Optimizing Antihydrogen Production via Slow Plasma Merging

TL;DR

The paper advances antihydrogen production by slow-merge mixing of large antiproton and positron plasmas in a nested trap, measuring time-resolved plasma properties to optimize yield. It reports a record production of $2.3\times10^{6}$ antihydrogen atoms per $15$ minutes, with $70\%$–$80\%$ of input antiprotons forming stable $\overline{\mathrm{H}}$, by controlling the mixing rate and cooling the positrons. A slow-extraction analogue clarifies how the antiproton entry radius into the positron plasma controls the beam-like fraction of antihydrogen, and the work demonstrates a path to higher yields and better beam quality for spectroscopy and fundamental tests.

Abstract

We measure the time-dependent temperature and density distribution of antiprotons and positrons while slowly combining them to make antihydrogen atoms in a nested Penning-Malmberg trap. The total antihydrogen yield and the number of atoms escaping the trap as a beam are greatest when the positron temperature is lowest and when antiprotons enter the positron plasma at the smallest radius. We control these parameters by changing the rate at which we lower the electrostatic barrier between the antiproton and positron plasmas and by heating the positrons. With the optimal settings, we produce $2.3\times 10^6$ antihydrogen atoms per $15$-minute run, surpassing the previous state of the art -- $3.1\times 10^4$ atoms in $4$ minutes -- by a factor of $20$.

Figures (7)

• Figure 1: Simplified cross section of the experiment and calculated magnetic field $B$ along the axis of symmetry. Drawing in (a) is scaled to match the axial position $z$ in (b). The downstream direction corresponds to increasing $z$. Abbreviations: Scintillating bar arrays (Tracking Scintillator), microchannel-plate phosphor screen (MCP), inner bore scintillator plates (Fast Scintillator), Cusp magnet coils (Magnet), gate valve (GV), and $\overline{\mathrm{H}}$ Beam Detector consisting of bismuth germanium oxide crystal (BGO) and scintillating bar layers (Hodoscope). We use the symbols $\overline{\mathrm{p}}$ and $\mathrm{e^+}$ for the antiproton and positron plasmas. Microwave meshes at $z=-0.52$ and $+0.48\,\mathrm{m}$ and field ionizers at $z=-0.12$ and $+0.12\,\mathrm{m}$ are not shown.
• Figure 2: Trapping potentials and annihilation vertex positions. (a) The electric potential at $r=0$ generated by the biased electrodes at the start and end of mixing, for the protocol used in Section \ref{['sec:time']}. $B$ is given again for reference. Length and space charge $\phi$ of the $\overline{\mathrm{p}}$ and $\mathrm{e^+}$ plasmas are suggested by the blue and red ellipses. (b) Vertices found using the Tracking Scintillator during $\overline{\mathrm{H}}$ production ("Mixing") and slow release downstream ("Null dump").
• Figure 3: Plasma diagnosed partway through mixing: (a) images, (b) fraction remaining $f$, (c) plasma temperature $T$, (d) count rate at the $\overline{\mathrm{H}}$ Beam Detector and $1/1000$ times the count rate at the Fast Scintillator. In (a), the side-length of the image squares is $4\,\mathrm{cm}$, or about $5\,\mathrm{mm}$ in the trap (see Appendix \ref{['sec:appx_plasma']}). The intensity scale is the same for all images in a given row. The color gradient corresponding to normalized intensity range $0$ to $1$ is given at the right of the image squares. The signals in (d) are the averages from 81 uninterrupted runs, with the standard deviation over these runs given as the error bars.
• Figure 4: Slow extraction of $\overline{\mathrm{p}}$ to the MCP. (a) Transverse profiles integrated in time: each pixel is summed over all the frames of the corresponding video. Each image is then normalized to its own maximum. (b) Temporal profiles integrated in space: intensity is the sum of all pixel values in that frame. The intensity of each trace in (b) is normalized to its integral, and the time axis is rescaled such that 0 is the start of the extraction and 1 is the time $\tau$ (in seconds) given in the key under each image. The color gradient bar and the side-length of the image squares in (a) are the same as in Fig. \ref{['fig:time']}.
• Figure 5: Mixing for a variable duration $\tau$: (a) $\overline{\mathrm{H}}$ Beam Detector counts (green) and $\overline{\mathrm{H}}$ formation efficiency $\varepsilon$ (black), (b) $\mathrm{e^+}$ plasma temperature $T_\mathrm{e}$ (red, open circles) and space charge $\phi_\mathrm{e}$ (black, filled circles). Background $\beta$ is subtracted from each point using the formula $\beta=\beta_0+\beta_{80}\cdot\varepsilon/80\%$, where $\beta_0$ is the number of counts per run in an equivalent window without $\overline{\mathrm{p}}$ annihilations and $\beta_{80}$ is the number when $\varepsilon=80\%$ and the gate valve is closed so that no $\overline{\mathrm{H}}$ can reach the $\overline{\mathrm{H}}$ Beam Detector. Error bars in (a) are statistical, i.e., $\sqrt{\mathrm{Counts/run}}$ or the equivalent for the ratio of Fast Scintillator counts that defines $\varepsilon$.