- ChAPTER 7 FREE-ELECTRON QUANTUM MODEL OF METALS 7-1 Particle in a Box In this chapter, we will modify the clasical model

we developed in Chapter 4. We will apply some quantum mechanics we introduced in Chapter 6 plus some additional concepts as we need them. As in the classical model, we will consider the conduction electrons to be free. We neglect their interaction with the positive ions and also with each other. We only require that the electrons remain within the metal. Thus the electrons behave like particles trapped in a box. We treated this problem in the last chapter for the one-dirnensional case. Let us now extend that treatment to three dimensions. Consider a cubic box with sides of length L. The wave function inside the box is simply a solution of Schrocdingers equation for a free particle. Outside the box, must be zero, If we put one corner of the box at the origin, as shown in Fig, 7.1 then we must require that the wave function tJ(z, y, z, t) satisfy the following boundary conditions: ,J(0,y,z,t) =0, iJ,(L, y, z, t) = 0, ,k(x,0,z,t) =0, i4(z,L,z,t) =0, t(z,y,0,t) 0, ,k(z,y,L,t) =0. Similar to Eq. (626), we try a wave function of the form. (z,y, z, t) = A sin(krz) sIn(ky) sin(kzz) exp(iwt), where k1, k, k are components of The wave vector, k kI+ ki + kk. (7=1) Problem 7-1. Show that Eq. (7-2) is a solution o the three-dimensional Schroedinger equation for a free particle l-q. (6-36) with U = 0.

If we require to satisfy the boundary conditiors of eq (7-1), we fInd that the cornponents of k must be restricted to values, = = where rz, n,, n are integers. As in the one-dimensionai case, we consider only positive non-zero values of n, n, and n.

problem 7_2. Show that the wave function given by (7-2) satisfies the bondany conditions given by Eq. (7-i) if the values of k1, k, and k are restricted to those of Ec. t7-)

7-2 Periodic Boundary Conditions There is another way to treat free conduction electrons in a metal which is actually preferable to paticles in a box. The cubic box of side L has surfaces. We remove the surfaces by stacking stacking identical cubes of side L in a simple cubic arrangement so that they fill all space as in Fig. 1-6. The cubes are repeated throughout all space. Each wave function also fills all space. Since we are setting up a model for electrons within a single cube of metal, we want everything about each cube to be identical. (Every cube represents the actual cube of metal which we are considering.) Therefore, each wave function must be identical in every cube. The boundary conditions of Eq. (7-1) are thus replaced with (z+ L,y,z,t) = tk(z, y + L, z,t) = iJ,(z,y, z, t), (7-5) iJi(z,y,z+L,t) = These are called periodic boundary conditions. We find that in this case, the wave function for a free electron given by Eq. (6-37), (x, y, z, t) = A exp(ikx + iky + ikz iwt), (7-6) can satisfy the periodic boundary conditions of Eq. (7-5) if we restrict the components of k to the values, = n1(2ir/L), = n,(2ir/L), (7-7) k = n(2ir/L).

Problem 7-3. Show that the wave function given by Eq. (7-6) satisfies the boundary conditions given by Eq. (7-5) if the values of k, k, and k are restricted to those of Eq. (7-7).

In this case, r. n, n, may take on negative values as well as positive values since these give us different wave functions i.. This is a principal advantage of using periodic boundary conditions in place of thos in Eq. (7-i). The wave vector k can now point in negative as well as positive directions, and we can now associate k with the direction of the electrons velocity. Using Eq. (77), we find that the allowed wave vectors are k (2x/L)(nl + n,j + rz2k). (7-8) Each allowed k corresponds to a different wave function of an electron.

7-3 Density of States If we plot all the allowed wave vectors in k-space (reciprocal space), we obtain a set of points which forms an sc lattice as shown in Fig. 7-2. The distance between adjacent lattice points is 2/L. Remember that for an sc lattice of lattice parameter the density of lattice points is given by a. In the present case, we have an sc lattice in k-space with lattice parameter 2/ L. Therefore, the

density of these points in ks pace is (2/L)3 = t,(2x)!. where V = L3, the volume of the cube. I ig. 7-2. Allowed wave vectors c conduction electrons in a metal.

Before we proceed further here, we must introduce another property of electrons. An electron spins about an axis. Since it has charge, the spinning motion gives rise to a magnetic moment (like current in a loop of wire). It has been found that the direction of this magnetic moment due to spinning is quantized and can take on only one of two possible values. These are called the spin states of the electron. For every allowed wave function / of an electron, there are two electron states, corresponding to the two possible directions which the magnetic moment of the electron can point. Every allowed value of k given by Eq. (7-8) correspond to a different wave function and thus represents two electron states. If the density of these allowed wave vectors k in k space is equal to V/(2ir)3, then the density of electron states in k- space is equal to twice that value, or 2V/(2ir)3. I) denoting the density of electron states in reciprocal space by g(k), we write g(k) 2V/(2ir)3. (7.9) The number of electron states within some volume 11k In k-space is given by the product 17k g(k). Let us next calculate the number of electron states with energy less than some value E. - For free electrons, E = p2/2rn = h2k2/2rn. (7-IC)) Solving for k, ________ k=./E/h2. (7-Il) Thus, the number of electron states with energy less than E is the same as the number of states with wave number less than the value of k given by Eq. (7-11). These states are contained within a sphere of radius k in k-space. The volume of this sphere is given by = irk3 = r(2mE/h2)312. (7-Il) To obtain the number of states N(E) within this sploIe, we multiply Vk by the density of states in k-space given by We next obtain the density of states g(E) as a function of E. By density of states g(E), we mean that the number of electron states between two energies L and 2, is given by the integral of g(E) from E to 2. The quantity V(E) Eq (7-13) above is the number of sta:es between energies 0 ad K. Thus, N(E) = J g(E) dE, (7-14) 0 d -v ( g(E) .L, (7-is) which, using Eq. (7-13), gives us g(E) (1f/2)(2m!h2)3/2Ez/2 (7-16)

a plot of g(E) is shown in Fig. 7-3. It is simply a parabola on its side. As stated above, by definition the number of electron states between two energies, E and E, is given by the area under the curve between E1 and E. This is shown by the shaded area in Fig. 7.3. 7- Paulis exclucion principle So far, We have only found the possible states of an electron in the metal. If we have N conduction electrons, we now must consider which states we will find these electrons in. Let us first consider the case of absolute zero temperature (T = 0). None of the electrons have any thermal energy and must therefor occupy the states of lowest energy available. We might expect all of the electrons to be in the same state (the state of lowest energy). However, this cannot happen because of another rule in quantum mechanics called Paulis exclusion principle Paulis exclusion principle states that no more than one electron can be in the same state at the same time. This is simply a fundamental principle of quantum mechanics which is analogous to the classical concept that two objects cannot be at the same place at the same time. however, Paulis exclusion principle only applies to a class of particles called fermions (Electrons, protons, and neutrons are examples of fermion) All other particles (called bosons) do not need to obey this principle. The helium atom, which of course is a composite of fermions, happens to be a boson, for example. Paulis exclusion principle has dramatic consequences for conduction electrons in a metal. At T 0. each electron tries to occupy the lowest available energy state. But since they cannot all be in the same state, some electrons must by necessity occupy states of higher energy. The situation is analogous to filling a glass with water. Because of gravity, all the water molecules would like to be at the bottom of the glass, but because of a volurne exclusion principle, only a small fraction are at the bottom. In fact, some water molecules are forced to be quite far removed from the bottom, depending on how much water we pour into the glass. Similarly, the energy of the highest state occupied by an electron in a metal depends on how many conduction electrons there are. With all the conduction electrons occupying states, the energy of the highest occupied state is called the Fermi energy Ef. The number of states N(EF) with energy between 0 and EF must therefore be equal to the total number of conduction electrons N. From Eq. (7-13), we thus have N = Vk c(k) = (73z)(2rr-zEr/h2)3/2. (7-17) Solving for E, we obtain = (T-i) where n = V/t, the density of conduction electrons in the metal. Note that EF does not depend on the volume V of the metal. The velocity of the electrons at the Fermi energy is called the Fermi velocity r1.