Physics Model and Formulas: Quantum Tab¶
Numerically Solving 1D Schrodinger’s Equation¶
The time-independent Schrodinger equation has the form
The inputs of the Schrodinger equation solver include: a finite 1D array with position \(x\), the corresponding potential \(V(x)\) with the same size, the effective mass \(m\), and an eigenstate range specified by the user, \(\left[E_\text{min}, E_\text{max}\right]\). The outputs are the eigenfunction, \(\psi\), and the eigenvalue, \(E\). The difference from standard form of the mass between spatial derivative is the requirement of Hermiticity for spatial dependent mass.
We solve the 1D Schrodinger’s equation numerically. Our method combines the Newton’s method that searches for eigenvalues \(E\) and the Numerov’s method that solves for the corresponding eigenfunction \(\psi(x)\) given any specific \(E\).
Shooting Algorithm for Eigen-problem¶
Initialize with a range for eigen energies.
For each possible eigen energy, solve for the wavefunction using the Numerov’s method for second order differential equations (if the mass is constant, otherwise Euler’s method), and check whether the solution satisfies boundary condition. If so, the energy is an eigen energy.
Use secant’s method to find eigen energy. Newton’s method is not chosen because the root finding converges usually ~10 interaction, and scant’s method (O(n^0.618)) compare to Newton’s (O(n^0.5)) doesn’t worth the extra wavefunction evaluation for numerical derivative.
One noticeable problem for shooting algorithm is that it can miss state pairs that are almost degenerate. When running the software and seeing missing of state, it is recommended to change the global field slightly and/or rotate the layer design and try again.
An example of solving simple Schrodinger equation can be found here.
Effective Mass in Band¶
Band theory predicts that movement of a particle in a potential over long distance can be very different from the movement of the same particle in vacuum. Usually, the movement is complicated; however, when the electron is in the highest energies of the valence band or the lowest energies of the conduction band, it can be shown that electrons behave as free electrons except with a different mass, the effective mass \(m_\text{eff}\).
A particle’s effective mass in each band can be approximated by Taylor expanding the band structure and ignoring higher-than-second-order terms, as the band structure can be expanded locally as
where \(k\) is the wave vector, and \(E_0\) is the edge energy of the band.
For QCL simulations, because of small layer thickness, constant effective mass approximation is sometimes not enough. This can be corrected by including non-parabolic effective mass, or effective mass with energy dependence. In this package, we will offer constant effective mass as simple solver and also non-parabolic effective mass computed using \(k\cdot p\) method.
Specifically in Zincblende crystal, the effective mass has the form including non-parabolic dispersion effect:
Where \(m_0\) is electron mass in vacuum, \(E_g\) is the bandgap at \(\Gamma\) point, \(\Delta_{\text{SO}}, E_P, F\) are parameters describing near-\(\Gamma\) behavior of the conduction band and valence band, \(\Delta E_c\) is the energy of electrons above conduction band, or effective kinetic energy. When \(\Delta E_c=0\) the model reduces to standard effective mass model without non-parabolic dispersion. See [VMRM01], [SCFS94].
For Wurtzite crystal, \(F=0\).
The computation of effective mass is implemented in File band.c (also see File band.h). The code structure is also capable of adding new crystal structures. See the material sections for details.
Scattering mechanism: LO phonon¶
The dominant scattering mechanism for inter-subband transition is Longitudinal optical phonon transition [FB89]. The scattering rate between state \(\psi_u\) and \(\psi_l\) is:
where \(k_u\) and \(k_l\) are upper and lower state electron momentum in the epitaxy layer plain, and \(Q_\theta\) is the phonon momentum. With the assumption that \(k_u = 0\), the formula reduces to:
(The denominator expression maybe problematic… it needs to be checked!)
Optical gain and threshold current¶
Using Maxwell-Bloch equation the optical gain from intersubband transition is
where \(\eta_{\text{inf}}\) is the injection efficiency, which is depend on transitions between all other states but is assumed to be approximatly constant, The Figure of Merit (FoM) is used to characterize the performance of a structure. \(J\) is the current density into the device, and with information of the loss of the optical cavity we can estimate a threshold current, assuming an reasonable \(\eta\) or just put it 1. This estimation is much underestimated.
To couple the design of quantum wells and waveguide, we define the gain coefficient as the ratio of gain the current density, and also assume \(\eta_{\text{inj}} = 1\), \(\omega = \omega_{ul}\), \(\gamma_\parallel = 0.1\omega\).
- FB89
R. Ferreira and G. Bastard. Evaluation of some scattering times for electrons in unbiased and biased single- and multiple-quantum-well structures. Phys. Rev. B, 40:1074–1086, Jul 1989. URL: https://link.aps.org/doi/10.1103/PhysRevB.40.1074, doi:10.1103/PhysRevB.40.1074.
- SCFS94
Carlo Sirtori, Federico Capasso, Jérôme Faist, and Sandro Scandolo. Nonparabolicity and a sum rule associated with bound-to-bound and bound-to-continuum intersubband transitions in quantum wells. Phys. Rev. B, 50:8663–8674, Sep 1994. URL: https://link.aps.org/doi/10.1103/PhysRevB.50.8663, doi:10.1103/PhysRevB.50.8663.
- VMRM01
I Vurgaftman, JR Meyer, and LR Ram-Mohan. Band parameters for III–V compound semiconductors and their alloys. Journal of applied physics, 89(11):5815–5875, 2001.