This project is directed at the development of improved InP substrate manufacturing technology with demonstrated benefits for discrete and integrated opto-electronic device yield and hence cost. Its objective is to support collaborative technology development within Europe to ensure that supply of critical materials of the highest quality can be guaranteed.
Substrate defect structure, surface characteristics and purity are significant factors which determine the current quality of metal organic chemical vapour deposition (MOCVD) grown epitaxial layer structures used in the fabrication of long wavelength indium phosphide/indium galluim arsenic phosphide optoelectronic devices. Consequently this project directs significant effort at the development of improved bulk crystal growth and substrate finishing technology which addresses these problems.
Crystal growth in a large scale liquid encapulated Czochralski (LEC) system has been achieved and the ability to produce both longer (up to 300 mm) and large diameter (up to 80 mm) ingots demostrated. Thermal characterization and improved furnace design have led to reduced crystal dislocation content in this system.
Semiinsulating material requirements demand high background purity in the crystals. Low residual carrier levels and impurity content determined by secondary ion mass spectrometry (SIMS) analysis have been demonstrated. This has allowed control over the concentration of iron dopant to a minimum level. The possibility of undoped semiinsulating indium phosphide has been investigated by use of heat treatments combined with electrical and electron paramagnetic resonance (EPR) characterisation. Modifications in bulk properties by high temperature treatments have been shown and semiinsulating behaviour extended to lower iron dopant concentrations.
Microprecipitate bulk defects and their relationship with dislocations have been studied by a variety of techniques. New etching methods have allowed the study of the influence of these defects on epitaxy and their behaviour during expitaxial processing.
Substrate surface quality has been an important topic in the work. New techniques for chemomechanical polishing have been developed resulting in overall improved surface specifications on geometrical tolerance, surface defect density and chemical cleanliness. The overriding imp ortance of epitaxial ready quality was established for MOCVD processing. This has been achieved in demonstrator substrates of semiconducting tin doped and semiinsulating iron doped material. New optical surface characterisation tools have been developed based on phase stepping microscopy. These have been applied to the quantitative assessment of surface roughness and patterningdue to device processing, with vertical resolution at the nanometre scale. Assessment surveys of the demonstrator substrates against state of art wafers from other sources show these to be comparable in characteristics such as background purity, bulk defect density and surface physical properties.
Technical Approach

Substrate defect structure, surface characteristics and purity are significant factors which determine the current quality of MOCVD grown epitaxial layer structures used in the fabrication of long wavelength InP/InGaAsP opto-electronic devices. Consequently this project directs significant effort at the development of improved bulk crystal growth and substrate finishing technology which addresses these problems.

The relationship between device behaviour and substrate quality has yet to be fully defined. It therefore is intended that substrates developed throughout this project should be fully characterised using appropriate measurement techniques. Selected samples will be taken through to device fabrication and lifetime assessment/yield evaluation. In this way feedback may be rapidly provided throughout the project in order to direct and optimise the substrate development for device yield. It is also believed that MOCVD growth technology is a significant factor which ultimately will limit device yield, and hence become increasingly important as opto-electronic integration takes place. To complement the substrate development activity therefore, this project also addressed the major issue of MOCVD growth and processing of whole wafers.

The main topic during the final year of the project will be the study of the detailed influence of substrate surface properties on epitaxy and device performance, including lifetime.

Key Issues

The key issue for this project is in the establishment of large area substrates of adequate quality which will underpin the development and production of opto- electronic devices for low cost CAC and CPN applications in the IBC.

Achievements

Crystal growth in a large scale LEC system has been achieved and the ability to produce both longer (up to 300 mm) and larger diameter (up to 80 mm) ingots demonstrated. Thermal characterisation and improved furnace design have led to reduced crystal dislocation content in this system.

Semi-insulating material requirements demand high background purity in the crystals. Low residual carrier levels and impurity content determined by SIMS analysis have been demonstrated. This has allowed control over the concentration of Fe dopant to a minimum level. The possibility of undoped semi-insulating InP has been investigated by use of heat treatments combined with electrical and EPR characterisation. Modifications in bulk properties by high temperature treatments have been shown and semi-insulating behaviour extended to lower Fe dopant concentrations.

Micro-precipitate bulk defects and their relationship with dislocations have been studied by a variety of techniques. New etching methods have allowed the study of the influence of these defects on epitaxy and their behaviour during epitaxial processing.

Substrate surface quality has been an important topic in the work. New techniques for chemo-mechanical polishing have been developed resulting in overall improved surface specifications on geometrical tolerance, surface defect density and chemical cleanliness. The overriding importance of 'epitaxial ready' quality was established for MOCVD processing. This has been achieved in demonstrator substrates of semi-conducting Sn doped and semi-insulating Fe doped material. New optical surface characterisation tools have been developed based on phase stepping microscopy. These have been applied to the quantitative assessment of surface roughness and patterning due to device processing, with vertical resolution at the nanometre scale.

Assessment surveys of the demonstrator substrates against state of art wafers from other sources show these to be comparable in characteristics such as background purity, bulk defect density and surface physical properties.

Expected Impact

The success of this project is likely to improve costs of O/E devices via improved intrinsic quality and better yields from large area substrates. Establishment of a capability for competitive quality wafer manufacture for O/E semiconducting substrates will provide users with a European supply option.

The use of semi-insulating substrates for integrated opto-electronics is gaining increasing importance and will fuel future demands for these substrate types, for higher bit rate applications.