Let C1 and C2 be 2 positively orientedsimple closed contours, with C2 lying inside the interior of C1. Let f be analytic in a domain D that contains both C1 and C2, together with the region between them.
∫C1f(z)dz=∫C2f(z)dz
The contour C₁ encloses the contour C₂, both positively oriented. The
function f is analytic on the shaded region between them, and its singular
points (×) lie inside C₂.
The integral over a closed contour depends only on which singularities the contour encloses, not on its shape.
To evaluate ∫Cf(z)dz, replace C by a small circle ∣z−a∣=ε around each singular point a inside C.
Corollary of DIT
Let z0 be a fixed complex number, and let C be a simple closed contour with positive orientation such that z0 lies inside C.
∫Cz−z01dz=2πi∫C(z−z0)n1dz=0,n∈Z,n>1
For n≥2, (n−1)(z−z0)n−1−1 is a single-valued antiderivative on C∖{z0}. Every closed integral is 0.
For n=1 the antiderivative is ln(z−z0). Its argument increases by 2π per loop around z0. This jump gives 2πi.
When to Use DIT
Consider deformation invariance when all of these hold:
C is a closed contour
The integrand is analytic on and inside C except at finitely many points
Those exceptional points lie inside C
Singularities Inside and Outside
Split the integrand by partial fractions into terms with 1 singular point each.
Term
Singular point a
Value of ∫C
z−ac
outside C
0
z−ac
inside C
2πic
(z−a)nc,n≥2
inside C
0
A singular point on C is not covered and needs separate treatment.
Singularities outside C never obstruct the deformation, since shrinking C only sweeps its interior.
∫Cf(z)dz=2πi∑ca
Here:
ca: coefficient of z−a1, summed over the singular points a inside C
Logarithm Shortcut
∫Cg(z)g′(z)dz cannot be evaluated directly as lng at the endpoints.
The shortcut ∫CF′dz=F(end)−F(start) needs F single-valued and analytic on a domain containing C. Here lng=ln∣g∣+iargg, and argg is multivalued when g has zeros inside C.
A valid but laborious version uses the change in argument. The ln∣g∣ part cancels on a closed curve.
∫Cg(z)g′(z)dz=iΔCargg
Partial fractions with deformation is the preferred route.
Examples
Line Integral of z from 0 to 1+i
Let C be the line segment from 0 to 1+i, parametrized by z(t)=t(1+i), 0≤t≤1. z′(t)=1+i.
∫Czdz=∫01t(1+i)(1+i)dt=(1+i)2∫01tdt
(1+i)2=2i and ∫01tdt=21
∫Czdz=2i⋅21=i
ML Bound on a Large Circle
Bound ∫C(z2+9)2z2dz for C:∣z∣=R>3.
L=2πR. On C, ∣z2∣=R2 and ∣z2+9∣≥R2−9>0.
∣f(z)∣≤(R2−9)2R2=M∫C(z2+9)2z2dz≤(R2−9)22πR3
The bound behaves like R2π, so it tends to 0 as R→∞.
Integral Using Partial Fractions
Show ∫Cz2+22zdz=4πi for C:∣z∣=2, positively oriented.
z2+2=(z−i2)(z+i2) and ∣±i2∣=2<2, so both poles lie inside C.
z2+22z=z−i21+z+i21
Each term integrates to 2πi by the key integral.
∫Cz2+22zdz=2πi+2πi=4πi
Using the argument change instead, g(2eiθ)=4e2iθ(1+21e−2iθ). The first factor changes its argument by 4π and the second by 0, so the integral is i⋅4π=4πi.