Scribd Logo
Upload

Welcome to Scribd!

  • Density of States Function (3D, 2D, 1D, 0D)

    Uploaded by

    J.V.'S CLASS
    220 views32 pages
    Tutorial on Density of states.

    Original Title

    Density of States Function (3D,2D,1D,0D)

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document
    Tutorial on Density of states.

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    220 views32 pages

    Density of States Function (3D, 2D, 1D, 0D)

    Uploaded by

    J.V.'S CLASS
    Tutorial on Density of states.

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    Download as pdf
    of 32
    Ae con be, 4.3 FREE ELECTRONS IN A BOX _ tors are covalently bound..in most cases the bonding is an admixture of covalent and ionic bonding. How the bonding mechanism influences the electrical Properties of a Semiconductor is discussed in Chaptzr 6 Figure 4.11d shows the energy bands in semimetals. By semimetals we mean those elements or materials whiose conduction and valence bands overlap slightly. Such is the case with Bi, As, Sb, Se, Te, and some of their alloys with transition metals. Had it not been for the stight overlap of the energy bands, the valence band would be ‘completely filled and the conduction band completely empty at absolute zero tempera- ture. Because of the slight overlap, a small number of electrons go from the valence band into the conduction band, creating an equal number of free holes and free elec- trons, respectively, in the two bands. By free holes we refer to the unoccupied stated in the valence band; and by free electrons we refer to the occupied states in the conduction band. As mentioned in Section 3.7, graphite, Vb elements, and Vib elements crystallize in complicated stractures which have a primary covalent bonding and a weak metallic bonding. Therefore, the electrical properties of these materials can be highly aniso- tropic. They may exhibit semiconductor properties (.e., they have an energy gap) along 2 certain direction or in a certain plane, and show semimetal properties along 2 different direction, Finally, when there is a considerable overlapping between the bands, the material becomes a metal. The situation is illustrated in Fig. 4.11e. Take magnesium as an example. A free magnesium atom has the electron configuration of neon plus 3s? va- lence electrons. At first we might think thatthe valence electrons would completely fill the 3nband in a solid. However, as Fig. 4.6 shows, the 3p band overlaps the 3s band. AAs 2 Stl, a considerable nunter of electrons occupy states in the 3p band instead of filling the 3s band. Both the situation illustrated in Fig. 4.11 and that in Fig. 4,11¢ are common in metals. 3 In Sections 4.1 and 4.2 we presented a qualitative discussion of the formation of entgy Ms bands in a solid and explained the difference in the electrical behavior of mgt“(als en ledge, we Proceed to analyze the problem of electric conduction on a quantitative bésis, Again we start with the simplest case first, that is, the case of free electrons” As we recall our discussion in Sections 4.1 and 4.2, the central and crucial factors that influence electric conduction are the total number of available states and the distribution of the valence Glectrons among these available states. The concept state is purely a quantum. mechani. cal one, and hence the electcon must be treated as # quantum particle. For a free particle, the energy is entirely made ef the kinetic energy, which is equal to p"f2m. Letting V = 0 i: Eq. (1.3.6), we obtain the following wave equation for a free electron: _# (ty gay, ey) | ; oe tae) 43.1) Scanned with CamScanner Following a similar procedure as used in Section 1.4, we first substt it Bp. (63.1) and then divide bah sides by y. Thus Ey (631) tonto ag ? ) =% 433) a ia ‘The first term on the left-hand side of Eq. (4.3.3) depends only on x, the sina sie” term only on y, and the third term only on z. However, the sum of the th al a constant &. The on! isfy this condi Seeual sa qual constant. The cnly way to satiy this condition i to et each em equal oa { | : = hex : i & 3 < 8 ama 8% we : Similar equations are obtained for'Y and Z. The solution of Eq (43.4) 8 given by i X= Asin kx + B cos kr 43.5) where A and B are two constants to be determined later and k, = VinGiih. To help us see how the analysis above can be applied to conduction electrons a metal, we draw the -energy curves together ep oa is 11a The penile BIE SHEET ya ogee tema poe als due to each individual atom. The procedure is the same as that Followed in obtain: ing Fig> FW have a chain of atoms instead of just two atoms. Since: there are no nuclei to the right of the solid boundary indicated by the vertical dashed — lines, the potential energy does not turn downward but approaches a constant value. AS: a 5 wy ; a Figure 42 (a) Energy of valence lee- tron (:aded area) neat the surface of a = petal and (Ga idealized model (be fee edecron mode) ofa metal suas o i FreeElectron Theory Chap. 4 i 120 Scanned with CamScanner 3.2) into, (4.3.3) ‘ he second, erms must equal toa (43.4) ven by 43.5) joms. Since. ical dashed it value. As: of valence elec- be surface of a d model (ihe fee- 5 surface, oy Chap. 4 Fass. done 8 salt, there is a dierence betwegn the potential ently of an elect fal and that outside the metal, For the present discussion, let us ignore the effect of the nuclea the boundary of the solid. In other words, we assume that the pot JPraside the metal is uniform and chosen to be zero (i.e, chosen to be the energy). However, at the boundary, there is a jump in the potential ene amount Vo. Thus we may replace Fig. 4.12a by Fig. 4.12b. In both dif energy band is that of valence electrons (e.g., the 3s band in sodium). The energy betiveen electrons in different states is entirely due to the differe kinetic energy. This simplified treatment of valence electrons is known electron theory. Let us temporarily forget about the energy band and inquire as 6 what'd the energy ofan electron. Consider an electron in a potential well (Fig. 4.13: the potential well, the wave equation in one dimensi 3 For an electron having energy < Vo, the solutioi of Eq. (4.3.6) is X = Cexp(-ax) + D exp (ax) where a = [2m(Vo — )/A7]"2. ‘The complete solution of the potential-well problem is illustrated in #4 Inside the well, the solution is in the form of sine and cosine functions; and'6if well, the solution is in the form of exponential functions. At the boundary x = L), the two solutions must have their values and the values of their matched. However, we shall not go through with the mathematical analysiged we present a qualitative description of the situation. We know classically thal tron having energy less than the potential barrier Vp is a bound electro. zero at x = +00, only one term in Eq, (4.3.7) survives. For x > L we haves X = Cexp(-ax) As discussed in Section 4.7, the difference Vy — for electrons havit ‘most energy in the band (Fig. 4.12b) is called the work function ,,.. For’ sell ‘ Figure 4.13 Electrons (with energy ) in a potential well with a barter (3) of height Vo and (b) of Sec.43 Free Electrons ina Box Scanned with CamScanner with ()a finite barir and (b) an infinite baie. ee (0) Infinite Vp Figure 4.14 Electron wave functions (for q = 5 and q = 1) inside a potential well ny lue of %,, ranges from 1.93 eV in Cs to 5.36 eV in Pt. For our discussion, we take = & to be 2 eV. The value of a is found to be 7.25 x 10” cm”. In other words, value of X drops to 1/2.72 of its value at the boundary if x is away from the ai indary for a distance of 1.38 A. Thus, for all practical purposes, we may set X the: boundary, which is equivalent to assuming that Vo = © in Eq. (4.3.6). The ‘ation. is illustrated in Figs, 4.13b and 4.14b. For a one-dimensional potential well 9g sides of infinite heights, the solution is X = Asin (kx) 43.9) foe 0 L. The sine function is chosen so that X i Gillifles the boundary condition X = 0 at x = 0. The other boundary condition X= 0” L requires that ; kaon (4.3.10) a L q is a positive integer (0, 1, 2,. . .)- he discussion above can be extended to a three-dimensional potential well. In electron theory of metals, we consider electrons as being confined to. a three- ial ‘potential well having sides of infinite heights. As our calculation of « “assumption that V) = © and hence y = 0 outside the potential well-is a For simplicity, we consider a solid of cubic shape having dimensions DFige ‘lectron wave function inside the solid is given by 2 WO, y. 2) = A sin (k,x) sin (ky) sin (K,2) 0 functions are chosen to satisfy the boundary conditions at x dy 0 satisfy the condition y = 0 at the other sides of the boundary, that the, values of Ky: 2nd k; must be Such that eae Scanned with CamScanner I % i é gs ~~ & i y (4.3.12) ELECTRON MODEL AND DENSITY OF STATES restate the physical model on which the free-electron theory of metal is based. 1 asd °o Let us Ina metallic crystal, the constituent atoms consist of an ion core and one or more valence electrons. The valence electrons are weakly bound tothe ion core. [nt Od der approximation, that is, the free-electron approximation, we completely neglect the} + oy ferces between the valence eléttrons and the Ton ris doing so, we treat the valence # clectrons as if they were a free-lectron gas. These electrons are allowed to move about 2g freely inside the metal, but their motions are confined within. the ment These two physical conditions are inconporated in the mathematical formulation of jee-clectron Y= 0 inside the cube box and demanding = 0 outside the cube theory by setting box. 1 the orbital motion of an electron around a nucleus, we use trey yOPP ! quant umber, and m. AS we recall, beeae of the Ur depends of the 4 serena energy te soto ofthe Sehringer equation is separable ony ne spher- Potent mee stem (r, 6, ). The thre quantum numbers, 1, and comply : veel ay the dependence of the electron wave furton on r.@, Ine et of a fee Seen ope itis simplest o use te ectangular coordinates x,y 2). The sreges ney and gy are tree quantum numbers which specif the depend’ of the integers du ection on and z. Inthe free-electron theory, therefore eC of ection i specified by four quantum numbers: gg and sy which specly the a cr tnat motion ofthe electron, ands, which specifies the spinning moon of the "In specifying / electron. ; ir, ys pursue our analysis further to find the energy ofan eleton. The solution of Eq, (4.3.11) has the following property: i fy aw (4. ins. ee em Gt hh. a 123 Free-Electron Model and Density of States er Scanned with CamScanner Sec. 4.4 Sing Aa (TIED A ae : / oes #4 a fe % wanlae* aft ae) Tan TB HENRY 42) Equation (4.4.2) is satisfied only if & = a we : Rete O-E() deded (443) ‘The las step in Bq, (44.3) is artved at through the use of Ea. (4.3.12). , In the classical theory, the kinetic energy of a particle is given by pm. Fora. oF fee pale wi V = 0, te tol energy i ently made othe Hl ie i ence ‘ 2 texee2 ot KE=s pm t+ PD 44.4) where ps. Py. and p, are the components of the electron momentum p. Comparing Eq. (4.4.3) with Eq. (4.4.4), we see that the momentum of an electron in the quantum theory is equal to p= tk 4.5) ‘and that the kk, and k, in Eq. (4.4.3) represent the three components of such that PoktR+8 (4.4.6) \Yy Since p is a vector, so must be k. We further note thatthe spatial variation of y shown in Fig. 4.14b is similar to that of E of an electromagnetic wave shown in Fig. 2.20c. In our present discussion, ‘we have not considered the time dependence of . If we include the time variation of {W in the solution of the Schrédinger equation, we indeed obtain a wavelike solution similar to that of Eq. (2.7.1). In other words, the electrons in a potential well have wayelike motions. Referring to Fig. 4.14b, we see that for g = 1, the wavelength X of the wave motion is equal to 2L. Since k= iL for g = 1, we have ; Aaa 447 From Eq. (4.4.3), 2m & = AR. In terms of 6, Eq. (4.4.7) becomes i: 4 Vint p (4.4.8) Equation (4.4.8) is the celebrated de Bogle relatioh expressed in Eq. (2.7.5), whose validity has been proved inthe electron diffraction experiment of Davisson and Germer. nr One function that is essential in the calculation of the electrical property of an | electric conductor is the density of state D(€). To find this function, we must Jearn how to count the number of proper modes of electron motion. First let us consider a two- dimensional case with q; = 0 and draw a two-dimensional lattice as shown in Fig. 4,16a. Each lattice point cepresents a definite set of integers for the values of q, and qa, which means a specitfe normai mode of oscillation, Therefore, to find the number off 4 Free-Electron Theory Chap. 4 Qf Gor WN, ¢ SU WW bor s. ae yumb trated ys wen aa’) ayn? wy 326 on any Scanned with CamScanner Qa” , > Figure 4.16 (a) Diagram used in ‘counting the normal Qex two-ineaon box. Each lattice point represents a set of isnre inne (43.127 With q, = 0. The numberor ni i a 4 , | modes im a rectangula ir z AY F "HD pa 101 tal mune tie poised ye os oe A three-dimensional k space that is obtained fr A three- a : from the two-dimensional by adds 5 § «y ty ng a third dimension and by converting g to kuhrough te use of Ea. ai, ne . apy formal modes inside a rectangular box with dimension AqAqs, we simpy ¢6 AZ umber of lattice points inside the box. If we choose Aqy = 2 and Ags = 2g trated in Fig. 4.16a, the number of normal modes is four, This is exaaly equal 3D a yy ris of the box. For a three-dimensional box, the number of normal modes AN’ Xt hw uy! AN’ = Aq, Age Ags Let us now estimate the value of q in Eq. (4.4.3). For an electron having 3.2 eV (topmost energy in Na as illustrated in Fig. 3.11), we find from Eq. (4, value of A = 6.9 x 107* cm. From Eqs. (4.4.7) and (4.3.10), the value calculated to be q = 2.9 X 10” for L = 1 cm. In other words, the vali of q cf any integer from 0 to 2.9 x 10’. For such a large valu¢ of g, we can treat the d variables g1, g2, and q3 as if they were continuous, and replace Eq (4.4.9) by aN’ = dgy dp dg al nt energy surface in the q space is 8p having energy &
  • Eo, the 2D density of states is a constant and does not depend on energy. If the 2D semiconductor has more than one quantum state, each quantum state has a state density of Eq. (12.20). The total density of states can be written as pr s(B) =r (E-E,) (aay « BS el net i a where E, are the energies of quantized states and o(E ~ Ey) is the step function. @D . wR] of) y 4 E=const. wah E=gonst oe dx ey a I coal onto) Fig. 12.6. Constant energy surfaces & aia) 3eaimensional, (0) 2-dimenson ce © dim ional system. The surfaces are a sph re, a circle, and a point for 3D, 2D, and 1D jimensi - ystems, respectively. the quantum wire, in which only one ae aln Jation is then given We next consider a 2 ee a DE aE i iin i wed, re ae a it) in k-space direction of mo! a aoe svolunte” (. = WP ke /(2m*). The “Vo ye 3 \ ©. F. Schubert a yy NM ch Scanned with CamScanner > “Mintroduction The density of states function describes the number of states that are available arrier concentratioris and. jin a system and is essential for determining the © energy distributions of carriers within a semiconductor. one, and zero systems of quantum wires In semiconductors, the free motion of carriers is limited to two, plying, semiconductor statistics to ‘spatial dimensions. When ap) yf states in quantum: wells (2D), ese dimensions, the density 0: nd quantum dots (OD) must be known. a + tantinita of Technology : Scanned with CamScanner Recalling from the density of states 3D derivation... k-space volume of single state cube in k-space: rom (HELE-( V is the volume of the crystal. Vinge i8 the smallest unit in k-space and is required to hold a single electron. Scanned with CamScanner Derivation of Density of States (2D) Recalling from the density of states 3D-derivation... 6 apace volume of single state cube in k-space: Vanm*(2\2(2)-(2}-(5) espace volume of sphere in k-space: Vee = 2 Sumber of filled states in sphere: wa tnesau{t 1 1) Fxdxt Sin gie-ne / ee A factor of two is added (Correction factor foe to account for the two redundancy in counting possible electron spins identical states +1-n,,/- of each solution. a thn, Scanned with CamScanner — egg e STE? Derivation of Density of States (2D) For calculating the density of states for a 2D structure (i.e. quantum well), we can use a similar approach, the Previous equations change to the following: k-space volume of single state cube in k-space: , -(2(- }-(@) ‘nase “(215 =| >| k-space volume of sphere in k-space: Vovrte = Number of filled states in a sphere: -ecmnnsice ve aK i Zo My w= tase —xax{ x3] ow “in gestae ap ae (1) He = x2x{t Qn Scanned with CamScanner Derivation of Density of States (2D) continued...... PP 2mE Substituting k= yields I Qn Wn The density per unit energy is then obtained using the chain rule: dN _ aN dle _ Lim dE dk dE mv Scanned with CamScanner Derivation of Density of States (2D) ~ F . _ The density of states per unit volume, per unit energy 1S found by dividing by V (volume of the crystal). &(E)zp becomes: Lm 2 m 8(B)ap. =e As stated initially for the electron mass, m— m’. Thus, Itis significant that the 2D density of states does not depend on energy. Immediately, as the top of the energy-gap is reached, there is 2 significant number of available states, Scanned with CamScanner Derivation of Density of States (AD) For calculating the density of states for a 1D strucnee Gerquannen wire), we San use a similar approach. The previous equations change to the following: E=constant —., n n n k-space volume of single state cube in k-space: Vin gle-sate = a) wh) L k-space volume of sphere in k-space: Vine = K v, a Ete Hinats seh nee x x(t sin gle —stave Number of filled states ina sphere: 2 KL a Scanned with CamScanner - Derivation of Density of States (1D) Continued..... vo Le hn Rearranging.....- jy =(2mE) ‘The density per unit energy is then obtained by using the chain rule: 1 12 an _ dN de_22m®) 2m (2mB)" «mL dE dk dE nn ha ECE 6451 Georgia Institute of Technology Scanned with CamScanner sy Derivation of Density of States (AD) The density of states per unit volume, per unit energy is found by dividing \ by V (volume of the crystal). 9(E);p becomes: Simplifying yields... ECE 6451 m © heQmE am vm m 80 Fe Pome 1 g(Ehv-Fg \2E Georgia Institute of Technology Scanned with CamScanner \ Derivation of Density of States (1D) As stated initially’for the electron mass, m—+m*. Also, because only kinetic energy is considered E—+Ec. Scanned with CamScanner exe the behavior OL Ie i Derivation of Density of States (0D) When considering the density of states for a 0D structure (i.c. quantum dot), no free motion is possible. Because there is no k-space to be filled with, electrons and all available states exist only at discrete energies, we describe the density of states for 0D with the delta function. Thus, g(E)op = 25(E-E.) - Georgia Institute of Technolosy Scanned with CamScanner Additional Comments The density of states has a functional dependence on energy- Degrees of freedom Density of states Poos| 3

    You might also like