NITROGEN-ENABLED HIGH GROWTH RATES IN HYDRIDE VAPOR PHASE EPITAXY

Abstract

Disclosed herein are methods and devices that use a low-thermal conductivity inert gas to shield reactant gases and thus enabling a cold wall operation within a hot wall system.

Claims

1. A reactor capable of the deposition of at least one layer of a semiconductor device by using hydride vapor phase epitaxy (HVPE) using at least one group III metal source, at least one group V hydride gas, and an inert carrier gas, wherein the reactor comprises at least one source zone region at a temperature of up to about 800 C., and at least one deposition zone region at a temperature of below about 700 C.; and a delivery means to the deposition zone region wherein the delivery means comprises the inert gas carrier and reactant gases comprising a group III metal and at least one group V hydride gas that are flowed towards a growth substrate at a volumetric flow rate.

2. The reactor of claim 1 wherein the deposition zone region is below about 650 C.

3. The reactor of claim 1 wherein the at least one layer of the semiconductor device is GaAs.

4. The reactor of claim 3 wherein the GaAs deposition is up to 528 m/h.

5. The reactor of claim 1 wherein the rate of deposition of the at least one layer of a semiconductor device is improved by up to 32 percent when compared to using H.sub.2 as a carrier gas in the reactor instead of the inert carrier gas.

6. The reactor of claim 1 wherein the volumetric flow rate for the inert carrier gas and group V hydride gas is up to 5000 sccm.

7. The reactor of claim 1 wherein the volumetric flow rate for the group V hydride gas is up to 100 sccm.

8. The reactor of claim 1 wherein the partial pressure for the group V hydride gas within the inert gas carrier and reactant gases is up to 0.01 atm.

9. The reactor of claim 1 wherein the partial pressure for the gas comprising the group III metal within the inert gas carrier and reactant gases is up to 0.0012 atm.

10. A method for growing at least one layer of a semiconductor device using a reactor comprising at least one source zone and at least one deposition zone wherein the method comprises hydride vapor phase epitaxy (HVPE) using at least one group III metal, at least one group V hydride gas, and an inert carrier gas.

11. The method of claim 10 wherein the at least one source zone is at a temperature of up to about 800 C., and wherein the at least one deposition zone is at a temperature of up to about 700 C.

12. The method of claim 10 wherein the at least one layer of the semiconductor device is GaAs.

13. The method of claim 12 wherein the GaAs growth rate is up to 528 m/h.

14. The method of claim 10 wherein the growth rate of the at least one layer of a semiconductor device is improved by up to 32 percent when compared to using H.sub.2 as a carrier gas in the reactor instead of the inert gas.

15. The method of claim 10 wherein a volumetric flow rate of the inert carrier gas and group V hydride gas is up to 5000 sccm.

16. The method of claim 10 wherein a volumetric flow rate for the group V hydride gas is up to 100 sccm.

17. The method of claim 10 wherein a partial pressure for the group V hydride gas within the inert carrier gas is up to 0.01 atm.

18. The method of claim 12 wherein a surface roughness of the GaAs layer is below 0.8 nm.

19. The method of claim 12 wherein an electron mobility of the GaAs layer is up to 1600 cm.sup.2/V-s.

20. The method of claim 12 wherein an electron concentration of the GaAs layer is up to 9.910.sup.18 cm.sup.3.

Description

DESCRIPTION OF THE DRAWINGS

[0008] Exemplary embodiments are illustrated in referenced figures of the drawings. It is intended that the embodiments and figures disclosed herein are to be considered illustrative rather than limiting.

[0009] FIG. 1 depicts an embodiment of a portion of an HVPE reactor that uses a low-thermal conductivity inert gas. FIG. 1 depicts the mole fraction of As.sub.4 within an HVPE reactor when using hydrogen versus when using a low-thermal conductivity inert gas.

[0010] FIG. 2 depicts an embodiment of a portion of a HVPE reactor that uses a low-thermal conductivity inert gas. FIG. 2 depicts the mole fraction of AsH.sub.3 within a portion of a HVPE reactor when using hydrogen versus when using a low-thermal conductivity inert gas. As depicted in FIG. 2, the concentration of AsH.sub.3 at a reaction surface in the HVPE reactor using a low-thermal conductivity inert gas is significantly higher than the same reactor using hydrogen.

[0011] FIG. 3 depicts an embodiment of a portion of an HVPE reactor that uses a low-thermal conductivity inert gas. As depicted in FIG. 3 the temperature is significantly lower at a reaction surface of the HVPE reactor when using a low-thermal conductivity inert gas when compared to using hydrogen.

[0012] FIG. 4 depicts growth rate as a function of GaCl partial pressure where the AsH.sub.3 carrier gas flow in all cases was 5000 sccm. The black squares indicate data using a N.sub.2 carrier gas and the red circles indicate data using H.sub.2 carrier gas.

[0013] FIG. 5 depicts GaAs growth rate vs. AsH.sub.3 carrier gas flow for either H.sub.2 or N.sub.2 carrier gas.

[0014] FIG. 6 depicts computational fluid dynamics (CFD) modeling of the temperature distribution of the growth chamber using H.sub.2 (left) and N.sub.2 (right), where the lower thermal conductivity gas insulates the AsH.sub.3 carrier gas flow in the center from the hot reactor walls, thus keeping the AsH.sub.3 at a lower temperature and minimizing the likelihood of thermal decomposition.

[0015] FIGS. 7a, 7b and 7c depict CFD modeling of As.sub.4 partial pressure inside the growth chamber. FIG. 7a depicts CFD modeling of As.sub.4 partial pressure inside the growth chamber where more As.sub.4 is generated with H.sub.2 carrier gas. FIG. 7b depicts AsH.sub.3 partial pressure inside the growth chamber showing that more AsH.sub.3 is delivered to the substrate with N.sub.2. FIG. 7c depicts a cross-sectional view of the plane of the substrate, showing the partial pressure of each species at the substrate position.

DETAILED DESCRIPTION

[0016] Disclosed herein are methods that use nitrogen as a carrier gas in HVPE. By using nitrogen as a carrier gas in HVPE, record fast growth rates reaching 432 microns/hour (about 40% faster than the current record using hydrogen as a carrier gas) have been observed.

[0017] The two largest cost barriers for III-V photovoltaics are those of both the growth and the substrate. Therefore, enhancing the growth rate of III-V semiconductors would increase the throughput of these devices, which decreases the total costs associated with growth with a simple modification to reactant gases. Additionally, use of a nitrogen ambient is potentially necessary to enable the nucleation and growth of GaAs epilayers on a monolayer-graphene coated GaAs substrate, because H.sub.2 is known to etch graphene at typical reactor temperatures. Remote-epitaxy through graphene and liftoff is a technology that is useful as a route to reduce substrate costs.

[0018] In an embodiment, disclosed herein are record GaAs growth rates approaching 530 m/h using nitrogen carrier gas and 400 m/h using hydrogen carrier gas in a dynamic-hydride vapor phase epitaxy reactor. Computational fluid dynamics modeling was performed to study the effect that either nitrogen or hydrogen carrier gas has on heat and mass transfer given the potential AsH.sub.3 decomposition at surfaces. As disclosed herein through experiments and support with modeling, nitrogen carrier gas enhances the growth rate compared to hydrogen carrier gas in a hydride-enhanced regime by minimizing the amount of arsine that decomposes before reaching the wafer surface.

[0019] Previous work using HVPE has demonstrated that using nitrogen as a carrier gas in HVPE slows the growth rate of III-V semiconductors. However, as disclosed herein, experiments using a dynamic-HVPE tool have resulted in a growth rate increase from 20%-100% when using nitrogen as a carrier gas instead of hydrogen, which indicates an exciting discovery contrary to conventional wisdom. Without being limited by theory, a hydride-enhanced growth mechanism does not require H.sub.2 to participate in the reaction, and H.sub.2 is potentially a product of the reaction AsH.sub.3+GaCl->GaAs+H.sub.2+HCl. This means that there are potential advantages to using an inert and/or higher density carrier gas, which thus makes the use of N.sub.2 enabling. A final advantage is that a nitrogen environment is non-flammable, which increases the safety of this growth technique compared to growth in a hydrogen ambient.

[0020] Previously, others in the art have stated that since argon and nitrogen are gases with heavier atomic weight, that the growth rate would be lower than that of hydrogen at the same temperature. However, as disclosed herein, the heavier gas can also affect the density and potentially thin the boundary layer thickness to enhance the growth rate. Without being limited by theory, hydrogen may be both a carrier gas and reactant that participates in the reaction to enhance the growth rate. Hydrogen may play a role in both the forward and reverse reactions.

[0021] Without being bound by theory, since the partial pressure of GaCl increases in nitrogen at the expense of HCl, solid GaAs tends to vaporize into the gas phase with the increase of the mole fraction of nitrogen and hence a reduction in growth rate is observed. However, the effect of partial pressure may be offset by the reduction in boundary layer thickness or difference in thermal conductivity between hydrogen and nitrogen.

[0022] Disclosed herein are methods and devices that use a low-thermal conductivity inert gas to shield the reactant gases and hence create a regime of cold wall operation within a hot wall system. In contrast to using a low thermal conductivity carrier gas that lowers the temperature of all of the reactant species, using methods and devices disclosed herein protects the stream of hot gases from the temperature of the reactor walls.

[0023] In another embodiment, disclosed herein is a method for the insulation of reactant gases by a low-thermal conductivity inert gas for hot wall vapor phase reactors. As disclosed herein, a heavy, inert gas with low thermal conductivity (such as nitrogen, argon, krypton, xenon, etc.) is flowed through an outer tube such that it surrounds reactant gases flowed through a central port towards a substrate in a hot wall vapor growth reactor. The reactant gases are insulated from the high temperatures of the reactor walls by the heavy outer gas, preventing their decomposition.

[0024] Growth in a hydrogen ambient at standard growth temperatures can either lead to arsine decomposition or, if gas velocity is increased to transport the reactants to the surface more quickly, reduce uniformity. Arsine can decompose into arsenic dimers and tetramers in a hot wall growth reactor at standard growth temperatures. However, it is advantageous to transport undecomposed arsine to the surface. High gas velocity can be used to transport arsine to the surface more quickly, in a hydride-enhanced growth mechanism. However, a high flow rate can lead to poor growth uniformity across a substrate, which impairs yield. Using an inert gas with low thermal conductivity flowed through an outer tube in order to encapsulate the arsine as it flows to the surface prevents decomposition without requiring exceedingly high reactant velocities or low pressures. A low-conductivity inert gas that blankets the reactant gases also eliminates the flammability risk of having a hydrogen ambient in a hot furnace.

[0025] As disclosed herein HVPE has demonstrated a hydride-enhanced regime in which a significant amount of uncracked AsH.sub.3 reaches the wafer surface, resulting in much higher growth rates due to the lower kinetic barrier to growth. In both the As.sub.x and the hydride-enhanced regime, the rate-limiting step to GaAs growth is the reduction of GaCl complexes that form on the surface. In the As.sub.x regime, GaCl can be removed either through reduction assisted by the H.sub.2 carrier gas to form HCl as a byproduct, or by a slower mechanism where three GaCl are converted to volatile GaCl.sub.3. Slower GaAs growth rates were historically observed with an inert carrier gas due to the slower reaction mechanism in the absence of H.sub.2. In contrast, the hydride-enhanced regime enables the direct reaction of AsH.sub.3 with GaCl to form GaAs without any influence of any carrier gas. The effect of inert carrier gas on growth rate has not been studied in the hydride-enhanced regime.

[0026] GaAs Growth Rates of 528 m/h Using D-HVPE with a Nitrogen Carrier Gas

[0027] Disclosed herein are experiments that confirm the observation of lower growth rates in the As.sub.x regime when using N.sub.2 as a carrier gas instead of H.sub.2, yet reveal a notable increase in growth rate when using N.sub.2 in the hydride-enhanced regime. Demonstrated herein are growth rates up to 528 m/h using N.sub.2 carrier gas, compared to 400 m/h using H.sub.2 carrier gas. CFD modeling was used to investigate the effect of carrier gas on internal reactor temperature and relative group V partial pressures to identify potential mechanism for these exceptional growth rates. Using methods disclosed herein, very similar material quality was achieved regardless of the growth regime used.

[0028] In an embodiment, materials were grown in a dual-chamber D-HVPE reactor. In an embodiment, a simplified schematic of the reactor can be found in K. L. Schulte, et al., High growth rate hydride vapor phase epitaxy at low temperature through use of uncracked hydrides, Applied Physics Letters, vol. 112, p. 042101, 2018; and U.S. Pat. No. 9,824,890 by D. L. Young, et al., entitled High throughput semiconductor deposition system. In an embodiment, GaCl and InCl were formed in-situ by the reaction of anhydrous HCl over elemental Ga and In for the group III sources. Hydride gases were used for the group V source, and n-type doping was achieved using hydrogen selenide (H.sub.2Se). The gas plumbing design is such that the carrier gas species is either H.sub.2 or N.sub.2, but not a combination of both. Changes in the AsH.sub.3 carrier flow (Q.sub.N.sub.2.sup.AsH.sup.3 or Q.sub.H.sub.2.sup.AsH.sup.3) were offset by changes to carrier gas flows through other parts of the system to maintain constant partial pressures of all reactant species. The reactor is heated in four independently controlled temperature zones: the two source zones where the metal chlorides are formed were held at 800 C., while the two deposition zones where the substrate is located were both maintained at 650 C. In an embodiment, the substrates used were (100) GaAs substrates miscut 6 towards the (111)A plane. Growth rates were determined by either a contact profilometer to measure the height difference between the GaAs epilayers and a GaInP etch stop, or by cross-sectional scanning electron microscopy (SEM) of a doped GaAs epilayer grown directly on an undoped GaAs substrate. Carrier concentrations and mobilities were determined using van der Pauw Hall-effect measurements, where four contacts were soldered near each corner of a 1 cm0.5 cm sample using In wire.

[0029] The modeling was performed with the commercially available computational fluid dynamics package CFD-ACE+, which allows the user to build and define the geometry in which reactions take place, set initial boundary conditions, and detail the composition and flow of the desired reactants. In an embodiment of the modeling experiments, the temperature was fixed at the reactor walls in the simulation to values of 800 C. at the source zone and 650 C. in the deposition zone, as it is in the real reactor. The effective rate constant for the decomposition of arsine at those temperatures were manually input into CFD-ACE+ from values previously reported in order to take into account AsH.sub.3 cracking into As.sub.4. The model assumes that any AsH.sub.3 that reaches the surface will react with available GaCl to deposit GaAs, an assumption which seems to align well with experimental results. The 3D geometry used for these simulations was a single growth chamber, where carrier gas flows through either a center tube as the carrier for AsH.sub.3 or through an inlet that approximates the flows through the actual top port and source boats.

[0030] It has been demonstrated that a higher Q.sub.H.sub.2.sup.AsH.sup.3 unlocks the hydride-enhanced regime by decreasing the amount of time that AsH.sub.3 spends at high temperature as it is injected into the reactor, which minimizes potential decomposition into As.sub.x. AsH.sub.3 begins to decompose at temperatures above 300 C., and can completely decompose at temperatures exceeding 500 C. Minimizing AsH.sub.3 decomposition, therefore, depends strongly on temperature, dwell time, and the surrounding surfaces of the substrate and walls of the reactor. GaAs growth rates were observed with N.sub.2 than with H.sub.2 while using these hydride-enhanced conditions, in contrast to other work using As.sub.x as the group V source. FIG. 4 shows the GaAs growth rate as a function of GaCl partial pressure (P.sub.GaCl) for both types of carrier gas, using conditions that are not limited by the availability of AsH.sub.3. The growth rates using both H.sub.2 and N.sub.2 carriers linearly increase with increasing GaCl partial pressure until the highest P.sub.GaCl tested, resulting in GaAs growth rates up to 528 m/h using N.sub.2 carrier gas and 400 m/h using H.sub.2 carrier gas and otherwise using the equivalent growth conditions, listed in FIG. 4.

[0031] The data shown in FIG. 4 are at the mass-flow controller maximum of 5000 standard cubic centimeters per minute (sccm) for either Q.sub.N.sub.2.sup.AsH.sup.3 or Q.sub.H.sub.2.sup.AsH.sup.3. It has been previously shown that maintaining a high AsH.sub.3 carrier gas flow is critical to unlocking high growth rates since the high resultant gas velocity minimizes cracking of AsH.sub.3 into As.sub.x species in our reactor. FIG. 5 shows how changing Q.sub.N.sub.2.sup.AsH.sup.3 and Q.sub.H.sub.2.sup.AsH.sup.3 affects the growth rate. In this experiment, increased Q.sub.N.sub.2.sup.AsH.sup.3 and Q.sub.H.sub.2.sup.AsH.sup.3 was offset by decreased carrier gas flow of an equal amount through other parts of the system to maintain constant reactant dilution for each data point. It was discovered that at a relatively low 500 sccm for Q.sub.N.sub.2.sup.AsH.sup.3 and Q.sub.H.sub.2.sup.AsH.sup.3, which is correlated with greater AsH.sub.3 decomposition, the growth rate with N.sub.2 is lower than H.sub.2 (1.68 m/h compared to 5.4 m/h), which is a similar percent decrease in growth rate reported for inert carrier gas at 650 C. However, as either Q.sub.N.sub.2.sup.AsH.sup.3 or Q.sub.H.sub.2.sup.AsH.sup.3 is increased, the growth rate of the samples grown with N.sub.2 becomes significantly higher than those grown with H.sub.2. This result implies that the extent of AsH.sub.3 cracking between GaAs growth using H.sub.2 or N.sub.2 is different.

[0032] CFD modeling was also used to better understand the influence of the carrier gas on the extent of AsH.sub.3 decomposition. In these simulations, carrier gas flows through either a center tube to simulate Q.sub.N.sub.2.sup.AsH.sup.3 and Q.sub.H.sub.2.sup.AsH.sup.3 or through a top inlet that approximates the flows through the rest of our reactor. The flow rates through the top inlet in CFD-ACE+ are 7500 sccm for the carrier gas and 10 sccm for GaCl, while the flows through the center tube were 65 sccm for AsH.sub.3 and 2500 sccm for the carrier gas. These flows correspond to the actual flows used for the data points in FIG. 2 at 2500 sccm for Q.sub.N.sub.2.sup.AsH.sup.3 and Q.sub.H.sub.2.sup.AsH.sup.3 FIG. 6 depicts the modeled temperature distribution inside the growth chamber for either H.sub.2 or N.sub.2. In the H.sub.2 case, the internal reactor temperature uniformly approaches the reactor wall boundary conditions of 800 C. in the source zone and 650 C. in the deposition zone, except for the lower temperature near the top inlet of the simulation. In the N.sub.2 case, the temperature towards the center of the reactor is much lower than that of the walls. The thermal conductivity at 650 C. is 0.423 W/m.Math.K for H.sub.2, compared to 0.0612 W/m.Math.K for N.sub.2. The model shows that the nearly 10 lower thermal conductivity of N.sub.2 greatly reduces heat transfer from the hot reactor walls. This effect decreases the temperature of the AsH.sub.3 that flows through the center of the reactor, subsequently decreasing the likelihood of thermal decomposition.

[0033] FIG. 7 depicts the partial pressures of As.sub.4 and AsH.sub.3 both as the species evolve down the length of the reaction chamber and as they appear at the plane of the substrate. The kinetic constants for the decomposition of AsH.sub.3 at the temperature of the two heater zones in these simulations were previously reported. FIG. 7a shows the As.sub.4 partial pressure for N.sub.2 and H.sub.2 carrier gas as it evolves through the reactor. A greater partial pressure of As.sub.4 can be observed in the H.sub.2 case, without being bound by theory, this is likely due to the higher temperature due to the higher thermal conductivity carrier gas as depicted in FIG. 6, and correspondingly increased likelihood for AsH.sub.3 to decompose into As.sub.4. FIG. 7b depicts a correspondingly low AsH.sub.3 partial pressure due to thermal decomposition, where the amount of both species at the substrate is shown in FIG. 7c. In the N.sub.2 case, AsH.sub.3 only begins to decompose near the tapering of the reactor walls, while a significant amount of AsH.sub.3 still reaches the substrate in the simulation, and a much lower partial pressure of As.sub.4 is observed. The AsH.sub.3 partial pressure is about 0.01 atm for N.sub.2 carrier gas, which is close to the AsH.sub.3 partial pressure used experimentally in FIGS. 4 and 5. These results indicate that the low thermal conductivity of N.sub.2 is able to thermally insulate AsH.sub.3 from the hot reactor walls and minimize thermal decomposition compared to H.sub.2, which is consistent with the higher growth rates with N.sub.2 disclosed herein.

[0034] FIGS. 6 and 7 together show that the temperature of the center of the reactor where AsH.sub.3 is transported is lower with N.sub.2, allowing comparatively more AsH.sub.3 to reach the wafer surface. In the hydride-enhanced regime, the mechanism for GaAs growth is

GaCl+AsH.sub.3GaAs+HCl+H.sub.2(1)

[0035] and the growth rate increases with enhanced preservation of AsH.sub.3. Without being bound by theory, because GaAs has a lower kinetic barrier to growth under AsH.sub.3 than As.sub.x, and less AsH.sub.3 decomposition occurs with the low thermal conductivity carrier gas, the resulting growth rate is higher for N.sub.2 compared to H.sub.2.

[0036] For low Q.sub.N.sub.2.sup.AsH.sup.3, significant AsH.sub.3 decomposition into As.sub.x occurs. In this case, As.sub.x is the predominant group V reactant, and

GaCl+ As.sub.4+H.sub.2GaAs+HCl(2)

[0037] is the reaction attributed to GaAs growth. In this case, H.sub.2 participates as a reactant and enables higher growth rates compared to GaAs growth with N.sub.2. This explains both the decrease in growth rate observed in FIG. 5 for decreasing Q.sub.N.sub.2.sup.AsH.sup.3, as well as the crossing of the growth rate trends at the lowest Q.sub.N.sub.2.sup.AsH.sup.3. Without the presence of H.sub.2, the growth likely proceeds according to

3GaCl+As.sub.42GaAs+GaCl.sub.3(3)

[0038] where the carrier gas plays no direct role in the reaction. For disproportionately large amounts of group III precursors, this reaction mechanism can yield higher GaAs growth rates compared to H.sub.2. However, for the GaCl partial pressure of 0.0012 atm used in these experiments, it is statistically improbable to assemble 3 GaCl molecules to create GaCl.sub.3, leading to the observed lower growth rate with N.sub.2 due to these kinetic limitations. Together, FIGS. 6 and 7 indicate that the lower thermal conductivity of N.sub.2 is responsible for the lower temperature in the center of the reactor, as well as the increased growth rates.

[0039] The average surface roughness, electron mobility, electron concentration of the samples in FIG. 4 were measured to verify the material for GaAs grown with either H.sub.2 and N.sub.2 carrier gas for a range of growth rates, displayed in Table 1. These samples were n-type GaAs grown on a nominally undoped GaAs substrate in order to ensure that any roughness could be directly attributed to the GaAs growth. The average roughness for these samples, calculated from a 1 m1 m AFM scan, reveals that the average surface roughness stays below 0.8 nm with no attempt to optimize the growth conditionssuch as V/III ratio. At the highest P.sub.GaCl tested, samples began to substantially roughen for both H.sub.2 and N.sub.2. Without being bound by theory, it is suspected that this is due to the decreasing V/III ratio as the P.sub.GaCl increases and the AsH.sub.3 flow remains constant at the mass-flow controller maximum. It is likely that the surface morphology would return to smooth if more group V precursor were available. It is also expected that these growth rates would continue to increase if the AsH.sub.3 did not become limiting, but these data represent the maximum growth rates we can achieve given our current mass-flow controller for AsH.sub.3 and AsH.sub.3 carrier flow, which precludes our ability to adjust V/III ratio as we further increase GaCl.

[0040] The electron mobility and concentration were determined using Hall-effect measurements of GaAs epilayers grown with the same H.sub.2Se flow between samples. The electron mobilities are high at all growth rates using H.sub.2 carrier gas, and only begin to degrade by 300 cm.sup.2/V-s between 483 m/h and 389 m/h with N.sub.2 carrier gas. Additionally, the value of electron mobilities are reasonably close between H.sub.2 and N.sub.2 carrier gas, which is a promising result for device quality at high growth rates with N.sub.2 since excellent device performance has been realized with H.sub.2.

TABLE-US-00001 TABLE 1 Average surface roughness, electron mobility, and electron concentration for the samples in FIG. 4 for a range of growth rates using N.sub.2 or H.sub.2 carrier gas. Average Electron Growth rate roughness Mobility concentration (m/h) Carrier gas (nm) (10.sup.3 cm.sup.2/V-s) (10.sup.18 cm.sup.3) 33 N.sub.2 0.35 1.6 2.1 213 N.sub.2 0.26 1.0 9.9 320 N.sub.2 0.36 1.2 7.6 389 N.sub.2 0.30 1.2 7.0 483 N.sub.2 0.38 0.9 4.5 528 N.sub.2 In progress 23 H.sub.2 0.78 1.4 2.5 167 H.sub.2 0.36 1.0 5.3 216 H.sub.2 0.29 1.1 7.4 264 H.sub.2 0.26 1.2 6.8 336 H.sub.2 0.37 1.3 4.7 400 H.sub.2 In progress

[0041] Using methods and devices disclosed herein, record growth rates up to 400 m/h using H.sub.2 and 528 m/h using N.sub.2 in a D-HVPE reactor have been demonstrated, with smooth surfaces and high electron mobility demonstrated up to growth rates of 336 m/h and 483 m/h for the respective carrier gases. The high throughput, lower cost, and safety benefits of the N.sub.2 compared to the incumbent H.sub.2 used in III-V As/P HVPE systems make N.sub.2 an attractive alternative for carrier gas. CFD modeling showed that enhanced AsH.sub.3 delivery to the wafer surface enabled by a lower internal reactor temperature using a N.sub.2, whereas more As.sub.x reaches the wafer surface with H.sub.2. However, for low AsH.sub.3 carrier gas flows, significant AsH.sub.3 decomposition occurs, and As.sub.x becomes the predominant group V reactant. In this case, H.sub.2 plays an enabling role in the growth mechanism, and subsequently a higher growth rate is observed with H.sub.2 instead of N.sub.2. This result indicates that it is possible to grow GaAs at high growth rates under a N.sub.2 environment which should reduce deposition cost by eliminating expenses associated with producing and handling H.sub.2 in a high temperature environment.