Light-Tight Shield for Skipper-CCDs

• The paper demonstrates that thin aluminum films (50–100 nm) achieve >99.6% optical suppression with no measurable X-ray attenuation.
• Methodologies include liftoff lithography and e-beam evaporation to deposit precise Al coatings on CCD dies, ensuring robust performance in harsh environments.
• Simulations and experiments confirm that optimized Al thickness and amplifier shielding are critical for mitigating internal photon and SEE backgrounds.

A light-tight shield for Skipper-Charge-Coupled Devices (skipper-CCDs) is a critical system component designed to suppress optical and near-infrared (NIR) photon backgrounds while maintaining high efficiency for soft X-ray detection. Skipper-CCDs achieve deep sub-electron resolution and have been identified as promising technology for high-precision energy reconstruction in space-based X-ray astronomy, where uncontrolled optical/NIR backgrounds can saturate the device and distort the true X-ray signal. The implementation of thin metallic (aluminum) coatings enables effective optical suppression with negligible X-ray attenuation in the operational energy range, and is complemented by additional shielding and best practices for suppressing internally generated photons from readout electronics.

1. Fabrication and Implementation of the Aluminum Light-Tight Shield

The light-tight shield for skipper-CCDs is fabricated via frontside deposition of thin aluminum (Al) films onto commercial 1.35 megapixel skipper-CCD dies. The process begins with patterning the aluminum shield using liftoff lithography: application and development of SPR955 photoresist define the intended shield regions, typically rectangular sectors in each device quadrant. Aluminum layers of 20 nm, 50 nm, and 100 nm thickness are deposited through e-beam evaporation (Temescal FC2000, 1 Å/s, 10⁻⁶ torr). The photoresist is then lifted off with Remover1165 and an O₂ plasma clean is performed (Botti et al., 31 Dec 2025).

The rationale for chosen thicknesses is grounded in the electromagnetic skin depth of Al at visible/NIR wavelengths (~10 nm). The 20 nm coatings test minimal blocking; 50 nm represents ~5 optical skin depths and is expected to provide >99% suppression; 100 nm further mitigates risks of pinhole defects and targets more aggressive NIR blocking, while remaining thin compared to X-ray attenuation lengths.

2. Quantitative Optical and Near-Infrared Suppression

Optical and NIR suppression by the Al films was evaluated across 650–1000 nm using a monochromator-controlled halogen lamp and integrating sphere to generate uniform, normal-incidence illumination. The shielded CCDs, maintained at 160 K in a high-vacuum environment (<10⁻⁴ torr), underwent timed exposures at various durations, with shutter-closed controls subtracted for dark current and readout noise (1.22 e⁻ for 10 samples, 0.25 e⁻ for 200 samples).

Signal rates per region and wavelength were extracted by fitting the slope of measured electrons per second as a function of exposure time. The light transmission factor, $T(\lambda)$, is then defined as the ratio of shielded to unshielded signal rates for each wavelength. Experimental results demonstrate that 20 nm films provide only 5–10% blocking, while 50 nm achieves $T<0.4\%$ across the measured range, corresponding to >99.6% suppression, and 100 nm achieves $T<0.1\%$ (>99.9% suppression).

The optical attenuation is well described by an exponential model,

$T(\lambda) = \exp[-\mu(\lambda) t]$

where $\mu(\lambda)$ is the absorption coefficient of Al and $t$ the film thickness (Botti et al., 31 Dec 2025). This regime is robust across the operational parameter space relevant to spacecraft backgrounds (wavelengths λ > 650 nm).

3. X-ray Transmission and Skipper-CCD Detection Efficiency

Preserving X-ray detection efficiency is essential for high-sensitivity instrumentation. Measurements of X-ray transmission were conducted using a $^{55}$Fe source (producing 5.9 and 6.4 keV X-rays) with the shielded and unshielded regions on the same CCD. Clusters corresponding to X-ray events (1450–1800 e⁻/event) were counted and event rate slopes were compared between shielded and unshielded regions.

No measurable X-ray attenuation was observed for Al layers up to 100 nm at 5.9 and 6.4 keV; the event rates were statistically identical (within 0.02 s⁻¹). These findings validate that thin metallic coatings at sub-µm scales are compatible with X-ray instrumentation mandates (Botti et al., 31 Dec 2025).

A summary of these results is presented below:

Al thickness (nm)	Efficiency (shielded, s⁻¹)	Efficiency (unshielded, s⁻¹)	Observed X-ray loss
20	0.19 ± 0.02	0.19 ± 0.02	None
50	0.22 ± 0.02	0.22 ± 0.02	None
100	0.20 ± 0.02	0.22 ± 0.02	None

4. Simulations of Broad-Band X-ray Attenuation

Geant4 v10.7 simulations were employed to extend the experimental results across a broader X-ray energy regime (0.1–25 keV) for both front-illuminated and back-illuminated CCD stacks. In the front-illuminated geometry, the full sensor stack includes the Al shield (varying thickness), SiO₂ layers, polysilicon, Si₃N₄, and Si bulk. The back-illuminated configuration features a thin dead polysilicon layer and partial charge collection layer atop the silicon substrate.

The key attenuation relationship follows Beer–Lambert law,

$I(E) = I_0 \exp[-\mu(E)\,t]$

where $\mu(E)$ is the energy-dependent X-ray attenuation coefficient for Al or the multilayer stack.

Simulation outputs indicate for front-illumination, the X-ray detection efficiency at 3.5 keV ($\eta_\text{front}(3.5\,\text{keV})$) reaches 60–65%, dictated by nonmetallic surface layers, and is largely unaffected ($\Delta\eta(E)<1\%$) by the addition of Al films ≤100 nm for energies above 1 keV. In the back-illuminated case, efficiency is 85–90% at 3.5 keV, and Al-induced losses remain negligible above 0.5 keV; below this energy, absorption in Al becomes relevant (Botti et al., 31 Dec 2025).

5. Suppression of Internal Photon Backgrounds: Amplifier Light and System Shielding

Internally generated backgrounds, notably amplifier-emitted infrared photons produced by hot-electron quasi-bremsstrahlung in the output transistor M1, require specialized mitigation beyond surface coatings. These photons (1–1.5 µm wavelength) can generate single-electron events (SEE) in adjacent pixels, with rates that scale linearly with readout time (Barak et al., 2021).

Operational strategies to suppress amplifier light include:

• Operating the CCD amplifier in the linear regime gate bias (e.g., VDD ≈ –21 V), which decreases hot-electron luminescence by a factor ~100 compared to saturation (≤ –22 V).
• Enclosing the entire readout board, including the M1 region, in a multi-layer, IR-opaque Al or Cu housing, precisely gasketed to the vacuum vessel to block line-of-sight photon paths.
• Employing multiple baffle stages, interior black-anodized or IR-absorbing surfaces, and direct mating of the amplifier box with the CCD skirt to minimize leakage.
• Maintaining deep cryogenic temperatures (e.g., 135 K) and implementing both internal (Cu) and external (Pb) shielding to further suppress extrinsic dark current and environmental backgrounds (Barak et al., 2021).

The semi-empirical model for SEE backgrounds is

$$\mu(t_{\rm EXP}, t_{\rm RO}) = \lambda_{\rm DC}\, t_{\rm EXP} + \left(\frac{\lambda_{\rm DC}}{2} + \lambda_{\rm AL}\right) t_{\rm RO} + \mu_{\rm SC}$$

with $\lambda_{\rm DC}$ the dark current rate, $\lambda_{\rm AL}$ the amplifier-light SEE rate, and $\mu_{\rm SC}$ the spurious charge per read. In optimal shielding and bias settings, amplifier-light backgrounds are reduced below the (already low) dark-current level.

6. Implications for Space Instrumentation and Best-Practice Design

The trade-off in selecting Al film thickness centers on maximizing optical/NIR suppression without compromising X-ray quantum efficiency. Films of 50 nm and 100 nm deliver >99.6% suppression over 650–1000 nm, and do not measurably attenuate X-ray signal for energies ≥1 keV in either experimental or simulated regimes. A 50 nm layer is optimal for minimal mass and complexity; 100 nm offers greater robustness against pinhole defects and long-term oxidative degradation.

For space missions in sun-illuminated or high-background environments, thin Al layers directly deposited on CCDs provide a lightweight, mechanically simple solution relative to bulk optical filters and facilitate stringent standards for background noise control in sub-electron-noise X-ray measurement (Botti et al., 31 Dec 2025). Future engineering efforts are directed at optimizing deposition methods (e.g., thermal evaporation, sputtering), integrating Al shields with thinned, back-illuminated or anti-reflection–coated CCDs, and qualifying the system through radiation and environmental testing for long-term space deployment.

A systematic approach incorporating shielded amplifier readout, multi-stage baffles, and vacuum vessel gasketed housing extends light-tightness and minimizes SEE backgrounds, crucial for the extension of skipper-CCD technology into ultra-sensitive rare-event searches in particle physics and astrophysics (Barak et al., 2021).

7. Summary Table: Key Features and Performance of Aluminum Shields

Feature	50 nm Al	100 nm Al	Significance
Optical/NIR suppression (650–1000 nm)	>99.6%	>99.9%	Eliminates signal swamping
X-ray loss (≥1 keV)	<1%	<1%	Preserves measurement SNR
Fabrication margin	Minimal mass/defects	Higher pinhole tolerance	Reliability in harsh env.
Application	Space-based CCD X-ray	Space-based CCD X-ray	Direct CCD integration

The confluence of empirical and simulation-based findings establishes thin Al films as an effective, low-cost, and robust light-tight shield for skipper-CCD–based X-ray instrumentation in demanding optical backgrounds (Botti et al., 31 Dec 2025, Barak et al., 2021).