« PreviousUpNext »Contents
Previous: 5.3 Single-particle Transition Rate    Top: 5 Electron-Electron Scattering    Next: 5.5 Implementation for Parabolic and Isotropic Bands

Investigating Hot-Carrier Effects using the Backward Monte Carlo Method

5.4 Implementation for Full-Band Structures

Up to this point, no assumption about \( \Epsilon (\kv _1) \), the dispersion relation of the primary electron, has been made. This fact allows us to construct a model in which for the high-energetic primary electron a full-band structure, as described in Section 2.1.4, is assumed. For the low-energetic partner electron the parabolic band approximation is used, as shown in the previous sections.

5.4.1 Total Scattering Rate

The total scattering rate \( \Gamma _1 \) is obtained by integration of the scattering rate (5.59) or (5.63) over the final states of the sample electron:

(5.67) \{begin}{align} \Gamma _1(\kv _1) = \int S(\kv _1,\kv _1’)\,\d ^3 k_1’ \label {eq:Gamma_1} \{end}{align}

The integral in (5.67) is approximated by a discrete sum in \( \kv \)-space.

(5.68) \{begin}{align} \Gamma _1(\kv _n, N_n) \approx \sum \limits _{m}^{N_n} S(\kv _n, \kv _m)\, V_m \label {eq:Gamma1-discrete} \{end}{align}

Here, \( \kv _m \) and \( V_m \) denote the center and the volume of the \( m \)-th tetrahedron, respectively. Whereas \( \kv _n \) denotes a discrete point in \( \kv \)-space.

The contributions of all neighboring tetrahedra of tetrahedron \( n \) are calculated and stored in a table. Recursively, all neighbors of these contributing tetrahedra are included in this table, see Fig. 5.1. The recursive search for contributing neighbors ends, when a tetrahedron contributes less than a pre-defined tolerance to the total scattering rate. The number of all tetrahedra found in this way for one particular point \( \kv _n \) is defined as \( N_n \).

(-tikz- diagram)

Figure 5.1: All contributing neighbors of a point \( \kv _n \) are recursively included. In the first step, all direct neighbours are added. For the second and third step, the tetrahedrons \( \mathrm {Tet}_{2,1} \) and \( \mathrm {Tet}_{2,1,1} \), respectively, are exemplarily illustrated. For the calculation of the scattering rate, all tetrahedra are represented by the center \( \kv _m \) and their volume \( V_m \).

5.4.2 Obtaining the Final State

To enable the selection of the after-scattering state, all partial sums of the form

(5.69) \{begin}{align} \Gamma _1(\kv _n, N) = \sum \limits _{m}^{N} S(\kv _n, \kv _m)\, V_m\„\qquad N\in [1,N_n] \label {eq:Gamma1-discrete-1} \{end}{align}

are pre-computed and stored in a table [P5]. This table of the partial sums (5.69) is stored for each discrete initial state \( \kv _n \) in the irreducible wedge of the Brillouin zone, for each band \( b \), and for a set of discrete Fermi levels \( E_F \) in the case of Fermi-Dirac statistics.

The final state is obtained by first, randomly selecting a tetrahedron \( N\in [1,N_n] \) using the pre-computed table of partial sums.

(5.70) \{begin}{align} \Gamma _1(\kv _n, N-1) \leq r < \Gamma _1(\kv _n, N) \{end}{align}

The uniformly distributed random number \( r \) is in the range

(5.71) \{begin}{align} 0 \leq r \leq \Gamma _1(\kv _n, N_n) \,.                \{end}{align}

Once a tetrahedron has been selected, a uniformly-distributed random state inside the tetrahedron is chosen using Barycentric coordinates.

Barycentric Coordinates for a Tetrahedron

A tetrahedron has four vertices \( \{V_0~,~V_1~,~V_2~,~V_3\} \) . The barycentric coordinates \( \{ \xi _0~,~\xi _1~,~\xi _2~,~\xi _3\} \) of a point \( P \) inside the tetrahedron can be calculated as [19]:

(5.72) \{begin}{align} \xi _i=\frac {\mathrm {volume}(P,V_{(i+1) \bmod 4},V_{(i+2)\bmod 4},V_{(i+3)\bmod 4})}{\mathrm {volume}(V_i,V_{(i+1)\bmod 4},V_{(i+2)\bmod 4},V_{(i+3)\bmod
4})}\,. \{end}{align}

Because of the condition

(5.73) \{begin}{align} \sum \limits _{i=0}^{3} \xi _i = 1 , \{end}{align}

one coordinate is redundant.

Random Selection of a Point inside a Tetrahedron

A uniformly distributed random point inside a tetrahedron can be obtained by randomly chosen barycentric coordinates [81].

(5.74–5.77) \{begin}{align}   \xi _0 &= r_0\\ \xi _1 &= r_1\,(1-\xi _0)\\ \xi _2 &= r_2\,(1-\xi _0-\xi _1)\\ \xi _3 &= 1-\xi _0-\xi _1-\xi _2 \{end}{align}

Here, \( r_i \) are uniformly distributed random numbers in \( [0,1[ \).

« PreviousUpNext »Contents
Previous: 5.3 Single-particle Transition Rate    Top: 5 Electron-Electron Scattering    Next: 5.5 Implementation for Parabolic and Isotropic Bands