Towards a New μ→eγ Search with the MEG II Experiment: From Design to Commissioning

Round 1

Reviewer 1 Report

The paper describes the motivation, the design, and the current status of the commissioning of the MEG II experiment. MEG II is one of the leading experiments in the search for charged lepton flavour violation and is employing many innovative detector developments. The content is thus of great interest to the community and the fact that the authors also discuss the issues experienced with the new detectors is greatly appreciated. The paper however requires quite a few language edits and will – even after that – profit from another round of careful proof-reading.

General comment:

In Figures 7,8, and 9 as well as many places in the text, there is a lot of emphasis on the difference between the early runs with a limited amount of DAQ channels and the 2021 run with all of them. Whilst it is certainly a big achievement to have the full DAQ running, I would try to focus a bit more on the directly subdetector-related results in the discussion of the subdetectors and maybe also reflect that in the figures.

Detailed comments/language edits:

L23: It is not clear what the subject of “producing” is

L24/25: Maybe also cite the recent full SM BR calculation using neutrino masses:
Eur. Phys. J. C, vol. 79, no. 1, p. 84, 2019, [Erratum: Eur. Phys. J. C 80, 438 (2020)] / Eur. Phys. J. C, vol. 80, no. 6, p. 506, 2020.  

L27: stringent experimental limits, I would then drop experimental in the following line
L30: I am not sure about the logic of this sentence – what do the 70 years (and the MEG limit) have to do with the new physics sensitivity of the process?
L35: Sensitivity for what?
L36: such as Mu3e
L36 and many other places: Make sure there is a space between words and the [xx] of citations.

L37: Maybe you can come up with a better title for the section?

L44: kinematics?
L52: dominating instead of well known?
L54: at the 100 ps level

L58: “the spectrum can be mitigated” does not really make sense

Figure 1 does not print very well, in particular the text.

L73: How do the beam monitoring devices address the target deformation systematic?
L75: I think this sentence is not grammatically correct.
L77: detectors is repeated, maybe you can find a more elegant wording

L100: The latter…
L103: As written, it is not clear whether “its” refers to the target or the dot pattern

L105: Maybe “inside the magnet” rather than “spectrometer internal volume”?
L108: “essential for” instead of “substantial to”
L113: distributions of what?
L115: consists of
L116: stereo wire?
L124: Move the sentence about the acceptance to before the COBRA discussion – the acceptance is larger because of the drift chamber not because of a change in field.
L140: LXe acceptance is jargon, make clear that you mean acceptance for positrons emitted back to back to a photon which can be detected in the LXe.
L149: is made from BC-422 (maybe also give a reference to the data sheet)
L149: Write out ESR
L150: Why the extra Tedlar wrapping?
L155: Maybe add what the cooling medium is and how much power you want to cool at which temperature.

L160: ensuring full azimuthal coverage

L163: the CDCH
L163: What limits the tracking efficiency to 65%? - if it is geometrical acceptance, maybe separate acceptance and reconstruction efficiency

L174: Make clearer what you mean by modular, as this seems to be unique

L179: What is helm-shaped?
L191: If the unit is really per cm, say that this is per cm of sense wire
L193: Maybe comment on whether this 20% gain loss is acceptable

L196: time of propagation

L197: the active cooling of the electronics…
L201: using instead of performing
L202: Here you suddenly change to first person singular

L206: condensation of ambient humidity on wires
L207: the wire volume
L222: achieve a

L236: Write out PDE once
L254: “In addition” does not really match the preceding sentence

L260: both the upstream

L260: The sentence staring here is very long and can easily be split

L273: 2 cm per side cubes

L276: will be installed

L286: complete
L288: It is not clear what “it” refers to
L290: Another one of the rare switches to active voice
L291: by developing
L313: a weighted sum algorithm similar to

L320: in order to forward
L321: algorithms can be run/executed/…
L322: selections, for example to

L326: the full detector connected to the data acquisition

L330: Is this SES or 90% CL or…? Say so
L332: Split the following sentence in two
L346: I know what you mean, but it reads weird to say that 41 ps fulfill the 40 ps spec…
L351: Equivalent to what? 1 MeV/c neutrons? Say so
L361: Any outcomes from this to report?
L365: What is the “stable lengthening”?
L367: Maybe split this sentence. Also make clear that the muon beam produces positrons to be detected.
L370: Anomalously
L375: resulting in
L376: I do not understand this sentence
L387: show momentum …

L387: Make sure numbers and units end up on the same line

L397: Is it not a degradation rather than an instability

L398: What does the muon beam environment actually imply for the MPPCs? Is there a significant flux of charged particles despite the magnetic field? Neutrons? Or all just photons?
L407: same level as what?
L407: of the readout electronics
L419: It refers rather to the collimator – the whole paragraph would profit from a careful re-wording

L424: of the MPPCs

L426: What does the last sentence mean? What exactly is 0.57 mm?
L459: calibration measurements
L459: Figure and the number should be on the same line
L467: parallel to what?
L478: within the SM
L498: reduced material?
L510ff: The fact that the full readout is now available is maybe stated a bit often.

Ref 9: The Mu3e TDR is published as Nucl.Instrum.Meth.A 1014, 165679, 2021

 

 

 

Author Response

We would like to thank the reviewer for the interesting comments and the appreciation for the efforts of MEG II Collaboration.

As stated a few times throughout the paper, the MEG II experiment is undergoing its first-ever datataking with a number of channels relevant to study the behaviour of its detector.

We would have loved to share a more in-depth analysis of the data we are collecting right now but we think that this is beyond the scope of this paper.

As recognised by the reviewer, in this paper we intended to share with the community the difficulties that the MEG II Collaboration faced in the construction of the experiment and that now all detectors are finally operative.

All changes not explicitly mentioned have been accepted. We are putting here the list of points that we think deserves a direct answer to the reviewer:

 

L37: Maybe you can come up with a better title for the section?

As far as we understood from reading the Journal Author instruction the main sections are forced to be "Introduction", "Material and Methods", "Results", "Conclusions". We would have loved to use something like "Design and construction" but looks not allowed. (Please let us know if it is not the case)

 

L115: "Why extra Tedlar Wrapping?"

- For extra light tightness.

Added: "for extra light tightness".

 

L163: What limits the tracking efficiency to 65%?

Reply: The 65% tracking efficiency is the combined hit reconstruction efficiency of the CDCH (about 75%) x the matching efficiency with the reconstructed hits of the pixelated TC (about 85%). The current number is obtained with full MC simulations and the most updated reconstruction algorithms. With further developments, we will improve the current result. In any case, we are not far from the 70% positron efficiency quoted in the MEG II Design paper ("The design of the MEG II experiment", Eur. Phys. J. C 2018, 78, 380: doi:10.1140/epjc/s10052-018-5845-6.). The current limiting factor is the pile-up discrimination at 7e7 mu/s beam intensity with the real noise situation on data. As quoted in the recent Symmetry paper ("The Search for μ+ → e+γ with 10^(–14) Sensitivity: The Upgrade of the MEG Experiment", Symmetry 2021, 13(9), 1591: doi.org/10.3390/sym13091591) the positron efficiency improves from 65% to 74% at 3.5e7 mu/s beam intensity with the current reconstruction algorithms.

 

no modification: we think this is too technical

 

L174: Make clearer what you mean by modular, as this seems to be unique.

Reply: The modular construction method is already roughly summarized. We first solder the two wire ends on the pads (few mm spacing) of two PCBs (wire-PCBs, each hosting 32 wires) and then we stack the PCBs on the endplates: one PCB on the US endplate and one PCB on the DS endplate. We radially alternate the wire-PCBs with PEEK spacers, whose thickness is adjusted through dedicated survey measurements to have the wires at the correct radial position in the chamber. This is repeated for all the endplate sectors and wire layers. Pressure-sensitive tape is used to keep the PCB and spacers in position. At the end of the assembly phase, we seal the endplates with special encapsulants and adhesives (Stycast and ThreeBond respectively) to have the He tightness.

 

Changed a paragraph to better describe the endplate structure: The wires are not strung directly on the chamber but soldered at both ends on the pads of two PCBs, which are then radially stacked by means of PEEK spacers in twelve $30^{\circ}$-sectors of the endplates. In between these sectors, a radial aluminium pillar provides the required mechanical robustness. Figure \ref{fig:cdchopen_pisa} shows the fully wired detector.

 

L426: What does the last sentence mean? What exactly is 0.57 mm?

The goal of this measurement is to align the MPPCs on the experimental coordinate which is defined by the position of the COBRA magnet.

 

Changed paragraph to: As a result, the MPPCs are aligned with respect to the COBRA magnet with a precision of 0.57 mm by the combination of the two measurements.

 

L376: I do not understand this sentence.

Reply: The use of a low-Z gas mixture such as He:Isobutane 90:10 is essential to minimize the Multiple Coulomb Scattering effects on 50 MeV positrons. However, the high He ionisation potential of 24.6 eV is such that a crossing charged particle produces only a small number of primary electron-ion pairs. The average number of ionization clusters produced by the passage of a 52.8 MeV positron in the counting gas is low: about 13/cm. In combination with the small drift cell size (few mm), this enhances the contribution to the single-hit resolution coming from the statistical fluctuation of the primary ionization along the track. This leads to a bias in the measurement of the distance of closest approach of the particle to the sense wire, if only the first cluster is used for timing. The Cluster Timing techniques consist in measuring the arriving time of all the individual clusters, thus minimizing the bias and improving the single-hit resolution. In order to detect the single ionization clusters in the signal waveforms, we need a very high gain (order 5e5) and fast FE electronics, the temporal separation between signals produced by different ionization clusters being a few ns. The citation should help to understand.

 

changed paragraph to make more clear: The gas gain was determined with cosmic rays in a clean environment, resulting to be ($4 \div 7) \times 10^5$, as expected and demonstrated to be sensitive to the single ionization cluster, thus allowing to explore cluster timing techniques.

 

L398: What does the muon beam environment actually imply for the MPPCs? Is there a significant flux of charged particles despite the magnetic field? Neutrons? Or all just photons?

Reply: The MPPCs are also exposed to electrons and positrons that come from the interaction between the material before the MPPCs. We have not yet pinned down the particles to induce the radiation damage.

 

no modification to the text: this is yet an open point.

 

L510ff: The fact that the full readout is now available is maybe stated a bit often.

 

paragraph changed: Sentence removed.

Reviewer 2 Report

Dear authors, please find in attachment my report.

Comments for author File: Comments.pdf

Author Response

We would like to thank the reviewer for the interesting comments and the clarity of the open questions.

All English-related suggestions are applied with minor changes needed to fit with similar ones by the other reviewer.

 

Q1: RDC LYSO design.

Question: Can you tell me which is the reason for the LYSO crystals?

The confusion probably comes from the fact that RDC needs to identify low energy positron from RMD decays from the abundant fraction of high energy positron from Michel decays. The energy deposit in the LYSO is different in the two cases because of the different energy of the particles. The plastic scintillator part instead provides the timing performances.

added: The energy deposit in the calorimeter part provides the additional discrimination power that is needed to separate RMD positrons from the higher energy positrons from Michel decays.

 

Q2: pTC results (Line 345-355) design

Question: It looks like more than 4 counters are crossed in average. Is this correct?

Yes, ~9 counters are crossed on average. The values quoted for the single counter resolution are obtained from laboratory measurements with a 90Sr source, while the overall resolution refers to data collected during the pre-engineering runs, where resolution turned out to be a little worse because of the larger noise contribution in the experiment.

Paragraph improved: The overall detector resolution is estimated to be 35 ps: this value is obtained by dividing the single hit resolution by the square root of the number of hits of a signal e+, which is 9 on average. The single counter resolution, measured in laboratory with a 90Sr source, is on average 72 ps and 81 ps for W=40 mm and W=50 mm scintillators. The discrepancy between these values and the overall resolution measured in the experiment is due to the larger noise contribution observed in the MEG II environment.

 

Question: Is your maximum neutron flux estimated including safety factors?

The initial predicted temperature was 15 degrees, so we already used some margin. Anyway, the estimate of the maximum neutron flux includes a factor 2 safety because they are calculated on 3 full years, while the actual exposition will be only in the second half of the year when we have beam time assigned.

No modification to the text.

 

Are you operating in vacuum? Can your cooling system perform below 10°C and allow additional temperature drop?

No, we are not operating in vacuum. Our cooling system can only reach 8 degrees.

No modification to the text.

 

Q3: Calorimeter System

Question: Have you tested consecutive heating-procedure to check if this annealing method works also if applied more than once?

We have not yet checked the consecutive heating procedure because we had few chances to test the annealing procedure. We need to anneal MPPCs with an empty cryostat but need LXe for PDE measurement. We will provide more long-term studies in the future.

No modification to the text.

 

Question: What is the problem for the resolution?

Answer: The degradation of the detector performance is limited because the position and energy resolution are not limited by the photo-electron statistics, and the time resolution is defined only by PMTs when the MPPC PDE is low.

The impact is estimated to be <0.5~mm / 0.6\% / 30~ps for the position, energy, and time resolution respectively, though these need to be checked with the data.

added paragraph: The degradation of the detector resolution was found to be limited because the position and energy resolution are not dominated by the photo-electron statistics. The time resolution is defined only by PMTs when the MPPC PDE is reduced.

 

Question: Can you explain what is the deterioration you are trying to describe?

We described the impact on the detector resolution based on the MC simulation in the text.

The effect on the resolution is limited as described in the previous answer, so no major effect on the experiment is expected as long as the signal to noise remains high enough for a calibration to be possible. The main source of sensitivity loss is the time required to perform the annealing process which directly translated in a loss of accumulated muon decays if it happens in the middle of the run. The collaboration is currently evaluating a run plan to limit this.

changed paragraph: The collaboration is evaluating how to include the time needed for the annealing process in the run plan to minimize any eventual sensitivity loss by the reduced statistics.

 

Question: the energy resolution looks 80% worse than expectation. Is this somehow related to the observed PDE loss?

Answer: The contribution to the worse detector energy resolution from the PDE loss is 0.6% based on MC simulation studies: the energy resolution is 0.8% with MPPC PDE 20% and it is deteriorated up to 1.0% with MPPC PDE 2%.

No change: Already answered in the previous questions

 

Q4: Camera

Question: Can you send me some details of the camera and its position in the apparatus?

The cameras are out of acceptance, thus they do not affect the tracking. - The position of the cameras and the LED is now added to the text in lines 104-105. From the references: one camera is a "digital CMOS photo-camera (IDS, mod. UI-3282SE), with a Sony IMX264 sensor having 2456×2054 pixels of 3.5 μm size, for a total sensor size of 8.473 mm×7.086 mm. A TUSS optical system, mod.LVK7518, with a focal length of 75 mm and a maximum aperture of f/1.8 is used", the other camera is "an industrial camera with a 1/2.3-inch CMOS sensor with 3856×2764 pixels, each 1.67 μm square, and a 50 mm lens".

 

Question: Is the camera small enough not to bother positron tracks or is it out of acceptance? Could you have a set of optical fibers?

The volume is gas-tight because it's filled with He, so bringing out optical fibers could create some difficulties.

added paragraph to properly explain the camera positioning: They are equipped with a LED and installed roughly one meter upstream and 10 cm off-axis, well out of beam path and detector acceptance.

 

Q5: CDCH and Momentum Resolution

Question: It is the increase on Nhits that provides the final momentum improvement?

In MEG 16 trapezoidal Drift CHambers (DCH) were placed below the muon stopping target to track the positrons. The DCH modules were spaced by 10.6° intervals and the sensitive area ranged radially from 19.3 cm to 27.9 cm and covered longitudinally |z| < 50 cm. The radiation length of the MEG II CDCH is slightly better than the MEG DCH: 1.5e(-3) X_0 vs. 2e(-3) X_0, despite the full azimuthal coverage around the muon stopping target of the CDCH. However, you are right, the major improvement for the momentum resolution is the total number of hits available for tracking, a factor of four more than MEG. This information is written in the 2.2.2 Section but there is also a smaller improvement due to the single hit resolution. The current positron momentum resolution is based on full MC simulations and the most updated reconstruction algorithms and it is 100 keV vs. 380 keV in MEG. Soon we will have the first measurements based on data.

No modification to the text: Not the only reason for resolution improvement

 

Question: Is momentum or angular resolution the most important parameter to provide a background reduction?

The angular resolution is the most important parameter for the reduction of the accidental background. This is due to the fact that the rate of accidental background scales quadratically with the angular resolution and only linearly with the momentum resolution. The current angular resolution is based on full MC simulations and the most updated reconstruction algorithms and it is 6.7 mrad vs. 9 mrad in MEG. Further developments in the reconstruction software will improve these results.

No modification to be done in the text

 

Question: Could you explain the procedure or point to a document to read?

Unfortunately, there is no paper describing the broken wire extraction procedure. We may suggest reading the Section 5.1 of the Marco Chiappini's PhD Thesis available at the following link: https://meg.web.psi.ch/docs/theses/chiappini_phd.pdf.

No modification to the text: Very technical

 

Q7: Hit timestamp with WDB.

Question: Please explain and add this to the text

The additional timestamp reported in lines 314-316 is used at trigger level as the counterpart of the timing which can be extracted from the DRS4 ASIC waveform.
Currently, we are running with an oversampling by 8 resulting in a time bin of size 1.56 ns.
This value is constrained by the limitation of the FPGA we are using (Spartan6) and the impossibility to perform time of walk compensation within the available trigger computation latency.

Added paragraph:  Operating with the maximum oversampling by a factor 8, the timestamp information available at trigger level is available with a 1.56ns time bin.