Complex Analysis: Handout-7 (Path Integrals)

S. Kumaresan
School of Math. and Stat.
University of Hyderabad
Hyderabad 500046
[email protected]

1 Path Integrals

Definition 1. Let γ be smooth path in U and f : [γ] → C be continuous. We define the

integral of f over γ by setting
Z Z b
f (z)dz := f (γ(t))γ 0 (t)dt.
γ a

If γ is just a path, using the above notation, we set

Z n−1
X Z tj+1
f (z) dz := f (γ(t))γ 0 (t) dt.
γ j=0 tj

Note that the integralsR on the RHS make sense, since γ 0 exists and is continuous on [tj , tj+1 ].
The complex number γ f is called the integral of f over the path γ.
R Rb
Note also that if γ(t) = t, t ∈ [a, b], then γ f (z)dz = a f (t)dt.
R
Show that γ cos z dz = cosh(1) − i sinh(1) where γ is the line segment from −π/2 + i to
π + i.
Example 2. Let γ, σ, τ : [0, 2π] → C∗ be given by γ(t) = eit , σ(t) = e2it and τ (t) = e−it .
Note that [γ] = [σ] = [τ ]. Let f (z) = z −1 for z ∈ C∗ . We have γ f = 2πi, σ f = 4πi and
R R
R
τ f = −2πi. Thus the path integral depends on the path and not on the trace. See also the
next exercise.
R
Ex. 3. Compute the following path-integrals γ f (z) dz:
20(1+i)
(1) f (z) := |z|2 and γ is the line segment from 2 to 3 + i. Ans. 3 .
−(1+i)
(2) f (z) := Re (z) and γ is the line segment from 1 to −i. Ans. 2 .
Theorem 4. Let f : U → C be continuous. Assume that there exists an F : U → C such that
F 0 = f on U . Let γ : [a, b] → U be any path. Then
Z
f (z) dz = F (γ(b)) − F (γ(a)).
γ
R
In particular, if γ is closed, then γ f (z) = 0.

1
 R
Corollary 5. If f : U → C is continuous and has a primitive in U , then γ f = 0 for any
closed path γ in U .

Example 6. Show that γ (z − a)n dz, where γ(t) = a + reit , 0 ≤ t ≤ 2π given by

R

(
n 6= −1
Z
n 0
(z − a) dz =
γ 2πi n = −1.

Ex. 7. Let γ(t) = eit , 0 ≤ t ≤ 2π. For n ∈ N, show that

Z
1 (2n)!
(2 cos θ)2n dθ = .
2π γ n!n!

Hint: Evaluate γ (z + z1 )2n dz

R
z using binomial expansion. Among the terms of the form
2n 2n−k −k


k z z , which is the term that will contribute to the integral? Look at Example 6.
R R
Proposition 8. Let γ be a path and f : [γ] → C be continuous. Then γ̃ f dz = − γ f dz.

Proposition 9. Let f : U → C be continuous with a primitive F in U . Let γj , j = 1, 2, be

paths
R in U both having Rthe same initial (resp. terminal)
R points z and w respectively. Then
γ1 f = F (w) − F (z) = γ2 f . Thus, in this case σ f depends only on the end points of σ for
any path σ in U .

Proposition 10 (ML Inequality). Let γ : [a, b] → C be a path. Let f : [γ] → C be continuous

with |f (γ(t))| ≤ M for all t ∈ [a, b]. We have
Z
| f dz| ≤ M L(γ). (1)
γ

Ex. R11. Establish the following:

(1) | γ z 2dz+4 | ≤ (RπR
2 −4) where γ(t) = Re
it for 0 ≤ t ≤ π and R > 2.
R −z
(2) | γ e dz| ≤ 2, where γ is the line segment from −i to i.
√
(3) | γ dz
R
z 4| ≤ 4 2 where γ := [i, 1].
R ez
(4) | γ z dz| ≤ 2πe where γ is the unit circle with the standard parametrization.

Ex. 12. Let f : C → C be continuous and bounded. Let γR (t) := Reit for 0 ≤ t ≤ 2π. Show
1
R f (z)
that limR→∞ 2πi γR (z−z0 )2 dz = 0 for each z0 .
R
Proposition 13. Let f : U → C be continuous. Then f has a primitive in U iff γ f = 0 for
any closed path γ in U .

Proposition 14. Let γ : [a, b] → C be a path. Let fn : [γ] → C beR continuous

R (n ∈ N) and
that fn converge uniformly on [γ] to a function f : [γ] → C. Then γ fn → γ f .

Proposition 15. Let γ(t) := z0 + reit for 0 ≤ t ≤ 2π. Then

(
2πi a ∈ B(z0 , r)
Z
dz
= (2)
γ z−a 0 a∈/ B[z0 , r].

2
Proof. Note that this is not Example 6. The trick here is to write
1 1 1 1 1
= = a−z = ,
z−a (z − z0 ) − (a − z0 ) (z − z0 )[1 − z−z ]
0 z − z0 1 − w
0

where w := a−z
z−z0 .
0
If a ∈ B(z0 , r), then |w| < 1 and we can substitute the geometric series
1
expansion for 1−w in the integral. Since the convergence is uniform on [γ], by Corollary ??,
we have
Z Z ∞
1 1 X
dz = (a − z0 )n (z − z0 )−n
γ z−a γ z − z0 n=0
∞
X Z
= (a − z0 )n
(z − z0 )−n−1
n=0 γ

= 2πi,
where we have used Example 6 to each of the terms of the series in the last but one line.
1 1 1 z−z0
If a ∈
/ B[z0 , r] we then write z−a = − a−z 0 1−w
, where w = a−z0 . Proceeding as above, we
get
Z Z ∞
1 1 X
dz = (a − z0 )−n (z − z0 )n
γ z−a γ a − z0 n=0
∞
X Z
−n−1
= (a − z0 ) (z − z0 )n
n=0 γ

= 0,
again by Example 6.
R f (w)
Theorem 16. Let γ be any path and f : [γ] → C be continuous. Then F (z) := γ w−z dw is
differentiable on C \ [γ]. In fact, F is analytic in C \ [γ].

Proof. The trick employed in the proof of the last proposition is implemented again.
Note that C \ [γ] is open.1 If z0 ∈ / [γ], choose r > 0 such that B(z0 , r) ∩ [γ] = ∅. If
z−z0
z ∈ B(z0 , r/3), then | w−z 0
| < 1/2 for w ∈ [γ]. See Figure ??. Thus the series
∞ 
f (w) X z − z0 n

f (w)
=
w−z w − z0 w − z0
n=0
of functions in w for fixed z and z0 converges uniformly on [γ]. Therefore,
∞
X
F (z) = an (z − z0 )n ,
n=0
R f (w)
where an := γ (w−z)n+1 dw.

1
We show that [γ] is closed in C. Let p ∈ C be a limit point of [γ]. Then there exist tn ∈ [a, b]] such
that γ(tn ) → p. Since tn ∈ [a, b], by Bolzano-Weierstrass theorem applied to [a, b], (tn ) has a convergent
subsequence, say (tnk ), converging to t0 ∈ [a, b]. Since γ is continuous at t0 , we deduce that γ(tnk ) → γ(t0 ).
Hence p = γ(t0 ), by the uniqueness of the limit. We may also show that [γ] is closed, by observing that [γ] is
compact.