GRANITE: High-Resolution Imaging and Electrical Qualification of Large-Area TPC Electrodes

Next-generation dual-phase time projection chambers (TPCs) for rare event searches will require large-scale, high-precision electrodes. To meet the stringent requirements for high-voltage performance of such an experiment, we have developed a scanning setup for comprehensive electrode quality assurance. The system is built around the GRANITE (Granular Robotic Assay for Novel Integrated TPC Electrodes) facility: a gantry robot on top of a $2.5,\text{m}\times1.8,\text{m}$ granite table, equipped with a suite of non-contact metrology devices. We developed a coaxial wire scanning head to measure and correlate localized high-voltage discharge currents in air with high-resolution surface images. We find that the identified discharge ‘hotspots’ are transient and show no significant correlation with static visual features. Next, we established a quantitative relationship between artificially induced abrasive surface damage on the wires and a reduction in the discharge inception voltage. This work provides a novel non-invasive tool for qualifying wires dedicated for use in electrodes for future low-background experiments.

💡 Research Summary

The paper presents GRANITE (Granular Robotic Assay for Novel Integrated TPC Electrodes), a comprehensive quality‑assurance platform designed to evaluate large‑area electrodes for next‑generation dual‑phase time projection chambers (TPCs). The system is built on a 2.5 m × 1.8 m granite table that provides a flat, vibration‑damped surface and stable environmental monitoring (temperature, humidity, pressure, dew point). A gantry robot (isel‑system) mounted on the table offers sub‑5 µm positioning repeatability and moves a suite of non‑contact metrology tools across an area of roughly 2 m × 1.4 m. The tools include a confocal microscope for sub‑micron 3‑D surface imaging, a high‑resolution industrial camera with a telecentric lens for detailed optical inspection, a laser distance sensor for sag and tension measurements, and a profile laser scanner for rapid surface profiling.

A key innovation is a coaxial wire‑scanning head that surrounds a test wire with a grounded copper cylinder while leaving a 3 mm slit for the camera view. By applying a controlled voltage difference (‑5 kV to ‑15 kV) between the wire and the cylinder, a nearly uniform electric field is generated around the wire. Finite‑element simulations confirm the field uniformity and quantify edge effects; a fitting formula (Eq. 3.3) corrects for small offsets during actual scans.

The measurement procedure involves cleaning stainless‑steel wires (California Fine Wire) with ultrasonic baths and a mild citric‑acid passivation, mounting them on a small frame, and then performing a raster scan. The robot moves the head in 4 mm steps; at each position an image is captured and ten current‑voltage readings are averaged. A “hotspot” is defined as a location where the dark‑regime discharge current exceeds ~10 µA, a regime that precedes visible corona and is chosen to avoid wire damage. By repeating scans at different voltages and after cleaning cycles, the authors build a statistical picture of discharge behavior.

The results show that discharge hotspots are transient: their locations shift between scans and do not correlate with static visual defects such as scratches or particles identified in the high‑resolution images. This suggests that the dominant trigger for dark‑regime discharge is not surface topography but dynamic processes such as charge accumulation or local gas ionization. An auto‑encoder analysis of the images finds only a weak link between image reconstruction error (or blur) and hotspot occurrence, indicating that current imaging resolution and dataset size are insufficient for reliable machine‑learning‑based defect prediction.

To establish a quantitative link between surface condition and electrical performance, the authors deliberately introduced abrasive damage to the wires, creating a range of surface roughness (Ra from ~0.1 µm to ~1 µm). Systematic scans reveal that wires with Ra ≥ 0.5 µm exhibit a reduction in discharge inception voltage (ΔU_c) of roughly 2 kV compared with smoother wires. This empirical relationship provides a practical criterion: surface roughness above a certain threshold can be used to reject wires before they are incorporated into large‑scale electrodes.

Overall, GRANITE integrates mechanical sag/tension assessment, high‑resolution optical inspection, and localized high‑voltage discharge testing into a single automated workflow. Compared with traditional separate inspections, it dramatically reduces the time and cost required to qualify the thousands of wires needed for a multi‑meter TPC such as the proposed XLZD (∼3 m diameter). By enabling early identification of electrically vulnerable wires, the system promises to improve the reliability and background suppression of future low‑background dark‑matter and neutrino experiments. The authors outline future work that includes scaling the system to full‑size electrode panels, automating roughness‑to‑voltage decision thresholds, and enhancing defect prediction through larger image datasets and deep‑learning models that incorporate 3‑D surface topology.