Growth of thallium-doped silicon from a tin-thallium solution
United States Patent 4270973
A method of growing single crystals of silicon doped with thallium for use as an extrinsic silicon photodetector of 3-5 um infrared radiation which will operate above 77 K.
Inventors:
Schmit, Joseph L. (Hopkins, MN)
Scott, Walter M. (Minnetonka, MN)
Application Number:
Publication Date:
06/02/1981
Filing Date:
04/28/1980
Export Citation:
Honeywell Inc. (Minneapolis, MN)
Primary Class:
Other Classes:
International Classes:
C30B19/02; (IPC1-7): C30B19/02
Field of Search:
156/600, 156/605, 156/624, 156/DIG.64, 148/171, 148/172, 148/185
View Patent Images:
&&&&&&PDF help
US Patent References:
3010857Nelson148/1852990372Pinter et al.156/6052904512Horn156/6052759895Belmont156/605
Primary Examiner:
Bernstein, Hiram
Attorney, Agent or Firm:
Dahle, Omund R.
Parent Case Data:
This is a continuation of application Ser.
No. 900,447, filed Apr. 27, 1978, now abandoned.
The embodiments of the invention in which an
exclusive property or right is claimed are defined as follows:
1. A method for epitaxially growing a thallium-doped silicon crystal which exhibits an infrared photoconductive response from a metallic solution, comprising the steps of:
heating and melting a mixture of tin and thallium in an ampoule to thereby provide a liquid metal solvent
providing at a first location in said solv
placing at a second location separate from the first location a
providing a temperature gradient in said solution from the silicon source first location in said solution to said second location in said solution with the temperature of the solution at said silicon source first location being hotter than at said second location whereby a thallium-doped silicon crystal grows at said second location.
2. A method for epitaxially growing from a metallic solution a thallium-doped silicon crystal which exhibits an infrared photoconductive response, comprising the steps of:
providing in an ampoule at an elevated temperature a liquid metal solvent of tin and thallium having a silicon source at a first locat
maintaining the solution at a lower temperature at a second location in said ampoule, the second location being remote from the first, whereby silicon dissolves from said silicon source into said liquid metal solvent and precipitates out and epitaxially grows a thallium-doped silicon crystal at said second location.
3. A method for epitaxially growing from a metallic solution a thallium-doped silicon crystal which exhibits an infrared photoconductive response, comprising the steps of:
providing in an ampoule a liquid metal solvent consisting of tin and thallium, and maintaining one end of the ampoule h
placing a source of silicon to be dissolved
placing a silicon seed crystal
dissolving silicon into the tin and thallium solvent from the silicon sour and,
precipitating out silicon and growing an epitaxial layer of thallium-doped silicon onto said seed crystal at said cooler end.
4. The method according to claim 1, 2 or 3 wherein the temperature of the solution at the location where the crystal grows is 1330° C.
5. The method according to claims 1, 2 or 3 wherein the temperature difference of the solution between the hotter location and the cooler location is about 50° C.
Description:
BACKGROUND AND SUMMARY OF THE INVENTIONThis invention is related to the arts of growing photoconductive semiconductors from metallic solutions. Particularly, this invention provides a method of growing single crystals of silicon doped with a concentration of thallium, which thallium concentration is sufficiently high to give a high absorption coefficient. Thallium has a sufficiently large ionization energy to make extrinsic silicon detectors of 3-5 um infrared radiation which operate above 77K. The growth of semiconductors from metallic solutions is well known. It is the basis for liquid phase epitaxy (LPE) processes, for example growth of gallium arsenide (GaAs) from gallium (Ga) solution. Thick layers can be grown by the solution growth (SG) for example, silicon (Si) can be grown from indium (In) solution. Both processes depend on the concentration of solute being different at different temperatures. In the LPE process a saturated solution is cooled, causing the solute in excess of the solubility limit at the lower temperature to precipitate out as an epitaxial layer. In the SG process, a temperature gradient is imposed across a molten solvent (e.g. In) such that more solute (e.g. Si) is soluble at one end than at the other. A source of solute is placed at the hotter end and a seed at the cooler end. The greater solubility at the hotter end will cause a concentration gradient to develop and the solute will diffuse down the concentration gradient to the cooler end where it will precipitate on the seed. The material grown will contain the solubility limit of solvent at the temperature at which the seed is maintained. In the example of Si grown from In solution, the Si grown will contain the solubility limit of In at the growth temperature. The present dopant used in Si for 3-5 um response is In. Indium has too small an ionization energy however, and therefore requires operation at &60K. Thallium has a larger ionization energy in silicon than has indium, and has the advantage of a higher operating temperature. The silicon-thallium system presents a unique problem. The solubility of silicon in thallium is infinitesimal even up to 1400° C. so that silicon cannot be grown from thallium solution. However, silicon is soluble in tin (Sn) and can be grown from its solution. Tin is not electrical therefore, by adding tin to thallium we can grow silicon doped to the solubility limit of both tin and thallium. The inactive tin can be ignored while the thallium is a deep acceptor suitable for 3-5 um infrared detection. The novel feature is the addition of a second metal, tin, to the melt thallium, for single crystal solution growth. The unique feature of the tin is that it is not electrically active in silicon but dissolves enough silicon so that silicon can be grown from it.
DESCRIPTION The increased emphasis on the development of a more sophisticated generation of infrared systems having a high density of detectors and signal processors on the same focal plane has led to a renewed interest in the deeper dopants in silicon. To avoid excessive cooling of the system one would like the detector well matched to the 3-5 um spectral interval. Studies made from a systems viewpoint indicate that the ideal dopant should have an activation energy in the range 0.21-0.29 eV, and be capable of achieving background limited IR performance (BLIP) operation at temperatures above 80K. Prior to this invention, the detector material which has been successfully integrated in a monolithic structure has been indium-doped silicon. A major problem with indium-doped silicon is the longer-than-desired cut-off wavelength of photoconductors which then requires cooling to about 50K for BLIP operation. Other dopants for silicon have been investigated to determine if a likely candidate could be found, without much success. Certain dopants which have been tried are sulfur and selenium. These show the presence of multiple levels and do not limit the long wavelength response sufficiently to make a significant difference. In addition, both of these impurities are relatively fast diffusers in silicon, complicating the focal plane fabrication process if contamination by the deeper impurity is to be avoided. An impurity having desirable characteristics for the infrared application is thallium. Its ionization energy of approximately 0.246 eV is of the range providing operation in the 3-5 um range at a much higher temperature than indium-doped photoconductors. As mentioned above, a special problem exists in the use of thallium as a dopant in the growing of thallium-doped silicon crystals for monolithic focal plane arrays. We are using a novel solution growth process to produce thallium-doped silicon crystals doped to the solubility limit with thallium. The technique uses a single crystal silicon seed, a liquid metal solvent consisting of a mixture of tin and thallium and a silicon source crystal all in a quartz ampoule. During growth, a temperature gradient is superimposed on the ampoule with the source hotter than the substrate. Typically, the temperature TH at the source is about 50° C. higher than the temperature TL at the substrate (seed crystal). The amount of silicon soluble in the tin-thallium solution is greater at the higher temperature so a concentration gradient of silicon is set up in the melt corresponding to the temperature gradient. The concentration gradient causes diffusion of the dissolved silicon to the seed crystal where it grows epitaxially. The grown silicon will contain the solubility limit of tin and thallium at the growth temperature. Tin is electrically inactive in silicon so no harm is done to the electrical resistance of the silicon by its presence. The tin is used as the transport medium during growth since the solubility of silicon in pure thallium is so low that growth from a thallium solution would be infinitesimally slow. One of the thallium-doped silicon crystals grown by this process was grown in 14 days at 1150° C. at a rate of approximately 0.2 mm per day. The crystal is .about.1 cm in diameter and .about.3 mm thick. It is a single crystal and there is no evidence of a second phase in the single crystal, although there are a few tin-thallium inclusions near the edge of the crystal. It contains a Tl concentration of .about.3×1015 cm-3. Another Tl doped silicon crystal was grown by this process in four days at 1330° C. at the rate of 1 mm per day. It contains a Tl concentration of .about.5×1016 cm-3. These crystals are mentioned by way of example and do not imply that these are the conditions to give maximum solubility. This invention can obviously also be used to grow large-area, thin layers of Si doped with Tl suitable for integration with silicon electronics. Although the solution growth process is explained in detail herein as a preferred method of growing the thallium-doped silicon, it is also possible to use the liquid phase epitaxy process in which after the tin-thallium solvent has added thereto the silicon solute, the solution is cooled, causing the silicon in excess of the solubility limit at the lower temperature to precipitate out as an epitaxial layer. The resulting thallium-doped silicon crystal could be grown on the same silicon source substrate or on a different silicon substrate. This would be of particular interest for growth of thin films.
& 2004-. All rights reserved.The influence of axial magnetic fields on the growth of III–V semiconductors from metallic solutions
Loading...
You have selected 1 citation for export.
RIS (for EndNote, ReferenceManager, ProCite)
Citation Only
Citation and Abstract
JavaScript is disabled on your browser. Please enable JavaScript to use all the features on this page.
, July 1992, Pages 305-314
The influence of axial magnetic fields on the growth of III–V semiconductors from metallic solutions
Opens overlay
Opens overlay
Opens overlay
Kristallographisches Institut, Universit?t, D-W-7800 Freiburg, Germany
Check if you have access through your login credentials or your institution.
Remember me
III–V semiconductor crystals are grown by the travelling heater method in a homogeneous axial magnetic field. First results demonstrate an influence not only on transport phenomena but also on growth kinetics and morphological stability. An interpretation of the growth results is given by an order of magnitude analysis using the dimensionless Rayleigh and Hartmann numbers.
Cookies are used by this site. To decline or learn more, visit our
Remember meThis page uses JavaScript to progressively load the article content as a user scrolls.
Screen reader users, click the load entire article button to bypass dynamically loaded article content.
Login via your institution
Remember me
&RIS&(for EndNote, Reference Manager, ProCite)
&RefWorks Direct Export
& Citation Only
& Citation and Abstract
JavaScript is disabled on your browser.
Please enable JavaScript to use all the features on this page.
JavaScript is disabled on your browser.
Please enable JavaScript to use all the features on this page. This page uses JavaScript to progressively load the article content as a user scrolls. Click the View full text link to bypass dynamically loaded article content.
2007, Pages 343–452
Chapter 7 – TRAVELING HEATER METHOD Crystal Growth Laboratory Department of Mechanical Engineering University of Victoria Victoria, BC, Canada V8W 3P6 BL Consulting Ltd. Shirley Road, Victoria, BC, Canada, V9A 6M3Traveling heater method (THM) falls into the category of solution growth, and is a relatively new, promising technique for commercial production of high quality, bulk compound and alloy semiconductors. Due to its importance, a number of experimental and theoretical studies are carried out for the THM growth process. This chapter presents the recent developments in modeling of THM, particularly during the last two decades. The basic theoretical considerations regarding the modeling issues are presented first. Then, two and three dimensional simulation models and numerical simulation results are presented for binary and ternary systems. The challenges in modeling of the THM growth of ternary systems are emphasized. The use of static and rotating magnetic fields has also found a great interest in THM. The models under static and strong magnetic fields, and also those under weak and rotating magnetic fields are analyzed.
Cookies are used by this site. To decline or learn more, visit our
No articles found.
This article has not been cited.
No articles found.Method for the growth of nitride based semiconductors and its apparatus
United States Patent 5637146
A method for the growth of semiconducting nitrides, such as GaN, InN, AlN, and their alloys, in an ultra-high vacuum chamber, wherein low energy atomic nitrogen is generated by a plasma-excited radical atom source, the atom beam is introduced to the heated substrate within a short distance, other gaseous reactants and dopants, such as TMGa, TMIn, TMAj, DEZn, CP2 Mg, SiH4, and similar organmetallic and hydride sources, are injected from a circular injector located between the substrate and the atom source, and therefore large area epitaxy with high growth rate is obtained.
Inventors:
Chyi, Jen-inn (3F, No. 38, Lane 8, Gao-Shuang Road, Ping-Chen, Taoyuan Hsien, TW)
Application Number:
Publication Date:
06/10/1997
Filing Date:
03/30/1995
Export Citation:
Saturn Cosmos Co., Ltd. (TW)
Chyi, Jen-inn (TW)
Primary Class:
Other Classes:
International Classes:
C30B25/10; (IPC1-7): C30B35/00
Field of Search:
117/9, 117/108, 117/200, 117/202, 117/953, 118/723VE, 118/723MR, 118/723MA, 118/723ER, 505/732
View Patent Images:
&&&&&&PDF help
US Patent References:
4950642Okamoto et al.505/7324683838Kimura et al.118/723VE4615756Tsujii et al.156/345
Foreign References:
JP3241173October, 1988118/728OXYGEN-RICH SUBSTANCEJPHA
Other References:
M.A. Herman, H. Sitter, "Molecular Beam Epitaxy: Fundamentals and Current Status," Springer-Verlag, pp. 29-42, 1989.
Shuji Makamura et al., "Candela-class high-brightness InGaN/AIGaN double-heterostructure blue-light-emitting diodes," Appl. Phys. Lett. 64(13), 28 Mar. 1994, pp. .
Shuji Nakamura, et al., "Highly P-Typed Mg-Doped GaN Films Grown with GaN Buffer Layers," Jpn. J. Appl. Phys., vol. 30, No. 10A, Oct., 1991, pp. L 1708-L 1711.
Hiroshi Amano, et al., "P-Type Conduction in Mg-Doped GaN Treated with Low-Energy Electron Beam Irradiation (LEEBI)", Jpn. J. Appl. Phys., vol. 28, No. 12, Dec. 1989, pp. L 2112 -L 2114.
W.E. Hoke, et al., "Evaluation of a new plasma source for molecular beam epitaxial growth of InN and GaN films," Journal of Crystal Growth 111 (-1028.
Shuji Nakamura, et al., "Hole Compensation Mechanism of P-Type GaN Films," Jpn. J. Appl. Phys., vol. 31 (1992), pp. .
"Superconducting Films of YBa.sub.2 Cu.sub.3 Ox and Bi-Sr-Ca-Cu-O Fabricated by Electron-Beam Deposition with a Single S" Terasaki, Japanese Journal of Applied P vol. 27, No. 8, Aug., 1988, pp. L.
Primary Examiner:
Nguyen, Nam
Assistant Examiner:
Garrett, Felisa
What is claimed is:
1. An apparatus for growing nitride based compound semiconductor epitaxial layers on at said apparatus comprising an ultra-high vacuum chamber including: a substrate manipulator for holding said at least one substrate, an atom source facing said substrate manipulator, said atom source being separated from said surface of said substrate manipulator at a distance of less than 15 cm for generating a flux of nitrogen atoms towards said at least one substrate, and at least one gas injector located between said substrate manipulator and said atom source for introducing other reactant and dopant gases into said chamber towards said at least one substrate.
2. The apparatus of claim 1 wherein said substrate manipulator, said at least one gas injector, and said atom source are arranged such that said substrate manipulator faces upward.
3. The apparatus of claim 1 wherein said substrate manipulator is supplied with at least one of a DC or a RF bias voltage to neutralize ions and reduce damage to the epitaxial layers caused by high energy nitrogen ions.
4. The apparatus of claim 1 wherein said at least one gas injector comprises at least two gas injectors, each for introducing respective gases to prevent cross contamination between said respective gases.
5. The apparatus of claim 1 wherein said at least one gas injector has the shape of a ring.
6. The apparatus of claim 1 wherein said at least one gas injector has the shape of a pair of semicircles.
7. The apparatus of claim 1 wherein said at least one gas injector has the shape of four quadrants.
8. The apparatus of claim 1, wherein said atom source is separated from said substrate manipulator at a distance of about 10 cm.
9. The apparatus of claim 1 wherein said nitride based compound semiconductor epitaxial layers are GaN based.
10. The apparatus of claim 1 wherein said atom source includes an aperture.
11. The apparatus of claim 10 wherein said aperture has a diameter of about 10 cm.
12. The apparatus of claim 1 wherein said atom source is arranged coaxial with said substrate manipulator.
13. The apparatus of claim 1 wherein said atom source generates said flux of nitrogen atoms within a half angle of about 15°.
14. The apparatus of claim 1 wherein said atom source is a RF excited plasma source.
15. The apparatus of claim 1 wherein said atom source and said at least one gas injector are arranged such that increasing said flux of nitrogen atoms does not decrease the mean free path of said reactant and dopant gases.
16. The apparatus of claim 1 wherein said substrate manipulator includes a heat source for heating said at least one substrate.
17. The apparatus of claim 1 wherein said at least one gas injector is coaxial with said substrate manipulator.
18. The apparatus of claim 1 wherein said at least one gas injector is located about 7 cm from said substrate manipulator.
19. The apparatus of claim 1 wherein said at least one gas injector includes a plurality of openings for introducing said reactant and dopant gases.
20. The apparatus of claim 1 wherein said at least one gas injector has a diameter of about 13 cm.
21. The apparatus of claim 1 wherein said atom source generates nitrogen atoms having an energy of less than 1 eV.
22. The apparatus of claim 1 wherein said atom source is arranged parallel to the surface of said substrate manipulator.
23. The apparatus of claim 1 wherein said substrate manipulator, said at least one gas injector and said atom source are arranged such that said substrate manipulator faces downward.
Description:
BACKGROUND OF THE INVENTIONThe present invention relates to a method for the growth of nitride based semiconductors such as GaN, InN, AlN and other alloys. The invention relates also to an apparatus for use with such a method. GaN and its related compound semiconductors are the key materials for blue-green light emitting diodes (LEDs) and semiconductor lasers. The preparation of high quality epitaxial layer of this material has been intensively pursued for over twenty years (see J. Pankove, U.S. Pat. No. 3,864,592; H. Kobayashi et. al., U.S. Pat. No. 4,473,938; K. Manabe et. al., U.S. Pat. No. 4,911,102). The main obstacle to the achievement of high efficiency light emitting diode is the preparation of highly conductive p-type GaN. Low energy electron beam irradiation (see H. Amano et. al., "P-type conduction in Mg-doped GaN treated with low-energy electron beam irradiation (LEEBI)," Jpn. J. Appl. Phs. 28, L, 1989) and thermal annealing (see S. Nakamura et. al., "Hole compensation mechanism of p-type GaB films," Jpn. J. Appl. Phys. 31, , 1992) have been used to activate the Mg-doped GaN epilayers. Dehydrogenation was proposed to be the critical step since highly conductive p-type GaN films could be grown in a hydrogen-free ambient (see M. E. Lin st. al., Appl. Phys. Lett. 63, 932-934, 1993; C. Wang and R. F. Davis, Appl. Phys. Lett. 63, 990-992, 1993; M. Rubin et. al., Appl. Phys. Lett. 64, 64, 1994). This invention presents the design of a chemical beam epitaxy system and the method for the epitaxial growth of large area epitaxial layers of GaN and its related compounds. Conventionally, there are two most effective methods to prepare GaN epilayers, namely molecular beam epitaxy (MBE) and organometallic vapor phase epitaxy (MOVPE). Because molecular nitrogen is too inert to react with Ga, radio-frequency plasma-assisted growth methods (see W. E. Hoke, P. J. Lemonias and D. G. Weir, "Evaluation of a new plasma source for molecular beam spitaxial growth of InN and Gan films", J. Crystal Growth III, pp. , 1991 and M. Liu, A. C. Frenfel, J. G. Kim, and R. M. Park, "Growth of zinc blende-GaN on β-SiC coated (001) Si by molecular bem epitaxy using a radio frequency plasma discharge, nitrogen free-radical source", J. Appl, Phys, 74, pp. , 1993) and microwave plasma-assisted growth methods (see C. H. Carter, Jr., U.S. Pat. No. 5,210,051, May 1993 have been most widely used for MBE while NH3 is used for MOVPE (K. Manade et al., U.S. Pat. No. 4,911,102, March 1990). The group III sources used are elemental metals evaporated from effusion cells for MBE and vapors from metal-organic compounds for MOVPE, respectively. Using a plasma-assisted growth method, MBE shows only limited success. Though the material obtained by MBE is of reasonably high quality, the growth rate is less than 0.6 um/h. This slow growth rate is attributed to the limited reactive nitrogen atom or ion flux provided by the plasma source for the growth of stoiohiometric films. The usable flux in a MBE system is determined by the efficiency of the nitrogen source and the distance between the substrate and the source. Because of the divergent nature of the nitrogen beam, the usable flux decreases dramatically as the substrate is located far away from the nitrogen source. Further, increasing the nitrogen rate into the plasma source raises the growth pressure above 10-4 Torr and reduces the mean free path of the reactants, which deteriorates the growth rate. Modification made on the conventional MBE chamber to accommodate the plasma source can not avoid the difficulties mentioned above which shorter distance between the substrate and the plasma source will either affect the uniformity of the epilayer or damages the epilayer when an electron cyclotron resonance (ECR) source is used. Currently, the only method that is employed for the production of GaN LEDs is MOVPE. Because high quality GaN can only be grown at 1000°-1100° C., thermal convection and gas phase pre-reaction have impeded the success of conventional MOVPE method. A two-flow reaction chamber has been proposed (see S. Nakamura, Jpn. J. Appl. Phys. 30, L, 1991) to overcome these barriers and produce high brightness blue LEDs. However, this method has disadvantages. Though the thermal convection and gas phase pre-reaction can be somewhat suppressed by reducing the pressure in the reactor, a large amount of pressing gas is still necessary for the growth of GaN. Morever, the nitrogen source, i.e. NH3, used in this method produces a great amount of hydrogen which is detrimental to the p-type GaN. Therefore, post-annealing on the MOVPE grown epilayers above 700° C. in a nitrogen ambient to activate p-type dopant Mg is necessary. Moreover, the manner of the introduction of the reactant gas to the substrate hinders uniform growth of epilayers over a large area, e.g. greater than two inch diameter. SUMMARY OF THE INVENTION Compared to the conventional plasma-assisted MBE, this invention will:
a. enhance the growth rate to a comparable level as that of MOVPE method because the nitrogen atom source is placed at about 10 cm from the substrate to obtain higher flux rather than the 15 cm in the conventional MBE case.
b. eliminate the damage and etching effects caused by nitrogen ions because low energy (estimated to be less than 1 eV) excited and ground state nitrogen atoms are produced by an RF plasma source.
c. enable the epitaxial growth on large substrate because large diameter atom source (6 inch in diameter) is commercially available.
d. reduce the size of the growth chamber required for growing substrates of comparable size because conventional Ga effusion cell needs longer distance to achieve high uniformity.
e. reduce the surface oval defects caused by Ga effusion cell. Compared to the two-flow MOVPE, this invention will:
a. minimize the content of hydrogen during the growth because no H2 and NH3 are used in this process. Therefore post-annealing may not be necessary.
b. eliminate the use of large amounts of pressing and reactant gases because there is much less thermal convection and pre-reaction.
c. Allow a lower growth temperature (≤800° C. compared to .about.1050° C.) because highly reactive nitrogen atom is used.
d. enable the epitaxial growth on large substrate because no complicated gas flow is involved in this apparatus.
e. enable effective surface cleaning at low temperature using hydrogen atoms generated by the plasma source. Using the apparatus and the process invented and described herein, the following effects, which leads to higher productivity and yield, can be obtained:
(1). A high growth rate (.about.1 um/h) can be obtained because high nitrogen atom flux is used.
(2). Large area epitaxy is possible because no complicated gas flow pattern is involved, large diameter atom source is used, and the reactant gases are injected from a symmetric and proper position.
(3). Lower growth temperature (≤800° C.) is sufficient because very reactive reactants are used in an ultra-high vacuum environment (&50 sccm in total).
(4). A small amount of gas is needed because MBE method is used and no pressing gas is needed.
(5). Post-annealing may be omitted because hydrogen content is greatly reduced.
BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram depicting a chemical beam epitaxy apparatus for the growth of GaN based semiconductors according to t FIG. 2 shows another configuration of the apparatus according to t FIGS. 3(a) to 3(c) shows possible configurations of the gas injector according to t and FIG. 4 is a schematic diagram of a GaN-based heterostructure.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Referring to FIGS. 1 and 2, an apparatus for the growth of nitride based semiconductors in accordance with the present invention comprises Ultra-high vacuum growth chamber 1, substrate manipulator 5, nitrogen atom source 6, and gas injector 13. Ultra-high vacuum growth chamber 1, same as conventional MBE chamber, is made of stainless steel and has a cryo-shroud inside. This chamber is evacuated by either oil diffusion pump or turbo-molecular pump to reach ultra-high vacuum level. Ports of various sizes are present for a vacuum gauge 2, residual gas analyzer 3, reflection high energy electron diffraction (RHEED) 4, substrate manipulator 5, atom source 6, valve, shutter mechanism feedthrough, 7 gas line feedthrough 8, view port, and other in situ analysis instruments. Substrate manipulator 5 can accommodate single wafer or multi-wafer substrate holder for production. Substrates can be heated up to 1000° C. as well as rotated to improve temperature uniformity across the substrate holder. Both DC and RF bias can be applied to the substrate via the feedthrough 9 on the manipulator flange. Nitrogen atom source 6, equipped with a gas inlet 10 and an RF 11 feedthrough, is arranged in the same axis as the manipulator and is parallel to the substrate. This RF excited plasma source dissociates high purity N2 into nitrogen atoms at excited and ground states. Since these atoms diverges with a half angle of about 15°, it is necessary to keep this atom source as close to the substrate as possible to minimize the loss of atom flux. The energy of these atoms is on the order of thermal energy (.about.300° C.), therefore, placing this source within 10 cm to the substrate will not result in any damage to the epilayer. This position can be adjusted externally by mounting this device on a bellow 12. The RF power and N2 flow rate can also be adjusted to obtain the desired growth conditions. Other nitrogen atom sources, e.g. nitrogen thermal cracker, can also be used. Gas injector 13 is used to inject the reactant and dopant gases except nitrogen. These gases are organometallic compounds and hydride gases, such as trimethylgallium (TMG), trimethylaluminum (TMA), trimethylindium (TMI), diethylzinc (DEZ), bis-cyclopentadiethylmagnesium (Cp2 Mg), and silane (SiH4). These gases are well mixed in the gas tube before they are injected out of the injector. The gas injector is made of electropolished stainless steel with many openings on it. The size and the position of the openings is arranged to distribute gases to achieve good uniformity and reasonable growth rate across the whole wafer. This injector can be a circular ring in shape, shown in FIG. 3(a) or split into two semicircles, shown in FIG. 3(b), or four quadrants, shown in FIG. 3(c), in order to improve the uniformity of the gas distribution. This device is also arranged in the same axis as the manipulator while situated in a position between the atom source and the substrate. This position and the diameter of this injection can be adjusted to achieve the desired uniformity of the films. The operation of the aforesaid apparatus is described as follows. A substrate, e.g., sapphire, is transferred onto the substrate manipulator and cleaned by high temperature heating or/and hydrogen plasma. The substrate is rotated to improve temperature uniformity and flux uniformity. With the main shutter opened, nitrogen atoms are first generated by the atom source and introduced to the substrate for the nitridation process. The growth of GaN and related compounds is commenced by introducing the desired organicmetallic source through the gas injector after the nitridation. The flow rate of each gas is controlled by mass flow controller or pressure regulation. The substrate temperature can be varied for different materials in order to achieve the optimum growth condition for each material. If necessary, a DC, periodic pulse train and/or RF power can be applied to the substrate. When the growth is finished, all the gases except nitrogen are stopped. The main shutter is kept open until the substrate temperature is below about 400° C. EXAMPLES (1). The construction of a plasma-assisted chemical beam epitaxy system as shown in FIG. 1 is described in the following. In a cylindrical stainless steel ultra-high vacuum chamber evacuated by a high capacity turbomolecular pump, a substrate manipulator which can accommodate a 5 inch platen is installed on the top of the chamber. There can be three 2 inch or one 4 inch substrate on the platen. The manipulator is able to rotate the substrate with a speed of 30 rotations per minute, heat the substrate up to 1000° C., and bias the substrate up to 1000 V. The substrate temperature is measured using a set of thermalcouples located at the back side of the platen. Opposite to the substrate manipulator in a distance of about 10 cm is a nitrogen plasma source which has an aperture of 10 cm in diameter. Utilizing RF plasma, about 30% of the nitrogen molecules fed into the source are decomposed into nitrogen atoms in the excited and ground states. These low energy atoms come out of the aperture of the atom source within a half angle of about 15°, therefore the whole platen on the substrate manipulator is under a flux of nitrogen atoms. Meanwhile, the group III reactant gases and dopant gases are introduced through a circular gas injector with a diameter of about 13 cm. To achieve a uniform growth across the five inch platen, the gas injector is concentrically located at about 7 cm from the substrate and has many openings evenly distributed toward the substrate. The flow rate of each gas is controlled by upstream pressure or mass flow controller. An integrated gas manifold with vent-run valves for each gas is mounted on the growth chamber close to the gas injector. GaN-based compound semiconductors can thus be grown in this chamber by introducing the reactive nitrogen atoms and group III metal-organic vapors onto the heated substrate. (2). A process of producing a GaN based heterostructure as shown in FIG. 4 is described in the following referring to the apparatus shown in FIG. 1. The single crystal (0001) sapphire substrate which has been cleaned by chemical treatment, is first placed on the platen with its front side facing down and loaded into a load-lock vacuum chamber. After degassing the substrate under a vacuum level of 10-9 Torr, the substrate is transferred onto the substrate manipulator in the growth chamber. Then, the substrate is subjected to another heat cleaning or hydrogen plasma cleaning process which is monitored by observing the RHEED pattern. After cleaning, the substrate is exposed to a flux of nitrogen atoms for 20 min at 500° C. to form a thin layer of AlN on the surface of the sapphire substrate. The epitaxial growth is initiated at 500° C. for the initial 25 nm-thick GaN buffer layer. The flow rate of N2 and TMG is 10 and 0.5 sccm, respectively. A second layer comprising n-type GaN layer is grown at 700° C. with TMG, SiH4 and N2 flowing into the growth chamber. A layer of n-type AlGaN, is grown by introducing a TMA source into the gas injector in addition to TMG, SiH4, and N2 gases. After growing this layer, the substrate temperature is lowered to 600° C. while only N2 is left on. As the temperature is stabilized, TMG, TMI and DEZ are introduced to the substrate to grow a Zn-doped InGaN layer. Then, the substrate temperature is raised again to grow a layer of p-type AlGaN using Cp2 Mg. Finally, the TMA flow is stopped for the growth of the p-type GaN layer.
& 2004-. All rights reserved.