Indium incorporation on c-plane and m-plane InGaN surfaces
Developing efficient nitride based optoelectronic emitters operating in the green is a challenging technical problem. Growth of InxGa1-xN active regions with xIn equal to approximately 0.25 is required, and such high concentrations can give rise to defects and poor surface morphology. Growth of InxGa1-xN active regions on various substrate orientations is being explored, and it is not clear what substrate will produce the most efficient green emitter.
Differences in the local atomic structures of m-plane and c-plane wurtzite nitride alloy surfaces are expected to give rise to differing efficiencies of incorporation for In on these two growth surfaces. Investigations of the energetics of indium incorporation on m-plane c-plane InGaN surfaces were therefore performed. The analysis assumes the presence of hydrogen consistent with conditions prevailing in MOCVD growth.
The systems investigated consist of four species: In, Ga, N and hydrogen. Chemical potentials are employed to facilitate the determination of the relative stability of surfaces with varying quantities of these species. The system is assumed to be in equilibrium with an InxGa1-xN compound with xIn = 0.25. Because the growth of such a compound necessitates N-rich conditions, the chemical potential of nitrogen is fixed at a value that is close to its maximum possible value. These two conditions are then employed to eliminate the N and Ga chemical potentials as variables. The chemical potentials of indium and hydrogen are then treated as the independent variable and allowed to vary. The hydrogen chemical potential is related to the temperature and partial pressure of hydrogen in the growth chamber. The indium chemical potential is limited by the onset of droplets if metallic indium. The relative energy of different systems is determined in the range of chemical potentials relevant for MOCVD growth of InGaN. The energies are determined by first-principles pseudopotential density functional calculations with full treatment of the In 4d and Ga 3d electrons as valence electrons.
The calculations show that for incorporation of indium in the surface layer to be energetically favorable the hydrogen partial pressure in the reactor must be kept low. The upper limit on pH for indium incorporation depends on the abundance of indium, but is typically below ~0.1 atm. Otherwise the surface becomes terminated by H and NH2 groups, and In incorporation is inhibited. This requirement is particularly important for the m-plane. A second finding is that the existence of an indium adlayer facilitates the incorporation of indium. Incorporation of In on Ga sites beneath an In adlayer is shown in the figure below for alloy growth on the m-plane. Finally we will show that the space of (In,H) chemical potentials for which incorporation of indium is favorable is somewhat larger for the c-plane than for the m-plane.
This work was supported by the VIGIL-program.
Northrup, J. E. Indium incorporation on c-plane and m-plane InGaN surfaces. International Workshop on Nitride Semiconductors; 2008 October 6-10; Montreux, Switzerland.