Support portal Careers Contact Investors
EN
EN · English JA · 日本語 KO · 한국어 ZH · 简体中文
Request a Demo
Request a Demo
Request a Demo
Support portal Careers Contact Investors
Language EN JA KO ZH

Victory Atomistic: atomistic simulation and quantum transport

When bulk parameters stop describing your device, Victory Atomistic solves it atom by atom and hands better-calibrated parameters back to your TCAD flow.

Talk to our Technical Experts Get downloads and documentation

What Victory Atomistic does

Victory Atomistic models your device atom by atom. It solves quantum transport using non-equilibrium Green's functions (NEGF). Structures run from a few atoms to around a million, the range where continuum device models give out. Each material starts as a tight-binding description, so confinement, strain relaxation and phonon modes are calculated, not fitted. That's the case for atomistic simulation when a new material or an advanced node arrives with no calibration history.

Capabilities

  • Quantum transport

    A self-consistent Schrödinger–Poisson solver sets the electrostatics. NEGF then gives you I-V characteristics with coherent and incoherent scattering included. When a question doesn't need full NEGF, you can answer it faster with top-of-the-barrier transport or the Quantum Transmission Boundary Method (QTBM). Force-field relaxation finds the strained geometry before any of it runs.

  • Scattering

    Carrier scattering is solved self-consistently in the Born approximation, not bolted on afterward. That covers electron-phonon (acoustic, polar and non-polar), surface roughness and impurities. Leave it out and you get a ballistic current: an upper bound, not a prediction.

  • Structure and materials

    You define the device by material, domain and region. Pseudoepitaxy and passivation rules then build the lattice, boundary conditions included. The tight-binding database holds established and novel microelectronics materials, from bulk to 2D layers. Band structure comes from those parameters or from Wannier90 Hamiltonians. Supported symmetries are zincblende, wurtzite, simple-cubic, diamond and trigonal, which covers graphene and BiTe.

  • Scale and performance

    A handful of atoms or a million, in the same tool. Mode-space and Büttiker-probe techniques cut the cost of the transport calculation itself. Multilevel MPI parallelization then spreads what remains, and massively parallel runs across hundreds of thousands of compute nodes are possible. You set up and review every run in Silvaco's TCAD interactive tools.

Where it sits in the TCAD flow

Victory Atomistic runs beside the main Victory chain, not inside it.

You solve the effect a continuum model can't capture, then carry the answer into Victory Device.

  1. Material and structure definitions
  2. Victory Atomistic Solved at atomic resolution
  3. Victory Device Sharper parameters for its classical and semi-classical models

Applications

Proof and resources

Read TCAD published papers Teaching or researching? See the University Program, used at 200+ universities

Evaluate and engage

  • Evaluating it

    Bring the effect your continuum model gets wrong: leakage it underestimates, or a threshold shift it can't reproduce. You'll know whether atomistic simulation can settle it before you put a single run behind it.

    Talk to our Technical Experts

  • Downloads and documentation

    Licensed users get installers and manuals through Support and Downloads. The product brief and the practical-simulation deck need no login.

    Downloads Documentation

  • Support

    With a current maintenance contract, you open a case in the support portal and follow it there until it's resolved.

    Open a support case Go to Support and Downloads

Questions about atomistic simulation

It models a device as the lattice of atoms it really is, rather than as a continuum with averaged material properties. Band structure, confinement and strain then emerge from how the atoms sit. Nothing has to be supplied as a fitted parameter, and fitted parameters are exactly what a new material doesn't have yet.

When bulk parameters like effective mass and bandgap stop meaning much, and the wave nature of carriers starts deciding transport. That happens at 5 nm and below, in nanosheets and nanowires, and in quantum dots and 2D-material channels. Before that point, a drift-diffusion simulator such as Victory Device is the faster tool.

Non-equilibrium Green's functions (NEGF) is a quantum formalism for transport through a device under bias, which is where a working nanoscale transistor always is. It returns I-V characteristics with coherent and incoherent scattering included, so tunneling and phonon interactions get calculated instead of patched in as correction terms.

Around a million atoms at the top end, and down to just a few. Reaching that ceiling takes method as much as hardware. Mode-space projection and Büttiker probes shrink the problem before MPI parallelization ever sees it, so large structures still finish in realistic simulation times.

Yes. The tight-binding database runs from bulk materials to 2D layers, and transition-metal dichalcogenides are among the devices you can simulate with it. If your material has no tight-binding parameter set yet, you can bring its band structure in from a Wannier90 Hamiltonian instead.

They give Victory Device sharper parameters for its classical and semi-classical models. You pay the atomic-resolution cost once, on the small piece that needs it, and the full device then runs at continuum speed. Atomistic simulation stays a calibration step in your flow, not a replacement for it.

Related TCAD products

See the full TCAD simulation suite