Systems and methods for growing a non-phase separated group-III nitride semiconductor alloy

Abstract

Systems and methods for MBE growing of group-III Nitride alloys, comprising establishing an average reaction temperature range from about 250 C to about 850 C; introducing a nitrogen flux at a nitrogen flow rate; introducing a first metal flux at a first metal flow rate; and periodically stopping and restarting the first metal flux according to a first flow duty cycle. According to another embodiment, the system comprises a nitrogen source that provides nitrogen at a nitrogen flow rate, and, a first metal source comprising a first metal effusion cell that provides a first metal at a first metal flow rate, and a first metal shutter that periodically opens and closes according to a first flow duty cycle to abate and recommence the flow of the first metal from the first metal source. Produced alloys include AlN, InN, GaN, InGaN, and AlInGaN.

Claims

1. A single-phase, group-III Nitride film comprising the formula In.sub.xM.sub.1-xN, wherein x is from about 37% to about 72%; M comprises gallium, aluminum or a mixture thereof; and the film is free of phase separation.

2. The group-III Nitride film of claim 1, wherein x is from about 45% to about 72%.

3. The group-III Nitride film of claim 1, wherein x is from about 37% to about 45%.

4. The group-III Nitride film of claim 1, wherein M is gallium.

5. The group-III Nitride film of claim 4, wherein M further comprises aluminum.

6. The group-III Nitride film of claim 1, wherein the film is free of surface defects and surface pitting.

7. The group-III Nitride film of claim 1, wherein an x-ray diffraction rocking curve of the film is less than about 416 arcseconds.

8. The group-III Nitride film of claim 1, wherein an x-ray diffraction rocking curve of the film is less than about 362 arcseconds.

9. The group-III Nitride film of claim 1, wherein a root-mean-square roughness of the film as measured by atomic force microscopy is less than about 0.76 nanometers.

10. The group-III Nitride film of claim 1, wherein a root-mean-square roughness of the film as measured by atomic force microscopy is less than about 0.53 nanometers.

11. A single phase group-III Nitride film comprising: aluminum in an amount from about 13.2% to about 24.8%; and gallium; wherein the film is free of surface defects and surface pitting, and wherein an x-ray diffraction rocking curve of the film is less than about 146 arcseconds.

12. The group-III Nitride film of claim 11, further comprising indium.

13. The group-III Nitride film of claim 12, wherein the film is free of phase separation.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIG. 1 illustrates an exemplary reaction chamber for growing an aluminum-indium-gallium-nitride alloy, according to various embodiments of the present invention.

(2) FIG. 2 illustrates an exemplary method for growing an aluminum-indium-gallium-nitride alloy, according to various embodiments of the present invention.

(3) FIG. 3 is a graph showing an exemplary modulation scheme for the growth of an aluminum-indium-gallium-nitride alloy, according to various embodiments of the present invention.

(4) FIG. 4 is a graph showing transient reflection high-energy electron diffraction intensity versus time for various gallium and indium shutter open times, according to various embodiments of the present invention.

(5) FIG. 5 is a graph showing an expanded view of the reflection high-energy electron diffraction intensity shown in FIG. 4.

(6) FIG. 6 illustrates the mechanism of indium surface segregation during the growth of an indium-gallium-nitride alloy, according to various embodiments of the present invention.

(7) FIGS. 7a-c are reflection high-energy electron diffraction intensity patterns during shutter open time, low shutter open time, and high shutter open time, according to various embodiments of the present invention.

(8) FIG. 8 illustrates the time of consumption of metal adlayer and the corresponding metal adlayer thickness versus gallium and indium shutter open time for the conditions shown in FIG. 4.

(9) FIG. 9 are X-ray diffraction scans, rocking curves, and corresponding atom force microscope morphologies of indium-gallium-nitride layers grown throughout the miscibility gap, according to various embodiments of the present invention.

(10) FIG. 10 are X-ray diffraction scans of independently grown indium-gallium-nitride layers, according to various embodiments of the present invention.

(11) FIG. 11 are X-ray diffraction scans of AlGaN alloys having varying aluminum contents, grown according to various embodiments of the present invention.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(12) Although preferred embodiments of the invention are explained in detail, it is to be understood that other embodiments are contemplated. Accordingly, it is not intended that the invention is limited in its scope to the details of construction and arrangement of components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments and of being practiced or carried out in various ways. Also, in describing the preferred embodiments, specific terminology will be resorted to for the sake of clarity.

(13) It must also be noted that, as used in the specification and the appended claims, the singular forms a, an and the include plural referents unless the context clearly dictates otherwise.

(14) Also, in describing the preferred embodiments, terminology will be resorted to for the sake of clarity. It is intended that each term contemplates its broadest meaning as understood by those skilled in the art and includes all technical equivalents which operate in a similar manner to accomplish a similar purpose.

(15) Ranges may be expressed herein as from about or approximately one particular value and/or to about or approximately another particular value. When such a range is expressed, another embodiment includes from the one particular value and/or to the other particular value.

(16) By comprising or containing or including is meant that at least the named compound, element, particle, or method step is present in the composition or article or method, but does not exclude the presence of other compounds, materials, particles, method steps, even if the other such compounds, material, particles, method steps have the same function as what is named.

(17) It is also to be understood that the mention of one or more method steps does not preclude the presence of additional method steps or intervening method steps between those steps expressly identified. Similarly, it is also to be understood that the mention of one or more components in a device or system does not preclude the presence of additional components or intervening components between those components expressly identified.

(18) In some embodiments, the present invention is directed to a relatively faster growth process when compared to conventional growth processes. In some embodiments, the present invention emphasizes the monitoring of adlayer accumulation. Various embodiments of the present invention can provide increased crystal quality and grain sizes and provide single-phase, group-III Nitride semiconductor alloy layers throughout the miscibility gap. In one exemplary embodiment, the present invention is a semiconductor growth process in which metal fluxes are periodically modulated, or shuttered, according to a duty cycle. As used herein, the process is termed Metal-Modulated Epitaxy (MME). In various embodiments, the MME process of the present invention can also use metal-rich fluxes. In still further embodiments, the processing temperature of the present invention is relatively lower when compared to conventional group-III nitride growth processes such as MOCVD and MBE. An aspect of some embodiments of the present invention is the abatement and resumption of the metal fluxes during the growth processes while maintaining the nitrogen flux constant. The flow of the metal fluxes from the metal effusion cells can be periodically stopped and restarted at certain times during adlayer accumulation in order to limit excess metal dose on the growth surface.

(19) Various embodiments of the present invention can also provide for wide substrate growth temperature ranges. In some examples, InGaN can been grown as low as 250.degree. C. and AlN can be grown as hot as 850.degree. C. The large window of substrate temperatures of the present invention provides for conditions that are cold enough to substantially avoid the intermediate regime, known to cause pitting in the growth surface. Furthermore, adlayer desorption in the intermediate regime results in an exponential metal adsorbate dependence on substrate temperature. Colder, non-intermediate regime growth thus inhibits substrate temperature related fluctuations in growth conditions.

(20) FIG. 1 is an illustration of an exemplary growth system that can be used to grow InGaN according to various embodiments of the present invention. Although the following describes the present invention in the context of growing group-III Nitride semiconductor alloy, including InGaN semiconductor alloys, it should be understood that the present invention is not limited to any specific group-III Nitride alloy such as InGaN, but rather, encompasses the growth of other group-III Nitride semiconductor alloys including, but not limited to, InN, GaN, AlN, InGaAlN, all of which are considered to be within the scope of the present invention. Further, the present invention is not limited to the use of gallium, indium, aluminum, but may also include other metals or elements with functional and/or physical properties that are analogous to the aforementioned metals, gallium, indium, aluminum.

(21) Returning to FIG. 1, a first step in an exemplary embodiment is to deposit nitride film 102 onto substrate 122. Substrate materials can include, but are not limited to, silicon, silicon carbide, sapphire, germanium, zinc oxide, lithium gallate, lithium tantalate, and lithium niobate.

(22) For example, lithium gallate can be a favorable substrate for nitride film 102 over sapphire due to lithium gallate's 1% lattice mismatch versus sapphire's 14-16% lattice mismatch. The dislocations created by this mismatch are partially mitigated by preparing GaN templates, which reduce or eliminate many defects by growing nitride film 102.

(23) However, in conventional processes, lithium gallate can decompose at the relatively higher temperatures used. In some embodiments of the present invention, because of the relatively lower temperature available for processing, lithium gallate can be used because the temperature is low enough where volatile substrates such as lithium gallate become feasible. Various embodiments of the present invention can provide for the growth of thin, low indium content InGaN quantum wells to be grown completely strained and without relaxation on the thick GaN layers. In some embodiments, nitride film 102 is grown by initiating nitrogen flux 116 from nitrogen source 110 and gallium flux 112 from gallium effusion cell 106 and periodically stopping and starting gallium flux 112 according to a flow duty cycle. In other embodiments, nitride film 102 is grown via conventional growth processes such as MOCVD and MBE. In some embodiments, it can be desirable to dope nitride film 102 with magnesium to produce a p-type (hole conducting) layer.

(24) Effusion cells 106, 108 and nitrogen source 110 provide an input of gallium flux 112, indium flux 114 and nitrogen flux 116, respectively. Once nitride film 102 is grown, metal adlayer 104 is periodically deposited with gallium and indium (and in the alternative, aluminum) via gallium effusion cell 106 and indium effusion cell 108. Gallium adlayer 104 is nitrided by nitrogen source 110. As previously discussed, it is generally preferable to grow InGaN alloys at relatively lower temperatures because of the lower energy necessary to break the InN bonds when compared to compared to the GaN bonds in the InGaN alloy. But, as less energy is supplied, the energy available to the reactant atoms to find optimal locations in the crystal lattice would be reduced. In order to achieve higher levels of indium in the InGaN alloy, conventional processes increase the metal flux to the highest levels feasible, even though increased metal flux can increase the probability of indium surface segregation, thereby reducing the indium content in the alloy. In some embodiments, it can be desirable to use different metals or infuse an additional metal to change the characteristics of the group-III Nitride semiconductor alloy. For example, an aluminum effusion cell (not illustrated) can be provided to grow aluminum-bearing semiconductor alloys such as, but not limited to, AlN, AlInGaN, and AlGaN.

(25) Various flux flow rates can be established to change the percentage indium in the InGaN alloy. In one example, the flow rate of nitrogen is 1.3 standard cubic centimeters per second and can be scaled with the deposition/growth area. In that example, to grow 22% InGaN, an indium flux of 2.3E-7 Torr beam equivalent pressure and a gallium flux of 6.0E-7 Ton beam equivalent pressure, or a ratio of indium flux to gallium flux of approximately 1:2.6, can be used. For 32% InGaN, an indium flux of 3.5E-7 Torr beam equivalent pressure and a gallium flux of 5.2E-7 Torr beam equivalent pressure, or a ratio of indium flux to gallium flux of approximately 1:1.48, can be used. For 45% InGaN, an indium flux of 4.8E-7 Torr beam equivalent pressure and a gallium flux of 4.4E-7 Torr beam equivalent pressure, or a ratio of indium flux to gallium flux of approximately 1.1:1, can be used. For 62% InGaN, an indium flux of 7.6E-7 Torr beam equivalent pressure and a gallium flux of 2.6E-7 Torr beam equivalent pressure, or a ratio of indium flux to gallium flux of approximately 2.93:1, can be used. For 72% InGaN, an indium flux of 8.3E-7 Torr beam equivalent pressure and a gallium flux of 2.2E-7 Torr beam equivalent pressure, or a ratio of indium flux to gallium flux of approximately 3.77:1, can be used. Some embodiments of the present invention provide for the ability to predict the final composition of the alloy, as the effect of desorption of the metals from the alloy is minimized or negated.

(26) To facilitate both lower processing temperature and to increase the rate at which indium is incorporated into the alloy, while reducing the probability of indium surface segregation, embodiments of the present invention shutter, or abate and resume, gallium flux 112 and indium flux 114 at periodic intervals, while maintaining nitrogen flux 116 constant. Stopping and restarting gallium flux 112 and indium flux 114 limits the excess metal adlayer 104 on the surface of nitride film 102 and provides time to allow the excess metal to be consumed within the crystalline matrix of Nitride film 102. In p-type GaN, or other group-III Nitride semiconductor alloys such as AlInGaN, a less Group-III metal-rich flux is used in order to suppress compensating N-vacancies, but these fluxes are still metal-rich enough to accumulate droplets over time. The use of effusion cell shuttering periodically switches the growth between N-rich and Group-III metal-rich conditions, preferably maintaining an optimal substitutional site for Mg while still obtaining higher crystal growth quality associated with Group-III metal-rich growth. Embodiments of the present invention can significantly enhance Mg doping, resulting in hole concentrations in the 10.sup.19 cm.sup.-3 range, which in some embodiments can be four (4) to forty (40) times higher than can be found in conventional growth methods. In some embodiments, reflection high-energy electron diffraction (RHEED) intensity analysis can be used to allow for the observation and control of fractions of adsorbed metal monolayers and can also provide a method of in situ growth rate determination.

(27) FIG. 2 is an exemplary process for forming an exemplary group-III Nitride semiconductor alloy, AlInGaN, according to various embodiments of the present invention. A substrate is loaded 200 into a processing chamber. In some embodiments, the substrate is SiC or sapphire but may also include, but is not limited to, zinc oxide, lithium gallate, lithium tantalite, silicon, germanium, and lithium niobate. Once the substrate is loaded 200 into a reaction chamber, the processing temperature of the reaction chamber is raised 202 to optimal processing temperatures. In some embodiments, the temperature can range from 250.degree. C. to 850.degree. C., and more preferably from 300.degree. C. to 800.degree. C., and most preferably from 350.degree. C. to 750.degree. C. The processing temperature can be raised in various ways including, but not limited to, heating elements placed on or near the substrate, heating elements placed within the reaction chamber, or heated gas.

(28) Once the processing temperature is raised 202 to an optimal or desired level, a GaN buffer layer is deposited 204. It should be understood that the use of a GaN buffer layer is merely exemplary and is not intended to limit the scope of the present invention to just GaN buffer layers, or, even to the requirement of a buffer layer. Step 204 can comprise the use of various metal fluxes in addition to, or in substitution of, gallium. For example, step 204 can comprise the initiation of gallium, indium, or aluminum fluxes, or various combinations thereof, to produce various layers, including, but not limited to, active layers and buffer layers. The designation of a particular metal in step 204 is merely exemplary and only intended to illustrate one exemplary embodiment of the present invention.

(29) Returning to FIG. 2, in some embodiments, the GaN buffer layer is deposited 204 by initiating a gallium flux and nitrogen flux, and thereafter, stopping and starting the gallium flux according to a flow duty cycle. In other embodiments, the GaN buffer layer is deposited via conventional deposition means such as MOCVD and MBE. Once GaN buffer layer is deposited 204 onto the substrate, growth of the aluminum-indium-gallium-nitride alloy commences by initiating 206 gallium, indium, aluminum, and nitrogen fluxes. In some embodiments, the transition from the formation of the GaN buffer layer to the AlInGaN alloy may be provided for by the initiation of the additional indium and aluminum fluxes. It should be understood that the incorporation of one or more gasses in step 206 is merely exemplary and not intended to limit the scope of the present invention to the use of the gasses designated. For example, step 204 can comprise the incorporation of one of the metal fluxes, e.g. gallium, indium, aluminum, or combinations thereof.

(30) During the growth of the AlInGaN alloy, the metal fluxes, i.e. gallium, aluminum and indium, are abated 208. If the processing is complete 210, the nitrogen flux is then abated 214 and the process is complete. If the processing is not complete, the gallium, aluminum and indium fluxes are recommenced 212 for a period of time and then abated 208 again. The metal fluxes can be stopped through the use of various means, including the use of shutters placed on the outlets of the respective effusion cells. In some embodiments, the timing of the commencement and abatement of the flow of one or more of the metal fluxes constitute the flow duty cycle of the particular metal. If the group-III Nitride semiconductor being grown comprises one or more metals, the respective metal can have their individual duty cycles. For example, if the alloy to be grown is InGaN, the indium can have a first flow duty cycle and the gallium can have a second flow duty cycle, or vice versa. The flow duty cycles for each metal can be the same or can be altered, depending on the particular characteristics of the metal to be incorporated and any desired physical properties. For example, it may be advantageous to start and stop the indium and gallium fluxes at the same times. In another example, if one metal is preferred over the other, one flow duty cycle can be longer or shorter depending on the desired incorporation. In another example, a quaternary alloy, such as AlInGaN, can have a third flow duty cycle for the aluminum that can be the same as the first and/or second flow duty cycle.

(31) The abatement 208 and restarting 212 of the gallium, aluminum and indium fluxes provides for the ability to grow the AlInGaN alloy in metal rich conditions while the shutters are open and gives the indium and/or aluminum time to incorporate into the AlInGaN alloy while the gallium, aluminum and indium fluxes are abated 208. Growth continues both in the abatement and non-abatement periods, providing a relatively high growth rate when compared to conventional growth methods. Further, because the infusion of aluminum and/or indium into the AlInGaN alloy is controlled via the use of shuttering, and the material is grown at relatively lower temperatures, various embodiments of the present invention provide for the ability to predict the indium and/or aluminum content of the AlInGaN alloy by the ratio of indium/aluminum to gallium flux. Finally, because the effects of thermal decomposition and spinodal decomposition, along with metal segregation, are reduced, the growth rate of AlInGaN alloy can be increased over conventional growth processes.

(32) Experimental Results

(33) In one experiment, a Riber 32 MBE system having standard effusion cells for indium and aluminum was used. To supply gallium, a Veeco SUMO effusion cell was used. A Veeco Unibulb plasma source was employed for nitrogen, and had a flow rate of 1.3 sccm at 350 W. Growth rates were approximately 800 nm/h for all InGaN growth reported herein. To provide a uniform starting point for all studies, Lumilog metal organic chemical vapor deposition grown GaN templates were used as substrates and were back-side sputtered with 1 .mu.m of tantalum (Ta) to provide uniform heating. Substrates were cleaned in H.sub.2SO.sub.4:H.sub.2O.sub.2 (4:1) before being loaded into the introductory chamber and outgassed at 160.degree. C. for 30 minutes. The InGaN alloys were grown between 400.degree. C. and 450.degree. C. with metal-rich fluxes. Metal fluxes used were normalized based on atomic number to account for ion gauge sensitivity differences between In and Ga and were between 0.24 and 0.3.times.10.sup.-7 Torr normalized BEP.

(34) InGaN was grown using a flow duty cycle, i.e. shutter modulation, scheme similar to that illustrated in FIG. 3 and FIG. 4. The indium and gallium shutters are opened and closed simultaneously, while nitrogen remains constant. This shutter scheme was used to study the RHEED response of In/Ga adlayers, and to achieve metal-rich growth of InGaN while preferably preventing droplet buildup. FIG. 3 is an exemplary time graph illustrating the shuttering of the metal fluxes in relation to time. As shown, the nitrogen flux is maintained constant while the gallium and indium fluxes are shuttered. Samples were grown with differing metal shutter open times and indium to gallium ratios to characterize adatom kinetics of the growing InGaN surface. A range of indium compositions throughout the miscibility gap of InNGaN were grown using conditions found during the experiment.

(35) In FIG. 4, representative RHEED transients observed during the growth of 20% InGaN for several metal shutter open times from 1.5 seconds to 3.25 seconds are shown with shaded regions indicating metal shutters open. These transient RHEED intensities were determined from one growth for consistency, with GaN interlayers grown between each InGaN layer to consume residual indium from the intentional surface segregated conditions and to provide a smooth template to facilitate comparable initial growth conditions.

(36) Several features across all shutter open times are apparent. First, a decrease of intensity on shutter open occurs, and is immediately followed by an increase in intensity. A flat, steady state intensity is also reached within a few seconds of metal shutter close which is visible across all shutter open times. Shutter-open-time-dependent features include a decreasing level of the steady-state intensity relative to a peak intensity observed during the shutter open period and an increasing width of a peak during the shutter open cycle for higher shutter open times. An expanded view of the 1.5 seconds, 2 seconds, and 2.75 seconds shutter open time RHEED transients are shown in FIG. 5 to help identify and explain specific features.

(37) Shown in FIG. 5 are expanded views of 3 transient RHEED signatures shown previously in FIG. 4. Specific features noted are colored in the shaded portion labeled OPEN and CLOSED, where the OPEN region indicates metal shutter open time. The region labeled CLOSED is a signature observed after the metal shutters are closed. Following the CLOSED region is a section of relatively flat intensity described herein as a steady state period. It can be seen from both FIGS. 4 and 5 that for a given metal shutter open time in a modulation scheme, the transient RHEED intensity, in some embodiments, begins and returns to the same intensity, regardless of truncation.

(38) Focusing on the 2 second metal shutter open time section, the white region on the left side of the OPEN region is a flat RHEED intensity reached after a previous cycle and before metal shutters are opened that will be referred to as a steady state. Upon opening the metal shutters, an oscillatory function composed of the OPEN and CLOSED regions occur. This behavior is attributed to a RHEED oscillation associated solely on the adsorption of one metal monolayer (ML), where the minimum intensity is deconstructive interference caused by an incomplete layer, and the following maximum intensity is constructive interference caused by a complete layer. The remaining decrease in intensity is a continuation of this sinusoidal oscillation as the second monolayer of metal is adsorbed, but is truncated by the closing of the metal shutters, halting the supply of metal to the surface. After the metal shutters are closed, the CLOSED region occurs, which is another sinusoid that resembles the time-opposite of the OPEN region. This signature is attributed to the consumption of the adsorbed metal layer into the film. After the green region occurs, where adsorbed metal is consumed into the film, a steady state is again reached. An illustration of this process is shown in the non-segregated column of FIG. 6, and a characteristic RHEED pattern of a sample grown under these conditions is shown in FIG. 7(a), which confirms a flat, dry surface by the bright streaks.

(39) Next, the zoomed-in view of the 1.5 seconds shutter open time RHEED transient of FIG. 4 is examined, and is shown at the top of FIG. 5. This case is also non-surface segregated, and is similar to the 2 second metal shutter open time. With the use of a smaller shutter open time, a comparable RHEED oscillation is observed during the OPEN region after opening of the metal shutters. This oscillation similarly corresponds to the adsorption of the first monolayer of excess metal. However, the metal shutters are closed before a full oscillation takes place, which indicates truncation of excess metal accumulation before a full monolayer forms. Upon closing of the metal shutters, this excess metal is consumed into the film during the CLOSED region, and a signal that resembles the time opposite of the OPEN region occurs. Although this film exhibits no form of indium surface segregation, the lack of accumulation of a full adlayer limits the beneficial effects of the metal adlayer on surface smoothness. This roughness is demonstrated by both the gradual increase in RHEED intensity over time in FIGS. 4 and 5, and the spotty RHEED pattern in FIG. 7(b). An illustration of this process is shown in the non-segregated column of FIG. 6 excluding the bottom two steps in the OPEN region.

(40) Finally, the zoomed-in view of the RHEED transient with the 2.75 second shutter open time is considered and shown in FIG. 5. Before analyzing transient RHEED intensities of this case, it can be important to determine the qualitative condition of the steady state period between modulation cycles, shown by the white region on the left. In this case, excess metal is supplied to the adlayer in a previous modulation cycle that indium surface segregation has occurred, resulting in indium atoms riding on top of the rest of the adsorbed metal. Upon consumption of the adlayer, residual indium metal remains on the surface, as indium nitride typically does not form in these experimental conditions at the substrate temperatures used for these growths. The presence of indium is supported by the diffuse RHEED pattern visible during the shutter closed portion of the modulation scheme, shown in FIG. 7(c). It should be noted that specified open times vary with targeted composition, flux flow rate, and substrate temperature, and that the illustrated shutter open times are merely exemplary.

(41) Upon opening the metal shutters in the blue region of the 2.75 second shutter open time transient in FIG. 5, the RHEED intensity quickly decreases to a minimum before steadily increasing again to a maximum. The time elapsed until reaching the minimum intensity is reduced in this case compared to the transients with smaller shutter open times in FIG. 4. Furthermore, the intensity during the steady state period is also decreased. These features are concurrent with the conclusion that indium is already present on the surface. The deconstructive interference produced by this incomplete layer results in a phase shift in the sinusoid associated with the RHEED oscillation. The behavior of the RHEED intensity after opening the metal shutters is then attributed to an oscillatory function associated with the continued accumulation of one monolayer of excess metal. Since only indium may be present as all residual gallium would have been consumed, it can be expected that this newly assembled adlayer is indium-rich, potentially forming In-rich InGaN.

(42) The RHEED intensity then continues to increase as the first ML begins to finish. However, the analysis of the RHEED transients becomes more complex than constructive/deconstructive interference considerations when gallium begins to replace indium for surface sites. As the first ML finishes and further impingent metal cannot spread laterally, a second monolayer begins to form. In this case, incoming gallium atoms will preferentially displace indium atoms in the indium-rich first monolayer. This process provides an increase in intensity after the first monolayer forms, as gallium has been shown to scatter the RHEED beam less than indium. The second monolayer is also able to hold more metal than the first monolayer, which causes a decrease in frequency of the sinusoidal RHEED oscillation. These two effects produce a tall, broad peak in the RHEED intensity just after one ML of excess metal accumulation. The substitution of gallium for indium on surface sites will also affect the composition of the growing InGaN film, producing indium-lean material. Combined with the growth of indium-rich InGaN from a mostly indium adlayer at the beginning of the modulation scheme, this surface-segregated growth can produce a multi-phase material.

(43) After the metal shutters are closed, the CLOSED region occurs, resembling a faster, time-opposite of the OPEN region. Similarly to the CLOSED regions of the smaller shutter open times, this period of the transient RHEED intensity indicates consumption of adlayer material into the film. However, in this case, indium has been displaced by gallium to the second monolayer, and the adlayer as a whole is no longer homogenous. As the adlayer is consumed, the indium-lean first monolayer incorporates first, leaving mostly indium on the surface. At the temperatures used for this growth InN would likely not form in these experimental conditions. This process leaves residual indium on the surface during the steady state period, which contributes to the diffuse RHEED pattern shown in FIG. 7(c) and a decreased RHEED intensity shown in FIGS. 4 and 5. An illustration of this process is shown in FIG. 6 under the surface segregated column.

(44) Thus, indium surface segregation is linked to adlayer accumulation on the growth surface as opposed to metal-rich growth in general. By limiting the quantity of this adlayer, surface segregation can be inhibited and metal-rich growth of InGaN can be conducted. However, the thickness of this adlayer can be limited too much, hindering the smoothing effect of a wetting layer and resulting in a rough film as shown by the spotty RHEED pattern in FIG. 7(b). It then can become important to identify the exact thickness of excess metal required to induce surface segregation, and adsorb as close to that thickness as possible during each modulation scheme.

(45) If indium surface segregation has occurred, some of the metal supplied during the metal shutter open time may not incorporate into the film. This effect can result in a smaller time required to reach a steady state after the metal shutters are closed relative to a non-segregated growth condition. This experiment is performed both to determine the exact amount of metal required to induce surface segregation, herein called the surface segregation onset dose, and to strengthen the theory suggested in the qualification section above. To extract the thickness of the consumed layer, a general MBE growth equation is considered. The relationship between growth rate, incoming metal flux, adlayer accumulation, and adlayer desorption can be described with Equation (1):
R.sub.Flux=R.sub.Growth+R.sub.Adlayer+R.sub.DesorptionEquation (1)
where R.sub.Flux is the rate of incoming flux, R.sub.Growth is the rate of film growth, R.sub.Adlayer is the rate of adlayer accumulation, and R.sub.Desorption is the rate of adatom desorption from the surface. The low temperatures used for the growth of these InGaN samples allows for the assumption that desorption of the adlayer is negligible. Furthermore, when the metal shutters are closed, the flux term can be eliminated which leaves Equation (2):
R.sub.Growth=R.sub.AdlayerEquation (2)

(46) Equation (2) has been used to determine growth rate from the consumption of the known thickness of the bilayer in GaN (2.3 ML). It can also be used in this case for a known growth rate to determine the thickness of an unknown adlayer on the surface of the growing InGaN film, by Equation (3):
R Growth=(D Adlayer t accumulation)=D Adlayer t consumption ##EQU00001##Equation (3)
where D.sub.Adlayer is the thickness of the incorporated adlayer, and t.sub.consumption is the time required to consume this adlayer. It has been calculated that this composition of .about.20% InGaN has a growth rate of approximately 800 nm/hr for a particular modulation scheme. However, this growth rate takes into account the steady-state time between modulation schemes, shown by the flat, unshaded regions in FIGS. 4 and 5 where no growth occurs. Furthermore, this measure of growth rate is dependent on each particular modulation scheme. If the steady-state period is eliminated by only considering the time where metal is present during growth, the film has been calculated to grow at approximately 900 nm/hr. This allows for a modulation-independent calculation of the thickness of the adlayer consumed into the film during the metal shutter closed period. With a known growth rate and a measurable time of adlayer consumption, the thickness of this consumed adlayer can be calculated and converted to monolayers.

(47) The CLOSED region of FIG. 5, the time to consume the metal adlayer, was monitored for two indium compositions and several metal shutter open times. The derivatives of the RHEED intensity plots were used to consistently highlight when flat, steady state was reached. The length of time of this region and shutter open times were averaged for several modulation cycles and plotted in FIG. 8. The secondary y-axis of FIG. 8 is the thickness of the metal adlayer accumulated onto the film calculated from the time elapsed to consume this adlayer.

(48) Shown in FIG. 8 are both the time of consumption and thickness of the metal adlayer for several metal shutter open times during the growth of 20% and 40% InGaN. For both indium compositions, the time to consume the metal adlayer increases linearly until a certain metal shutter open time. After this point, the time to consume the adlayer decreases to a lower value, then linearly increases at a different slope than before. This break in linearity is attributed to the onset of indium surface segregation, which causes indium to segregate on top of the gallium in the adlayer. In a segregated case, once the gallium-rich first monolayer of the adlayer is consumed, mostly indium remains on the surface and is prevented from incorporating into the film due to the substrate temperature being greater than needed for InN decomposition. This would cause the steady state regions shown in FIGS. 4 and 5 to be reached sooner since the thickness of the adlayer that can consume into the film is effectively decreased.

(49) The existence of a surface segregation onset dose (1-2 ML) less than the droplet accumulation thickness (2-2.5 ML) is significant. Because an intermediate regime does not exist at these low temperatures, unshuttered MBE growth will either produce droplets or result in dry, nitrogen rich growth. Thus, the present invention uses modulation to grow high-quality, metal-rich InGaN growth. The modulation (or shuttering) provides for the accumulation of a certain amount of excess metal for its beneficial effects while preventing indium surface segregation.

(50) The understanding and suppression of indium surface segregation has enabled higher indium content films to be grown when compared to conventional growth processes, and within the InNGaN miscibility gap. Shown in FIG. 9 are x-ray diffraction scans of InGaN samples with varying indium contents from 22% to 72% and associated rocking curve full widths at half maximum (FWHM) for the <0002> symmetric reflections. Atomic force microscopy images with 30 nm height scales are shown on the right with inset root-mean-square roughnesses. Film thicknesses were measured from the thickness fringes visible in the x-ray diffraction scan for all samples. Each of the samples shown in FIG. 9 exhibit single phase x-ray diffraction spectra with rocking curve FWHMs comparable to the GaN templates used to substrate these films (.about.350 arcseconds). Shown in FIG. 10 are x-ray diffraction scans of single-phase InGaN samples, all with thicknesses of approximately 250 nm.

(51) This process has also been shown to be successful for AlGaN alloys, which are used for ultraviolet light emitters. Shown in FIG. 11 are x-ray diffraction scans of AlGaN samples with varying indium contents from 13% to 25% and associated rocking curve full widths at half maximum (FWHM) for the <0002> symmetric reflections. Film thicknesses were measured from the thickness fringes visible in the x-ray diffraction scan for all samples. Each of the samples shown in FIG. 11 exhibit single phase x-ray diffraction spectra.

(52) Numerous characteristics and advantages have been set forth in the foregoing description, together with details of structure and function. While the invention has been disclosed in several forms, it will be apparent to those skilled in the art that many modifications, additions, and deletions, especially in matters of shape, size, and arrangement of parts, can be made therein without departing from the spirit and scope of the invention and its equivalents as set forth in the following claims. Therefore, other modifications or embodiments as may be suggested by the teachings herein are particularly reserved as they fall within the breadth and scope of the claims here appended.