You are here

Polarization-induced diodes in graded nanowires

The bedrock of solid state electronic devices is the ability to control the conductivity of semiconductors. This is traditionally done by adding small amounts of impurity atoms to an otherwise nearly impurity-free material. In 2002, Jena et al. developed a fundamentally new way to generate conductivity in semiconductors accomplished by grading the composition of AlGaN along its c-axis, the direction along which the polarization dipoles lie in wurtzite 1. In a single crystal of GaN or AlN the net polarization charge within each unit cell is canceled by neighboring unit cells resulting in a net polarization charge only at the top and bottom surfaces (Fig. 1a). However, if the composition is linearly graded, from GaN to AlGaN, then neighboring unit cells have a different sized polarization dipole: the polarization charge is not completely canceled but a constant background polarization charge is left behind. For example, grading from GaN to AlGaN along the Ga-face direction (polarization dipole going from + to - along c-direction, c-axis bond is ordered cation to anion) each subsequently stacked unit cell with increasing %Al has a larger dipole magnitude (Fig. 1b), partially canceled by the previous unit cell, but leaving behind a net positive polarization charge that is constant throughout the graded region. This polarization charge is fixed by the crystal structure, but must be compensated by free charge carriers to achieve charge neutrality. Solving the Poisson equation for such a graded region leads to a build-up of electrons in the graded regions to compensate the bound positive polarization charge. Those electrons occupy the conduction band and are therefore mobile. However, they entered the crystal not from bulk impurity states, but from surface donors. This n-type conductivity is observed even without any donor doping, which removes ionized impurity scattering. One could also achieve bound negative polarization charge simply by grading from AlGaN to GaN along the Ga-face direction, or from GaN to AlGaN along the N-face direction. The bound negative charge requires positive charges (holes) to accumulate in the graded region resulting in p-type behavior, as was shown by Simon et al. in 2010 2.

In nanowires, the large surface to volume ratio provides two advantages in this regard. First, the strain due to grading between two lattice mismatched materials is accommodated by the free surface allowing access to the full composition range of AlGaN. Second there is a larger ratio of surface states to bulk states, which can boost the polarization doping efficiency. With these ideas in mind, we adapted the polarization grading concept described above into nanowires. In particular we grew nanowires that were doubly graded, from GaN to AlN and back to GaN. Grading in one direction generates an n-type conducting layer, and grading in the reverse direction a p-type layer. Side by side, this is a simple pn-diode. However, it can be formed even without any impurity doping. The diodes exhibit clear rectification (Fig. 2) and show very little temperature dependence since, unlike impurity doping, polarization doping does not freeze-out. The polarization-induced diodes can be made into LEDs by inserting AlGaN quantum wells into the junctions and emit UV light (Fig. 3).

For more information about this work, please see:
S.D. Carnevale, T. F. Kent, P.J. Phillips, M.J. Mills, S. rajan  and R.C. Myers. Nano Lett. 12, 915-920 (2012) .

References

  1. Jena, D. et al. Realization of wide electron slabs by polarization bulk doping in graded III-V nitride semiconductor alloys. Appl. Phys. Lett. 81, 4395-4397 (2002).
  2. Simon, J., et al. Polarization-induced hole doping in wide-band-gap uniaxial semiconductor heterostructures. Science 327, 60-64 (2010)