FIGURE 6.2.1 Contour forthe Proof of Theorem 6.2.4 6.2 Taylor Series, 271 When n = 0in (5), we i the formula gives dg erpret the zero-order derivative as f(zo) and O! = 1 so that (zo), Substituting (5) into (1) yields son "Eg — a, © ‘This setiesis called the Taylor series for f centered at zo. A Taylor series with center 2% =0, f@= LO oO a, o is referred to as a Maclaurin series. ‘We have just seen that a power series with a nonzero radius R of convergence ‘epresents an analytic function. On the other hand we ask: Since a power series converges in a circular domain, and a domain D is generally not circular, the question comes down to: Can we expand f in one or more power series ‘that are valid—that is, a power series that converges at z and the number to which the series converges is (z)—in circular domains that are all contained in D? The {question is answered in the affirmative by the next theorem, THEOREM 6.2.4 Taylor's Theorem Let f be analytic within a domain D and let zo be a point in D. Then f has the series representation ® valid for the largest circle C with center at zo and radius that lies entirely within D. Proof Let z be a fixed point within the circle C and let s denote the variable of integration. The circle C is then described by |s — zal = R. See FIGURE 6.2.1. To begin, we use the City integral formuta to obtain the value of fat z: f $04 bf Le ian, “ini k@=@-e- wo Lf fo “ari LeG= wo) i, By replacing z by c ~ z)/(8 ~ 20) in @) of Section 6.1, we have @-2" *e=a6- aand so (9) becomes 1 ff), za f SO) fO=TF Pezuet oT Par ar® @-wF f so Gea [SO ccpiag a or $e @-uy i) SP t 2ni (s — z)(s = zo)" Utilizing Cauchy's integral formula for derivatives, (6) of Section 5.5, we can rewrite (10) as ( ‘ on . He = feed + LEM — 2) 4 LD Ge — eg LOM asl + RD, (1D «eo 50), ee Bal) = Oi § ore nar * Equation (11) is called Taylor's formula with remainder R,. We now wish to show that R,(z) > 0.as n — 00. This can be accomplished by showing that | R,(z)| + 0 as n > 00, Since f is analytic in D, by Theorem 5.5.8 we know that | f(z)| has a maximum value M on the contour C. In addition, since z is inside C, |z — zol < R and, consequently, Is — zl =|s — zo — ( — zo) 2 Is — zol — Iz - zo] = R-d, where d = |z — zo| is the distance from z to zo. The ML-inequality then gives ze anf (s) a" M MR (d\" fate | ee SM op we MR tri Sea ae aye |S 2 ROR R= Raa (3) . Because d < R, (d/R)" > Oasn — coweconclude that | Rq(z)| > 0as n > 00. It follows that the infinite series | Ral) Shea) + FG — 29) 4 Fee aay. converges ‘0 f(@). In other words, the result in (8) is valid for any point z interior We ean find the radius of convergence of a Taylor series in exactly the same of th in 7 we simplify matters even further by noting that we ne ection: However, we can : , but an isolated singularity nae but is, nonetheless, analytic at all other ints th hood of the point. For example, z = 5i is an isolated singularity of f(@) = 1/(e ~ 5i). If the funetion f is coat ae a of at any point zo is necessarily R = oo. Using he following Maclaurin series representationsNote: Generally te formu in (8) wed es Just reson. > EXAMPLE 1_ Radius of Convergence at ‘Suppose the function f(z) = is expanded in a Taylor series with center 20 = 4—2i. What sits radius of convergence R? Solution Observe that the function is analytic at every point except at z = —1 +i, Seriatat ees eae tee 4-2iis lz—zol = V(-1-4P +1 - (2)? = This last number is the radius of convergence R for the Taylor series centered at 4-21. = If two power series with center zo, y ‘a(z— zo)» and Yae- zo) ms represent the same function f and have the same nonzero radius R of convergence, then 0 aah = 4 200,1,2, Stated in another way, the power series expansion of a function, with center zo, is, unique. On a practical level this means that a power series expansion of an analytic function f centered at zo, irrespective of the method used to obtain it, is the Taylor series expansion of the function. For example, we can obtain (14) by simply differen- tiating (13) term by term. The Maclaurin series for e® can be obtained by replacing, the symbol z in (12) by 2”. EXAMPLE 2 Maclaurin Series EXAMPLE 2 “Meee Find the Maclaurin expansion of /(2) = 7 Solution We could, of course, begin by computing the coefficients using (8). How- ever, recall from (6) of Section 6.1 that for zl < 1, L$ztPaPter. (a5)274 CHAPTER 6 Series and Residues I we differentiate both sides of the last result with respect to 2 then iar ae Gd a, des ti taetere Vt att ag tae See. 08 convergence of the last power series or 041422432 +° a Since we are using Theorem 6.2.2, the radius ofc is the same as the original series, R = 1- 7 {Wecan often build on results such as (16). For example, if we wantthe Maclaurin 3 z 3. ‘expansion of = — , we simply multipl 16) by 2°: pansion of f@) = Ga ply multiply (16) by 2 tte tact = aay ‘The radius of convergence of the las series is still R = 1. EXAMPLE 3 Taylor Series 7 ear Expand f(2 in a Taylor series with center zo = 2, z Solution. In this solution we again use the geometric series (15). By adding and subtracting 2i in the denominator of 1/(1 — 2), we can write 1 1 2) Wea 1-21 We no wie: m8 power es by using (15) with the symbol 2 1-21 replaced by the expression ayy] L Fe BREA 2+ Tae BW? + os ay Because the distance from the center z9 = 2i to the nearest singularity z = 1 is v5. we conclude that the circle of convergence for (17) is |z — be Taielby be oistatbepeceepscions nS tea2H1 8 sine ant 1) we represen te sane funtion fe) = 1/(1 = 2) ye . uitferen power series Thefisseies (15) has enter n= Oa adis Oconee Js | Thesecond seis (17) his ee y= 2/ and radius of convergence R = The two diferent circles of convergence aeilustaedin FIGURE 6.2.2 Thee n ofthe two circles, shown in color, is the region where both series eune eed seas sneered pnt apn th ee ome ton carverge inthe shaded region. ies teal 2"). Outside the colored region at least one of the (° — aa ce EXERCISES 6.2 _ answersto selected odd-numbered problems begin on page ANS-16. In Problems 1-12, use known results to expand the given function in a Maclautin series. Give the radius of convergence R of each series. z Le $O= Te 2 f= 1 3. f= = JO= Waa 4 10=q-5 5. f(z) 6 f(z) = ze" 7. f@) 8B. f(z) = coshz 9 F(@) = 0055 10. f(z) =sin3z Ul. f(z) =sinz? 12. f(z) = cos? z (Hint: Use atrigonometric identity) In Problems 13 and 14, use the Maclaurin series fore to expand the given function in a Taylor series centered at the indicated point zo. (Hint: z = z 29 + z0.] 13. f(z) =e, 0 =3i M4. f@)=@-Ne*, w=1 In Problems 15-22, expand the given function in a Taylor series centered at the indicated point zp. Give the radius of convergence R of each series. 21. fle) =c0sz, zo=/4 2. fle) =sinz, w= n/2 In Problems 23 and 24, use (7) to find the first three nonzero terms of the Maclaurin series of the given function. 23, fle) =tanz In Problems 25 and 26, use partial fractions as an aid in obtaining the Maclaurin series for the given function, Give the radius of convergence R of the series. g 26. f@) 28, f=276 CHAPTER 6 Series and Residues cout actually expanding, determine the radius of convergence R of Centered atthe indicated point. In Problems 27 and 28, wi the Taylor series of the given Funct A452 1 245i 28. f(z) = cotz, zo TH 29, What is the radius of convergence 2 of the Maclau 1e Maclautin series in Problem 242 2 see series in Problem 23? 30, What is the radius of convergence of the function in Taylor series centered at each of the n Problems 31 and 32, expand the gven Eee atc ich series. Sketch the region within indicated points. Give the radius of convergence R of ea ‘which both series converge. 1 32. f= 7 wo= lth w=3 1 3h f= zp abe In Problems 33 and 34, use results obtained in this section to find the sum of the given power 3. Exe $2 1d the Maclaurin series (14) by differentiating the Maclaurin series (13). i= [orn 35. 26 Terr nt i) dy te mee bed Maca ob inp Mas sore teins 37 nd 36 pine eae of Shen xeon wine ena von tts aces : 4/10, 38. 141) two terms 31 60 hem . o( 58, Focus on Concepts 39, Every function f has a domain of definition. Describe in words the domain of the function f defined by a power series center at zo 40. IF F(e) = DRo auc! and g(2) = DE bee! then the Cauchy product of f and g is given by Jee) = Sree where a= abe Write out the first five terms ofthe power series of f(2)g(2). 41, Use Problem 40, (12) ofthis section, and (6) from Section 6.1 to find the first four nonzero term of he Macau series of el — 2. Whats the radius of convergence of the 442, Use Problem 40, and (13) and (14) of this section to find the first four nonzero terms of the Maclaurin series of sin z cos z. Can you think of another way to obtain this series? 43, The function f(z) = sec is analytic at = 0 and hence possesses a Maclaurin series representation. We could, of course, use (7), but there are several alternative ways ‘obiain the coefficients of the series ee a age sece = ay bare bare? age foe ‘One way is to equate coefficients on both sides of the identity 1 — (see2)cos2 oF turin series of f. What is the radius of tata? tae 4--)(1-8 Find the first three nonzero terms of the Macla ‘convergence 2 of the series?278 6.3 Laurent Series CHAPTER 6 Series and Residues i ts Computer Lab Assignment ee Ta robles St-S4 ase CAS tiniest thes nonzero ems of he Tylor seis fhe ‘given function centered at the indicated point zo. In Mathematica fle) = coszLnz, = 1 +i Sh. fl) = etanz, x sake) 2 83. f=, w= 1 54. F@) Projects : 55. (a) You will find the following real function in most older calculus texts: ee, x #0 FOND A ssurngeio, aid in showing that f is infinitely Do some reading in these calculus texts as an aid f : differentiable at every value of x. Show that f is not represented by its Maclaurin expansion at any value of x #0. (b) Investigate whether the complex analogue of the real function in part (a), eve, 20 0 2=0, f@), is infinitely differentiable at z = 0. INTRODUCTION 1f a complex function f fails to be analytic at a point z = zp, then this point is said to be a singularity’ or singular point of the function. For example, the complex numbers z = 2i and z = —2i are singularities of the function L@ = z/(@ +4) because f is discontinuous at each of these Points. Recall from Scetion 4.1 thatthe principal value of the logarithm, Ln z, is analyti¢ at all points ticePt those points on the branch cut consisting of the nonpositive x-axis; thats, the branch point z = 0as well as all negative real numbers are singular points of Ln z. In this section we will be concemed with a new kind of ‘power series” expansion Cf F about an isolated singularity zo. This new series will imvolec negative as well 8 nonnegative integer powers of z — 29, ighborhood, or punctured open disk, 0 < Iz —zol < Rof tic: Forexample, we have just seen that z = 2i andz (2? +4), Both 2i and —2i are isolated singularities the neighborhood defined by |z — 2i| < 1, except ighbothood defined by |z — (—2i)] < l,exceptat and jean eis ee in the deleted neighborhoods 0 < |z— 2i| < 1 ~2i are singularities of f(z) since f is analytic at every point az = 2i, and atevery pointin the n \ 0 must contain points on the lar point z = zo of a function f is nonisolated iat least one singularity of f other than zo. For Oisa Nonisolated singularity of Ln z, Q A New Kind of Series “enainly {cannot be expanded ina pons iborhood of 29 contai in example, the branch point z= = zo is a singularity of a function f, then ‘Sr Series with zo as its center. However, about\.o Laurent series an isolated singularity z = zo, it is possible to represent f by a series involving both negative and nonnegative integer powers of z — zo; that is, aa @ 5 + tag + anc — 20) + cae — 20) + °° @=wP 2-2 As a very simple example of (1) let us consider the function f(z) = 1/(z — 1). As can be seen, the point z = 1 is an isolated singularity of f and consequently the function cannot be expanded in a Taylor series centered at that point. Nevertheless, f can be expanded in a series of the form given in (1) that is valid for all z near 1: 0 1 pte ‘The series representation in (2) is valid for 0 < |z — 1] < 00. Using summation notation, we can write (1) as the sum of two series L@ = Yare-ay* + \raule - zo)! @ ia ray IRs + +040-(@-1)40-(2-1P +e. 2) Fst ‘The two series on the right-hand side in (3) are given special names. The part with negative powers of z — zo, that is, Vare-20)* mi is called the principal part of the series (1) and will converge for |1/(z — 0)| < r* or equivalently for |z — zo| > 1/r* = r. The part consisting of the nonnegative powers of z~ 20, ay zy = 20} io & Lace -z0y', () i is called the analytic part of the series (1) and will converge for |z — zo| < R. Hence, the sum of (4) and (5) converges when z satisfies both |z — zo| > r and |z — zo| < R, that is, when z is a point in an annular domain defined by r < |z — zp| < R. By summing over negative and nonnegative integers, (1) can be written compactly as LO= VY ale - zo). sre ‘The principal part of the series (2) consists of exactly one nonzero term, whereas its analytic part consists of all zero terms. Our next example illustrates a series of the form (1) in which the principal part of the series also consists of a finite number of nonzero terms, but this time the analytic part consists of an infinite number of nonzero terms. EXAMPLE 1 Series of the Form Given in (1) SOT eee ‘The function f(z) = “y= is not analytic at the isolated singularity z= 0 and hence cannot be expanded in a Maclaurin series. However, sin z i an entire function, and from (13) of Section 6.2 we know that its Maclaurin series,FIGURE 6.3.1 Contour for ‘Theorem 63.1 FIGURE 6.3.2 C; and C; are ‘concent circles, converges for |z| < 00. By dividing this power series by z* we obtain a series for f ‘with negative and positive integer powers of z: principal analytic et et peavey ces te sas ta © ‘The analytic part of the series in (6) converges for |z| < 00. (Verify.) The principal parti valid for |z| > 0. Thus (6) converges forall z except at z = 0; that s, the series representation is valid for 0 < [2] < 00. = ‘A series representation of a function f that has the form given in (1), and (2) and (6) are such examples, is called a Laurent series or a Laurent expansion of f ‘about zo on the annulus r < |z— zo] < R. THEOREM 6.3.1 _Laurent's Theorem Let f be analytic within the annular domain D defined by r < |z—zol < R. ‘Then f has the series representation $@O= Vy a - zo) o valid forr < e~ tl < R. The coeficens a ars given by aff 2ni Jc (s — zo)! where C isa simple closed curve that les entirely within D and ha zp i ts interior See FIGURE 6.3.1. ds, k=0, 41, +2, ..., ® Proof Let Ci and C; be concentric circles with center zo and radii ry and Re, where r as. © As in the proof of Theorem 6,2.4, we can write Lf £0 = ini fy ae = Lae ~ 2, ao s Il (s where = poner k=0,1,2, a) Oni Jo, (= wayLf £0) oni 6.3 Laurent Series 284 ‘We then proceed in a ‘Manner simil 1 £6) Die. G2) —@ aay B (0 (9) of Section 6.2: J¢ so} 1 Pride, 2-20) | _ sae zz Jf £0 fist 2ni te, z= 29 Laat eo. Where £0 ge Tf, GoprT a ESL oy 13) and 1 LNs ~2F 4 Brie wa Ic, zs Now let d denote the distance from z to zo, that is, Iz — zo = d, and let M denote the maximum value of |/(2)] onthe contour Cy. Using js — zal 71 and the inequality (10) of Section 1.2, le-sl= le 20-6 ~ 22 le~ aol -ls-aol = d ‘The ML-inequality then gives ey LONS= 20)" | Met 10h = area SG a et HL ten Because ry < d, (r1/d)" + 0.as n —+ 00, and so |Ry(z)| > 0.as n > 00. Thus we have shown that aa) where the coefficients as are given in (13). Combining (14) and (10), we see that ©) yields a, S . $0=D Gat ds u(z — z0)*. (as) Finally, by summing over nonnegative and negative integers, (15) can be written as f(2) = DP 2o ae(z — Z0)*. Moreover, (11) and (13) can be written as a single integral: f@ =$ ——it 3 fa = or Oe where, in view of(5)of Section 5.3, we have replaced the contours Cy and Cs by any Simple closed contour C in D with z in its interior. = 1, £2, . = = -vy the prineipal part (4) is zero and case when d-1 = 0 for k = 1, 2,3, ny the principal i cues oa Tye ses Ts Leto a be ‘considered as a generalization of a Tuylor series,“Tec defi hy (6) a seldom used ‘Seetadinhe Remora atte ender secon @ FIGURE 6.3.3 Annular domains for Example 2 CHAPTER 6 Series and Residues, ‘The annular domain in Theorem 6.3.1 defined by r < z—zol < R need not have the “ring” shape illustrated in Figure 6.3.2. Here are some other possible annular domains: (Wr=O, Riinite, (ir #0, R= oo, and (ii) r= 0. Inthe first case, the series converges in annular domain defined by an 0 < [z~ zal < R. This isthe interior of the circle [z — zo] = ® except the point zai in other words, the domain is a punctured open disk. In the second case, the annular domain is defined by lz — zal and consists ofall poins exterior to the circle [2 ~ zol = r. In the third case, the domain is defined by 0 < [z — zol. This represents the entire complex plane except the point zo, The Laurent series in (2) and (6) are valid on this last type of domain, "The integral formula in (8) for the coef =00. ts of a Laurent series are rarely used in actual practice. As a consequence, finding the Laurent series of a function in a specified annular domain is generally not an easy task. But this is not as disheartening as it might seem. In many instances we can obtain a desired Laurent series either by temploying a known power series expansion of a function (as we did in Example 1) for by creative manipulation of geometric series (as we did in Example 2 of Section 6.2). The next example once again illustrates the use of geometric series. EXAMPLE 2 Four Laurent Expansions EXAMPLE? ooo Expand f(z) = ——1— in a Laurent series valid for the following annular domains. w-) @O arTeaesn [-@-)+@-1%-@- +e.) 14+@-)-@-1F + ‘The requirement that z # 1 is equivalent to 0 < [z— 1], and the geometric Series in brackets converges for |z ~ 1| < 1. Thus the lat series converges for 2 satisfying 0 < Jz — 1] and |z 1] < 1, thatis, for 0 < |z— 1] < 1. (@) Proceeding as in part (b), we write pe eyes L fO" Fe) "EP 1p 1 -eol cH “@- GP eG Because the series within the brackets converges for | 1/(2 — 1)] < 1, the final series converges for 1 <|z—1|. = EXAMPLE 3__Laurent Expansions Pm aan ee Scere oe ce ee i ies valid for (a) 0 <|z—1] <2 Expand f(@) = page ay it Laure seres valid for (@) 0 < |= — 1] and (b) 0 <|z—3] <2. in parts (c) and (d) of Example 2, we want only powers of z — 1 Seton prs oad) of Esl, we mam uy pve f= writing 1 aE Bt A GF 24G=0) * 3G= |_ 1 ry SO = GaG=D =OM CHAPTER 6 Series and Residues jon 6.1 with the symbol z replaced by (z — 1)/2, @-1F ,@- +S at and then using (6) of Secti S@ ted Ege D oe a6 () To obtain powers of z — 3, we write z — 1 = 2+ (@—3)and We now factor? from this expression a z = R+e@-317 LO= GaypE= “= 1 Hey “elt a] - : z A hs pint we can obtain power serie for [1+ a ] by orn te binomial expansion,* (2) (2-3 (2-3) 3)? (2-3-4) (2-3)° po= gts + P(E) +S ( z) PPR ES) + in the brackets is valid for |(z — 3)/2| < 1 or Je—3| <2. Multiplying this series by —— O<[e-3/ <2: f@= The binomial seri FELT HVE # Laurent series that is valid for 3 +7ge-D 1 ‘d % ger. EXAMPLE 4 A Laurent Expansion ae SU OT EON UN eth ae ae . 847 Expand f(@) = <7 ina Laurent series valid for 0 < lel <1. Solution By partial fractions we ean rewrite f as 8: f= St 1 =ot+ ‘Then by (6) of Section 6.1, 2-2) 27 Te 9 Tag TIF 9 +9P HO, The foregoing geometric series conve Bye emetic ras mnverges for |z| <1, but after we add the term I/z 1 JO +949 492 4... is valid for0 < [el <1, ‘ Inthe precedin ‘ foreach Laurent se iat de {he point atthe center ofthe annular domain of valid! of Theorem 6.3.1 shows thar i neaty ofthe function fA re-examination 8 Need not be the case, “tere bia ibe eer wt 4 gy EXD 4 wt ear MDa y Me NO=95 . gel Bane is6.3 Laurent Series 285 EXAMPLE 5 A Laurent Expansion Expand f(z) = in a Laurent series valid for 1 < \z ~ 2] <2. i 2-1) Solution The specified annular domain is shown in FIGURE 6.3.4. The center of this domain, z = 2, is the point of analyticity of the function f. Our goal now is to find two series involving integer powers of z ~2, one converging for | < |z ~ 2| and the other converging for |z ~ 2] < 2. To accomplish this, we proceed as in the last ‘example by decomposing f into partial fractions: FIGURE 6.3.4 Annular domain for Example 5 a Now, 2 2B 27 _@=2P ] @-2 | (-29 Bp to ‘This series converges for |(z — 2)/2| < 1 or |z — 2| < 2. Furthermore, A@ 1 1 2+@-H Gp apis Ns cea ieee ee @-3" @-2 Ga” converges for |1/(z ~ 2)] < 1 or 1 < |z ~2|. Substituting these two results in (17) then gives 1 2 ab ae a G2 @-2PF *z- is valid for z satisfying lz — 2| < 2 and 1 <|z—2); in other words, for | < lz —2| <2 27 EXAMPLE 6 A Laurent Expansion in a Laurent series valid for 0 < |z| < 00. Expand f( Solution From (12) of Section 6.2 we know that forall finite z, that is, |2| < 00, eae Fett at as)CHAPTER 6 Series and Residues We obtain the Laurent series for / by simply replacing z in (18) by 3/2, z # 0, as) in Sections 6.5 and 6.6. Replacing the complex variable s with the usual | @ In conclusion, we point out a result that will be o I ‘symbol z, we see that when k = —1, formula (8) for the Laurent series | 1 ; coefficients yields a1 = >> ¢ ‘f(2)dz, or more importantly, ic f fede= Ixia. 2) 2 i) Regardless how.a Laurent expansion of a function fis obtained in specified annular domain it is the Laurent series; that is, the series obtain is unique. Z oe EXERCISES 6.3 Answers to selected odd-numbered problems begin on page ANS-1T. In Problems 1-6, expand the given function in a Laurent series valid forthe given anne e082 9 sinz 1 s@=S0< 121 3B O ina Laurent series valid for the indicated annular 7. 0<(2<3 8 O0<|2-3/ <3 9% kel >3 10. |2—3)>3 MW. Ve te=4) <4 RB i2 lel > 15. 0<[e~1)< <1 1. 0< [2-2 <1 In Problems 17-20, expand f(g) = 2 expand f(2) = —__2 See Gey in a Laurent series valid for the given 17. 0 1 n-o In Problems 23 and 24, expand f(z) in a Laurent series valid forthe given annular domain, 23. 0< [2-2 <1 24. 0<|z-1<1 paste @-2G- In Problems 25 and 26, expand f(z) = in a Laurent series valid for the given annular domain, ae ten3_a-D 44 28. O<|z)<1 3» O<|z—1]< lint: —— = —————_. tel 26. 0<|e—11 1 [ain - eet 2242 In Problems 27 and 28, expand f(@) = annular domain, h Uetenil 28. 0 < [221 in a Laurent series valid for the given In Problems 29 and 30, use cos = 1 — 5 vesing = 2— 5 + 5 — sand long i aI 1” ST division to find the firs three nonzero terms of a Laurent series of the given function f valid for 0 <[z| axle — 20) = ae 20) + Doauz—zo}¥ A) i= te tt is the Laurent series representation of f valid for the punctured open disk 0 <|z—zol < R. We saw in the preceding section that a Laurent series (1) consists cof two parts. That part of the series in (1) with negative powers of z — zo, namely, a a te @ Dare - 20) t= = is the principal part of the series. In the discussion that follows we will assign different names tothe isolated singularity z = zo according to the number of terms in the principal par. O Classification of Isolated Sin point z = zo of a complex function the principal part (2) ofits Laurent an infinite number of terms, gular Points Anisolated singular J is given a classification depending on whether expansion (1) contains zero, a finite number, ot (0 If the principal part is zero, that is, all the coefficients a in (2) , then z= zois called a removable singularity, CE ee (i If the principal part contains a fini called a pole. If, in this case, the then we say that z = 20 is a pole the principal part (2) contains ex order 1 is commonly called a si ite number of nonzero terms, then z last nonzero coefficient in (2) of order n. If z actly one term wit imple pole, (iii) If the principal part (2) contain an it Wat nia pt) cular tnitely many nonzero terms, then z = 20 ois isa_,.n >, 2o is pole of order 1, then ith coefficient ay. A pole of6.4 Zeros and Poles i) 0 Laurent Series for 0 < |z— zo] < R Removable singularity 40 + a1(& — 20) +an(z — 29)? +++ Pole of order n iz ca wot op (@ — zo) +- @= wy * Gay tty tate wt Simple pole Fag tot ale — 29) + le — 29 + Essential singularity |... 4 42 ai 2 +o aay(z — 20) + ax(z — 20)? +++ Gea tag tot ale — 20) + axle - 20 TABLE 6.4.1 Forms of Laurent series Inprut puapaph Reveal it sever ines » EXAMPLE 1 Removable Singularity Proceeding as we did in Example 1 of Section 6.3 by dividing the Maclaurin series for sin z by z, we see from sing 2 eae es 3 7 ata @ that al the coefficients inthe principal part ofthe Laurent series are zero. Hence is a removable singularity of the function f(z) = (sinz)/z. = Ifafunction f has aremovable singularity atthe point z = zo, then wecan always supply an appropriate definition for the value of f (zo) so that f becomes analytic at z = zo. For instance, since the right-hand side of (3) is 1 when we set z = 0, it makes sense to define (0) = 1. Hence the function f(z) = (sin z)/z, as given in (3), is now defined and continuous at every complex number z. Indeed, f is also analytic at z = 0 because it is represented by the Taylor series 1 — 22/3! + 24/5! — centered at 0 (a Maclaurin series) EXAMPLE 2 Poles and Essential Singularity (a) Dividing the terms of sin: for 0 < [zl < 00. From this series we see that a_y #0 and so z = O is a simple ste of the function f(2) = (sin2)/z%.Inlike manner, wesee that = Oisapole of aerier3 of the function f(z) = (sin2)/z* considered in Example 1 of Section 6.3. of Section 6.3 we showed that the Laurent expansion of ample 3 (0) In Brame 2 cg 3) valid for 0 < le Il <2 was fO=VE- principal past f@O=—"7Qr1f He-Y 8 16 1 £0, we conclude that z = 1 isa pole of order 2 +40, Since a-2jon 6:3 we see from (19) thatthe principal part ofthe 6 of Section 6.3 we $0 ye valid for 0 < [el © oo contains an init (©) InExample ion f(2) ; expansion of the function f°) = ows that z = O is an essential sin, number of nonzero terms. sy i tibet Fis 0 Zeros Recall, a number zo is zero of ee = f if Fo) = 0. We say that an analytic function f has a zero of order m at z= zo tp isa e10 off and of its rst I derivatives Mo} but Fe) £0, ‘A zero of order n is also referred to as a zero of multiplicity 1, For exam S(2) = (@ — 5P we see that f(5) = 0, FS) = 0, £5) = 0, but £5) Thus f has a zer0 of order (or multiplicity) 3 at zy = 5. A zero of order 1 is cal simple zero. ® ple, for #0, lled a ‘The next theorem is a consequence of (4). THEOREM 6.4.1 Zero of Order n A function f that is analytic in some disk lz—zol < R has a zero of order n at z= 20 if and only if f can be written F@) = & — 20)" $2), © Where is analytic at z = zp and $(zo) # 0. +2, ... , it follows from (4) that the have the form f@ nC — 20)" + an si(z — za) "+ agate — zy"? 4 = © 20)" [an + anaae — 29) + n42(% — 20)? +++ With the power-series identification $0) = ay + dnsa(e — 20) 4 Gn42(@ = 20)? +++ We conclude that ¢ is an F€G0)/n\ 0 from (a), Funetion and that $¢20) = ay + 0 because de EXAMPLE g Order Of a Zero (13) of seu cttetion (2) = zig 2 ce ection 6.2, has a zero at z = 0, If we replace z by 2 oO |O@) = ant 6.4 Zeros and Poles 294 is analytic and (0) = 1, When compared to: in aos Feared ommared 105), the result in (6) shows that «= 0 O Poles We can characterize a pole of order n in a manner analogous to (5). THEOREM 6.4.2 Pole of Order n A function f analytic in a punctured disk 0 < | z—zol < R has a pole of order natz = zo if and only if f can be written oo fo= Gye o where @ is analytic at z = zo and $(z0) # 0. Partial Proof As in the proof of (5), we will establish the “only if” part of the preceding sentence. Since f is assumed to have a pole of order n at zo it can be expanded in a Laurent series an 4 Oty (20 (wer 2-20 valid in the disk 0 < |z — zo] < R. By factoring out 1/(z — zo)", (8) confirms that f ‘can be written in the form @(z)/{( — zo)". Here we identify tag tay(z—z0) ++", (8) $a2le — zo)? +412 = Zo)! + anlz zo)" Fae— za" +---, OC) as a power series valid for the open disk |z — zal < R. By assumption, z = zo is a pole of order n of f, and so we must have a.» # 0. If we define (za) = an, then it follows from (9) that ¢ is analytic throughout the disk |z ~ zol < R. = O Zeros Again A zeroz = zo of an analytic function f is isolated in the sense that there exists some neighborhood of zo for which f(z) # 0 at every point z in that neighborhood except at z = zo. As a consequence, if zis a zero of a nontrivial analytic function f, then the function 1/f(2) has an isolated singularity at the point 7 are following result enables us, in some circumstances, to determine the poles of a function by inspection. THEOREM 6.4.3 Pole of Order n If the functions g and h are analytic at 2 1 and g(zo) # 0, then the function f (2 ‘9 and h has a zero of order n at g(z)/h(z) has a pole of order n mnction ft has zero of order 1, (5) gives f(z) = (z ~ z0)"$(2), the fu i" Proof Beate ie o and O(za) # 0. Thus f can be written where ¢ is analytic at z £0) : 8/00), 7 @-ay tO) = zp and 6(20) # 0, it follows that the function g/ Since g and pare ana 0 implies g(co)/4(2o) # 0. We conclude from is analytic at zo. Moreover, farm 642 tha the Function f has a pole of order at Zo, = £@CHAPTER 6 Series and Residues When 1 = 1 in (10), we see that a zero of order 1, or a simple zero, in the denominator h of f(z) = 8(2)/h(2) corresponds to a simple pole of f- jon of the rational funct (2) Inspec n 224+5 {0 = e+e —F shows that the denominator has zeros of order 1 at z= 1 and z = —5, and a zero of order 4 at z = 2. Since the numerator is not zero at any of these points, it follows from Theorem 6.4.3 and (10) that f has simple poles at z = 1 and , and a pole of order 4 at z = 2. z= (b) In Example 3 we saw that z = O is a zero of order 3 of z sin z?, From Theorem 6.4.3 and (10) we conclude that the reciprocal function f(z) = 1/(zsin 2”) has a pole of order 3 at z = 0. | Remarks Me | @ From the preceding discussion, it should be intuitively clear that if a fune- | tion f has a pole atz = zo, then |f(2)| =» 00 as z —> zo from any direction. From (i) of the Remarks following Section 3.1 wecan write lim f(z) = 00. to do this, you may encounter the term meromorphic. A function f is meromorphic if it is analytic throughout a domain D, except possibly | for poles in D, It can be proved that a meromorphic function can have | fat most a finite number of poles in D. For example, the rational function | io Qi 1/(* + 1) is meromorphic in the complex plane. A |G 1 you peruse other texts on complex variables, and you are encouraged EXERCISES 6.4 _nswersto selected odd-numbered problems begin on page ANS-18, In Problems 1-4, show that 2 = ion of f(0) so that fis anal a removable singularity of the given function. Supply 8 az enact 2 fata 4 fqn oi oose? sin? Jn Problems 5-10, determine the zeros and their order fo the given function, 5. f@)=@ 42-18 6 fe 7 §@) 9% fe) 6 ee 8 (a) = site 10. f(z) = zet6.4 Zeros and Poles 283 4n Problems 11-14, the indicated number isa zero of the given function. Use a Maclaurin or ‘Taylor series to determine the order of the zero, MW. f@) = x1 ~ cos? 2); 13 f@) qn Problems 15-26, determine the order of the poles forthe given function. 14. $@ mmitcte z= 3e-1 1 fey = —2e= 1 1O= srs Leas =1 17. f= — 1 = _ 10- ese 8 10 = ery 19. f@)=tanz 20. fey= SE 2. fa=$ z wm, sya 55+ am fq) = ERE 2 In Problems 27 and 28, show that the indicated number is an essential singularity of the given 1 funtion 2. f@ (2e=0 28 70)=(¢~Ye0s(5):2= 29, Determine whether 2 = 30. Determine whether 2 = is an essential singularity of f(z) = et", isan isolated or nonisolated singilarity of f(z) = tan(1/z). Focus on Concepts 31, In part (b) of Example 2 in Section 6.3, we showed thatthe Laurent series representation of f@ valid for el > Lis See gga h yg Lp ly, Ongtatats ‘The point z = Ois an isolated singularity of f, and the Laurent series contains an infinite ‘number of terms involving negative integer powers of z. Discuss: Does this mean that z= 0 isan essential singularity of 7 Defend your answer with sound mathematics. 32, Suppose f and g are analytic functions and f has a zer0 of order m and g has zero of ‘order n ai z = zo. Discuss: What is the order ofthe zeto of fg at zo? of f + g at 29? 33. An interesting theorem, known as Picard’ theorem, states that in any arbitrarily small neighborhood of an isolated essential singularity zo, an analytic function f aesumes ‘every finite complex value, with one exception, an infinite numberof times. Since =~ 0 is an isolated essential singularity of f(z) = e", find an infinite number of = in any neighborhood of z = 0 for which f(2) = i. What isthe one exception? Tha is, what is the one value that f(e) = e¥ does nt take on? ed neighborhood ofan isolated singularity zo, Cassi ‘Suppose | (2) is bounded ina deleted neighborhood sgularity zo, Classify sammie of he thee Kinds folate singular ltd on page 289, Jus youranswer with sound mathematics. tom f(z) has a zero of order n at z 35, Suppose the analytic function f( nate [/Cor"sm a postive integer, has a zero of order mn at z = z. In this problem you are guided through the start of the proof of the proposition: ‘The only isolated singularities ofa rational function fare poles or removable singularities. o. Prove that the function