MATETPRO - Matériaux et Procédés pour des Produits Performants

GaN LEDs process and growth technology development on structured silicon substrates – MOSAIC

Submission summary

Light emitting diodes (LEDs) manufacturers and research laboratories are engaged in a race for higher energy efficiency (lumen per watt) but also lower fabrication costs (€uros per lumen). This last point is one of the major issues delaying the mass commercialization of white LEDs the general lighting applications.
White LEDs available today are constituted by two main parts: a blue LED which emits light under electrical injection and a luminophor (« phosphor ») which absorbs part of the blue photons and re-emits them in yellow. The combination of blue and yellow generates a white light adequate for general lighting. The blue LEDs are made of AlGaInN semiconductors which are generally grown on sapphire substrate.
The reduction of the LED fabrication costs is now on the strategic path for major industrial players in this field. The “Silicon substrate” approach is one of the solutions envisaged today to replace the traditional sapphire substrate. The use of silicon presents two main advantages:
- Cost, size, quality and availability of Si substrates,
- Simplification of the growth substrate removal (reduction of patent dependence to “Laser Lift Off” technique used for LED process on sapphire substrate)
The efficiency LEDs grown on silicon does not yet reach those of sapphire grown LEDs. This is mainly due to crystalline quality of the GaN epitaxial layers, especially the larger threading dislocation density. In addition, thermal strain forbids the growth of thick GaN layers that would enable reduced dislocation densities.

In the last 10 years, the CRHEA has developed specific buffer layers which compensate the thermal strains and which permit thicker GaN layers on silicon, but with a maximum thickness of 3 µm: beyond that level crack formation is observed. These cracks are detrimental for the device fabrication such as LEDs.
One possible solution to avoid crack formation in thick layers and hence obtain higher quality GaN layers on Silicon substrates is to pattern the substrate before epitaxy. This pattern consists in mesas separated by trenches ….like a mosaic. This solution seems to be promising since a Chinese company (Lattice-power) has announced recently blue LEDs on silicon substrates with performances close to state of the art sapphire based diodes. For that, they grew thick GaN layers on structured silicon.
In the frame of the MOSAIC project, we intend to develop a dedicated substrate patterning that will permit growth on silicon of very thick and high quality GaN layers compatible with high efficiency LED.

In addition, the “mosaic” structure will be less size dependent and emphasize the scalability of this process.
As the challenges are both technical and economical, we will work on demonstrating the potential of mosaic substrate through a complete LED fabrication process flow and analyze the advantages of using Silicon.

Project coordinator

Monsieur Benjamin Damilano (CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE - DELEGATION REGIONALE COTE D'AZUR) – bd@crhea.cnrs.fr

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.

Partner

SILSEF SILSEF
CRHEA CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE - DELEGATION REGIONALE COTE D'AZUR
CEA-LETI COMMISSARIAT A L'ENERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES - CENTRE DE GRENOBLE

Help of the ANR 894,754 euros
Beginning and duration of the scientific project: December 2010 - 36 Months

Useful links