Can buried quantum dots make wide-gap nitrides conduct holes?
Embedding tiny GaN quantum dots in aluminium-rich AlGaN raised the top of the valence band so magnesium acceptors release holes far more easily, giving good p-type conduction and a better deep-ultraviolet LED.
Source
Quantum engineering of non-equilibrium efficient p-doping in ultra-wide band-gap nitrides
Study at a glance
- Design
- Other — DFT calculations of Mg acceptor levels near GaN quantum dots in AlN, followed by growth of Mg-doped AlGaN with buried GaN dots, temperature-dependent Hall measurements, optical characterisation and a comparison of two deep-UV LEDs.
- N
- No sample size; grown films with about 50%, 60% and 70% Al content, plus two LED devices (quantum-engineered versus uniformly doped p-layer).
- Population
- Mg-doped AlGaN semiconductor films with embedded GaN quantum dots
- Outcome
- Acceptor activation energy, hole concentration, resistivity, band gap, and LED current-voltage and electroluminescence
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The authors first used density functional theory to compute how much energy a magnesium acceptor needs to free a hole when it sits at different distances from a GaN quantum dot inside AlN. They then grew Mg-doped AlGaN containing periodically buried GaN dots using an interrupted, non-equilibrium deposition sequence repeated 40 times, and checked the structure with X-ray diffraction, electron microscopy and mass spectrometry. Temperature-dependent Hall and resistivity measurements gave the acceptor activation energy, and two deep-UV LEDs that differed only in their p-type layer were compared.
What they found
Calculations showed the activation energy depends strongly on where the Mg sits: doping inside the dot was poor, while sites in the matrix near the dot interface gave much lower values. In the grown material with 60% aluminium, the activation energy came out at roughly 21-43 meV, far below the 630 meV typical of Mg in AlN, with a room-temperature hole concentration of about 1.25 × 10^18 per cubic centimetre; similar values were found at 50% and 70% aluminium. The LED with the quantum-engineered p-layer turned on at lower voltage, emitted more strongly and showed a single peak at the designed wavelength, while the conventional device showed an extra peak from electron overflow.
The limits
What it doesn't show
The calculations used a small supercell with very strong quantum confinement, so the computed energies at farther doping sites could not be tested and the authors only argue that larger dots would do better. The two methods of extracting activation energy disagreed by about a factor of two, which the authors attribute to dot size spread and contact problems at low temperature. The LED comparison involves just one device of each type with no reported efficiency figures or lifetime tests, and whether the approach transfers to other semiconductors is only proposed.
Key terms
- Acceptor activation energy
- The energy needed for an acceptor impurity to capture an electron from the valence band, releasing a mobile hole.
- Valence band maximum
- The highest energy level of the filled valence band; raising it closer to the acceptor level makes holes easier to create.
- Quantum dot
- A nanometre-scale semiconductor island that confines carriers in all three dimensions.
- p-type doping
- Adding impurities that create mobile positive holes as the main charge carriers.
- Hall measurement
- A technique using a magnetic field and the resulting sideways voltage to find carrier type, concentration and mobility.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Why do GaN quantum dots lower the acceptor activation energy in AlGaN?
Common questions
Why is p-doping so hard in aluminium-rich nitrides?
Their valence band sits very low, so the Mg acceptor level is far above it and few acceptors ionise at room temperature, where thermal energy is only about 26 meV.
Why not just put the Mg inside the GaN dots?
Calculations showed quantum confinement pushes the dot's valence band down, so Mg in the dot centre had a higher activation energy than in bulk GaN; sites in the matrix near the interface worked best.
Why was non-equilibrium growth needed?
Under equilibrium conditions GaN dots do not form inside AlGaN and Mg does not concentrate at interfaces, so gas sources were switched on and off in sequence to force that structure.
More on Non-equilibrium and stochastic thermodynamics