114 lines
5.7 KiB
TeX
114 lines
5.7 KiB
TeX
\chapter{Background}
|
|
\label{chap:Background}
|
|
|
|
\section{Partial-wave decomposition}
|
|
\label{sec:partial_wave}
|
|
|
|
For spherical symmetric two-body potentials $V(r)$, where r is the distance between the particles, the Hamiltonian has an $SO(3)$ symmetry. Consequently, the time-independent Schr\"odinger equation,
|
|
\begin{equation}
|
|
\left[{-}\frac{1}{2\mu}\nabla^2 + V(r)\right]\psi(\vec{x}) = E \, \psi(\vec{x}) \, ,
|
|
\label{eq:SE}
|
|
\end{equation}
|
|
can be separated into radial and angular parts by writing out the Laplacian operator in spherical coordinates,
|
|
\begin{equation}
|
|
\nabla^2 \equiv \frac{1}{r^2} \frac{\partial}{\partial r} \left( r^2 \frac{\partial}{\partial r} \right) + \frac{1}{r^2} \left[ \frac{1}{\sin \theta} \frac{\partial}{\partial \theta} \left( \sin \theta \frac{\partial}{\partial \theta} \right) + \frac{1}{\sin^2 \theta} \frac{\partial^2}{\partial \phi^2} \right] \, ,
|
|
\end{equation}
|
|
which gives the angular eigen-equation,
|
|
\begin{equation}
|
|
{-}\left[ \frac{1}{\sin \theta} \frac{\partial}{\partial \theta} \left( \sin \theta \frac{\partial}{\partial \theta} \right) + \frac{1}{\sin^2 \theta} \frac{\partial^2}{\partial \phi^2} \right] Y^m_l(\theta,\phi) = l(l+1) \, Y^m_l(\theta,\phi) \, ,
|
|
\end{equation}
|
|
having the well-known spherical harmonics $Y^m_l$ as eigenfunctions and $l(l+1)$ as corresponding eigenvalues. Substituting this back into Eq.~\eqref{eq:SE} gives the radial Schr\"odinger equation,
|
|
\begin{equation}
|
|
\left[ \frac{\dd ^2}{\dd r^2} - \frac{l(l+1)}{r^2} + 2 \mu V(r) - p^2 \right] u(r) = 0 \, ,
|
|
\label{eq:SE_rad}
|
|
\end{equation}
|
|
where $p=\sqrt{2 \mu E}$ and $u(r)$ is the \emph{reduced} radial wave function implicitly defined via,
|
|
\begin{equation}
|
|
\psi(r,\theta,\phi) = R(r) Y^m_l(\theta,\phi) = \frac{u(r)}{r} Y^m_l(\theta,\phi) \, .
|
|
\end{equation}
|
|
|
|
\subsection{Free solutions}
|
|
|
|
In a region of zero-interaction, that is $V(r)=0$, Eq.~\eqref{eq:SE_rad} has solutions that are given by some linear combination of Ricatti-Hankel functions,
|
|
\begin{equation}
|
|
u(r) \sim a \, \hat{h}^-_l(pr) + b \, \hat{h}^+_l(pr) \, ,
|
|
\label{eq:free_wf}
|
|
\end{equation}
|
|
where $\hat{h}^+_l(pr)$ and $\hat{h}^-_l(pr)$ are the Ricatti-Hankel functions of the first and second kind respectively. These Ricatti-Hankel functions can be defined using the familiar spherical Bessel functions,
|
|
\begin{subalign}
|
|
\hat{h}^+_l(z) &= \ii z \, h^{(1)}_l(z) \\
|
|
\hat{h}^-_l(z) &= {-}\ii z \, h^{(2)}_l(z)
|
|
\end{subalign}
|
|
|
|
When $V(r)=0$ everywhere, we can fix the coefficients $a$ and $b$ to satify the boundary condition $u(r) \to 0$ for $r \to 0$. This gives the ``free solutions'' of the the radial Schr\"odinger equation,
|
|
\begin{equation}
|
|
u(r) = \hat{j}_l(pr) = \frac{i}{2} \left[\hat{h}^-_l(pr) - \hat{h}^+_l(pr)\right] \, .
|
|
\end{equation}
|
|
$\hat{j}_l(pr)$ are called the Ricatti-Bessel functions and they form a complete basis for any given $l$.
|
|
|
|
\subsection{The radial basis}
|
|
|
|
Next, we define the partial-wave ``radius'' state $\ket{r,l,m}$ such that,
|
|
\begin{equation}
|
|
\braket{r,l,m | \psi_{l,m}} = u(r) \, ,
|
|
\end{equation}
|
|
where $\psi_{l,m}$ is a state with definite angular momentum quantum numbers $(l,m)$ and $u(r)$ is the corresponding reduced radial wave function.
|
|
This leads to the relation,
|
|
\begin{equation}
|
|
\braket{\vec{x} | r,l,m} = \frac{\delta (x - r)}{r} Y_l^m(\hat{x}) \, .
|
|
\end{equation}
|
|
|
|
Now, consider a free solution $\ket{E,l,m}$ with energy $E=p^2/\mu$. Its overlap with the radius state $\ket{r,l',m'}$ is,
|
|
\begin{spliteq}
|
|
\braket{r,l',m' | E,l,m} &= \int d^3 x \, \braket{r,l',m' | \vec{x}} \braket{\vec{x} | E,l,m} \\
|
|
&= \iiint x^2 \, dx \, d(\cos{\theta}) \, d\phi \, \frac{\delta (x - r)}{r} Y_{l'}^{m'}(\theta,\phi) \left[\ii^l \sqrt{\frac{2\mu}{\pi p}} \frac{1}{r} \hat{j}_l(pr) Y_l^m(\theta,\phi) \right] \\
|
|
&= \ii^l \sqrt{\frac{2\mu}{\pi p}} \hat{j}_l(pr) \, \delta_{l,l'} \, \delta_{m,m'}
|
|
\end{spliteq}
|
|
where the relation
|
|
\begin{equation}
|
|
\braket{\vec{x} | E,l,m} = \ii^l \sqrt{\frac{2\mu}{\pi p}} \frac{1}{r} \hat{j}_l(pr) Y_l^m(\theta,\phi)
|
|
\end{equation}
|
|
is used~\citep[Eq. 11.8]{taylor}.
|
|
|
|
\section{Hankel transform}
|
|
|
|
We can transform between the ``partial-wave momentum space'' and the usual partial-wave configuration space using the Hankel transforms. We adopt the ``symmetric'' convention given by,
|
|
\begin{subalign}
|
|
u(k) &= \sqrt{\frac{2}{\pi}} \int_0^\infty dr \, \hat{j}_l(kr) u(r) \, ,\\
|
|
u(r) &= \sqrt{\frac{2}{\pi}} \int_0^\infty dk \, \hat{j}_l(kr) u(k) \, . \label{eq:inv_hankel}
|
|
\end{subalign}
|
|
|
|
To verify the constant pre-factors, we show that the normalization is preserved under this transform. We know that in configuration space, the radial part of the wave function obeys the normalization condition,
|
|
|
|
\begin{equation} \begin{split}
|
|
1 &= \int_0^\infty \, dr \, r^2 \, R^*(r) \, R(r) \\
|
|
&= \int_0^\infty \, dr \, u^*(r) \, u(r)
|
|
\end{split} \end{equation}
|
|
|
|
And in momentum space, it obeys,
|
|
|
|
\begin{equation}
|
|
1 = \int_0^\infty dk \, u^*(k) \, u(k)
|
|
\end{equation}
|
|
|
|
Now, we show that we end up with a properly normalized $u(r)$ in configuration space,
|
|
given we start with a normalized $\Tilde{u}(k)$ in momentum space.
|
|
|
|
\begin{equation} \begin{split}
|
|
\int_0^\infty \, dr \, u^*(r) u(r)
|
|
&= \frac{2}{\pi} \int_0^\infty \int_0^\infty \int_0^\infty
|
|
dr \, dk \, dk' \,
|
|
\hat{j}_l(kr) \, \hat{j}_l(k'r) \, u^*(k) \, u(k') \\
|
|
&= \int_0^\infty \int_0^\infty
|
|
dk \, dk' \,
|
|
\delta(k-k') \, u^*(k) \, u(k') \\
|
|
&= \int_0^\infty
|
|
dk \, u^*(k) \, u(k) \\
|
|
&= 1 \\
|
|
\end{split} \end{equation}
|
|
|
|
where we have used Eq.~\eqref{eq:inv_hankel} in the first step
|
|
and evaluated the integral w.r.t. $r$ using \cite[p183]{taylor}
|
|
$$\int_0^\infty dr \, \hat{j}_l(kr) \, \hat{j}_l(k'r) = \frac{\pi}{2} \delta(k-k')$$
|
|
in the second step.
|