# Hard asymptotic analysis of an integral [closed]

Can anyone help me in taking $x\rightarrow \infty$ for $$\exp\left(i{\int^{n}_{n_{0}}p\sqrt{1+\frac{x^2}{p^2}}dx}\right)$$

I don't really know how to proceed because it is an integral? It won't let me post pictures but here it is ( the symbols are different to what I used)

$$\omega^2=k^2 + \phi'^2 (n-n_{0})^2$$ $$p\equiv \frac{k}{\sqrt{\phi' }}$$ $$\tau = \sqrt{\phi'}(n-n_{0})$$

and

$$e^{i\int^{n}_{n_0}\omega dn} \rightarrow\left(\frac{2\tau}{p}\right)^{ip^{2}/2}e^{i\tau^{2}/2}e^{ip^{2}/4}$$

as $\tau \rightarrow \infty$.

-

## closed as unclear what you're asking by Jonas Meyer, Newb, Johanna, Adam Hughes, Jack D'AurizioMar 6 at 4:22

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

Doesn't seem right. $x$ is a dummy variable, it cannot approach anywhere except integration limits. Did you mean $n \rightarrow \infty$ instead? –  Kaster Mar 9 '13 at 10:02
no, mmm, I will post the paper up. –  anon Mar 9 '13 at 10:08
Hard to say without the information about $\omega dn$, but even know you might notice that $\tau \ne n$, so you' re taking wrong limit. –  Kaster Mar 9 '13 at 10:22
yeah the limit would be x or $\tau$ –  anon Mar 9 '13 at 10:36
So, as I said, you have to either say $n \rightarrow \infty$, or change your integral as $\int_{n_0}^x \ldots dx$ –  Kaster Mar 9 '13 at 11:00

First of all, find the integral $$\int \sqrt{1+\frac {x^2}{p^2}}dx = \left | \text{do substitution t = \frac xp}\right | = p \int \sqrt{1+t^2}dt = \\ = \left | \text{do substitution t = \sinh q}\right | = p \int \cosh^2 q\,dq = \\ = \left | \text{use \cosh^2 q = \frac {1+\cosh 2q}2}\right | = p \int \left( \frac {1+\cosh 2q}2\right) dq = \frac p2 \left ( q+\frac {\sinh 2q}2\right) +C = \\ = \left | \text{use \sinh 2q = 2 \sinh q \cosh q}\right | = \frac p2 \left ( q + \sinh q \cosh q\right)+C$$ Now substitute everything back $$\frac p2 \left ( q + \sinh q \cosh q\right) = \frac p2 \left ( \text{arcsinh } t + t \sqrt{1+t^2}\right) = \frac p2 \left ( \text{arcsinh } \frac xp + \frac xp \sqrt{1+\frac {x^2}{p^2}}\right)$$ So final integral is $$\int \sqrt{1+\frac {x^2}{p^2}}dx = \frac p2 \left ( \text{arcsinh } \frac xp + \frac xp \sqrt{1+\frac {x^2}{p^2}}\right) + C = F(x)$$ Obviously $$\int_{n_0}^x \sqrt{1+\frac {x^2}{p^2}}dx = F(x)-F(n_0)$$ Due to the form of $F(x)$ one can prove that $\lim_{x \rightarrow \infty}F(x) = \infty$, so $F(x)-F(n_0) \approx F(x)$ when $x \rightarrow \infty$

Now find asymptotical expressions of each term in $F(x)$ $$\text{arcsinh } \frac xp = \ln \left ( \frac xp + \sqrt{1+\frac {x^2}{p^2}}\right) \approx \ln \left ( \frac xp + \frac xp\right) = \ln \frac {2x}p \\ \frac xp \sqrt{1+\frac {x^2}{p^2}} = \frac xp \frac xp \sqrt{1+\frac {p^2}{x^2}} \approx \frac {x^2}{p^2}\left ( 1 + \frac {p^2}{2x^2} \right) = \frac {x^2}{p^2} + \frac 12$$ And finally $$\exp \left (ip \int_{n_0}^x \sqrt{1+\frac {x^2}{p^2}}dx\right) \approx \exp \left [ ip F(x)\right] \approx \exp \left [ \frac {ip^2}2 \left ( \ln \frac {2x}p + \frac {x^2}{p^2} + \frac 12 \right)\right ] = \\ = \exp \left (\frac {ip^2}2 \ln \frac {2x}p \right) \cdot \exp \left ( \frac {ix^2}2\right) \cdot \exp \left ( \frac {ip^2}4 \right) = \left ( \frac {2x}p\right)^{ip^2/2} \cdot e^{ix^2/2} \cdot e^{ip^2/4}$$

-
The $p^2$ term comes from $$\def\f{\frac{x^2}{p^2}}\frac xp\sqrt{1+\f}=\f\sqrt{1+\frac{p^2}{x^2}}=\f+\frac12+O\left(x^{-2}\right)\;.$$ –  joriki Mar 9 '13 at 12:16
Thanks @joriki, I've fixed it. –  Kaster Mar 9 '13 at 22:07
@user21119, if that's what you wanted eventually, please, do not hesitate to mark you question as solved. –  Kaster Mar 9 '13 at 22:11
@Kaster: see the comments on my answer. Are you okay with me temporarily deleting your answer until the due date? You will get your points back at that time. –  robjohn Apr 21 '13 at 20:29
@robjohn Absolutely. My intent was to help, not to ruin something. –  Kaster Apr 22 '13 at 1:17

Since $x$ is a dummy variable, it makes no sense to say $x\to\infty$. Perhaps you meant $n\to\infty$.

Using the substitution $x=\tan(\theta)$ gives $$\int\sqrt{1+x^2}\,\mathrm{d}x=\frac12\left(x\sqrt{1+x^2}+\log\left(x+\sqrt{1+x^2}\right)\right)+C$$ Thus, we have, using the substitutions described in the question \begin{align} \int_{n_0}^n\omega\,\mathrm{d}n &=\int_0^\tau\sqrt{k^2+\frac{k^2x^2}{p^2}}\frac{p\,\mathrm{d}x}{k}\\ &=\int_0^\tau\sqrt{p^2+x^2}\,\mathrm{d}x\\ &=p^2\int_0^{\tau/p}\sqrt{1+x^2}\,\mathrm{d}x\\ &=\frac{p^2}{2}\left(\frac{\tau}{p}\sqrt{1+\frac{\tau^2}{p^2}}+\log\left(\frac{\tau}{p}+\sqrt{1+\frac{\tau^2}{p^2}}\right)\right)\\ &=\frac{p^2}{2}\left(\frac{\tau^2}{p^2}\sqrt{1+\frac{p^2}{\tau^2}}+\log\left(\frac{2\tau}{p}\right)+\log\left(\frac{1+\sqrt{1+\frac{p^2}{\tau^2}}}{2}\right)\right)\\ &=\frac{p^2}{2}\left(\frac{\tau^2}{p^2}\left(1+\frac{p^2}{2\tau^2}+O\left(\frac1{\tau^4}\right)\right)+\log\left(\frac{2\tau}{p}\right)+O\left(\frac1{\tau^2}\right)\right)\\ &=\frac{p^2}{2}\left(\frac{\tau^2}{p^2}+\frac12+\log\left(\frac{2\tau}{p}\right)+O\left(\frac1{\tau^2}\right)\right)\\ &=\frac{\tau^2}{2}+\frac{p^2}{4}+\frac{p^2}{2}\log\left(\frac{2\tau}{p}\right)+O\left(\frac1{\tau^2}\right) \end{align} Therefore, $$e^{i\int_{n_0}^n\omega\,\mathrm{d}n} \sim\left(\frac{2\tau}{p}\right)^{ip^2/2}e^{i\tau^2/2}\,e^{ip^2/4}$$

-