Choose timezone
Your profile timezone:
US-Japan meeting
→
Asia/Tokyo
Description
Use following zoom.
US-Japan meeting
Zoom Meeting ID
63107134849
Host
Koji Nakamura
Passcode
48320804
Useful links
Join via phone
Zoom URL
US–Japan 4D Tracking / LGAD R&D Meeting Minutes
Date: November 4, 2026
Format: Zoom
Main topics: Simulation, monolithic LGAD, readout electronics, sensor production, gain studies, and future collaboration.
1. Next US–Japan Workshop
The next workshop is being considered for Osaka, Japan, between late February and April 2027.
The first half of February is difficult for Japanese participants due to examinations and thesis defenses. On the US side, the ePIC Workshop in Santa Cruz during the first week of March and CMS Week during the third week of March are potential conflicts.
Late March or April was considered preferable, as it would also allow more time to advance ongoing design, simulation, and development activities.
Action: Koji will consult the Osaka organizers and circulate possible dates through the US–Japan mailing list. The week of April 5 is primary candidate so far.
2. BNL Simulation Activities
2.1 Garfield++ simulation (Alec Haines)
Alec presented the development of a Garfield++-based simulation framework for LGAD and AC-LGAD devices.
The framework imports electric and weighting fields from TCAD simulations and calculates charge transport and induced signals. A reader for Silvaco TCAD output has been developed, extending the existing interface capabilities.
HEED is used to model particle ionization, including MIP signals. The framework enables studies of position-dependent signal response and charge sharing.
The main advantage is faster evaluation compared with full TCAD transient simulations, particularly for repeated geometry and parameter studies.
Further validation against TCAD and studies of mesh dependence remain important.
2.2 Resistive electrode simulation (Alec Haines)
Alec also presented an extension of Niccolò Cartiglia's resistive surface simulation framework for AC-LGAD layout optimization.
The model represents a 4×4 pixel array using a two-dimensional resistive network with 301×301 nodes.
It can evaluate charge sharing, spatial resolution, signal leakage to neighboring electrodes, and inter-electrode resistance.
Both laser-like signals with fixed charge and MIP signals with Landau fluctuations can be simulated. Different electrode layouts, including resistive strips and trenches, can be investigated efficiently.
This approach complements Garfield++ by providing rapid layout optimization, while Garfield++ offers a more detailed treatment of charge transport and signal formation.
2.3 Allpix Squared simulation (Alessandro Tricoli and collaborators)
Progress in integrating TCAD-based detector models into Allpix Squared was presented.
The work includes importing electric and weighting fields, particularly from Silvaco, and simulating particle interactions, charge transport, and induced signals.
Comparisons with resistive network models show encouraging agreement.
The longer-term objective is to establish an integrated simulation chain covering particle interactions, sensor response, and readout electronics.
Further developments, including gain suppression and improved electronics modeling, were discussed.
3. CMOS Readout Electronics (Nadira Chowdhury)
Nadira presented simulation studies of a CMOS inverter-based transimpedance amplifier for monolithic LGAD readout.
The simulations use the open-source SKY130 PDK with KiCad and ngspice, operating at a 1.8 V supply voltage.
The input waveform is based on LGAD signals generated by Allpix Squared simulations.
The studies focus on the dependence of transimpedance gain, bandwidth, and transient response on transistor dimensions and feedback resistance.
The target performance includes approximately 500 MHz–1 GHz bandwidth, sensitivity to input charges of tens of fC, timing jitter below 20 ps, and low power consumption.
The current model does not yet include realistic output loading or extracted layout parasitics. Noise and jitter evaluations are also ongoing.
Future work will extend the model to multiple channels, detector coupling, and electronics-induced crosstalk.
A comparison with Japanese ASIC simulation results was discussed. Although the Japanese PDK is not publicly available, selected performance results may be shared without disclosing proprietary design details.
A dedicated electronics discussion was proposed for a future meeting.
4. Deep-Junction LGAD Optimization (Chandanie Kutwaru)
Chandanie presented TCAD studies of alternative deep-junction electrode and implant configurations.
A modified design with separated deep N-well and deep P-well regions showed comparable signal amplitude and a faster peak response of approximately 0.5–0.75 ns, compared with about 1 ns for the previous structure.
The simulations also indicated approximately a factor-of-two reduction in coupling to the electronics.
Further studies investigated narrower electrodes with 5 µm width and 5 µm spacing, together with additional P-sub and N-sub implants.
The N-sub configuration increased the signal on one electrode but affected other electrode responses. P-sub implants improved the isolation of N++ electrodes, although additional optimization is required.
Future work will focus on implant geometry, electrical isolation, crosstalk, mesh dependence, and validation using 3D TCAD simulations.
5. LGAD Sensor Production (Gabriele Giacomini)
Gabriele presented the ongoing sensor production plans at BNL.
The production aims to accommodate requests from both DRD3 and the US–Japan collaboration.
The designs will include configurations compatible with thin and thick substrates, with different electrode pitches and geometries.
The final metal layouts have not yet been fixed, allowing additional input from participating groups. Some early-out wafers will proceed ahead of the final metal definition.
For the US–Japan collaboration, the current priority is a double-mesh design intended to demonstrate the feasibility of placing readout electronics at the sensor periphery.
Deep-junction devices cannot be implemented in the present production and will be considered in a future run.
KEK requests
Koji expressed interest in two sensor options.
Thick LGAD sensors for low-power electronics: For applications such as FCC-ee, a larger signal charge may be more important than achieving the best possible timing resolution. Thick LGADs could provide larger signals, potentially allowing substantial reduction or elimination of the preamplifier power consumption. The importance of timing information for FCC-ee remains under discussion.
Fine-pitch AC-LGAD sensors: Koji proposed exploring compatibility with the fast ASIC being developed at SLAC/Fermilab, particularly a 100×100 pixel matrix with a 50 µm pitch, corresponding to a 5×5 mm² active area.
Gabriele indicated that such a design could be considered, provided that detailed requirements for the pixel array, electrode geometry, and distance to the scribe lines are supplied.
Action: KEK will provide the relevant geometry and layout requirements. Gabriele will circulate the design list for review by consortium members.
6. LGAD Gain-Layer Dose Study (Gabriele Giacomini)
Gabriele presented gain measurements of LGAD devices fabricated on four wafers with different gain-layer implantation doses.
The gain layer was implanted at 1 MeV, and the devices exhibited breakdown voltages of approximately 90, 120, 170, and 200 V.
The measurements were performed using a charge-sensitive preamplifier and an Amptek PX5 system with a shaping time of 0.8 µs.
The purpose of the study is to isolate the effect of gain-layer dose while keeping other fabrication parameters controlled.
The discussion highlighted the importance of comparing devices at the same gain but different bulk electric fields, as this may reveal differences in charge collection, diffusion, and gain fluctuations.
Koji emphasized that the optimal operating point should not be determined solely by the maximum gain. Experience with HPK AC-LGADs has shown that timing performance and its stability against bias-voltage and temperature variations are important considerations.
A sensor with a narrow optimal bias range may be difficult to operate reliably, even if excellent timing resolution can be achieved under carefully controlled conditions.
Further studies should therefore consider gain, timing resolution, and operating stability together.
7. Test Beam and Future Collaboration
The detailed test-beam discussion was postponed because some key participants were no longer available.
A dedicated meeting of approximately 30–40 minutes will be arranged to discuss the upcoming test beam and the overall US–Japan collaboration plans.
An additional topical meeting focusing on electronics development and simulation was proposed for December 2026 or January 2027.
Possible topics include the Japanese ASIC developments, comparison with SKY130-based studies at BNL, and future integration of sensor and electronics simulations.
Koji also mentioned ongoing single-event burnout studies and 3D TCAD work, which may be presented once the results are ready.
The next general collaboration meeting is expected in approximately one month.
Future meeting times should better accommodate participants on the US West Coast and in Hawaii.
8. Action Items
|
Responsible |
Action |
Timeline |
|
Koji / Osaka organizers |
Coordinate possible dates for the next US–Japan Workshop |
Soon |
|
Alessandro |
Circulate information about the BNL Ozaki Program |
Soon |
|
Gabriele |
Distribute the sensor design list and collect feedback |
Soon |
|
KEK / Koji |
Provide detailed requirements for fine-pitch AC-LGAD sensors |
Follow-up |
|
BNL simulation team |
Continue deep-junction optimization and 3D TCAD studies |
Ongoing |
|
BNL electronics team |
Extend TIA simulations to realistic noise, parasitics, and multichannel response |
Ongoing |
|
Koji / Alessandro |
Organize a dedicated test-beam planning meeting |
Soon |
|
Koji / Alessandro |
Arrange a topical electronics meeting |
Dec.–Jan. |
|
Koji |
Grant Alessandro editing permissions for the Indico agenda |
Soon |
There are minutes attached to this event.
Show them.