Semiconductor device manufacturing: process – Bonding of plural semiconductor substrates – Subsequent separation into plural bodies
Reexamination Certificate
2000-01-06
2002-07-16
Mulpuri, Sav itri (Department: 2812)
Semiconductor device manufacturing: process
Bonding of plural semiconductor substrates
Subsequent separation into plural bodies
Reexamination Certificate
active
06420242
ABSTRACT:
FIELD OF THE INVENTION
The invention relates generally to the formation of thin films. In particular, the invention relates to transferring a thin film from its growth substrate to another substrate.
BACKGROUND ART
Compound semiconductors, such as the III-V semiconductors, are known to offer superior performance in some special applications, for example, for high-speed and high-temperature electronics and for optical emitters and detectors in particular optical wavelength bands. For efficient semiconductor operation, it is generally required that the semiconductor be crystalline, that is, have a regular atomic arrangement. The technologies for the growth of singly crystalline large-scale bodies have been greatly advanced for silicon (Si), to a reduced extent for gallium arsenide (GaAs), and to a yet lesser extent for indium phosphide (InP). For other compound semiconductors, such as GaN, bulk crystalline substrates are not readily available. Sometimes, the unavailability results from an immaturity in the technology. However, bulk crystalline GaN substrates are intrinsically very difficult to grow because of the high vapor pressure of nitrogen above molten GaN. For these materials, the usual practicable procedure involves epitaxially growing the compound semiconductor upon a crystalline substrate of another material that is more easily formed into a crystalline substrate, that is, heteroepitaxy.
Gallium nitride (GaN) is a very interesting III-V semiconductor having a bandgap corresponding to the required bandgap for blue lasers and other optical devices emitting in the blue region of the spectrum. Semiconductor optical emitters in the red, yellow, and even green portions of the spectrum are known, but blue emitters are not widely available but are greatly desired both for their very short emission wavelength, enabling a dense, data recording or reading, and also for the completion of a three-color optical spectrum, thus enabling a full multicolor display. An active device based upon GaN needs to be epitaxially grown upon a substrate, but singly crystalline substrates of GaN or other equally difficult compound substrates are not readily available. The alloy system (Al, In, Ga)N provides bandgap control over the entire visible spectrum.
In the case of gallium nitride, it has been discovered that GaN thin films can be grown on substrates of sapphire, which is a form of Al
2
O
3
. A plane of the hexagonal crystal structure of GaN is closely matched to a crystallographic plane of sapphire. Foreign growth substrates are known for other compound semiconductors. High-quality sapphire substrates of up to 150 mm diameter are available at high but reasonable prices. Once the GaN thin film has been epitaxially formed over the sapphire substrate, it may be processed into electronic and opto-electronic devices based upon the semiconductive properties of GaN. However, this fabricational approach does not produce the commercially best devices. In the case of GaN, the sapphire growth substrate introduces difficulties in the fabrication process. For example, sapphire cleaves in a basal plane which is perpendicular to the direction in which GaN epitaxially grows on sapphire. As a result, the GaN/sapphire composite cannot be as easily diced as silicon. The lack of good cleavage with GaN on sapphire is particularly a problem with GaN edge-emitting laser requiring a highly reflective (i.e., smooth) end face, typically provided in other materials by a cleaved face. Reactive ion etching has been used to for etching vertical reflector walls, but control of verticality is a problem. Sink et al. have disclosed an alternative process in “Cleaved GaN facets by wafer fusion of GaN to InP,”
Applied Physics Letters
, vol. 68, no. 15, 1996. However, this process requires abrading away most of the entire sapphire substrate, a tedious task and one prone to harm the adjacent GaN thin film.
The processes described above do not address the need for a GaN opto-electronic chip to be integrated with an electronic semiconductor chip, most particularly, of silicon.
For these reasons, a number of technologies have been developed to detach thin films of compound semiconductors from their growth substrates and to reattach them to other substrates, whether they be of silicon, other semiconductors, or non-semiconductive materials.
Yablonovitch originated the technology of transferring GaAs-based thin films from a GaAs growth substrate to a silicon substrate. The process is described by Yablonovitch et al. in “Extreme selectivity in the lift-off of epitaxial GaAs films,”
Applied Physics Letters
, vol. 51, no. 26, 1987, pp. 2222-2224 and by Fastenau et al. in “Epitaxial lift-off of thin InAs layers,
Journal of Electronic Materials
, vol. 24, no. 6, 1995, pp. 757-760. This same process is described in U.S. Pat. No. 4,883,561 to Gmitter et al. In this process, an epitaxial sacrificial layer is first grown on the substrate, and then the desired film is epitaxially grown on the sacrificial layer. The as-grown film is separated from its growth substrate by selectively etching away the sacrificial layer with a liquid etchant which attacks neither the substrate nor the GaAs thin film, thereby lifting off a free-standing film. The free-standing film can then be bonded to a substrate of silicon or other material by one of a variety of methods, as has been described by Yablonovitch et al. in “Van der Waals bonding of GaAs on Pd leads to a permanent, solid-phase-topotaxial, metallurgical bond,”
Applied Physics Letters
, vol. 59, no. 24, 1991, pp. 3159-3161. The so bonded GaAs thin film can then be further processed to form devices or circuits that integrate the functionalities of GaAs and of the substrate material. Quantum wells and other advanced structures can be grown on the GaAs thin film prior to liftoff. Similar results of bonding InAs thin films onto GaAs substrates have been reported by Fastenau et al., ibid.
These prior-art processes have not addressed the important compound GaN as well as other non-GaAs compound semiconductors. Further, the prior-art processes rely upon a liquid etchant dissolving from the side s a very thin sacrificial layer between the growth substrate and the epitaxially formed film. Such a separation process is geometrically disadvantageous and results in separation times that are commercially un economic for large-area films.
Kelly et al. have reported an alternative separation process for GaN films in “Optical process for liftoff of group III-nitride films,”
Physica Status Solidi
(a), vol. 159, 1997, pp. R3, R4. In this process, a GaN film is epitaxially grown on a sapphire substrate. The resultant structure is then irradiated from the sapphire side with an intense laser beam at a wavelength of 355 nm. This wavelength is within the sapphire bandgap so that the radiation passes through it, but the irradiation wavelength is somewhat outside of the absorption edge of GaN. As a result, a significant portion of the laser energy is absorbed in the GaN next to the interface. The intense heating of the GaN separates the gallium from gaseous nitrogen, thereby separating the GaN thin film from the sapphire substrate. It is known that GaN undergoes incongruent decomposition at temperatures above about 800° C.
The process of Kelly et al., however, suffer s difficulties. We observe that the 355 nm radiation of Kelly et al. has sufficient power that the overlying GaN film is explosively blown away during the laser irradiation, and fracturing of the film often occurs. Obviously, such an explosive process does not dependably produce uniform films. Even if the explosive process is acceptable, the area of the high-energy laser beams required for this process is limited. The beam of Kelly et al. has a diameter of 7 mm. If 7 mm portions are being separated, then it is impossible to obtain film fragments of larger size. It is greatly desired to obtain larger film segments by a process that is not so violent.
SUMMARY OF THE INVENTION
The invention may be summarized as a method of transferring a crystalline film from a g
Cheung Nathan W.
Sands Timothy D.
Wong William S.
Mulpuri Sav itri
Pennie & Edmonds LLP
LandOfFree
Separation of thin films from transparent substrates by... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Separation of thin films from transparent substrates by..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Separation of thin films from transparent substrates by... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2907161