Past exam of the short options paper shared by Prelims and Parts A and B of the physics course of the University of Oxford / 2018 / S19 / 2

The accelerating supply transfers $I_{e}V=50$ kW into the beam, but the electrons need not be discarded with that kinetic energy. An electrostatic recovery system decelerates the used beam before collection and returns most of its energy to the electric field or supply. The net required power then covers unrecovered beam energy and losses, rather than the whole circulating kinetic-energy flux.

For example, $99\%$ recovery leaves $0.01(50\mathrm{kW})\simeq 500$ W of unrecovered beam power, plus auxiliary loads. This explains the stated two-order-of-magnitude scale. It does not imply that gun heating, vacuum equipment, magnets and all electrical losses are automatically included in a $500$ W total plant budget.

Regenerative braking in an electric vehicle converts part of its mechanical kinetic energy back into electrical energy stored in a battery, rather than dissipating it all in friction brakes. An industrial recuperative heat exchanger transfers heat from a hot exhaust stream to incoming process fluid, reducing the required new heat input.

Both recover energy that would otherwise be rejected, while remaining subject to conversion losses and limits on the usefulness of the recovered form. Neither is a source of net energy over a complete cycle.

Express temperatures as energies $\Theta =k_{B}T$. A longitudinal cathode-energy spread of order ${\Theta }_c$ remains of that order during an ideal common electrostatic acceleration. At final mean speed $v_0$, a small energy spread produces $\delta v_{\parallel }\simeq \delta E/(m_e{v}_0)$. In the accompanying beam frame, the corresponding random kinetic-energy scale isThe last equality uses the nonrelativistic acceleration energy $E_e=m_e{v}_0^2/2$. This is the approximate temperature convention supplied in the hint; exact numerical factors depend on the emitted distribution and the definition of spread. At $50$ keV, relativistic corrections are modest but not strictly zero.

Electrostatic acceleration narrows longitudinal velocity spread relative to the fast mean beam even though it does not remove each electron's transverse cathode motion. In ideal magnetized cooling, slow proton motion couples to the electron guiding-centre dynamics, making this very small longitudinal spread relevant to the attainable ion temperature. Practical voltage ripple, space charge, alignment errors and ion heating can set much larger limits.

The stated conversion gives ${\Theta }_c\simeq 0.10$ eV. Hence the ideal longitudinal bath temperature isThis estimate equates the cooled proton thermal-energy scale with the effective electron bath scale, not the two particles' random speeds. The proton root-mean-square speed for a one-component convention $m_p{\sigma }_v^2=k_{B}T$ would be about $2.2\mathrm{m}{\mathrm{s}}^{-1}$.

The result is an ideal floor under the hinted cooling model. It is not a promise of a sub-millikelvin stored beam in a real ring, where noise, beam self-fields, intrabeam scattering and other heat sources enter the balance.

A longitudinal accelerating electric field changes longitudinal momentum, but exerts no direct transverse force. With a constant guiding magnetic field and adiabatic transport, the transverse gyro-energy is consequently retained from the cathode:More generally the magnetic moment of the orbit gives ${\Theta }_{\perp }/B\simeq \text{constant}$, so the equality assumes the same relevant field strength, not merely that a field is present all along the line.

In the strong-magnetization regime, many electron gyro-orbits occur during an important long-range electron-ion interaction and the gyroradius is small. The ion then interacts predominantly with a guiding centre; rapid transverse cyclotron motion does not act like an unmagnetized bath at the full transverse electron temperature. This allows a proton equilibrium scale far below $T_c$ despite the hot transverse gyro-motion.^[ref]

That conclusion is conditional on magnetization and collision scales. A finite gyroradius, nonadiabatic fields or transverse beam errors can restore heating and degrade cooling; it is not generally true that transverse temperature can never affect the cooled beam.