0% found this document useful (0 votes)
236 views13 pages

Notes On FD Integral3rdEd Revised 080411

This document provides an overview of Fermi-Dirac integrals, which commonly arise when computing properties of semiconductors. It defines the general form of the Fermi-Dirac integral Fj(ηF) and discusses specific examples that appear in calculating the electron density of one-dimensional, two-dimensional, and three-dimensional semiconductors. The document also covers the derivatives, asymptotic behavior, and approximate expressions for common Fermi-Dirac integrals. Matlab functions for numerically evaluating the integrals are referenced.

Uploaded by

Neel Atiq
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
236 views13 pages

Notes On FD Integral3rdEd Revised 080411

This document provides an overview of Fermi-Dirac integrals, which commonly arise when computing properties of semiconductors. It defines the general form of the Fermi-Dirac integral Fj(ηF) and discusses specific examples that appear in calculating the electron density of one-dimensional, two-dimensional, and three-dimensional semiconductors. The document also covers the derivatives, asymptotic behavior, and approximate expressions for common Fermi-Dirac integrals. Matlab functions for numerically evaluating the integrals are referenced.

Uploaded by

Neel Atiq
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Notes on Fermi-Dirac Integrals 3rd Edition

Raseong Kim and Mark Lundstrom Network for Computational Nanotechnology Purdue University December 10, 2008 (Last revised on August 4, 2011)

1. Introduction Fermi-Dirac integrals appear frequently in semiconductor problems, so a basic understanding of their properties is essential. The purpose of these notes is to collect in one place, some basic information about Fermi-Dirac integrals and their properties. We also present Matlab functions (see Appendix and [1]) that calculate Fermi-Dirac integrals (the script F defined by Dingle [2] and reviewed by Blakemore [3]) in three different ways. To see how they arise, consider computing the equilibrium electron concentration per unit volume in a three-dimensional (3D) semiconductor with a parabolic conduction band from the expression,

n=

EC

g(E) f (E)dE = 1 + e(
0 EC

g(E)dE
E EF ) k B T

(1)

where g(E) is the density of states, f0 (E) is the Fermi function, and EC is the conduction band edge. For 3D electrons with a parabolic band structure,

g3D

( 2m ) (E) =

* 3/ 2

2 2 3

E EC ,

(2)

which can be used in Eq. (1) to write

(2m ) n=
2

* 3/ 2 2 3

EC

1 + e(

E EC dE

E EF ) k B T

(3)

By making the substitution,

= (E EC ) kBT ,
Eq. (3) becomes

(4)

(2m k T ) n=
* B

3/ 2

2 2 3

1 + e
0

1/ 2 d
F

(5)

where we have defined

F (EF EC ) k BT .
By collecting up parameters, we can express the electron concentration as n = N 3D where 2 F1/2 ( F ) ,

(6)

(7)

N 3D

2 m*k BT = 2 h2

3/ 2

(8)

is the so-called effective density-of-states and

F1/ 2 ( F )

1/ 2 d 1 + exp( F ) 0

(9)

is the Fermi-Dirac integral of order 1/2. This integral can only be evaluated numerically. Note that its value depends on F , which measures the location of the Fermi level with respect to the conduction band edge. It is more convenient to define a related integral,

F 1/ 2 ( F )

1/ 2 d , 1 + exp( F ) 0

(10)

so that Eq. (7) can be written as


n = N 3DF 1/2 ( F ) .

(11)

It is important to recognize whether you are dealing with the Roman Fermi-Dirac integral or the script Fermi-Dirac integral. There are many kinds of Fermi-Dirac integrals. For example, in two dimensional (2D) semiconductors with a single parabolic band, the density-of-states is m* g 2D ( E ) = 2 , (12)

and by following a procedure like that one we used in three dimensions, one can show that the electron density per unit area is
nS = N 2DF 0 ( F ) ,

(13)

where N 2D = and m * k BT , 2 (14)

F 0 ( F ) =
0

0d
1+ e
F

= ln (1 + eF )

(15)

is the Fermi-Dirac integral of order 0, which can be integrated analytically. Finally, in one-dimensional (1D) semiconductors with a parabolic band, the density-of-states is 2 m* g1D ( E ) = 1 , E EC (16)

and the equilibrium electron density per unit length is


nL = N1DF 1/2 ( F ) ,

(17)

where N1D = and 1 2 m* k B T , (18)

F 1/ 2 ( F ) =

1 + e 0

1/ 2 d
F

(19)

is the Fermi-Dirac integral of order 1/ 2 , which must be integrated numerically.

2. General Definition
In the previous section, we saw three examples of Fermi-Dirac integrals. More generally, we define 3

1 j d F j ( F ) , ( j + 1) 1 + exp( F ) 0

(20)

where is the gamma function. The function is just the factorial when its argument is a positive integer,
( n) = (n 1)! (for n a positive integer) .

(21a)

Also

(1 / 2) = ,
and ( p + 1) = p( p) . As an example, lets evaluate F 1/ 2 ( F ) from Eq. (20):

(21b)

(21c)

F 1/ 2 ( F )

1 1/ 2 d , (1/ 2 + 1) 1 + e F 0

(22a)

so we need to evaluate (3 / 2) . Using Eqs. (21b-c), we find,

1 (3 / 2) = (1/ 2 + 1) = (1/ 2) = , 2 2
so F 1/ 2 ( F ) is evaluated as

(22b)

F 1/ 2 ( F )

1 + e 0

1/ 2 d
F

(22c)

which agrees with Eq. (10). For more practice, use the general definition, Eq. (20) and Eqs. (21ac) to show that the results for F 0 ( F ) and F 1/ 2 ( F ) agree with Eqs. (15) and (19).

3. Derivatives of Fermi-Dirac Integrals


Fermi-Dirac integrals have the property that dF j d F = F j 1 , 4 (23)

which often comes in useful. For example, we have an analytical expression for F 0 ( F ) , which means that we have an analytical expression for F 1 ( F ) ,

F 1 =

dF 0 1 = . d F 1 + e F

(24)

Similarly, we can show that there is an analytic expression for any Fermi-Dirac integral of integer order, j, for j 2 ,

F j ( F ) =

e F

1+ e

P j 2 e F ,

( )

(25)

where Pk is a polynomial of degree k, and the coefficients pk ,i are generated from a recurrence relation [4] (note that the relation in Eq. (26c) is missing in p. 222 of [4])

pk ,0 = 1 ,
pk ,i = (1 + i ) pk 1,i ( k + 1 i ) pk 1,i 1 i = 1, , k 1 ,

(26a) (26b) (26c)

pk , k = pk 1,k 1 .
For example, to evaluate F 4 ( F ) = eF generated from Eqs. (26a-c) as [4]

(1 + e )
F

P2 ( eF ) , polynomial coefficients are

p0,0 = 1 , p1,0 = 1, p2,0 = 1,


and we find

p1,1 = p0,0 = 1 , p2,1 = 2 p1,1 2 p1,0 = 4, p2,2 = p1,1 = 1 ,

(27)

F 4 ( F ) =

e F

(1 + e )
F

p2,i eiF =
i=0

eF

(1 + e )
F

(1 4e

+ e 2F .

(28)

4. Asymptotic Expansions for Fermi-Dirac Integrals


It is useful to examine Fermi-Dirac integrals in the non-degenerate ( F << 0 ) and degenerate ( F >> 0 ) limits. For the non-degenerate limit, the result is particularly simple,

F j ( F ) e ,
F

(29)

which means that for all orders, j, the Fermi-Dirac integral approaches the exponential in the non-degenerate limit. To examine Fermi-Dirac integrals in the degenerate limit, we consider the complete expansion for the Fermi-Dirac integral for j > 1 and F > 0 [2, 5, 6]

F j ( F ) = 2

j +1 F

( 1) e nF t2 n ( j + 2 2n ) 2n + cos ( j ) n j +1 , n =0 n =1 F
n 1

(30)

where t0 = 1 2 , tn = =1 (1)

n = ( 21 n ) (n ) , and (n ) is the Riemann zeta function. 1

The expressions for the Fermi-Dirac integrals in the degenerate limit ( F >> 0 ) come from Eq. (30) as F j ( F ) Fj +1 ( j + 2 ) [7]. Specific results for several Fermi-Dirac integrals are shown below.

F 1/ 2 ( F ) F 1/ 2 ( F )
1 2

2 1/ 2 F

(31a) (31b) (31c) (31d) (31e)

3/ 4 F 2 , 3

2 F 1 ( F ) F ,
5 8 F / 2 , F 3/ 2 ( F ) 15 1 3 F 2 ( F ) F . 6

The complete expansion in Eq. (30) can be related to the well-known Sommerfeld expansion [8, 9]. First, note that the integrals to calculate carrier densities in Eqs. (1) and (3) are all of the form

H ( E ) f ( E ) dE .
0

(32)

Taylor expansion of H ( E ) about EF as [9]

If H ( E ) does not vary rapidly in the range of a few k BT about EF , then we can write the

( E EF ) . dn H ( E ) = n H ( E ) E =E F n! n = 0 dE
n

(33)

Using this Taylor series expansion, the integral in Eq. (32) can be written as (see [9] for a detailed derivation)

H ( E ) f 0 ( E ) dE =

EF

H ( E ) dE + ( k BT ) an
2n n =1

d 2 n 1 H (E) , dE 2 n 1 E = EF

(34)

where

1 1 1 an = 2 1 2 n + 2 n 2 n + , 3 4 2

(35)

and it is noted that an = 2t2 n . Equation (34) is known as the Sommerfeld expansion [8, 9]. Typically, the first term in the sum in Eq. (34) is all that is needed, and the result is

H ( E ) f 0 ( E ) dE

EF

H ( E ) dE +

2
6

( kBT )

H ( EF ) .

(36)

If we scale E by k BT in Eq. (34), E k BT , then Eq. (34) becomes

H ( ) f0 ( ) d =

H ( ) d + an
n =1

d 2 n 1 H ( ) . d 2 n 1 = F

(37)

Then the Sommerfeld expansion for the Fermi-Dirac integral of order j can be evaluated by letting H ( ) = j ( j + 1) in Eq. (37), and the result is

F j ( F ) = 2 Fj +1

t2 n . 2n n = 0 ( j + 2 2 n ) F

(38)

Equation (38) is the same as Eq. (30) except that the second term in Eq. (30) is omitted [5]. In the degenerate limit, however, the second term in Eq. (30) vanishes, so the Eqs. (30) and (38) give the same results as Eqs. (31a-e).

5. Approximate Expressions for Common Fermi-Dirac Integrals


Fermi-Dirac integrals can be quickly evaluated by tabulation [2, 7, 10, 11] or analytic approximation [12-14]. We briefly mention some of the analytic approximations and refer the reader to a Matlab function. Bednarczyk et al. [12] proposed a single analytic approximation that evaluates the Fermi-Dirac integral of order j = 1/ 2 with errors less than 0.4 % [3]. AymerichHumet et al. [13, 14] introduced an analytic approximation for a general j, and it gives an error of 1.2 % for 1/ 2 < j < 1/ 2 and 0.7 % for 1/ 2 < j < 5 / 2 , and the error increases with larger j. The Matlab fuction, FD_int_approx.m, [1] calculates the Fermi-Dirac integral defined in Eq. (10) with orders j 1/ 2 using these analytic approximations. The source code of this relatively short function is listed in the Appendix. 7

If a better accuracy is required and a longer CPU time is allowed, then the approximations proposed by Halen and Pulfrey [15, 16] may be used. In this model, several approximate expressions are introduced based on the series expansion in Eq. (30), and the error is less than 10-5 for 1/ 2 j 7 / 2 [15]. The Matlab function, FDjx.m, [1] is the main function that calculates the Fermi-Dirac integrals using this model. This function includes tables of coefficients, so it is not simple enough to be shown in the Appendix, but it can be downloaded from [1]. There also have been discussions on the simple analytic calculation of the inverse FermiDirac integrals of order j = 1/ 2 [3]. This has been of particular interest because it can be used to calculate the Fermi level from the known bulk charge density in Eq. (11), as F = F 1/ 21 ( n N 3 D ) . Joyce and Dixon [17] examined a series approach that gives F 0.01 for F max 5.5 [3], and a simpler expression from Joyce [18] gives F 0.03 for F max 5 [3]. Nilsson proposed two different full-range ( 10 F 20 ) expressions [19] with F 0.01 and F 0.005 [3]. Nilsson later presented two empirical approximations [20] that give F 0.01 for

F max 5.5 and F max 20 , respectively [3].

6. Numerical Evaluation of Fermi-Dirac Integrals


Fermi-Dirac integrals can be evaluated accurately by numerical integration. Here we briefly review the approach by Press et al. for generalized Fermi-Dirac integrals with order j > 1 [21]. used for F 15 , and the double exponential (DE) rule is used for larger F . Double precision (eps, ~ 2.2 1016 ) can be achieved after 60 to 500 iterations [21]. The Matlab function, FD_int_num.m, [1] evaluates the Fermi-Dirac integral numerically using the composite trapezoidal rule following the approach in [21]. The source code is listed in the Appendix. This approach provides very high accuracy, but the CPU time is considerably longer. An online simulation tool that calculates the Fermi-Dirac integrals using this source code has been deployed at nanoHUB.org [22]. Note that the numerical approach we consider in this note is relatively simple, and there are other advanced numerical integration algorithms [23] suggested to improve the calculation speed. In Fig. 1, we compare the accuracy and the timing of the three approaches that calculate F j ( F ) . The Fermi-Dirac integral of order j = 1 2 ( F 1/ 2 ( F ) ) is calculated for 10 F 10 with F spacing = 0.01 using approximate expressions (FD_int_approx.m and FDjx.m) and the rigorous numerical integration (FD_int_num.m) with double-precision. The relative errors of the approximate expressions are calculated as (F 1/ 2,approx F 1/ 2,num ) F 1/ 2,num , where F 1/ 2,approx and F 1/ 2,num represent the results from the approximate expression and the numerical integration respectively. The elapsed time measured for each approach (using Matlab commands tic/toc for Pentium 4 CPU 3.4 GHz and 2.0 GB RAM) clearly shows the compromise between the accuracy and the CPU time. 8

In this approach, the composite trapezoidal rule with variable transformation = exp (t e t ) is

Fig. 1. (a) Relative errors from the approximate expressions for F 1/ 2 ( F ) with respect to the numerical integration (FD_int_num.m). (A) Relative error from FD_int_approx.m. (B) Relative error from FDjx.m. All Matlab functions are available in [1].(b) The absolute values of the relative errors in the log scale. The elapsed time measured for the three approaches clearly shows the trade-off between the accuracy and the CPU time.

References
[1] [2] [3] R. Kim and M. S. Lundstrom (2008), "Notes on Fermi-Dirac Integrals (3rd Edition)," Available: https://www.nanohub.org/resources/5475/. R. Dingle, "The Fermi-Dirac integrals F p ( ) = ( p !)
1

p ( e + 1) d ," Applied
1

[4] [5] [6]

Fermi--Dirac integrals F p ( ) = 1 ( p + 1) p (1 + e ) d ," J. Math. Phys., vol. 42,

Scientific Research, vol. 6, no. 1, pp. 225-239, 1957. J. S. Blakemore, "Approximations for Fermi-Dirac integrals, especially the function F 1/ 2 ( ) used to describe electron density in a semiconductor," Solid-State Electron., vol. 25, no. 11, pp. 1067-1076, 1982. M. Goano, "Algorithm 745: computation of the complete and incomplete Fermi-Dirac integral," ACM Trans. Math. Softw., vol. 21, no. 3, pp. 221-232, 1995. R. B. Dingle, Asymptotic Expansions: Their Derivation and Interpretation. London: Academic Press, 1973. T. M. Garoni, N. E. Frankel, and M. L. Glasser, "Complete asymptotic expansions of the
0

[7]

[8] [9] [10] [11]

[12] [13]

[14] [15]

[16]

[17]

no. 4, pp. 1860-1868, 2001. J. McDougall and E. C. Stoner, "The computation of Fermi-Dirac functions," Philosophical Transactions of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol. 237, no. 773, pp. 67-104, 1938. A. Sommerfeld, "Zur Elektronentheorie der Metalle auf Grund der Fermischen Statistik," Zeitschrift fr Physik A Hadrons and Nuclei, vol. 47, no. 1, pp. 1-32, 1928. N. W. Ashcroft and N. D. Mermin, Solid State Physics. Philadelphia: Saunders College Publishing, 1976. A. C. Beer, M. N. Chase, and P. F. Choquard, "Extension of McDougall-Stoner tables of the Fermi-Dirac functions," Helvetica Physica Acta, vol. 28, pp. 529-542, 1955. P. Rhodes, "Fermi-Dirac functions of integral order," Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, vol. 204, no. 1078, pp. 396-405, 1950. D. Bednarczyk and J. Bednarczyk, "The approximation of the Fermi-Dirac integral F 1/ 2 ( ) ," Physics Letters A, vol. 64, no. 4, pp. 409-410, 1978. X. Aymerich-Humet, F. Serra-Mestres, and J. Milln, "An analytical approximation for the Fermi-Dirac integral F3/ 2 ( ) ," Solid-State Electron., vol. 24, no. 10, pp. 981-982, 1981. X. Aymerich-Humet, F. Serra-Mestres, and J. Millan, "A generalized approximation of the Fermi--Dirac integrals," J. Appl. Phys., vol. 54, no. 5, pp. 2850-2851, 1983. P. V. Halen and D. L. Pulfrey, "Accurate, short series approximations to Fermi-Dirac integrals of order -1/2, 1/2, 1, 3/2, 2, 5/2, 3, and 7/2," J. Appl. Phys., vol. 57, no. 12, pp. 5271-5274, 1985. P. Van Halen and D. L. Pulfrey, "Erratum: "Accurate, short series approximation to Fermi-Dirac integrals of order -1/2, 1/2, 1, 3/2, 2, 5/2, 3, and 7/2" [J. Appl. Phys. 57, 5271 (1985)]," J. Appl. Phys., vol. 59, no. 6, p. 2264, 1986. W. B. Joyce and R. W. Dixon, "Analytic approximations for the Fermi energy of an ideal Fermi gas," Appl. Phys. Lett., vol. 31, no. 5, pp. 354-356, 1977. 10

[18] [19] [20] [21] [22] [23]

W. B. Joyce, "Analytic approximations for the Fermi energy in (Al,Ga)As," Appl. Phys. Lett., vol. 32, no. 10, pp. 680-681, 1978. N. G. Nilsson, "An accurate approximation of the generalized Einstein relation for degenerate semiconductors," Phys. Stat. Solidi (a), vol. 19, pp. K75-K78, 1973. N. G. Nilsson, "Empirical approximations for the Fermi energy in a semiconductor with parabolic bands," Appl. Phys. Lett., vol. 33, no. 7, pp. 653-654, 1978. W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Numerical Recipes: The Art of Scientific Computing, 3rd ed. New York: Cambridge University Press, 2007. X. Sun, M. Lundstrom, and R. Kim (2011), "FD integral calculator," Available: http://nanohub.org/tools/fdical. A. Natarajan and N. Mohankumar, "An accurate method for the generalized Fermi-Dirac integral," Comput. Phys. Commun., vol. 137, no. 3, pp. 361-365, 2001.

11

Appendix
FD_int_approx.m
function y = FD_int_approx( eta, j ) % Analytic approximations for Fermi-Dirac integrals of order j > -1/2 % Date: September 29, 2008 % Author: Raseong Kim (Purdue University) % % Inputs % eta: eta_F % j: FD integral order % % Outputs % y: value of FD integral (the "script F" defined by Blakemore (1982)) % % For more information in Fermi-Dirac integrals, see: % "Notes on Fermi-Dirac Integrals (3rd Edition)" by Raseong Kim and Mark % Lundstrom at http://nanohub.org/resources/5475 % % References % [1]D. Bednarczyk and J. Bednarczyk, Phys. Lett. A, 64, 409 (1978) % [2]J. S. Blakemore, Solid-St. Electron, 25, 1067 (1982) % [3]X. Aymerich-Humet, F. Serra-Mestres, and J. Millan, Solid-St. Electron, 24, 981 (1981) % [4]X. Aymerich-Humet, F. Serra-Mestres, and J. Millan, J. Appl. Phys., 54, 2850 (1983) if j < -1/2 error( 'The order should be equal to or larger than -1/2.') else x = eta; switch j case 0 y = log( 1 + exp( x ) ); % analytic expression case 1/2 % Model proposed in [1] % Expressions from eqs. (22)-(24) of [2] mu = x .^ 4 + 50 + 33.6 * x .* ( 1 - 0.68 * exp( -0.17 * ( x + 1 ) .^ 2 ) ); xi = 3 * sqrt( pi ) ./ ( 4 * mu .^ ( 3 / 8 ) ); y = ( exp( - x ) + xi ) .^ -1; case 3/2 % Model proposed in [3] % Expressions from eq. (5) of [3] % The integral is divided by gamma( j + 1 ) to make it consistent with [1] and [2]. a = 14.9; b = 2.64; c = 9 / 4; y = ( ( j + 1 ) * 2 ^ ( j + 1 ) ./ ( b + x + ( abs( x - b ) .^ c + a ) .^ ( 1 / c ) ) .^ ( j + 1 ) ... + exp( -x ) ./ gamma( j + 1 ) ) .^ -1 ./ gamma( j + 1 ); otherwise % Model proposed in [4] % Expressions from eqs. (6)-(7) of [4] % The integral is divided by gamma( j + 1 ) to make it consistent with [1] and [2]. a = ( 1 + 15 / 4 * ( j + 1 ) + 1 / 40 * ( j + 1 ) ^ 2 ) ^ ( 1 / 2 ); b = 1.8 + 0.61 * j; c = 2 + ( 2 - sqrt( 2 ) ) * 2 ^ ( - j ); y = ( ( j + 1 ) * 2 ^ ( j + 1 ) ./ ( b + x + ( abs( x - b ) .^ c + a ^ c ) .^ ( 1 / c ) ) .^ ( j + 1 ) ... + exp( -x ) ./ gamma( j + 1 ) ) .^ -1 ./ gamma( j + 1 ); end end

12

FD_int_num.m
function [ y N err ] = FD_int_num( eta, j, tol, Nmax ) % Numerical integration of Fermi-Dirac integrals for order j > -1. % Author: Raseong Kim (Purdue University) % Date: September 29, 208 % Extended (composite) trapezoidal quadrature rule with variable % transformation, x = exp( t - exp( t ) ) % Valid for eta ~< 15 with precision ~eps with 60~500 evaluations. % % Inputs % eta: eta_F % j: FD integral order % tol: tolerance % Nmax: number of iterations limit % % Note: When "eta" is an array, this function should be executed % repeatedly for each component. % % Outputs % y: value of FD integral (the "script F" defined by Blakemore (1982)) % N: number of iterations % err: error % % For more information in Fermi-Dirac integrals, see: % "Notes on Fermi-Dirac Integrals (3rd Edition)" by Raseong Kim and Mark % Lundstrom at http://nanohub.org/resources/5475 % % Reference % [1] W. H. Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, % Numerical recipies: The art of scientific computing, 3rd Ed., Cambridge % University Press, 2007. for N = 1 : Nmax a = -4.5; % limits for t b = 5.0; t = linspace( a, b, N + 1 ); % generate intervals x = exp( t - exp( -t ) ); f = x .* ( 1 + exp( -t ) ) .* x .^ j ./ ( 1+ exp( x - eta ) ); y = trapz( t, f ); if N > 1 % test for convergence err = abs( y - y_old ); if err < tol break; end end y_old = y; end if N == Nmax error( 'Increase the maximum number of iterations.') end y = y ./ gamma( j + 1 );

13

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy