Device miniaturization comes with a dramatic increase in heat generation inside the integrated circuits. High electric fields in devices generate hot carriers which transfer their energy to the lattice, leading to hot spots. These effects are strongly enhanced in nanoscale devices which sustain high current densities and include interfaces with materials of low thermal conductivity like SiO2.
Our goal is to study these problems in ultimate devices (< 10 nm) in which heat transfer and electron-phonon coupling are significantly affected by the confinement and the low dimensionality.
In our previous ANR project, we developed modules including Poisson / Schrödinger / transport equation (NEGF or Boltzmann) solvers. In the present project, we plan to perform atomistic calculations and to add new modules to the software platform in order to predict quantitatively the role of hot carriers and self-heating on the electrical performances of realistic devices.
Atomistic theory of hot carrier dynamics
The confinement of electrons and phonons considerably influences the low-field transport of carriers in ultimate devices. IEMN and INAC will study the processes controlling the dynamics of hot carriers:
- Emission and absorption of acoustic or optical phonons
- Decay of optical phonons into acoustic ones and generation of heat (optical phonons do not carry heat as acoustic ones)
- Non-equilibrium distribution of carriers and phonons
Thermal properties at the atomic scale
The heat generated by the scattering of hot electrons by the lattice will dissipate ultimately to the heat sink placed at the back side of the device. This dissipation will be impacted by:
- The nano-dimensions of the present devices (thin layer, fin, nanowire)
- The presence of thermal insulators on the dissipation path, such as buried oxides
Two different approaches will be used to calculate the thermal conductivity of the channel region in nanowires and planar devices (both Si and III-V materials). The first calculations will be performed using atomistic codes based on dynamical matrices computed via a modified valence force model within a Landauer approach. Further we will extract the thermal conductivity from approach-to-equilibrium molecular dynamics (AEMD) using interatomic potentials.
Development of a common module for the 3D heat equation
The prediction of the electrical characteristics of nano-devices including thermal effects is a very complex problem. Ideally, transport of both electrons and phonons should be described in a NEGF formalism to account for non-equilibrium distributions and non-local interactions, in a 3D framework for realistic systems. Such an approach is extremely challenging, well-beyond present capabilities of any international group at the moment but must be seen as a long-term objective for the development of our numerical simulation platforms.