Al₂O₃‑Coated Microchannel Plates via Atomic Layer Deposition Deliver Five‑Fold Gain and Extended Lifetime

Abstract

Microchannel plates (MCPs) are critical in high‑gain electron multipliers used across photon and particle detection. Traditional MCPs suffer from high dark current, limited lifetime, and sub‑optimal secondary‑electron‑emission (SEE) performance. In this study, we deposit a conformal Al₂O₃ SEE layer inside MCP pores by atomic layer deposition (ALD), systematically characterizing morphology, elemental distribution, and structural uniformity with SEM and EDS. We find that a 8–10 nm Al₂O₃ coating yields a five‑fold increase in average gain compared with conventional MCPs and demonstrates superior sensitivity and longevity under accelerated UV exposure.

Introduction

MCPs—arrays of 4–25 µm diameter, 0.2–1.2 mm long, high‑aspect‑ratio pores—are indispensable in MCP‑PMTs, night‑vision optics, electron microscopy, and X‑ray imaging. Their performance hinges on three geometric parameters: bias angle (5–15°), length‑to‑diameter ratio (20:1–100:1), and open‑area ratio (60–80%). Conventional MCPs are fabricated from lead‑silicate glass via complex glass‑drawing and etching processes. After hydrogen‑reduction, a conductive layer and a low‑yield SiO₂ SEE layer form on the pore walls. While widely adopted, this approach incurs high surface roughness, gas adsorption, and resistive‑gain coupling, leading to noisy dark currents, reduced charge extraction, and shortened operational life.

Atomic layer deposition offers atomic‑scale control over film thickness and composition, enabling uniform coatings inside complex geometries. Prior work has shown that depositing high‑SEE materials (e.g., Al₂O₃) inside MCP channels can raise secondary‑electron yield (SEY) and mitigate surface contamination. However, achieving uniform coverage over millimeter‑long pores remains challenging. This paper demonstrates that an extended‑precursor ALD protocol, coupled with a stop‑flow purge strategy, delivers a highly uniform Al₂O₃ layer on 24 µm‑diameter, 40:1 aspect‑ratio MCPs, and quantifies the resulting performance gains.

Experimental and Calculation Methods

The test rig (Fig. 1) comprises a gold cathode, MCP, and PCB anode inside a vacuum chamber (2 × 10⁻⁴ Pa). Electrodes receive high‑voltage supplies via feedthroughs; a picoammeter records MCP output current. A low‑intensity mercury lamp provides UV photons for gain measurements, while a high‑power lamp accelerates lifetime testing.

ALD was performed on both silicon wafers (for SEY measurement) and MCPs. MCPs (thickness = 1.2 mm, pore = 24 µm, bias = 10°) were pre‑heated to 200 °C for 1 h to promote precursor adsorption. Two deposition strategies were compared: (1) extended precursor exposure per cycle (sample F); (2) stop‑flow purge (sample G). The Al₂O₃ ALD sequence was TMA/N₂/H₂O/N₂ (0.05 s/10 s/0.05 s/10 s). Films of 4–10 nm (samples B–E) and 60 nm (sample F) were deposited; sample G used 600 cycles with 3 s stop‑flow pauses.

Post‑ALD, 200 nm copper electrodes were evaporated onto both MCP surfaces to facilitate electrical measurement. SEM and cross‑sectional EDS characterized film thickness and elemental distribution. A 50 mm MCP was half‑coated to enable direct performance comparison between ALD‑processed and pristine regions.

Results and Discussion

Figure 3 illustrates MCP output current versus primary photoelectron energy at a fixed 1400 V bias. Current rises linearly below 400 eV and plateaus thereafter, reflecting the SEY peak of SiO₂ at ≈400 eV. The open‑area ratio (~60%) allows additional electrons to re‑enter channels at higher fields, sustaining output current beyond the SEY peak.

SEY curves (Fig. 4) confirm Al₂O₃ reaches a maximum yield of 3.6 at 400 V, surpassing SiO₂. The 400 eV SEY drop is offset by the MCP’s high open‑area ratio, which recycles reflected electrons at higher fields.

Uniformity assessment (Fig. 5) used cross‑sectional EDS to map Al distribution along pore walls. The extended‑precursor method yielded near‑uniform Al content throughout the 40‑length pore, while the stop‑flow approach produced a pronounced gradient—low at pore ends, high in the middle—due to precursor depletion during purge.

SEM thickness profiling (Fig. 6) corroborated the EDS findings: the extended‑precursor deposition achieved a consistent 8–10 nm film along the entire pore, whereas the stop‑flow sample exhibited thickness variations up to 15 nm.

Electrical characterization (Fig. 7) demonstrates that 8–10 nm Al₂O₃ coatings provide >5× output current relative to uncoated MCPs. Thicker films (≥12 nm) showed diminishing returns, likely due to increased electron scattering and reduced channel conductivity. The half‑coated MCP illuminated a phosphor screen, confirming superior imaging from the coated region.

Lifetime tests (Table 2) exposed MCPs to high‑power UV. Conventional MCPs lost ~50 % output current after sustained illumination, whereas ALD‑MCP maintained ~6 nA, evidencing a markedly extended operational life. Dark current increased marginally for ALD‑MCPs (1.0 pA→1.2 pA) due to the high‑yield Al₂O₃, but remained negligible compared to total signal.

Conclusions

We have shown that a 8–10 nm Al₂O₃ layer deposited by an extended‑precursor ALD protocol uniformly coats high‑aspect‑ratio MCP pores, boosting SEY to 3.6 and delivering five‑fold gain over standard MCPs. The improved uniformity eliminates charge‑recombination hotspots, extends device lifetime under high‑intensity UV, and preserves imaging fidelity. This approach offers a scalable, process‑intelligible pathway to high‑performance MCPs for next‑generation photon and particle detectors.