% vim: tw=50 % 26/01/2023 09AM \subsection{The electrostatic potential} For general $\rho(\bf{x})$ we cannot determine $\bf{E}(\bf{x})$ using Gauss's law alone. \myskip (M3') implies that $\bf{E}$ can be written in terms of an \emph{electrostatic} (or \emph{electric}) \emph{potential} $\Phi(\bf{x})$: \[ \boxed{\bf{E} = -\nabla \Phi} \] The \emph{potential difference} (or \emph{voltage}) between two points $\bf{x}_1$ and $\bf{x}_2$ is \begin{align*} \Phi(\bf{x}_2) -\Phi(\bf{x}_1) &= \int \dd \Phi \\ &= -\int_{\bf{x}_1}^{\bf{x}_2} \bf{E}(\bf{x}) \cdot \dd \bf{x} \end{align*} and is path-independent because $\nabla \times \bf{E} = \bf{0}$. \myskip The electric force on a particle of charge $q$, \[ \bf{F} = q\bf{E} = -q\nabla \Phi \] is a conservative force associated with the potential energy \[ \boxed{U(\bf{x}) = q\Phi(\bf{x})} \] (M1) implies that $\Phi$ satisfies \emph{Poisson's equation} \[ \boxed{-\nabla^2 \Phi = \frac{\rho}{\eps_0}} \] The solution can be written as an integral (over all space, assuming decay at infinity) \[ \Phi(\bf{x}) = \frac{1}{4\pi\eps_0} \int \frac{\rho(\bf{x}'}{|\bf{x} - \bf{x}'|} \dd^2 \bf{x}' \] This is the convolution of $\rho(\bf{x})$ with the potential of a unit point charge, $\frac{1}{4\pi\eps_0|\bf{x}|}$, which is the solution of \[ -\nabla^2 \Phi = \frac{\delta(\bf{x})}{\eps_0} \] satisfying $\Phi \to 0$ as $|\bf{x}| \to \infty$. \myskip $\bf{E}$ is unaffected if we add an arbitrary constant to $\Phi$. We usually choose this such that $\Phi \to 0$ as $|\bf{x}| \to \infty$. If $\rho(\bf{x})$ does not decay sufficiently rapidly, this may not be possible, for example a line charge $E_r \propto \frac{1}{r}$ and $\Phi \propto \ln r$. \subsubsection{Point Charge} Potential due to a point charge $q$ at the origin: \[ \Phi(\bf{x}) = \frac{q}{4\pi \eps_0|\bf{x}|} = \frac{q}{4\pi \eps_0 r} \] \subsubsection{Electric Dipole} Two equal and opposite charges at different positions. Without loss of generality, consider charges $-q$ at $\bf{x} = \bf{0}$ and $+q$ at $\bf{x} = \bf{d}$. \begin{center} \includegraphics[width=0.15\linewidth] {images/c42b3c889d5a11ed.png} \end{center} Potential due to the dipole: \[ \Phi(\bf{x}) = \frac{q}{4\pi\eps_0} \left( -\frac{1}{|\bf{x}|} + \frac{1}{|\bf{x} - \bf{d}|} \right) \] Apply Taylor's theorem for a scalar field, \[ f(\bf{x}) = f(\bf{x}) + (\bf{h} \cdot \nabla) f(\bf{x}) + \half (\bf{h} \cdot \nabla)^2 f(\bf{x}) + O(|\bf{h}|^3) \] to $f(\bf{x}) = \frac{1}{|\bf{x}|} = \frac{1}{r}$: \begin{align*} \Phi(\bf{x}) &= \frac{q}{4\pi\eps_0} \left( -\frac{1}{r} + \frac{1}{r} - (\bf{d} \cdot \nabla) \frac{1}{r} + O(|\bf{d}|^2) \right) \\ &= \frac{q}{4\pi\eps_0} \frac{\bf{d} \cdot \bf{x}}{|\bf{x}|^3} + O(|\bf{d}|^2) \end{align*} In the limit $|\bf{d}| \to 0$ with $q\bf{d}$ finite, we obtain a \emph{point dipole} with \emph{electric dipole moment} \[ \bf{p} = q\bf{d} ,\] potential \[ \Phi(\bf{x}) = \frac{\bf{p} \cdot \bf{x}}{4\pi\eps_0 |\bf{x}|^3} \] and electric field \[ \bf{E} = -\nabla\Phi = \frac{3(\bf{p} \cdot \bf{x})\bf{x} - |\bf{x}|^3 \bf{p}}{4\pi\eps_0|\bf{x}|^5} \] In spherical polar coordinates aligned with $\bf{p} = p\bf{e}_z$, \begin{align*} \Phi &= \frac{p\cos\theta}{4\pi\eps_0 r^2} \\ E_r &= -\pfrac{\Phi}{r} = \frac{2p\cos\theta}{4\pi\eps_0 r^3} \\ E_\theta &= -\frac{1}{r} \pfrac{\Phi}{\theta} = \frac{p\sin\theta}{4\pi\eps_0 r^3} \\ (E_\phi &= 0) \end{align*} Note \begin{itemize} \item $\Phi$ and $\bf{E}$ are not spherically symmetric. \item They decrease more rapidly with $r$ than for a point charge. \end{itemize} A point dipole $\bf{p}$ at the origin corresponds to \begin{align*} \rho(\bf{x}) &= -\bf{p} \cdot \nabla \delta(\bf{x}) \\ \Phi(\bf{x}) &= \bf{p} \cdot \nabla \left( \frac{1}{4\pi\eps_0 |\bf{x}|} \right) \end{align*} \subsubsection{Field lines and equipotentials} \emph{Electric field lines} are the integral curves of $\bf{E}$, being tangent to $\bf{E}$ everywhere. \myskip Since $\nabla \cdot \bf{E} = \frac{\rho}{\eps_0}$, field lines begin on $+$ charges and end on $-$ charges. In electrostatics, $\bf{E} = -\nabla \Phi$, field lines are perpendicular to the equipotential surfaces $\Phi = \text{const}$. \begin{center} \includegraphics[width=0.6\linewidth] {images/d68b78729d5d11ed.png} \end{center} \subsubsection{Dipole in an external field} Consider a dipole $\bf{p}$ in an external electric field $\bf{E} = -\nabla \Phi$ generated by distinct charges. With $-q$ at $\bf{x}$ and $+q$ at $\bf{x} + \bf{d}$, the potential energy of the dipole due to the external field is \begin{align*} U &= -q\Phi(\bf{x}) + q\Phi(\bf{x} + \bf{d}) \\ &= q(\bf{d} \cdot \nabla) \Phi(\bf{x}) + O(|\bf{d}|^3) \end{align*} In the limit of a point dipole, \begin{align*} U &= \bf{p} \cdot \nabla \Phi \\ &= -\bf{p} \cdot \bf{E} \end{align*} and sis minimised when $\bf{p}$ is aligned with $\bf{E}$. \subsubsection{Multipole Expansion} For a general charge distribution $\rho(\bf{x})$ confined to a ball $\{V \colon |\bf{x}| < R\}$, \[ \Phi(\bf{x}) = \frac{1}{4\pi\eps_0} \int_V \frac{\rho(\bf{x}')}{|\bf{x} - \bf{x}'|}\dd^3 \bf{x}' \] External potential $(|\bf{x}| > R$) -- expand \begin{align*} \frac{1}{|\bf{x} - \bf{x}'|} &= \frac{1}{r} - (\bf{x}' \cdot \nabla) \frac{1}{r} + \half (\bf{x}' \cdot \nabla)^2 \frac{1}{r} + O(|\bf{x}'|^3) \\ &= \frac{1}{r} \left[ 1 + \frac{\bf{x}' \cdot \bf{x}}{r^2} + \frac{3(\bf{x}' \cdot \bf{x})^2 - |\bf{x}'|^2 |\bf{x}|^2}{2r^4} + O \left( \frac{R^3}{r^3} \right) \right] \end{align*} Leads to the \emph{multipole expansion} of the potential (see Example Sheet 1, Question 11) \[ \Phi(\bf{x}) = \frac{1}{4\pi\eps_0} \left( \frac{Q}{r} + \frac{\bf{p} \cdot \bf{x}}{r^3} + \half \frac{Q_{ij} x_i x_j}{r^5} + \cdots \right) \] First three multipole moments: \begin{itemize} \item total charge (monopole - scalar) $Q = \int_V \rho(\bf{x}) \dd^3 \bf{x}$ \item electric dipole moment (vector) $\bf{p} = \int_V \bf{x} \rho(\bf{x}) \dd^3 \bf{x}$. \item electric quadripole moment (traceless symmetric 2nd order tensor) \[ Q_{ij} = \int_V (3x_i x_j - |\bf{x}|^2 \delta_{ij})\rho(\bf{x}) \dd^3 \bf{x} \] \end{itemize} For $r \gg R$, $\Phi$ and $\bf{E}$ look increasingly like those of a point charge $Q$, unless $Q = 0$, in which case they look like those of a point dipole, unless $\bf{p} = \bf{0}$, etc (See Example Sheet 1, Question 11).