Lecture 16
Deformation Theorem 103
and
γ1b
f(z)dz =
P2P4
+
γ2b
+
P3P1
f(z)dz.
Figure 16.2 P1
P2
γ2a γ2b
P3
P4 P1
P2 P3
P4 γ1a
γ1b
Adding these equations, we find
γ1
f(z)dz =
γ1a
+
γ1b
f(z)dz
=
γ2a
+
γ2b
f(z)dz =
γ2
f(z)dz.
Example 16.1.
Determine the possible values ofγ
1
(z−a)dz,where γ is any positively oriented, simple, closed contour not passing through z=a.Observe that the function 1/(z−a) is analytic everywhere except at the pointz=a.Ifalies exterior toγ,then the integral is zero by Theorem 15.2. Ifalies insideγ,we choose a small circleγr centered ataand lying within γ.Then, from Theorem 16.1, it follows that
γ
dz z−a =
γr
dz z−a.
Now, since, onγr, z =a+reiθ (ris fixed),dz=rieiθdθ, and hence
γr
dz z−a =
2π 0
1
re−iθrieiθdθ = 2π
0
idθ = 2πi.
Hence, we have
γ
dz z−a =
2πi ifais inside γ 0 ifais outside γ.
Example 16.2.
Evaluateγ
1
z2−1dz,where the contourγ is given in Figure 16.3.
Figure 16.3
·
γr·
−1 1
γ
Clearly, the integrand 1/(z2−1) fails to be analytic at z=±1.Since the point 1 lies outside ofγ,the integral overγis the same as the integral over that small circleγrenclosing−1.Using partial fraction expansion, we find
γ
dz z2−1 =
γr
dz z2−1 =
γ
1
2(z−1)− 1 2(z+ 1)
dz
= 0−1
22πi = −πi.
Now we state the following result that extends Theorem 16.1.
Theorem 16.2.
Letγ, γ1,· · ·, γnbe simple, closed, positively oriented contours such that eachγj, j= 1,· · ·, nlies interior to γ,and the interior of γj has no points in common with the interior ofγk ifj =k (seeFigure 16.4, where for simplicity we have taken each γj to be a circle). If f is analytic on the closed region containing{γ}, {γ1},· · ·,{γn}and the points between them, then
γ
f(z)dz = n j=1
γj
f(z)dz. (16.1)
Figure 16.4 γ1 γ2 γ
γ3
γn
Deformation Theorem 105
Proof.
The proof of Theorem 16.2 is exactly the same as that of Theorem 16.1, except here we need to divide γinto several parts.Example 16.3.
Findγ
(5z−2)
z2−z dz,where γis given in Figure 16.5.
Figure 16.5
0
·
1·
γ
The integrandf(z) = (5z−2)/(z2−z) is analytic everywhere except for the zeros of the denominator,z= 0 and z= 1.Letγ1andγ2 be two small circles enclosing these points. Then, using Theorem 16.2, we get
Figure 16.6
0
·
1·
γ γ1
γ2
γ
f(z)dz =
γ1
f(z)dz+
γ2
f(z)dz.
Thus, we have
γ
5z−2 z(z−1)dz =
γ1
2 z + 3
z−1
dz+
γ2
2 z + 3
z−1
dz
= 2×2πi+ 0 + 0 + 3×2πi = 10πi.
Example 16.4.
Let γ be a simple closed contour that contains the distinct pointsz1, z2,· · ·, zn in its interior. We shall show that
γ
dz
(z−z1)(z−z2)· · ·(z−zn) = 0.
For this, we write the partial fractional decomposition 1
(z−z1)(z−z2)· · ·(z−zn) = A1
z−z1+ A2
z−z2 +· · ·+ An
z−zn
,
which is the same as 1 =
n j=1
Aj(z−z1)· · ·(z−zj−1)(z−zj+1)· · ·(z−zn).
Thus, comparing the coefficients of zn−1 on both the sides, we find that A1 +A2 +· · · +An = 0. Next, let γj, j = 1,2,· · ·, n be small circles with centerzj, that lie in the interior ofγ.Then, from Theorem 16.2 and Example 16.1, it follows that
γ
dz
(z−z1)(z−z2)· · ·(z−zn) = n j=1
γj
Aj
z−zj
= 2πi n j=1
Aj = 0.
Problems
16.1. Use an antiderivative to show that for everyγ extending from a pointz1 to a pointz2
γ
zndz = 1 n+ 1
z2n+1−zn+11
for n= 0,1,2,· · ·.
16.2. Letγbe the semicircle from−3ito 3iin anticlockwise direction.
Show that
γ
dz z =πi.
16.3. Evaluate each of the following integrals where the path is an arbitrary contour between the limits of integrations
(a).
i/2 i
eπzdz, (b).
π+2i 0
cos z
2
dz, (c).
3 1
(z−3)3dz.
16.4. Show that ifzi is the principal value, then
γ
zidz = 1 +e−π 2 (1−i), whereγ is the upper semicircle fromz= 1 toz=−1.
16.5. Let f and g be analytic for all z, and let γ be any contour joining the pointsz1 andz2.Show that
γ
f(z)g(z)dz = f(z2)g(z2)−f(z1)g(z1)−
γ
f(z)g(z)dz.
Deformation Theorem 107 16.6. Let f = u+iv be continuous and possess continuous partial derivatives with respect toxandy in a neighborhood ofz= 0.Show that f is differentiable atz= 0 if and only if
rlim→0
1 πr2
|z|=r
f(z)dz = 0. (16.2)
16.7. Use Green’s Theorem to show that Area of D = − i
2
C
zdz.
16.8. For each of the following functionsf,describe the domain of ana- lyticity and apply the Cauchy-Goursat Theorem to show that
γ
f(z)dz= 0,whereγis the circle|z|= 1 :
(a).f(z) = 1
z2+ 2z+ 2, (b).f(z) =ze−z. What about (c).f(z) = (2z−i)−2?
16.9. Let γ denote the boundary of the domain between the circle
|z| = 4 and the square whose sides lie along the lines x = ±1, y = ±1.
Assuming thatγis oriented so that the points of the domain lie to the left ofγ,state why
γ
f(z)dz= 0 when (a).f(z) = z+ 2
sinz/2, (b).f(z) = z 1−ez.
16.10. Let γ denote the boundary of the domain between the circles
|z| = 1 and |z| = 2. Assuming that γ is oriented so that the points of the domain lie to the left of γ, show that
γ
f(z)dz = 0 for the following functions:
(a).f(z) = ez
z2+ 9, (b).f(z) = cotz.
16.11.Letγ be the unit circle |z|= 1 traversed twice in the clockwise direction. Evaluate
(a).
γ
Log (z+ 2)dz, (b).
γ
dz 3z2+ 1.
16.12.Let r, R be constants such that 0 < r < R. Denote by γr the circlez=reiθ, 0≤θ≤2π.Show that
1 2πi
γr
R+z
(R−z)zdz = 1.
Hence, deduce that 1 2π
2π 0
R2−r2
R2+r2−2rRcosθdθ = 1.
16.13.Suppose that f is of the form f(z) =
n j=1
Aj
zj +g(z),
where g is analytic inside and on the simple, closed, positively oriented contourγcontaining 0 in its interior. Show that
γ
f(z)dz= 2πiA1. 16.14.Establish the following Wallis formulas
(a).
2π 0
cos2ktdt = 2π (2k)!
22k(k!)2, (b).
2π 0
cos2k+1tdt = 0, (c).
2π 0
sin2ktdt = 2π (2k)!
22k(k!)2, (d).
2π 0
sin2k+1tdt = 0.
Answers or Hints
16.1. !
γzndz= n+11 (zn+12 −z1n+1).
16.2. !
γdz/z= Logz3i
−3i (since dzdLogz = 1z on γ) = ln|3|+iπ2 − ln|3| −iπ2
=iπ.
16.3. (a). (1 +i)/π,(b).e+e−1,(c). −4.
16.4. Principal value zi =eiLogz. Hence, !
γzidz =−!0
πei(iθ)ieiθdθ =
−(1−2i)(e−π+ 1).
16.5. Use [f(z)g(z)]=f(z)g(z) +f(z)g(z).
16.6. As in Lecture 15, we have
!
|z|=rf(z)dz=! !
|z|≤r(−vx−uy)dxdy+i! !
|z|≤r(ux−vy)dxdy, and hence, from the Mean Value Theorem of integral calculus, we find
1 πr2
!
|z|=rf(z)dz=−(vx(x, y) +uy(x, y)) +i(ux(ˆx,y)ˆ −vy(ˆx,ˆy)), where (x, y) and (ˆx,y) are suitable points. Thus, (16.2) holds if and only ifˆ the Cauchy-Riemann conditions (6.5) hold atz= 0.Now, since the partial derivatives ofuandvare continuous, it is equivalent to the differentiability off atz= 0.
Deformation Theorem 109 16.7. In Green’s Theorem, we letP =xandQ=y,to obtain 2! !
Ddxdy=
!
Cxdy−ydx, and hence 2D = !
Cxdy−ydx. Suppose the parametric equation of C is x = x(t), y = y(t), so that z(t) = x(t) +iy(t) and x(a) = x(b), y(a) = y(b) (C is closed). Then 2D = !b
a(xy −yx)dt.
Now !
Czdz =!b
a(x−iy)(x+iy)dt =!b
a[(xx+yy) +i(xy−yx)]dt=
12[x2(t) +y2(t)]b
a+i!b
a(xy −yx)dt = 0 +i!b
a(xy −yx)dt = 2iD, and henceD=2i1 !
Czdz=−i2!
Czdz.
16.8. (a). f(z) = (z−(−1+i))(z1 −(−1−i)) and thus analytic if z=−1 +i or
−1−i; furthermore,!
γ dz
z2+2z+2 = 0, (b). ze−z is entire. (c). Analytic if z =i/2,cannot apply Cauchy-Goursat Theorem; however, sincef has an antiderivative onγ, !
γ
(2zdz−i)2 = 0.
16.9. (a).f(z) = [(sinz/2)(1)−(z+2)(1/2) cosz/2]/sin2z/2 if sinz/2= 0;
sinz/2 = 0 if and only if e−iz/2(eiz−1) = 0; i.e., if and only if eiz = 1, or z = 2kπ, k ∈ Z; none of this point is in the domain. (b). f(z) = [(1−ez)−z(−ez)]/(1−ez)2ifez= 1;ez= 1 if and only ifz= 2kπi, k∈Z;
none of this point is in the domain. Finally, note that γ = γ1+γ2 and
!
γ1f dz=!
γ2f dz= 0.
γ1
γ2
16.10.(a). f(z) =ez/(z2+ 9) is analytic ifz2+ 9= 0; i.e.,z=±3i.But
±3i are not in the domain, and hence, by the Cauchy-Goursat Theorem,
!
γ1f(z)dz=!
γ2f(z)dz= 0.Hence,!
γf(z)dz= 0.
γ γ1
γ2
(b). cotzis analytic if sinz= 0; i.e.,z=kπ, k∈Z.Butkπ are not in the domain for allk∈Z.Hence, as in (a),!
γcotzdz= 0.
16.11. Let γ be the circle |z| = 1 taken in the anticlockwise direction.
(a).!
γLog (z+ 2)dz= 0 by the Cauchy-Goursat Theorem, (b).!
γ
3zdz2+1 =
13
!
γ
√
3
2i(z−i/√3)−2i(z+i/√3√3)
dz= √6i3(2πi)−√6i3(−2πi) = 0.Thus,
−!
γLog(z+ 2)dz=!
γLog(z+ 2)dz+!
γLog(z+ 2)dz= 0 and−!
γ
3zdz2+1 =
!
γ
3zdz2+1+!
γ
3zdz2+1 = 0.
16.12.Use partial fractions. Compare the real parts.
16.13.Use Theorem 16.1 and Examples 14.6 and 16.1.
16.14. Let γ be the unit circle |z| = 1. Then, 2πi1 !
γ
z+1zn dz
z = 22πn×
!2π
0 cosntdt. Now expand (z+ 1/z)n using the binomial formula, and use Problem 16.13.