Why the Aurora Is Made Locally: Field-Aligned Acceleration above the Poles — Epoche B2
Introduction: a popular but incorrect picture Textbook diagrams often suggest that Earth's magnetic field simply guides solar-wind particles down to the poles, where they strike the atmosphere and glow. This picture is attractive but wrong in a crucial way. The electrons that actually produce the bright, structured, curtain-like aurora are accelerated locally , at altitudes of a few thousand kilometres, by electric potentials aligned with the magnetic field. The energy does ultimately come from the solar wind, but it is first stored as magnetic energy in the stretched magnetotail and then released suddenly, rather than delivered by direct particle entry. This essay traces the causal chain that leads to that conclusion, and each link in it can be checked against a number. Two terms are needed at the outset. The solar wind is the continuous outflow of ionised gas from the Sun's corona, reaching Earth at roughly $400\ \mathrm{km/s}$; it is a plasma, meaning a gas hot enough that its electrons have been stripped from their nuclei, so that it carries currents and responds strongly to magnetic fields. And the electron volt ($\mathrm{eV}$) is the natural energy unit here: the energy an electron gains when it falls through a potential difference of one volt. A thousand of them is a $\mathrm{keV}$. Cause one: direct entry cannot supply the energy The naive picture fails twice over, on energy and on numbers. On energy: solar-wind electrons carry typical thermal energies of roughly $10\ \mathrm{eV}$, while the electrons that produce visible auroral arcs arrive at the atmosphere with $1$–$10\ \mathrm{keV}$ — a hundred to a thousand times more. This is not measured indirectly; it is read straight off the particle detectors on polar-orbiting spacecraft, which count electrons and sort them by energy. Since a static magnetic field can bend a particle's path but never change its speed — the magnetic force $q\mathbf{v}\times\mathbf{B}$ is always perpendicular to the motion, and a perpendicular force does no work — mere guidance along field lines cannot make up the difference. Something must do work on the electrons, and only an electric field can. On numbers: the field lines converge sharply as they approach Earth, and a converging field reflects particles rather than delivering them. The reason is an adiabatic invariant — a quantity that stays almost constant when a particle's environment changes slowly compared with its own gyration around the field line. For a charged particle spiralling in a magnetic field, that quantity is the magnetic moment $\mu = m v_\perp^2 / 2B$, where $v_\perp$ is the speed across the field. Writing the pitch angle $\alpha$ for the angle between the velocity and the field, so that $v_\perp = v\sin\alpha$, and recalling that the total speed $v$ cannot change, the invariance of $\mu$ becomes $$\frac{\sin^2\alpha}{B} = \text{constant along the field line.}$$ As the particle moves into stronger field, $\sin^2\alpha$ must rise in proportion; when it reaches $1$, the pitch angle is $90^\circ$, all the motion is across the field, and the particle turns around. This is the magnetic mirror. A particle therefore reaches the atmosphere only if its equatorial pitch angle is small enough that it never mirrors first — that is, only if it lies inside the loss cone , defined by $$\sin^2\alpha_{\mathrm{LC}} = \frac{B_{\mathrm{eq}}}{B_{\mathrm{ionosphere}}}.$$ Put in real fields. On a field line crossing the equator at eight Earth radii, a dipole gives $B_{\mathrm{eq}} \approx 3.1\times10^{4}\ \mathrm{nT}/8^3 \approx 60\ \mathrm{nT}$ (the field of a dipole falls as the inverse cube of distance), while at the foot of that line, near the pole, $B \approx 6\times10^{4}\ \mathrm{nT}$. The ratio is $10^{-3}$, so $\sin\alpha_{\mathrm{LC}} \approx 0.032$ and $\alpha_{\mathrm{LC}} \approx 1.8^\circ$. The fraction of an isotropic population inside so narrow a cone is $1-\cos\alpha_{\mathrm{LC}} \approx 6\times10^{-4}$ — about one particle in seventeen hundred. Direct funnelling therefore delivers neither enough energy per particle nor enough particles. One qualification belongs here, since the expanded argument makes it visible. The faint, structureless diffuse aurora really is produced by ambient hot electrons scattered into the loss cone by plasma waves, without local acceleration. It is the bright discrete arcs — the curtains and rays that people photograph — that require the mechanism described below. Cause two: the tail stores energy, reconnection releases it In a plasma of high conductivity the magnetic field is "frozen in": field lines and plasma move together, as if the field were threaded through the fluid. This is what allows the solar wind to drag Earth's field into a long tail on the night side. But the freezing can break down in thin sheets where the field reverses direction, and there the field lines can cut and reconnect into a new topology, converting stored magnetic energy into bulk flow and heat. Dungey (1961) proposed that this happens twice in a cycle: on the dayside, where the solar wind's own field meets Earth's, and again far down the tail. The dayside step has a switch. Earth's field points northward at the sub-solar magnetopause; when the interplanetary magnetic field carried by the solar wind turns southward, the two are antiparallel, and reconnection proceeds efficiently. Each reconnected field line, now attached to Earth at one end and to the solar wind at the other, is swept over the poles into the tail, where it piles up in the two magnetic lobes . The stored energy density of a magnetic field is $B^2/2\mu_0$, and the lobes grow over roughly half an hour to an hour. The magnitude is worth checking. With a lobe field of $20\ \mathrm{nT}$, the energy density is $$u = \frac{(2\times10^{-8}\ \mathrm{T})^2}{2\times(4\pi\times10^{-7})} \approx 1.6\times10^{-10}\ \mathrm{J\,m^{-3}},$$ which seems negligible until multiplied by the volume of the lobes: a tail some $20$ Earth radii a