When an n-type nanowire (NW) is placed on a p-type substrate, does it really form a standard p–n junction? A group of researchers from Harvard University suggest that the answer, in general, is "no". Especially for materials in which doping of both types (p and n) is difficult to achieve (e.g. ZnO, GaN), it's of paramount importance to answer this question.

The most common application for NW-substrate junctions today is the light-emitting diode (LED). The key challenge in creating NW light sources lies in their assembly rather than material synthesis. One aspect of this challenge, which has received considerable attention, deals with the geometric arrangement of NWs into large-scale patterns. The other aspect, which has received almost no attention, is the control of the interface properties resulting from making contact between NWs and planar surfaces. Semiconductor light sources derive their behaviour from the properties of the interface between two semiconductors: the p–n junction LED and the so-called double heterojunction laser are the most obvious examples.

In this study, in collaboration with the group of Prof. Venky Narayanamurti, also at Harvard University, the issue of p–n junction formation was addressed in the NW on-substrate geometry for the purpose of creating an ultraviolet NW LED and understanding the underlying physics affecting the luminescence properties and current-voltage characteristics, an aspect so far largely neglected in previous studies of such sytems. LEDs were assembled using n-type doped GaN NWs in contact with a p-type silicon substrate. The results show that, in general, junctions resulting from the intimate mechanical contact with a semiconductor substrate, due to Van der Waals type forces, are far from ideal compared with those epitaxially grown in planar systems, and practically form a tunnel junction rather than a standard p-n junction.