Bandgaps: Think they’re constant? Think again!

A team of European researchers has found that the semiconductor bandgap is not necessarily constant, and can vary with distance to the surface.

The partnership between researchers from the UK, France, Spain and Denmark has looked at the surfaces of some compound semiconductors that can support a quasi two-dimensional electron gas (Q2DEG). Here, electrons can move freely parallel to the surface but are confined to this region. Traditional semiconductors like silicon and GaAs have a depletion of carriers close to the surface. For a long time, this has been thought of as the norm. However, increasingly, new materials are being found which exhibit an electron accumulation. For example, Philip King, one of the leading researchers in this study pointed out that he and other scientists have previously observed the presence of a Q2DEG zone at the surface of In-rich InGaN and InAlN alloys. In this latest investigation, the localized surfaces of InAs and CdO were investigated for quantum-well states supposedly intrinsic to these materials. Both materials exhibited similar phenomena. It was also noted by the researchers that very similar results have previously been observed by Colakerol et al [1] regarding InN.

The scientists used the ASTRID synchrotron to obtain Angle-Resolved Photo-Emission Spectroscopy (ARPES) data. This powerful technique directly images the electronic structure of the Q2DEG at the surface, and supplies information regarding the interactions between particles.

The InAs(111)B sample was grown by Molecular Beam Epitaxy (MBE) and was silicon-doped to 6x1017cm3. An amorphous arsenic cap was grown over this structure and removed in-situ by annealing at 3500C.

The CdO was grown by metal-organic vapor phase epitaxy (MOVPE). The figure below shows the quantum well states at the CdO surface. Similar trends were exhibited InAs.

The researchers were surprised to find that interactions between the particles at the surface caused the bandgap to become smaller close to the surface of the material compared to in the bulk.

Phil King commented that it was not the fact that the particle interactions caused a reduction of the bandgap that was unexpected, but the fact that the changes occurred over macroscopic distances within the sample (i.e. the bandgap changes approaching the surface of the materials). These results showed the presence of a complicated interplay between several degrees of freedom within the materials.

He added “The magnitudes of the changes in bandgap are also much larger than might naively be expected considering traditional semiconductors, which is testament to the greater flexibility and variety of properties that can be achieved in some of these emerging compound semiconductor materials.”

Regarding the results, he also explained, “The conventional one-electron picture of surface space-charge in semiconductors was different to the electronic structure observed from ARPES, indicating that many-body interactions play an unexpectedly large role in the Q2DEG in these materials.”

The relevance of this discovery should serve several purposes. The fact that the bandgap of these semiconductors becomes smaller when approaching the surface is essentially due to interactions between the electrons within the surface electron accumulation layer. Essentially these results could provide a stepping-stone in the advancement of bandgap engineering; the authors suggest even an entirely new route to spatially-inhomogeneous bandgap engineering, eventually leading to tuning the functionality of electronic devices.

Furthermore, as King said, “The surface electronic properties of materials are crucial in any device application, as an electrical contact must always be made to the surface of a material (indeed, in his 2000 Nobel prize lecture, Herbert Kroemer remarked that "the interface is the device")”.

“This so-called surface electron accumulation may, for example, make obtaining Schottky (rather than Ohmic) contacts difficult, but could potentially be beneficial for certain device applications such as terahertz generation or chemical sensors.”

Traditionally, the way to tune the bandgap of materials is to alloy two or more semiconductors together that have different bandgaps to start with. The findings in this paper show that the bandgap of a single material may be spatially modulated by controlling how strong the interactions are. This could be controlled by changing the doping levels in the material.

King says he hopes these results could add an additional tool to the arsenal of band structure engineering.

Further details on this work will be available in the journal ‘Physical Review Letters’, soon to be published. [2]

On asking whether the results could be applied to other compound semiconductor materials, King explained that these results should directly apply to other materials that exhibit 2QDEGs, such as GaAs/AlGaAs and GaN/AlGaN heterostructures. However, it is not yet known how strong the effects are in these cases. The next priority of the researchers is to try and use this information to build up a more detailed picture of the many-body interactions within the surface electron accumulation layers. King said that, “Initial results on this look promising, and support the findings we report here, but more work on this is still required.”

[1] Colakerol, L, et al, Phys. Rev. Lett. 97, 237601 (2006)

[2] King, P.D.C. et al., “Surface Band-Gap Narrowing in Quantized Electron Accumulation Layers”, Phys. Rev. Lett. in press”