Skip to main content
Log in

An advanced equation assembly module

  • Original article
  • Published:
Engineering with Computers Aims and scope Submit manuscript

Abstract

We present an advanced equation assembly module which has been developed for the simulation of semiconductor devices based on the Finite Boxes discretization scheme and is currently used in the general purpose device and circuit simulator Minimos-NT. Such simulations require the solution of a specific set of nonlinear partial differential equations which are discretized on a grid. The resulting nonlinear problem is solved by a damped Newton algorithm that demands the solution of a linear equation system at each step. The presented module is responsible for assembling these systems and takes into account several requirements of the simulation process. The underlying concepts, namely the representation of boundary conditions, physically motivated variable transformation, preelimination and numerical conditioning, are presented together with some examples.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9

Similar content being viewed by others

References

  1. Institut für Mikroelektronik, Technische Universität Wien, Austria (2004) Minimos-NT 2.1. User’s Guide. http://www.iue.tuwien.ac.at/software/minimos-nt

  2. Selberherr S (1984) Analysis and simulation of semiconductor devices. Springer, Wien

    Google Scholar 

  3. Stratton R (1962) Diffusion of hot and cold electrons in semiconductor barriers. Phy Rev 126(6):2002–2014

    Article  Google Scholar 

  4. Bløtekjær K (1970) Transport equations for electrons in two-valley semiconductors. IEEE Trans Electron Devices ED-17(1):38–47

    Article  Google Scholar 

  5. Grasser T, Kosina H, Gritsch M, Selberherr S (2001) Using six moments of Boltzmann’s transport equation for device simulation. J Appl Phys 90(5):2389–2396

    Article  Google Scholar 

  6. Wachutka G (1990) Rigorous thermodynamic treatment of heat generation and conduction in semiconductor device modeling. IEEE Trans Computer-Aided Design 9:1141–1149

    Article  Google Scholar 

  7. Grasser T, Selberherr S (2001) Fully-coupled electro-thermal mixed-mode device simulation of SiGe HBT circuits. IEEE Trans.Electron Devices 48(7):1421–1427

    Article  Google Scholar 

  8. Grasser T, Tang T, Kosina H, Selberherr S (2003) A review of hydrodynamic and energy-transport models for semiconductor device simulation. Proc IEEE 91(2):251–274

    Article  Google Scholar 

  9. VanRoosbroeck W (1950) Theory of flow of electrons and holes in germanium and other semiconductors. Bell Syst Techn J 29:560–607

    Google Scholar 

  10. Schroeder D (1994) Modelling of interface carrier transport for device simulation. Springer Berlin Heidelberg New York

  11. Simlinger T (1996) Simulation von Heterostruktur-Feldeffekttransistoren. Dissertation, Technische Universität Wien. http://www.iue.tuwien.ac.at.

  12. Fischer C (1994) Bauelementsimulation in einer computergestützten Entwurfsumgebung. Dissertation, Technische Universität Wien. http://www.iue.tuwien.ac.at.

  13. Scharfetter D, Gummel H (1969) Large-signal analysis of a silicon read diode oscillator. IEEE Trans.Electron Devices ED-16(1):64–77

    Article  Google Scholar 

  14. Deuflhard P (1974) A modified newton method for the solution of ill-conditioned systems of nonlinear equations with application to multiple shooting. Numer Math 22:289–315

    Article  MATH  MathSciNet  Google Scholar 

  15. Bank R, Rose D (1981) Global approximate Newton methods. Numer Math 37:279–295

    Article  MATH  MathSciNet  Google Scholar 

  16. Van der Vorst H (1992) BI-CGSTAB: A fast and smoothly converging variant of BI-CG for the solution of nonsymmetric linear systems. SIAM J Sci Stat Comput 13(2):631–644

    Article  MATH  MathSciNet  Google Scholar 

  17. Saad Y, Schultz M (1986) GMRES: A generalized minimal residual algorithm for solving nonsymmetric linear systems. SIAM J Sci Stat Comput 7(3):856–869

    Article  MATH  MathSciNet  Google Scholar 

  18. Saad Y (1990) A basic tool kit for sparse matrix computations. Moffett Field, CA 94035: Technical Report, RIACS, NASA Ames Research Center

  19. Ascher U, Markowich P, Schmeiser C, Steinrück H, Weiss R (1986) Conditioning of the steady state semiconductor device problem. Tech Rep 86-18, University of British Columbia

  20. Fischer C, Selberherr S (1994) Optimum Scaling of Non-symmetric jacobian matrices for threshold pivoting preconditioners. In: International workshop on numerical modeling of processes and devices for integrated circuits NUPAD V, (Honolulu), pp 123–126

  21. Barrett R, Chan MBT, Demmel J, Donato J, Dongarra J, Eijkhout V, Pozo R, Romine C, Van der Vorst H (1994) Templates for the solution of linear systems: building blocks of iterative methods. SIAM, Philadelphia

    Google Scholar 

  22. McMacken J, Chamberlain S (1989) CHORD: A modular semiconductor device simulation development tool incorporating external network models. IEEE Trans Computer-Aided Design 8(8):826–836

    Article  Google Scholar 

  23. Rollins J, Choma J (1988) Mixed-mode PISCES-SPICE coupled circuit and device solver. IEEE Trans Computer-Aided Design 7:862–867

    Article  Google Scholar 

  24. Engl W, Laur R, Dirks H (1982) MEDUSA - a simulator for modular circuits. IEEE Trans Computer-Aided Design Int Circuits Syst 1(2):85–93

    Article  Google Scholar 

  25. Mayaram K, Pederson D (1992) Coupling algorithms for mixed-level ciruit and device simulation. IEEE Trans Computer-Aided Design 11(8):1003–1012

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Stephan Wagner.

Appendix

Appendix

As an example consider a simple ohmic boundary condition for the electron concentration n i = N 0 with N 0 being a constant value depending only on the doping of the semiconductor. To further simplify the example we only consider the static Poisson equation and the static continuity equation for electrons. Furthermore, the separate variable for the electron contact current \(I_{C_{n}}\) is omitted. For a one-dimensional device (index 0 refers to the incomplete point at the contact, 1 to the first complete point), the segment constitutive relations read

$$f_{\psi_{0}}^{\rm S} = \Psi_{01}+\hbox{q}\cdot n_{0}\cdot V_{0}\ne 0,$$
(41)
$$f_{\psi_{1}}^{\rm S} = \Psi_{10} + \Psi_{12} + \hbox{q}\cdot n_{1}\cdot V_{1} = 0,$$
(42)
$$f_{n_{0}}^{\rm S} = I_{01} \ne 0,$$
(43)
$$f_{n_{1}}^{\rm S} = I_{01} + I_{12} = 0,$$
(44)

with

$$\Psi_{i, j} = \Psi(\psi_{i},\psi_{j}) = -\Psi_{ji}$$
$$I_{i, j} = I(n_{i}, n_{j}, \psi_{i}, \psi_{j}) = -I_{ji}.$$

At the boundary, the constitutive relations are

$$f_{\psi_{0}} = \psi_{0} - \psi_{\rm C} = 0,$$
(45)
$$f_{n_{0}} = n_{0} - N_{0} = 0,$$
(46)
$$f_{I_{\rm C}} = I_{\rm C} + f_{n_{0}}^{\rm S} = 0,$$
(47)
$$f_{Q_{\rm C}} = Q_{\rm C} + f_{\psi_{0}}^{\rm S} = 0.$$
(48)

The boundary constitutive relations will be used to determine the quantity values at the boundary while the segment constitutive relations will be used to build up an expression for the boundary charge Q C and for the boundary current I C. This is done by the boundary models which set the appropriate entries in the transformation matrix T B . The solution vector x contains the following quantities

$$\mathbf{x} = \left( \psi_0, \psi_1, \ldots n_0, n_1, \ldots \psi_{\rm C}, I_{\rm C}, Q_{\rm C}, \ldots \right)^{\rm T}.$$
(49)

For voltage controlled contacts with V 0 applied to the contact one gets (50), when applying the current I 0 to the contact \(f_{\psi_{\rm C}}\) changes to (51).

$$f_{\psi_{\rm C}} = \psi_{\rm C} - V_{\rm 0} = 0,$$
(50)
$$f_{\psi_{\rm C}} = I_{\rm C} - I_{\rm 0} = 0.$$
(51)

The contact related entries in the three matrices (A b + T A s) are assembled as follows (the complete equations such as Eqs. 42 or 44 are assembled in A s only and have T b diagonal entries of 1):

A b

ψ0

n 0

I C

Q C

ψC

see

T b

ψ0

n 0

I C

Q C

ψC

ψ0

-1

0

0

0

0

(45)

ψ0

0

0

0

0

0

n 0

0

-1

0

0

0

(46)

n 0

0

0

0

0

0

I C

0

1

-1

0

0

(47)

I C

0

1

0

0

0

Q C

0

0

0

-1

0

(48)

Q C

1

0

0

0

0

ψC

0

0

-1

0

0

(51)

ψC

0

0

0

0

0

A b

ψ0

n 0

I C

Q C

ψC

see

transferred to

ψ0

−∂Ψ01/∂ψ0

-q·V0

0

0

0

(41)

Q C

n 0

− ∂I 01/∂ψ0

−∂I 01/∂n 0

0

0

0

(43)

I C

I C

0

0

0

0

0

  

Q C

0

0

0

0

0

  

ψC

0

0

0

0

0

  

The compiled linear equation system for iteration step k is

A

ψ0

n 0

ψC

I C

Q C

RHS

ψ0

−1

0

1

0

0

\(f^{k}_{\psi_{0}}\)

n 0

0

−1

0

0

0

\(f^{k}_{n_{0}}\)

ψC

0

0

0

−1

0

\(f^{k}_{\psi_{\rm C}}\)

I C

−∂I 01/∂ψ0

−∂I 01/∂n 0

0

−1

0

\(f^{k}_{I_{\rm C}}\)

Q C

−∂Ψ01/∂ψ0

−q·V0

0

0

−1

\(f^{k}_{Q_{\rm C}}\)

As the constitutive relations for the quantities ψ0, n 0, and I C are preeliminated one ends up with the following equation for ψC:

j x, y

ψ

RHS

ψC

−∂I 01/∂ψ0

\(f_{\psi_{\rm c}}^{k}-f_{I_{\rm C}}^{k}-f_{\psi_{0}}^{k}\cdot \partial I_{01}/\partial\psi_{0} + f_{n_{0}}^{k}\cdot \partial I_{01}/\partial n_{0}\)

Rights and permissions

Reprints and permissions

About this article

Cite this article

Wagner, S., Grasser, T., Fischer, C. et al. An advanced equation assembly module. Engineering with Computers 21, 151–163 (2005). https://doi.org/10.1007/s00366-005-0319-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00366-005-0319-5

Keywords

Navigation