Device type: Cirular/Square silicon nanowire – channel dir: < 110>, conf dir: < 100> and channel dir: < 100>, conf dir: < 100> Dimensions: 2.5nm to 3.5nm Strain condition: longitudinal uniaxial tension/compression (-6% to 6%)

Comparing to the very well-known bulk silicon strain physics and a few available papers that report strained nanowire DFT or TB results, I believe there might be a problem in the strain model. The response of the simulator to longitudinal uniaxial tension is just as it is expected from a compressive strain (CB gamma/off-gamma valley splitting reduces and CB edge switches from Gamma to off-Gamma point with tensile strain, along with a reduction in hole mass due to band-warping). Similarly, exact opposing results with respect to known strain effects is observed for uniaxial compression (CB valley splitting increases, hole mass increases due to bandwarping and “camel back” hole band becomes the top valence band).

Although having some problems in certain cases, the previous version of the Bandstructure Lab tool was able to produce results that are at least predicted from known strain physics in bulk silicon, by means of mass change due to bandwarping and energy splitting change due to bandsplitting.

I believe the problem may occur from a simple sign error in the strain input line to the simulator. I would greatly appreciate if you can recheck your strain inputs to the sp3d5s* TB model and verify its consistency.