1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Fresnel Integrals, Contour Integration

  1. Feb 15, 2014 #1
    1. The problem statement, all variables and given/known data

    Please let me know if this kind of posting of exact problems from a textbook isn't allowed; if that's the case I'll delete it immediately.

    From Boas's Mathematical Methods in the Physical Sciences, Third Edition: The Fresnel integrals, [itex] \int_0^u sin (u^2)\,du [/itex] and [itex] \int_0^u cos (u^2)\,du [/itex], are important in optics. For the case of infinite upper limits, evaluate these integrals as follows: Make the change of variable [itex] x = u^2 [/itex]; to evaluate the resulting integrals, find [itex] \oint e^{iz}z^{-1/2}\,dz [/itex] around the contour shown. Let [itex] r \rightarrow 0 [/itex] and [itex] R \rightarrow \infty [/itex] and show that the integrals along these quarter-circles tend to zero. Recognize the integral along the y-axis as a gamma function and so evaluate it. Hence evaluate the integral along the x-axis; the real and imaginary parts of this integral are the integrals you are trying to find.

    (Note: I already solved the problem correctly, but I don't know how to prove the bold part above; see work below.)

    (The "contour shown" looks like a quarter circle in the first quadrant, except that there is a small quarter circle (of radius [itex]r[/itex]) cut out around the origin to avoid the singularity there.)

    2. Relevant equations

    [itex] \oint f(z)\,dz = 0 [/itex] when the function is analytic on and inside a simple closed curve.

    3. The attempt at a solution

    The contour integral is the sum of four integrals, and it is equal to zero since the singularity at the origin is outside the contour. I omit the integrals along the x and y-axes because I'm just wondering about the ones along the two semicircles:

    [itex] \int_0^{π/2} e^{iz}Rie^{iθ}/R^{1/2}e^{iθ/2}\,dθ [/itex]


    [itex] \int_{π/2}^{0} e^{iz}rie^{iθ}/r^{1/2}e^{iθ/2}\,dθ [/itex]

    These were obtained by substituting [itex] z=Re^{iθ} [/itex] and [itex] z=re^{iθ} [/itex], respectively, into the contour integral given in the problem statement.

    I can see that in the second integral, the [itex]r[/itex] in the numerator cancels with the [itex]r^{1/2}[/itex] in the denominator, and since [itex]r[/itex] (and therefore [itex]z[/itex]) are tending to zero, the integral tends to zero. Is this sufficient?

    The second integral tending to zero as [itex] R \rightarrow \infty [/itex] makes less sense. Doesn't the numerator tend to infinity, and the denominator stay a constant 1, as [itex] R \rightarrow \infty [/itex]? And substituting [itex] z = Re^{iθ} [/itex] into the exponential only makes things worse. I considered rewriting the exponentials using Euler's formula and examining their behavior, but I don't see how that would help.

  2. jcsd
  3. Feb 15, 2014 #2


    User Avatar
    Homework Helper

    The numerator goes to 0
    You forgot to change a z
    $$\int_0^{\pi/2} e^{i R e^{iθ}}Rie^{iθ}/(R^{1/2}e^{iθ/2})\,dθ$$
  4. Feb 16, 2014 #3
    How does that work when [itex] R \rightarrow \infty [/itex] ?
  5. Feb 16, 2014 #4


    User Avatar
    Homework Helper

    $$e^{i R e^{iθ}}\rightarrow 0$$
    very fast
    compare to e^-x
  6. Feb 16, 2014 #5


    User Avatar
    Science Advisor
    Homework Helper

    It's not that simple. You wind up needing to estimate ##\int_0^{\pi/4} e^{-R\sin\theta} d\theta##. It doesn't even really go to zero all that quickly. But it does go to zero.
    Last edited: Feb 16, 2014
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted