next up previous contents
Next: Bibliography Up: Dissertation Siddhartha Dhar Previous: Appendix A


Appendix B

From (4.118) and (4.119) we get

$\displaystyle a_{44} = \frac{\Omega_{100}-\Omega_{110}}{210}$ (6.10)

After elimination of $ a_{44}$, the system of equations (4.118) to (4.122) can be expressed in matrix form as

$\displaystyle \ensuremath{{\underline{S}}}\cdot \ensuremath{{\underline{Z}}} = \ensuremath{{\underline{B}}}$ (6.11)

where

$\displaystyle \ensuremath{{\underline{S}}} = \begin{pmatrix}S_{00} & S_{01} & S...
... -1/2 & 3/8  1 & 1 & 1  1 & 1/4 & -13/32  1 & 1/4 & -13/32 \end{pmatrix},$ (6.12)
$\displaystyle \ensuremath{{\underline{Z}}} = \begin{pmatrix}z_{0}  z_{1} z_...
... \Omega_{101} - 105a_{44}/4  \Omega_{11\sqrt{2}} + 105a_{44}/4 \end{pmatrix}$ (6.13)

To determine the four coefficients from the overdetermined system (6.11), we solve (4.118) to (4.120) exactly and minimize the error in  (4.121) and (4.122). Since $ S_{00} = S_{10} = S_{11} = S_{12} =
S_{20} = 1$, using the new variables the equations can be written as

$\displaystyle z_0 + S_{01}z_1 + S_{02}z_2 - b_0 = 0,$ (6.14)
$\displaystyle z_0 + z_1 + z_2 - b_1 = 0,$ (6.15)
$\displaystyle z_0 + S_{21}z_1 + S_{22}z_2 - b_2 \rightarrow \ensuremath{{\mathrm{Min}}},$ (6.16)
$\displaystyle z_0 + S_{21}z_1 + S_{22}z_2 - b_3 \rightarrow \ensuremath{{\mathrm{Min}}}.$ (6.17)

Next, we eliminate $ z_0$ by subtracting (6.15) from (6.14), (6.16), and (6.17).

$\displaystyle (S_{01} - 1)z_1 + (S_{02} - 1)z_2 + b_1 - b_0 = 0$ (6.18)
$\displaystyle (S_{21} - 1)z_1 + (S_{22} - 1)z_2 + b_1 - b_2 \rightarrow \ensuremath{{\mathrm{Min}}}$ (6.19)
$\displaystyle (S_{21} - 1)z_1 + (S_{22} - 1)z_2 + b_1 - b_3 \rightarrow \ensuremath{{\mathrm{Min}}}$ (6.20)

To eliminate $ z_1$, (6.18) is multiplied by $ \tau =
(S_{21}-1)/(S_{01}-1)$ to obtain

$\displaystyle (S_{21} - 1) z_1 + \tau (S_{02} - 1) z_2 + \tau (b_1 - b_0) =0.$ (6.21)

Subtracting (6.21) from (6.19) and (6.20) gives

$\displaystyle f_1 = uz_2 + v \rightarrow \ensuremath{{\mathrm{Min}}},$ (6.22)
$\displaystyle f_2 = uz_2 + w \rightarrow \ensuremath{{\mathrm{Min}}},$ (6.23)

with

$\displaystyle u = (S_{22}-1) - \tau (S_{02}-1),$ (6.24)
$\displaystyle v = b_1 - b_2 - \tau (b_1-b_0),$ (6.25)
$\displaystyle w = b_1 - b_3 - \tau (b_1-b_3).$ (6.26)

The least square minimum can be obtained by setting the derivative to zero,

$\displaystyle \frac{\partial f_1^2}{\partial z_2} + \frac{\partial f_2^2}{\partial z_2} = 2  (uz_2 +v) u + 2 (uz_2 + w) u = 0,$ (6.27)

which gives

$\displaystyle z_2 = -\frac{v +w}{2u}.$ (6.28)

With the definitions (6.24) to (6.26) and using (6.12) we get

$\displaystyle z_2 = -\frac{16}{35} (-b_0 - b_1 + b_2 + b_3),$ (6.29)
$\displaystyle z_1 = \frac{2}{3} (b_0 + b_1) - \frac{5}{12}z_2,$ (6.30)
$\displaystyle z_0 = \frac{2}{3} (-b_0 + b_1/2) - \frac{7}{12}z_2.$ (6.31)

Substituting back the value of $ b_i$'s from (6.13) into (6.29) to (6.31) gives equations (4.123)-(4.126).


next up previous contents
Next: Bibliography Up: Dissertation Siddhartha Dhar Previous: Appendix A

S. Dhar: Analytical Mobility Modeling for Strained Silicon-Based Devices