|Publication number||US20050247948 A1|
|Application number||US 11/184,190|
|Publication date||Nov 10, 2005|
|Filing date||Jul 19, 2005|
|Priority date||Feb 21, 2000|
|Also published as||EP1187229A1, EP1187229A4, US6979844, US20020158253, US20030183835, WO2001061766A1|
|Publication number||11184190, 184190, US 2005/0247948 A1, US 2005/247948 A1, US 20050247948 A1, US 20050247948A1, US 2005247948 A1, US 2005247948A1, US-A1-20050247948, US-A1-2005247948, US2005/0247948A1, US2005/247948A1, US20050247948 A1, US20050247948A1, US2005247948 A1, US2005247948A1|
|Inventors||Tetsuji Moku, Kohji Ohtsuka, Masataka Yanagihara, Masaaki Kikuchi|
|Original Assignee||Tetsuji Moku, Kohji Ohtsuka, Masataka Yanagihara, Masaaki Kikuchi|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (12), Referenced by (1), Classifications (9)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to a light-emitting semiconductor device composed primarily of gallium nitride (GaN)-based semiconductors, and to a method of making the same.
The compound semiconductors composed primarily of GaN have been used extensively for fabrication of light-emitting devices such as diodes that are capable of glowing in blue. Examples of such compound semiconductors include, in addition to GaN itself, gallium aluminum nitride (GaAlN), indium gallium nitride (InGaN), and indium gallium aluminum nitride (InGaAlN).
A typical prior art light-emitting device of the kind under consideration comprises a baseplate of electrically insulating material such as sapphire, a buffer layer overlying the baseplate and composed for example of GaxAl1-xN, where x is greater than zero and not greater than one (as taught by Japanese Unexamined Patent Publication No. 4-297023), an n-type semiconductor region of GaN or other compound semiconductor composed principally of GaN and grown epitaxially on the buffer layer, an active layer of another compound semiconductor composed principally of GaN (e.g. InGaN) and grown epitaxially on the n-type semiconductor region, and a p-type semiconductor region grown epitaxially on the active layer. The n-type semiconductor region is connected to a cathode, and the p-type semiconductor region to an anode.
The common practice in the manufacture of light-emitting devices is first to form wafers on which there are fabricated matrices of desired devices, and to cut them into the individual devices as by dicing, scribing, or cleavaging. The noted sapphire baseplate of the light-emitting devices has been a cause of trouble in such dicing of the wafers because of its extreme hardness. Sapphire itself is expensive, moreover, adding much to the manufacturing costs of the light-emitting devices.
There have been additional difficulties in connection with the sapphire baseplate. Being electrically insulating, the sapphire baseplate makes it impossible to form a cathode thereon. This inconvenience was conventionally circumvented by exposing part of the n-type semiconductor region through the active layer and p-type semiconductor region for connection to a cathode. The results were a greater surface area of the semiconductor and a corresponding increase in the costs of the light-emitting devices.
A further inconvenience arose from the fact that current flows through the n-type semiconductor region not only vertically (normal to the plane of the sapphire baseplate) but horizontally (parallel to the sapphire baseplate plane). The dimension of the n-type semiconductor region for the horizontal current flow is as small as four to five micrometers, so that the resistance of the horizontal current path of the n-type semiconductor region was very high, adding substantively to the current and voltage requirements of the prior art devices.
A still further inconvenience concerns the etching-away of parts of the active layer and p-type semiconductor region in order to expose part of the n-type semiconductor region for connection to the cathode. The n-type semiconductor region had to be dimensioned sufficiently large to allow for some errors in etching, necessitating a correspondingly elongated period of time for it to be grown epitaxially.
It has been suggested to use a conductive baseplate of silicon carbide (SiC) in substitution for the sapphire. Permitting a cathode to be formed thereon, the SiC baseplate offers such advantages over the sapphire baseplate as a smaller surface area and easier separation of the wafer by cleavaging. Offsetting these advantages, however, is the fact that SiC is even more expensive than sapphire. Another shortcoming is the difficulty of placing the n-type semiconductor region in low-resistance contact with the SiC baseplate, so that the current and voltage requirements of the light-emitting device incorporating the SiC baseplate were just as high as those of the device with the sapphire baseplate.
The present invention aims at the provision of a light-emitting device, and a method of fabrication thereof, such that the device is efficiently manufacturable at a lower cost than heretofore and is improved in performance too.
The present invention will be briefly summarized with use of the reference characters used in the subsequent detailed description of the best mode of carrying out the invention. As used in this summary, however, the reference characters are meant purely for an easier understanding of the invention and should not be taken in a limitative sense.
Briefly stated in its perhaps broadest aspect, the light-emitting semiconductor device according to the invention comprises a baseplate (11) of low resistivity composed of a silicon compound or silicon with impurities, a buffer layer (12) formed on the baseplate and having a first sublayer (12 a) of AlxGa1-xN, where x is greater than zero and not greater than one, and a second sublayer (12 b) of GaN or AlyGa1-yN, where y is less than x and greater than zero and less than one, a semiconductor region (10) formed on the buffer layer and having a plurality of sublayers of compounds composed primarily of GaN or GaN-based compound semiconductors for emission of light, a first electrode (17) formed on the semiconductor region, and a second electrode (18) formed on the baseplate.
As stated in claim 2, the sublayers of the semiconductor region (10) may include a first semiconductor sublayer (13) of a first conductivity type formed on the buffer layer (12) and made of a compound composed primarily of GaN, an active sublayer (14) on the first sublayer, and a second semiconductor sublayer (15) of a second conductivity type, which is opposite to the first conductivity type, formed on the active layer and also made of a compound composed primarily of GaN.
As stated in claim 3, the buffer layer (12) may consist of an alternation of a first set of sublayers (12 a) of AlxGa1-xN, and a second set of sublayers of GaN or AlyGa1-yN.
As stated in claim 4, the first set of sublayers (12 a) of the buffer layer 12 should each be from 5×10−4 to 100×10−4 micrometers, and the second set of sublayers thereof from 5×10−4 to 2000×10−4 micrometers.
As stated in claim 5, the light-emitting semiconductor device of the above summarized configuration may be fabricated by a method comprising the steps of providing a baseplate (11) of a single crystal of silicon containing impurities and having a low resistivity, forming by a vapor phase growth on the baseplate (11) a buffer layer (12) in the form of an alternation of a first set of sublayers (12 a) of AlxGa1-xN, where x is greater than zero and not greater than one, and a second set of sublayers (12 b) of GaN or AlyGa1-yN, where y is less than x and more than zero and less than one, forming by vapor phase growth on the buffer layer a semiconductor region (10) containing a plurality of GaN-based compound semiconductor layers for emission of light, and forming a first electrode (17) on the semiconductor region (10) and a second electrode (18) on the baseplate (11).
The invention as set forth above yields the following advantages:
The invention of claim 2 provides a light-emitting device of even more favorable performance characteristics.
According to the invention of claim 3, the buffer sublayers of AlxGa1-xN, which is relatively small in difference in lattice constant from silicon, are provided one directly on the baseplate and another between the buffer sublayers of GaN or AlyGa1-yN, resulting in improvement in the flatness of the buffer layer and the crystallinity of the semiconductor region.
According to the invention of claim 4, the first set of buffer sublayers are each so determined in thickness as to provide a tunnel effect in terms of quantum mechanics, limiting the resistance of the buffer sublayers and reducing the power and voltage requirements of the device.
The invention of claim 5 enables an easy and inexpensive fabrication of the light-emitting semiconductor device of the improved performance characteristics.
The light-emitting semiconductor device according to the invention will now be described in detail in terms of the blue-light-emitting GaN-based compound diode illustrated in
The lamination of the light-emitting semiconductor region 10, the baseplate 11 and the buffer layer 12 constitutes a substrate or base body 16. An anode 17 is formed on one of the two opposite major surfaces, or on the top as seen in the attached drawings, of the base body 16, or on the semiconductor region 15, and a cathode 18 on the other major surface, or on the bottom, of the base body. The buffer layer 12, the n-type semiconductor region 13, the active layer 14, and the p-type semiconductor region 15 are grown epitaxially on the baseplate 11, in that order and with their crystal orientation aligned.
The baseplate 11 is made of a single crystal of silicon containing impurities that determine its conductivity type. The baseplate 11 has an impurity concentration ranging from 5×1018 cm−3 to 5×1019 cm−3, and a resistivity ranging from 0.0001 ohm-cm to 0.01 ohm-cm. Made from n-type silicon into which is introduced arsenic, the baseplate 11 is low in resistivity that it serves as a current path between anode 17 and cathode 18. Additionally, being as thick as approximately 350 micrometers, the baseplate 11 functions as a support for the semiconductor region 10 and the buffer layer 12.
Thoroughly covering one surface of the baseplate 11, the buffer layer 12 is shown as an alternation of two buffer sublayers 12 a and another two buffer sublayers 12 b. In practice, however, the buffer layer may be constituted of as many as fifty sublayers 12 a and another fifty sublayers 12 b in alternation.
The first set of buffer sublayers 12 a are made from substances that can be defined by the chemical formula, AlxGa1-xN, where x is greater than zero and equal to or less than one. Examples of such substances are aluminum nitride (AlN) and aluminum gallium nitride (AlGaN). The first set of buffer sublayers 12 a are made from AlN (x being one in the general formula above) in this particular embodiment of the invention. Each such sublayer 12 a is an extremely thin, electrically insulating film.
The second set of buffer sublayers 12 b are extremely thin, insulating films of an n-type semiconductor that is either GaN or any of substances expressed by the formula, AlyGa1-yN, where y is less than x and greater than zero and less than one. In use of AlyGa1-yN for the second set of buffer sublayers 12 b, it is recommended that y be made greater than zero and less than 0.8, in order to prevent an increase in the resistance of these sublayers.
The first set of buffer sublayers 12 a should each have a thickness ranging from 5×10−4 micrometers to 100×10−4 micrometers, or 5-100 Angstroms, preferably from 10×10−4 micrometers to 80×10−4 micrometers. If less than five Angstroms in thickness, the first set of buffer sublayers 12 a would fail to keep the overlying n-type semiconductor layer 13 sufficiently flat, and, if more than 100 Angstroms in thickness, would fail to provide the desired quantum-mechanical tunnel effect, resulting in an undue increase in the resistance of the buffer layer 12.
The second set of buffer sublayers 12 b should each have a thickness ranging from 5×10−4 micrometers to 2000×10−4 micrometers, or 5-2000 Angstroms, preferably 10-300 Angstroms. If each less than five Angstroms in thickness, the second set of buffer sublayers 12 b would fail to provide a desired degree of electrical connection between the neighboring sublayers 12 a, causing an undesired increase in the resistance of the total buffer layer 12. If each more than 2000 Angstroms in thickness, on the other hand, the second set of buffer sublayers 12 b might fail to hold the overlying n-type semiconductor layer 13 sufficiently flat.
Speaking more strictly, and in this particular embodiment of the invention, the two sets of sublayers 12 a and 12 b are each 50 Angstroms. The total thickness of the buffer layer 12 is therefore 5000 Angstroms.
A method of fabricating the light-emitting semiconductor device according to the invention will now be explained on the assumption that the first set of buffer sublayers 12 a are of AlN, and the second set of buffer sublayers 12 b are of GaN.
The known metal organic chemical vapor deposition (MOCVD) method is recommended for alternate fabrication of the AlN and GaN buffer sublayers. A monocrystalline silicon substrate or baseplate 11 may first be placed in a MOCVD reaction chamber to have oxide films removed from its surfaces by thermal annealing. Then a first buffer sublayer 12 a of AlN may be formed to a thickness of approximately 50 Angstroms on one of the major surfaces of the baseplate 11 by introducing trimethyl aluminum (TMA) and ammonia (NH3) gases into the reaction chamber for approximately twenty-seven seconds. Actually, after heating the baseplate 11 to 1120° C., the TMA gas, or aluminum in effect, was supplied at a rate of approximately sixty-three micromoles per minute, and the NH3 gas, or NH3 itself, at a rate of approximately 0.14 micromoles per minute.
Then, with the heating temperature of the baseplate 11 maintained at 1120° C., the supply of the TMA gas suspended, and the gases of trimethyl gallium (TMG), NH3, and silane (SiH4) were introduced instead into the reaction chamber for approximately fifteen seconds. There will thus be created a second buffer sublayer 12 b of n-type GaN to a thickness of fifty Angstroms in overlying relationship to the first buffer sublayer 12 a on the baseplate 11. The SiH4 gas is intended for introduction of Si, an n-type impurity, into the sublayer being formed. The TMG gas, or Ga in effect, was introduced at a rate of approximately sixty-three micromoles per minute; the NH3 gas, or NH3 itself, at approximately 0.14 moles per minute; and the SiH4 gas, or Si in effect, at approximately twenty-one nanomoles per minute.
In the case where there are fifty first buffer sublayers and fifty second buffer sublayers, as in this embodiment of the invention, the foregoing process of AlN sublayer creation may be repeated fifty times, and that of GaN sublayer creation as many times, in order to form a buffer layer 12 consisting of one hundred alternating AlN and GaN sublayers. These numbers should not, however, be taken in a limitative sense: The buffer layer may be constituted of, for instance, fifty alternating such sublayers.
Next comes the step of successively fabricating the n-type semiconductor region 13, active layer 14, and p-type semiconductor region 15 on the buffer layer 12 by the MOCVD method.
First, for formation of the n-type semiconductor region 13, the baseplate 11 with the buffer layer 12 thereon was put into the MOCVD reaction chamber, into which were then introduced TMG, NH3, and SiH4 gases. The SiH4 gas is intended for introduction of Si, an n-type impurity, into the n-type semiconductor region 13. More specifically, the baseplate 11 with the buffer layer 12 thereon was heated to 1040° C. Then the TMG gas, or Ga in effect, was introduced at a rate of approximately 4.3 micromoles per minute; the NH3 gas, or NH3 itself, at approximately 53.6 millimoles per minute; and the SiH4 gas, or Si in effect, at approximately 1.5 nanomoles per minute. The n-type semiconductor region 13 was thus formed to a thickness of approximately two micrometers.
It may be noted that the n-type semiconductor region 13 is very thin compared with the thickness, from four to five micrometers or so, of the conventional LEDs. The impurity concentration of the semiconductor region 13 was approximately 3×1018 cm−3, sufficiently less than that of the baseplate 11. The formation of the semiconductor layer 13 at as high a temperature as 1040° C. is possible thanks to the interposition of the buffer layer 12.
Then the active layer 14 of p-type InGaN was formed on the n-type semiconductor layer 13. To this end, with the heating temperature of the baseplate 11 set at 800° C., there were introduced into the reaction chamber both trimethyl indium gas (hereinafter referred to as the TMI gas) and bis-cyclo pentadienylmagnesium gas (hereinafter referred to as the Cp2Mg gas) in addition to TMG and NH3 gases. The Cp2Mg gas was intended for introduction of Mg, a p-type impurity, into the active layer 14.
More specifically, for the fabrication of the active layer 14 as above, the TMG gas was introduced at a rate of approximately 1.1 micromoles per minute; the NH3 gas at approximately sixty-seven millimoles per minute; the TMI gas, or In in effect, at approximately 4.5 micromoles per minute; and the Cp2Mg gas, or Mg, at approximately twelve nanomoles per minute. The active layer 14 thus formed had a thickness of approximately 20 Angstroms and an impurity concentration of approximately 3×1017 cm−3.
Then the p-type semiconductor region 15 of p-type GaN was formed on the active layer 14. The heating temperature of the baseplate 11 was raised to 1040° C. toward this end, and there were introduced into the reaction chamber TMG, NH3, and Cp2Mg gases. The TMG gas introduced at approximately 4.3 micromoles per minute; the NH3 gas at approximately 53.6 micromoles per minute; and the Cp2Mg gas at approximately 0.12 micromoles per minute. The thus-formed p-type semiconductor region 15 has a thickness of approximately 0.5 micrometers and an impurity concentration of approximately 3×1018 cm−3.
The MOCVD growth method set forth above has proved to make possible the fabrication of LEDs such that the crystal orientation of the monocrystalline silicon substrate or baseplate 11 is favorably followed by the buffer layer 12. Additionally, the n-type semiconductor region 13, active layer 14, and p-type semiconductor layer 15 are all aligned with the buffer layer 12 in crystal orientation.
Then, for formation of the first electrode or anode 17, nickel and gold were vacuum-deposited on the top of the semiconductor body 16, that is, on the p-type semiconductor region 15 in low-resistance contact therewith. Disc-like in shape as depicted in
The second electrode or cathode was formed on the entire bottom surface of the baseplate 11, as indicated at 18, rather than on the n-type semiconductor region 13. Vacuum deposition of titanium and aluminum was used for cathode formation.
In use of the blue LED fabricated as above, the cathode 18 may be mechanically and electrically connected, as by soldering or with use of an electrically conductive adhesive, to, for instance, an electrode on a circuit board. The anode 17 may be electrically coupled to an external electrode as by wire bonding.
Constructed and manufactured as in the foregoing, the blue LED according to the invention gains the following advantages:
Notwithstanding the foregoing detailed disclosure, it is not desired that the present invention be limited by the exact details of such disclosure. The following is a brief list of possible modifications of the illustrated embodiments which are believed to fall within the purview of the instant invention:
The present invention provides LEDs and like light-emitting devices of low resistance and low power loss.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5239188 *||Nov 4, 1992||Aug 24, 1993||Hiroshi Amano||Gallium nitride base semiconductor device|
|US5585648 *||Feb 3, 1995||Dec 17, 1996||Tischler; Michael A.||High brightness electroluminescent device, emitting in the green to ultraviolet spectrum, and method of making the same|
|US5764673 *||Sep 25, 1997||Jun 9, 1998||Mitsubishi Denki Kabushiki Kaisha||Semiconductor light emitting device|
|US5831277 *||Mar 19, 1997||Nov 3, 1998||Northwestern University||III-nitride superlattice structures|
|US5923950 *||Jun 10, 1997||Jul 13, 1999||Matsushita Electric Industrial Co., Inc.||Method of manufacturing a semiconductor light-emitting device|
|US5937274 *||Dec 18, 1997||Aug 10, 1999||Hitachi, Ltd.||Fabrication method for AlGaIn NPAsSb based devices|
|US6110277 *||Apr 14, 1998||Aug 29, 2000||Temic Telefunken Microelectronic Gmbh||Process for the fabrication of epitaxial layers of a compound semiconductor on monocrystal silicon and light-emitting diode fabricated therefrom|
|US6121121 *||Jul 27, 1999||Sep 19, 2000||Toyoda Gosei Co., Ltd||Method for manufacturing gallium nitride compound semiconductor|
|US6153010 *||Apr 9, 1998||Nov 28, 2000||Nichia Chemical Industries Ltd.||Method of growing nitride semiconductors, nitride semiconductor substrate and nitride semiconductor device|
|US6165812 *||Jan 14, 1997||Dec 26, 2000||Matsushita Electric Industrial Co., Ltd.||Gallium nitride compound semiconductor light emitting device and process for producing gallium nitride compound semiconductor|
|US6255198 *||Nov 17, 1999||Jul 3, 2001||North Carolina State University||Methods of fabricating gallium nitride microelectronic layers on silicon layers and gallium nitride microelectronic structures formed thereby|
|US6566677 *||Mar 21, 2001||May 20, 2003||Sanyo Electric Co., Ltd.||Nitride-based semiconductor device and manufacturing method thereof|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US7462504 *||Jul 18, 2006||Dec 9, 2008||Lg Electronics Inc.||Surface-emitting type light-emitting diode and fabrication method thereof|
|International Classification||H01L33/00, H01L33/04, H01L33/32|
|Cooperative Classification||H01L33/007, H01L33/32, H01L33/04|
|European Classification||H01L33/00G3B2, H01L33/32|