3.2 ITO Deposition by Reactive r.f. Sputtering

The basic principles of reactive r.f. sputtering technique are described below. This is followed by a discussion about the calibration of the system used in this work.

3.2.1 Principles

A sputtering system consists of an evacuated chamber, a target (cathode) and a substrate table (anode). The electric field inside a sputtering chamber accelerates electrons which collide with Ar atoms producing Ar+ ions and more electrons and a characteristic purple/blue plasma. These charge particles are then accelerated by the electric field: the electrons towards the anode and the Ar+ ions towards the cathode (ITO target). When an ion approaches the target, one of the following may occur:

1. It may undergo elastic collision and be reflected.
2. It may undergo inelastic collision and be buried into the target.
3. It may produce structural rearrangement in the target material.
4. The impact may set up a series of collisions between atoms of the target leading to the ejection of one of these targets; this process is known as sputtering.

Thus the sputtering process can be likened to a break in a game of "atomic" billiards. The excited ion, representing the cue ball, strikes the atomic array of the target - the neatly arranged pack - scattering them in all directions. Some of these will be ejected in the direction of the original approaching ion i.e. normal the target surface. It is this ejected particle which is useful for deposition on the surface of the wafer. Hence the sputter process essentially involves knocking an atom or molecule out of the surface of a target. Under the right conditions, the sputtered species will travel through space until it strikes and condenses on the surface of the substrate. For further detail see references [102,103].

A r.f sputtering system allows the deposition of non-conductive materials at a practical rate. In such a system, the r.f. power alone is capable of generating the plasma and accelerate ions to the target to cause sputtering.

3.2.2 The Nordiko 1500 r.f. Sputtering System

Figure 3.4 shows the schematic of the sputtering chamber and the associated r.f. power supply of the Nordiko 1500 system. Both the top and the bottom electrodes are shielded by guard rings. The diameter of the top electrode is 20cm while that of the target electrode is 10cm and the distance between them is 7cm. The r.f. generator is operated at 13.56 MHz. Further information is available in the Nordiko r. f. sputtering handbook [104].

Figure 3.4: Schematic of the r.f. Sputtering Chamber

The material to be sputtered is made into a target and mounted onto a circular copper backing plate using Ablebond 84-1MI heat resilient adhesive [105]. In this case the target consists of a circular disk of hot pressed 99.999% purity ITO (90% In2O3 + 10% SnO2) which is 4 inch diameter and approximately 0.25 inch thick (available from [106]). During deposition, the sample is inverted and placed into substrate table facing the target. There are two shutters which separate the target and the substrate. These help prevent contamination of the target during sample loading and unloading, protect the sample during pre-conditioning as well as provide means of controlling ITO deposition thickness during sputtering.

The sputtering procedure is commenced by evacuating the chamber to pressures lower than 1e-6 Torr. Ar, being a noble gas which does not react with either the target or the semiconductor wafer, is then introduced into the chamber at a specified pressure. This is followed by allowing O2 into the chamber at a set rate. The r.f. suupply is then switched on and stabilised to the required power and induced d.c. bias levels; this bias is an indication of the sheath potential and is a good sign of the ion bombardment energy. During this time the substrate is shielded by the top shutter. Once pre-conditioning is complete, the top shutter is opened marking the beginning of the deposition process.

3.2.3 Controllability and Calibrations

The specifics of the system and the repeatability of the sputter conditions will be discussed next. The chamber is first "roughed" before being opened to the high vacuum pump. Figure 3.5 shows the chamber pressure dropping as it is pumped by the rotary pump through the "roughing valve"; the pressure was monitored using the thermocouple gauge 1. It is seen that the pressure drops to below 0.1 torr (the pressure which has to be attained before the high vacuum valve can be opened) in less than 10 minutes.

Figure 3.5: Chamber pressure vs. time after opening the roughing valve.

Figure 3.6 shows the schematic diagram of the by-pass pumping system of the Nordiko 1500 sputtering machine along with all the associated vents and pressure gauges. In order to standardise the oxygen content of the plasma during the sputter depositions, it is necessary to measure the partial pressure of the gas rather than the flow rate although the two are proportional for a given system, total pressure and flow rate of the other constituent gases - in this case, Ar.

Thus it was necessary to install a second pressure gauge, Penning 505, into the sputtering chamber in order to monitor actual pressure there and to measure the partial pressures of Ar and O2 gases respectively prior to exciting the plasma as the existing Hastings gauge did not cover the necessary range. The alternative is to monitor the respective flow rates of the two gases, but as this relies on the absolute pressure gradients it is not very useful for standardising the deposition conditions. However, once the partial pressures are calibrated against the flow rates, the latter was used to monitor the partial pressures during the deposition because the ion gauges become unstable once the plasma is excited.

Figure 3.6: Schematic of the by-pass pumping system

Figure 3.7 shows the chamber pressure vs. time once the high vacuum valve is opened and it is pumped by the diffusion pump. It is seen that the actual chamber pressure, as monitored by the newly installed Penning 505 gauge, is always slightly higher than the pressure between the diffusion pump and the chamber. Figure 3.7 also shows that the time taken to reach the "base pressure" (usually 1e-6 torr) is approximately 2.5 hours.

Figure 3.7: Chamber pressure vs. time after opening high vacuum valve

Once the chamber was pumped down to the base pressure, argon gas was flowed in at 150 mlcm-2s-1 and the high vacuum valve adjusted to maintain a chamber pressure of 5e-3 torr. At this stage, the "throttle pressure", P(Throttle), was found by switching off the argon flow and was found to be 2e-5 torr. Ar flow was restored and O2 was then flowed in at the required rate (over the range 15 to 105 mlcm-2s-1); the high vacuum did not need to be readjusted at any of these stages to maintain the set pressure of 5e-3 torr. The Ar was then switched off again to find the chamber pressure, P(Chamber) due to the O2 flow alone under these conditions. The partial pressure of O2, PO2, was then found using (eqn. 3.5), as follows:

Figure 3.8 shows the PO2, the oxygen partial pressure, as a function of the O2 flow rate. As expected, the partial pressure was found to be a linear function of the flow rate.

Figure 3.8: O2 Partial Pressure vs. O2 Flow Rate; the Ar flow rate was kept constant at 150 mlcm-2s-1 throughout.

3.2.4 Sample Preparation Prior to ITO Deposition

Following usual cleaning of samples in TCE, acetone, methanol and D.I. water, these were etched in a 1 HCl : 1H2O solution for up to 1 minute (in case of InGaAs samples, the soaking period was reduced to 20 seconds as InGaAs is etched in this solution). This pre-etching promotes ITO adhesion to semiconductor substrates and is particularly useful for subsequent reliable patterning of ITO by wet chemical etching techniques. A mild solution of HF (10%) for 10 seconds has also known to have been used for this pre-etching prior to ITO deposition.

Title

Dedication

Acknowledgements

Abstract

Symbols

Contents

---

This page is hosted by Dr. Shabbir A. Bashar, e-mail: shabbir@betelco.com