3.7 Thickness Monitoring and Calibration

Monitoring metal deposition thickness or wafer etch depth are regular but vital parts of the whole fabrication process. Post deposition thickness measurement is also the only easy means of monitoring the ITO sputtering where no in-situ detector is available. In the course of fabrication, the doping on wafers vary according to their depth and etching processes have to be carefully monitored and controlled. This can be done by etching for a predetermined length of time if the etch rate of a particular solution is known. Similarly, the accuracy of film thickness monitors often need to be cross checked in the case of metalisation. However, definite results are obtained if the wafer is tested after the deposition or etching is complete. Two methods were widely used for doing so in this work; the first method can be used for measuring actual thickness and is applicable for both etching and depositions while the second is only suitable for judging etch thickness. These are described next.

3.7.1 Talystepping

The Taylor-Hobson Talystep (model no. 5) is a highly sensitive purpose built instrument designed for the micro-electronics industry which can measure vertical features down to 100Å within a few percent accuracy [115]. This method involves physically traversing the surface with a fine stylus suspended from an electro-magnetically sensitive spring. Vertical movement of the stylus is amplified electronically and recorded on paper by a spark pen. After leveling the sample with respect to the traverse plane, the stylus is scanned either across the test feature or over its edge. An accurate surface profile can be built up in this way and to some extent the instrument will also reveal the surface roughness up to a given range. Lateral features wider than a few mm such as edge profile of mesas or metal pads can also be inferred at very low stylus scan speed - this dimension being limited by the physical width of the stylus. The maximum traverse length is 1 mm.

A more advanced version of this principle has been developed and incorporated into what is known as an "Atomic Force Microscope" or AFM [116]. This instrument is linked to a computer and as well as printing out surface profiles, the AFM is capable of producing 3-dimensional graphics of a relatively large surface (1mm by 1mm) which can manipulated to alter the perspective.

3.7.2 Electrical Method

The electrical method relies on the breakdown voltage of a reverse biased Schottky diode to estimate the doping concentration and hence the thickness of an etched sample whose doping profile is known. It is particularly suited to situations when the entire surface needs to be etched or the mesas are too small where either no edges remain or those that do would be unsuitable for Talysteping.

Thus the breakdown voltage is measured by placing two gold (tipped) probes firmly on the test surface. These can be treated as two Schottky diodes connected in a manner such that one is reverse biased if the other is forward biased. Thus, the reverse breakdown voltages, Vbr, of either of the diodes, derived from Poisson's equation for a Schottky diode can be expressed as [117]:

where,
Ebr = field required to cause breakdown

It is seen that the Vbr is inversely proportional to the doping concentration. Knowing the doping profile, allows the estimation of the current surface doping from the breakdown voltage (for a given bias, the current will be increased if the doping is greater and vice versa). Hence, the need for further etching can be judged on this basis.

In practice, the probes must be kept clean and sharp with a reasonable separation between them. Figure 3.10 shows the plot of a typical Vbr as a function of etch time in ammonia solution for an AlGaAs/GaAs HBT layer; typical etch rates for GaAs and Al0.3Ga0.7As are 30Å/sec and 25Å/sec respectively. The large change in the Vbr is used as guide to judge the etch depth and therefore it relies on significant variation in the doping profile of the wafer between successive layers.

Figure 3.10: Plot of a typical breakdown profile as a function of etch time for an AlGaAs/GaAs HBT wafer (no. 3075-3); an ammonia solution (8NH3 : 3H2O2 : 400H2O) was used at room temperature and below 60% humidity.

This technique has been used successfully over the course of this work to fabricate AlGaAs/GaAs HBTs and other devices where selective etching was not a viable option.

Title

Dedication

Acknowledgements

Abstract

Symbols

Contents

---

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