0
$\begingroup$

To explain my problem I start with two functions to be sine transformed. This question is a problem of current research in the field of electrolyte transport theory. The first Function is given by:

$$Y_1(r)=\frac{1}{r\cdot\biggl[\text{e}^{+A\cdot r}+B\cdot\text{e}^{-C\cdot r}\biggr]}$$

and

$$Y_2(r)=\frac{1}{r^2\cdot\biggl[\text{e}^{+A\cdot r}+B\cdot\text{e}^{-C\cdot r}\biggr]}$$

with constants $A,B,C\in\mathcal{R}^+$ beeing positive real numbers without 0. The symbol $r\in\mathcal{R}^+_0$ denotes the variable in coordinate space. Note that the important condition $C>A$ holds! The sine transform of the two functions are given by:

$$Z_i(k) = \int_{0}^{\infty}Y_i(r)\cdot\text{sin}(k\cdot r)~\text{d}r\quad\forall\quad i\in[1,2]$$

with $k\in\mathcal{R}^+_0$ beeing the variable in Fourier space. The two integrals should exist unter the condition $C>A$.

But I'm having trouble solving the integrals. Unfortunately, I haven't been able to find a solution using Maple either. The integral tables available to me didn't help me either. Can anyone solve these two integrals or know of literature sources with corresponding solutions? Can someone perhaps solve these two integrals with another CAS like Mathematica? If a solution is possible, I would be very happy if the solution could be roughly outlined. Thank you in advance.

$\endgroup$
1
  • $\begingroup$ the integral $Y_2$ is divergent; the integral $Y_1$ is convergent also for $C<A$, as long as $C,A$ are both positive. $\endgroup$ Commented Aug 26 at 12:49

2 Answers 2

2
$\begingroup$

Consider the whole thing a function of $A$. Differentiating this function boils the first integral down to $$ \frac A{2i}\int_0^\infty (e^{ix}-e^{-ix})\frac{e^{Ax}}{e^{Ax}+Be^{-Cx}}\ dx =\frac A{2i}\int_0^\infty (e^{ix}-e^{-ix})\frac{1}{1+Be^{(-A-C)x}}\ dx $$ Now for $B<1$ expand the second term in the integral as a geometric series and interchange the sum with the integral. The upcoming Integrals are elementary and after differentiating you get a geometric series again which reduces the whole thing to a rational function in $e^{(A+C)x}$. Hence the orgiginal function is an double integral of a very simple rational function which can be easily evaluated.

$\endgroup$
4
  • $\begingroup$ I am surprised: this procedure gives an answer that does not involve special functions? $\endgroup$ Commented Aug 27 at 8:01
  • $\begingroup$ It is not a complete answer for my question. But I have to admit that my question was too vague to get a complete answer. The reference to the integration method actually helped me, even if my question was not answered conclusively and completely. $\endgroup$ Commented Aug 27 at 12:29
  • 1
    $\begingroup$ OK, it's just that if you "accept" an answer, you imply that it is correct, which in this case would mean that the integral can be reduced to the integral of a rational function; I don't see how that can be reconciled with the exact expression in terms of special functions for $B=1$, do you? $\endgroup$ Commented Aug 27 at 12:52
  • 1
    $\begingroup$ No, I don't! I have retracted the correctness of the answer. Regardless of the fact that it also helped me further. But I see the need to provide more information about the context of the integrals. Otherwise it will be difficult for the community to provide a correct answer. I will expand my original question to the physical context. But I am afraid that the question should then be re-framed in a physics form, as it will then go beyond the purely mathematical context. $\endgroup$ Commented Aug 27 at 14:12
1
$\begingroup$

The integral $Y_2$ is divergent (it has a non-integrable pole at $r=0$).

Concerning $Y_1$, Mathematica gives a complicated closed-form answer for $B=1$, in terms of the generalised Stieltjes constant $\gamma_n(a)$, the Euler gamma constant $\gamma$, the Gamma function $\Gamma(s)$, and the zeta function $\zeta(s)$: $$\int_0^\infty\frac{\sin k r}{r (e^{Ar}+e^{-Cr})}\,dr$$ $$=\frac{1}{2(A+C)}\Re\biggl[(k+i A) \gamma _1\left(\frac{A-i k}{2 (A+C)}\right)-(k+i A) \gamma _1\left(X\right)+2 i (A+C) \biggl((\ln (A+C)+\gamma +\ln 2) \left[\zeta \left(0,X\right)-\zeta \left(0,Y\right)\right]+\ln \Gamma \left(X\right)+\ln \Gamma \left(Y\right)\biggr)\biggr],$$ $$X=\frac{3 A+2 C-i k}{2 (A+C)},\;\;Y=\frac{2 A+C-i k}{2 (A+C)}.$$

$\endgroup$
2
  • $\begingroup$ Your note about the divergent behaviour of Integral $Z_2$ is very interesting. The integral is part of the Onsager solution for point charges multiplied by a sigmoid function. Perhaps I should have provided more information about the context of this problem. However, the question for the integration procedure is answered for both integrals. Thank you for your help. $\endgroup$ Commented Aug 27 at 7:34
  • $\begingroup$ Thank your very much for your answer. Your answer has given me another valuable insight into the solution to this problem. But unfortunately I cannot restrict the integral to the case B=1. $\endgroup$ Commented Aug 27 at 12:31

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .