CHAPTER 5 Angular Momentum 5.1 SIMULTANEOUS SPECIFICATION OF SEVERAL PROPERTIES In this chapter we discuss angular momentum, and in the next chapter we show that for the stationary states of the hydrogen atom the magnitude of the electron’s angular ‘momentum is constant. As a preliminary, we consider what criterion we can use 10 decide which properties of a system can be simultaneously assigned definite values. In Section 3.3 we postulated that if the state function V is an eigenfunction of the operator A with eigenvalue s, then a measurement of the physical property A is certain to yield the results. If W is simultaneously an eigenfunction of the two operators A and B, thatis,if AW = s¥ and BY = r'¥, then we can simultaneously assign definite val- ues to the physical quantities A and B. When will it be possible for ¥ to be simultane ously an eigenfunction of two different operators? In Chapter 7, we shall prove the following two theorems. First, necessary condition for the existence of a complete set of simultaneous eigenfunctions of two operators is that the operators commute with each other. (The word complete is used here in a certain technical sense, which we ‘won't worry about until Chapter 7.) Conversely, if A and B are two commuting opera- tors that correspond to physical quantities, then there exists a complete set of functions that are eigenfunctions of both A and B. Thus, if [A 8] = 0, then can be an eigen- function of both A and B. Recall that the commutator of A and B is defined as [A,B] = AB — BA [Eq. 3.)|. The following commutator identities are helpful in evaluating commutators, these identities are easily proved by writing out the commutators in detail (Problem 5.1): «yr 1,2,3,... (5.2) KA, 8) 63 [4.8] +(4.0], [A+BC)=(4.0) + (B.C) ar [ABC + BLA), (AB.C)= (4,018 + 418.2) 65" constant and the operators are assumed to be linear. EXAMPLE __ Starting from [0/ax.x] = 1 [Eq. G.)], use the commutator identities (5.1)— (535) to find (a) [£, 5]; (b) (4, 62] (€) [&, H] for a one-particle, three-dimensional system. (a) Use of (5:3), (5:1), and [4/ax, x] = 1 givesSection 5.1. Simultaneous Specification of Several Properties 95 aaa [x 2] [adam 66) (b) Use of (5) nd (5.6 gives (3, #2) = [%, lB. + Bald, BD ee 2 [ale 62) (6) Use of (5.4), (5:3), and (5.7) gives (A [ET +0) = (4,7) + (2.0 y.29) = (87) 68) = [x (2m + BF + PD] = (I/2m)lé, f2] + (I/2m)L i, (69) ‘The above commutators have important physical consequences Since (2, p,] * 0, ‘we cannot expect the state function to be simultaneously an eigenfunction of £ and of i, Hence we cannot simultaneously assign definite values to.x and p,,in agreement with the uncertainty principle. Since £ and H do not commute, we cannot expect to assign definite values to the energy and the x coordinate at the same time. A stationary state (which has a definite energy) shows a spread of possible values for x, the proba- bilities for observing various values of x being given by the Born postulate. Fora state function that is not an eigenfunction of A, we get various possible outcomes when we measure A in identical systems. We want some measure of the spread or dispersion in the set of observed values A,.If (A) s the average of these val- ues, then the deviation of each measurement from the average is A, - (A). If we aver- aged all the deviations, we would get zero, since positive and negative deviations would cancel. Hence to make all deviations positive, we square them. The average of the squares of the deviations is called the variance of A, symbolized in statistics by o% and in quantum mechanics by (A.A)* (aay = oh =(4- (A) = fora — (aw adr (5.10) where the average-value expression (3.88) was used. The definition (5.10) is equivalent to (Problem 5.4) (aay = (4) = (4? an)96 Chapter5 Angular Momentum ‘The positive square root of the variance is called the standard deviation, o, or AA. The standard deviation is the most commonly used measure of spread, and we shall take it as the measure of the “uncertainty” in the property A. For the product of the standard deviations of two properties of a quantum- ‘mechanical system whose state function is ‘Y, one can show that (Problem 7.58) AAAB>5| (5.12) Afera. BY az| If A and B commute, then the integral in (5.12) is zero, and AA and AB may both be zero, in agreement with the previous discussion. ‘As an example of (5.12), we find, using (5.6) and |z,2 [ell (Ea. 1.329), = 5 il fo ar| Ax Ap, >4h 13) Equation (5.13) is the quantitative statement of the Heisenberg uncertainty principle. 1 Ax dp, =4 Sport. BJV dr [foraya 2 EXAMPLE __ For the ground state of the particle in a three-dimensional box, use the follow- {ng results of Eqs (391), (3:92), (3.39), the equation following (389), and Problem 3.39 (3) = 4/2, () = a3 — 1/207), Cp.) = 0, (Pi) = Ha? tocheck that the uncertainty principle (5.13) i obeyed. We have (Ax? = (27) ~ (xP = aX(1/3 ~ 1/207) ~ ab/4 = aXe? — 6)/1207 Ax = a(a? — 6)'7/(12)"2 (Ape? = (0) ~ (pe) = H/4a?, Ape = hf2a Ar ap,= 34 = 0.5684 > ta ‘There is also an uncertainty relation involving energy and time: AE A= (4) Some texts state that (5.14) is derived from (5.12) by taking if a as the energy oper- ator and multiplication by ¢ as the time operator. However, the energy operator is the Hamiltonian H,and not if 3/41. Moreover, time is not an observable but i a parameter ‘quantum mechanics, Hence there is no quantum-mechanical time operator. (The noun observable in quantum mechanics means a physically measurable property of a system.) Equation (5.14) must be derived by a special treatment, which we omit. (See Ballentine, Section 12-3.) The derivation of (5.14) shows that A/ is to be interpreted asSection 5.2 Vectors 97 the lifetime of the state whose energy is uncertain by AF. It is often stated that Ar in (6.14) is the duration of the energy measurement; however, Aharonov and Bohm have shown that “energy can be measured reproducibly in an arbitrarily short time.” [See Y. Aharonov and D. Bohm, Phys. Rev., 122, 1649 (1961); 134, B1417 (1964) ] Now consider the possibility of simultaneously assigning definite values to three physical quantities: B, and C. Suppose [A, 8) =0 (5.15) [4,é]=0 (5.16) Is this enough to ensure that there exist simultaneous eigenfunctions of all three oper- ators? Equation (5.15) ensures that we can construct a common set of eigenfunctions for A and B; Eq, (5.16) ensures that we can construct a common set of eigenfunc- tion tor A and @ It theve wo sts of eigenfunctions are the same, then we wil havea common set of eigenfunctions forall three operators. Hence we ask: Is the set of eigen- functions of the linear operator A uniquely determined (apart from arbitrary ‘multiplicative constants)? The answer is, in general, no. If there is more than one inde- pendent eigenfunction corresponding to an eigenvalue of A (that is, degeneracy), then any linear combination of the eigenfunctions of the degenerate eigenvalue is an eigen- function of A (Section 3.6). It might well be that the proper linear combinations needed to give eigenfunctions of B would differ from the linear combinations that give eigenfunctions of C. It turns out that, to have a common complete set of eigenfunctions of all three operators, we require that [8, C] = 0, in addition to (5.15) and (5.16). To have a complete set of functions that are simultaneous eigenfunctions of several opera- tors, each operator must commute with every other operator. 5.2. VECTORS In the next section we shall solve the eigenvalue problem for angular momentum, which isa vector property. We therefore first review vectors. Physical properties (for example, mass, length, energy) that are completely speci- fied by their magnitude are called sealars. Physical properties (for example, force, velocity, momentum) that require specification of both magnitude and direction are called veerors. A vector is represented by a directed line segment whose length and direction give the magnitude and direction of the property. ‘The sum of two vectors A and B is defined by the following procedure: Slide the first vector so that its tail touches the head of the second vector, keeping the direction of the first vector fixed. Then draw a new vector from the tail of the second vector to the head of the first vector. See Fig, 5.1. The product of a vector and a scalar, cA, is defined as a vector of length |c times the length of A with the same direction as A if cis positive, or the opposite direction to A if cis negative. To obtain an algebraic (as well as geometric) way of representing vectors, we set up Cartesian coordinates in space. We draw a vector of unit length directed along the positive x axis and call iti. (No connection with i = V1.) Unit vectors in the positive yand z directions are called j and k (Fig. 5.2).To represent any vector A in terms of the three unit vectors, we first slide A so that its tail is at the origin, preserving its direction98 Chapter5 Angular Momentum ix @ OCH A+ BBA FIGURE 5.1 Addition of two vectors. uring this process. We then find the projections of A on the x, y, and z axes: A,, Ay, and A, From the definition of vector addition, it follows that (Fig. 52) AMAGtAj+ AK (a7 ‘To specify A, itis sufficient to specify its three components:(A,, A,, A,).We can there- fore define a vector in three-dimensional space as an ordered set of three numbers. ‘Two vectors A and B are equal if and only if all the corresponding components are equal: A, = B,, A, = B,, A, = B,. Hence a vector equation is equivalent to three scalar equations. To add two vectors analytically, we add corresponding components: A+B=AG+Aj+Ak+ Bi+ Bj + Bk A+B=(A, + Bit (A, +B) + (A, + Bk 6.18)" Also, if cis a scalar, then cA = cA,i + cAyj + cAK (S.19)* ‘The magnitude A of a vector A is its length and is therefore a scalar. Often the notation [Al is used for the magnitude of A. FIGURE 5.2. Unit vectorsi.J,k, and components of ASection 5.2 Vectors 99 “The dot product or scalar product A.» B of two vectors is defined by A+ B=|Al|B|cos 0= B-A (520) where 0 is the angle between the vectors. The dot product, being the product of three scalars, is a scalar. Note that |A| cos @ is the projection of A on B. From the definition Cf vector addition, it follows that the projection of the vector A + B on some vector C is the sum of the projections of A and of B on C. Hence (A+B)-C=A-C4B-C (621) ince the three unit vectors i,j,and k are each of unit length and are mutually perpen- jcular, we have jick Using (5.22) and the distributive law (5.21), we have Ad + Aj + Ade) (Bit Bj + BR) A+B A,B, + A,B, + A.B, (523)" jok=kei cos (77/2) = 0 (5.22) AB where six of the nine terms in the dot product are zero. Consider the dot product of a vector with itself. From (5.20) we have A-A= (AP 624)" Using (5.23), we therefore have lal = (At + a2 + Ayre 625)" For three-dimensional vectors, there is another type of product. The cross prod- uct or vector product Ax Bis.a vector whose magnitude is [a x Bl =|Al[B| sino (526) whose line segment is perpendicular to the plane defined by A and B, and whose direc- tion is such that A, B, and AX B form a right-handed system (just as the x, y, and z axes form a right-handed system). See Fig, 5.3. From the definition it follows that BxA=-AxB (5.27) ‘Also, it can be shown that (Taylor and Mann, Section 10.2) AXx(B+C)SAXBHAXC (5.28) Axa B FIGURE 5.3 Cross product of two vectors.100 Chapter Angular Momentum For the three unit vectors, we have ixisjxj=kxk=sin0=0 Ixj=k, Jxis-k xk Kxj=-h kx Using these equations and the distr property (5.28), we find AKBH(AG+ A+ AM) x (B+ Bi + BLO AX B=(A,B,— AB)I+ (A,B, ~ A.B) + (A.B, ~ A,B Asa memory aid, we can express the cross product as a determinant (see Section 8.3): ijk AxB=|4, 4, A, \e &) (29) B, By B, a" ‘We define the vector operator del as aya 2p tene a ax thay k 630) From Eq. (3.23), the operator for the linear-momentum vector is p = —iAV. ‘The gradient of a function g(x, y, z) is defined as the result of operating on the function with del: ve 4 ag, ag 28 ax ay” az grad g(x, y, z) = Val, yz) (3* ‘The gradient of a scalar function is a vector function. The vector Vg(x, yz) represents the spatial rate of change of the function g; the x component of Vg is the rate of change of g with respect to.x, and so on. Tt can be shown that the vector Vg points in the direc- tion in which the rate of change of g is greatest. From Eq. (4.26), the relation between force and potential energy is, F = -WV(x,y,2) = (5:32) Suppose that the components ofa vector are each functions of some parameter A, = A{t), Ay = Aj(t), A, = Aff). We define the derivative of the vector with respect o 38 ahaa, dAy dt a ae dA, a Vector notation is a convenient way to represent the variables of a function. The wave function of a two-particle system can be written a8 W(X}, Ys Zi Xo Jos 22) If is the vector from the origin to particle 1, then r, has components x, 1, 2 and specifi- cation of r, is equivalent to specification of the three coordinates x;, 21. The same is true for the vector r, from the origin to particle 2. Therefore we can write theSection 5.2 Vectors 101 wave function as #(F), 2). Vector notation is sometimes used in integrals. For ex- ample, the normalization integral over all space in Eq. (3.57) is often written as SS MC ces te OP dey dt. ‘Vectors in n-Dimensional Space. For many purposes itis useful to generalize the definition of a vector to more than three dimensions. A vector A in three-dimensional space can be defined by its magnitude |A| and its direction, or can be defined by its three components (A,, A,, A.) in a Cartesian coordinate system. Therefore we can define a three-dimensional vector as a set of three real numbers (A,, Ay, A,) in a par- ticular order. A vector B in an n-dimensional real vector “space” (sometimes called a hyperspace) is defined as an ordered set of n real numbers (By, Bs,..., B,) .., where Bj, Bry. By ate the components of B. Don’t be concerned that you can’t visualize vec- tors in an n-dimensional space. ‘The variables of a function are often denoted using n-dimensional vector nota- tion. For example, instead of writing the wave function of a two-particle system as, ‘U(r,.t2), we can define a six-dimensional vector q whose components are 4) =X), 42= G5= 21,du= 0. 4s= Yo Go= Zo and write the wave function as #(q). For an n-particle ‘system, we can define q to have 3n components and write the wave function as (q) and the normalization integral over all space as { |Y(@)? da. ‘The theory of searching for the equilibrium geometry of a molecule uses n-dimensional vectors (Section 15.11). The rest of Section 5.2 is relevant to Section 15.11, and need not be read until you study Section 15.11. ‘Two n-dimensional vectors are equal if all their corresponding components are equal; B= C if and only if By = C,, By = Ca, ..., By = Cy- Thus, in n-dimensional space, a vector equation is equivalent to » scalar equations. The sum of two n-dimensional vec- tors B and D is defined as the vector (B, + Dj, By + Dz,..., B, + D,)-The difference is, defined similarly. The vector KB is defined as the vector (KB), KB2, ..., KB,), where kis a scalar. In three-dimensional space the vectors kA, where k > 0, allie in the same direction. In n-dimensional space the vectors kB all lie in the same direction. Just as the numbers (4,, A,, A.) define a point in three-dimensional space, the numbers (B,, Br, ..,B,) define a point in n-dimensional space. ‘The length (or magnitude or Euclidean norm) |B| (sometimes denoted |B) of an ‘n-dimensional real vector is defined as [|= (B+ By)? = (By + B+ + BI? |A vector whose length is 1 is said to be normalized. “The inner product (or scalar product) B + G of two real n-dimensional vectors B and G is defined as the scalar B-G = B,C, + BG + + BG, If B- G = 0, the vectors B and G are said to be orthogonal. The cosine of the angle © between two n-dimensional vectors B and C is defined by analogy to (5.20) as cos @ = B - C/|B| |C|. One can show that this definition makes cos @ lic in the range =lto1 In three-dimensional space, the unit veetors i= (1,0,0), j = (0,1, 0), k = (0,0, 1) are mutually perpendicular. Also, any vector can be written as a linear combination102 Chapter5 Angular Momentum of these three vectors [Eq, (5.17)]. In an n-dimensional real vector space, the unit vectors €; = (1,0,0,...,0),€2 = (0, 1,0,-.-,0),...+€, = (0,0,0,...,1) are mutually ‘orthogonal. Since the n-dimensional vector B equals Bye; + Bye: + ~~ + Byes, any ‘n-dimensional real vector can be written as a linear combination of the m unit vectors ©, €n,---, This set of n vectors is therefore said to be a basis for the n-dimensional real vector space. Since the vectors €;, €,..-»€, are orthogonal and normalized, they are an orthonormal basis for the real vector space. The scalar product B - ¢; gives the component of B in the direction of the basis vector e,. A vector space has many possi- ble basis sets Any set of n linearly independent real vectors can serve as a basis for the ‘n-dimensional real vector space. A three-dimensional vector can be specified by its three components or by its length and its direction. The direction can be specified by giving the three angles the vector makes with the positive halves of the x, y, and z axes These angles are the direc- tion angles of the vector and lie in the range 0 to 180°. However, the direction angle with the z axisis fixed once the other two direction angles have been given, so only two rection angles are independent. Thus a three-dimensional vector can be specified by its length and two direction angles. Similarly, in n-dimensional space, the direction angles between a vector and each unit vector e,, €,..., €, can be found from the above formula for the cosine of the angle between two vectors. An n-dimensional vector can thus be specified by its length and n — 1 direction angles. ‘The gradient of a function of three variables is defined by (5.31). The gradient Vf of a function (qi, qos... 4,) Of variables is defined as the n-dimensional vector whose components are the first partial derivatives of f: VE = (af /Aai)er + (Af [Agee + + (AF/8qn)en ‘We have considered real, n-dimensional vector spaces. Dirac’s formulation of quantum mechanics uses a complex, infinite-dimensional vector space, discussion of which is omitted. 5.3 ANGULAR MOMENTUM OF A ONE-PARTICLE SYSTEM. In Section 33 we found the eigenfunctions and eigenvalues for the linear-momentum operator jn this section we consider the same problem for the angular momentum of a particle. Angular momentum is important in the quantum mechanics of atomic structure, We begin by reviewing the classical mechanics of angular momentum. Classical Mechanics of One-Particle Angular Momentum. Consider a moving particle of mass m, We set up a Cartesian coordinate system that is fixed in space. Let r be the vector from the origin to the instantaneous position of the particle. We have reixtjy+ke (5.33) where x, y, and z are the particle's coordinates at a given instant. These coordinates are functions of time, and defining the velocity vector v as the time derivative of the posi- tion vector, we have (Section 52)Section §.3. Angular Momentum of a One-Particle System 103. v= dx/dt, v= dy/dt, — v, = de/at We define the particle’s linear momentum vector p by p=my (635) Pe= vz, y= mv, pe = mv; (536) ‘The particle’s angular momentum L with respect to the coordinate origin is defined in classical mechanics as Lerxp 637 ijk Le|x yz (8.38) IPs Py Pe Le YPe~ 2Pyy Ly = Pe ¥Pey Le = Py YP (539) where (5.29) was used. L,, L,, and L, are the components of L along the x, y, and z axes. The angular-momentum vector L is perpendicular to the plane defined by the particle's position vector r and its velocity v (Fig, 5.4). “The torque + acting on a particle is defined as the cross product of r and the force F acting on the particle: r= r x F, It is readily shown that (Halliday and Resnick, Section 12-3) 7 = dL/dt. When there is no torque acting on the particle, the rate of change of its angular momentum is zero; that is, its angular momentum is constant (or conserved). For a planet orbiting the sun, the gravitational force is radially directed Since the cross product of two parallel vectors is zero, there is no torque on the planet and its angular momentum is conserved. One-Particle Orbital-Angular-Momentum Operators. Now for the quantum- mechanical treatment. In quantum mechanics, there are two kinds of angular ‘momenta: orbital angular momentum results from the motion of a particle through space, and is the analog of the classical-mechanical quantity L; spin angular momen- tum (Chapter 10) is an intrinsic property of many microscopic particles and has no classical-mechanical analog, We are now considering only orbital angular momentum, FIGURES L=r xp108 ChapterS Angular Momentum We get the quantum-mechanical operators for the components of orbital angular ‘momentum of a particle by replacing the coordinates and momenta in the classical equations (5.39) by their corresponding operators (Eqs (3.21)-(3.23)}. We find (5.40) (5.41) (5.42) (Since $f, = p,¥,and so on, we do not run into any problems of noncommutativity in constructing these operators) Using P= [LP =L-L= t+ i+ it (5.43) ‘we can construct the operator for the square of the angula-momentum magnitude from the operators in (5.40)-(5.42). Since the commutation relations determine which physical quantities can be simultaneously assigned definite values, we investigate these relations for angular momentum. Operating on some function f(x, y, z) with L,, we have i= a(t) ‘Operating on this last equation with L,,we get a; af, ef (yf 4 yeh LL f = -# (» ox dean 7 Similarly, 2% eh a ae he gh wert Zl) oa Subtracting (6.45) from (5.44), we have LL, ~ bybf = - w(t - *) Lig -#(e ax “ay, (bab) = ih, 46) where we used relations such as af ef azax ax dz Gane which are true for well-behaved functions. We could use the same procedure to find [Z,, £,] and [£,, L,], but we can save time by noting a certain kind of symmetry inSection 5.3 Angular Momentum of a One-Particle System 105 (5.40)-(5.42). By a cyclic permutation of x, y, and z, we mean replacing x by y, replacing y by z, and replacing z by x. If we carry out a cyclic permutation in Z.,, we get Ly; a ‘cyclic permutation in L, gives L,; and L, is transformed into L., by a cyclic permuta- tion, Hence, by carrying out two successive cyclic permutations on (5.46), we get [L,Lj=ith,, (Lab,) = itl, (5.48) ‘Now we evaluate the commutators of L? with each of its components, using com- mutator identities of Section 5.1. (2 be = =t =t =0 (6.49) Since a cyclic permutation of x, y, and z leaves L? = 1} + £} + L? unchanged, if we carry out two such permutations on (5.49), we get (2,L,)=0, [£L,)=0 (6.50)* ‘To which of the quantities L?, L,, L,, L, can we assign definite values simultane- ously? Because L? commutes with each of its ‘components, we can specify an exact value for L? and any one component. However, no two components of L. commute with each other, so we cannot specify more than one component simultaneously. (There is one exception to this statement, which will be discussed shortly.) It is tradi- tional to take L, as the component of angular momentum that will be specified along with L?, Note that in specifying L? = |L? we are not specifying the vector L, only its magnitude. A complete specification of L requires simultaneous specification of each of its three components, which we usually cannot do. In classical mechanics when angu- Jar momentum is conserved, each of its three components has a definite value. In quan- tum mechanics when angular momentum is conserved, only its magnitude and one of its components are specifiable. We could now try to find the eigenvalues and common eigenfunctions of £? and L, by using the forms for these operators in Cartesian coordinates. However, we would find that the partial differential equations obtained would not be separable. For this reason we carry out a transformation to spherical coordinates (Fig, 5.5). The coordi- nate ris the distance from the origin to the point (x, y, z) The angle 0 is the angle the vector r makes with the positive z axis. The angle that the projection of rin the xy plane makes with the positive x axis is , (Mathematics texts often interchange # and ¢.) A litle trigonometry gives x=rsinBcos, y=rsindsing, z= rcos (51) z Petty+e, coso=—— yee, oa tang = y/x (552)106 Chapter 5 Angular Momentum FIGURE S.5 Spherical coordinates ‘To transform the angular-momentum operators to spherical coordinates, we must transform a/ax, a/@y, and a/@z into these coordinates. [This transformation may be skimmed if desired. Begin reading again after Eq. (5.64).] “To perform ths transformation, we use the chain rule. Suppose we have a function of 7,8, and d: f(r, 6, d). If we change the independent variables by substituting rerayz), 9=O5y%2), b= Hn y2) into f we transform it into a function of x,y, and z: Flex, ¥. 2) 00 Y 2) 6 YN] = C942) 7 cos 0 + 2 tan? 4. Using (552), we have g(x,y, 2)= ©)... OE). GG), 6 ; 3)... OB). @.G)., (2),,-GAG),, GE), GE), 6 Tome hae goons opt equations eet an tore -(9),3°@),3°@).a 59 such as (2r/ax),.. Taking the partial derivative of the first equation in (5.52) with respect toxat constant y and z, we have (654) o(@), =2-a anton (%), - saves 657)Section 53 Angular Momentum of a One-Particle System 107 Differentiating P= x + y*+ 2? with respect to y and with respect to z, we find a ar (B),,- serie Goo «9 From the second equation in (5.5), we find 0 ~sne(2) (2) -s (559) Also, (8. Qt 6m From tan 6 = y/x,we find ap sind (36) _ cosd (09 ().- Ae G)crme G), tan Substituting (5.57), (5.59), and (5.61) into (5.56), we find a 3, cos 086 3 f= sino cos 6 = a (502) alg a, cosdsing a | cosd a Fyn 8s TS tne 8 (563) a a _ sind a greg eS (5.64) At last, we are ready to express the angular-momentum components in spherical coordinates. Substituting (5.51), (5.63), and (5.64) into (5.40), we have , a sing L, alr sinosin (20s 02 r 3) a pen go , SOsBsingd a | com 2)| - + ome as + coso(sinasing 2 7 a0 t rain 30 a(sing d+ cot cos 2.) (5.65)108 Chapter 5 Angular Momentum. Also, we find in(eoso = cot sind 6.66) L.= -it o (6.67) By squaring each of £,,£,, and L, and then adding their squares, we can con- struct L? [Eq. (5.43)]. The result is (Problem 5.16) (® a, i @# ) = (+ cote 5 + 5.68) Sts aa oa Although the angular-momentum operators depend on all three Cartesian coor- dinates, x,y, and z, they involve only the two spherical coordinates 0 and 4. One-Particle Orbital-AngularMomentum Eigenfunctions and Eigenvalues. We now find the common eigenfunctions of L? and De wich we denote by ¥. Since these operators involve only @ and 4, Y is a function of these two coordinates: ¥ = ¥ (8,4). (Of course, since the operators are linear, we can multiply ¥ by an arbi- tary function of r and still have an eigenfunction of L? and L...) We must solve LY(0,) = bY, 6) (5.69) L¥(0,4) = ¥@,4) (5.70) where b and care the eigenvalues of 2, and L? Using the £, operator, we have a = ih ag Y(0. #) = bY(0. 4) (6.71) Since the operator in (5.71) does not involve 8, we try a separation of variables, writing, ¥ (6,4) = S()T(b) (5.72) Equation (5.71) becomes ne - wih 36 [S(@)T(#)] = bS(@)T(4) a ~ikS(6) me) = oS(OT(#) aT) _ ib T(#) T($) = Act (5.73) where A is an arbitrary constant. Is T suitable as an eigenfunction? The answer is no, since it is not, in general, a single-valued function. If we add 27 to ¢, we wil still be at the same point in space, andSection 5.3 Angular Momentum of a One-Particle System 109 hence we want no change in T'when this is done. For T'to be single-valued, we have the restriction T(@ + 2x) = T() Aetitgiteit = Ago eth =] (5.74) “To satisfy &* = cos a + isina = 1,we must have a = 2am, where m= 0, £1, £2, + ‘Therefore, (5.74) gives 2mbjh = 2nm b=mh, m= 2,-1,0,1,2, (5.75) and (5.73) becomes T($) = Ae™*, 0, +1, +2,. (5.76) ‘The eigenvalues for the z component of angular momentum are quantized We fix A by normalizing T. First let us consider normalizing some function F of r, 0, and ¢. The ranges of the independent variables are (see Fig. 5.5) O |m, the magnitude [J(/ + 1)]!%A of the orbital angular momentum Lis greater than the magnitude mjf of its z component L.,,except for ! = 0. If it were possi- ble to have the angular- momentum magnitude equal to its z component, this would ‘mean that the x and y components were zero, and we would have specified all three ‘components of L. However, since the components of angular momentum do not com- ‘mute with each other, we cannot do this. The one exception is when /is zero. In this, case, L? = L? + 2 + L2 has zero for its eigenvalue, and it must be true that all three components L, Ly,and L, have zero eigenvalues, From Eq, (5.12), the uncertainties in angular-momentum components satisfy anal,» 4|[vts.t,1¥ de] S\[vetv ar (5.10) and two similar equations obtained by cyclic permutation. When the eigenvalues of i, L,,and L, are zero, £,W = 0, L,¥ = 0, LW = 0, the right-hand sides of (5.110) and the two similar equations are zero, and having AL, = AL, = AL, = 0 is permit- ted. But what about the statement in Section 5.1 that to have simultaneous eigenfunc- tions of two operators the operators must commute? The answer is that this theorem refers to the possibility of having a complete set of eigenfunctions of one operator be cigenfunctions of the other operator. Thus, even though £., and £., do not commute, it is possible to have some of the eigenfunctions of L. (those with = 0 = m) be eigen- functions of L,. However, it is impossible to have all the L, eigenfunctions also be eigenfunctions of Z..Section 5.4 The Ladder-Operator Method for Angular Momentum 115 FIGURE 5.6 Orientation of L, Since we cannot specify L., and L,, the vector L can lie anywhere on the surface of a.cone whose axis is the z axis, whose altitude is mf, and whose slant height is Vill + 1) h (Fig. 5.6). The possible orientations of L with respect to the z axis for the case / = 1 are shown in Fig, 5.7. For each eigenvalue of £?, there are 2/-+ 1 different cigenfunctions Y7, corresponding to the 2/ + 1 values of m. We say that the £? eigen- values are (2 + 1)-fold degenerate. The term degeneracy is applicable to the eigenval- ues of any operator, not just the Hamiltonian. Of course, there is nothing special about the z axis; all directions of space are equivalent, If we had chosen to specify L? and L., (rather than L..), we would have got- ten the same eigenvalues for L, as we found for L. However, itis easier to solve the L., eigenvalue equation because 1, has a simple form in spherical coordinates, which involve the angle of rotation # about the z axis. 5.4 THE LADDER-OPERATOR METHOD FOR ANGULAR MOMENTUM. ‘We found the eigenvalues of £? and L, by expressing these orbital angular-momentum, operators as differential operators and solving the resulting differential equations. We now show that it is possible to find these eigenvalues using only the operator com- mutation relations. The work in this section applies to any operators that satisfy the angular-momentum commutation relations. In particular, it applies to spin angular momentum (Chapter 10) as well as orbital angular momentum. vin = mao FIGURE 5.7 Orientations of. for116 Chapter Angular Momentum We used the letter L for orbital angular momentum. Here we will use the letter M to indicate that we are dealing with any kind of angular momentum. We have three linear operators Mf, M,, and M,, and all we know about them is that they obey the commutation relations [similar to (5.46) and (5.48)] (8M) = iM, (MM) = iM, (Ute Me) We define the operator 1? as inst, (5.111) BP = Nie + WE + ME (5.112) ‘Our problem isto find the eigenvalues of M? and M, ‘We begin by evaluating the commutators of M? with its components, using Eqs. (G11) and (5.112). The work is identical with that used to derive Eqs (549) and (550),and we have [8° #4.) = (iP, M,] = (WP, M,) = 0 (6.13) Hence we can have soutunsosdhertncion of Mand Me Next we define two new operators, the raising operator Wt. and the lowering operator Mt M. =m, + ist, (aia) M_=M,- iM, (5.115) ‘These are examples of ladder operators. The reason for the terminology will become clear shortly. Let us investigate their properties We have (M, + iM,)(ML, — iM) = M(M, — iN) + iN, (My — — iN A, + iM, + ME = NP — BE + LM, MM = MP N+ BM, (6.16) ‘Similarly, we find My. MM, = WP 02 AM, (17) For the commutators of these operators with M., we have (1,81) = [+ im, +a, —inm, - hit, (04 M,) M1 St, = SEM, ~ Ri, (6.118) where (5.111) was used. Similarly, we find MM_ + KM (5.119) ‘Using Y for the common eigenfunctions of M? and M.,,we have SPY =c¥ (5.120) MY = bY G.21) where cand b are the eigenvalues Operating on Eq. (5.121) with W,,we getSection5.4 The Ladder-Operator Method for Angular Momentum 117 MMY = M.bY Using Eq. (5.118) and the fact that M, is linear, we have (1, — AMY = BILLY MM ,Y) = (b + AMY) (5.122) ‘This last equation says that the function #7, is an eigenfunction of Mf, with eigen- value b + h.In other words, operating on the eigenfunction ¥ with the raising operator 'M., converts Y into another eigenfunction of M, with eigenvalue # higher than the eigenvalue of ¥. If we now apply the raising operator to (5.122) and use (5.118) again, we find similarly MH, (MRY) = (b + 2YWBY) Repeated application of the raising operator gives M(MEY) = (b + KAYMEY), k= 0,1,2,... (5.123) It we operate on (5.121) with the lowering operator and apply (5.119), we find in the same manner M,(M_Y) = (b ~ AMY) (5.124) M(MEY) = (b — kay(MEY) (5.125) ‘Thus by using the raising and lowering operators on the eigenfunction with the eigenvalue 6, we generate a ladder of eigenvalues, the difference from step to step being f: b=, b-A, by b+h, b+, ‘The functions WEY are eigenfunctions of M, with eigenvalues b + kh [Egs (5.123) and (5.125)]. We now show that these functions are also eigenfunctions of 2, all with the same eigenvalue c: MSY = (b = kAMIEY (6.126) MP MRY = cMEY, k= 0,1,2,.. (5.127) ‘To prove (5.127), we first show that 4? commutes with I, and I (MP, M1.) = (MP, Mt, © iM,) = (WP, ML) © (0, M1, We also have (5.128) If we operate on (5.120) with Wand use (5 125), we ast MEMPY NP(MIEY) (5.129) which is what we wanted to prove.118 Chapter Angular Momentum, Next we show that the set of eigenvalues of ff, generated using the ladder opera- tors must be bounded. For the particular eigenfunction ¥ with M7, eigenvalue &, we have MY = bY and for the set of eigenfunctions and eigenvalues generated by the adder operators, wwe have MY, = bs 6.130) where ¥, = MEY G.131) b= b+ kh (5.132) (Application of IM, or M_ destroys the normalization of Y, so Y, is not normalized. For the normalization constant, see Problem 10.20.) Operating on (5.130) with N7,,we have MEY, = BMY DY, = bY, (5.133) Now subtract (5.133) from (5.127), and use (5.131) and (5.112): PY, ~ MY, = CY, — bY: (MA? + MY, = (€ — (5.134) “The operator M12 + M12 corresponds to a nonnegative physical quantity and hence has nonnegative eigenvalues. (This is proved in Prob. 7.7.) Therefore, (5.134) implies that ¢ — b> Oandc!? > |p). Thus CR = by = cl, = 0, +1, 42, (5.135) ‘Since c remains constant as k varies, (5.135) shows that the set of eigenvalues b is bounded above and below. Let Bmax and bin denote the maximum and minimum val- ues Of by. Ymax ANd Yaip are the corresponding eigenfunctions: MYnax = PmasYenax (5.136) MY = BoiaYoin 6.137) Now operate on (5.136) with the raising operator and use (5.118): MYnax = Paarl Yoon ,.(0. Yas) = (boas + 9X Yous) (6.138) This last equation seems to contradict the statement that Pacis the largest eigenvalue of Mf, since it says that M1, Yq,, is an eigenfunction of M, with eigenvalue by,, + #.The only way out of this contradiction is to have M, Yquq Vanish. (We always reject zero as an eigenfunction on physical grounds.) ThusSection 5.4 The Ladder-Operator Method for Angular Momentum 119 My Yags = 0 (5.139) Operating on (5.139) with the lowering operator and using (5.117), (5.136), and (5.120), we have o= Yams = (M? — M12 — BM YYrax, = (€ ~ Bins — Abn) ¥rae Brae — Ab nax = 0 Fras + Ara (5.140) A similar argument shows that MYoin = 0 (5.141) and by applying the raising operator to this equation and using (5.116), we find Phin — RD in © ‘Subtracting this last equation from (5.140), we have Boar + Ab pax + (Pain ~ Prin This is a quadratic equation in the unknown by,,,and using the usual formula (it still ‘works in quantum mechanics), we find ax = ~Bmins brie = Prin ~ Fe ‘The second root is rejected, since it says that By, is less than By. SO Box (5.142) Prin = Moreover, (5.132) says that bax and Bqin differ by an integral multiple of A: Boar ~ Pain = My — = 041, 2500 (5.143) Substituting (5.142) in (5.143), we have forthe M, eigenvalues Snax = jh, 7 = 04,1552, (5.144) ro = =i B= hl J HWA f + Manes(F = KG ~ Wh fh (5145) and from (5.140) we find as the 47 eigenvalues c= +I, fF =OELE (5.146) Thus MPY = jj + PY, 7 = 0,41,3,2, (5.147) MY =mf¥, mjy=-j,-7 + Me FMI (5.148)120 Chapter Angular Momentum. ‘We have found the eigenvalues of M? and M, using just the commutation rela- tions. However, comparison of (5.147) and (5.148) with (5.108) and (5.109) shows that in addition to integral values for the angular-momentum quantum number (J = 0, 1, 2,....) we now also have the possibility for half-integral values (j = 0, , 1,3,...). This perhaps suggests that there might be another kind of angular momentum besides orbital angular momentum. In Chapter 10 we shall see that spin angular momentum can have half-integral, as well as integral, quantum numbers. For orbital angular momentum, the boundary condition of single-valuedness of the T(d) eigenfunctions [see the equation following (5.73)] eliminates the half-integral values of the angular- momentum quantum numbers. [Not everyone accepts single-valuedness as a valid boundary condition on wave functions, and many other reasons have been given for rejecting half-integral orbital-angular-momentum quantum numbers; see C. G. Gray, Am. J. Phys, 37, 559 (1969);M.L. Whippman, Am. J. Phys, 34, 656 (1966).] ‘The ladder-operator method can be used to solve other eigenvalue problems; see Problem 5.31 5.5 SUMMARY For a complete set of eigenfunctions to be simultaneously eigenfunctions of several ‘operators, each operator must commute with every other operator ‘The standard deviation AA measures the uncertainty in a quantum-mechanical property A, where (A.A) = (42)— (A). For the properties x and p,, we have Ar ap, > 1h. ‘The vector B can be written as B= B,i + B, j + B.k, where i, j.and k are unit vec- tors along the x, y, and z axes,and B,, B,, and B, are the components of B. The magni- tude of B is [Bj = (B2 + B2 + B2)"*, The dot product of two vectors that make an angle @ with each other is B- A = B.A, + B,A, + B.A, = [B||A|cos 8. The cross product is given by (5.29). ‘The classical-mechanical definition of orbital angular momentum is L = r x p. The operator L* commutes with LL and but L Ly and L, do not commute with one another. When expressed in spherical coordinates, the operators BL and £,, depend only on the angles 6 (the angle between the z axis and r) and ¢ (the angle between the projection ofr in the xy plane and the x axis) and not on the radial coordinate r “The ranges of the spherical coordinates are 0 to 7 for 8,0 to 2ar for # and 0 to 00 for 1r-The infinitesimal volume element in spherical coordinates is dr = r? sin 6 dr dO dé. ‘The common eigenfunctions and eigenvalues of L? and L., are given by L°¥7" = UL + 1ye¥? and L,Y} = mAY?, where the angular-momentum quantum numbers are 1=0,1,2,...,andm=—l —I+1,...,/— 1,4 and the functions ¥ (0, d) are spherical harmonics. For operators M,,.M,,,, and? = N12 + SP + SP that obey the angular- momentum commutation relations, use of the ladder operators My ~ M, + iM, and M_ = M, ~ iM, gives the possible M* eigenvalues as j(j + 1)f?,where jcan be inte- gral or half-integral, and gives the M7, eigenvalues as m ft, where m ranges from —j 10} in integral steps.