A Langmuir Probe Investigation of Electron Cyclotron Resonance Argon-Hydrogen Plasmas

Journal of Electronic Materials, Jun 2005 by Stoltz, A J, Sperry, M J, Benson, J D, Varesi, J B, Et al

We report on the physical attributes of an argon-hydrogen plasma and the effects that induced changes in these attributes have on the physical and electrical characteristics of the plasma itself. Changes in the plasma conditions of these argon-hydrogen plasmas due to variations in microwave power, DC biasing, gas concentrations, and pressures were measured. We determined that increasing the hydrogen flow increases the sheath potential of the plasma, thereby increasing the arrival energy of ions at the surface of a sample placed in the plasma. Even with the decrease in plasma density from an increase in hydrogen input flow, we found the ion current is maintained in the predominately hydrogen plasma and is likely compensated by the high velocity and long mean free path of the hydrogen. We also observed that increasing total pressure also results in hydrogen ions dominating the total number of ions reaching the Langmuir probe and therefore the sample during processing. Last, a model based on the ion/electron energy ratio was developed and used to determine the relative ion concentrations of hydrogen and argon ions.

Key words: Langmuir probe, electron cyclotron resonance (ECR), Ar-H^sub 2^ plasmas

INTRODUCTION

Argon-hydrogen plasmas, generated in reactors with either the electron cyclotron resonance (ECR) or the inductively coupled plasma (ICP) configurations, have been demonstrated to be very effective in etching CdTe, CdZnTe, and HgCdTe materials for focal plane array applications.1"8 The effects on the material induced by changes in the process chemistry have been well documented.7'8 Little is known of the effects induced within the argon-hydrogen plasma itself by changes in the various process parameters associated with the operation of the reactor.9 These parameters include microwave or ICP power, DC bias, gas ratio, and total pressure.

A Langmuir probe is useful for sampling currents within a plasma.10-12 With such a probe, the floating point potential, plasma potential, and the electron temperature can be measured directly without making any assumptions about the plasma. From these values, ion current and electron density can be calculated. Using either the Bohm or the Laframboise collision model, along with assumptions about ion mass, ion density can then be determined. We report here a Langmuir probe investigation of an ECR argon/hydrogen plasma.

In this study, we independently varied several ECR input parameters. These include microwave input power, the DC bias, hydrogen input flow, and total pressure. Langmuir probe current/voltage relationships are being used to examine the sensitivity of each of these parameters on argon-hydrogen plasmas.

EXPERIMENT

A model 357 PlasmaQuest ECR reactor (Nexx Systems, Wilmington, MA) was used to perform all of the plasma processing in this paper. The DC bias source is driven at 40.68 MHz, and the microwave source is driven at 2.45 GHz with a 875 Gauss resonance zone. Unless stated otherwise, the plasma conditions are as follows: (1) the argon flow rate was varied over the range of 0 to 80 seem, (2) the hydrogen flow rate was varied over the range of 0 to 50 seem, (3) the system pressure was varied between 1.0 and 10.0 mtorr, (4) the microwave power was between 150 to 600 W, and (5) the DC power input was varied over a range of 0-180 W. The sample cooling chuck was held at a constant 00C with 10 torr of backside He cooling. A more complete description of this system is described in Refs. 2 and 3. No RF interference is expected as a result of the high frequencies, and the primaryreference configuration is employed in this system to further reduce RF interference.11-12

A Scientific Systems Langmuir probe (Dublin, Ireland) with a reference probe was used in these experiments. The main probe tip is 10-mm long and has a radius of 0.19 mm. Version 3.16 of the probe's included SmartSoft software was used for the data acquisition and data analysis. The probe was positioned 2.75 cm from the center of the sample chuck and 1 cm above it at a 70� angle from the normal. It is important to note that the values measured are directly above where the sample would normally sit during processing and not in the center of the plasma giving results more indicative of particles reaching the sample. A 3-in. silicon wafer was placed on the sample chuck, to avoid damaging the backside cooling O-ring. Since a process pressure of 1.0 to 10.0 mtorr was used in these experiments, we assume collisionless sheath conditions.11

MICROWAVE POWER VARIATIONS

The microwave klystron is the primary source of energy for ionizing gases in a plasma.13,14 We first examined variations in microwave power. Figure 1 shows the ion density and the ion current per area (the surface area of the probe) as a function of microwave input power. As the microwave power increases, both the ion density and the ion current also increase. This increase is consistent with the literature in that the greater the microwave power, the greater the amount of ionization that occurs.13 Over the range of 250-450 W, increasing power does not increase the ion density. The power is still being absorbed and therefore changing the plasma; however, it is not creating further ionization. We believe that is due to another kinetic dynamic other than ionization, perhaps disassociation of hydrogen, as energy due to disassociation does not contribute to ionization; however, other mechanisms are possible.