This tool version is unpublished and cannot be run. If you would like to have this version staged, you can put a request through HUB Support.
Transmission and the reflection coefficient of a five, seven, nine, eleven and 2n-segment piece-wise constant potential energy profile
- Quantum-Mechanical Reflections
- Quantum-Mechanical Reflections in Nanodevices
- Double-Barrier Structure
- From one well, to two wells, to five wells, to periodic potentials
- Bands as a function of the geometry of the n-well potential
Improvements / modifications in subsequent releases:
- 1.1.2 - the energy and transmission coefficent axis are exchanged, so the resonance peaks now line up with the spatial resonances in the barrier structure.
- 1.1.2 - bug-fix: transmission through a single barrier can be simulated now in the "n" barrier case. The code no longer provides an empty output.
- 1.1.2 - the adaptive energy refinement was improved through a different algorithm. The tool no longer utilizes the Matlab built-in adaptive integration routine but an adaptive resonance finding and grid refinement technique as used in the NEMO1D tool or the Resonant Tunneling Diode Tool.
- 1.1.2 - The single barrier case has been corrected and should be functional.
- 1.1.2 - The tool now has a progress update for the adaptive resonance finding.
- 1.1.4- The tool now has the tight-binding Green's function based formalism built into it. This will enable the user to make a comparison between the Transfer matrix method and the single band tight-binding calculation.
- 1.1.5 - spatially varying effective masses are introduced.
- 1.1.6 - Fixed a plotting problem for very small numbers that cannot be properly represented in Rappture.
- 1.1.7 - the effective mass treatment in the tight binding approach has been corrected. The Transfer matrix approach still appears to have some problems when masses are varied across the device.
- 1.1.7 - the natural lattice constant is set to 0.5 for the tight binding calculation such that there is no rounding of the barrier and well heights which will result in deviations from the Transfer matrix method. Now if there is no effective mass variation the two methods give virtually the same results for the default structures.
- 1.1.8 - the effective mass variation in the Transfer Matrix approach is now properly implemented. Tight Binding and Transfer Matrix method now deliver virtually the same result.
- 1.1.8 - proper tool label in the effective mass assignment window.
- 1.1.9 - A problem pertaining to the display of the geometry adjustment in the text log when a tight-binding calculation is called, has been corrected. Rappture plots remain unaffected.
- 1.2- Composite plots have now been introduced. Potential profile, transmission data and band-structure are shown in combination on the same plot.
- 1.2.1-Composite plots now have resonances. X axis is normalized to improve transition from one slider element to another.
- 1.2.1- Transmission and reflection data zoom into not more than seven orders below one.
- 1.2.1- Peaks in transmission are now reported as resonances above the highest barrier also.
- 1.2.1- Outputs rearranged to make them easier to locate.
- 1.2.1- The transmission co-efficient at zero energy is corrected for the case of a flat potential profile. It goes to one for transfer matrices, which can also be seen analytically.
- 1.2.2- Transmission defined as a real quantity to avoid small imaginary values which might cause errors in plotting.
- 1.2.2- An error in the tight-binding calculation pertaining to the potential profiles affect on the BCs has been corrected.
- 1.2.2- Bulk bandstructure now displayed as an output.
- 1.2.2- The adaptive mesh algorithm was refined to significantly reduce the number of energies needed in cases of very sharp resonances. For example in the case of 30 barriers the number of energy nodes was reduced by over half from around 12,000 points to just over 5,000 points. The computation time was reduced by around 50% in those cases. In the case of small number of barriers the reduction of number of nodes is slightly less and the mesh refinement algorithm itself requires more time. The code slows down by about 10-20%. Those compute times are within 2-3 seconds, so the slow-down is not that significant.
|Vasileska||... core transfer matrix engine in matlab|
|Luisier, Agarwal||... core effective mass engine in matlab|
|Wang,Agarwal||... GUI implementation|
|Klimeck||... GUI and output design|
|Sellier||... adaptive mesh refinement|
Dragica Vasileska lecture notes on Quantum mechanics http://www.eas.asu.edu/~vasilesk .
Cite this work
Researchers should cite this work as follows: