Semiconductor device manufacturing: process – Making device or circuit emissive of nonelectrical signal – Compound semiconductor
Reexamination Certificate
1999-10-12
2002-06-18
Christianson, Keith (Department: 2813)
Semiconductor device manufacturing: process
Making device or circuit emissive of nonelectrical signal
Compound semiconductor
Reexamination Certificate
active
06406931
ABSTRACT:
BACKGROUND
The present invention relates generally to the growth of compound semiconductor thin films, and more specifically to controlling certain properties thereof by introduction of buffer layers.
The III-V semiconductors are known for the versatility of their electronic and photonic properties, relatively easy growth processes, and, in most cases, ease of doping. As a result, they form the basis for much of current optoelectronics.
The GaN system, by which is meant gallium nitride itself and III-V semiconductor alloys based on gallium nitride, but which include additional III-V components, such as In, Al, or As. These alloys can be ternary alloys, quaternary alloys, or still more complex alloys. The main requirement is that they can be deposited with a single crystal structure, that is, without separation into distinguishable phases.
Unique among the III-V semiconductors, the GaN system includes alloys with bandgaps reaching far into the ultraviolet. Al
0.9
Ga
0.1
N has a bandgap of about 6 electron volts, corresponding to a photon wavelength of about 0.18 &mgr;m (visible light is from 0.4 to 0.7 &mgr;m). Accordingly, these materials are of great interest in the development of ultraviolet photonic devices.
In particular, GaN system devices which emit light are of primary interest, as alternate sources of monochromatic or near-monochromatic photons are either unavailable or inappropriate for widespread use.
At this point, both GaN-based light emitting diodes (LEDs) and GaN-based laser diodes (LDs) have been demonstrated, with function restricted to the 2-3 eV regime (roughly 0.35 &mgr;m to 0.5 &mgr;m emission). This limitation is essentially structural in nature.
It is a challenge to grow the GaN system in large single crystal form, so it is difficult to develop devices which are structurally and chemically matched to their substrate. Further, the GaN system has quite short lattice constants compared to the other III-V materials, so using a closely related hetero-substrate is not a practical option either.
As is often done in such cases, a search is made for a substrate with good intrinsic qualities, and upon which the desired material will grow epitaxially, even though the crystal structure and orientation may differ dramatically. In this case, considerable success has been achieved by growing thick (
~
3 &mgr;m) layers of GaN on sapphire substrates. This causes the generation of a variety of structural defects in the GaN-based layer, but the layer is epitaxial, nearly totally relaxed to its nominal lattice constant, and structurally consistent across the substrate. This thick layer of GaN then serves as a pseudo-substrate for device growth and fabrication.
Typically, however, fabrication of devices does not simply require a mechanically and chemically compatible substrate surface, but also makes particular requirements on the electronic structure and properties of the substrate, it is well-known that, even when extrinsic dopants are not added, an as-grown GaN film on sapphire usually exhibits slightly n-type behavior, with equivalent carrier electron densities of 10
16
to 10
18
cm
−3
or more. These residual electrons are not associated with defects of uniform structure, nor are they well-understood. It is believed that these residual electrons are associated with the various structural defects characteristic of the highly defective GaN on sapphire lattice structure.
Given that some types of electronic devices require layers with well characterized dopant levels having concentrations as low as 10
17
cm
−3
, it can be seen that the uncompensated levels found in GaN on sapphire can dramatically alter, or even dominate, the desired device structure.
There has not previously been a clear correlation between how the GaN layer was grown (e.g., temperature, rate, presence of a buffer layer, etc.) and the residual electron density. As a result, attempts to compensate the residual electron density by adding p-type dopants (typically Mg) have generally failed, first by misestimating the amount of Mg dopant to add, second by the extremely poor activation properties of Mg in the GaN system, and finally by changes in the degree of Mg activation in further processing steps.
The end result of these difficulties is that prior art GaN-based devices have generally exhibited reduced performance (at times degraded to nonfunctionality) when grown on substrates comprising a large layer of GaN on a sapphire wafer. However, for the reasons summarized above, such a pseudo-substrate is the best choice for a variety of reasons. There is therefore a need for a procedure or treatment which will (at least) render it possible to reliably and robustly reduce the level of residue electrons in a layer of GaN grown on sapphire to levels that do not compromise device performance or manufacturability. More generally, there is a need for a procedure or treatment which can control the level of residue electrons in such a pseudo-substrate to a desired level.
SUMMARY OF THE INVENTION
The present invention relates to methods to design and grow GaN-based layers on a sapphire substrate so that the residue electron density can be controlled to a desired value. This is accomplished by introducing a buffer layer of low temperature deposited AlN between the sapphire and the GaN layer. All other growth parameters being equal, the thickness of the AlN layer will determine the final residue electron density. Alternately, the residue electron density can be controlled by varying the length and temperature of an annealing step introduced after the AlN layer is grown, but before the GaN layer is grown on top. GaN layers having nominally identical structure and chemistry can thereby be grown with resistivity values ranging from essentially intrinsic properties to below
~
0.1 &OHgr;-cm.
REFERENCES:
patent: 5290393 (1994-03-01), Nakamura
Ebel, R. et al., “Buffer layers for the growth of GaN on sapphire by molecular beam epitaxy”, J Crystal Growth pp. 433-436, May 1999.
Figiel Jeffrey J.
Han Jung
Bieg Kevin W.
Christianson Keith
Dodson Brian W.
Sandia Corporation
LandOfFree
Structural tuning of residual conductivity in highly... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Structural tuning of residual conductivity in highly..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Structural tuning of residual conductivity in highly... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2923946