Computing $-\hbar^2 \kappa^2 \int_{-\infty}^{\infty} \exp(-\kappa^2 |x|) \frac{d^2}{dx^2} \exp(-\kappa^2|x|)dx$

44 Views Asked by At

I am trying to compute $$-\hbar^2 \kappa^2 \int_{-\infty}^{\infty} \exp(-\kappa^2 |x|) \frac{d^2}{dx^2} \exp(-\kappa^2|x|)dx$$

where $\kappa = \frac{\sqrt{m \alpha}}{\hbar}$. Thus far, I have defined $f(x) = \exp(-\kappa^2 |x|)$ and have asserted

$$ f(x) = \begin{cases} \exp(-\kappa^2 x) & : x > 0 \\ \exp(\kappa^2 x) & : x < 0 \end{cases} $$

to show

$$ \dot{f}(x)\begin{cases} -\kappa^2 \exp(-\kappa^2 x) & : x > 0 \\ \kappa^2 \exp(\kappa^2 x) & : x < 0 \end{cases}= \kappa^2 f(x) (-\text{sgn}(x)) $$

where I have been intentionally sloppy about the behavior of the system at $0$. (I am currently reading a physics textbook and the theory of distributions was never introduced in a rigorous context.) We have previously defined the step function as

$$ \theta(x) = \begin{cases} 1 & : x > 0 \\ 0 & : x < 0 \end{cases} $$

where (as the author stated) $\theta(0)$ is irrelevant. Furthermore, I have already proven that $\frac{d\theta}{dx} = \delta(x)$. In terms of this function,

$$ \dot{f}(x) = \kappa^2 f(x) (\theta(-x) - \theta(x)) $$

Thus, we integrate by parts (once again sloppily assuming that convergence is guaranteed and that the boundary conditions align).

$$ \begin{align} -\hbar^2\kappa^2 \int_{-\infty}^{\infty} f(x) \frac{d^2 f}{dx^2} dx &= -\hbar^2\kappa^2 \left[f(x) \frac{df}{dx}\right]_{-\infty}^{\infty} \\ & - \hbar^2 \kappa^4 \left[\int_{-\infty}^{\infty} f(x) \frac{df}{dx} (\theta(-x) - \theta(x)) - f(x)^2 (\dot{\theta}(-x) + \dot{\theta}(x))dx\right] \\ & = - \hbar^2 \kappa^4 \left[\int_{-\infty}^{\infty} \kappa^2 f(x)^2 - f(x)^2 (\dot{\theta}(-x) + \dot{\theta}(x))dx\right] \\ & = - \hbar^2 \kappa^4 \left[\int_{-\infty}^{\infty} \kappa^2 f(x)^2 dx - 2 f(0)^2 \right] \\ & = - \hbar^2 \kappa^4 \left[\int_{-\infty}^{\infty} \kappa^2 f(2x) dx - 2 \right] \\ & = 2 \left(\frac{m \alpha}{\hbar}\right)^2 - \hbar^2 \kappa^6 \int_{-\infty}^{\infty} f(2x) dx \\ & = 2 \left(\frac{m \alpha}{\hbar}\right)^2 - 2 \hbar^2 \kappa^6 \int_{0}^{\infty} \exp(-2\kappa^2 x) dx \\ & = 2 \left(\frac{m \alpha}{\hbar}\right)^2 + \hbar^2 \kappa^4 \int_{0}^{\infty} \exp(-2\kappa^2 x) dx \\ & = 2 \left(\frac{m \alpha}{\hbar}\right)^2 - \hbar^2 \kappa^4\\ & = \left(\frac{m \alpha}{\hbar}\right)^2 \end{align} $$

The book tells me that the correct answer is $\left(\frac{m \alpha}{\hbar}\right)^2$, so I assume my general approach is correct. How can I make the use of distributions more rigorous? How can I try to quantify asymptotic growth rates for $f(x)$? I even tried setting up a recursion relationship by introducing

$$I_{\alpha} = \int_{-\infty}^{\infty} f_{\alpha}(x) \frac{d^2 f_{\alpha}}{dx^2} dx$$

where $f_{\alpha} (x) = \exp(-\alpha\kappa^2 |x|)$ and then trying to use a generating function to find series from

$$I_{2^n} = 2^n k^4 I_{2^{n+1}} + 2 \kappa^2$$

However, it did not appear to be particularly helpful. I feel that the book heavily suggested differentiating the Integrand, so I decided that I would either use Feynman's trick or integration by parts. The latter felt more natural given the second derivative in the middle of the integrand.

Edit: I should clarify that I am asking my question on Math SE (as opposed to Physics SE) because I deeply dissatisfied with the lack of mathematical explanation in the derivation. Does anyone have advice for making the use of distributions rigorous. Every time I say sloppily, is there a theorem or result I should quote? I have not yet had a formal course in the theory of distributions although I had to learn a little bit when I studied Green's functions in PDEs.

Are there any combinatorial approaches for describing distributions? I may be overzealous to apply generating functions, but I had hoped that they would provide bounds on more general wavefunctions that do not arise from simple potentials such as the delta function.

1

There are 1 best solutions below

5
On

You are making the situation far too complicated. We don't neet the theory of distributions. The integrand is an even function and $\kappa^2 >0.$ Let$c=\kappa^2.$ The integral is $$2\int_0^\infty e^{-cx}c^2e^{-cx}dx=2c^2\int_0^\infty e^{-2cx}dx$$ $$=\frac{2c^2}{2c}=c$$