Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Integral Help exp (-x)/x with 0 to t limits

  1. Jun 17, 2004 #1
    I am trying to get this integral but no clues about how to proceed
    [tex] \int exp(-x)/x dx [/tex]

    with limits from 0 to t ( any real number, non infinity)

    I know of the limits are from t to infinity, there is a standard integral known as Ei(x). please help/share ideas in solving this!

    Regards and Thanks

  2. jcsd
  3. Jun 17, 2004 #2
    Mathematica doesn't like it, it tosses Ei[-x] back at me, whatever that is. Not sure if there's an analytical solution.
  4. Jun 17, 2004 #3


    User Avatar
    Science Advisor
    Homework Helper

    I'm not sure it is valid at x=0. If we looks at the infinite series:

    [tex]\frac{e^{-x}}{x} = \frac{1}{x} - \sum_{n=0}^{\infty} (-1)^n\frac{x^n}{(n+1)!}[/tex]


    [tex]\int \frac{e^{-x}}{x} dx = \ln x - \sum_{n=0}^{\infty} (-1)^{n}\frac{x^{n+1}}{(n+1)!*(n+1)} + C[/tex]

    This is clearly not valid for x = 0, which kind of makes sense as [itex]e^{-x}/x[/itex] is not valid at x = 0.
  5. Jun 17, 2004 #4


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Since 1/x is non-integrable in any interval that includes 0 (and exp(0)=1) since the integral diverges, so would be (1/x)*exp(-x).

    ie. the integral diverges to infinity.
  6. Jun 17, 2004 #5
    Thanks guys: I have a few questions on this:

    1. If I integrate within limits say 0.001 to t, is it going to introduce a lot of error in subsequent calculations that i carry out?

    2. Actually I am trying to obtain this integral

    [tex] \int f(r,x)dx [/tex] with limits 0 to t
    where f(r,x) is the Laplacian inverse of F(r, s) where F(r,s) is

    [tex] \frac{K_{o}(r\sqrt{s})}{\sqrt{s}K_{1}(\sqrt{s})} [/tex]

    Ko and K1 are modified Bessel functions of zero and first order
    Some simplifications resulted in the integral that I had put initially..

    Can anyone please suggest a way out?
  7. Oct 15, 2011 #6
    If we know t>0 then what is the best way to integrate this problem, without computer help?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook