Hermiticity of operator on square integrable functions

184 Views Asked by At

I'm trying to solve this problem for my Quantum Mechanics class but its my first time dealing with infinite dimensional spaces so I'm a little lost here. I need to show that the operator defined by $$ i\frac{d}{dx}\delta(x-x_0) $$ is not Hermitian in the $L^2(\mathbb{R})$ space.

I tried calculating the integrals $$ \int\limits^{\infty}_{-\infty}\phi^*(x)i\frac{d}{dx}\phi(x)dx\\ \int\limits^{\infty}_{-\infty}\left(\phi(x)i\frac{d}{dx}\right)^*\phi(x)dx $$ but i didnt get anywhere, maybe i made a mistake but I'm not sure.

Any help very welcome!

1

There are 1 best solutions below

2
On BEST ANSWER

The Hermitian adjoin $S^\dagger$ of an operator $S$ satisfies $\langle u,Sv\rangle=\langle S^\dagger u,v\rangle$.

With $S\phi:=i\frac{\mathrm{d}}{\mathrm{d}x}\big(\delta(x-x_0)\phi(x)\big)$ we can work out

$$ \begin{array}{ll} \langle u,Sv\rangle & \displaystyle = \int_{\mathbb{R}}\overline{u(x)}~i\frac{\mathrm{d}}{\mathrm{d}x}\Big(\delta(x-x_0)v(x)\Big)~\mathrm{d}x \\[5pt] & = \displaystyle i\int_{\mathbb{R}} \overline{u(x)} \Big( \delta'(x-x_0)v(x)+\delta(x-x_0)v'(x) \Big)\,\mathrm{d}x \\[5pt] & = \displaystyle i\int_{\mathbb{R}} \delta'(x-x_0)\overline{u(x)}v(x)\,\mathrm{d}x+i\int_{\mathbb{R}}\delta(x-x_0)\overline{u(x)}v(x)\,\mathrm{d}x \\[5pt] & = \displaystyle -i\Big(\overline{u'(x_0)}v(x_0)+\overline{u(x_0)}v'(x_0)\Big)+i\overline{u(x_0)}v'(x_0) \\[5pt] & = -i\overline{u'(x_0)}v(x_0) \\[5pt] & = \displaystyle \int_{\mathbb{R}}\overline{i\delta(x-x_0)\frac{\mathrm{d}}{\mathrm{d}x} u(x)}\, v(x)\,\mathrm{d}x \end{array} $$

Therefore, $\Big(i\frac{\mathrm{d}}{\mathrm{d}x}\delta(x-x_0)\Big)^\dagger = i\delta(x-x_0)\frac{\mathrm{d}}{\mathrm{d}x}$ which is a different operator (as you can show by example); in other words $S^\dagger\ne S$ so $S$ is not Hermitian.

One could also use the facts $(iS)^\dagger=-iS^\dagger$ and $(AB)^\dagger=B^\dagger A^\dagger$ along with the fact $\delta(x-x_0)$ is Hermitian and $\frac{\mathrm{d}}{\mathrm{d}x}$ is anti-Hermitian (via by-parts integration and functions decaying at $\pm\infty$).