2.2 Micro- and Nanoelectronics Device Simulation Tools

The field of MNDS offers highly specialized publicly available simulation tools, of which an overview is presented in the following in alphabetical order.

Archimedes [5][33] is a two-dimensional Monte Carlo Boltzmann transport based simulation tool for submicron and nanoscale semiconductor devices. Various physical effects and transport models can be investigated for electrons and heavy holes with respect to a rich set of materials. Heterostructures as well as electrostatic and magnetic fields by solving Poisson’s and Faraday’s equation are supported. The tool is released under the general public license ( GPL) and coded in the C programming language. Additionally, an online, GUI-based version of Archimedes is provided via the nanoHUB platform.

Genius [34] is available as a publicly accessible version under the GNU GPL. The FLOSS version supports two-dimensional device simulation based on the drift-diffusion ( DD) model. Lattice heating is taken into account by, for instance, a temperature corrected DD model. A rich set of functionality is provided, such as various mobility models, an energy transport model, and several impact-ionization models.

Gold Standard Simulations [35] specializes in simulating statistical variability in nano-CMOS devices and provides corresponding commercial simulations tools. More specifically, the tools support the physical simulation of statistical variability, statistical compact model extraction, and statistical circuit simulation.

Minimos-NT [36][37] is the successor of the Minimos [38] simulator and is commercially supported. Minimos-NT is a general-purpose semiconductor device simulator, providing a general-purpose, multi-dimensional semiconductor device simulator. The simulator supports stead-state, transient, and small-signal analysis of arbitrary devices. Also, mixed-mode device and circuit simulations based on compact models are supported.

nanoHUB is a platform hosting scientific tools, primarily in the field of computational nanotechnology [39][40]. At the time of writing this thesis from the total number of 325 tools 17 tools (corresponding to 5.2%) are tagged as open source1 . Figure 2.1 gives an overview of the accumulated code lines of each project. The Count Lines of Code [41] tool has been used to quantify the code base implemented in languages such as C/C++, Python, Matlab, and Fortran. Irrelevant data has been - to a large extent - ignored, like comments and building instructions. Of the 17 open source tools 41% have between 100 and 1 000, 35% offer 1 000 to 10 000, and 24% provide 10 000 to 100 000 lines of code. Therefore, the majority of the available free open source tools can be considered to be small to medium scale-size projects, further underlining the lack of FLOSS-based device simulation tools of considerable size. Overall, nanoHUB provides free registration for an online account, enabling the execution of tools directly from within a web browser. The computational resources are provided by the nanoHUB facilities, thus no compilation and/or installation procedure is required.


pict


Figure 2.1: The accumulated code lines of the 17 open source nanoHUB simulation tools are depicted. Seven tools (41%) have 100 to 1 000 code lines, six tools (35%) have 1 000 to 10 000, and four tools (24%) offer 10 000 to 100 000 code lines, respectively.


NanoTCAD ViDES [42] supports the simulation of nanoscale devices through the self-consistent solution of the Poisson and the Schrödinger equations by means of the Non-Equilibrium Green’s Function formalism. The tool allows for the simulation of transport in graphene nanoribbons, carbon nanotubes, and two-dimensional (bilayer) graphene field-effect transistors. The simulator is distributed as a Python module, utilizing high performance C and Fortran based subroutines. The package is released under the BSD License.

Silvaco [43] provides a broad set of commercial simulation tools for TCAD, interconnect modeling, and analog/mixed-signal/radio frequency analysis. A broad set of modeling and analysis tools are provided, allowing for a wide range of simulations and evaluations.

Synopsis [44] provides a plethora of commercial simulation tools, covering a variety of application categories, such as TCAD, verification, manufacturing, and system-level design. Extensive pre- and postprocessing facilities are provided, such as structure and mesh generation as well as visualization.

ViennaSHE [45][46] is a multi-dimensional, self-consistent semiconductor device simulator based on the deterministic solution of the Boltzmann Transport Equation using Spherical Harmonics Expansions. ViennaSHE provides a standalone simulation application as well as an API for utilizing the simulator by other implementations. The tool is released under the MIT License and written in C++.

The presented open source simulation tools are highly specialized. However, they share the requirement for certain pre- and postprocessing software components. For instance, each tool requires visualization capabilities to enable investigations of the simulation results. Another typical requirement is the generation of the simulation domain and the access to material parameters. A detailed analysis of the available open source simulation tools leads to the conclusion that these tools treat these aspects in a marginal manner. For instance, only a static set of material parameters is supported, which is hard coded into the simulation code. This specific aspect introduces the need for a flexible material mechanism, providing simulation tools access to various material database backends. With respect to the commercially distributed simulation tools, the advantages and disadvantages, as introduced in Section 2.1, apply here in a similar manner.