Scribd
Upload
40 views24 pages

SSP Unit 2 Part 1

Solid state physics

Uploaded by

ASWIN S
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
40 views24 pages

SSP Unit 2 Part 1

Solid state physics

Uploaded by

ASWIN S
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 24
V 11.1 BLOCH THEOREM In the case of free electron theory we assumed that the electrons move in a region of Com potential inside a one or three dimensional potential well. Although the free electron theory is able to explain the phenomena such as electrical conductivity, thermionic emission ete, to explain why some materials are good electrical conductors while some are good electrical insulators and still some others are semiconductors? ¥ In order to understand the difference between the conductors and insulators, it is rece incorporate the variation of potential inside the crystal due to the presence of positive ion cores in the free electron model (Fig. 11.1a). It appears more realistic to assume the potential inside a metallic crystal to be periodic with the periodicity of the lattice, somewhat as shown in Fi 11.1b. The potential is minimum at the positive ion sites and maximum between the two! The corresponding one dimensional Schrodinger equation can be written as: St Vee BODO Electron motion DO OO > Surface potential NAD nuclei b) Fig. 11.1 One dimensional Periodic potential in a crystal the periodic potential, ba icing proving Bk geet equation takes the form ch has shown that the one dimensional sol ae YO) = exp (ikx) uy(x) see dimensions itis given by VAG) = exp (ik.r) yr) (4) spe equation 3 (or 4) is known the free electron is periodic vi iodic function u( where or (7) is periodic with the Periodicity of the lattice j on u(x) or ul), w Trerefore, considering only the one dimens; ice in one and three dimensions, respectively. gfaoms in linear chain of atoms of length L, then we oe write (2) = ya + Na) 6) anddepends on the exact nature of the potential field, From eqs. 3 and 5, we have Yule + Na) = uy(x + Na) exp [ik(x + Na)] = exp (ikNa) u(x) explikx) = yx) exp (ikNa) © Thisis frequently referred to as the Bloch condition. Similarly, the complex conjugate of eq. 6 canbe written as wile + Na) = exp(-ikNa) yi (x) @ Theeqs. 6 and 7 give us yal + Na) yaa + Na) = Yi) Wi) @) Thisindicates that the probability of finding the electron is same everywhere in the whole chain probability of : ; . Sf atoms, ie. itis not localized around any particular atom but is shared by all atoms in the chain Whole crystal in three dimensions). Thus, eq. 6, gives us z exp (ikNa) = 1 kNa = 27x integer = 20 2nn _ 2mn salsa atoms, When 1 360 Solid State Physics : This is equal to the total number of atoms in the chain of atoms (or in the unit : dimensions). 11.2 THE KRONIG-PENNEY MODEL V In the preceding section, we qualitatively discussed periodic potential. However, in order to know yx the exact E-k relationship it is necessary to consider a well defined one. dimensional periodic { potential. For the purpose, we shall consider Kronig-Penney model. They suggested asimpler Ye potential in the form of an array of square wells nyt as shown in Fig. 11.2. ‘The corresponding Schrodinger equations for the two regions I and II are of the form the behaviour of an electron in an dy , 81 peeeeeely 70 O | as b + 0 baw Hence, eq. 27 becomes are 2ap cos ka = Bb sin aa + cos oa (28) Beant pete nee sain cy, -28) where B-a@ Since, Vo >> E, so that a B?-a?= serve Substituting this in eq. 28, we obtain cos ka = aa 2 in a = An? may,4 9 5 cos aa a SINS cowie @) ee 4 4x? ma (0) where naV,b hand side is. jon 29 is schematically represented in Fig. 11.3. The quantity on its right aes a function of a. The cosine term on the left hand side of the equation can only B Fpin ca + 60s oa oa cae tion 38 only certain y, Aled fy jot ales » 0 hori; : ie 29, itis clear that for a Specify f o ang hee lines in the ae Moreover, since cos ka jg «Vale of g ni seal figure, 4 ‘ais positive, negative or it ig Med VEN perio, nergy E(ag Owed, Consequence of this of the en is an even per a Ih : energy a8 a function of k. Tf K jg 4 tction of e et ja <1) which Corer ae be real, 8 petiod op 3 dingly, the total anergy B for which | cos ka 5}, ay allow pite forbidden bands. The alloweq a imaginan fig 11-4. the forbidd the mapnituy od Of 2da. Fi nitude of Fig. 11.4 shows a 4 energy band, Ont SHOU be less than 1 Y values of & are fen energy ban (it cee 0, a 3la-2nla-mla Wa Inia nia Fig. 11.4 Energy versus wave vector for a one dimensional lattice The analysis of eq. 29 leads us to the following inferences: mits a wave mechanical solution to exist as shown in Fig. 11.5. 1. Allowed range of aa pe! : x ‘ons in a periodic lattice is characterized by the bands of allowed Thus, the motion of electr energy separated by forbidden regions. a Peahicosre As the value of aa increases, the width K of the allowed band also increases while the width of the forbidden band decreases, This is because of the cal that the first term on the right cae eq. 29 on an average decreases increasing aa. For further understanding It ee the influence of P 0” te aa spectrum. The quantity Ie oa cated be a measure of the potenti ‘potential ength. If P is large, : ari (eq, n oo angle as shown in Fig. 11.6. Thy F at a steeper Hence, the ratio of th ses +1 and -1 region bands are wider. Ras arowarrower ad a Paibued band increases. let us take the ¢, ol potent ral ds to infinity. However, as forbidden band to eee. Pen wee at pes a stay inihin ie limit +1, it follows from eq. of the eq. P+0 ae ; fe +. 7 Fig. 11.6 (a) Right side of eq. 29 for P = 6x as a function of aa (b) P >» (c)P> 0 sin @a 30, ie. sinaa>0 aa = oa =+nnand ¢ = 2 However, from eq. 12, we have i =n? _ 82>mE n2h? a at yz ES amas This is equivalent to the case of. ee energy spectrum of a particle in a constant’ of atomic dimensions. This is expected because for large P tunnelling through the bar almost improbable. If on the other hand, P is made equal to zero, then the eq. 29 leads to C08 Oa = cos kor or @ =k PF = i? = 82> mE ao ge pees cvalent to the case of a fren omple: Prove that for the Kr, Patticle. He, © Die © ‘Onig-P, NCe, no er : eo) band at k = 0 is enney potential with nergy level exists; allowed. Pec 1, the energy of the lowes zo eem 47? ma? solution: For k = 0, the eq. 29 reduces to Eemag] Ga * 0S Qa=1 a P__1~cos aa eke sin Ga oe a (aa)? cos aa =1- SA 4 3 r sin az = an 20 8. a 2, Thus PB _ (aa) _@a ai aoe or P= 95 82?mE , oan 2mE @ is BaP. Therefore, pee or E aetna ? RILLOUIN ZONES bee i ave Galues of Kof the clectons in 205, Sas i the origin a way to to discuss the construct cos ka = 1 fers, Pear in electron momentum (or equivalently ine] vrease in el so that it can jump discontinuously f the in ie region between the first and the vied the second Brillouin zone and’ on, Ws lite eel ai vector k) will increase the ene! ‘of one allowed band to the bottom of the ELIS sec of & for which discontinuities occur is cal” Wietce shows two Brillouin zones for one dimension: 1 st zone 2nd zone Fig. 11.7. The first two Brillouin zones for a one dimensional lattice 5 In Two Dimensions Letus consider the motion of an electron in the field of a two dimensional square lattice, Keep the treatment just qualilatve, We know that the de Broglie wavelength of a free ae momentum p and energy Eis given by sea tal without being diffracted. However, m Fermi level wi cee ith E = B,), have comparable wavelengtt Re 4). Such electrons will be Giffracted in the same way as m according to Brage diffraction law where n= 1, 2,3 ni Ice pi tel a of electrons (Xrays or neutrons). oe ve the electron in a Crystal in terms of wave! Iisa vector quantity and its magnitude is relat nA = 2a sin 0, jorsion square lattice Fig. 1] reflection from vertical Tows WF sion is oF te ane Rpg ea of 5 ea ions Occurs. hen a ean (eq. 37) by saying that nent of the vector k in the x- ° ° és 1 1 ; i Positive ions, ° ° 5 i gi 4 9 © © ke nna sin 9 1 ee T - zag f a a Fig. 11.8 Bragg reflection from the vertical rows of ions occur when k, = nla iS =m ky = ksin @ = Similarly, reflection from horizontal rows occurs when forfirst zone, n = +1, therefore, k= eS and ky B, C and D as shown in x and y axes by the points A, B, Can e evaluated according to the following procedure. ors are given as: Thus, the first zone is bounded along be Fig 11.9. The rest of the boundary may ne i Consider a square lattice whose primitive oD vet azaiand b=bj : ® di, j are the unit vectors along the x, y axes, respectively: ‘The | Se lattice translation vectors can be written as: ding primitive recij 2n; ar = 28 and OF = 2E Fig. 11.9 The first three Brillouin zones for 2 two-dimensional square lattice so that the reciprocal lattice vector G can be written as =m ak +m b* = 22 [ny i+ mail a Further, the wave vector k in terms of its component vectors k, and ky can be v itten k= kyl + ky Substituting eqs. 39 and 40 in the Bragg’s condition 2k. G + G? = 0, we obtaim ee, 2 (bed + kj) - 2E (mitmj + AE (n, i+ my =0 An or 7a (on 2 + maky) + AE (n} +3) =0 where ii = jj = 1, and ij ou 0. Further simplifying this, we obtain miky + Mak, = 7 (n? + n}) fol sa x ony eve hor dates e a isis strated in Fig, 11.9, nus he e808 in k-space that yy, rillouin zone. When x Clectrons fist Brill 5 é S Ta, th CAN Occupy wi, without being diftra, "the elec rove UC beg ditfrased i ca my directions due to ditacsetete when Beil tel ay direction inside the Ff i 101 we mori jon of motion, until when f° ae Perot: oleae Prevented from moving in they move diagonally (at 45° with 4, and k, a V2. the electrons ne jon the possible aad the second zone begins, ¥ AX€s) inside the squi For the second zone, both the e integers equations for the second zone veers (ny and ny) in eq. 4 boundaries are: 1 are equal to +1. Therefore the M=41,n=41 giving ket hy= Qn a ny=-1 +M2= 41 giving -—, + & als m= +1, ny =-1 giving k — k = 2% a ny = -1, my = =1 giving ~k, ky = an The above four equations describe a set of four lines at 45° to the k, and ky axes passing through E,F,G and H as shown in Fig. 11.9. Similarly, the third zone is defined by giving m and n> the Yalues such as 0, +1, and +2. z The second Brillouin zone contains electrons with k values from 7/a (that do not fit into the first zone) to 2/a, or the region between the square ABCD and EFGH as shown in Fig. 11.9. In Three Dimensions Similar to the two dimensional case, fasily evaluated using the equation the form of the Brillouin zones in three dimensions can area 2 ko tm k= E (nt tnd tm) my ky + "y a follows from this equation that letsect the k,, ky and k, axes at tained by adding a pyramid (like zone as shown in the the points + a triangle i © ) 1 1 } 2 2 se bee foe hep Fig. 11.10 The first and second Brilliouin zones for some simple lattices Example: Find the ratio between the kinetic energies of an electron in a two dimensi lattice (a) when k, = ma and (b) when k, = /a and ky = 0. Solution: Given: Case I: = ma and case II: k, = 7/a and k, = 0. We know that of an electron inside a BZ is given by For case I, the energy becomes Pet B= grt (ke +g) = Similarly, for case I words, any two poi, : nts Wo lattice vector, LAVA k-space rey 0 i x Band Theory 31. Hf as shown in Fig. 11.11a, me eney ted 10 each wed 10 Pi BY 8 translation The points 12%: Let us take yee 8 it pid Go are reci equal tog °P2and P, the CHN6gF i ears ere ge iprocal lattice ~Gy and Py ie will have versus k for a Vectors, + is related the same energy square ' to Py by , because P) 4 Wanslation ~G), where © @ Fig 11.11 (a) The translational symmetry of the energy E(k) in k-space for a square lattice. (b) Mapping, of the second zone into the first zone. (c) The rotational symmetry of the energy E(k) im k- space for a square lattice sf we translate four pieces of the second 908 88 ‘Lean be shown that the pieces just fitani> ns that the first and the second ZO0ES &% he first zone contains all NECESSAEY Making use of this translational symmetry, tackward directions by the respective lattice vectors, the first zone as illustrated in Fig. 11-11b. This mea equivalent. The same is valid for higher-order Zones: Hence, t! information of higher zones: Ths Inversion Syminetty* a e oa respect (0 the inversion round the origi k=0. Shows that she band * a the point Pi is equal to that at Phe ‘ Accordingly, in Fig. 11-118, ener®Y at 376 Solid State Physics i mensional lattice. It is clear that i {llouin zone for one-dimension: i h a a a eee of the band (lower k value) and negative close to the x aan k~ ma). The figure also shows the curves £ versus k, dE/dk versus k and kin qualitative form. Re ice ‘The effective mass m* may be equi of a band and E versus k curve is parabolic (i.e. only under the free electron approxi most conductors, m* = mas the band is only partially filled. In semiconductors, ingyy certain conductors (e.g. bismuth) where full or almost full valence bands are im effective mass differs from 7. 11.7 THE NEARLY FREE ELECTRON MODEL | In this model, the crystal potential is assumed to be very weak as compared to Kinetic energy so that the electrons behave essentially like free particles. The wea of perturbing effect on the free electrons int potential introduces only a small amount Therefore, this model demands the application of very elementary perturbation # simplicity, only the one dimensional case is discussed. It is convenient to choose energy so that the mean value of the potential function is zero, i.e. * = m) only when the energy is not ih Vix)dx = 0 lo where a is the periodicity of the lattice. The unperturbed wave functions correspond O are the plane waves (2) = Tie where the wave functions are normalized over a microcrystal containing N atoms. The! electron energies are exp (ikx) E9(k) = Heke (Oe era This is free electron case where k can take an; into ac cae y value. Now suppose, we take into periodicity of the latti inehVaes hrs ee ae zs bea ice keeping V = 0 and restricting k to lie in the first Brillowl Wilx) = jniroduce the periodic poteng, Li al, the real wave j Band Theory 37 function Vea) sg, becomes G)+ Ba k) are constants and are fo Kae (bby, fed foibrst order are given by "nd 0 be small from erturbati a ion theory. Their values Av (k) = = (&) ~ BNR) on Ee: < KIM k> = [gs = JOLY Ode peperturbed energy, corrected to second order is % E(k) = E(k) + © kiko F BMK) — BOR) om» pease of eq. 52, the first order term is zero. Now: let us take into account the real periodic potential which may have the same periodicity ssthat of the lattice. In terms of the Fourier series, it can be expressed as V(x) = PAA exp(-igx) and Vy = Vi Then the integral 58 becomes eK IMK>=Vq ifR-R = rk = 0 otherwise Consequently, the wave function becomes Va where n #0 2m E(k) a rs bed energy iS we? 2) <8 * Bo ey ae * aie With Ay = 1 and Ay = 378. Solid State Physics Wee ? BK) - EK) = oe (k+ gy) 2 nh 2m “gta |Y (ee ) | hen («+ 1) *2nma At or near the zone boundary where n is an integer. The perturbation theory therefore breaks down whenever the w is an integral multiple of tn/a and the wave is reflected (k = -K’). Near a zone bo necessary to write W(x) = Ap explikx) + A, exp i(k + g)x where Ap and A, are constants. The eq. 64 means that one Fourier coefficient V, near the boundary k = +a/a. Substituting eq. 64 in eq. 1, we obtain [E*) — E(h) + Vo(2)]Ao explikx) + [E%Ck + g) — E(k) + V(x)1Ay exp i (k + ghe=t If we multiply eq. 65 either by exp(ikx) or by exp[-i(k + g )x] and integrate, we: following simultaneous equations, i.e. [E"(k) - E(k) Ao + Vp Ay = 0 Vy Ao+ [E%Ck + g) — E()IA, = 0 For a non-trivial solution, the determinant of coefficients must vanish. The resulting: equation has two solutions 2E(K) = [E(k + g)) + [{B%k) - E(k + g)}? + 4IVal°y? At the Zone boundary, k = tnz/a and E%(k) = Ek + 8). Therefore, the two en er(m) = 29(2) +101 e(22) = e°(22) ~ Wal BY Bap at the first zone boundary (n = 1) is AS opin! = F57.22V,, ; Band Theory 319 The ds i The free eee 1 Partbola ' 1 ' ' \ ' — Allowed enerey band ~ _tAllowe { —|Allowed energy band ak Bar ‘Bl Ist zone @) ng. 1S Energy band in the NFE model, (a) the extended 201 ne scheme, (b) the reduced zone scheme serposition of two plane waves traveling in opposite di in oj i ig osite directions and thus represents a standing Near k = 0, the difference in the unj i ir perturbed energies sreglsible and hence there is no effect onthe aneanan ae oe When E(k) — E(k + g) >> Vp, then E(k) = E°(k) or E%(k + g), i.e. away from Bragg reflection, the energy values are practically plane waves, the same as that of the free electrons. BS TIGHT BINDING APPROXIMATION itis nother extreme view point to see the electrons in a periodic potential, In the NFE case, we observed that the electron wave functions outside the ion core look very nearly as plane waves. However, near the ion cores they look like atomic orbitals, This suggests an entirely different tceme for the construction of wave functions, In 1928 5 Bloch suggested thatthe rystal wave Be ie with the correct symmetry could be eonstrucied 0 the linear combination of atomic attitals (LCAO). Letus start Ai an electron in a free atom and then suppose that a number of such atoms are brought together to form a crystal. For simeligy rosea Avo dimensional arrangement of atoms O° 4 soir ac Sani * iquare lattice is shown in Fig. 11.16. Let? a Rube the position vector of the electro? (atp) e Besos in tno crystal, The powenbal eh . ina free atoms is V(r ~ Ro crystal is V(r). In Fig. 1- rm nied by dashed and the Se ely. In tight binding aPP"® F “ ‘sumed that the electron WAN. by the nearest arom am i ‘atoms of the lattice: n for a free Fig. 11.16 Anaten te Oo SE re vo), Crystal potential Fig. 11.17 Atomic and crystal potentials approximated as y (r — Ry) So that, in general, the crystal wave function may be wi linear combination of the form V(r) = E expCik-Ry W0(r — Ra) The summation in eq. 68 is over all atoms of the crystal, which for simiplicity we: infinite. Since, we have considered that the electrons are in a periodic potential, a Bloch function because it can be shown that Viel + Rm) = E explik-Ry Wo (7 — Rn + Rm) = explik-Rm) E explik-(Ry- Rm)IWolr = (Ra ~ Rm) = exp(ik-Rn) Yr) This shows that the Bloch condition (eq. 3) is satisfied. y(r) given by eq. 68 is the Schrodinger equation for the whole crystal: Ayr) = Eyylr) The crystal Hamiltonian H may be written as the sum of the Hamiltonian for @ H=Ho +H, : BY the denominator here J vic Cde By oy ny XPLiK(R, ! Ralf ys pepsin the overlap of electron waye vie Ra)Wo(r = Ry) Ve functio ‘ ns betw i en the neighbouring atoms such that J ¥i0--a Svea fn and * Jui OW (r)dr = N where N is the total (76) 1 numberof ele: s st lerefore, eq. 74 becomes electrons (ato; ms) in the crystal. Th . Therefore, orl . E(k) = N j Vi LH + Hy lyy (rar =1 [ yepfosilien w ] vien{- ge v? + vole Ro ystnne +H j VECLV(r) — Volr= Radlva(rddr 1 SE explik(Re- Ral] vile —Ra)™ = 2 J vivid [V(r) — Volr — Rad] yolr ~ Ral dr explit(Ra- Re) vitrr Balls =Fot ye a =Oton=N-1to contain N identical (0, Hence, eq. 77 reduces (0 (78) Vo(r — Rall yotr - 20) a Ivo - ae Hete, we expect that every term in the summation etituting n= s, which may be evaluated most easily Dy § a = Vole) yotr) ar pyanbige tte Jy ve) — Volo] E(k) = Fo+ mm i vmmetric (ie: Funher scimasion that Wo is ess That the contibuions due foal 5 if we make an aPPEPr cron state een or m= 0 ene integral im eo CS be i : Metals for which the grou" oo curs ay beeen - J vier ee van te a= 382 Solid State Physics and for the nearest neighbour atoms J voce wen — vai volr r= B s overlap integrals. Since Vr) ~ Va(y) ig, where a and B are constants and are known a: i V the eq. 78 can be written in a simp] therefore cand fare positive. Consequently, E(k) = Ey- @- B x exp (-ik-Rm) ‘where the summation is to be carried out over the nearest neighbours only. It is observed eq. 81 consists of a constant term Ey ~ o together with the term dependent on k. It is th term which transforms the discrete atomic levels into the energy bands in the solid, In order to demonstrate the application of the above method, let us take the case of as cubic crystal and consider only the nearest neighbour interaction. For an atom which is the origin, there are six nearest neighbours whose lattice vectors are: R, (+a, 0, 0), (0, +a, 0), (0, 0, +a) Substituting this in eq. 81, we obtain E(k) = Ep ~ &- 2B (cosk, a + cosk, a + cosk; a) For small value of k, we may expand the cosines and retaining upto the second term, & becomes S Ka? ka Bw) =,-a-29|(1- F)+(1- oe )-0-42] =F) 0-6 Bi + +2) Since, the last term on the right hand side i i ast ter of eq. 83 is periodic and hence E(k) is With the periodicity 2n/a, Thus, the first Brillouin zone is defined as alae & 3 wd Brillouin zone is a cube. Further, al : 1 Thus, the bottom ofthe band is given by nn Swi the enerey is minimum at =H EWnin = ER) scion = Ey ~ a0 68 (a) near bottom of band Mig. 11.18 Constant energy surfac, (©) neat top of band es for simple cu bic lattice j ice in tight binding mod Thetwo dimensional a e imensional view of the eni ; ; ergy cont F e dectron model are sh ‘ Ours for simple cubi consi Own in Fig. 11.19. In tight binding modi, dee tight binding and free 5 ouin zone of a simple ~ E®@soriom = 128 a ma Wa ma = -ala N d ala (a) tight binding approximation Ly 4 -nla (b) nearly-free electron approximation Fig. 11.19 Electron energy contours for simple cubic lattice Subic lattice k curve along a cube axis are shown in Fig. 11.20. Further, from 4. 80, it an, Page pee, width of the band becomes greater as the orsiangt the wave actions on the ighbouring atoms increases, Thus, the inner electronic levels of the free so nei, A ‘ to the outer levels. MS give rise to narrow bands in the solids and increase as we proceed to the outer le 386 Solid State Physics 11.10 THE PSEUDOPOTENTIAL METHOD a ‘The pseudopotential method is just an extension of the OPW method. 1959 noticed that the effective potential Veq(r) defined by eq. 95 is weaker than, Vand called it the pseudopotential. They observed that the contribution from the right side of eq, 95 is negative, while the contribution from the second term is posi the two terms cancel each other and lead to a weak effective potential, Veri(r), Physically, the cancellation between the true attractive potential ang orthogonalization term may be viewed as an elimination of rapid oscillations 50 electron appears to see only the weak pseudopotential. The true crystal po function, and their corresponding pseudo counterparts are as shown in Fig. 11.7 hand, mathematically, the above cancellation may be viewed as a transform wave equation to a pseudo wave equation. Accordingly, let us write the true a valence state as a linear combination of OPW’s as in eq. 91. Further, suppose: plane wave part of this expansion. Then, eq. 88 can be rewritten as: Fig. 11.22 The pseudopotential concept (a) the actual potential and the corresponding. seen by electrons, (b) the corresponding pseudopotential and pseudofunction O1(r) = Za, expi(k + g)-r or lof) = Za, k+ 9) This defines the pseudo wave function as a linear combinati P ee ce sonsiicnan ination of ordinary AEE WA) = (1 ~ P) Ip) ff similar to eq. 94, we May define posite alectroms in metals can (in many cases) be regarded as quasi-free particles. Because of the same reason, it is possible to use the perturbation theory to find out the energy eigen states. Free electron wave function may be used as a zero approximation.

You might also like