SPECIAL FUNCTIONS FOR SCIENTISTS AND ENGINEERS W. W. BELL Lecturer in Theoretical Physics Department of Natural Philosophy University of Aberdeen ‘ ae 2, D. VAN NOSTRAND COMPANY LTD MA STk5 LONDON tS PRINCETON, NEW JERSEY TORONTO MELBOURNED. VAN NOSTRAND COMPANY LTD Windsor House, 46 Victoria Street, London D. Van Nosrranp Company Inc. Princeton, New Jersey Van Nostranp REGIONAL OFFICES: New York, Chicago, San Francisco D. Van Nostranp Company (CANADA) Ltp Toronto D. Van Nostranp Austratia Pty Ltp Melbourne Copyright © 1968 D. Van Nostrand Company, Ltd Published simultaneously in Canada by D. Van Nosrranp Company (Canada), Ltd No reproduction in any form of this book, in whole or in part (except for brief quotation in critical articles or reviews), may be made without written authorisation from the publisher. Library of Congress Catalog Card No. 68-26453 Printed in Great Britain by Butler & ‘Tanner Ltd, Frome and LondonPREFACE This book is primarily intended for use in undergraduate courses by physics and engineering students in universities and technical colleges. It should, however, also prove useful to those working in other sciences, since it does not contain any specific examples drawn from the fields of physics, engineering, chemistry, etc. We believe it to be preferable to give here the principal results concerning special functions likely to be encoun- tered in applications, and to leave the applications themselves to the par- ticular context in which they arise. The book is designed for those who, although they have to use mathematics, are not mathematicians themselves, and may not even have any great mathematical aptitude or ability. Hence an attempt has been made, especially in the earlier chapters, to sacrifice brevity for completeness of argument. The level of mathematical know- ledge required of the reader is no more than that of an elementary calculus course; in particular, no use is made of complex variable theory. Equally, no attempt is made to attain a high level of mathematical rigour; for example, we use results like 1/0 = 00, 1/00 = 0, which, to be rigorous, should really be written in terms of limits. In Chapter 1 we discuss the solution of second-order differential equa- tions in terms of power series, a topic of crucial importance for the follow- ing chapters. This is because very often in applications we have a partial differential equation for some quantity of interest, and this equation may, by the method of separation of variables, be split into several ordinary differential equations whose solutions we thus require. If, as is often the case, these equations do not have solutions in terms of elementary func- tions, we have to investigate these solutions, and this is best done by the method of series solutions. These solutions lead to the definitions of new functions, and it is the study of the properties of such ‘special’ functions which is the main subject-matter of this book. Chapter 2 is devoted to the gamma and beta functions, two functions defined by integrals and closely related to one another; these functions are not only used in later chapters, but are also encountered in many other contexts. Chapter 3 is concerned with a study of the Legendre polynomials.vi PREFACE They are introduced as solutions of Legendre's equation, which very often arises when a problem possesses spherical symmetry. Such problem ean occur, for example, in quantum mechanics, clectronugnetie theory, hydrodynamics and heat conduction. Chapter 4 is devoted to Bessel functions. ‘These functions occur in a very wide range of applications, such as loudspe: fraction, the vibration of circular membranes and plates, the scattering of sound by circular cylinders, and in general in many problems in both classical and quantum physics involving circular or cylindrical lmnundaries of some kind. Kelvin’s functions, introduced in Section 4.9, are of most interest to the electrical engineer, while the spherical Bessel functions of Section 4.10 are of prime importance in quantum mechanical scattering theory. The Hermite polynomials, discussed in Chapter 5, have their main application in the quantum-mechanical harmonic oscillator, but other applications do occur. Likewise, the Laguerre polynomials of Chapter 6 are most useful in the quantum mechanical study of the hydrogen atom, but they also find applications in, for example, transmission line theory and seismological investigations. Chapter 7 treats the Chebyshev polynomials which not only are of interest for their use in polynomial approximations to arbitrary functions, but also occur in electrical circuit theory. Chapter 8 is devoted to two generalisations of the Legendre polynomials —the Gegenbauer and Jacobi polynomials. These occur less frequently than the functions mentioned above, but they do have application in various branches of physics and engineering, e.g., transformation of spherical har- monics under co-ordinate rotations. In Chapter 9 we discuss the hypergeometric function. This is the most general of all the special functions considered. Indeed, all the other special functions considered, and many elementary functions, are just special cases of the hypergeometric function. Chapter 10 contains a brief summary of various other functions which occur in applications. These are often defined hy integrals which cannot be evaluated in terms of known functions, and very often all the useful information concerning these integrals is contained in a table of values of the integral. Appendix | is concerned with the definition of Euler’s constant, encoun- tered in the chapter on Bessel functions, and Appendix 2 treats explicitly the convergence of the solutions to Legendre’s equation found in Chapter 3. ‘Vhe remaining appendices summarise certain properties of the special functions which have been proved throughout the book.PREFACE vii The problems at the end of each chapter should be considered as an integral part of the text: every reader is urged to attempt most, if not all, of the questions, Hints and solutions are provided for all but the easiest of the problems, but the reader should not seek help from the hints before he has made an effort to solve the problem for himself. The proofs of certain theorems are indicated as being able to be omitted at a first reading. This does not mean to imply that the proofs of such theorems are difficult, but rather that they are fairly lengthy and that inclusion of such a proof on a first reading would tend to destroy the con- tinuity of presentation of the material. W. W. B. Aberdeen, 1967LIST OF SYMBOLS Symbol Function Section in cohich symbol first appears Bex, y) Beta function 21 ber, %, bei, x Kelvin’s functions 4.10 C(x) Fresnel integral 10.3 Cx) Gegenbauer polynomial or ultraspherical 8.1 polynomial Ci(x) Cosine integral 10.2 D,(2) Debye functions 10.5 E(k, 4) Elliptic integral of second kind 10.6 G Complete elliptic integral of second kind 10.6 Ei(z), E(x) Exponential integrals 10.2 E,(z) Generalised error function 10.3 erf x Error function 10.3 F(R, $) Elliptic integral of first kind 10.6 mF nts 25» + + Om By Bay + +» Bos) Hypergeometric function 91 H,(x) Hermite polynomial 5.1 H,.0(x), H,,(x) Hankel functions or Bessel functions of the 45 third kind fiq(x2), h(x) Spherical Hankel functions 411 F(x) Modified Bessel function 47 Tal) Bessel function of the first kind 44 Fal) Spherical Bessel function 411 K(k) Complete elliptic integral of first kind 10.6 K,(x) Modified Bessel function 47 ker, x, kei, x Kelvin’s functions 4.10 L(x) Laguerre polynomial 61 IMs) Associated Laguerre polynomial 67 HG) Logarithmic integral 10.2 M(z, B, x) Confluent hypergeometric function O41 Memn(2) Whittaker’s confluent hypergeometric 93 function Ny (2) Neumann function or Besscl function of the 44 second kindSymbol Pix) Pre) Py P(x) Qx) Sx) si(z), Si(x) T@ TGs, 0), 965, o) y S) I(k, $, a) II(k, a) Pax) List OF SYMBOLS Function Legendre polynomial Associated Legendre function Jacobi polynomial Legendre function of the second kind Fresnel integral Sine integrals Chebyshev polynomial of the first kind Chebyshev polynomial of the second kind. Spherical harmonic Bessel function of the second kind Spherical Reasel function Gamma function Incomplete gamma functions Euler’s constant Riemann’s zeta function Elliptic integral of third kind Complete elliptic integral of third kind Weber-Hermite function Section in which symbol first appears 31 3.8 8.2. 3.10 10,3 10,2 7 7A 3.1 41 4. 24 10.1 4. 10.4 10.6 10.6 5.721 2.2 23 24 3.5 3.6 3.7 3.8 3.9 CONTENTS PREFACE List or SyMBoLs CHapPTeR 1 Series SOLUTION OF DIFFERENTIAL EQuaTIONS Method of Frobenius Examples Problems Cuaprer 2 Gamma AND Bera Funcrions Definitions Properties of the beta and gamma functions Definition of the gamma function for negative values of the argument Examples Problems CrapTer 3 LeGENDRE PoLYNOMIALS AND FUNCTIONS Legendre’s equation and its solution Generating function for the Legendre polynomials Further expressions for the Legendre polynomials Explicit expressions for and special values of the Legendre polynomials Orthogonality properties of the Legendre polynomials Legendre series Relations between the Legendre polynomials and their deriva- tives; recurrence relations Associated Legendre functions Properties of the associated Legendre functions 3.10 Legendre functions of the second kind 3.11 Spherical harmonics 3.12 Graphs of the Legendre functions 3.13 Examples Problems. 4 30 37 42 46 48 50 52 55 58 62 65 70 82. 85.xii CONTENTS CuapTer 4 Besser. Functions 4.1 Bessel’s equation and its solutions; Bessel functions of the first and second kind 4.2. Generating function for the Bessel functions 4.3 Integral representations for Bessel functions 44 Recurrence relations 4.5 Hankel functions 4.6 Equations reducible to Bessel’s equation 4.7. Modified Bessel functions 4.8 Recurrence relations for the modified Bessel functions 4.9 Integral representations for the modified Bessel functions 4.10 Kelvin’s functions 4.11 Spherical Bessel functions 4.12 Behaviour of the Bessel functions for large and small valucs of the argument 4.13 Graphs of the Bessel functions 4.14 Orthonormality of the Bessel functions; Bessel series 4.15 Integrals involving Bessel functions 4.16 Examples Problems Cuaprer 5 Hermite PoLyNoMIALs 5.1 Hermite’s equation and its solution 5.2 Generating function 5.3. Other expressions for the Hermite polynomials 5.4 Explicit expressions for, and special values of, the Hermite polynomials 5.5 Orthogonality properties of the Hermite polynomials 5.6 Relations between Hermite polynomials and their derivatives; recurrence relations 5.7 Weber-Hermite functions 5.8 Examples Problems Cnapter 6 Lacuerre PotyNoMiars 6.1 Laguerre’s equation and its solution 6.2 Generating function 6.3 Alternative expression for the Laguerre polynomials 6.4 Explicit expressions for, and special values of, the Laguerre polynomials 92 99 101 104 107 108 110 113 116 120 121 127 133 137 141 148 154 156 157 158 160 161 162 163 164 166 168 169 170 17165 6.6 6.7 68 6.9 6.10 7A 72 73 74 75 8.1 8.2 8.3 91 9.2 9.3 9.4 10.1 10.2 10.3 10.4 10.5 10.6 10.7 CONTENTS Orthogonality properties of the Laguerre polynomials Relations between Laguerre polynomials and their derivatives; recurrence relations Associated Laguerre polynomials Properties of the associated Laguerre polynomials Notation Examples Problems CHAPTER 7 CHEBYSHEV POLYNOMIALS Definition of Chebyshev polynomials; Chebyshev’s equation Generating function Orthogonality properties Recurrence relations Examples Problems CHapTeR 8 GEGENBAUER AND JACOBI POLYNOMIALS Gegenbauer polynomials Jacobi polynomials Examples Problems Cuapter 9 Hyperczometric FuNcTIoNs Definition of hypergeometric functions Properties of the hypergeometric function Properties of the confluent hypergeometric function Examples Problems Cuapter 10 Orxer Spectra FUNCTIONS Incomplete gamma functions Exponential integral and related functions The error function and related functions Riemann’s zeta function. Debye functions Elliptic integrals amples Problems xiii 172 173 176 177 182 182 185 187 190 192 193 194, 196 197 198 200 201 203 207 210 212 216 218 218 221 223 224 224 225 228CONTENTS APPENDICES: 1 Convergence of Legendre series 2 Kuler’s constant 3. Differential equations 4 Orthogonality relations 5 Generating functions Hints anp SoLutions To PRosLems BiptioGRaPHY INDEX 230 231 233 234 236 237 243 245SERIES SOLUTION OF DIFFERENTIAL EQUATIONS 11 METHOD OF FROBENIUS, Many special functions arise in the consideration of the solutions of equations of the form dy dy PO) + OE + Rey = 0. ce) We shall restrict ourselves to equations of the type di dy 2 3 4 agin + r(aly <0, (12) where g(x) and r(x) may be expanded as power series in x, (1.3) (1.4) convergent for some range of x including the point x = 0. The basis of Frobenius’ method is to try for a solution of equation (1.2) of the form 2(8,8) = (aa tage ast? +... tant +...) =D ae'™ (15) = with ay <0. (We may always take a “0, since otherwise we should just2 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 have another series of the type (1.5) with a different value of s, the first coefficient of which we could now call ay.) From equation (1.5) we have MS age i ae dv SS dz 2 and ua Date bans fom Iwtn 8, so that if we require = to satisfy cavation (1.2) we must have as aot g(x) in 4 r(x)e = 0, and this reduces to S als bem)(s em Txt 4 g(a) Day(s fm) att a0 ao + r(x) > anti” = 0 2, which, on using equations (1.3) and (1.4) for (x) and r(x) and cancelling a common factor of x*, becomes > a,(s + n)(s +m — x + S > nds + x" ab ae. (1.6) For the infinite series on the left-hand side of equation (1.6) to be zero for all values of x in some range, we must have the coefficient of each power of x equal to zero. This requirement gives rise to a set of equations as follows. Requiring that the coefficient of x° be zero, and noting that x® arises only from the choice m = 0), 1 = 0, gives ags(s — 1) + aogos + aot = 0. (1.7) Requiring that the coefficient of x* be zero, and noting that x? arises in equation (1.6) by choosing in the first term m = 1, and in the second and third terms n -- m= 1 (ie., m =0, n = 1 or m= 1,0 =0) gives a(s + Us + {argos + 1) + aos} + {arto + agrs} = 0. (1.8) We may now write down a general equation from the requirement that ient of xt should be zero. Remembering that in equation (1.6) x¢ 0 the first term by choosing n = # and in the second and third terms§1t METHOD OF FROBENIUS 3 by choosing 1 -+m=i (ie, n=i, m=0 or n n=i—2,m =2, etc, up ton = 0, m =i) we obtain a(s + is +%—1) i—1, m=1 or {ag +2) + aeag(s +t = 1) + aags +4 — 2) +... + aoges} + {ato taant-.. tarp =0 G> 1). (1.9) ‘Thus af(s +i) + fa agls + i —1) + @-2g,s +t — 2) +... + aos} Havars + aeata boos + ar} = 0 @>1) (1.10) where we have collected all the terms in a, together and denoted their coefficient by Als +i) =(5 | ils + 8-1) + als +i) +70 (uy We now sce that equation (1.7) reduces to ao{s* + (Go — 1s + ro)} = 0 which, with the given assumption that 4, 0, becomes s* + (Go — Ils +70) = 0. (1.12) ‘This is called the indicial equation; it is quadratic in s and hence will yield two roots which we shall denote by s, and sg. In many cases of interest these roots will be real; this we shall henceforth assume, together with the fact that 52 > s. Noting that equation (1.12) is just f(s) = 0, we have immediately SS) = (8 — sl ~ 5). (1.13) We might now expect that these two values of s would lead to the two independent solutions of the original differential equation. We shall see that this is true apart from certain exceptional cases, viz. when (a) the two roots of the indicial equation are equal; or (b) the two roots of the indicial equation differ by an integer. The set of equations (1.10) can now be used to determine the coefficients @, dy, ... in terms of ao. Fquation (1.10) with # = 1 gives af(s +1) + aegis + aor, = 0 so that we have ans +) fs +1) a= (1.14) where f(s) ~~ (q:5 | 4) is a polynomial of first degree in s, sew4 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS CH. 1 Equation (1.10) with 7 = 2 gives aaof(s + 2) + (aagi(s | 1) + aogas} + fears | aera} — 0, which, when we use equation (1.14) for a, becomes aaf(s +2) 4 (29 ot 11) + aus} + G29 + onl yielding ay = ok als + Wh(s) + Qesf(s + 1) + rn(s) + refs +} ° f(s + YFG + 2) _ hfs) = FeLi) fG $3) (1.15) where hi(s) = —{ar(s + 1ha(s) + gasf(s + 1) + raha(s) + ref(s + 1)} is a polynomial in s. In general it is not difficult to see that we obtain ana hs) FS FS TFs + 2)..-fls + 8) where As) is a polynomial in s. If in equation (1.16) we now substitute s = s, we shall obtain expressions for a; in terms of dp, leading to a solution with ay as an arbitrary multipli- cative constant; similarly we may substitute s = s,, so that we have the two solutions required. This will work, however, only if the denominator of equation (1.16) is never zero, i.e., provided f(s, +i) #0 and ‘Fs +4) 0 for ¢ equal to any positive integer. (1.17) Now, equation (1.13) tells us that f0) = (6 — sls — 5) fle +i) = (6 +85 +8) (1.16) so that and hence fia +i) =i, — 5 +i) and Plu -+ 8) = (52 — 55 +i), (1.18) so that equation (1.17) will be satisfied if, and only if, s, — 5, is not a positive integer, i.e., provided the roots of the indicial equation do not differ by an integer. (We recall that we have labelled the roots so that Sa > Sy) If the roots do differ by an integer, the procedure given above will work for the larger root: if s, — s, = r (a positive integer), then from equations (1.18) we have$i METHOD OF FROBENIUS 5 fs +i) =i{-r +4) (1.19) S(sa + i) = i(r 4-1), (1.20) and we see that f(s, +2) =0 for ir, while f(s, +i) + 0 for any positive integral 7. Hence the procedure will work with s = s, but not with 4 = 5,. How, in fact, do we find the second solution? If we use the relation- ship (1.16) before fixing the value of s we have the series i sa ae) ag NY ESD IED his) ' Fe TIC+ 2-H +o} (121) We know by the method of construction of a), ag... that if all the terms on the right-hand side are well defined, then this must satisfy te + sala) tor (x)z = aof(s)x* (nince equations (1.10) are satisfied, the coefficient of each power x*## with é> 1 must vanish, and the coefficient of x is, by equations (1.7) and (1.11) just ao/(). This equation may be rewritten in the form d? d {or B+ aot + rephetn) afte (1.22) Unfortunately, all the terms on the right-hand side of equation (1.21) are not well defined for s = s,; there are zeros in the denominators for i> r. However, if we multiply 2(x, s) by f(s + 1) we shall just cancel the factor in the denominators of the later terms which vanishes when s = s,. And since f(s +1) =(s +r —s)(s tr — ) =(s + 5, — 25)(s — 5) we might just as well multiply by (s — s,). Also, since this factor is inde- pendent of x, equation (1.22) now takes the form {3 + sao +1006 —anete9) = alo — ayo dt + HDG, +76 ats 9) = als ~ 8, = als — 5X5 — 9) (1.23) using equation (1.13). Setting s = s,, we have {a + sat. + reohtes ~ sdats 5) = 0 which shows that 2(7, s2) is a solution, a fact we already know. Similarly, setting s-- s, gives [(s — s,)2(x, s)], ., a8 a solution, In fact6 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS CH. 1 it turns out that this series is not independent of 2(x, s,) but is just a multi- ple of it. This comes about as follows. The factor s -- s, cancels zeros in the denominator for terms with i > r, but will provide zero in the numerator for all terms with i < r, thus the first power in the series is just x**" = x*, which is just the first term in 2(x, s,). And since we use the same rules for calculating any coefficient in terms of preceding ones in the two cases, we must just obtain the same series in both cases, apart from a constant multiplicative factor. If we differentiate both sides of equation (1.23) with respect to s, we obtain {od + sod +109} [4 to —sdet.99] - asl —s)* £@ — 8:)x*} + 2(s — g\(s — se (1.24) and we see that the right-hand side of this equation is zero when s = s,, so that we must have [(d/ds){(s — s1)2(x, 5)}],..5, a8 a solution of the differential equation. It may be proved that this solution is in fact independent of the first solution (i.e., it is not merely a constant multiple of the first solution), but we shall not do so here. ‘Thus we can take as independent solutions [5 ~ s)2(% sea, (1.25) d and [Ste - spate ot]. (126) Another type of situation may arise when the roots of the indicial equa- tion differ by an integer. As well as f(s +i) = 0 for s = s, and i =r, it may happen that h,(s,) — 0. In this case a, is indeterminate (since it has a zero in both numerator and denominator) so that we may in fact use it as another arbitrary constant, thus obtaining the two independent solutions from the one series. We shall see in detail how this happens when we come to consider particular examples. ‘The one remaining case is when the indicial equation has equal roots, say s = s,. Here f(s) = (s — s,)?, and if we make use of equation (1.22) we obtain d® d {x qa t Cz, 4 Hahats 8) = a(s — s,)?x*. (1.27) Setting s = s, makes the right-hand side zero, so that 2(x, s,) is a solu- tion. But if we differentiate both sides with respect to s, we obtaingia METHOD OF FROBENIUS 7 fom tno} [ee] = aelage — s,)x* + (s — 51)? Sel. (1.28) Again the right-hand side is zero when s = s,, so that we have for a solution [(d/ds){a(x, 5)}]_...,- This, in fact, may be shown to be independent from 2(x, 5), so that in this case we have the two independent solutions (5% 5) and [Sa dh (1.29) Let us now summarise bricfly the methods of obtaining independent solutions which have been described above. Form the series 2(x, s) by the use of the relations (1.10). Then there are four distinct cases. (1) If the roots s, and s of the indicial equation are distinct and do not differ by an integer, then the two independent solutions are given by a(x, 51) and a(t, 2). (2) If the roots s, and sy (s, > s,) differ by an integer and one of the coefficients in the series for 2(x, s) is infinite when s = s,, the two indepen- dent solutions are given by [6 — ses, Mann, and [Ete -se.03]. (3) If the roots s; and s, (s, > s,) differ by an integer and one of the coefficients in the series for 2(x, s), say a, is indeterminate when s — s,, the two independent solutions are obtained from 2(x, s,), keeping a, and a, as arbitrary constants. (4) If the roots of the indicial equation are equal, say, s = s,, then the two independent solutions are given by a(x, 5,) and [fee do. "Ihe question of convergence of the series obtained must, of course, be considered. It may be proved (although we shall not do so here) that the range of values of x for which the solution is convergent is at least as large8 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS CH, 1 as the range of values of x for which both q(x) and r(x) may be expanded as convergent power series. In certain cases q(x) or (x) may not be expanded as power series in x for any range of x (e.g., 1/x or exp (1/x)). It may perhaps then be possible to make a change of variable, such as x’ = x — a, or x’ = 1/x, so that q(x) and r(x) may be expanded in powers of x’ and a solution found in the form (1.5) with x" replacing ». Finally, let us remark that when carrying out the actual solution of a differential equation of type (1.1), it is not usually profitable to transform it to an equation of the type (1.2); all that we require for the above results to hold is that this transformation should be possible. Let us now illustrate the above methods by means of some examples. 1.2 EXAMPLES Example 1 dy bay TY =O We must first verify that this equation is of the form aty dy 2d = Ga + aaa)g, t rely =0 where q(x) and r(x) may be expanded as power series. Multiplied by +/2, the given equation becomes dy tidy to aot 2rd TI 9 so that g(x) = 4, r(x) = 4x and the required condition is obviously satisfied. Since g(x) and r(x) are already in power series form, valid for all values of x, it follows from the remarks made at the end of the previous section that any series solutions obtained will be convergent for all values of x. We now revert to the original equation. ~ 2 Set a= > agen = > ager" = n=0 so that de S a,(s + nxt a and ia =D ae tale ea Tere, a§1.2 EXAMPLES 9 Then we have s 2a,(s + n)(s fn — test > a,(s + n)xtint S Day(s -} (s+ m= ARE LS ag(s + nator noo =o I + D> aaa, (1.30) & where in the last summation we have replaced the label » by the label n — 1 in order that the general power should look the same in all terms, In order that 2 should satisfy the differential equation, we must have the coefficient of each power of x in (1.30) equal to zero. The power with n = 0 appears only in the first two terms; thereafter the powers with n = | appear in all three terms. Hence we have the equations: 2a(s + Os +0 —1) + als +0) =0 (1.31) and 2a,(s + n)(s + —1) + a(s +2) +a). = 0 (n> 1). (1.32) ‘These equations simplify to as(2s — 1) =0 (1.33) and a,(s + n)(Q(5 +0) ~ 1} + dy = 0. (1.34) Equation (1.33) gives the indicial equation s(2s —1) =0 with roots s = 0 and s = $. Since these are distinct and do not differ by an integer, we know that we may proceed according to prescription (1) on page 7. Equation (1.34) gives the recurrence relation for the coefficients in 2(x, 8): ana oF As +2) 1 (1.35) and hence a, a © (s-F 125-1 1)10 SERINS SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 Ho — P64 22843) (5 + Ils + HRs + 1)(2s + 3)’ a a 4, “CFDR +5) and in general =( 1) ___% — om = (“WES DESD...G mG +R +3). "Thus we have the series 2(x, s) given by (s + 1)(s + 2)(s +} 3)(2s +- 1)(2s + 3)(2s + 5) 2 a(x, 8) = age ~ | ome aay te as, 8) = aor? { GMs + CHG +H +3) 7 ” (AY GaG53) mye b143) 0. @erQnay ) (1.36) Since the roots of the indicial equation are s =: 0 and s = 4, the two independent solutions are given by 2(x, 0) and 2(x, 4) which, by equation (1.36), are x xt x8 : a=) = ast “iitih2d3 paar ts +(e aD +} (1.37) and 2(x, 3) = ot 3, 3 fa BB oae... cee ' 7 PEN 6. On » (1.38) We may rewrite these series more compactly if we note the following results: 1.2.3....m=n, (1.39) 1.2.3.4...(2n —1).20 1.3.5...0n N= 9 on 0.20 (2n)! ~ 271.2.2.2.3...2(n — 1).20 _ ny! Pat (140)§12 EXAMPLES RSF MEE a 53.5.7... Qn 41) 1.2.3.4.5.6.7...(2n +1) “Oe 24.6.... 20 1 Qn +1)1_ n+)! SO Rent Dal? and 2.4.6...2n = 2m). Using these results we sce that equation (1. 7 reduces to me) =% 2 (1° ii a (2%)" =o on ai while equation (1.38) becomes a(x, 3) = aye? 2" a = Dea & 2 nae" 2 (i ee er (1.41) (1.42) (1.43) (1.44) ‘Thus the general solution is given by the gencral linear combination of 2(x, 0) and 2(x, 3), viz. 2x)" (24) Ay (-1y aa y= x 1) (ay + Be a Yona pi where A and B are arbitrary constants. Example 2 dty dy _ al =a) S24 (J —y 0. We first verify that this equation is of the form d 4 x? bed + ate) +r(x)y =0 where q(x) and r(x) may be expanded as power series in x. ‘The given equation is obviously equivalent to ely ye ae Xde 1 eh12 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 Hence q(x) = 1 and r(x) = —x/(1 — x). 9(x) is already in power series form, convergent for all values of x, but r(x) is not. However, it may be expanded as a power series in x by the binomial theorem, and this series will be convergent for those values of x such that |x| <1. It follows from the remarks at the end of the previous section that any series solutions which we obtain will be convergent for at least the same values of x, viz. |x| <1. We now revert to the original equation. Set > agg «eS so that En 2 a,(s + n)xtint a < . and Get 7D, oe bas Hm Dees Hence, d?z dz ea ge + ae — 8 = S a,(s-}n)(s+n—1)xtin—? — > a,(s-n)(s+-n—1)xtt" Sa S + > alstayart — Salona —S ager® & < & « 2 = > als tay staat — >. anal6-bn—1)lo-tn Dera S 4 + > als taper? — Yay (etn —Tpereet — Dag at 4S & A For 2 ta he a solution, we require the coefficient of each power of x to he zero, so that we have ags(s — 1) + aps =0 (1.45) and a,(s + n)(s +n — 1) — ay_(s + 2 — 1)(s +2 — 2) + a,(s +2) an a(s tn —1) ~an~=0 (n> 1). (1.46) "These equations simplify to ays? = 0 (1.47)g12 EXAMPLES 13 and a,(s +n)? ~ ay s((¢ +n — 1)? + 2 (1.48) Equation (1.47) gives the indicial equation s%=0 (1.49) while equation (1.48) gives the recurrence relation (in = 1+ (s + )* Equation (1.49) has coincident roots s = 0; so we know (by prescription (4) on page 7) that the two independent solutions are given by 2(x, 0) and [(d/ds)2(*, s)] no Equation (1.50) gives Gy = Any (1.50) at), 1 E41 a a CAMA | HIG ++ BN 5 + 28 G++ 2)?” and in general af tM $ DED G12 Ete = PED 6+ AG +2). Fa so that - (Ah (e402 t1)...een—1t41} aes) =a (HGH... Ga)? ~} (1.51) ‘The first solution is then 1124-1} 224-1} 3241} a(x, 0) =a + S 2 1.2.5.10.17... = a,[1 +> Pe coe . = ayy,(x), say. (1.52) For the second solution we require (d/ds)2(x, s). Now, from equation (1.51) we have {(a- 1)P+1} =| 4 Re) 4 Se fet} (6441p... 6+n—1)?+41} = al a dl: + 2% (et) 4e42)". eR -| wd AV ME. s tn-Netiy] | oe 2 ie i (O42)... (Sent ] (153)14 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS — cli. 1 For the first term we note that du 4 deny — 4 emt = (In) ef ™* = (In xx (1.54) and hence [f-] ins. (1.55) ‘The second term we evaluate by the technique of logarithmic differen- tiation. If we denote the term in square brackets by f,(s) we have d ld gs nals) “hO ate) 30 that Sale =fs) fan Ko. (1.56) But In f(s) = In fs? + 1} + In (5 + 1)8 + 1} +... + In (6 4 198 + 1} —In(s + 1)*—In(s + 2)*... —In(s | m)? = S neem 1) 41) — S ing + mys Hence a x s+m—1 1 =? 2 feta ri - tah so that a ah OG 1m PEN Gf stmt ey eee 2 and [S00], _ A? + 2? + 2 12.22. 4 We recall that ~ el" 4, where we are using the notation In.x loge x.§1.2 EXAMPLES 15 Denoting this expression by ¢,, we have 2.1.2.5.10.17. {= MNS ma? o~ wi Law ry oO ‘Thus, setting s = 0 in equation (1.53) gives the second solution d [aa], = a 2 — 1)2 = agin [1 | Dy (e+ He rn {o 1)? + Be] ay > eat” cot = agyi(a) Inv + ay > ea" at = ayya(x), Say, where ¥alX) = ya(%) Ine + > Cn". (1.58) = ‘Then the general solution is given by ¥ = Ay(x) + By.(x) where A and B are arbitrary constants. Example 3 wd s 32 4 8 — ayy =o. We see immediately ee this is of the form (1.2) with q(x) = r(x) = 3 — x, which are already in power series form, so that they have a power series expansion valid for all values of x. It follows from the remarks at the end of the previous section that any serics solutions obtained will be convergent for all values of x. Writing z = S° a,x leads in the same manner as in the previous examples to the following equations, which have to be satisfied if z is to bea solution of the given equation: age — 1s —3) = 0 (1.59) alsin West a 3) aa--O Ml). (1.60)16 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 Equation (1.59) gives the indicial equation (s — 16 — 3) =0 (1.61) with roots s = 1, s = 3. These differ by an integer, so that we know that we are dealing with an exceptional case. Equation (1.60) gives the recurrence relation By -— ae 1 1 = (6a Ts tay FD (1.62) so that we have a= fe = ay o, ..,.,.. = EFD6—1 s+ DE —2s—N and in general ay a SF IG + 2)... (G+ m — IS — Ys s...(6 +n — 3) (1.63) We see explicitly from equation (1.63) that when s =: 1 all the a, with n > 2are infinite, so that we have to apply method (2) described above on page 7. We have . 8 #(e 8) = aoe + Dea eG Da so that (s — 1)s( (2,8) = ages — nee =}. ~ xt +2 asp ibn Is 2)e+l)...@ bm af (1.64) ‘The first solution is given by Us — 1s@ sent » ax =a > (2... m(—1).1.2...(n — 2) 2. Perel = 2 ate al E = ay(x), say. “The second solution is given by ((d/ds){(s -~ 1)3(x, 8) }Jeaa-§12 BXAMPLES 7 From equation (1.64) we have See — nate} = a4) fe “1+ £ — a 2 xt +> 0. (Fn — 16 — 2s + 1) s( + tte +5 ds(s — 2)" To evaluate the last derivatives we use logarithmic differentiation: denoting 1 a4 1)... @ bn —1\6—2)s +1)... +n —3) by f(s), we have again $10) = fle) 0. fs) and since Info) = Ins —In( +1) In +2)... ~Ine +m —1) —In(s — 2) — Ins — In(s +1)... —In(s +n — 3) we have Sin fo18 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 1 d he = oa ge an U(r .n(—l).1.2...(7—2) 1 ~~ alfa — 2) so that d 1 C1 1 1 [4.00] 1 me Qn ares +3 = Gy say. It is also easy to show that esa). ds s(5 2) Jenn os = dn a)or ds ‘ and, as before, we have ‘Thus, finally, we have from equation (1.65) ; [Ste — de.93],_ | = extn ype) + ase{t x + > oa} = aaya(x), say, and the general solution is then y = Ayx(x) + Bys(x) where A and B are arbitrary constants, Example 4 dey dy 2D 5 (gt — any — ay = xtjga f(t — 2a) 2y = 0. Again this is of the form (1.2) with q(x) —2 + «and (x) = —2, so that the series solutions to be obtained will be valid for all values of x. Writing «= x* > a,x" and requiting = to be a solution of the given a equation leads, as before, to the system of equations: ags — 2's — 1) =0 (1.66) a(s = 1)s = (1.67) a(s-+n—2Bstn W4-afs tn 2)=0 (w>2). (1.68)§12 EXAMPLES 19 Equation (1.66) gives the indicial equation (6 ~ 26 —1) =0 with roots s = 2, s = 1. These differ by an integer, so we are dealing with an exceptional case. When s = 1, equation (1.67) is satisfied irrespective of the value of ay; i.e., a, is indeterminate, so that we are dealing with case (3) of page 7. We then know that the two independent solutions are obtained from the one value of s, namely s = 1. Equation (1.68) gives the recurrence relation (1.69) (the factor s + — 2 may be cancelled since it is non-zero for n > 2 and s =1ors = 2), which with s = 1 becomes a, =~ (n > 2). (1.70) a a ‘Thus a = 4% ge @K a= z as 5 and in general = (—1) a am = (Il Fe Ba a and aan = (I 59 Ge pty so that we have s(x, 1) = x[acfi + 2a} a> Ys ae no ‘That is, we have the general solution given by ¥ = Ay(x) + Bys{x) where 2. aint yn@ =e + > (ye " 2 2.4.6...20 2 inst ae) => CYT 33 Ge) and 4 and B are arbitrary constants.20 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS CH. 1 In the examples given so far, it has always been relatively easy to write down the general term of any series appearing in the solutions. This is by no means always so; it may be impossible for us to find in any simple manner an expression for the general coefficient from the recurrence relation. This is so, in general, when the recurrence relation contains three or more terms instead of the two they have had up till now. All that we can do in such a situation is to give the first few terms of the power series solution, and hope that this will be of some use. We illustrate by the following example. Example 5 d?y dy 2 iyi ee eb gf) (ot - Ny + Seg, hay = 0. Rewriting this in the form 2 3 ety, 3 ye 0, det et 1 de PD we see that it is of the form (1.2) with (x) = (3x%)/(x? — 1) and +(x) = x9/(x? — 1). Both q(x) and r(x) may be expanded as power series in x (by the binomial theorem) convergent for |x | <1. Hence any series solutions we obtain will he convergent for at least this range of valnes of x. We now revert to the original equation, and trying for a solution of the form a= Sag leads, in the same manner as in previous examples, to the set of equations ays(s — 1) =0 (7b as + 1)s=0 (1.72) ags(s + 2) — a,(s + 2)(s + 1) =0 (1.73) Ogg + Ana (5 + 2 — 2) — ag (s+ al(o-+-n—1)=0 (n> 3). (1.74) Equation (1.71) gives the indicial equation (is —1)=0 with roots s =0 and s = 1. Equation (1.72) then shows that a, is inde- terminate if s = 0, so that we are dealing with case (3) above—the two independent solutions may both be obtained from the root s = 0, taking @, and a, as arbitrary constants.PROBLEMS 21 With s 0, equation (1.73) gives a,=0 and equation (1.74) becomes ya + (1 — 2)ay_y — n(0 — Vay = giving Ong +00 — 2)» a= n(n — 1) It is impossible for us to use the recurrence relation (1.75) to obtain a general expression for a,; but we may proceed to calculate as many terms as we please. (n > 3). (1.75) Thus a = t 3m 6 = dao + da, +8 a = bay + 8a, and so on as far as we please. We have calculated the series solution up to terms in x*: = ay + ay + (So + ban)es + tease! + (bao + 8a,)a° + afl +t bee) + ayfoe + 4x? + nt pe te PROBLEMS (1) Obtain solutions of the following differential equations in ascending powers of * stating for what values of x the series are convergent: dty | dy . O Sas" ae 7 = 8 se dy (ii) 3 +@tD ty =0 (iii) af 2 Fy =0522 SERIES SOLUTION OF DIFFERENTIAL EQUATIONS cH. 1 (iv) 9x(1 — oe - 12 4 4y =0; oy 1 2982 4B 4» <0; . (vi) a a“ + (et we + (4 — 2a)y =05 (wii) (Q— ee 4+e%_y <0; (viii) #2 -~ ae + (x3 — 5)y =0; . 4 (ix) ed a —y =0. (2) Determine whether or not it is possible to obtain solutions of the following equations in terms of ascending powers of x: i) Be =0; y Or aera xy =05 (iii ee Sead yy = 0. (3) Obtain solutions of the following equations in terms of ascending powers of 1 /x: (i) 2n%(% — i + x(3x + vg —2y =0; Gi) 2004 ed + (2x + a +2y =0.GAMMA AND BETA FUNCTIONS 2.1. DEFINITIONS We define the gamma and beta functions respectively by: Va) = [ewe (2.1) ° Bix, y) = {. P-L — pede. (2.2) The first definition is valid only for x > 0, and the sccond only for x > 0 and y > 0, because it is for just these values of « and y that the above integrals are convergent. We shall not prove this statement, but shall at least make it plausible. Consider the integral in definition (2.1). It is known that at infinity the behaviour of an exponential dominates the behaviour of any power, so that e~' f°-1-» 0 as t > oo for any value of x, and hence no trouble is expected from the upper limit of the integral. Near the lower limit of the integral we have e~' = 1; hence if this approximation is good between t = 0 and t =c, we may write equation (2.1) in the form T(x) = f etde 4 vr e-4e-t det = [el + r en4?-1 dt, ode so that if the first term is to remain finite at the lower limit we must have x>0. ‘A similar argument applies to the beta function, the behaviour at ¢ — 0 leading to the restriction on x and at ¢ -- | to the restriction on y.24 GAMMA AND BETA FUNCTIONS cH. 2. 2.2. PROPERTIES OF THE BETA AND GAMMA FUNCTIONS Theorem 2.1 Tl) = 1. PRoor We have, from the definition (2.1), rd) = i enl-t de Theorem 2.2 Tw +1) =a (a). (w > 0) Poor T(x +1) = fj wat 7 (by the definition (2.1)) = [cove]? ~ FP erga dt 0 Jo (on integrating by parts) =0+% [rete at ° (where the first term vanishes at the upper limit since the exponential dominates the power, and at the Jower limit since x > 0) = xT(x) (using definition (2.1) again). Theorem 2.3 If x is a non-negative integer, then T(x + 1) = x! ProoF P@ +1) = «P(x (by theorem 2.2)§2.2 PROPERTIES OF TIE BETA AND GAMMA FUNCTIONS 25 = a(x — 1)0( ~- 1) (using theorem 2.2 again) = a(x — 1)(x — 2)0(w — 2) (by further use of theorem 2.2) = x(x — 1)(~ — 2)(x — 3)... 3.2.19) (by repeated use of theorem 2.2, and remembering that » is integral, so that continued subtraction of unity will eventually lead to 1) xIT(1) x! (by theorem 2.1). Theorem 2.4 T(x) = 2 re eo M1 dt, ° PRoor In definition (2.1) substitute u? for t: ¢ = u2, so that dt = 2u du; when t —0, « —O and when t = 00, « — oo, so that we have Te) = \. eM w)F-4.2u du =2 |" emu du ° =2 { nha dt ° (since in a definite integral we may choose the variable of integration to be anything we please). Theorem 2.5 we is TO)P(y) cos’?! 9 sinv-19 do = L ) i cost" sint19 00 = VO). PRooF We prove this result by considering in two different ways the double integral T= ff exp (-08 ~ wy dt du k (where 2 is the first quadrant of the fu-plane, shown unshaded in Fig. 2.1).26 GAMMA AND BETA FUNCTIONS cH 2 Fic. 2.1 The tu-plane Firstly, we have I= f Pr exp (—#% — utye2—1ut—1 de du ot u-0 I oe ew j en pte at. | eu) du ° ° aM(x).4T) (using theorem 2.4) = WET). (2.3) Next we change variables to plane polar co-ordinates r and 0 in the tu-plane so that 1 =r cos 0, u =r sin @ and the element of area dt du is replaced by the clement of area r dr d9. ‘Then we shall have Ie J exp (—r*.cos?@ - 7 sin” 6)(r cos 0)"*—"(r sin 8)"—¥r dr dO 2 = P fe e771 cos 9 72-1 sin 6 r dr dO raol ono = i en rte dy j * cos? 9 sin—1 9 dO ° ° 1 =1T@ | »f cos*#-1 9 sin®”-1 9 do (2.4) ° (using theorem 2.4 again). Equating the two expressions (2.3) and (2.4) for I leads immediately to the required result, which will be true for x > 0 and y > 0, since it is for this range of values that the gamma functions involved are defined.§2.2 PROPERTIES OF THE BETA AND GAMMA FUNCTIONS 27 Theorem 2.6 TS) — vin PROOF Put x = y = } in theorem 2.5, and we obtain foe _T@)P@) ° 2rd) = Hla}, (using the fact that [(1) = 1). Performing the integration on the left-hand side gives = 5 we and hence TQ) = £7 But from the definition (2.1) we sce that I(x) must be positive (since the integrand is positive), so that we can discard the negative square root and obtain P'(}) = yx as required. Conoutany Per dt = 44/2. Proo ° Setting x = 3 in theorem 2.4 gives I'(}) = 2 \. e-® dt, so that the required result follows immediately on the use of theorem 2.6. Theorem 2.7 y — P@TO) BED = Te psy PRooF Substitute ¢ — cos’ in the definition (2.2) of B(x, 3): ¢ — cos” so that dt = —2 cos @ sin 6 dO; also, when t = 0, cos @ = 0 so that 0 = 2/2 and when t = 1, cos # = 1 so that # = 0. Hence we have: ° Blx,y) = j (cos? 6)#~1 (sin? 6)'-1, —2 cos 6 sin 6 dO “ cos#~! 9 sin*”-1 9 d0 afGAMMA AND BETA FUNCTIONS 28 9 FO) 20 +) (by theorem 2.5) _ TOI) TG ea) Theorem 2.8 B(x, y) = BU, 2)- PRoor This result follows inunediately from theorem 2.7. ‘Theorem 2.9 (i) Ble + 1,9) = Bes». . y (i) Bly +1) = Bs 9). PRooF (i) We have L@ + )LQ) POSED Tee T=y) (by theorem 2.7) al @TQ) («+ 9) +9) (by theorem 2.2) — _* Pel) x ty Te +y) © = pablo) (by theorem 2.7). (ii) Exactly similar to (i).§2.2 PROPERTIES OF THE BETA AND GAMMA FYNCTION: Theorem 2.10 (Legendre Duplication Formula) T(2x) = Paar +H. Proor We have, by theorem 2.7, L(x) P(e) Te yay BO) - fiera =e ° (by definition (2.2) 1 1 =fiz pal by ta 945 ds (on making the substitution ¢ = a +s) in = a f, (1 = styrot ds = 25. (1 —s1ds = geese fia ~ uj??? du (on making the substitution u = s*) = Qo fia = wt? du = 2 *OB(, 3) pens FLO) TQ + «) (by theorem 2.7). Hence, using theorem (2.6), Tw) 2-282 Tex) “Ter” and thus 12x) — a PEQl(e |).30 GAMMA AND BETA FUNCTIONS cu. 2 COROLLARY If xis a positive integer : (2a)! , V+) = fev PRoor In the theorem we may use theorem 2.3 to rewrite ['(2x) as (2x — 1)! and T(x) as (x — 1)! Hence we have st 2 te +d, which, on multiplying both sides by 2x, becomes 2x(2v — 1) = oats <1) + 3 this equation may be expressed more simply in the form (2x)! = 7aMe +2 Vv and thus , 2x)! te += va 2.3. DEFINITION OF THE GAMMA FUNCTION FOR NEGATIVE VALUES OF THE ARGUMENT From theorem 2.2 we have TQ) = tye +1). (2.5) Apart from x = 0, where the denominator vanishes, the right-hand side ot this equation is well detined for those values of x such that the argument of the gamma function is positive (since this was the condition for definition (2.1) to hold), that is, for + 1 > 0, i.e. for x > —1. We also note that we may say that I'(0) is infinite, for as x —» 0 we have M(x | 1)—> Tl) = 1, and hence I(x) = (1/x)I'(w + 1) > ©. So far the left-hand side is only defined for x > 0, but we can now use the right-hand side to extend this definition to x > --1, The argument we are using is as follows: (i) Equation (2.5) was proved true for x > 05 (ii) Right-hand side is well defined for x > --1, left-hand side for x > 0; (iii) Use right-hand side to define left-hand side for x >- 1.§2.3 DEFINITION OF THE GAMMA FUNCTION 31 ‘This means that we now have I(x) defined for x > —1, so that the right- hand side of equation (2.5) is well defined for x + 1>—1, ie. x > —2; we may thus use it to define the Icft-hand side for x > —2; and this process may be repeated to define I(x) for all negative values of x. Theorem 2.11 T(m) = <0, if m is zero or a negative integer. PRoor We have already remarked above that P'(0) — co. From equation (2.5) we have ret T0) = © =i 1 and r(—2) ete. BI(-1) = «, It is now possible to draw a graph of P(x). Plotting of points combined with information from the various theorems leads to a graph as shown in Fig. 2.2. rin Fig. 2.2. ‘The gamma funetion32, GAMMA AND BETA FUNCTIONS cH. 2 Theorem 2.12 P(e)Pd — x) = sin 2x" Proort We shall prove this result in three stages: (i) We shall prove the infinite product expansion for sin 0, viz. sind =| [{1 — (0%/’n4)}; net (ii) We shall show that this implies that Joy oy, Sd 7 20! (iii) We shall use result (ii) to prove the required result, (i) We know that : _ 0 0 sin 9 = 2 sin 5 cos 5 0. ix 0 ~ 2ssin 5 sin G + ) (2.6) and hence, by applying this result to both factors on the right-hand side, we obtain . . 6. (x o . {am 0. [x a, bo sin 0 = 2{2 sin $ sin (5 + £) (+5) sm (§+F+9)} = 2%sin 2 sin . 27) _ t+0 2 2 It is left to the reader to verify that applying the result (2.6) to each of the factors on the right-hand side of equation (2.7) gives us e=2 6 7” +O. da+ 8 3a +8 4a +8 sin 02" sin $ sin *5 sin 24 ao sin sin 5749 gin 9748 gig 7 tf. (28) 28 23 If this Process is carried out n times altogether, the general - will be “+ The proof of this theorem may be omitted on a first reading; alternatively, parts (i) and Gi), which are standard trigonometie identities, may Le assumed and. used to prove the result in part (iii),§2.3 DEFINITION OF THE GAMMA FUNCTION 33 sin =2°t sin 2 sin 219 gin ZF gig P= DATE pp ? ? with p = 2". ‘The form of this equation is plausible as the correct generalization of equations (2.6), (2.7) and (2.8); if desired it may be proved by the method of induction. The last factor in equation (2.9) is (@=1)n - sin Similarly, the second last factor is sin P28 sin fa C= 0) and in general the rth factor before the end is sin (> *), We now regroup the factors appearing in equation (2.9) by taking together the second and the last, the third and the second last, and so on. Then equation (2.9) becomes sind = 2-1 sin 8 sin? sin =} sin 2a +8 ain a... p P P {in (2 — De Gp — In +0 nade e749 19) ? p ‘Vhe factor inside each curly bracket is of the form sin (A + B)sin(A — B) = }{cos 2B -- cos 24} = sin? A — sin? B and hence equation (2.10) becomes sin 0 = 2?-1 sin! {sine 2 _ sin? a {sine 2 aint i. . Pp b Pp. P P. {sine woe - sint con? cos. (2.11) Dividing both sides by sin (0/p), letting 0 > 9 and oe that lim (sin x/x) -= 1 so that no34, GAMMA AND BETA FUNCTIONS cu. 2 im 808 si am ain Op) ~ EG =p, we obtain p=D sin? sin? om sin? ores (2.12) If we now divide equation (2.11) by equation (2.12), we obtain ne sin? sin? (9 /p) sin? (0/p) - sin {1 - { ~ Son Gin 1 8 O/D) Jp (2.13) We now let - > oo, We take into account the three results sin (0/p) oo =, (6) Tim SOP). tiny {Seave ne rp ¥ 1 poo Sin*(r2/p) pew O/p \sin (@x/p)J (r/p)* = 08/r'x?, (a) jim 2 sin’ . 6 e) Tim cos = 1 (©) Tim cos 5 and see that therefore equation (2.13) gives : oF oF oF sin 6 = of = “0 — sal! ~ pm © gt = “ly (1 - o). (ii) It is trivial to verify the result that 14a tan? sind = ag PE Hence 14, sino) Sind ~ a6” cos (072) ca 2 sin® (0/2) ™ ¥ sin (0/2) os (672) ri 2 sin? (0/2) ay” sind32.3 DEFINITION OF THE GAMMA FUNCTION 35 d = Ggllt 2 + In sin? (0/2) — In sin 0} = So In sin (6/2) — In sin 6} - pe Lo ( ~ aaa) TT° (- za} = ff2ino 42> 10(1— 52) 4250 (1 +54) nt m1 (e005) Sting +25 Iman —0) +25 Inne +6 n= — Smt 8) ~ dno +9} net (iii) Let us first assume that 0 (yu du ° n=O 2 2S ay fee ame n=O ° = S ye 1 yt tit 1 pie : =z, eyt" erage < 1 1 -ScrG- a} (remembering that 0 -< x <2 1)§24 EXAMPLES 37 > CD ~ Sa, (ent) — (eas) x sin mx (by part (ii) of the proof). We now remove the restriction 0 —I,n>0). (iii) in exp (2ax ~ x®) dx = } Vx exp (a’). a/2 (2) Prove that j tan 0.40 = EP {(L 4- n)/2} P{(1 — n)/2} if [a] <1. ° va T{(L + 0)/2} 2 T+? (4) Express each of the following integrals in terms of the gamma or beta functions and simplify when possible: 6) i - i)" dx; (ii) f. (at) dx, (a >0); (iii) f (6 — x)" —aytde, (b > a,m>0,n>0); iz a2 (3) Prove that f sin” 6 do — £ cost 8 d6 — (iv) fia — a")? dy, (m> 1p > —1,>0); dx . (” dt 0 aca > of ae (5) Show that the area enclosed hy the curve xt + yt = 1 is (T()}2/2/m. (6) Evaluate '(—4) and T(—}. (7) Show that @ P@)P(-s) = xsinax (i) PG + 4)PG — 4) = cos3 LEGENDRE POLYNOMIALS AND FUNCTIONS 3.1 LEGENDRE’S EQUATION AND ITS SOLUTIONS Legendre’s differential equation is (a wf 2 thy = (3.1) We shall write k in the form /(J -+ 1) for reasons which will become clear as we proceed. This equation, a ~9 - ax pt Ul + Uy = 0, (3.2) is of the form of equation a2 with g(x) = —2x*/(1 — x) and r(x) = {UL + 1)x*}/(1 — x*). "he binomial theorem may be used to ex- pand these as power series in x for values of x such that x? <1, i.e. such that —1 a,x and requiring 2 to be a solution of equation = (3.2) leads, as in Chapter 1, to the system of equations as(¢ — 1) = 0, (3.3) as + 1)s =0 (3.4)§3.1 LEGENDRE’S EQUATION AND ITS SOLUTIONS 43 and Anaa(s + 0 + (8 +m +1) — agl(s + alls +m +1) — UL + 1} =0 (@@>0). (5) Equation (3.3) gives the indicial equation s(s — 1) = 0 with roots s = 0 and s = 1 (which differ by an integer, so that we know we are dealing with one of the exceptional cases). Equation (3.4), with s = 0, is satisfied irrespective of the value of a; this means that @, is indeterminate and hence the one root of the indicial equation (s = 0) leads to two independent solutions with the two arbitrary constants a) and @,. Equation (3.5) with s = 0 becomes Gq 42 = (3.6) which, since a(n +1) ~ Ub +1) =n? ta PI = (0? —F) + (n—D) =(2~DOt)+@—) =(n— In +141), may be written in the form. _ alam ta+)) Ona = 80 By (3.7) "Thus we have 1 =1 a= ald, a= al M49), — 2 +3 1— 31+ 4 a ven aoe a : = all =D DE +9), _ 1d — 3) + 2 +. 4) = 1.2.3.4 “a 2.3.4.5 and in general ay = (—1)"aq MU=2)(—4) «C= 2n42MEH IV) E21) Gy (2x)! , and naa = ( 1)Pag (FAYE 3). ED AME YEA). L420) 9) (2n +1)!44 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 ‘Therefore a(x, 0) =a + > (—1P 12)... (1—2n-42)(U41)(L43) . .. (U4-2n 1). (ny! + ads + = (-1 (1-13) « «(U2 1)(L-42)(U4-4) (U2) ovat} oo nb A . = aoyi(x) + ay2(x), say, so that y,(x) and y.(x) are the linearly independent solutions. We know already that y,(x) and y,(x) will be convergent for —1 sey We wish to find the coefficient of #', so we must take r + p = J, and hence for a fixed value of r we must take p — / — r. But p only takes on values such that 0 < p < 1, so we must only consider those values of r satisfying 0 <1—r sey Cr DEQ eye Codd)48 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 If we now change the variable of summation from r to k = 1 — r, we obtain coefficient of # 2 (21 — 2k)! => Sgn a *Cf—1)K(2x)"* + GRP aaa - c (21— 28)! =A! sept angtaae = 2 Ba — HEC — ai (where [1/2] is as defined above in equation (3.12) qt wey (21 — 22)! - = 2 Vag pig 2 Dd — win” imo = Pix) Hence the required result is proved. The restriction on the values of x comes from the condition for convergence of the binomial expansion of {1 — «(2x —N}-"2, viz. | (2% — t)| <1. When |x| <1 this condition may be shown to be equivalent to | t] <1. (by the definition (3.17)). 3.3 FURTHER EXPRESSIONS FOR THE LEGENDRE POLYNOMIALS Theorem 3.2 (Rodrigues’ Formula) 1 d PA*) = pa Gal* ~ Proor We may expand (x* — 1)! by the binomial theorem: (1 = Ste(—-ye, S Henee, 1d 1 X (x2 - 7 = — t 1 )ry2!=2r 2 ae = oi aw, Coty, G.19) But the ith derivative of a power of « less than / is zero, so that 20if22 - 27 1/2.§3.3 FURTHER EXPRESSIONS FOR LEGENDRE POLYNOMIALS 49 iz toy Thus we may replace S by > if Lis even and by > if J is odd, ie., wn to int Pa by >, where [J/2] is as defined above, Also, a ag” = PCP ~ Wp — 2)... (p 1+ Ir Pl gt (e=oe a era = 2) so that ad™ Gaant* and hence, substituting into equation (3.19), we have 1 & 1 Boe ae 2”) yg ana — = a 4 ri pio Geant” > 4 (21 i we = 20) Brill — NT — 2” = Pia) (by the definition (3.17)), Theorem 3.3 (Laplace’s Integral Representation) P(x) = : f {x + V(x? — 1) cos 0}¢ dd. ° PRooF It may be shown by elementary methods (for example by means of the standard substitution t = tan (0/2)) that ote = 7. (3.20) ol +Acsd V(i—ay If we now write A = —{u/(x? — 1)}/(1 — ux), expand both sides of equation (3.20) in powers of w and equate the coefficients of corresponding powers of u, we shall obtain the desired result: 1 1 Ti hoot” 7 aver (i — ux) = (1 — ux)[l —ufe + V(x? — 1) cos 0) ) cos 0 =(1 - ux) due +b (x? — 1) cos OF tab50 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 (since, by the binomial theorem (1 — a)-1 = > a"). no Meee us) © Vil = ux)? = u(x? — 1} (1 — ux) * VG = 2ux Futy Hence, substituting into equation (3.20) gives i, > wee + Vie? — 1) 00s 0}! d0 = ai hea i= ie, St c {x ++ V(x? ~ 1) cos 0 dd = 2S wP(x) (by theorem 3.1). Equating coefficients of u* gives xP(x) = c fe + V(x — 1) cos 6} d6 ° which is the required result. 3.4 EXPLICIT EXPRESSIONS FOR AND SPECIAL VALUES OF THE LEGENDRE POLYNOMIALS From the definition (3.17) we may write down explicitly the Legendre polynomial of any given order. We give here the first few polynomials: *, 43x? ~ 1), 4(Sx* — 3x), P(e) = 4(35x4 — 30x? + 3). (3.21) Theorem 3.4 (i) Pl) = 1. (ii) P(—1) = (1). (iii) PA) = HO 4-1). (iv) PI) = (DTD)§3.4 VALUES OF THE LEGENDRE POLYNOMIALS 51 (9) Pal) = (1) 0 (vi) Pay (0) = 0. (By Pi(1) we mean [{dP,(x)}/dx],.1)- Proor (i) Set x = 1 in theorem 3.1 above and we obtain > era, > Pi). But 1/1 —t) = > # (by the binomial theorem or by considering the that is right-hand side as the sum to infinity of a geometric progression), so that we have > t= S #P((1), For this to be true for all values of ¢ in some = range (in this case |? | ~< 1) we must have the coefficients of corresponding powers of f equal, ie, P(i) = 1. (ii) Exactly similar to (i) but setting x = —1 in theorem 3.1. (iii) P(x) satisfies Legendre’ eauation (3.2), so that we have (a= wt we (x) — rx rye) 4 + UL + 1)P(x) = 0 (3.22) and setting x = 1 in this equation gives —2Pi(1) + Kl + DP) = 0, which reduces on using part (i) above, to Pi(1) = H( + 1). (iv) Exactly similar to (iii), but setting x = —1 in equation (3.22) and using part (ii) above. (v, vi) Set x = 0 in theorem 3.1 above and we obtain ¢, a +8) 2 oy Expanding the left-hand side by the binomial theorem gives us =(1+e)78 1 =14+( pe a&52 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 eee UDC DA ay 5... (20 — 1) Oe (- ay ls 3. lI . (cy EEE 7 (21 = 2)(21 — 1)2h — 22 2D! 9 (Var at i Mes 2nt Ce ihe IMs ‘Thus we have a 20 ~ Sy Gat => iro and equating coefficients af corresponding powers of ¢ on bath sides gives Py) = (I Pxy1(0) = 0. 3.5 ORTHOGONALITY PROPERTIES OF THE LEGENDRE POLYNO- MIALS Theorem 3.5 iflém jifl=m. FL PeyPat) dx {2 (This result may be written more concisely if we introduce the Kronecker delta, defined by _ foiflem bm BF, Then the statement of the theorem is ‘ P(x)P, ly ss : 5, J POMPAOD ds ay heed oO 2 d+.§3.5 PROPERTIES OF LEGENDRE POLYNOMIALS 53 PRoor P{(x) and P,,(x) satisfy Legendre’s equation (3.2), so that we have (a Pry oP se DP, =0 ae de = dP, AP, — xm _ 20m | = and (1 = at) — Dag + mlm + 1)Pn = 0, F which may be rewritten in the form d Pi A gl elt + vA =0 (3.23) and Ha - wat + m(m +1)P_ = 0. (3.24) Multiplying equation (3.23) by P,(x), equation (3.24) by P(x), sub- tracting equation (3.24) from equation (3.23) and integrating with respect to x from —1 to +1 gives P,Peade oe} rade oS} & 4 {0 +1) — mm +1} f, P,P, dx =0. But since Af. - ah = Po — 8B p, Ha - Ph, we shall have ce + Pa - 2h] dx 4 (I? 41 — m* — m) f, P.Py dx = 0. On integrating the first term we obtain dP, dP |t [ra a el + (L= my + m +1) f, P,Py, dx = 03 4 Throughout this proof we denote Pi(a) by Py and Pa(x) by Pm. The argument x is to be understood in every case.54 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 but now the first term vanishes at both upper and lower limits, because of the (1 — x*) factor, so that we have 1 (I= mt + m +1) j P,P,, dv = 0, and when 1 34m we can cancel = the factor outside the integral to obtain 1 j PP, dx =0. 1 1 It remains to prove that j {P,(x)}? dv = 2/(20 + 1). For this we use -1 the generating function of theorem 3.1: ' : Wes Pa) so that ace as a=) -{> > ti Pe)" = > P(x). S "P,() a md = > £+™P(x)P,,(x). amo We now integrate both sides with respect to x between —~1 and +1: 1 1 2 1 de em i, (1 + 22) — 2tx de a f, P(s)Pa() de. But the left-hand side is a standard integral of the form J {1(a + ba)} dav = (1/8) In (a + By), while the right-hand side contains only terms for which J = m, by the first part of this proof. Hence 1 2 |" Sef * [-kma +t ~ 2tx) 7D! 2 dx which gives def, Pyar = hin +e 2 +4 = In (1 -} #8 + 24) acd mint ~~ t)? ++ zn +t)?§ 3.6 LEGENDRE SERIES 55 = Hina +8) —In(1— a} at ef 1428 } atop tae tity tae (using the series expansion of In (1 + 1) 1 23 e Yarn 8 +E 4.4} 2 2 arf taty.} yt 421-1 Equating coefficients of corresponding powers of ¢ on both sides gives J eras — 575 as required. 3.6 LEGENDRE SERIES Theorem 3.6 If £(x) is a polynomial of degree n, then f(s) =D GP) with, = (+4) J £G)P,() de. r=0 a Also, if £(x) is even (or odd), only those ¢, with even (or odd) suffixes are non-zero. Proor We have f(x) = a,x" + aqy_yx"-2 +... + a,x + dy, say, and we may write P,(x) = hax” + Ryox*—? +... (since by equation (3.17) P,(x) is a polynomial of degree , containing only odd or even powers of x according to whether 1 is odd or even), so that if we take f(x) — (a,/k,)P,(x) we obtain either zero (in which case the first part of the theorem is proved) or else a polynomial of degree 2 — 1. This means that £(x) = GPa) + Bn-x(2) where ¢, = @,/k,, and gq_x(x) is a polynomial of degree n — 1. ‘The same argument may be applied to g,_,(x) to give n= natPa-a(¥) + Bua)56 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 and hence £(x) = cyPa() + Cn—aPua(#) + Sn-a()s The same argument may now be applied to g,_2(x), etc., to yield eventually £(x) = CuPa(x) + Cn1Pu a(t) +--+ eyPs(%) + coPo(x) n = > «Pe. Foard If we now multiply both sides of this equation by P,(x) and integrate from —1 to+1, we obtain f © FPA) x= > e |) BRL) dv. 70 But, by theorem 3.6, the integral on the right-hand side is non-zero only for the value of r which equals s, in which case it is 2/(2s + 1). 1 2 Therefore {. £(x)P(x) dx = ¢, BHT 1 giving c, = (r + 4) j f(*)P,(x), as required. “1 Now suppose that f(x) is even. Then, since P,(x) is even when r is even, and odd when r is odd, so the integrand £(x)P(x) is also even when r is even and odd when r is odd. But an odd function integrated over the range —1 to +1 will give zero, since positive and negative values cancel one another. Hence ¢, is zero when r is odd. Similarly when f(x) is odd, c, is zero when r is even. ‘COROLLARY If f(x) is a polynomial of degree less than J, then f * £(3)PAa) dx = 0. PRooF Suppose f(x) is of degree n; then by the theorem we have f(x) = > oP), so that fi fer@ar= > a f° RIPE ax =i moth =0 by theorem 3.5, since r < 2 < /, so that r is never equal to /.§3.6 LEGENDRE SERIES 37 The results of the above theorem may be extended to functions which are not polynomials, We shall not prove this extension, but will merely quote the following result (the proof is not difficult, but is fairly lengthy). ‘Theorem 3.7 ‘Suppose £(x) satisfies the following conditions in the interval —1 . ¢,P(), where 1 Ga FD]. API) ds, os converges to f(x) if x is not a point of discontinuity of {(x) and to E+) + £@—-)} if x is a point of discontinuity.t Also, at the end points of the range, x = +-1, the series converges to £(1—) and {(—1+) respectively. This series is called the Legendre series for £(x). + By f+) we mean Lim f(x + €) where ¢—» 0 through positive values and by f(#o—) we mean Lim f(x) — «) where again ¢—> 0 through positive values. See, for example, Fig. 3.1. y Fic. 3.1. Left+ and right-hand limits at a point of discontinuity58 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 3.7. RELATIONS BETWEEN THE LEGENDRE POLYNOMIALS AND ‘THEIR DERIVATIVES; RECURRENCE RELATIONS Theorem 3.8 bey @ Pa) = Sl — Pa. Ci) P46) = FHP, 0) + gp GPG. (it) (C+ Psa) ~ 21+ I)xP(x) + IP,a(s) = 0. (iv) Prats) — Pala) = 21+ 1). (¥) sPi(x) ~ Py) = IP). (vi) Pi) — #P,a(@) = P.a(9). (vii) (et — PY) = BPs) ~ IPs). (si) (@# — DPa) = + Rika) = (C+ SP), ) See NPR) = Fj PeaGPOO) ~ POP io Proor (i) We know that P(x) is a polynomial of degree J, containing only even powers of « if J is even, and only odd powers of x if J is odd. Hence P{(x) is a polynomial of degree / — 1 containing either odd or even powers of x according to whether / is even or odd. Hence by theorem 3.6 we have Pi) = GaP, a(e) + G1-2Pi-a() wp foP fee bea Phi) too +(e Gen) wih = +4) [PUPA de =+nf{[aoort] —[° Pler ax} (on integrating by parts) = (8 + D{RCP.A) — P(-1)P(-1) — 0} (where the integral vanishes by the corollary to theorem 3.6, since P,(x) is a polynomial of degree s — 1, and s — 1 is always less than 7) = +H (-1"5 (by theorem 3.4 (i) and (ii).§3.7.| LEGENDRE POLYNOMIALS AND THEIR DERIVATIVES 59 But s takes on the values / — 1, 1 — 3, . .., so that s + 1 takes on the values 2/ — 1, 2/ ~ 3, are always odd, irrespective of the values of Lor s. Hence ( —1 and we have a= s+) —(-1} = (2s +1). Hence eraayy = Al 2r 1) 1 =21—4r —1 and so we have Pix) = (21 — 1)P_a(x) + (21 — 5)Pr_o(#) + + (Ql = 47 — Phiy ie) $e. _ (3P,(x) if Lis even "AP (x) if Lis odd Ld=)) = D, Ql 47 — Pails). (ii) xP,(x) is a polynomial of degree J + 1, odd if is even and even if / is odd. Hence by theorem 3.6 we have Pie) = 6 sPhesle) a aP ae) +. + (pts ere) where 1. = EB | xPCOPLO) de =6 +9 [Peter ae. But this integral is zero by the corollary to theorem 3.6 if r +1 <1 (since then xP,(x) is a polynomial of degree less than 1), i.e., ifr <1 — 1. Hence we have 2PA%) = 2Pte(#) + G-1Pial)- (3.25) To determine ¢,41 and c_, we set x = 1 in equation (3.25) and in the same equation differentiated with respect to x, viz. PAs) + xPi(x) = CryxPhia(®) + 42P tals). (3.26) Setting x -= 1 in equations (3.25) and (3.26) and using theorem 3.4 (i) and (iii) gives Dei + tr (3.27) and 1 M1) = aE NE + 2D} 1 a A A. (6.28)60 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 Solving equations (3.27) and (3.28) for c,, and ¢,_, gives i+] qT Seog i * Bi so that on insertion of these values in equation (3.25) we obtain qT 1 SPS) ay Pasa) Uo gy Pica) 1 2+1 (iii) Multiply both sides of equation (ii) above by (2 + 1) to obtain (21 + 1)xPi(x) = (+ 1)Pria(x) + 1P,_1(). On rearrangement, this equation becomes (E+ I)Pisa(x) — (20 + 1)xP(x) + Pax) = 0, which is the required result. (iv) We may use result (i) for both Py ,,(x) and P/_ (x): Phas() = (20 + 1)P(x) + (21 — 3)Py_o(x) + 21 — 7)P,a(a) ++. Pix) = (21 — 3)Py_o(x) + (2 — 7)Pi_a(x) ++ Subtracting these two equations gives Phas(s) ~ Phas) = (21 + DPQ). (v) If we differentiate result (iii) with respect to x we obtain (CL 4 UPh a(x) — 2E + 1){Pix) + ePi(s)} + UPF_a(x) = 0 and if we now use (iv) to substitute for P/ ,.(x) we shall have (L-1){PL s(x) + 21+ 1)Pi(x)} — (20+-1){P(a) +P (@)} + LP} a(x) = 0. Collecting terms in P/_,, P, and Pi gives (21 + 1)Pr_a(*) + 2! + 12 + 1 — 1)P*) — 21 4 1)ePi(e) = 0, which reduces to Pi_,(x) + IP,(x) ~ xPi(x) = 0, and on rearrangement this equation becomes xPi(x) — Pi_a(x) = Pi(x). (vi) If we multiply (iv) by » we obtain Pj x(x) — 2Pf_a() = (20 + 1)xP\(x) and substituting this expression for (2/ + 1)xP,(x) in (iii) gives (E+ WPraal) + UP aa(s) = *P ii (2) ~ #Pi1 (2). (3.29) If we now rewrite (v) with J replaced by J +- 1, we have xP i,1(%) — Pix) = (I+ 1)Pi ls), and substituting this value of (7 + 1)P;.,(x) into equation (3.29) gives xPf, (x) — Pix) 4 IP, (x) = 8Pf, (x) Ph (3).§3.7 LEGENDRE POLYNOMIALS AND THEIR DERIVATIVES 61 When rearranged, this equation gives Pix) ~ #Pi_s() = IPs), which is the required result. (vii) If we multiply (v) by x we obtain aP{(x) — aP}_,(x) = lxP(x) and on subtracting (vi) from this we obtain w9Pi(x) — Py(x) = bePi(x) — IP,a(x) which may be rewritten in the form (x? = DP{@) = WP Ax) — IPs). (viii) If we replace / by J +1 in results (v) and (vi) we obtain Phy x(4) — Pile) = (E+ 1)Paga(s) and Phas(s) — 8Pi@) = (0+ PG). Eliminating P/,.,(x) between these (by multiplying the second equation by « and subtracting from the first) gives —Pi(x) + 9®Pi(x) = (0 + Pras) — (fF + 1)xP(x), which reduces to (68 — NPI) = C+ DE) — + AP), (ix) Using (iii) we have 1 Pr DPAO) —PuPauld) = ppp ql (+ DPI) — PPL D}RO) — Pe {(2k + DyPs0) — #0) = Fle — PGP) + FeO) — Pra(x)Py)}s multiplying both sides of this equation by k + 1, we obtain (b+ I) fPesae)Pay) — Pele) Pria(y)} = CRE y)Pu(a)PA9) HP G)PLa(9)—Pra@)PAy)}- 3.30) If (& -+ 1){Peas(2)Pi(y) — Pi(x)Prua(y)} is denoted by f,, equation (3.30) may be written in the form fy (2k 4 U(e = y)PAR)PAC9) ff ae (3.31)62 LEGENDRE POLYNOMIALS AND FUNCTIONS cH, 3 (Strictly, this equation is true for k > 1, since it is for this range of values of k that all quantities involved are well defined; but we may also make it true for k = 0 provided we define f_, = 0, for then fo = Pa(x)Po(y) — Po(x)Pa(s) =x— (by equations (3.21)) and x — y is also equal to (2k + 1)(« — y)P.(x)Px(y) + fea with k = 0.) If we now sum the set of equations (3.31) from k=0 to k=l, we obtain Dy fm Dy Gh NG —s)PAOIPLO) | fo : 7 = DORM —)POIRO) + D for = > e+e — POP) + > & Hence , - , - > = > e+ fe — YPO)PAD) so that - we : = (#9) > Ck + IPP) and if we remember the definition of f, we obtain i CD 6920) — Ple)Pan} = Ch + DPAO)PAD). (x 3.8 ASSOCIATED LEGENDRE FUNCTIONS Theorem 3.9 If xis a solution of Legendre’s equation dy dy _ = a — eZ MU + Dy =O then (1 — x3)"/2 (dez/dx™) is a solution of the equation 2 a — of — a + {e+ =; m ly <0 (known as the uocated Leeds equation).§3.8 ASSOCIATED LEGENDRE FUNCTIONS 63 PROOF Since z is a solution of Legendres equation, we must have (- a ait 410 4 1)x =0. (3.32) Now let us differentiate equation as m times with respect to x: an dt] od" f de de a fy yayd?2| od" fd) ds za “ neat aE} ++ Nam =O which, when we use Leibniz's theorem for the mth derivative of a productyf becomes dams d deg m(m— 1) d PN akc _ (a tga — 9 oS Ma ams dd de dns ~2 fo ou 28 Hye = (since higher derivatives of 1 — x? and x vanish). Cotlesing terms in ae dmg /de+ and da/dx, we obtain 1 UL | 1) = mle 1) — ame 0, SS (1x ne 5 — 2xe(m + yee which, on denoting d™z/dx™ by z,, becomes and? dz, (1 2) — 20m + Et + + 1) — mm + Je, = 0. (3.33) If we now write d™z ma (t xg, = (1 aye equation (3.33) becomes : a = 29 fet — 2-0) — 20m + a fal — 29-4} + {Ul +1) — mm + 1ja,(1 — x2)-™/2 = 0, (3.34) Bat dz, ea = yon) = FAI — at) ye = 5a — x2)-'mid-2, Dy = se = x82 maya 1 — xtyiad-a td nce dy jun Cri agit64 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 so that ae geil! — 8} _ ad — tyme +H Fa — x8)-(mi—1, (24) mf 2a — wt)Onr—a 4 ga] — 2) -tmi2)—a e ae - sa — x8) -(ms2)-2, ~29} aw st yim Sent — set) toa) + m1 = x8)ronid-a fe may(L — x2)-(9-1 | mayx%{m + 21 — x2)-lnrad—2, Hence equation (3.34) becomes Axe Ba gyn ama(h — 29) mel — 98) se m(m + 2)(L — 2) -!72 Agta — 20m + nef _ waynes - mx(1 — x) “ta st {UL 41) — mlm + Vax — 2)-™/2 = 0 which, on cancelling a common factor of (1. - x*)-"/? and collecting like terms, becomes tz, day (get {2mx —2(m + Ke 2 M1 + {m mee — Meet Dm 104.1) ~ mine) bay = 0. (3.35) 'Vhe coefiicient of day/de is just —2x, while the coetticient of 2 is (n? 2m — 2m? — Dx? a4) ee we +m —m*—~m =Il41)— QS — m? x —W ye, ‘Thus equation (3.35) reduces to dts, axfte S m* he (Langa - 28 UE) pape 0§3.9 PROPERTIES OF ASSOCIATED LEGENDRE FUNCTIONS — 65 so that , satisfies the associated Legendre equation which, by the defini- tion of 25, proves the theorem. COROLLARY The associated Legendre functions Pits defined by Py(x) = (1 = x? aye =, Pils) (3.36) satisfy the associated Legendre equation. PRoor This result follows immediately from the theorem, since P,(x) satisfies Legendre’s equation. Using Rodrigues’ formula (theorem 3.2), it is possible to rewrite definition (3.36) in the form qm Poe) = salt — xt Set — ay The right-hand side of this expression is well defined for negative values of m such that / + m > 0, ice., m > —I, whereas the original definition (3.36) of P7(x) was only valid for m > 0. Thus we may use this new form to define P7'(x) for values of m such that m > —1. It is easy to verify that if we consider m positive, the function P7"(x) defined in such a way is a solution of Legendre’s associated equation as well as P?(x). In fact, it is not an independent solution; it may be proved. that Pome) = (0 er) 637) (see problem 3 at the end of this chapter). 3.9 PROPERTIES OF THE ASSOCIATED LEGENDRE FUNCTIONS Theorem 3.10 (i) PPx) = Pi(x). (i) PP(e) = 0 if m > 1. PRooF (i) This result is immediately obvious from definition (3.36). (ii) Since P(x) is a polynomial of degree /, it will reduce to zero when differentiated more than / times. Thus PQ) 0 for m > J, and the required result then follows from definition (3.36).66 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 Theorem 3.11 (Orthogonality Relation) 1 20+ mts J_, Propet) ar @+ i — mi Proor We first prove that if J <1’ t J Preyer G) as =o. This proof follows exactly the same lines as that of the first part of theorem 3.5, so we shall not repeat it here. All that now remains to be proved is that . 21 + mi! molt da . J, 2r}* ae G+ it— mr Assume first that m :> 0; then from definition (3.36) we have fi (prey ex =fia =n{ poh {esr} dx - [eo fer “f (frelfe- or Srohe (on integrating by parts) a fi, {em Ps oie {a - =» “pya)} as (3.38) (the first term vanishing at upper and lower limits because of the (1 — x*) factor). Now, from equation (3.33) with m replaced by m — 1 we have that anit a aut (1°) Pua) — mee Pe) $+ {1+1) — (m1) m} = Pix) = 0 which, when multiplied by (1 — #7", becomes i (22) = 2ma( 1 — x?)"—1 70) (i= EL mlm pL 28" ae Pla) =0§3.9 PROPERTIES OF ASSOCIATED LEGENDRE FUNCTIONS 67 and this equation may be rewritten in the form E{o - = Sree} = -c mem eng — 2 Sr Substituting this result in equation (3.38) gives JL Pre} as tp gm et =f) keco}a + me — m4 na — x {2 pea ae = (I4m)\(I—m-+1) f 5 (ax {roa} dx =Cemt-min” {PRA}? de. Applying this result again gives FL Prey eae = Cnt msytem—1y-maay] (Pre) as = (ebmytm 1m ym 2) {era} dy, and repeating the process m times in all we obtain PL, rey} ae -1 = (l4m)(l4+m—1)... (141).(l-m+1)(I-m +2)... Lf {PP(x)}? dx “1 2 = Cim\ mat). (DAE 1). (m4 m1) 375 (using theorem 3.5) aCtm 2 “(= mast which is the required result. Suppose now that m <0, say m = —n with n > 0, Then f, {PP(a)}? de = fi, (Pe"(x)}? de Oe Gowns =f {or ay era ae (by equation (3.37)) srr68 LEGENDRE POLYNOMIALS AND FUNCTIONS cH, 3 nye pt = {By fl ereres (= n)e(l¢a)t 2 (ai (Qe (by the result just proved, since n > 0) =m! 2 “(Fm aei _(@4m! 2 (= my 1 which is the required result. Theorem 3.12 (Recurrence relations) (@) Pris) — Tiare) + UH) —m(m—1)}P PM) = 0. (8) ON NAEP) = Us MPRA) + m+ DPR (ii) VL = *)PP@) = ap 7 ahh) — PRY). Civ) Vl = PP) = gH + mtd +m — DPE) —(l—m + 1d ~ m + 2)PPGa)}. ProoF (i) This is the fundamental relationship linking three associated Legendre functions with the same / values and consecutive m values. Let us denote (d"/da")P,(x) by P{(x) so that definition (3.36) may be written in the form PR(x) = (1 — tyre), (3.39) Now, in equation (3.33) we know that we may take = P,(x) and hence 2, == P{"(x), so that we obtain (1 = 292 pom(ay — 2m + Det PEM) dx? ‘dx! + UL +1) ~ m(m + 1PM) = 0. Using the definition of P{"x), this equation becomes (1 = 8) PIMF%R) — Alon 4 1)ePIMEMS) {UL} 1) = m(m | 1)}PO(x) = 0§3.9 PROPERTIES OF ASSOCIATED LEGENDRE FUNCTIONS 69 which, on multiplying throughout by (1 — x*)"”*, gives (CL = 8) PIF 24) Dom 4 DoE — v2) 2PM) + {I + 1) = men + yi — styn2Pim(x) = 0. Hence, using equation (3. 3), we have Prix) — 2m + 1x — a PIMA) + {Ul +1) = min + SPM) = which, when m is replaced by m ~ 1, becomes rs 2) Prei(a) - say eh + {UL + 1) ~ (mm — Tym}PPm x) = 0; this is the required result. (ii) This is the fundamental relationship between associated Legendre functions with equal m values but consecutive / values. By theorem 3.8 (iii) we have (E+ D)Prea(e) — Ql + IePi(x) + IP, (e) = 0 which, when differentiated m times (making use of Leibniz’s theorem for the second term), gives (E+ IPE) — (20 + IEP) + mPP-M(2)} + IP(x) = 0. (3.40) Similarly by theorem 3.8 (iv) we have PEe) — PRs) = (21 + 1)Pix) which when differentiated m — 1 times gives PEM) — PEA) = 2 + PPM). (3.41) Using equation (3.41) to substitute for P(x) in equation (3.40) gives (1+ APEC) — (20 + 1)ePI™(a) — m{PE(x) — PIA(H)} + IPC) = 0. Multiplying this equation throughout by (1 — x*)"“? and using equation (3.39) gives C+ Y)PH@) — (214-1) Pix) — mP Hh (x) + PM (x) + IPR (x) = 0. Collecting like terms gives (Lp 1 = m)PFa(8) — 20 + DxPi(x) + E+ m)PL Ae) = 0 which, when rearranged, is the required result. (iii) Multiply equation (3.41) throughout by (1 — x#)"/2 and we obtain (1 PPEC) (LPR) GLE TL -x8APP-MR)70 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 which, on using equation (3.39), becomes Piia(x) — Pita) = (20 + 1)v(1 — «PP Ha). (3.42) Replacing m by m + 1 gives PRY) ~ PRAM) = (21 + 1)v( — 2*)PPa) which, when divided by 2/ + 1, is just the required result. (iv) We use « to replace xP7'(x) in (i) by Sp + PRG) =m 4 DPR AG) so that we obtain FE) gaya 1 PENG) | (Em | DPR UL +1) = m(m — 1)}PPH@) = 0. If we now use equation (3.42) for P/'-1(x), we obtain PE) — er ay re yl + PEM) +O — m+ PRG) MU) = ml — 1) es rg PP) ~ PES =0. By straightforward algebraic manipulation this reduccs to v(1 — #*)PP (x) {+ myE + m 4 1)PIR AG) — C= m)E — m + YP GD} which, when m is replaced by m — 1, is just the required result. 3.10 LEGENDRE FUNCTIONS OF THE SECOND KIND In the first section of this chapter we obtained two independent series solutions y,(x) and y,(x) of Legendre’s equation. We obtained solutions finite for —1 < x < 1 (indeed, finite for all finite values of x) by taking 7 integral; then for 7 even y,(x) reduced to a polynomial, while for / odd ,(x) reduced to a polynomial, In both these cases the other series remains infinite; it may be shown to be convergent for | x | < 1 and diver- gent for |x| > 1. In some physical situations we wish two independent solutions valid for the region | x | > 1; one of these is, of course given by P(x), while a second solution is given hy the following theorem (note that it is still infinite for x = +1).§3.10 LEGENDRE FUNCTIONS OF THE SECOND KIND a ‘Theorem 3.13 A second independent solution of Legendre’s equation is given by (2) 21 — Oi) = BPO) ES egg aarti (> 1) Qua) = 41m 4 1A itis oad 1? if tis even, Qu(2) as called the Legendre function of the second kind. 1 ther ee Proort In Legendre’s equation, set y = 2P,(x) so that 2 is a new dependent variable. We have dy de ~ Pi dty | oft a2 P aa ~ PO) at? dc de * at and hence the equation becomes (1 = ere) +20 — SE - rast +11 + I)2P(x) = + oe — 2P (0)8 Collecting terms in z, a and d?z/dx*, we have ‘ - ae = 208P 1 neo} te Slo. — 2992 — axryay} + 0 99h IES = 0, which, on using the fact that P, satisfies Legendre’s equation, becomes (1 29705 + {aa — 292 — aero} =o + The proof of this theorem may be omitted on a first reading. ‘The proof ends on page 77.72 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 d’s/dx? ,dP,/dx 2x — “defies > PG) and this is equivalent to d ds’ in (®) + 2s InP) + ¢ fina —x)=0, which when integrated gives =0, In = + In {P(x)}* + In (1 — x*) = constant. ‘Therefore “vere — x*) = constant = A, say, so that wl A de {P(2)}*(1 — x*) and hence s= de 4 lear ‘This means that we have a solution of Legendre’s equation given by Oe) = Px) leat (643) ‘We must now show that this is of the form stated in the theorem. We first dispose of the case 1 = 0: 040) = ol nama=m es — “[i(5 tra =H-In(1 — x) + In(1 +x} l+ex lax If now / £0, we remember that P,(x) is a polynomial of degree J, so that it may be written in the form P(x) = hie — an)(x — a9)... (x ~~ a). =4}In ‘Thus |. : J (CPE (Lay xhi(e a) (x (ea)?$3.10 LEGENDRE FUNCTIONS OF THE SECOND KIND 73 =7c5 tats? + Meet ea! (34) on splitting into partial fractions. We may determine ap, by and ¢, fairly simply. Multiplying both sides of equation (3.44) by (1 — x*)}{P,(x)}* gives = a1 + x){Pix)}* + bo(1 — x){Pi(a)}? G, d, + (1 ~ x*){(P(x) { wate C—O" 2 ea tea Setting x =1 in this equation and remembering that P,(1) = 1 gives a ~ J, while setting » — -1 and remembering that P,(—1) — (—1)! gives by = }. We now show that a=[§ iA PG Tyr lw To prove this we note that a 2 »df Salle — 2) 51C9)} = 2G — aE) + (@ — 0) SE = 0 when x = a, provided that f(x) is finite at x = a,. ‘The only terms on the right-hand side of equation (3.44) which are not finite at x = a are c;/(x — a) and d,/{(x — «)*}. Hence we have [ae - Tw ~ [ie ott eal. ~ [dear], = [elena =. Thus, if we write P(x) = (x — x)E(x) we have oe [é a ay.74 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 2x 2L'(x) - la=sicey ~~ tel _ pe 20 — 2E0)] LT La) dena, YouL(u) — (1 — a)E(a)) Ue FGF (8-45) But we know by setting P(x) = (« — «,)L(x) in Legendre’s equation that (1 = x) — a) L(y} — 20S (0% — aL} dx? dx +10 + Ns — a)Ls) — 0, which, on performing the differentiations, becomes (1 — 25){(8 — aL") + 2'(0)} — Que — aL) 4+ L65)} +10 ++ Ie — SLs) = 0, and setting x = «, in this equation gives (1 —a2)21/(@,) — 2aF(n) = 0 so that by substituting back into equation (3.45) we obtain c; = 0. ‘Thus from equation (3.44) we have ra; oo (= #){P,@)}? 20 — x) * A +a) 7 Gwe — a, where the d, are constants whose values will not concern us. ‘Thus 1 1 . \a—aemer*- —}In(1 — x) + $n(l +x) — De rot so that from equation (3.43) we have 1 Ons) = APA) In But (x — a,) is, for all «,, a factor of P(x), so that P,(x)/(x — a,) is a 1 polynomial in x of degree /—1. Thus > d, {P(x)}/(x — ,) is a poly- fol§3.10 LEGENDRE FUNCTIONS OF THE SECOND KIND 1s nomial of degree / — 1; let us denote it by W,_,(x). Then we have Ow) = $P(*) Int To determine W,_,(x) we semenber that Q,(x) isa solution of Legendre’s equation so that 2 WG), (3.46) E{a 2992} 410+ ng =0, which, on use of — a > gives 54 {a — 29 dpqein b 4-41). 4P(@) In 2 - ‘ {a - ae ~WL+1)M.=0. (3.47) ae = Pi(x) inp te 22 + Pi da 3 ~ Efe — 29 mt SE 2s} 1 ts x But = 4 pg) In ieee = Pi(x) int so that d{a - 2927) In =h>e— aa — x%)Pi(x)} + (1 — 9°)Pi (), gat 2Pi(x). Hence equation (3. a becomes gin te [a — enpiey 410+ npc] +2710 _ afa-s 2M. Ma} ~ I +1)Wi4 =0 which, on remembering that P,(x) satisfies Legendre’s equation, reduces to ifa tt +104 DW = 29 (3.48)1 LEGENDRE POLYNOMIALS AND FUNCTIONS on. 3 Now, by theorem 3.8 (i) we have dP, = Fe = LH PAG) + AL — SP) +. (21 — 4r — 1)P,op (2) ++. ng=1 = Dl 4-12.40) (3.49) Ss so that if we assume for W;..(x) (which we know to be a polynomial of degree J — 1) an expression of the form Wal) = a0Pra(2) + @:Pra() +. ad) = DaP.2r() 3.50) and substitute equations (3.49) and (3.50) into equation (3.48), we shall obtain ad-1)) d : ude) Dead POF oa} M41) DePieril) ads) — 2521 ~4r = 1)P2,1@)- G51) 0 But by Legendre’s equation we have d 7 & {(L = 2°)Pjapa()} + (FP — 2r — 1) ~ 27) Pay a(@) = 0, so that equation (3.51) becomes ng Daf—C = 27 — 1 = 21) UE + YIP ale) SS A@—1)1 = >2e! — 47 — 1)Ph-,-a(2). "The coefficient of each polynomial must be the same on both sides, so that we obtain {-—(I — 2r — 1) — 2r) + WI + 1)ja, = 2(21 — 47 — 1). (3.52)§3.10 LEGENDRE FUNCTIONS OF THE SECOND KIND 7 But —(I — 2r — 1)(1 — 2r) + +1) = —(I = 2r)8 + (I= 2r) + Ul 4-1)

land |y| <1.78 LEGENDRE POLYNOMIALS AND FUNCTIONS cu. 3 Theorem 3.15 (Neumann’s Formula) _1pt Po) 200) =3 J FO wy. Theorem 3.16 The results contained in theorem 3.8 (ii — ix) remain true when P,(x) is replaced by Q(x). Theorem 3.17 The associated Legendre functions of the second kind defined by re) = 1 — 28999 040) satisfy Legendre’s associated equation. 3.11 SPHERICAL HARMONICS In many branches of physics and engineering there is interest in the equation 1 a. 0! 1 ie g sins) + ar ou ¥ ule y¥=0, (354) solutions of which we shall call spherical harmonics. (This cquation usually arises in the solution of a differential equation such as Laplace’s or Schré- dinger’s in terms of spherical polar co-ordinates r, 6, ¢, so that the range of the variables involved is 0< 0< 2, 0< $ < 2n and we often require a solution which is finite and continuous for these values—the continuity implying that the value of ¥ at ¢ = 2z is the same as at $ = 0.) One method of finding a solution of equation (3.54) is the so-called method of separation of variables—we look for a solution of the form ¥(6, $) = O(6)®(¢). Inserting this expression into the given equation gives 0) fd d0\) , 0(0) de Bo (a (0 Se)} Sw age + DIA) —0 or, dividing throughout by @(0)0($) and multiplying by sin? 0, sindd/. d0\ 1d . ° § (sin 4p) gage TMF Dsinto = 0 which, when rearranged, becomes sing d (sin o 3) b+ 1) sint0 = — do 4 1 ~ do @ ag?§3.11 SPHERICAL HARMONICS ” Now the left-hand side of this equation is a function only of the variable 6, while the right-hand side is a function only of the variable ¢. Since these two variables are independent, it follows that left-hand side and right-hand side must separately be a constant which we shall denote by m*. ‘Thus we have ont Hsin os 6) +. Ul + 1) sin? 6 = m® (3.55) 1d . and -papo™ (3.56) Equation (3.56) is just e = —mb (3.57) while equation (3.55) simplifies to sip (sin 088) + {a + — hehe m0. (38) Equation (3.57) has the general solution @ = Ae" + Be-m where, if the solution is to be continuous, we require (27) = (0), so that m must be an integer (which we may take conventionally to be positive). In equation (3.58) we make the change of variable cos 9 = x. Then we have — sin @ do = 1d d andhence = — d oy 4d af and sin 05, = sin?@.— = — (1 — 84). Accordingly, equation (3.58) becomes d do’ m Shay yxyd 41 = {fa 22h {ra y- 0 which we recognise as Legendre’s associated equation; it will have a solution finite at 0 = 0 and x (x = +1 and — 1) only if /is integral. In that case the finite solution is given by © = Pf'(x) = PP(cos 6). ‘Thus the general solution which is finite at both 6 = 0 and x and is continuous must be Y(0, $) — (Ae™ + Be-™)PP(cos 0) which, because of equation (3.37), we may write in the form W ~: Aye™*P™(cos 0) |- Age ™P-P(cos 0)80 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 where A,=A and A, = Crt If now we denote IP, 4) = ePI(cos 6) (3.59) we may write the general solution in the form ¥ = Ayi@, ¢) + Avi" 4). Of course, this is a solution of the original equation (3.54) for any value of m, and since (3.54) is homogeneous we have the solution 1 w= DS APE, 8) + Mr, H}- For many purposes it is more useful to consider a multiple of yf" (which we shall denote by Yj") as the basic solution; a multiple chosen so that the solutions are orthogonal and normalised in the sense that ne Jae [7 eosin (V7, 8)" YPC, 8) = Febmm (8.60) *), (where the * denotes complex conjugation). We may readily prove that this is accomplished by taking 1 (21 + 1) — wi! . ve { Alam) J (0, $) 1+ 1 — to. {Ce nem 0a ms ai m) ‘emp Pens 6). (3.61) YFG, $) =(—-1"" FT =(-1)" For then we have J, a6 [" a0 sin ofvr0, oievir0, 4) o *), vegme 2 /f(QL + AAP + 1Y(1— my! Ul — m'yt =(y zn { 4 +m)! (=m)! } °° c= ag f PP" (cos 6) P®(cos 6) sin 0 d0 ° ° Vv (2x) (using the fact that P?"(x) is real), maam ll [S2LE ANA + IKE — mt = av 4 +m)! (0 +m)! J PrOPr ax “1§3.11 SPHERICAL HARMONICS 81 (since the first integral vanishes unless m' = m, in which case it is equal to 2a; and in the second integral we have made the substitution x = cos 0) am [f(2h + AYQr + 1d =m)! (= my =CN) Me 4(L-+ m)(U + m)! Jim fi Preyere) de _ iG $12 +190 — mi = mh, 21 +m)! 4+ m)! (0 my! ie I+ le — m) (by theorem 3.11) = Sb mm "The factor (—1)" in definition (3.61) of ¥7'(6, 4), which we shall take as the basic spherical harmonic, was not necessary for the orthonormality property; however, its introduction is conventional (although the reader is warned that in the topic of spherical harmonics different authors may employ different conventions). Theorem 3.18 (YP, H}t = (IYO. 8). Poor (7H, $)* —(—1\m 1 21+1(l—m)l\ EM Femi 2 Teme Fes (by equation (3.61)) on 2f4-1(l—m)! —tmgy =m = Taal PS eh i + mye ™(cos 0) (by equation (3.37)) mn im A+ + my!) ing pm = II Seay A ate mtr 6 =( Iryn, 4) (by equation (3.61).82, LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 We may use the definitions of YP(9,4) and P?'(cos 6) to obtain the following explicit expressions for the first few spherical harmonics: I Ye? = J) sin? @e%2#, Yet — -/@) sin 6 cos 6 e+, Yp= J (Ze cos*@ — 1). (3.62) 3.12 GRAPHS OF THE LEGENDRE FUNCTIONS In this section we give graphs of some of the functions encountered in this chapter. Play Fic. 3.2. Px), 0< x < 1,1 = 1, 2, 3,4§3.12 GRAPHS OF THE LECENDRE FUNCTIONS 83 Poin) iP, 0 Ps > Pr 1 10! Pa. Wi 2-5 4-6 io * Fic. 3.3 Pi(x), « > 1,1 = 1,2, 3, 4 (Note that the seale on the vertical axis is logarithmic.) Pitcos 8 3.4 Pi(cos0),0° 0+ 3/2,f 1,2,3,4LEGENDRE POLYNOMIALS AND FUNCTIONS Pin Fic. 3.5 Pix), 0< #< 1,l=1, 2,3 lx) Fic. 3.6 Q(x), 0+. x8 1,1 - 6,1,2,3,4§ 3.13 EXAMPLES 85 atx) 10°} 10% ———% 10% 2 10° a 10" es ese s ere 9 ox Fic. 3.7 Qi(x), x > 1,2 =0, 1, 2, 3, 4 (Note that the scale on the vertical axis is logarithmic.) 3.13 EXAMPLES Example 1 1 2U(t +1) . — 21) ‘Show that fi Prsa(*)Pr_a(x) dx = GF Dal py 1 Deduce the value of f P, (3) P,_(x) dx. ° ‘We use theorem 3.8(ii) to dispose of the x? appearing in the integrand, and we may then use the orthonormality property of theorem 3.5. ‘Thus, JL erenearacydyf* GP Gn 6P. .6)) de86 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. 3 1 1+2 1+1 =f eae me (2) {P+ Palen} ae toyt -f, ee) Pes) pp Ps) de (the other terms vanishing by theorem 3.5, since they are of the form 1 f P(x)P,,(x) dx with 1 = m) -1 +1) : ~ I EHC 1) JP 1(x)}* dee ta | ~~ (21 + 3)(21 — 1) 27 (by theorem 3.5) _ 2 +1) ™ (4P — 121 +. 3y We know that P,(x) is a polynomial of degree /, so that the above inte grand is a polynomial of degree 2 + (J + 1) + (J —1) = 2(1 + 1), ie, of even degree. ‘Thus the integrand is even, so that 1 1 fener ae) de = 2 oF, 98 at6) dr and hence fi stP. a) iale) do = ge (4 “ iar 3y Example 2 1 Evaluate j Oe ° By theorem 3.8 (iv) we have Jj Pied de = a7 J inte) — Pie) ax “a rei [Pane) — Pe], Sapp) — PAW) — P.O) + PL)§ 3.13 EXAMPLES 87 _ ai (err de ~ +1 ae + DHE (ayers gy! } are = 1/2}! (by theorem 3.4(i) and (v), remembering that J 4-1 and J —1 are both even) al api a (nye 221 |- 1)/2}%, _ ty tf ya! _! | +i 2 EG-1)/232 1 eayerg= pt © 2TH = GT = 4)! (aya — 1)! © 2+ DIA ar Example 3 If £(x) = | x | for —1 1, show that P(x) < P,..(x). We prove this by the method of induction: assume first that P,_,(x) < P,(x) and use this to prove that P,(x) < P,4,(x). Then since the result is trivially true for 1 = 0 (Pa(x) = 1 and P,(x) == x) it will be true for all /. We may assume throughout this proof that P,(x) > 0. For, from Rod- rigues’ formula (theorem 3.2), it is readily seen that if x > 1 (which it is in this example) then P,(x) > 0 for all values of /. Now, by theorem 3.8(iii) we have a+ bes erp tye 1 Pte, P,§ 3.13 EXAMPLES 89 Thus PP, Ti 1+i FP Ql+1) 1 iti fai (since x > 1 and we have assumed that P,_,/P, <1) 141 = Tei =1, Since P, is positive for all /, this implies that P,,, > P, and hence the method of induction guarantees the result to be true for all values of J. Example 5 Prove that {Q,(x)P1_(*) — Q1a(8)Px)} = (FINO a(@)P ial) ~ Q-a#) Pra} and deduce that (Q,()P,_.(x) — Qrs(2)P\(x)} = —1. From theorem 3.8(iii) with / replaced by 7 — 1 we have IPi(x) — (21 — 1)xPra(x) + U — 1)Pia(2) = 0 (3.65) and by theorem 3.16 we have also IOs) — 21 — I)xQrals) +E ~ Qrals) =0. (8.66) Multiplying equation (3.65) by Q,_,(x), equation (3.66) by P,_(x) and subtracting gives HP) Q1-4(4) — Q1()Pial#)} += IEP) O11) — O12) Pra} = 0, which is equivalent to HOWU)Prale) ~ Q1-1(%)P,(3)} = = I Qra@)Pra(@) — Qs-a(4Pa()} as required. If we now define F(2) = KQi#)Pials) — Qral®)Pulx)} this result may be written in the form FQ = Fl 1), and hence we have F() =F 1) KI—2)=. ‘Thus F() -- FQ) =F).90 LEGENDRE POLYNOMIALS AND FUNCTIONS cH. so that 1{Qi(x)Pr_a(e) — Qra(e)Pu(x)} = Qx(x)Po() — Qolx)Pi(x). But we know that Pix) = 1, Lyx) = x, l+s Ode) = 4 n -—%, Oy(x) = xin vane 1 so that we obtain l4x 1 HOME a6) = QP} = fo in AE 1 — pon tS =—-1, which is the required result, PROBLEMS ' 2 (1) Show that JP) de = By x pt 2uL +1 (2) Show that J © GL = 2 9PIC)PAle) ds — a 9s, (3) IfD= 4 use Leibniz’s theorem to prove that 1 eeptn(y? — 1 = (tye py mgt — 4 (1 = xtyentin( (I mpmer—1) Ocmed and hence deduce that P; "s) = (—1)"G dm) mr. @um =, OSES), expand f(x) in the form 2 c Pa). . (2m)'(1 — x#yr2 (5) Show that > Pein)! = Drini( — Bie 4 Bye nooPROBLEMS 91 1 (6) If uy =| s-1P,(x)P,_.(x) dx, show that nq + (0 — 1)ua1 = 2, a and hence evaluate tp. (7) Show that > (2r + 1)PAe) = Prale) + Px). = (8) Show that (1 — x) > (27 + 1)P,(x) = (# + 1){Palx) — Pra sla)}- (9) Assuming the result of theorem 3.14, prove theorem 3.15. (10) Show that Pya}Qi(2) — P42) 042) = 5g (11) Show that YP(0, 4) = J Fe tae (12) If x > 1, prove that took O40) = ari f, Cah and deduce that (i) by making the substitution ¢ = See} we obtain 20) =| ex ea i) cosh (i) 00 =F, > Sa ae4 BESSEL FUNCTIONS 4.1 BESSEL’S EQUATION AND ITS SOLUTIONS; BESSEL FUNCTIONS OF THE FIRST AND SECOND KIND Bessel’s equation of order n is ay dy ES et ix? — iy = ate ba + (x? — ly = 0 (4.1) (where, since it is only n? that enters the equation, we may always take x to be non-negative). Since this is of the form of equation (1.2) with g(x) = 1 and r(x) = x? —n?, we may apply the methods of Chapter 1 and be assured that any series solution obtained will be convergent for all values of x. © Setting a(x, 8) = > a,x'*” and requiring « to be a solution of equation r= (4.1) leads, by the same method as in previous cases, to the system of equations f(s — 1) +5 — n}ay = (4.2) {+s -+ (6 +1) — 23a, =0 (43) and (8 +7)(s tr —1) 4-6 +9) — na, + ag =0 (> 2). (44) Equation (4.2) gives the indicial equation s?—n? = 0 with roots s = 2. We shall obtain two solutions from these which will be inde- pendent, apart, possibly, from the case when the two roots differ by an integer, i.c., when 2n is integral.§41 BESSEL’S EQUATION AND ITS SOLUTIONS 93 Equation (4.3) is {( + I? —n%}a, =0 and, since s?= mn, we cannot also have (s-+1)*=w"f, so that (5 + 1)*— n? 40 and hence a, = 0. Equation (4.4) is just {6 +7)? —n¥}a, + a, 2 =0 which gives oar ‘Taking s = n, this becomes oe Ora Tbr = an prea) — fat ” (> 2. @5) ‘Thus we have a a 20m 62) > 28 4. 1)" a, a, ay mw % ~~ 6(2n +6) 2?.3(n +3) = a ) _ 25. 3ia + (in + 2)(n and in general ao ae (OW geri ie 4 eG ay + The only situation in which we can have s? = (5 + 1)? =ntisn =4,5 = —4. For this ease itis true that s -= —} makes a, indeterminate and leads to two inde~ pendent solutions, but the results of the text still hold good, since zero is a possible choice for a, and we obtain the second independent solution from 5 = 4.94 BESSEL FUNCTIONS cH. 4 We simplify the appearance of this expression somewhat by noting that (+ 1 +2)... (nr) Seen br Del h OKn DED _ TQ +r+1) =~Ta +) on using repeatedly the fact that xT'(x) = T(w + 1) (theorem 2.2). ‘Thus we now have te (OD aeriR a a Also, if we use the fact that a, — 0 together with equation (4.5), we obtain =O Hence, substituting these values for the a, into the series for 2(x, s), we obtain as a solution of Bessel’s equation S Poe tl) erin, > WY pit@ ar PY ‘This is a solution for any value of ays let us choose a, = 1/(2*N(n + 1)} and we obtain the solution which we shall denote by J,(x) and shall call the Bessel function of the first kind of order n: Sole) = > { Van@ir try =y(a) ” (46) From the remarks at the beginning of this section it is clear that the infinite series in equation (4.6) will be convergent for all values of x. So far we have dealt with the root of the indicial equation s = n. The other root s = —-1 will also give asolution of Bessel’s equation, and the form of this solution is obtained just by replacing n in all the equations above by —n, so that we obtain the solution to Bessel’s equation Jal) = xe Vance 3G) . 47) Suppose now that 7 is non-integral. ‘Then, since r is always integral, the factor T(—n +r + 1) in equation (4.7) cannot have its argument equal to a negative integer or zero, and hence must always be finite and non- zero. Thus equation (4.7) shows that /_,(x) contains negative powers of x (they arise for those values of r such that 2r 0. Then ena) Jab) = SD aes r=0 ° . from equation (4.7). But I'(--n + r + 1) is infinite (and hence 1/{f(—n +-r + 1)} is zero) for those values of r which make the argument a negative integer or zero, ive., for r = 0,1, 2,...(m — 1) (remembering that this is possible because nis integral). Hence the sum over 7 in the above expression for J_,(x) can equally well 1 be taken from x to infinity, and then 2 arn J (2) = ze Penner ) = s ("Ge + ae COG saneeyG) 2 1 i\ tm Jefe) = zo mPa + 9 ry) (by equation (4.6)) But96 BESSEL FUNCTIONS cH. 4 so all that remains in order to complete the proof is to show that (m + n)!P(m + 1) = mIP(n + m +1) for n and m integral. But (m +n)!E(m +1) = (m + n)(m = m'P(m +n (on using repeatedly the result that I(x 1) = xT'(x)), and thus the result is proved. Now consider 7 < 0; in this case we may write n = —p with p > 0. ‘Then what we require to prove is that I(x) = (—1)-*F(#) or (-1)J,0) = Js which, of course, since p is positive, is just the result we have proved above. m1)... (m+ 1m!P(m +1) 1) Let us summarise what we have proved so far: we have shown that if n is not an integer then J,(x) and J_,(x) (defined by equations (4.6) and (4.7), respectively) arc independent solutions of Bessel’s equation (so that the general solution is given by AJ,(x) + BJ_,(x)) while if n is an integer they are still solutions of Bessel’s equation but are related by Jax) = (Tul) Theorem 4.2 The two independent solutions of Bessel’s equation may be taken to be Sul) and Yq(x) = COS In) — J-n(2) sin a7 for all values of n. Proor We consider separately the cases 2 non-integral and n integral. (@) non-integral. Here sin nx 0, so that ¥,(x) is just a linear combination of J,(x) and ]-.(x). But from the above discussion we know that in this case J,(x) and J») are independent solutions, so that J,() and a linear combination of Tilx) and J (xe) must also he independent solutions. Hence J,(x) and Y,,(x) must be independent solutions.§4.1 BESSEL’S EQUATION AND ITS SOLUTIONS 97 (ii) n integral. Tn this case sin nz = 0 and cos nx = (—1)", so that by theorem 4.1 we have cos na Ja(x) — Jal*) = (—1)Yal#) — (—1)Yala) = 0. Hence Y,(x) has the form 0/0 and so is undefined. However, we may give it a meaning by defining it as Y,(x) = lim Y,(x) = lim 7a) ~ J) (48) rn sin w= _ [(/2»){c0s va J,(x) - J Al—n — (8/4) sin oz), 4 (by L’Hopital’s rule)} __ [eatsin va J,(x) + 608 va (8/00) J.(x) ~ (8/0) (2) =n Pr c08 7g __ cos af(8/20)T Mven = 10/2 on ‘7 COS NI 1 = SLC /ONT8) = (C/T aw (+9) We must now prove two things; firstly that Y,,(x) as defined by equation (4.9) is in fact a solution of Bessel’s equation and, secondly, that it is a solution independent from J,(2). To accomplish the first of these we note that J,(x) obeys Bessel’s equation of order v: d. Je 5 oe 4 (et yng, <0 which, when diferentisted with respect to v, gives a3 ¥ Fe a ae ~ ooh — 2, =0. (4.10) Also, of course, J_,(x) a Bessel’s equation of order ». so that we have in exactly the same way ef Td a (et 9 ee 2, =0. (411) t L’Hépital’s rule states that if f(a) = = g(a) =0, then im 2 £(x) _ f(a) i a ee 8) aa) where f(a) and g(a) are the lowest-order derivatives of f(x) and g(x), respec tively, which are not both zero at x = a.98 BESSEL FUNCTIONS cH. 4 Multiplying equation (4.11) by (—1)" and subtracting from equation (4. im gives of “(oh (4 y tah 4 of (2h ae (ay aa) + (et fe — ay 21} 2, =~ (-IT-} <0. Setting » = » in this last equation and using equation (4.9) gives da? d 2 Gaza Vals) + eg Yale) + Ge? — 2°) Yale) — 2.0) ~ (D7 ae) = 0. But now nis an integer, so that we may use the result of theorem 4.1 that J -(*) = (—1)"J.(x) to obtain d d a8 TaYale) +25, Yale) + (88 — 2 )¥a(x) = 0 which just states that Y,,(x) satisfies Bessel’s equation of order ». "That Y,(x) is independent from J,(x) may be seen from the result of the next theorem, which implies that at x = 0 Y,(¥) is infinite, while we know that J,(x) is finite. Theorem 4.3 (Explicit expression for Y,, (x) for n integral) ¥90) =2 {in 7 —} >> 2} 70) oe oer) Sita ~1)! (:)~" as at, vt 2, where y is Euler’s constant (see Appendix 2). Sn Mi Proor We shall omit the proof of this theorem, merely noting that the method to follow is to use the series expansions (4.6) and (4.7) in equation (4.9). Euler's constant appears because of properties of the derivative of the gamma function.§ 4.2 GENERATING FUNCTION FOR THE BESSEL FUNCTIONS 99 ‘Theorem 4.4 When n is integral, Y9(x) = (—1) ¥,(x). PROOF From equation (4.9) we have, v0) =2[27.0) -(- 27 ‘, = 3 [aes ona], =1[-2 749) +¢ ve ane). = (92 [2 7.0) -(-r 2709], =(—1)"¥,(@). We shall call J,(x) and Y,(x) Bessel functions of order » of the first and second kinds respectively. (Y,(x) is sometimes called the Neumann function of order m and denoted by N,(x).) As we have seen, J,(x) and Y,(x) (with a considered positive) provide us with two independent solutions of Bessel’s equation, J,(x) always being finite at x =0 and Y,(x) always infinite at x = 0. 4.2 GENERATING FUNCTION FOR THE BESSEL FUNCTIONS Theorem 4.5 eo {ir} > > eh n= PRooF We expand exp ele - , )} in powers of ¢ and show that the coefhi- cient of f* is J,(x): on {be} en (4) o0(-) 3% (Gey > C bx/t) s/t100 BESSEL FUNCTIONS cH. 4 - S @er yer Bosh = > oy (4.12) We now pick out the coefficient of 7", where first we consider 1 > 0. For a fixed value of r, to obtain the power of t as i" we must haves = 7 — n. ‘Thus for this particular value of r the coefficient of 1” is sey (3) - Hest We get the total coefficient of @ by summing over all allowed values of r. Since s = r — mand we require s >> 0, we must have r >> m. Hence the total coefficient of 2” is 7 ven (18 tro 2. (x/2)274" 2 (2) rig — at 2, — parayipt (where we have set p = 2, (x/2)224" - ap oO rp cen pt (remembering that both p and n are integral, so that we may use the result of theorem 2.3 that T(p +n +1) = (6 +m!) =IJnlx) (by equation (4.6)). Ifnown <0,westill have the cocfficientof 7" for a fixed value of r given by 9) Fae but now the requirement that s > 0 with s =r — n is satisfied for all values of r, Thus the coefficient of ” is just 2 ]\eren tron 2) tr—n 2! “ty 6) Faery =(-)- x y ay +) —n) = ayy aco (by equation (4.7)) = Jul) (by theorem 4.1).§4.3 INTEGRAL REPRESENTATIONS FOR BESSEL FUNCTIONS 101 4.3 INTEGRAL REPRESENTATIONS FOR BESSEL FUNCTIONS Theorem 4.6 Jilx) = Frees (nf — xing) dg (n integral). PRooF Since J _,(x) == (—1)"Jq(x) for m integral, the result of theorem 4.5 may be written in the form exp Ber SD} ate) > (6 (ANI). If we now write t = c¥* so that 1 aoe t 4 =e* —e-# = 2ising this equation becomes erHed = Fa) 4D fol + (De HD). East But when 1 is even eM £ (—1)r ei = oF 4 o™ = 2 cos md, while when n is odd we have of 4 (= 1h eo = ins — ei — Di sin nd. ‘Thus we have 5 = Jn) + DS) ZeosndIu(s) + > 2isin np Jala) = Jo(x) + S 2. cos 2k Ju(x) iD, 2sin 2k ~ 1)6Jor-al2). a & Equating real and imaginary parts of this equation gives cos (x sin 6) = Ju(x) + > 2 cos 2k Jus(x) (4.13) sin (x sin ¢) = > 2sin (2k ~ 1)bJor-s(2)- (4.14) Go Af we multiply both sides of equation (4.13) by cos nd (7 2- (0), both sides102 BESSEL FUNCTIONS cH. 4 of equation (4.14) by sin nd (1 > 1), integrate from 0 to x and use the identities a 0 (m 1) f cos mp cos ng dg == 4/2 (m =n 0) ° x (m=n=0) 7. . 0 (mn) id si dé = an |, sin mg sin ng ib x/2 (man x40), we obtain the results [fos ma cos ( sin 8) a — {i ale) : > . 0 ; and f * sin nf sin («sin $) db = £ 1) : ae Adding these last two equations gives f {cos nd cos (x sin ¢) + sin nd sin (x sin 4)} db = a7J,(x) for all positive integral n. Hence i cos (np — x sin ) dé = 2J,(x) o which is the required result for positive n. If n is negative, we may set n — —m where m is positive, so that the required result is Ficos (me een a ° (where m is positive). But feos m$ xsing) dp ° 0 =| cos {mee — 0) ~ xsin (a — 0)}. —a0 (where we have changed the variable by setting 6="—$4) = fies {-miv + mo — x sin 0} a0 7 = j {cos (m0 — x sin 0) cos ma 0 4- sin (m0 — x sin 0) sin mz} dO§4.3 INTEGRAL REPRESENTATIONS FOR BESSEL FUNCTIONS 103 =(-1)" [7 cos (ma — x sin) dO o — (-)"Ja(*) (since we know the result to be true for positive m) a] al) mf nf). Theorem 4.7 Gx)" t pyr ieltt ~ Il) = yesreey pl trite @> — 2. Proor Consider the integral J defined by 1 I= I (l— pen dt 2 fap SOF ay 2 rn = ye a re (l= eyed r=Q 0° 1 Now, if is odd, the integrand in f (1 — &yte de is an odd function -1 of t, so that the integral is zero; while if r is even (say, equal to 2s), the integrand is even, so that we have f (b-eptra = [- (1 = epee ae a a F _ — p-tyae d, af. (1 — rete a = fie = uyttuet du ° (where we have made the change of variable u = t?, du = 2t dt) = Bn + hs +4) (by the definition of the beta function; we must have n > — } to ensure the convergence of the integral (see Section 2.1)104 BESSEL FUNCTIONS cu. 4 Pe +pre +) ~ Fa +s +1) (by theorem 2.7). (ix) T@ + OTs +4) 4, 2! Ta +s+1) 1 Qo! = 4 hat ie V Gy Rear e oi) 251V™ (using the corollary to theorem 2.10) mn 2. (—1)'(x/2)2048 Hho + Devin) > te ya (by equation (4.6)). G =ro env ( 1 ) re) "Thus J,(x) = A) waren +H - Ware 7 5G) S- d= een 44 RECURRENCE RELATIONS Theorem 4.8 @ 4 ene = 2h. ) Sef.) = Paul) ii) Ju) — Jnl) — "Jala. Gv) Jabs) = 2Jal) ~ Juste). (©) ues) = HJu- a2) ~ Joan) (i) Jus) Jule) = 270).§ 4.4 RECURRENCE RELATIONS 105 PRooF (i) From equation (4.6) we have JAX) = > ( Yr Tr +h ()"" so that 1 d d= 1 Fe OO} = 5 2 (-19 rH mae -sy Yer Ppa 2r + Andersen 1 =e > CU ie ER ee yen Fee ro (using the result of theorem 2.2 that T(x +: 1) = x1(x)) ee 1 /ap\ 27 ln 1) =" 2 CU iret ali) = ¥Jn ald) (by equation (4.6)). (Note that in this theorem no restriction is placed on n: it may be integral or non-integral, positive or negative.) (ii) We have Ste Hd) = ys 1 er rit + + r+ 1) 2" 1 rye 1 = seo Tare 4 FH ya 1 = -> (— VY area ee pee (since the factor 7 in the numerator makes the term with r =0 vanish; we remember that 0! = 1) & 1 1 = > Cay Diy Beret 44 1xtern— & (where we have sets =r 1)106 BESSEL FUNCTIONS cH. 4 ys 1 1 = 2 pga pee vent = 3 Oe ral " = — anal) (on use of equation (4.6)). (iii) From result (i) above, by carrying out the differentiation of the product on the left-hand side, we have mt1Pu(3) + 8°78) = 2a -a(2) which, on dividing throughout by x", gives Tat) + Fale) =Jns@s and hence Sale) = Ju-als) — 7 Jal). (iv) We carry out the differentiation in result (ii) above to obtain nah D7, (x) + 8 -i(8) = 8a sal) which, on multiplying throughout by x", gives Bole) + Ju) = June). "Thus Fels) = Tals) ~ Joni. (v) Add results (iii) and (iv) and the result follows immediately. (vi) Similarly, if we subtract result (iv) from result (iii) we obtain the required relationship. Theorem 4.9 All the results of theorem 4.8 remain true when the Bessel functions of the first kind are replaced by the corresponding Bessel functions of the second Rind. PROOF We shall prove that result (i) remains true for Y,(x); a similar method will prove result (ii), and then results (iii-vi) follow in the same way as they did in theorem 4.8,§4.5 HANKEL FUNCTIONS 107 Thus what we have to prove here is that (d/dx){x" Y,(x)} = #" Y,,-1(x). We must consider separately the cases of integral and non-integral n. (a) Non-integral n. Here we may write ¥,(x) = 2° ni Jul) ~ J?) sin m7 so that Servis) Frag loosmnd terpuo)} — Ltery of) = giles er Jab) = (Ha al)] (where we have used theorem 4.8 (i) for the first derivative and theorem 4.8 (ii) for the second) [cos nz Jn a(x) + J n—a()] ~ sin nx x"[cos {(n — 1) + 2} Jnalx) +I -n-(2)] ™ sin {(n — I) + a} =<: “Ge Tyg 008 = De) + Ten] = (0 Dan) Joa) sin (n — Iw = Yq (0). (b) Integral n. Here we note that Y,(x) — lim Y,(x) and that by part (a) (d/dx){x"¥,(x)} = x" Y, a(x). Taking the limit of this result as » —» n gives the required result. 4.5 HANKEL FUNCTIONS We define the Hankel functions (sometimes called Bessel functions of the third kind) by gi) = Jal) + i¥.(2) Hy) = Jule) V(x).108 BESSEL FUNCTIONS cH. 4 ‘These are obviously independent solutions of Bessel’s equation. Both are, of course, infinite at « = 0; their usefulness is connected with their behaviour for large values of x, which we shall investigate in a later section. Theorem 4.10 Alll the recurrence relations of theorem 4.8 remain true when ],(x) is replaced by either H,((x) or H,((x). PRoor Again we prove relation (i) only, the remainder following as before. But we know relation (i) to hold for both J,(x) and Y,(x), so that we have Ef TOD) = 2"Tusl) and hee Y,(2)} == 4" Yn_a(2). Hence d id : LT} LV) = 24Teals) Yn a), so that te Fale) VON = 24a a) Vn al) which, on remembering the definitions of H‘(x) and H2(x), gives 4 font (ay} = HY) (taking the plus sign) and A poti2\} = HE (6) (taking the minus sign). 4.6 EQUATIONS REDUCIBLE TO BESSEL’S EQUATION, Theorem 4.11 The general solution of 2 t & + (xt — ny =0 is AJ,{Ax) + BY,(Ax).§4.6 EQUATIONS REDUCIBLE TO BESSEL’S EQUATION 109 PRooF Make the substitution t = Ax dy _ ,dy so that ae A ‘at dy ody and wn a ae Then the given equation becomes dy. dy oF Ds nay = ae t lg tomy a0 which is Bessel’s equation of order n, so that the general solution is ¥ ~ ASidt) + BY.) and hence, replacing ¢ by Ax, the general solution of the original equation is ¥ = Af, (x) + BY, (Ax). Theorem 4.12 The general solution of eae 2a + {pty + (a? — ny )y =0 dx? is Ax*],(Bx”) + Bx Y,(Bx’). PROOF We first make the change of variable y = x*s. "Then Baw F parts dy de yds . and age = Gye FG + ale — De® Be so that the given equation becomes dz dz fg 2 \ te Dott ee 2) ag} 2 ate ge | 2 | oe — Dx | (1 Daye 2 | at + (Byte 4 (2 = uty} = 0 which, on collecting terms in d2x/dx?, dz/dx and x becomes dz dz ob fae + — Qala ati b Qa + (1 — 2a} sb foun — 1) + (LH Zand Byte?) (at ny) fs = 0110 BESSEL FUNCTIONS cH. 4 which simplifies to 4 2S + & + {BYytx® — nYy%}e = 0. (4.15) Now change the independent variable x by setting t = dz ded¢ de apr ‘Then dx 7 dt de?” at dz dz d dz ae iy gyyr-2 FF 4 yyy A oe ae dee 7 ae ds dz - yaa YA yyy rey — Uatt yet part dts dz = pry Be ~ 1x" yaad ee + Hy ~ Weta so that — (4.15) os 7 2 ay — ae E 4 yar pryret” — ate = 0, which, on coleting like terms and camsating acommon factor of y2, gives pt z dz at 242 — nt}e = 0. oa ty + (Pt — ne <0. But this is just the equation of theorem 4.11 with solution 2 = AJ,(ft) + BY,(ft). Hence the solution of the original equation is obtained by substituting back the relationships y = as and t = x’: y = Ax'],(Bx!) + Be Y,(A2). 4.7, MODIFIED BESSEL FUNCTIONS Consider the differential enum a y 4 2 = (xt + ny =0. (4.16) This has the form of the equation in theorem 4.11 with 42 — —1. ‘Thus it has the general solution y = Ajj,(ix) + BY, (ia). Now, the solutions J,(ix) and Y,(ix) have the disadvantage of not§47 MODIFIED BESSEL FUNCTIONS 111 necessarily being real when x is real. However, we may take the constant multiple of J,(ix) defined by A(x) = iJ n(ix) and use it as one of the independent solutions of equation (4.16). It is easy to show that it is a real function of x: 6 fein 7-1) \2, A(x) = i-"Ja(ix) wae ay 1 syn (2 2"" <0} Vaneau") oi DO ag S 1 aren ; ~ 2 ri +r +i) 6) : (4.17) I,() 80 defined is called the modified Besse! function of the first kind, and equation (4.16) is called Bessel’s modified equation. Theorem 4.13 If nis integral, I_,(2e) = Iy(x). PRooF T (2) = iJ fix) = i(—1)"Ja(ix) (by theorem 4.1) = i(—1)"i"L,(@) —1)",(%) = 1,(x). We may obtain the second independent solution to Bessel’s modified equation by considering Y,(ix); or alternatively we may employ a method similar to that used for the definition of Y,(:x). I,(x) and I_,(x) are both solutions of equation (4.16). When 7 is non- integral they are independent solutions (since J,(ix) and J_,(ix) are in- dependent solutions). However, when 7 is integral I_,(x) = I,(x). Define now a Tn(*) — Lala) Ke) j= (4.18) When n is non-integral, this is well defined, and I,(x) and K,(x) together112 BESSEL FUNCTIONS cH. 4 provide independent solutions of equation (4.16). When » is integral, K,(x) is indeterminate as it stands and has to be defined by x1_(x) ~ L(x) Ka) = lim sin va _s =(@n (2) — (2/2) Mean, [C08 7,5 (by L’Hopital’s rule) oy x) ay, | Ske) — 50] ‘As in theorem 4.2, it may be shown that this provides a second in- dependent solution of Bessel’s modified equation for integral values of 1. ‘The explicit series for K,(x) may be obtained from those for J,(x) and Y,(x) by use of the following theorem. Theorem 4.14 K,@) = Fe Talis) +i¥i0)} 7 Fein). Proor rae = te (by definition (4.18) wit] (is) — i-*Julis) 2 sin am But from the definition of Y,(x) in theorem 4.2 we have that J -ahist) = 008 nat J(ix’) — sin noe ¥,(ix). Hence in sin nat Y,(ix) Sioo{ivgiay 4 teens “y.00)}. (4.19) —i cos my — i-tn-t = i cos mz + i.i-* = —i cos nm + i(e*"2)-™ (writing i = e”*)§4.8 RELATIONS FOR MODIFIED BESSEL FUNCTIONS 113 = —icosna + ie = —icos nz ++ i(cos nz — i sin nz) = sin mn. Thus, using equation (4.19) we have +1 GY (ix) 1 Sais}. Kn(x) 4.8 RECURRENCE RELATIONS FOR THE MODIFIED BESSEL FUNCTIONS The modified Bessel functions satisfy recurrence relations similar to, but not identical with, those of theorem 4.8. Also, in contrast to the func- tions J,(x) and Y,(x), which satisfy the same relations as one another, 1,(x) and K,(x) satisfy different relations from one another. Theorem 4.15 () Len(a)} = aha) Gi) A feen(a)) = eh ul0), Gi) 146) — Zl) — 2 (iv) 1) = 2a(a) + Zonal. (0) His) = Haile) + Ina. (63) Leal) ~ Tous) = 220). PROOF We shall prove results (i) and (ii), The remaining results follow from (i) and (ii) exactly as in theorem +8, () From theorem 4.8 (i) we have d SO VO) = #3). Replacing x by ix gives d atin G9} — iT in114 BESSEL FUNCTIONS cu. 4 which, on using the fact that I,(x) = i-*J,(ix), becomes ie finscrinT,(x)} = irxrin—aT,_(x) and this equation, on cancelling a factor of i*-1 throughout, gives d Gi} = x"lya(2). (i) From theorem 4.8 (ii) we have d Soba) = — Yale, which, on replacing x by ix, becomes gt HG) = Hie a (ix). ‘Thus 1d. «as nut Pag eR L()} = —itactie Tn (8) and hence 1d), in Lag AO) = ix nal) so that d ee a} = nails) Theorem 4.16 (i) 4 {x"K,(a)} = —x"Ky a(x). (i) Ape eK.) = -9 Ky al). il) Kia) = —Kya(s) ~ 7K). (iv) Kila) = 2K) — Kal). (%) Kile) = Male) + Kens). (6) Kes) ~ Keats) = — KC).§4.8 RELATIONS FOR MODIFIED BESSEL FUNCTIONS 115 PROOF Again, we prove only results (i) and (ii). Also, in the same way as in theorem 4.9, we prove the results for non-integral n and appeal to con- tinuity for a proof in the integral case, (i) From theorem 4.14 we have K,(x) = ey HH Gx) and by theorem 4.10 we know that 47{(«) satisfies the first recurrence relation of theorem 4.8. Hence ArH} = (2), which, on replacing x by ix, gives ate HO(x)} = Px (ix), "Thus 1d, 2 sas Fgh OK} — relink, 6) so that diye = ~ Serle) = Ky a6) and hence He KO) = KG). (ii) Similarly H{?(x), by theorem 4.10, satisfies the second recurrence relation of thcorcm 4.8 so that A fH) = — H.C) which, on replacing x by ix, gives a dix) and hence, using theorem 4.14, 1d vc astl nyt Bia (2)} = - neni AK, (a), eH} = ime (in),116 BESSEL FUNCTIONS cu. 4 ‘Therefore : d . itt fe"K,(2)} = in 2K (0), from which a common factor of i-**-? cancels to give Loe eK (9)} — Kaul). 4.9 INTEGRAL REPRESENTATIONS FOR THE MODIFIED BESSEL FUNCTIONS. Theorem 4.17 1 x)» ft i = —_{*

-}. 7 vm (NT ae at (ii) Kas) = e565) i e-*(t2 — 1st de (n > -4,x > 0). PROOF (i) By equation (4.17) we have I(x) = i"J,(in) and by theorem 4.7 we have 1 yn! 2)n—$ git Sal) - ware) {a Byte de (2> -)) so that ‘x \n ce fT Le) =" are ep () fa ryte at 1 e\" [t Lo = Ware 1 (3) ja ~ Ht ome dt. (ji) We first show that the integral Pax r eet —Iytdt (9 > dx >0) 4.20) 1 satisfies Bessel’s modified equation 5 ty) dat | de PE =O§4.9 REPRESENTATIONS FOR MODIFIED BESSEL FUNCTIONS 117 We have oe | +s]. dx 1 1 and aP nn — Iyer? in e-*(i2 — 1-H de ae : net in tee? — 1ytde 1 ha \. t emm(t2 — 1h de 1 so that ait Pa = (x8 +9 =x ° {n(n — 1) — nxt +84? on — xt — xt — nene(e2 — Ay de =a i {x(t® — 1) — 2m + Yet? — 1a = yttl i {xe-(e2 — 1)! — (w+ Ae® — 1)" e-*H} de ed = nt aa aby = nth f plone? 1) de = Ax" Hey? — Hye =0, since the integrated part vanishes at the upper limit, f = 00, because of the factor e-* (remembering that we are considering here x > 0) and at the lower limit, t = 1, because of the factor (t? — 1)"+# (remembering that 2 + 4>0). Thus P satisfies Bessel’s modified equation and therefore must be of the form P = AL(x) + BK,(x). We show now that A = 0. We do this by considering the limit as x—> oo, From the series (4.18) for I,(x) we see that I,(x) consists of a power scrics with positive cocflicients, and hence I,(x) > 00 as x—> 00, Consider now P(x). We have, from the definition, P(x) > 0, and we shall show that P(x) is less than some quantity which tends to zero as x tends to infinity so that P(x) itself must tend to zero as x tends to infinity. Since the asymptotic behaviour of an exponential dominates that of a118 BESSEL FUNCTIONS cH. 4 power, we must have (t# — 1)"-! X). "Then for x >- X we have Pls) 0 as x —> oo, again using the fact that the exponential dominates the power. ‘Thus we have shown that P(x)—> 0 as x—> 0, whereas I,(x)—> oo. Hence P(x) cannot contain any multiple of I,(x) and we must have P(x) = BK,(x). ‘To determine the constant B we ¢xamine the behaviour of both P(x) and K,(x) as x—> 0. For K,(x) only the lowest power of x is important in this limit, and we may use the definition Ke) =F 10) sin nz together with T(x) = 2 riT@- 4 op ()"" to see that the lowest power of x in I,(x) is Twp Gy and hence in K,(x) is x 2\e . Hence as # > 0, Ki) —> sry gin = which, on making use of the result of theorem 2.12 that Tard — x) — ina?§4.9 REPRESENTATIONS FOR MODIFIED BESSEL FUNCTIONS 119 may be written in the form K(x) > P(nj2n-t ‘To study the behaviour of P(x) near « = 0 we make the change of variable t=1+" ® 1 so that dt =— du. x Also, when f= 1, =0 and when t = 0, u = ©. "Thus, using equation (4.20), P(x) =a i eee + “ lay ° ¥ x x 1 * 2x\nF 1). For small values of x we may make the approximations e-* <1 and 1 P ~tyan=1 PO) —> 5, |, etwer du 1 = 5T en) (by the definition of the gamma function), Thus for small values of x the result P(x) = BK,(x) reduces to Lig. h(n)? Pen) = pe se re giving B= Rome But by theorem 2.10 we have Ten) == ronn +1) so that (vayE@ya' 2M (n+ 1) Vn120 BESSEL FUNCTIONS cH. 4 Hence Po) =e D Ky, giving Va Ke) = a PO) and thus, using the ‘into of P(x), Ke) = Tas rep) J Sones peta 4.10 KELVIN’S FUNCTIONS Consider the differential equation d' dy es 4 = ~ (Ate? + ny == 0. (4.21) ‘This is of the form of bo 's modified equation d 0 4 aD Suset + any =0 with 2? = ik®. ie since the general solution of Bessel’s modified equation is y = AI,(Ax) + BK,(Ax), the general solution of equation (4.21) must be y = ALfitkx) + BK,(ithx)+ Also, since 1,(x) = i-*J,(ix), we may take the independent solutions of equation (4.21) as J,(i*/*kx) and K,(i3/*kx). Of course, when » is real J,(i°/*x) and K,(i"x) are not necessarily real; we obtain real functions by the following definitions: ber, x = Re J,(i?/2x) bei, x = Im J,(i?/*x) Ini**x) = ber, x +i bei, x. (4.22) ker, » = Re i"K, (ix) kei, x = Im i-*K, (x) i-*K,(i"2x) = kery x + ikei, x. (4.23) + Strictly speaking, i# is a two-valued function with values e#/4 and c!5*/4, We remove this ambiguity by taking the value e!*/4,§4.11 SPHERICAL BESSEL FUNCTIONS 121 (If 2 == 0 the notation used is often ber x, etc.) ‘Thus the general solution of equation (4.21) is given by y = A\(ber, kx + ibei, kx) + A,(ker, ke +ikei, kx). (4.24) 4.11 SPHERICAL BESSEL FUNCTIONS Consider the equation af + ae + (kext — L-+ Dy = 0. (4.25) "This is of the form of the equation of theorem 4.12 with 1-—22=2 ye aye a? — nyt = — Kl +1). Solving these equations for «, 8, y, n gives a= -hB-hy—ia-Tth and hence by theorem 4,12 the general solution of the above equation is given by ¥ = Ax Yrij(lex) + Bu Y1<4(ke) = Ayjfkx) + Aayi(he) where we have defined the spherical Bessel functions j,(x) and y,(x) by it) = (2) rt (426 wt) =) rue (4.27) and A,, A, are new arbitrary constants related to A, B by A, /(2k/n)A, Ay = V(2k/)B- Spherical Hankel functions may also be defined in a way exactly anal- ogous to the definition of Hankel functions: HPS) = 42) ++ inte) (#28) HPC) — JM) — iy. (429) In applications it turns out that the spherical Bessel functions of most interest are those of integral order. We shall show that such spherical122 BESSEL FUNCTIONS cH. 4 Bessel functions may in fact be written in closed form in terms of ele- mentary functions. First, however, we prove recurrence relations anal- ogous to those in theorems 4.8, 4.9 and 4.10 for the Bessel functions. Theorem 4.18 Tf £,(x) is any of jax), nlx), AD(x) or W2(x), then Stree} = ah) (8) Sf0°6)} = 3 yal) n Gi) a) = 6, 40) — Fee, (iv) 60) = 28) — fal). (8) Ce + DEL) = a) — (0 + Dhl) (6) fabs) ++ fale) = FQ, PRoor We first prove the results for j,(x). (i) From theorem 4.8 (i) with » replaced by 2 -| § we have d Gt Tan} = Ing a) which, on using definition (4.26), becomes d be ‘ inh Sefer versinco} = x V2/a)ia 0) and this simplifies to Ley) = a0. (ii) From theorem 4.8 (ii) with » replaced by n + 4 we have Fans} = 2 Yarn Again, use of definition (4.26) gives Eom tyes} = Vase 0)§4.11 SPHERICAL BESSEL FUNCTIONS 123 and hence d : . Ge VAD} = — ea sal). (iii) If we carry out the differentiation on the left-hand side of (i) above, we obtain a Yin() + (+ Vxfale) = a Yn al). On dividing throughout by x"+4, this gives Hee) ie) = jules) and thus Jnl) = jn al) — Nie) (iv) Carry out the differentiation on the left-hand side of (ii) above, and we obtain eM (x) — mx" Ya(X) = — Mn gal). Cancelling a common factor of x-" then gives Galt) ~ "5sl8) = — jul) and hence dn) = Sula) — Jasa(a)- (v) Multiply result (iii) by n, result (iv) by (n +1) and add, and we obtain the required result. (vi) Subtract result (iv) from result (iii), and we obtain the required result. The proofs for y4(x), A?(), and Ji? (x) follow in the same way, since these functions bear the same relationship to Y,,,(x), Hp},(x) and HO) (x) that j,(x) does to J,,,(x) and by theorems 4.9 and 4.10 the various Bessel functions all satisfy the same recurrence relations. Theorem 4.19 (i) gol®) = sin x cos x (i) yu)124 BESSEL FUNCTIONS cu. 4 (ii) AD) = AS ir (iv) WP) = PRoor (i) We have, by equation (4.26), i) = veerante = I(Z) > aga)" (from equation (4.6); but, by the corollary to theorem 2.10, we have (2r +2)! j= — ‘5 Mert +) = geag yp piy™ so that we now have . y eee 41)! att doe) = Me i os Yr SDs a r-+1) a, - 3 Yay ay" =e Veep ro wy “> (yr Gap e* (on recognising the infinite series for sin x) (ii) We have x wt) = /(Z)v@ (by definition (4.27) = |Z) BELO aL (by the definition of ¥, (x) in theorem 4.2) Me)§4.11 SPHERICAL BESSEL FUNCTIONS 125 = -/®@ > Or ap srena) (by the series (4.7)). But again, by the corollary of theorem 2.10, we have M(-$+r+) =F +h ! 2, © 22r! so that Qe] yt Jobe) = ie i> CV Feantva 2 2 Orem 1< xe sm CY ay 1 —~ = cos x, x (on recognising the infinite series for cos x). (iii) Hx) = jl) + ive(x) (by definition (4.28)) 1. 1 = —sinx —i-cosx x * =~ £(c08 x isin x) le =-i-. x (iv) HP (x) = jol*) — ivolx) 1. 1 == —sinx +i- cos x x x i eo = = (cos # ~ isin x)126 BESSEL FUNCTIONS cH. 4 Theorem 4.20 (Rayleigh’s Formulae) If nis a non-negative integer, then ‘sin x’ (8) inl) = (— yet ay(@), % (i) yalx) = -crye(LAy(2 * os) (iii) AP) = —i(—1)" Gs £) (2) (iv) ADR) = «y(t 4y ( —). Proor We give the proof for (i) only, the proofs for the other three results being similar. ‘We shall use the method of induction; i.e., we shall assume the result true for n = N, say, and then prove it true for n = N + 1. Then since theorem 4.19 guarantees the result true for 2 = 0, it follows that it must be truc for all positive integral 1. Assuming the result true for m == N implies that dx(x) =(— yrer(t s ay" (sm ‘) From theorem 4.18 (ii) we have jy als) = — 2" fe (9)} -watecoGa) FD) (using the fact that we are assuming the result to be true form = N) =(-vravn(t EL) (82) = (-prravn(t aya) which thus proves the result true for n = N +1.§4.12 REHAVIOUR OF THR BESSEL FUNCTIONS 127 Hence, by the remarks above, the result must be true for all positive integral n. We may use the results of this theorem to write down explicitly the first few spherical Bessel functions of integral order. We obtain in this way: sinx cosx iC) == = =| : 3. o1\. 3 ids) = (Ga) sin — cos, . 15 1 sinx —~ (8 = °) cos x; (4.30) sin x, is - & - ) sin x. (431) It is to be noted that because of the relationships (4.26) and (4.27) all the above information concerning spherical Bessel functions of integral order provides an equal amount of information concerning Bessel func- tions of half-odd integral order. 4.12 BEHAVIOUR OF THE BESSEL FUNCTIONS FOR LARGE AND SMALL VALUES OF THE ARGUMENT Theorem 4.21 As x—> 0 we havet (toe) ~(2)° cos {x +08} «ip Yacy~(2)' sin fe — + nes Gi) 1%) ~ (2) exp [ifs ~ @ +93} Gv) 90) ~ (2)! exw [-if —@ + v3} t Here we use the symbol ~ rather loosely to mean ‘behaves like’, A more precise definition is: (x) ~ g(x) as x > aif lim (F(x)/ela)) 1128 BESSEL FUNCTIONS cu. 4 (0) Ids) ~ Fass (vi) K,(x) ~ “Ge i (vii) fala) ~= £ sin (- (viii) yg(x) ~ a cos (« "); (x) Aaa) ~ — Leap file — (rn/2)3 (2) H4s) ~+ exp [ifs ~ (n/2)})- PRoor We shall first prove result (vi) and then deduce the remainder of the results from it. (vi) By theorem 4.17 (ii) we have x\" ett — 1)r44 Ke) = TG en) re (eI de We now make the change of variable f=1+4"= x 1 so that dt == du. x Also, when ¢ = 1, u =0 and when t = ©, u = o, Thus K,(x) = mee a) i otto 2 2 ay “Fee a) Os ok +1)" we de 2) -1 +f “e r ) nok =e zy ~e€ fre ul +a du. For large values of x we have u/2x small, so that we may take as an approximation ( u 1 | 2e, ~l,§ 4.12 BEHAVIOUR OF THE BESSEL FUNCTIONS 129 Then _ vm ee et Kole) ~ ren ve5" Jf eet an vn =e vane te th (by definition (2.1) = ( ,) - =(3) «* (iii) From theorem 4.14 we have K(x) =5 BiH, In’ M(ix) where, if n is non-integral, we take “he value of the many-valued function yee given. by l/h v, Then Hy (ix) = = 2 exp {(—ier/2)(m +- 1)} Kal) and hence, on relacing x x by —ix, HO (x) = 2 exp ( —in/2)(n + 1)} Ka(-—ix) Thus 196) ~2 (C ix 2yn | w(- ° se = (2) (i) exp {(—in/2\(n + 1)} 7x, = (2) "exp ((—in/2)(n + 1)} el , (on writing —i = e-**/2) 2 7 = (2) ep ie-@ + Hey (iv) Since Hx) = Ja) V(x) and Hx) = Jule) + i¥.(2) we have AD(e) = {H@))* ~[(2) ev tie —@ + nen] (a) exp Life 1 2)e/2)i]-130 BESSEL FUNCTIONS cH. + (i) We have Jule) = Re Hy'(z) ~Re(2)' exp fife —@ + Dea = (2) cos fe (nt ay}. (ii) Similarly Yala) = Im HE) b mim (2) exp fee ~ + 0/21 = (2) sin {e- (n+ oy}. (v) By definition we have I,(x) — i-*J,(ix) so that, by part (i) above, we have Ie) ~i-(Z.) cos {iv — + D5} =e (2) dep (aoe 2) Heap fH AH wnt oy: =i Jeena fe + (n + x/2)i} (keeping only the dominant term for x —> co) inet ams exp (x).im## (writing exp (ix/2) =i) 1 ~ (Gas) 1 (i) 40) = (Z) Fauld (by definition (4.26)) “OG of 9 (by result (i) above) 4 ( =) a heos (x -F ee§4.12 BEHAVIOUR OF THE BESSEL FUNCTIONS 131 1. ne = jain (»— 9). 7 wi) 940) = (2)! ¥en (by definition (4.27)) YC) abo (by result (ii) above) tine) --tal5-6-9) =~ Leos (2), (ix) AP) — fale) + inl) aero ate =~ foo (x 7) +isia (©-9)} =~ exp fie — on) @) WO) = Geo}* ~ [dew Gte—ne]” a Lexp [ite — (2 Theorem 4.22 As x —> 0 we have O10) ~ eq) v@)’132 BESSEL FUNCTIONS cu. + i 1. /2\n ) - trea?) (n 40) Yua)~y 9” = Inx (n =0); (iii) f.0) ~ ea (n integral) sf Cn=1 yo. wi (n integral). (iv) yale) ~ — PRoor nS (i) We consider the series (4.6) for J,(x). As x—» 0 only the lowest power of x will be important. But this lowest power of x is just given by the term with r = 0, so that we have Jee) ~ Te eal) (ii) When n is not an integer we have Yo(x) = C8 Sole) _ Jax) " sin nz and we pick out the lowest power of x occurring in this expression. This will be contained in J_,(x) and will be given by the result of (i) above, so that we obtain 1 1 fx\-" Yul) ~ — na on in(3) --19¢7, using the result of theorem (2.12) that Tq) — 2) = When zn is an integer we pick out the dominant term from the series for Y,(x) given by theorem 4.3. Remembering that a power of x dominates a logarithm, we see that Y,(z) ~2 ins (n =0) + By nll (n double factorial) we mean n(n ~2)(n — 4). {53-1 if mis odd * (6.4.2 if n is even.§.4.13 GRAPHS OF THE BESSEL FUNCTIONS 133 and ve) ~— Lai) @ +o ro (iii) We have from definition (4.26) i) = I (2) Vere ral)” _ (v7) Bt a ne vr: $40) (using repeatedly the fact that [(n + 1) = nT(n)) a ae (nF 1Qn 1)... 3:1.va (where we have made use of the result of theorem 2.6 that '($) = v/n) x" ~ (Gn + it (vy) ale) = al (2) Yan) ~JG)-se +9@)"} - vat +» = ~ eal — 4m — 4)... 4-408) 1 (2n —1)(2n — 3)... 3.1.x rr _ n=! =-"a 4.13 GRAPHS OF THE BESSEL FUNCTIONS We give here some graphs of various Bessel functions; they bring out rather clearly the behaviour for large and small x discussed in the previous section.134 BESSEL FUNCTIONS cH. 4 Fic. 4.1 Ja(x), n = 0,1,2 Yh *D Fic. 4.2 Yn(x), 1 = 0, 1,2$4.13 GRAPHS OF THE DESSEL FUNCTIONS Ala 1 2 se Fic. 43 In(x), 2 = 0, 1,2 Ad i 2 3 * Fic. 44 Ka(s),n =0,1,2 135136 Ava BESSEL FUNCTIONS Fic. 4.6 ya(x), 2 = 0, 1,2§ 4.14 onTHONORMALITY oF THE BESSEL FUNCTIONS 137 x (Grophs oscillate for large x) yeber x yebei_x Fic. 4.7. ber x, bei x Fic. 4.8 ker x, kei x 4.14 ORTHONORMALITY OF THE BESSEL FUNCTIONS; BESSEL SERIES Theorem 4.23 Fi anearniee de = 9 In l8.a), if &, and €, are roots of the equation J,(Ea) 0.138 BESSEL FUNCTIONS cH. 4 PROOF We first show that if £, and &; are distinct roots of the equation J,(Ea) = 0 then Jj sfaeadgutee) de = 0. Jo{Ex) and J,(E,), being Bessel functions, must satisfy the following equations w Slee) +a —Jn(Sex) + (Bix? — 0°) Jn(Exx) = 0 and a d fet 9X a gel aE) + eT IalEx) + (G8* — m)Ja(Ex) = 0, which may be written in the form af {8 1, aha + (Gx? = n)lEsx) = 0 (432) and T(E x) 1 (82 u2 VT fbx) — 0. (4,33) and Wax Ode Ia, yy 1 (Sfx UnkSsx) = 2. (4.33) Multiplying equation (4.32) by (1/x)Ja(G#), equation (4.33) by (1/x)Ja(Ex) and subtracting, we obtain Jub {nf Jal6o} — Jaleo) {a Jaleo} +E — 8) aJa(Em)IulSsx) = Applying the result u(dv/dx) = (d/dx)(ue) — v(du/dx) to both first and second terms gives 4 {reed p60} —{476a)}s 2 1060) ~ 8 frsteeowd sateen} + Sracer}e E629 + (8 ~ B)xFalEeJalEa) = 0 and thus {rex 8 tl6ed} — 2 {pled E Jee} + (= DeJalEa ales) = 0.§ 4.14 ORTHONORMALITY OF THE BESSEL FUNCTIONS 139 Integrating from 0 to a, we obtain [Jaleaded JG) — Jed JGa) |, + =a | a2aean fle) dx = 0. ‘The first term will now vanish at the upper limit since InlEia) = Jn(Eia) = 0, Accordingly (& ~ &9) ["afa(da)Ja(Ga) de = 0 and hence [sateen Jala) de = 0 provided bh, ‘To complete the proof we require to prove that * a [i sUstear ar =F Gaal if Jaga) = 0. Now, if we denote J,(E2) by = we have we! + xa’ + (Ex? — nts 0 which, on multiplication by 22’, becomes 2wtg"a! + Qua’? + 2(E%4% — nB)ea’ = 0 and this is readily seen to be equivalent to d Gay's — nig? 4 byte) — 2etent = gg 28 — nts? + Elxtat) — Deteat = 0, Integrating this equation from 0 to a gives [xte'? — nat + Seta] — 282 j xat de =0 ° which, on replacing x by J,(éx), becomes [=+{2.7.e9}* —meten? + een], 2 — 28+ |” x(7.(¢a)}4 dx = 0 and hence, using the facts that J,(£a) == 0 and 2J,(0) == 0, [efire}'], TAgx)}ide 0.140 BESSEL FUNCTIONS cu. 4 Hence a 1 d z f[strdeontas = 5s[e"{ZIoe9} But, by theorem 4.8(iv), we have d n ign") = pI obs) — Jnalbe) which may he rewritten in the form. Elles) = "Jlb) — onl) so that Ji etneentae = sh [er{2 sate ~ eFanted | = at Un alle) (again using the fact that J,(£a) = 0) =F Tanlba)}e Theorem 4.24 If f(x) is defined in the region 0 < x < a and can be expanded in the form > caFulGex) where the €, are the roots of the equation J,(a) = 0, then il 2 f* xfeyFulba) dx = OnE)? PRrooF We have fe) = Dele) & so that ACEH) — D> eee FulEoe)JalB8) i i=§415 INTEGRALS INVOLVING BESSEL FUNCTIONS 141 and hence fi 09.6) ox = de [esta dx = Ye, "Junalést)®, SoS rutare, (by theorem 4.23) : = 45 Unaibsa)} ‘Thus . 2 fatto n(é) ds OF a Tank}? which completes the proof. The series >, c.J,(Ee4), with ¢, as given above, is called the Bessel series G for f(x). The conditions under which f(x) may be expanded in this way are given in the next theorem, which we quote without proof. Theorem 4.25 If f(0) is defined for 0 m>—I). PRooF Consider the integral I> f. (Leon g y(t) dt142 BESSEL FUNCTIONS cu. 4 and substitute the series (4.6) for J, to obtain ‘ ayeem—igmea SS ¢— yy (ae r=fia-*# em 2. Ue” S (x/2)n+#r ‘ 21mg 2m +2r+1 - 2 Yam +r iy Jo (Fo Bye de =F ty OPP [aye tum rs0 (on making the substitution u = ¢?). 1 But [d= wyrm-tume di ° = Bin —m,m+r+1) (by the definition of the beta function, pro- vided n — m >0 and m+r+1>0. This last condition is equivalent to m> —1, since r takes on values between zero and infinity.) _ Ta —mlm +r +1) ~ Tw+rtl) « (by theorem 2.7). Hence 1 = 4 (n —m) > (-1y ep 5 =a —m(3)* xe Yon eh = Wa - m(5)” “Tal ‘Thus (s/ayon Jas) = 2g which is the required result, Theorem 4.27 i e *J(bx) dx = (a>0). 1 Vee)§ 4.15 INTEGRALS INVOLVING BESSEL FUNCTIONS 143 PRooF From theorem 4.6 we have Jol) = } f cos (x sin ¢) dg. Hence [, e-*Jo(bx) dx = i eo } i cos (bx sin $) dp de = if [{. e-* Hexp (ibr sin 4) + exp (—ibx sin 4)} ar] as adolde (interchanging the order of integration) a fc [ee (Hla ~ ibsin gx} | exp {—(a + ib sin Dailen “Ralol” a Fibsing Band lo 1 1 1 ake “wang tay tama aft =F Nh Feng But this last integral may be evaluated by elementary means (e.g., by the substitution 1 == cot ¢) to give ° oe #7 s; oN Jabs) bx = Fala? +B) 1 ~ ya? + bY Theorem 4.28 i) Ta(bx) de = i (if n is a non-negative integer). Q Proor We first prove the result for n = 0 and n = 1, and then show that if the result is true for n = N, it is also true for = N + 2, thus proving the result for all non-negative integral 7. For n = 0 we take the limit as a—» 0 of the result of theorem 4.27, obtaining J; 700) dy -144 BESSEL FUNCTIONS cH. 4 For n = 1 we make use of theorem 4.8(ii), which says that eT} = Tale) so that, by taking 1 = 0, we have Je) = Fle and replacing x by bx gives yl) = lb) which is equivalent to 1 8 jab) = —J,(bx). Hence {Fre as = — $[ytom]” 1 b since J,(c0) = 0 and J,(0) =1, results which arg implied by theorems 4.21(i) and 4.22(i). Té we now use theorem 4, so) and integrate from 0 to 00, we obtain [709] - ° ma. * Tn-als) —Jnals)} dx. Tak 20) — Jul) — 0, we have £20} — Fak} i a) ~ Jnsa(s)} de and thus ff Jase) ax = |" Jat) ae Replacing n — 1 by N and x by by gives the result JF Fnslbey de = [ Jns) de ‘Thus if \ Jalbx) dx = i we also have { Tov afbx) dee = i and hence the proof is complete.§4.15 INTEGRALS INVOLVING BESSEL FUNCTIONS 145 Theorem 4.29 (0 JP Jabape ere de = CD at 2E(n + 2) ayn Va @ +o? (a > 0). Gi) sr Trlbex)aet+ eo ® dx = Proor (i) From equation (4.6) we have = 1 (bag\ 2 1) = > Oey (B) so that [7 Joey om ae 2 1 Beran pe “2 (YFG TF =a(3) J erate da. © But j ext tin dee ° ° portin = -t = Jp eget (writing t — ar) P(r + 2n + 1), =e (by definition (2.1) of the gamma function), so that i. "Ta(bx)a” o-% dae 7 1 b\ aren = 2Y ater e a(2) gore +20 +1) S 1 20 (2r + 2n 1 a yf eee quimi (using the result of theorem 2.2 that Im {-1) = nl'(n)).146 BESSEL FUNCTIONS cH. 4 But by theorem 2.10 we have PQs) _ ea te) Tete so that P TalBx)x" 0° dx _ s crt Lory a yee (a) Va Na) ozs tat be Xe 1 sien y| we - Tete n(ey = (Yaa 2 >< ) (By. jut ora = weal + my , shafts ff) BEDE) Hey.) (by the binomial theorem) a(n tre ay 1S) yy Met D Ot Deed. br Dio = get, (IS Tati () ie Ta +r+ yey. =e DOT Sper laa)’ on repeated use of the result xI'(x) — P(x -|- 1) (orem 2.2). neg, 2M 2 Hence { Tul bx)xe e- dx = min +)—>5 @y “ipa 7 _ a +4) Va (a+ bt (ii) This result follows immediately from result (i) if we differentiate both sides with respect to a.§ 4.15 INTEGRALS INVOLVING BESSEL FUNCTIONS 147 Theorem 4.30 () [> Jeet exp (—as8) de = oar exp (—B#/4a); @\" Tidbx) 3+ exp (ax?) de = geal” + exp (—B?/4a); PROOF The proof is very similar to the proof of theorem 4.29; so we shall not give it here, Theorem 4.31 © 0 (6 >a) (i i sin ax Jo(Bx) dx — pe m bea 1 (ii) Jjcos ax Jo(bx) dx = {ve aw «(>a ° 0 (b i. sin ax Jo(bx) dv 0 ita-: 6. °148 BESSEL FUNCTIONS cH. 4 Equating real parts of both sides gives © 1 f cos ax Jobe) de = ys — f cos ax Ju(bx) dx = 0 ifb a 4.16 EXAMPLES Example 1 Use the generating function to prove that Sil +9) = >, TV)» We have from theorem 4.5 that em fuera )}= > neve so that J,(x + y) is the coefficient of @ in the expansion a wo {ue +33(¢—")} But on{te +af/-B} ae Spowfo(—D} = > Ser. > Loy = > L@Loyr For a particular value of r we obtain #” by taking s — n — r, so that, for this value of r, we obtain the coefficient of 2 as J,(x)Jn—,(y)3 hence the total coefficient of /” is obtained by summing over all allowed values of 7. « Accordingly the coefficient of # is equal to >. J,(x)Ja-r(y) and thus Jal +9) = Y Tad)§ 4.16 EXAMPLES. 149 Example 2 Show that pre dx =i, o # i From theorem 4.28 we have i Tid) dx = 1. But from theorem 4.8(vi) we have 2 Fal) —Iu-sl0) + Jnsil) so that an [Sob Sols) Tole) gy. J Jn-a(x) de + . Tus (x) dx =i+ 1 “a and hence i Lol) gy * Example 3 If & are the solutions of the equation J_() = 0, show inat 5 5 JAE) stn (0 edEx) A with L 2 fic 4 ln x) Jo(E,x) de GP [[xtexJiea) dy ° ED ‘Yo evaluate the integral appearing in the numerator we use the series (4.6) for Jy: a 1 ‘tx\ IE DOM ar (3)150 BESSEL FUNCTIONS cH. + => ergy) (and we note that since J,(¢,) = 0 we shall have 1 (& > (1 og (§ ‘* =). (4.34) We shall also require the integral f. x" Inx de, ° Integrating by parts gives 1 1 » opt 4 1 i oe ae fgets o -1 jo on+T1 * 1 ~ [she +I ~~ @ en LaF (4.35) Thus f * ln x Jo (6st) dx : = S- ly rye CO ~ ora (using equation (4.35)) “8 pact ‘wame) -3 201 ef al -al Sra) 3} 1 + # (using equation (4.34). ‘Thus we have 1 EE) c=§ 4.16 EXAMPLES 151 and hence Tdges) ESE}? Example 4 Show that —Juls) V8) ~) Ys) = 4 where A is a consiani, und by considering ihe behaviour for smuil x show thai A=2n. We have Se, —T¥a} alnYn —In¥o =Ta¥ +In¥n Iu Vn Ie¥n =In¥n —In Vn (all functions having argument 2); but both J,(x) and Y,,(x) satisfy the equation wy” + ay’ + (a® — n2)y = 0 so that EUs LY} =Jn{-2¥ 4 — (@? =) ¥,} — (-, — (et — Yn = -a¥s —SiYe)- Hence, writing Jn, — JY, = 2, we have dz B & 7 = d. di = dy pte When integrated this gives Inz + Inx = constant which is equivalent to In sx = constant and hence sx constant A, say.152 BESSEL FUNCTIONS cH. 4 ‘Thus 2 = 7 so that, by the definition of z, ‘oa A In¥n ~In¥n = 5 For small valucs of x, we havc, by theorem 4.22, Joe) ~1, Vols) ~2 tne so that, considering the reiationship for n = 0, we have In¥n —Jn¥a ~ 12. a2 == Hence 2_A me wo 2 giving A-= and thus 1 a _2 Sols) V8) — Tila) Vals) = 2 Fern at S720) — Juan} We shall show this by first proving that 70} = S[ E40 Jost Iould3] and then integrating this relationship from 0 to x. We have de {C3 —IuaTand the) = (GP.§4.16 EXAMPLES 153 102 —SucsDues) + $CaTs —SocDnen — Jo-aFo) (it being understood that the argument of all Bessel functions here is 2) = 5 —JuTon) + ${2F, Maca ~ Jue) — (Fa Ions Seale 2a} (using theorem 4.8(iii), (iv) and (v)) 213 —Jo-sTua) +5 2Taadns = tf, Integrating from 0 to x now gives at : {lara = [Sur — por.) = SUR) — Jn alas) Example 6 Use the generating function to show that J,(—x) = (—1)'Jq(x)- We have, from theorem 4.5, cole se ve Siomeeb-3] -offo- 2a = doar = > Cent, Equating the coefficients of # on both sides gives Ax) (Wal).154 BESSEL FUNCTIONS cH. 4 PROBLEMS (1) Show that A tafale nus} = U8) — Jol) alt ‘ (2) Show that (i) f ‘Jolt cos 8) cos 6 do =" *, pale 1 — cos x ii) J, Jue 0s 6) do = (3) Show that J*(2)J_.(2) —Ja(x)J a(x) = A/x where A is a constant; by considering the behaviour for large values of x, show that A = (2sin nz)/x. (4) Find the solution of the differential equation (vag? tay =0 ° which is zero at x = 0. 2 yn (5) Show that ‘> 5 Jala) = Jol v(a? — 2ax)}. (©) Use the generating function to prove that 1 = {Jo(x)}? + 2a(x)}* + 2fFa(e)}? +... and deduce that | Jo(x) | <1, |Jn(x)|<1/v2 (1 =1,2,3...). (7) Show that [ "220 gy = ater! +H} (8) Prove that > Tallix)t = exp {e( _ ve k(x) and deduce that 1() = > Fusn(x)- (9) Show that exp &(: + a) = S L(x). 0 (10) Show that Ju(s), Ya(x), Ky(x) and I(x) all satisty the differential equation dy 2d%y Qn? $1d4y nt + 1 dy e= ) . Gite oe atte de ea. 2"PROBLEMS 155 (11) Expand x? ina series of the form S eddies) valid for the region 0 < ~ (—1)* ir bl (14) Prove that @ fj ebers ae = x bei x; ° (ii) fj tbci tar = —x ber’ x. ° (18) Show that, for large values of x, . Loy x a). (bers i pgaye os (5, —$)5 (ii) bee ven et! sin (3, A)HERMITE POLYNOMIALS 5.1 HERMITE’S EQUATION AND ITS SOLUTION, Hermite’s equation is given by dy, dy we 2x. ‘dz and in applications we are required to find solutions which are finite for all finite values of x and are such that exp (}x%) y(x) > 0 as x —> 00. ‘The methods of Chapter 1 are applicable, and if we try for a solution of the form - 2ny — 0, (.1) api Mes ale, 5) a Sy — we find that the indicial equation has roots s = 0 and s = 1, with a, in- determinate when s = 0, so that s = 0 gives the two independent series solutions. The recurrence relation for the coefficients then has the form Sra Ar =) a (x + 1)r +2) ‘This recurrence relation may now be used to construct the two series solutions. However, from these series it can be shown that both solutions behave like exp (x%) for large values of x. Thus they cannot satisfy the requirement that exp (4x%) y(x) > 0 as x —> 00; this can only be satisfied if the series terminate. From equation (5.2) we see that this will be so if, and only if, n is a non-negative integer, for then a, .. and all subsequent coefficients in the corresponding series will vanish. We shall now write the (5.2)§5.2 GENERATING FUNCTION 157 series in descending powers of x. Using equation (5.2), rewritten in the form __ © EN +2) a =~" yt we obtain the series n(n — 1) n(n — 1)(n — 2)(n — 3) aden — 0 “4 FS a 2.2.4 wt “gym aN at), 1 rey W24..m tf ma Se 1yM Da 2 Dy a 27.2.4, An if nis even (where [dn] = pce — 1) if n is odd) nt = an 2 CY; Fa = ‘The standard solution is obtained by making the choice of 2” for the arbitrary constant a,; the solution is then denoted by H,(x) and called the Hermite polynomial of order n: eS Hs) = 5 1 pg ane 63) 5.2. GENERATING FUNCTION Theorem 5.1 ->5 ala). PROOF We wish to pick out the coefficient of ¢* in the power series expansion of exp (2¢~ — 2), Now, Qe _ gtr ett e e _ -3& ors 2 o > 2158 HERMITE POLYNOMIALS cH. 5 For a fixed value of s we obtain t” by takingr + 2s =n, ie. = — 2s, so that for this value of s the coefficient of 2” is just given by (2x) —1) C) (n — 2s)!sV" ‘The total coefficient of 2 is obtained by summing over all allowed values of s, and, since r= — 2s, this implies that we must have n —2s>0, ie. s < 4n. Thus, if n is even, s goes from 0 to 4m, while if n is odd, s goes from 0 to 4( — 1); that is, in all cases, s goes from 0 to [4n] with [jn] defined as above. Thus we have: Td coefficient of = 2,(- DG yr gone Jane), (by definition (5.3)). 5.3 OTHER EXPRESSIONS FOR THE HERMITE POLYNOMIALS Theorem 5.2 do H,{s) = (—1p ee ProoF By use of the generating function of theorem 5.1 and Taylor’s theorem, which states that we have on . #,(x) = [z= oe] an = |— (2-1) — iz “ Tea on q = ~ eH e ‘Line i But @ a ae -—)= ate —),§5.3. OTHER EXPRESSIONS FOR HERMITE POLYNOMIALS 159 so that an an a — 9 = (1 ge — and we have of Ore H,(x) = (—1)het [iz e Tee 1 on =(-Ipe"Re Theorem 5.3 19 ~> (on (jes n(e) = 2" yexp (— a5) # PRooF We have 1 doe =tett 2 de and hence G é Ge pm gle, 1d S 1/1 a?\* ‘Thus exp (42 exp (2t%) — > 8) ot mon! I Expanding both sides in powers of t, we have 1d?) S 2m < i” oo (fa) 2 a= Dm (using the generating function property of theorem 5.1). + ‘Theexponentiatof an operator is defined by its power series expansion. ‘Thus, for ex- d Sifd\" Sa a d <1 df ores 0( > (ae) Dat dyn 82 that oo( 2) re > n! dx"160 HERMITE POLYNOMIALS cH. 5 Equating the coefficients of ¢" on both sides now gives (oCig) r tuo ->foo( 48) 5.4 EXPLICIT EXPRESSIONS FOR. AND SPECIAL VALUES OF, THE HERMITE POLYNOMIALS We may use either the definition (5.3), or thcorem 5.2 or theorem 5.3 to write down an explicit expression for the Hermite polynomial of any order we choose. For the first few orders we obtain and hence H,(x) = 16x — 482 + 12 IL fx) — 32x6 — 160x9 + 120%. Theorem 5.4 Qn)! Hy(0) = (1p OC), Hy, (0) =. PRooF From the generating function of theorem 5.1 with x = 0, we have < ” ete >, 0) which, on expanding the left-hand side in powers of ¢, reduces to DOT 2 Oh & Equating coefficients of corresponding powers of ¢ on both sides gives H,(0) = 0 if nis odd (which is equivalent to Hy ,,(0) = 0 with n a non-negative integer) and —1)" =, = Hap(0) << (with n a non-negative integer Jt aly nme§5.5 PROPERTIES OF HERMITE POLYNOMIALS 1oL which yields (20 Hafo) = (1 2, 5.5 ORTHOGONALITY PROPERTIES OF THE HERMITE POLYNOMIALS Theorem 5.5 FAL, Ax)H (x) de == 2014/0) nme Proor We have er S ny i a and eet tee _ = 3 0 i so that |" e- H(a)Ffa(x) dx is the coeficient of (sm)/(nim! in the expansion of P eo ete FH dy, But “ [er eeet ent as ae r exp[—&e — (6 +9} £6 + 9] dx =e |" exp [fe — 6 + 9) dy ~em cr exp (—u*) du (changing the variable of integration tow = x ~ (s+ ¢)) em (by the corollary to theorem 2.6) vs angmyn Dwr162 HERMITE POLYNOMIALS cH. 5 Hence the coefficient of (¢'s")/(n!m!) is zero if m én and is (x/7)2"n! when m =n. or, making usc of the Kronecker delta, ifm én ifm=n cr €o# Hy) Heyl) de = (4/7)2" Be 5.6 RELATIONS BETWEEN HERMITE POLYNOMIALS AND THEIR DERIVATIVES; RECURRENCE RELATIONS Theorem 5.6 (i) Hy(s) = 2nH, a(x) (n> 1); Hoa) = 0. ii) Haale) = 2xH,(x) — 2nH, (x) (n> 1); Aix) = 2xH(x). Proor (i) If we differentiate both sides of the generating function relationship with respect to «, we obtain So td >, His), = 5, exp ts — #%) = 2t exp (2tx — 1) 2 > Hunt, n=0 a =2> He S Ea 2 Agr Equating coefficients of #" gives for n = 0 Hx) =0 and, form > 1, Hy) 2Hy (2) at (mw —I)! which reduces to A,(x) = 2nH, (2).§5.7 WEBER-HERMITE FUNCTIONS 163 (ii) Differentiating both sides of the generating function relationship with respect to ¢ gives d ar = —-pP=— — apex te — B= 2 ~ Hels) and thus, performing the differentiations, Sontht (2% — 22) exp (2tx ~ #2) = > Ay Hal). We now note that the term with » = 0 does not contribute to the summation on the right-hand side (we remember that 0! = 1) and thus, on use of theorem 5.1, the above equation becomes Ax — 2) DE H,{3) = > won (a) which is equivalent to ae Hols) — 235 "His) = a which may be written in the form a > 5 Fale) — 2D ea Di Fils) = > Eta ita) ‘Thus, equating coefficients of t* for n > 1, we have 2v. ny vy — 2 Ana) 1 yy ais) abe’ (m= Alo nfo? which on multiplying throughout by n! becomes 2xH,(x) — 2nHys(%) = Hn sa(%)- Similarly, equating coefficients of 1° gives 2xH(x) = H,(x). 5.7 WEBER-HERMITE FUNCTIONS An equation closely related to Hermite’s equation is de J 4 (- dx? If we make the substitution - xy =0. (5.4)164 HERMITE POLYNOMIALS cH. 5 we find that the above equation reduces to dz dz np + (A= 1)z =0. (5.5) This is Hermite’s equation of order m with 2n = A — 1. If, as is usually the case in applications, we are looking for solutions of equation (5.4) which are tinite for all values of x, we must have a solution of equation (5.5) which does not tend to infinity any faster than exp (x?/2) as x tends to infinity, and by the discussion of Section 5.1 this implies that 4a — 1) given by B(x) ="? H,(x). ¥,,(x) is called the Weber-Hermite function of order n. 5.8 EXAMPLES Example 1 Prove that, if m s Fal) nao ™ =an> i a=o 1 = 2 2 Goatees setting r =m +n. ‘The coefficient of t” is therefore 1 * Gemini)§5.8 EXAMPLES 165 and hence the coefficient of 2" /n! is nt Ge my in-n), which proves the required result. Qn Example 2 Evaluate P wen Hi, (x)Hya(x) dx. We have, from theorem 5.6(ii) 2H, (x) = mH ya() + Hn si) so that er xent Hy(x)H, (x) de J ‘ eW* (nH x(a) + Mla gals) Ha) de = Kin — WV) aa, m # E20 + IM) O nt, (by theorem 5.5) = (W/m) Bua, + (V2 1)! Example 3 Show that P,(x) = te-" H,(xt) dt. (wal I, From equation (5.3) we have Halt) = >, (1 sg ere so that 2 ’ eH _f£_ (xt) di cami Je . . Holst) ‘a . = want! eo" > (Hy Flo ease WL gnatrsa trae 0 = > (varia — 2nh [oma tun 2a Ayn tr / S (varia api hm ori) (by theorem 2.4)166 HERMITE POLYNOMIALS cH. 5 SA eer — tye On — 2n)! Ly (are — 2r)i 22°F — 7) (by the corollary to theorem 2.10) _% 1 (2n — 21)! a ~ > Cy Brin — 27) — 7)! ro = P(x), (by equation (3.17)). PROBLEMS (1) Show that in wten® (H,(x)}* de = (V/n)2"aln + 4). (2) If f(x) is a polynomial of degree m, show that f(2) may be written in the form £() = > olla) rd 1 where "FE [7 er eee) dx, Deduce that ec e-* £(x)H,(x) dx = 0 if £(x) is a polynomial of de- gree less than n. (3) Show that e~* = z f e~" cos 2at dt. By differentiating this result Zn times with respect to x, show that gait ro 2 e—* 2" cos 2xt di, va 0 and obtain a similar expression for Hay .a(3). ie" f° Deduce that H,(x) = 2¢=i" j enh gm dp, ya Jo» (4) Use the result of problem 3 to show that = (x2 s Hal) HV) yy =(1—1)-"2exp ou (# + yy? S Onmt and deduce that Hanl(2) = > THN (1 ety exp Batt + 9):PROBLEMS 167 (5) Use the result of problem 4 to show that [> Gey eo de = 2 +s -« 4 . dye" * (: = 2 () SriJic T+ae = {. ive) Toa (6) Prove that () 2n, (8) = xP a(x) + Vilas Gi) 20¥(4) — InP," (a) = Vale): Gili) Pax) = x Pe) — Pa (7) Evaluate @ [7 Fal) ass Gi) re V(x) P(x) dee.6 LAGUERRE POLYNOMIALS 6.1 LAGUERRE’S EQUATION AND ITS SOLUTION Laguerre’s equation is dey saat and in applications we usually require a solution which is finite for all finite values of x and which tends to infinity no faster than e*/? as x tends to infinity, The methods of Chapter 1 apply, and if we look for a solution of the a-— oe +ny=0 (6.1) form 2(x, 8) = > @, x*+* we obtain an indicial equation with double root mb s = 0, and the recurrence relation _) (s+r—a) Bra = OE ‘The two independent solutions are then given by a(x, 0) and [Z| Aslsn0 "The second of these we know from Chapter 1 contains a term of the form In x, and so is infinite when x = 0. Since we require a solution finite for all finite x we can only obtain this from 2(x, 0). In this case the recurrence relation takes the form (=n) G4. = Gy (sn (6.2)§6.2 GENERATING FUNCTION 169 However, the infinite series obtained from this relation may be shown to behave like e* for large values of x, so that, from our remarks above, it does not behave well enough as x tends to infinity. ‘The way round this difficulty is to make the series terminate, and from equation (6.2) we see that this ‘happens if n is a positive integer. For, in this case, a, £0 but @n4, and all subsequent coefficients will vanish. In such a case the relation (6.2) is best written in the form Gai = and we obtain the solution a1 yaa fet Meet. (aye) att —rt Dey, } (the highest power of x being +") =a S (ary MD aS » al =a 2 -Y gaa We define the standard solution as that for which a, = 1; we shall call it the Laguerre polynomial of order 2, and denote it by L,(x): +1) L(x) = > (-1)yr ae (6.3) 6.2 GENERATING FUNCTION Theorem 6.1 exp {—at/(1 —1)} (1 — 2) 2, L,(x)t" Proor We have exp { at/(1 — t)}- (I ' 1 at170 LAGUERRE POLYNOMIALS cH. 6 1 Bu Gyn (ene: Ne $y C+ Mr +20 + 3) =l+(r4¢D4" Sa +... (by the binomial theorem) Sr + Me so that we now 1 have 1 (+s) = —a)}= yy CN ergs, i =q or xt/(1 — t)} 2 (~1) Gist wnt For a fixed value of r the coefficient of rin this expansion is obtained by takingr +s = 1, i.e.,s =n — r. Thus, for this value of r, the coefficient of i is given by nl . CY Gm ai" ‘The total coefficient of t* is obtained by summing over all allowed values of r. Since s = n — r, and we require s > 0, we must have 7 < n. Hence the total coefficient of 2” is given by ae 1y wera = Ly(3) (by equation (6.3)), which proves the required result. 6.3 ALTERNATIVE EXPRESSION FOR THE LAGUERRE POLYNOMIALS Theorem 6.2 L(x) = fe e*). PROOF Leibniz’s theorem for the nth derivative of a product states that d™ nt du do ae) = Zeon Sayin! daw da”