Gallium nitride (GaN), an III-V compound semiconductor, holds a dominant position in the solid state lighting industry and has emerged as the preferred choice in the power electronics market. GaN devices such as an LED has a heterostructure grown on a substrate by metal-organic vapor phase epitaxy or similar techniques. The heterostructure of a blue LED includes oppositely doped n- and p-type GaN layers that sandwich one or more active layers (quantum wells). When operated in a forward biased direction, electrons drop down from the conduction band of n-type GaN layer. They move across the junction between the n- and p-type GaN layers and recombine with holes from the valence band of the p-type GaN layer. When an electron fills a hole it releases energy in the form of a photon (a packet of light).
Epitaxial growth is the key enabling technology for the fabrication of GaN based devices. The heteroepitaxial growth of GaN is carried out on a carrier substrate (epitaxial wafer) made of a single crystal material such as sapphire, silicon (Si), silicon carbide (SiC), or GaN. Thus GaN devices can be classified by the type of carrier substrate used to epitaxially grow GaN. GaN-on-Sapphire, GaN-on-SiC, GaN-on-Si, and GaN-on-GaN devices refers to GaN devices epitaxially grown on sapphire, silicon carbide (SiC), silicon (Si), and gallium nitride (GaN), respectively.
The performance and efficiency of GaN devices depend very much on the epitaxial growth of GaN. In the LED lighting industry, most innovation sits in the choice of epitaxial wafers. Most GaN LEDs are grown on sapphire. The material properties and chemical structure of sapphire support a relatively high quality GaN epitaxial growth and a practical device efficiency. Other noticeable advantages of sapphire include low material and manufacturing costs and high processability in a wide range of temperatures. A major disadvantage associated with this architecture is that there is a large mismatch (13%) between its crystal lattice structure and that GaN. Such a mismatch introduces microcracks (called “threading dislocations”) into semiconductor dies. A high density of threading dislocations can affect LED efficacy and reliability.
Silicon carbide (SiC) was commercially introduced by Cree as a high performance alternative to sapphire for heteroepitaxial growth of GaN LEDs. Compared with sapphire, SiC has a closes lattice match to GaN (3.4% lattice mismatch), a simpler nucleation layer structure, and a significantly higher thermal conductivity. These attributes make SiC a superior substrate material for high power LED packages which have a high operating temperature and power density. Additionally, GaN-on-SiC LEDs are less sensitive to electrostatic discharge (ESD). ESD is one of the common failure mechanisms for GaN-on-Sapphire LEDs. The disadvantages associated with SiC are high cost of substrate production and limited wafer size.
The GaN-on-GaN architecture elimiates the concerns of lattice and CTE mismatches that're present in LEDs using foreign carrier substrates such as silicon carbide (SiC), sapphire and silicon (Si). These allows GaN-on-GaN LEDs to operate reliably at very high current densities. An extremely low defect density (1000x fewer defects than conventional LEDs) allows GaN-on-GaN LEDs 5 higher flux densities than conventional LEDs.
While silicon was a very attractive low cost alternative, it remained very challenging to grow high quality epitaxial layers on silicon substrates due to the intrinsically large lattice mismatch (17%) and high thermal expansion coefficients (54%).
Epitaxial growth is the key enabling technology for the fabrication of GaN based devices. The heteroepitaxial growth of GaN is carried out on a carrier substrate (epitaxial wafer) made of a single crystal material such as sapphire, silicon (Si), silicon carbide (SiC), or GaN. Thus GaN devices can be classified by the type of carrier substrate used to epitaxially grow GaN. GaN-on-Sapphire, GaN-on-SiC, GaN-on-Si, and GaN-on-GaN devices refers to GaN devices epitaxially grown on sapphire, silicon carbide (SiC), silicon (Si), and gallium nitride (GaN), respectively.
The performance and efficiency of GaN devices depend very much on the epitaxial growth of GaN. In the LED lighting industry, most innovation sits in the choice of epitaxial wafers. Most GaN LEDs are grown on sapphire. The material properties and chemical structure of sapphire support a relatively high quality GaN epitaxial growth and a practical device efficiency. Other noticeable advantages of sapphire include low material and manufacturing costs and high processability in a wide range of temperatures. A major disadvantage associated with this architecture is that there is a large mismatch (13%) between its crystal lattice structure and that GaN. Such a mismatch introduces microcracks (called “threading dislocations”) into semiconductor dies. A high density of threading dislocations can affect LED efficacy and reliability.
Silicon carbide (SiC) was commercially introduced by Cree as a high performance alternative to sapphire for heteroepitaxial growth of GaN LEDs. Compared with sapphire, SiC has a closes lattice match to GaN (3.4% lattice mismatch), a simpler nucleation layer structure, and a significantly higher thermal conductivity. These attributes make SiC a superior substrate material for high power LED packages which have a high operating temperature and power density. Additionally, GaN-on-SiC LEDs are less sensitive to electrostatic discharge (ESD). ESD is one of the common failure mechanisms for GaN-on-Sapphire LEDs. The disadvantages associated with SiC are high cost of substrate production and limited wafer size.
The GaN-on-GaN architecture elimiates the concerns of lattice and CTE mismatches that're present in LEDs using foreign carrier substrates such as silicon carbide (SiC), sapphire and silicon (Si). These allows GaN-on-GaN LEDs to operate reliably at very high current densities. An extremely low defect density (1000x fewer defects than conventional LEDs) allows GaN-on-GaN LEDs 5 higher flux densities than conventional LEDs.
While silicon was a very attractive low cost alternative, it remained very challenging to grow high quality epitaxial layers on silicon substrates due to the intrinsically large lattice mismatch (17%) and high thermal expansion coefficients (54%).
