AVS1996 Session PS-TuP: Deposition and Plasma Ion Implantation Poster Session

Tuesday, October 15, 1996 6:30 PM in Ballroom A

Tuesday Evening

Time Period TuP Sessions | Topic PS Sessions | Time Periods | Topics | AVS1996 Schedule

PS-TuP-1 The Effects of Radicals on the Crystallinity of Polycrystalline Silicon Films formed by using ECR SiH\sub 4\/H\sub 2\ Plasma
M. Hori, R. Nozawa, H. Takeda, M. Ito, T. Goto (Nagoya University, Japan)
The effects of ions on polycrystalline silicon (poly-Si) films prepared by electron cyclotron resonance (ECR) SiH\sub 4\/H\sub 2\ plasma have been investigated. The energy of ion incident on the substrate was changed by applying dc biases to the substrate. It was found that the positive biases improved the crystallinity and the surface roughness of the deposited films using X-ray diffraction (XRD) technique and atomic force microscopy. The best crystallinity of films was obtained with a substrate bias of +50 V, a substrate temperature of 300 C, a total pressure of 0.5 Pa, a hydrogen dilution rate of 90.9 % and a microwave power of 300 W under our experimental apparatus. Furthermore, we investigated the roles of species in the plasma for the poly-Si formation using a shield and permanent magnets. The shield set at 10 mm above the substrate shut out the species bombarding the substrate with high energy and the light radiation of plasma. Therefore, the film deposited below the shield was formed by the species with low energy. The XRD pattern of the film deposited with the shield showed that the films had amorphous structure. On the other hand, the permanent magnets were set so that the only charged species incident on the substrate might be pushed away and so the film was deposited by only neutral species and by the light radiation of plasma. These films showed polycrystalline structures.These results indicate that the films can be crystallized in the condition without charged species. Therefore, we found that neutral species with high energy and the light radiation of plasma play important roles in the poly-Si formation at 300 C in ECR plasma CVD.
PS-TuP-3 How Many Individual Chemical Steps Are Responsible for the Reactions Observed in ECR-Microwave Plasmas?
S. Webb, R. Blumenthal (Auburn University)
. Pulsed supersonic, plasma sampling mass spectrometry experiments on ECR-microwave plasma deposition of diamond have revealed that a fast interconversion of chemical species occurs, with ~50% of the ethylene feed gas is converted into a C\sub 2\H\sub 2\ species. Under normal ECR-microwave plasma conditions, with pressures at or below 1 X 10\super -3\ Torr and mean free paths of several centimeters, logic dictates that only a few chemical steps must be responsible for the observed fast interconversion. Substitution of D\sub 2\ for the normal hydrogen gas results a D substitution on the carbon backbone for each addition step. The extent of isotopic substitution of the carbon backbone of the interconverted plasma gas is then a direct measure of the number individual H atom addition and abstraction steps that are responsible for the observed chemical transformations.
PS-TuP-4 Determination of the Structure of HMDSO/O\sub 2\ Plasma Deposits using High Resolution XPS
M. Alexander, R. Short, F. Jones (University of Sheffield, United Kingdom); M. Stollenwerk, W. Michaeli (IKV, Germany); C. Blomfield (Kratos Analytical, England)
Deposits formed from plasmas of hexamethyldisiloxane (HMDSO) have applications in physical and chemical surface modification. We have shown that XPS can be used to follow the changes in the chemistry of the deposit with the addition of oxygen to the plasma\super 1\. The Si2p peak position was used to provide information on the chemical state of the silicon atoms in the deposit. In the work presented here, we have investigated the additional information that may be gained from the application of a high resolution XP spectrometer to the analysis of these materials: a Kratos AXIS 165 with magnetic lens and charge neutralisation. Curve fitting of the Si2p envelope was undertaken to allow the different chemical states of the silicon to be determined. Increasing the proportion of oxygen in the plasma produced a change in the deposit from one with an organic, polymeric chemistry, to an inorganic silica-like deposit. An inert gas was added to the HMDSO/O\sub 2\ plasma to produce deposits more closely resembling SiO\sub 2\.1. M. R. Alexander, R. D. Short, F. R. Jones, M. Stollenwerk, J. Zabold, and W. Michaeli, J. Mater. Sci., 31 pp 1879-1885 (1996).
PS-TuP-5 Langmuir Probe Analysis of a Cylindrical, Highly Asymmetric PECVD System
R. Turkot, D. Ruzic (University of Illinois, Urbana)
Thin film deposition onto non-planar, even irregularly shaped, objects often encounters uniformity problems. PECVD techniques are frequently used in place of evaporation or other methods to address this problem. Here, electrode size and relative placement play an important role in determining the uniformity of films deposited. This study analyzes experimentally and theoretically the system used to deposit SiO\sub 2\ films onto cylindrically shaped containers. The apparatus consists of a two-plasma system in which 2 "different" plasmas are separated by the object being coated. The powered electrode is placed inside the container and a seal separates the gases on the outside from those inside. Gas pressure and composition for each plasma are independently controlled. The second electrode is a cylindrical ground shield that surrounds the system. Langmuir probe analysis inside and outside the container show evidence of the well known unequal electrode effect pulling potentials near the container wall and the inner electrode well above plasma potential and that normally expected at such surfaces. These potential rises alter the electron and ion densities normally seen in the sheath, and therefore the chemistry occurring there. Analysis of the system reveals that the inner plasma composition and pressure has little effect on plasma parameters outside the container. The potential rise at the container surface, caused by the unequal electrode areas, prevents positive ions from striking the surface, thus limiting the plasma-surface interactions to electrons, and negative ions. Electrodes of different sizes are utilized and their effects on the plasma and films noted. The sheath voltages predicted by Koenig and Maissel due to unequal electrode areas are larger than those seen in our system, but may be limited due to the lack of plasma density in the sheath due to the increased voltage at the container surface.
PS-TuP-6 Low Temperature ECR Plasma Enhanced Chemical Vapor Deposition of Silicon Nitride
S. Speakman (Thin Film Consultancy); C. Phillips, J. Ashe (Xaar Printing Technologies, Ltd.); B. Mercer, J. Spencer (Plasmaquest, Inc.)
Plasma enhanced chemical vapor deposition of silicon nitride is normally carried out at temperatures above 350\super o\C. Films deposited at lower temperatures with conventional RFPECVD generally exhibit high hydrogen levels, low density, and low hermeticity. However, new applications in compound semiconductors and piezoelectric micromachines are requiring processing temperatures below 100 \super o\C while achieving high density, low stress, and high hermeticity. High density plasmas provide the high ion current necessary to densify films during deposition, but the low processing pressures (approximately 10 mtorr) result in poor temperature control of the substrate. Without heat sinking, a wafer under ECR CVD will attain a temperature in excess of 200\super o\C within seconds due to the heat from ion bombardment. This temperature will be attained regardless of the chuck temperature. A divergent beam ECR source with magnetic multipolar plasma confinement has been used to deposit dense, hermetic films of silicon nitride at wafer temperatures as low as 40\super o\C using a process of SiH\sub 4\, Ar, and N\sub 2\. By heat sinking the wafer by helium backside cooling, film properties were found to be most strongly influenced by ECR power and pressure, parameters which correlate strongly with ion current. Substrate temperature was only a minor factor in determining film properties over wafer temperature range of 40 to 150\super o\C. Without heat sinking, the wafer surface temperature exceeded 200xC within 30 seconds of initiating the process. Stress was \<\2 x 10\super 9\dynes cm\super-2\ compressive and the wet etch rate in 10:1 buffered oxide etch at room temperature was \<\1 \Ao\sec\super -1\. Using the technique of ERDA hydrogen content of \<\10\%\ and a stoichiometry of Si:N of 0.65 were measured. FTIR analysis confirmed the stoichiometry of the films. The properties of this film suitable for a wide range of micromechanical applications.
PS-TuP-7 A Comparison of Hollow Cathode Electron Sources for an Enhanced Magnetron Sputtering System
N. Williams, J. Cuomo (North Carolina State University); G. Hufnagel (Commonwealth Scientific Corporation)
A comparison of two hollow cathode electron sources has been made as electron sources for an enhanced magnetron sputtering system. One is a commercial electron source manufactured by Commonwealth Scientific Corporation (CSC) and the other is a thermionic emitting hollow cathode electron source previously described in the literature. The CSC hollow cathode electron source provides a source of electrons of more than 5A in reactive environments. Under the same conditions, a thermionic emitting hollow cathode would have a shorter lifetime. This improvement upon the existing thermionic emitting hollow cathode electron source provides less maintenance, longer lifetimes, and higher emission currents. We have determined that the CSC hollow cathode electron source can provide a higher plasma current density which has been determined by various methods. The two systems are evaluated comparing the coupling efficiency, plasma density, deposition rate, operating pressure, and magnetron voltage. The thermionic emitting hollow cathode yields more efficient operation at lower pressures, however at pressures between 1-10 mTorr the CSC hollow cathode is comparable in output. The effects of these two systems upon the properties of various metals such as aluminum and titanium has also been investigated.
PS-TuP-8 Diagnostic of a Magnetron Discharge with Enhanced Ionization
M. Choquier, A. Vanderbecq (Université de Mons-Hainaut, Belgium); A. Pointu (Université de Paris Xi, France); J. Dauchot, M. Wautelet, M. Hecq (Université de Mons-Hainaut, Belgium)
A conventional magnetron discharge with the addition of an RF inductively coupled plasma located above the magnetron cathode is analyzed by optical emission and atomic absorption as well as Langmuir probe. The aluminum target, powered in the DC mode and 1.3 inch in diameter, is sputtered in pure argon. The deposition rate is measured by a quartz microbalance. The influence of the discharge pressure and the electric power applied to the magnetron and to the induction coil on the ArI, ArII, AlI, AlII emission and absorption lines, electron energy and density and deposition rate is studied via an "experimental design procedure". The argon metastable density and aluminum atomic density are calculated from the absorption coefficients and correlated to the parameters of the discharge.
PS-TuP-9 Sheet Electron Beam Plasma Atomic Layer Deposition
M. Shaheen, M. Hersam, D. Ruzic (University of Illinois, Urbana)
SEBALD (Sheet Electron Beam Plasma Atomic Layer Deposition) is a novel method of thin film deposition which is a variant on the conventional PECVD and an alternative to remote PECVD. Both CVD and PECVD suffer from a limited control over growth rates relative to MBE. SEBALD is proposed as a means to overcome this limitation through the use of an electron beam generated plasma, which dimension is confined to a narrow electron sheet that passes over the substrate at a controllable height. Variations in beam shutter bias, cathode voltage, sheet beam to substrate distance, gas type and pressure can vary the type and energy of the species arriving at the substrate. The bulk plasma electron temperature at 5 cm from the sheet beam remains low, 0.4 eV, allowing better control over the plasma chemistry and ion acceleration to the substrate. A model of the e-beam plasma kinetics, silane chemistry and surface deposition is used to guide the choice of the experimental parameters so as to effectively select a specific radical for deposition. Amorphous silicon films of several monolayer thickness have been grown as a proof of principle at a deposition rate of 10 A/min for wafer to sheet distance of 3 cm in a 5% silane in helium plasma. A variety of surface analysis techniques including SEM, AFM and SIMS are used to characterize the films.
PS-TuP-10 Sheath Thickness in Very-High-Frequency Plasma CVD of Hydrogenated Amorphous Silicon
W. van Sark, H. Meiling, E. Hamers, J. Bezemer, W. van der Weg (Utrecht University, The Netherlands)
When very-high-frequency (VHF) glow discharges are used to obtain homogeneous hydrogenated amorphous silicon films on glass substrates, an optimum combination of pressure and frequency is required. With frequencies in the range of 19 to 80 MHz pressures from 0.65 to 0.35 mbar are used. It was suggested that the sheath thickness plays an important role here. We use two recently developed experimental methods to determine the sheath thickness, i.e. gap-induced inhomogeneity and in-situ energy-resolved mass spectrometry. The gap-induced inhomogeneity method is based on the observation of a reduced deposition rate if a gap exists between the glass substrate and the metal substrate electrode. The reduction scales with gap thickness and, more importantly, with the sheath thickness. In-situ energy-resolved mass spectrometry is employed to measure the ion-energy distributions at the grounded electrode. From the appearance of charge-exchange peaks in silane plasmas we are able to deduce the sheath thickness. Both methods have been employed to determine the sheath thickness in VHF silane/hydrogen glow discharges. For the optimum frequency-pressure combinations it appears that all sheath thicknesses are about equal and amount to about 2 mm. This was found independently using both methods.
PS-TuP-11 Experimental Studies of Contaminant Ion Effects on PSII Fabricated Diodes
R. Speth, J. Shohet, J. Booske, S. Gearhart, H. Liu, W. Wang, R. Mau (University of Wisconsin, Madison)
Plasma Source Ion Implantation\sup 1\ (PSII)has long been regarded as a possible alternative to classical ion implantation. The benefits of PSII, such as low energy high dose implants and its comparatively small footprint, become increasingly tempting for integrated circuit fabrication as device dimensions shrink. An open question concerns the effect of plasma-sputtered contaminant ions on device characteristics. A suite of experiments is currently underway which will quantify the allowable level of contamination in a diode array. The array of diodes is fabricated on 3 inch N type <100> P-doped silicon wafers. Asided from the implantation step the processing of the wafers is standard. The implantation step which occurs just after the oxide etch includes both the implantation of the boron ions and the contaminant ions. In each wafer processed the energies of both ions will be the same to simulate the conditions found in the PSII chamber. Controlled ion beam implantation will be utilized to introduce measured quantities of both the boron and contaminant ions. As a result, control of the absolute as well as the relative dose of the impurities can be achieved. The first contaminant to be studied is aluminum since it represents a preferred material for constructing PSII chambers, and it is expected that aluminum will be sputtered from the chamber walls. The experiment is a 2\sup 3\ factorial design varying implant energy, aluminum concentration and their effects on the performance of test diodes of various dimensions. The measured responses are ideality factor, breakdown voltage and reverse bias leakage current. *Supported by NSF grant EEC-8721545. \sup 1\ J.R. Conrad, J.L. Radke, R.A. Dodd, F.J. Worzala and N.C. Tran, J. Appl. Phys. 62, 4591 (1987).
PS-TuP-12 Plasma Purification by Ion Cyclotron Resonance for Plasma Source Ion Implantation Doping of Semiconductors
T. Snodgrass, M. Jenkins, J. Shohet (University of Wisconsin, Madison); M. Kushner (University of Illinois, Urbana-Champaign); J. Booske, R. Stewart (University of Wisconsin, Madison)
Using Plasma Source Ion Implantation\super 1\ (PSII) to create the shallow source and drain structures required for next generation devices may be a necessity\super 2\. A critical concern for the use of PSII in doping these next generation semiconductor devices is to avoid implanting contaminant ions. For example, future devices are predicted to require that heavy metal doses be kept less than 3 x 10\super 9\ atoms per square centimeter\super 3\. When doping semiconductors using conventional beam line accelerator technology, the implantation is very pure because bending the beam with a magnetic field very efficiently selects a single mass species. In order to provide a similar purity implantation for PSII a process utilizing ion cyclotron resonance for plasma purification prior to implantation is being investigated. Plasma purification is done using ion cyclotron resonance to selectively expel unwanted ions from the plasma where they will be neutralized upon collision with the chamber wall and no longer an implantation hazard\super 4\. With a computer simulation we can determine the necessary field strengths and uniformity for plasma purification, cleaning efficiency and frequency/mass resolution of the method. Initial results to confirm the theory using mass spectrometry will also be shown. \super *\ Work supported by the National Science Foundation under Grant EEC-8721545. \super 1\ J. R. Conrad, J. L. Radtke, R. A. Dodd, F. J. Worzala, N. C. Tran(1987), J. Appl. Phys. 62(11) 4591. \super 2\ Semiconductor Industry Association, The National Technology Roadmap for Semiconductors, (1994). \super 3\ N. Natsuaki, T. Kamata, K. Kondo, Y. Kureishi (1995), Nucl. Inst. Meth. Phys. B, 96, 62. \super 4\ J. L. Shohet, E. B. Wickesberg, M. J. Kushner (1994), J. Vac. Sci. Technol. A, 12(4), 1380.
PS-TuP-13 Ultra-Shallow p+/n Junction Fabricated by Boron Solid-Source and Plasma Source Ion Implantation
S. Gearhart, H. Liu, J. Booske, R. Mau, Unknown ERC-Thrust-Area-4-Team (University of Wisconsin, Madison)
Using a new plasma-aided solid-source implantation method, ultra-shallow ( <100 nm ) p+/n junctions have been fabricated for the application of sub-micron CMOS source/drain formation . This process avoids the hazards and costs of handling highly toxic and reactive gases. In this method, a 7\Ao\ thick B layer is first sputtered onto the Si wafer from a boron target. The wafer is then immersed in an Ar plasma and biased at -3 kV in a Plasma Source Ion Implantation ( PSII ) chamber. Argon ions bombard the wafer and drive B atoms into the Si substrate by means of ion beam mixing. Dopant activation and damage removal are achieved via rapid thermal annealing (RTA). The B profiles are measured by SIMS. Recent results show that a 7\Ao\ thick B layer implanted with 3-keV, 4*10\super 15\ cm\super -2\ Ar\super +\ and annealed at 950\super o\C for 10s shows a peak B concentration of 3*10\super 20\ cm\super -3\, and the concentration is more than two orders of magnitude lower at a d! epth of 20 nm. The sheet resistance was measured to be 3000 ohm/square. For comparison, the same amount of B was deposited on an n-type (100) ~1500-2000 ohm-cm Si wafer. Then it was implanted with 3-keV, 1.3*10\super 15\ cm\super -2\ Ar\super +\ and annealed at 1050\super o\C for 10s. The sheet resistance of this B-doped layer was 300 ohm/square. Currently, Ar preamorphization in PSII, various Ar implantation doses, and different RTA temperatures are being investigated to achieve better junction properties. Measured characteristics of test diode structures will also be discussed. This work is supported by the NSF under grant number EEC-8721545.
PS-TuP-14 Plasma Assisted Deposition of Pd Nanoclusters
P. Brault, A. Thomann (GREMI, Orleans, France); B. Rousseau, H. Szwarckopf (CRMD, Orleans, France); R. Boswell (PRL, ANU, Canberra, Australia); J. Rozenbaum (GREMI, Orleans, France)
We are presently studying the deposition of palladium on silicon and graphite using an argon plasma. Both systems can be considered as model surface catalysts : Numerous studies have been already conducted on Pd/graphite using atomic beam methods. On the other hand, Pd deposition on the silicon native oxide can modelize the Pd/SiO2 catalyst. The argon plasma ions sputtered a biased helicoidal palladium wire (Vb=0 to -350 V, PAr =1 to 100 mTorr) located in a vacuum chamber between a high frequency (100 MHz) excited antenna and the substrate. For coverages between 0.02 and 1 monolayer (as deduced from XPS), a broadening and shift of the Pd3d core level line occurs at low Pd coverage indicating that Pd islands are growing. For coverage greater than one monolayer, the Pd film recover the electronic structure of bulk palladium. Scanning tunneling microscopy was performed on Pd films at coverages above one monolayer and reveals three dimensial islands with size between 50 and 300 \Ao\ depending on the plasma parameters. For Pd deposition on graphite, we use first the plasma for inducing a roughness growth. This roughness rms height ranges from 2 to 20 \Ao\ when increasing Ar plasma duration. Further studies will be devoted to examine the roughness effect on the electronic structure of the Pd islands. It is expected that the roughness will driven growth of a larger number and smaller Pd islands on the graphite surface.
Time Period TuP Sessions | Topic PS Sessions | Time Periods | Topics | AVS1996 Schedule