(i)??Boundary and interference effects: by definition, for an infinite TPC no boundary effects (like wall stops, scattering and diffusion) would occur. Interference effects, on the other hand, are generated by decay electrons, affecting the muon stop reconstruction in a time dependent manner. The experiment has to balance these two competing systematic effects, by carefully selecting muons within a clean fiducial volume, without introducing interference distortions. For the final, best MuCap run R07 their total uncertainty added up to \newcommand{\LS}{{\Lambda_{\rm singlet}}} \newcommand{\si}{{\rm s^{-1}}} \delta \LS = 3.3~\si. The obvious remedy for boundary effects is a larger TPC volume, coupled with potential geometry improvements, as well as the reduction in the beam stopping volume. While the dimensions perpendicular to the TPC drift field are only constrained by practical considerations, the drift time cannot be much longer than Tobs, in order to avoid reducing the acceptable beam rate. We expect that the TPC can be operated with drift fields up to 10?kV cm?1, as demonstrated at higher density in MuSun [, 97: #ropaac190bib097], which would double the drift velocity and allow drift distances of 20?cm. To improve the tracking quality, a geometry of independent pads like MuSun has proven advantageous, as the MuCap cathode wires strung in the direction of the muon tracks provide very limited information. As MuSun demonstrated, full digitization of all signals instead of simple threshold timing information, adds powerful tracking capabilities

(ii)??Gas impurities: For the R07 run, transfer to gas impurities amounted to \newcommand{\LS}{{\Lambda_{\rm singlet}}} \newcommand{\si}{{\rm s^{-1}}} \delta \LS \sim 1~\si. Gold coating of inner vessel surfaces, improved gas chromatography (already achieved in MuSun) and/or spectroscopy and, most importantly, full digital readout of the TPC signals [, 97: #ropaac190bib097], for in situ detection of capture recoils, should reduce this uncertainty to below 0.1 \newcommand{\si}{{\rm s^{-1}}} \si

(iii)??Electron detector effects: a significant uncertainty of \newcommand{\LS}{{\Lambda_{\rm singlet}}} \newcommand{\si}{{\rm s^{-1}}} \delta \LS \sim 1.8~ \si was included in the MuCap error budget, because of incompletely understood discrepancies between alternative electron track definitions. Because diffusion processes in hydrogen introduce systematic problems when applying tight vertex cuts, MuCap concluded that precision tracking is not essential. Thus a new experiment should use scintillators or scintillating fibers with SiPM readout, which are simple, robust and more stable than the wire chambers used in MuCap. Then an instrumental uncertainty below 1 ppm similar to MuLan [, 50: #ropaac190bib050] can be expected

(iv)??\newcommand{\ppm}{{pp\mu}} \ppm molecular effects: although capture from \newcommand{\ppm}{{pp\mu}} \ppm molecules amounts only to 3% in 1?MPa hydrogen gas, the uncertainty introduced by the inconsistent determinations of the ortho-para rate \lambda_{\rm op} [, 8: #ropaac190bib008] shown in figure, 3: #ropaac190f03, introduces a \newcommand{\LS}{{\Lambda_{\rm singlet}}} \newcommand{\si}{{\rm s^{-1}}} \delta \LS \sim 1.8~ \si uncertainty [, 16: #ropaac190bib016]. As shown in figure 2 of [, 15: #ropaac190bib015], the poor knowledge of \lambda_{\rm op} also leaves unresolved the question whether the previous measurement of ordinary muon capture in liquid hydrogen [, 90: #ropaac190bib090] or, alternatively, the measurement of radiative muon capture [, 9: #ropaac190bib009] strikingly deviates from theory. The high density cryogenic TPC developed for the MuSun ?D experiment, could settle both issues with a first precise measurement of \lambda_{\rm op} when filled with protium gas of about 10% liquid density

High Energy Physics and Nuclear Structure (New York: