The NMRC's Thomas Swan MOVPE reactor is of a conventional horizontal design capable of growing onto two 2" wafers and is normally operated at atmospheric pressure. The system is equipped with 4 hydride cylinders and 7 alkyl sources. A carbide-coated graphite susceptor is inductively coupled to a 5 KW rf power supply enabling a maximum growth temperature in excess of 950°C. Particular attention was paid to the design of the inner liner quartz tube to achieve good uniformity across 2 wafers. An extremely valuable feature of the NMRC's MOVPE reactor are the use of Epison sensors for monitoring the TMIn pick-up rate from the bubblers. They measure and control the concentration of TMIn in the carrier gas after the bubbler and as a result ensure both a high degree of composition uniformity within a structure and excellent reproducibility from run-to-run. The entire system is operated via a personal computer using full control software and data logging capabilities.

An essential factor in achieving high quality growth is having a wide range of characterisation techniques immediately available. Within the MOVPE clean room suite the following characterisation techniques are readily available.

Double Crystal X-Ray Diffraction

Fourier Transform Photoluminescence (PL), Room Temp and 77K

Absorption at Room Temp

Hall Measurements, Room Temp and 77K

C-V Etch Profiler

Normarski Microscope

Scanning Electron Microscope equiped with Cathodoluminescence and Energy

Dispersive X-Ray Analysis (EDX)

Given below is the structure of an epiwafer that was grown to produce LEDs that emit at 940 nm. A particularly important aspect in this variety of structure is that the correct PL emission energy is achieved across the entire wafer.

This wafer was fully characterised by making measurements of (i) surface morphology by optical inspection by Normarski microscopy, (ii) photoluminescence peak wavelength across the full wafer, (iii) layer thickness by cross-sectional scanning electron microscopy, (iv) composition variation from double crystal X-ray diffraction spectra across the wafer and (v) the carrier concentration from a C-V profile.

Fig 2. The PL spectrum of the LED wafer taken from the wafer's centre. Note that due to the high quality both heavy-hole and light-hole transitions are observed

WAVELENGTH DISTRIBUTION % Total Area Mean +/- Std.Dev. 8.90 0.940 +/- 0.001 35.58 0.939 +/- 0.002 80.06 0.939 +/- 0.004

Fig 3. Shows a map of the peak PL wavelength of the LED wafer indicating the very high uniformity of the wafer

Fig 4. The free carrier concentration through the LED structure. The measurement was taken from a ¼ wafer test piece that is placed alongside the full wafer. The test piece was grown onto a p-type substrate to aid measurements of layer thickness from SEM micrographs

Fig 5. The DXRD spectrum of the LED wafer. From the peak splitting between the substrate and epilayer the Al composition of layer 2 can be determined

The III-V group has been involved in GaAs Schottky diode fabrication for over 10 years. The epitaxial structure given below in Fig. 6 is that typically used to fabricate varistors. To achieve the ideal device characteristics it is important that the structure is grown with a low defect density and uniform doping profile, particularly in layer 1. Since C-V profiling, to determine the free carrier concentration, is a destructive technique, an alternative must be found to map the carrier concentration. In this instance we use Raman microscopy where positions of plasmon features (see Fig.7) are sensitive to the free carrier concentration. As shown in Fig.8 the uniformity of the doping in layer 1 is exceedingly high.

Large area (4µm diameter) Schottky test structures fabricated from the material presented in Fig. 8 using a 5 level Ohmic and a Pt/Au Schottky contacts gave the following results.