Contact Block Reduction (CBR) method

Dr. Denis Mamaluy is working on the quantum transport simulation in semiconductor nano-structures and nano-devices. He is one of the authors of the CBR method that is based on the Green's function formalism.

An introduction to the quantum transport simulation using the CBR method can be found here.

Our group is applying the CBR method to simulate a variety of nano-size devices, including:

• double-gate MOSFETs (view an animated example)

• FinFETs (view the latest results), Tri-gate FETs (3D simulation results)

• quantum dots

• nano-wires

• Single-electron transistors (in partnership with NIST)

Background: The quantum-mechanical transport calculation of 3D structures presents a significant challenge due to profound computational difficulties. Although a large variety of other methods developed over last decade to address this issue, many commonly used approaches are insufficient.  These methods cannot be used to accurately describe devices with more than two Ohmic contacts, or have severe restrictions on device geometry. Most importantly, presently available methods suffer from extremely long simulation times and often can only be performed on large and expensive computer clusters. Thus, there exists a profound need for a fast, universal and fully 3D quantum-mechanical transport simulation tool, since many novel nano-size devices (e.g. FinFETs) possess essentially 3D geometry.

We aim to fulfill that need, using a tool based on a novel, numerically efficient CBR technique, which is based on the Green’s function formalism. The CBR method is applicable to fully self-consistent quantum transport calculations in arbitrarily shaped 2D and 3D structures, with any number of leads, using either the effective mass approximation or the mutli-band Hamiltonian description. Within the CBR method the open system is treated as truly 2D or 3D; no extra assumptions on the device symmetry are implied.

Research Description: Using the CBR technique, we are able to dramatically improve simulation times for the quantum-mechanical transport calculation. Using our tool, a fully converged self-consistent calculation for a typical 2D double-gate FET structure requires only about 50-60 minutes per bias point on a regular PC, which is an order of magnitude faster than other quantum transport simulators presently available (see presentations given at Freescale Semiconductor Inc., Synopsys Inc. and at Purdue University). At the same time, the CBR results are in very good correspondence with the results from other simulation groups and experimental data.

Presently we are working on extending our tool to incorporate the self-consistent, quantum-mechanical transport models including strain, high-k/metal gate (using k.p or tight-binding descriptions), surface orientation effects, rigorous treatment of gate leakage effects, interface roughness and unintentional doping effects. The CBR simulator also includes the effect of electron-electron interaction, which becomes increasingly important for many small nano-devices, and scattering on phonons through a simplified NEGF model.

The list of the most recent publications on the CBR method can be found here.