STR Group

Modeling of Crystal Growth and Devices

Home About Consulting Products Learn Publications Events Distributors Contact



In this example we will show how CVDSimTM can be applied to the modeling analysis of AlN/AlGaN HVPE technology with the major focus on the content of group-III mono- and trichlorides in the vapor, rate limiting species, and relative incorporation of Al and Ga into the alloy. Original quasi-thermodynamic model was used to describe chemical reactions on the Al/Ga metal surfaces and AlN/AlGaN crystal surfaces. To verify the results, the growth system design published in [1] was used.

Figure 1 Distribution of the AlCl (left) and AlCl3 (right) molar fractions

First of all, computations showed significant difference in operation of Al and Ga metal sources in terms of mono- and trichloride formation. At the typical Al source temperature of 550 oC, yield of AlCl3 dominates over that of AlCl by the factor of about 20 and this domination sharply increases at lower temperatures, see Fig. 1. At the same time, yield of GaCl3 is by several orders of magnitude less than that of GaCl at the Ga source temperature of 850 oC and becomes distinct only at temperatures close to 600 oC.

As the model was verified against the data found in [1] and [2], it became clear that the use of pure N2 carrier gas results in a uniform 25% decrease of the AlN growth rate due to lower reactive species diffusion coefficients in N2 as compared to H2, see Fig. 2. Also, both the computed and experimental data for AlN growth rate versus NH3 flow rate exhibit a sharp common boundary between the linear and saturated regions. Thorough investigation of this behavior showed that while GaN grows from GaCl under near-equilibrium conditions, the growth of AlN from AlCl3 occurs under essentially non-equilibrium conditions.

Figure 2 Left – computed AlN molar fraction in AlGaN vs. Al content in group-III vapor for pure carrier N2 (solid line) and in 10% addition of H2 (dash line) in comparison with data of Ref. [1]. Right – computed AlN growth rate at the substrate center (solid line) and averaged over the substrate surface (dash line) vs. NH3 flow rate in comparison with data of Ref. [2].

Within the quasi-thermodynamic model, this behavior can be attributed to the fact that due to the thermodynamic properties of the species involved in reaction, quasi equilibrium pressure of either AlCl3 or NH3 will have to be close to zero, depending on their supply. This means that under typical growth conditions the crystal growth rate will linearly depend on the partial pressure of one of the species and, at the same time, will be practically independent of the other. Back to Figure 2, the linear region corresponds to the deficient NH3 while the saturated region corresponds to the deficient AlCl3. Detailed analysis is given in [3].

It is essential, however, that such behavior is not characteristic of GaN HVPE. Particular thermodynamic properties of the species involved GaN growth result in quasi-equilibrium pressures of both GaCl and NH3 far from zero, which means that GaN grows under near-equilibrium conditions and GaN growth rate depends simultaneously on the GaCl and NH3 flow rates. Moreover, GaCl growth is sensitive to all the species directly involved in the reaction of GaN growth, as it was shown theoretically and experimentally in [4].

Figure 3 Distinct AlCl3 concentration boundary layer (left) and no visible GaCl concentration boundary layer (right).

Above differences not only explain many experimentally observed features of the technologies but also predict some new features, which generally result from a high sensitivity of the Ga incorporation into the alloy to variation of the species flow rates vs. high stability of the AlN incorporation that is sensitive only to the rate-limiting species, normally AlCl3.

Once again, if you are interested, [3] and [4] with detailed and thorough analysis of the problem are available for reading on our website.


[1] A. Koukitu, F. Satoh, T. Yamane, H. Murakami, and Y. Kumagai. J. Cryst. Growth 305, 335 (2007).

[2] B. Armas, M. de Icaza Herrera, and F. Sibieude, Surf. Coat. Technology 123, 199 (2000).

[3] "Modeling analysis of AlN and AlGaN HVPE",
A. S. Segal, D. S. Bazarevskiy, M. V. Bogdanov, and E. V. Yakovlev
Phys. Status Solidi C 6, No. S2, S329–S332 (2009)

[4] "Surface chemistry and transport effects in GaN hydride vapor phase epitaxy",
A.S. Segal, A.V. Kondratyev, S.Yu. Karpov, D. Martin, V. Wagner, M. Ilegems,
Journal of Crystal Growth 270 (2004) 384–395 (01)


Subscribe to STR Newsletter

Download free Software DEMO Versions and Documentation

STR 2020. All rights reserved.