AVS1996 Session MS+PS-WeM: Advanced Plasma Equipment: RF Sensors and Controls

Wednesday, October 16, 1996 8:20 AM in Room 201A

Wednesday Morning

Time Period WeM Sessions | Abstract Timeline | Topic MS Sessions | Time Periods | Topics | AVS1996 Schedule

Start Invited? Item
8:20 AM MS+PS-WeM-1 Downstream Microwave Plasma Sources for Semiconductor and Industrial Processing
R. Post (Applied Science and Technology, Inc. (ASTeX))
Downstream microwave plasma sources are widely use in semiconductor processes and have made the largest impact in photoresist strip. This technology has become the standard for single wafer processing where its success derives from simplicity, reliability, and low damage attributed to low charge concentration in the process chamber. We will review the implementation of this technology for single, 8" wafer processing, including radical production, power distribution, material and thermal problems, chamber and microwave requirements for advanced semiconductor processing. Future semiconductor processing opportunities are for 300 mm, isotropic etch, and wafer cleaning. These new opportunities will drive the technology further to higher power and reactive gases. We can take several approaches to this challenge. The simple tube type applicator, consisting of a dielectric tube passing through a cavity, can be extended with the availability of sapphire tubes and active cooling. Alternatively, large area sources provide a different path. They reduce the power density on the dielectric surface and, thus, mitigate the challenge of higher power applicators, but they reintroduce issues which made downstream tube applicators successful in the first place. These large area source issues include greater complexity, downstream charge, and UV flux on the processing surface. We will examine these two strategies for next generation processing. Additionally, while the semiconductor industry is currently the largest market for this technology, a wide range of industrial processing is possible with downstream technology, including surface treatment for adhesion, sterilization, and polishing to name a few. The use of high power, 915 MHz, can increase the throughput for industrial processes which seek to take advantage of this attractive technique for low temperature, dry chemistry.
9:20 AM MS+PS-WeM-4 Characterization of a Metal Etching Process in an Inductively Coupled Discharge using Measurements of Discharge Impedance and I/V Sensors
R. Patrick, C. Lee, S. Hilliker, R. Moeller (Lam Research Corporation)
The RF characteristics of a BCl\sub 3\/Cl\sub 2\ discharge used to etch aluminum films for semiconductor interconnect applications have been studied in a Transformer Coupled Plasma (TCP\super TM\) etching system. The impedance (both real and imaginary parts) of the discharge was obtained as a function of process condition by studying the capacitor positions of the matching network attached to the TCP plasma source. By correcting for the contribution of the excitation coil, estimates were made of the plasma resistance and reactance as a function of generator power and chamber pressure. In addition, measurements of RF current and voltage were made using I/V probes incorporated into the match as a function of TCP power measured at the generator, pressure and total gas flow. It was found that the RF current and voltage at the input to the coil were both linearly dependent on power, inversely dependent on total pressure and virtually independent of total gas flow. The current-voltage measurements were correlated with measurements of etch rate and uniformity for both aluminum and oxide films as a function of generator power, pressure and total gas flow.
9:40 AM MS+PS-WeM-5 Uniformity of Aluminum Etch in a Chlorine Plasma: Fluid Simulaton and Validation
D. Beale, P. Jones, S. Siu, R. Patrick (Lam Research Corporation)
A 3D reacting flow simulation of Aluminum etching was validated by comparing predicted vs. measured etch profiles, then used to predict conditions which yield good uniformity. Measurements for this study were made in Lam Research Corporation's TCP 9600 reactor using 12 mT pressure and Cl\sub 2\ as the etch chemistry. A simple Arrhenius model of the nearly transport-limited Cl\sub x\+Al->AlCl\sub y\ reaction was used. The flow simulations, which were done with the commercially available fluid dynamics software package CFD-ACE, correctly predicted a trend from edge-fast to center-fast etching as the mass flow rate through the showerhead was increased.The compressible flow fluid simulation, which accounted for sonic injection, Knudsen flow and heat generated by the surface reaction, explained observed etch profiles in terms of convection (from both the injectors and pump) and diffusion through the stagnant layer of reaction products. The relative importance of convection, diffusion and loading was determined for each case by calculation of the Thiele Modulus and fluid Peclet number. Direct comparisons were then made to measured etch profiles as a function of pressure, residence time, and chamber aspect ratio. The flow simulation was extended to approximately include plasma dissociation. The simulation was used to predict the range of reactor dimensions and process conditions which produce good uniformity.
10:00 AM MS+PS-WeM-6 RF Techniques and Harmonic Generation
P. Miller (Sandia National Laboratories)
Many workers have gained an appreciation for complexities involved in delivery of rf power to plasma reactors. While the rf requirements of processes commonly are specified solely in terms of rf power, additional electrical factors can be critical to success. As wafer sizes grow and process requirements become more stringent, it becomes increasingly important to understand in detail factors such as harmonic generation. Signals at multiples and submultiples of the rf excitation frequency are generated by the plasma because the plasma is electrically nonlinear. These harmonics are dependent on, and, thus, indicative of the plasma state. This presents diagnostic opportunities for fingerprinting processes and for endpoint detection. Additionally, the harmonic levels can be enhanced or suppressed due to their interaction with rf-delivery circuits. Since we can manipulate those circuits, this presents an opportunity for control of processes. In this paper we will describe early experiments demonstrating the reality, importance, and control of harmonics in laboratory reactors; describe techniques for analyzing harmonics and circuits; relate how a broken process was fixed by attending to harmonic issues; and present recent experimental results analyzing harmonic generation in terms of a fundamental plasma sheath model [M. E. Riley, Sandia report SAND95-0775 (1995)]. ________ Work supported by the U.S. Department of Energy and by SEMATECH.
10:40 AM MS+PS-WeM-8 Remote Plasma Impedance Measurements in an Inductively Coupled Plasma
G. Melden, J. Cecchi (University of New Mexico)
An inductively coupled plasma source with a flat spiral induction coil designed to operate at 13.56 MHz. was constructed with provisions for measuring the induction coil voltage and current. A novel design feature is the inclusion of a second current transformer that measures the capacitively coupled current from the coil to the plasma. Both of the induction coil leads pass through this second current transformer so that the capacitive current is indicated by the transformer directly. A circuit model for the source match network was developed to account for losses and to allow an accurate assessment of the plasma electrical characteristics. Circuit elements in the model were deduced by least squares fitting of data obtained at differing frequencies and loads. The frequency range assessed was 12.75 to 14.5 MHZ. Test loads comprised movable conducting rings with series resistors; the "no load" condition was also included. With the match network fully characterized, the electrical measurements were reduced to disclose the complex plasma impedance. The use of this measurement as a diagnostic tool in the chlorine poly etch process is explored using response surface methodology. Controlled variables include pressure, source power, and chuck bias. The correlation of the complex plasma impedance to plasma characteristics and uniformity, as well as to etch characteristics on patterned poly is discussed.
11:00 AM MS+PS-WeM-9 Techniques for Measuring Plasma Impedance and Power in an Inductively Coupled Plasma Source
J. Caughman, L. Berry, D. Hoffman (Oak Ridge National Laboratory)
The amount of power that is deposited in the plasma during the fabrication of microelectronics is an important processing parameter for closed loop and model based control systems. The plasma impedance changes that occur during processing will change the efficiency of the plasma source and the amount of net power that reaches the plasma. The plasma impedance and power are being measured in an inductively coupled plasma source. The plasma reactor consists of a flat spiral coil driven at 13.56 MHz. The coil is separated from the 30 cm diameter chamber by a 2.54 cm thick quartz window. The matchbox consists of a "T" type matching network where the capacitors are driven by stepping motors with precision position encoders. The electrical characteristics of the matching network have been precisely measured so that capacitor position can be related to plasma impedance and coupled power to the plasma. These measurements complement those made by a sensor located between the matchbox and the coil. Results indicate that the efficiency of the plasma source varies widely with processing gas and pressure and can range from 50-85%. The voltages and currents in the system can vary widely as well and depend on match settings and plasma conditions. Voltages of up to 10 kV and currents of up to 30 A are possible and can approach or exceed the specifications of the matching components. Plasmas of argon, hydrogen, helium, sulfur hexafluoride, and silane have been studied, and processing implications will be presented. ORNL is managed by Lockheed Martin Energy Research Corp. for the U.S. Department of Energy under contract no. DE-AC05-96OR22464.
Time Period WeM Sessions | Abstract Timeline | Topic MS Sessions | Time Periods | Topics | AVS1996 Schedule