasmas during active discharge and afterglow periods. Negative plasma potentials are observed when using a hollow cathode to create a plasma while implanting at ultralow energies (<500eV). The kinetics of ion generation and decay in BF3 during the pulse on and off periods have been discussed.

Plasma diagnostics in pulsed plasma doping (P/sup 2/LAD) system

As semiconductor devices continue to shrink in size, demands for the formation of ultra-shallow junctions (USJ) are increasing. Pulsed plasma doping (P/sup 2/LAD) has emerged as a scaleable and cost effective solution to dopant delivery, since it is capable of high dose rates at ultra-low energies (0.02-20 kV). In P/sup 2/LAD, a pulsed plasma is generated adjacent to the silicon wafer using pulsed biases. Typical pulse widths range between 5 and 50 /spl mu/s, and pulse repetition rates are between 100 and 10000 Hz. Time-resolved Langmuir probe measurements showed that cold plasma is present during the afterglow period, which may play an important role in process control. Probe measurements also showed the presence of primary electron and electron beams during the initial pulse-on stage in both Ar and BF/sub 3/ plasmas. Ion mass and energy analysis indicated that BF/sub 2//sup +/ is the dominant ion species in the BF/sub 3/ plasmas, with BF/sup +/ as the second-most abundant ion species.

Optical emission spectroscopy of rf discharge in SF6

In this paper we present experimental results obtained for a rf discharge in SF6 and for SF6 with Ar and/or N2. The data for power dependence of some emission lines usually used in actinometry are acquired and their applicability and excitation kinetics discussed. We also present the spatial (time averaged) variations of some emission lines. From such data the spatial dependence of the high‐energy tail of the electron energy distribution function may be obtained.

Beam angle control on the VIISta 80 ion implanter

Advanced integrated circuit design requires precise control of beam incidence angle. This requirement has led to the development of an automated angle control system on Varian Semiconductors high current VIISta 80 ion implanter. In this paper we show beam incidence angle and angular spread measurements for 200 and 300 mm ion beams on the VIISta 80 ion implanter. Multiple beam measurements are sampled across the wafer plane for each beam setup. Beam angle computation results are compensated for prior to wafer implantation for optimal incident angle control. Beam, bare wafer and device performance data were used to confirm the accuracy of this measurement and control system. Excellent measurement accuracy and repeatability has been demonstrated. Data will be shown which includes arsenic, boron and phosphorus implants from both drift and decel operation. Benefits and process differences will be shown with active beam angle correction as compared to classical open loop methods. Mechanical tilt angle accuracy, repeatability and verification data will also be discussed.

Dissociative excitation of hydrogen in rf and dc glow discharges through H2

Abstract The spectral and spatial profile of H β from the low pressure rf and dc glow discharges in hydrogen is studied in order to reveal the excitation mechanism of the fast excited H fragments. Measurements were performed both for the normal and abnormal dc glow discharges. Spatial distributions of the Balmer β radiation reflect the local plasma conditions in the discharge, especially the excitation efficiency which is used to determine the excitation kinetics in hydrogen discharges. Spectral H β profiles were measured and used to determine the kinetic energy of excited H atoms and to check which of the mechanisms describes best the results observed in our experiment. We have also calculated the number densities of vibrationally excited levels by solving a set of vibrational master equations for the conditions similar to those of our experiments, as excitation from the vibrationally excited ground-state hydrogen molecules may be used to explain the changes in the intermediate wing component of the line profile with the changing current.

Electron energy distribution functions in weakly ionized argon

The authors present measurements of the electron energy distribution functions (EEDF) for electrons in argon discharges at moderate and high E/N values (E being the electric field and N the gas density), for homogeneous electric fields and a low-current diffuse glow regime. Results were obtained for electric field to gas density ratios (E/N) from 500 Td to 50 kTd (1 Td=10-21 V m2). A multigrid energy analyser with a retarding grid potential was used to measure distribution functions of electrons sampled through an aperture in the anode. Experimental data are used to make a comparison with the two-term Boltzmann calculations for E/N<1 kTd, and the single-beam model predictions, normally used to model electron kinetics at high values of E/N.

Comparison of surface vacuum ultraviolet emissions with resonance level number densities. II. Rare-gas plasmas and Ar-molecular gas mixtures

Vacuum ultraviolet (VUV) emissions from excited plasma species can play a variety of roles in processing plasmas, including damaging the surface properties of materials used in semiconductor processing. Depending on their wavelength, VUV photons can easily transmit thin upper dielectric layers and affect the electrical characteristics of the devices. Despite their importance, measuring VUV fluxes is complicated by the fact that few materials transmit at VUV wavelengths, and both detectors and windows are easily damaged by plasma exposure. The authors have previously reported on measuring VUV fluxes in pure argon plasmas by monitoring the concentrations of Ar(3p54s) resonance atoms that produce the VUV emissions using noninvasive optical emission spectroscopy in the visible/near-infrared wavelength range [Boffard et al., J. Vac. Sci. Technol., A 32, 021304 (2014)]. Here, the authors extend this technique to other rare-gases (Ne, Kr, and Xe) and argon-molecular gas plasmas (Ar/H2, Ar/O2, and Ar/N2). Results...

Comparison of surface vacuum ultraviolet emissions with resonance level number densities. I. Argon plasmas

Vacuum ultraviolet (VUV) photons emitted from excited atomic states are ubiquitous in material processing plasmas. The highly energetic photons can induce surface damage by driving surface reactions, disordering surface regions, and affecting bonds in the bulk material. In argon plasmas, the VUV emissions are due to the decay of the 1s4 and 1s2 principal resonance levels with emission wavelengths of 104.8 and 106.7 nm, respectively. The authors have measured the number densities of atoms in the two resonance levels using both white light optical absorption spectroscopy and radiation-trapping induced changes in the 3p54p→3p54s branching fractions measured via visible/near-infrared optical emission spectroscopy in an argon inductively coupled plasma as a function of both pressure and power. An emission model that takes into account radiation trapping was used to calculate the VUV emission rate. The model results were compared to experimental measurements made with a National Institute of Standards and Techn...

Measurements and modeling of electron energy distributions in the afterglow of a pulsed discharge in BF3

In this paper we use experimental data (Radovanov S. and Godet L., J. Phys.: Conf. Ser., 71 (2007) 012014) for time-resolved electron energy distribution function in boron trifluoride (BF3) discharges together with cross-sections for electron excitation processes and attachment in order to explain electron dynamics in the pulsed plasma doping system. A Monte Carlo simulation (MCS) was used to perform calculations of the electron energy probability function (EEPF) in pulsed DC electric fields as found in practical implantation devices. It was found that in the afterglow, electric field in the plasma is not zero but still has a significant reduced electric field (E/N) albeit below the breakdown condition. Our analysis assuming free diffusion conditions in the afterglow led to the calculation of EEPF for a range of E/N corresponding to different afterglow times of a pulsed DC discharge. Calculated and experimental EEPF agree fairly well for a given set of cross-sections (see paper by Radovanov and Godet quoted above) and assumed initial distributions. In addition we have modeled the kinetics of production of negative ions in the afterglow as observed by experiment and found an increase in the production of negative ions in the early afterglow. Electron attachment in BF3 with 0.1% of F2 is a possible explanation for the observed rate of negative-ion production as predicted by our Monte Carlo simulation. However, the most likely cause for the increase in detected number density of ions is the collapse of the field-controlling electrons.