Off-axis: AHCAL + MS Discussion

Wednesday 23 Sept 2026, 09:00 → 11:00 Europe/Zurich

Aaron White (CERN), Hidetoshi Otono (Kyushu University (JP)), Hiroaki Menjo (Nagoya University (JP))

09:15 → 09:35 AHCAL Upgrade

AHCAL Upgrade Discussion – Meeting Minutes

1. Scintillator tiles and reflective wrapping

• The current reference tile size was discussed. A size of approximately 40 × 40 mm² is considered a practical baseline, while the exact granularity should still be optimized.
• IHEP has existing facilities and tooling for producing scintillator tiles of similar dimensions. If a different tile size is chosen, a new mold would be required, with an estimated additional cost of approximately 7 kCHF.
• China has equipment for automated reflective-foil wrapping. Modification of the wrapping setup for a different tile size may require an additional cost of a similar order.
• Chinese-made reflective foil could significantly reduce the cost compared with 3M ESR. The estimated cost is roughly one-quarter to one-third of the 3M option, although the light yield may be approximately 10–15% lower.
• Increasing the scintillator thickness was mentioned as a possible way to compensate for the reduced light yield without substantially increasing the total cost.
• Formal IHEP participation in the collaboration may not be required for production. It may be possible to place an order directly with the required dimensions and specifications. However, additional manpower, potentially including students from our side, may be needed during production. The expected production timescale is of the order of a few months.

Action: Zhen will provide more detailed information on the scintillator and reflective foil costs, including quotations or supplier information, so that the cost estimate in the Technical Proposal can be updated.

2. Front-end electronics and ASIC choice

• A major contribution to the current detector cost comes from the front-end readout electronics. The present estimate assumes RADIOROC-based electronics, with a cost of approximately 477 kCHF.
• Replacing RADIOROC with another ASIC could substantially reduce the overall cost.

• H2GCROC is currently considered a strong candidate because:

• it is already available,

• it is used for the CMS HGCAL system,
• its unit cost is substantially lower than RADIOROC.
• CMS cannot currently guarantee that spare H2GCROC chips will be available, so procurement directly from the manufacturer may be necessary.
• A quotation obtained by Hidetoshi indicated a price of approximately 100 EUR per chip, making H2GCROC financially feasible.
• CALOROC remains a possible future option, but its large-scale availability is expected only in a few years. Therefore, H2GCROC is presently regarded as the more realistic baseline option.
• Switching from RADIOROC to H2GCROC was estimated during the discussion to potentially reduce the total cost by approximately 300–400 kCHF.
• The backplane PCB may also be significantly simplified if the front-end ASICs are located outside the detector, providing an additional possibility for cost reduction.

3. Budget and production strategy

• The current Japanese-side funding is expected to cover approximately 200–300 kCHF.
• An additional funding request of approximately 100–200 kCHF has already been submitted.

• A staged construction strategy is acceptable:

• the full detector does not necessarily need to be installed at the beginning of Run 4;

• a partial detector could be installed first;
• additional layers or modules could be added during Run 4 as further funding becomes available.
• The next approximately two years can therefore be treated as an R&D and prototyping phase, while additional funding for mass production is pursued.
• The off-axis detector Technical Proposal should be updated by around November. The next two months should therefore be used to improve the readout cost estimate and identify possible production and collaboration options.

4. Tile-size optimization

• 40 × 40 mm² is currently considered the baseline tile size.
• This size already provides reasonable performance in existing studies and is attractive from the production point of view because IHEP has suitable existing tooling.
• Nevertheless, the granularity should be optimized before committing to mass production.

• Candidate geometries such as:

• 30 × 30 mm²

• 40 × 40 mm²
• 50 × 50 mm²

should be compared.
* ETH has already performed a study using the 40 × 40 mm² geometry. For νe CC identification, its performance was reported to be significantly better than the previous CNN-based study, by roughly a factor of two to three in the comparison discussed during the meeting.
* The planned workflow is:

1. prepare simulation samples for different geometries;
2. perform the computationally intensive training with ETH;
3. evaluate the resulting performance on the Japan/China side;
4. choose the final tile dimensions.
5. Yasu will organize a meeting with ETH to discuss the detailed workflow.

6. The approximate timescale discussed was:

7. simulation preparation: about 2 weeks;

8. pre-training: about 2 weeks;
9. evaluation: about 1 week per configuration.
10. Unless a smaller tile size provides a clear improvement in physics performance, the 40 × 40 mm² geometry remains the preferred baseline because of its lower channel count and simpler production.

5. Run 3 analysis and neutral-hadron simulation

• The neutral-hadron background simulation remains an important task for the Run 3 AHCAL analysis.
• Haichuan is expected to contribute significantly to this activity over the coming months.
• The detailed scope of the neutral-hadron simulation work should be discussed at the next analysis meeting.
• Computing resources are available on the China side, including a GPU cluster, and additional commercial/cloud GPU resources could be considered if needed.
• The Japan and China groups could also contribute to the evaluation/post-processing stage after the ETH training.

6. Action items

• Yasu

• Organize a meeting with ETH to define the detector-geometry optimization workflow.

• Prepare simulation samples for different tile sizes.

• Update the Technical Proposal cost table and detector description once updated quotations are available.

• Zhen

• Obtain detailed cost information or quotations for scintillator production, reflective foil, wrapping, and potentially PCB production.

• Clarify IHEP production capabilities, lead time, tooling costs, and manpower requirements.

• Japan / China / ETH

• Compare detector performance for different tile granularities and determine whether there is sufficient benefit to move away from the 40 × 40 mm² baseline.

• Haichuan

• Proceed with the neutral-hadron background simulation for the Run 3 analysis, with the detailed task definition to be discussed at the next analysis meeting.

7. Overall direction

The current working direction is to develop a more cost-effective Run 4 AHCAL design based on approximately 40 × 40 mm² scintillator tiles, H2GCROC-based front-end electronics, and the use of IHEP production capabilities where possible. In parallel, the detector granularity will be optimized with ETH using machine-learning-based performance studies. The outcome of these studies, together with updated cost estimates, will be incorporated into the next revision of the Technical Proposal.

09:35 → 09:55 Muon Spectrometer

09:55 → 10:15 Discussion

Feel free to request to be added.

Speakers: Aaron White (CERN), Annabelle Marie Parry (University of Liverpool (GB)), Enrique Kajomovitz Must (Department of Physics), Enrique Kajomovitz Must (Technion, Israel Institute of Technology), Hidetoshi Otono (Kyushu University (JP)), Hiroaki Menjo (Nagoya University (JP)), Mikito Sakuma (Nagoya University (JP)), Shikma Bressler (Weizmann Institute of Science (IL)), Sune Jakobsen (CERN), Tomohiro Inada (Kyushu University (JP)), Dr Umut Kose (ETH-Zurich), Yasuhiro Maruya (Kyushu University (JP))