Master contour integrals, Cauchy's theorem, Cauchy's integral formula, and their applications.
đ Integrationđ Lesson 93 of 100đ Free Course
Advertisement
Why It Matters
âšī¸ Why It Matters
Contour integration is the computational heart of complex analysis. It transforms the problem of evaluating difficult or impossible real integrals into a routine procedure: find singularities, compute residues, sum them up. Cauchy's integral theorem reveals that for analytic functions, the integral depends only on the endpoints (not the path), while Cauchy's integral formula reconstructs a function's value at any interior point purely from its boundary values â a remarkable result with no real-variable analog. These tools evaluate integrals like âĢââââx2+1cosxâdx in minutes, solve boundary value problems in physics, and form the basis of the residue theorem that underpins all of advanced complex analysis. Without contour integration, Fourier analysis, quantum field theory, and much of mathematical physics would lack their most elegant computational methods.
Core Definitions
DfContour (Path)
A contour Îŗ is a piecewise smooth curve in C, parameterized as z(t)=x(t)+iy(t) for aâ¤tâ¤b. The curve is smooth if zâ˛(t) is continuous and zâ˛(t)î =0. A contour is closed if z(a)=z(b), and simple (non-self-intersecting) if it crosses itself at most at the endpoints.
DfContour Integral
The integral of f along a contour Îŗ:z(t), aâ¤tâ¤b, is:
âĢÎŗâf(z)dz=âĢabâf(z(t))zâ˛(t)dt
This reduces complex line integration to a real definite integral.
DfSimply Connected Domain
A domain D is simply connected if every closed curve in D can be continuously deformed to a point without leaving D. Equivalently, D has no "holes." The complement of a disk CâD(a,r)â is not simply connected; the interior of a circle is.
DfWinding Number (Index)
The winding number of a closed contour Îŗ around a point z0â not on Îŗ is:
IndÎŗâ(z0â)=2Īi1ââŽÎŗâzâz0âdzâ
It counts how many times Îŗ winds around z0â (positive = counterclockwise).
DfRemovable Singularity
A singularity z0â of f is removable if limzâz0ââ(zâz0â)f(z)=0. In this case, f can be redefined at z0â to be analytic.
DfSimple Pole
A singularity z0â of f is a simple pole if limzâz0ââ(zâz0â)f(z)=cî =0. The residue is this limit: Res(f,z0â)=c.
If f is analytic inside and on a simple closed contour C, and the domain interior to C is simply connected, then:
âŽCâf(z)dz=0
Proof Sketch: By Green's theorem, âŽCâudxâvdy+iâŽCâvdx+udy. The Cauchy-Riemann equations make both integrands have zero curl, so both contour integrals vanish. âĄ
General version (Goursat): The theorem holds even without assuming fⲠis continuous â analyticity alone suffices.
ThCauchy's Integral Formula
If f is analytic inside and on a simple closed contour C, and z0â is interior to C, then:
f(z0â)=2Īi1ââŽCâzâz0âf(z)âdz
Proof Sketch: Define g(z)=zâz0âf(z)âf(z0â)â for zî =z0â and g(z0â)=fâ˛(z0â). Then g is analytic inside C. By Cauchy's theorem, âŽCâg(z)dz=0. Expanding and rearranging yields the formula. âĄ
Geometric meaning: The value of an analytic function at an interior point is completely determined by its values on the boundary.
ThResidue Theorem
If f is analytic inside and on a simple closed contour C except at isolated singularities z1â,z2â,âĻ,znâ inside C, then:
âŽCâf(z)dz=2Īik=1ânâRes(f,zkâ)
Proof Sketch: For each singularity zkâ, draw a small circle Îŗkâ around it. By Cauchy's theorem applied to the multiply connected region (deforming the contour), the integral over C equals the sum of integrals over the small circles. Each integral over Îŗkâ equals 2ĪiRes(f,zkâ) by the Laurent series expansion. âĄ
ThDeformation of Contour
If f is analytic in a region between two simple closed contours C1â and C2â (with C2â inside C1â), and f has no singularities between them, then:
âŽC1ââf(z)dz=âŽC2ââf(z)dz
provided both contours are oriented the same way (counterclockwise).
Significance: You can deform contours to pass around singularities conveniently, as long as you don't cross any singularities.
ThEstimation Lemma (ML Inequality)
If âŖf(z)âŖâ¤M on a contour Îŗ of length L, then:
ââĢÎŗâf(z)dzââ¤ML
Application: This is used to show that integrals over large arcs vanish as the arc radius Rââ, which is essential for improper real integrals.
ThJordan's Lemma
If g(z) is analytic in the upper half-plane and âŖg(z)âŖâ0 uniformly as âŖzâŖââ in the upper half-plane, then for a>0:
RââlimââĢCRââg(z)eiazdz=0
where CRâ is the upper semicircle of radius R.
Application: Justifies discarding the semicircular arc when evaluating Fourier-type integrals âĢââââf(x)eiaxdx.
Worked Examples
đExample 1: Direct Parameterization
Evaluate âŽCâzdzâ where C is the unit circle âŖzâŖ=1, oriented counterclockwise.
Verification: Since 1/(z2+1) is analytic at z=â (behaves like 1/z2), and the sum of all residues including the residue at infinity is zero, the residues at Âąi cancel. â
Practice Problems
đProblem 1: Evaluate a Contour Integral
Evaluate âŽâŖzâŖ=2âz(zâ1)z+1âdz.
đĄSolution
Step 1: Singularities at z=0 (inside âŖzâŖ=2) and z=1 (inside âŖzâŖ=2). Both are simple poles.
Step 2: Residue at z=0:
Res(f,0)=limzâ0âzâ1z+1â=â11â=â1
Step 3: Residue at z=1:
Res(f,1)=limzâ1âzz+1â=12â=2
Step 4:âŽ=2Īi(â1+2)=2Īi
đProblem 2: Cauchy Formula for Derivatives
Evaluate âŽâŖzâŖ=3â(zâĪ/2)2coszâdz.
đĄSolution
Step 1:f(z)=cosz is entire, z0â=Ī/2 is inside âŖzâŖ=3, n=1.
Step 3: Sum of residues: 21ââ4+29â=21â8+9â=22â=1
Step 4:âŽ=2Īiâ 1=2Īi
Common Mistakes
Mistake
Correction
Example
Forgetting the 2Īi factor in the residue theorem
The residue theorem states âŽ=2ĪiâRes, not just âRes
âŽzdzâ=2Īi, not 1
Using the wrong orientation
Counterclockwise is positive; clockwise gives a minus sign
âŽclockwiseâ=â2ĪiâRes
Including singularities outside the contour
Only sum residues of singularities inside C
If z0ââ/int(C), don't include it
Wrong residue formula for higher-order poles
For order n: (nâ1)!1âlimdznâ1dnâ1â[(zâz0â)nf], not just the limit
Need the derivative for nâĨ2
Forgetting the ML inequality
Large arcs vanish only if âŖfâŖâ Lâ0; must verify the arc contribution vanishes
Check âŖf(z)âŖ=O(1/Rp) with p>0
Incorrect parameterization
dz=zâ˛(t)dt, not just dz=dt
For z=Reiθ: dz=iReiθdθ
Misidentifying pole order
Factor the denominator carefully; check the limit
(zâ1)2(z+1)1â has a pole of order 2 at z=1
Interview / Exam Questions
Q1: State Cauchy's integral theorem. What are its hypotheses, and why is the simply connected condition important?
A1: Cauchy's theorem states: if f is analytic inside and on a simple closed contour C, and the interior of C is simply connected, then âŽCâf(z)dz=0. The simply connected condition ensures there are no singularities "hidden" inside the contour. For example, âŽâŖzâŖ=1âzdzâ=2Īiî =0 because 1/z has a pole at z=0 inside the contour â the domain Câ{0} is not simply connected. The theorem holds because the Cauchy-Riemann equations make the integrand's "curl" vanish, and Green's theorem converts the contour integral to a double integral of this zero curl.
Q2: Why does Cauchy's integral formula imply that analytic functions are infinitely differentiable?
A2: Cauchy's formula f(z0â)=2Īi1ââŽCâzâz0âf(z)âdz expresses f as an integral of an analytic integrand. Differentiating under the integral sign (justified by analyticity) gives fâ˛(z0â)=2Īi1ââŽCâ(zâz0â)2f(z)âdz, and repeated differentiation yields the generalized formula for f(n). Since this works for all n, f is infinitely differentiable. In real analysis, a function can be differentiable once but not twice (e.g., f(x)=x3/2). Complex differentiability is far more restrictive.
Q3: How do you compute the residue of f(z)=Q(z)P(z)â at a simple pole z0â?
A3: If z0â is a simple zero of Q(z) (i.e., Q(z0â)=0 but Qâ˛(z0â)î =0), and P(z0â)î =0, then:
Res(f,z0â)=Qâ˛(z0â)P(z0â)â
This comes from f(z)=Qâ˛(z0â)(zâz0â)P(z)â+analytic terms near z0â, so the residue (coefficient of 1/(zâz0â)) is P(z0â)/Qâ˛(z0â).
Q4: Explain the deformation of contour principle and give an example.
A4: If f is analytic in a region between two simple closed contours C1â (outer) and C2â (inner), and f has no singularities between them, then âŽC1ââfdz=âŽC2ââfdz (both counterclockwise). This is because the region between C1â and C2â is multiply connected, but Cauchy's theorem applied to this annular region shows the two contour integrals are equal.
Example: To compute âŽâŖzâŖ=3âz(zâ1)dzâ, we can deform to two small circles: one around z=0 and one around z=1, each giving 2Īi times the local residue.
Q5: When is the residue theorem more useful than direct parameterization?
A5: The residue theorem is superior when: (1) The contour encloses singularities â you only need the residues, not the full parameterization. (2) The integral has multiple poles â the theorem converts integration to algebra (summing residues). (3) You need to evaluate integrals over large contours (as Rââ) â residues capture the essential behavior while the contour contribution vanishes. Direct parameterization is simpler for elementary functions with no singularities inside the contour (where the integral is zero by Cauchy's theorem) or for very simple cases like âŽzdzâ.
Q6: What is the ML inequality, and when do you use it?
A6: The ML inequality states: if âŖf(z)âŖâ¤M on a contour Îŗ of length L, then ââĢÎŗâf(z)dzââ¤ML. It's used to bound contour integrals and show that contributions from arcs vanish. For example, to show âĢCRââz2+1eizâdzâ0 as Rââ (where CRâ is a semicircle): on CRâ, âŖeizâŖ=eâRsinθâ¤1 and âŖz2+1âŖâĨR2â1, so âŖfâŖâ¤R2â11â and L=ĪR. Thus ââĢCRââââ¤R2â1ĪRââ0.