The demand for highly sensitive, low-power, and miniaturized hydrogen (H₂) gas sensors is driven by the increasing use of H₂ as a clean and renewable energy source. The inherent dangers of H₂ leakage—due to its colorless, odorless nature, low ignition temperature, and explosive concentration range—necessitate sensors capable of detecting concentrations well below the explosion limit. While PdNP-decorated FETs offer advantages like current amplification, CMOS compatibility, and miniaturization, their sensitivity is limited by weak capacitive coupling between the PdNPs and the FET channel. Existing designs attempt to enhance this coupling by reducing the distance between the sensing layer and the channel (e.g., ultra-thin planar channels or nanowire channels with side-gates). However, the fundamental problem of indirect capacitive coupling due to floating PdNPs remains unresolved. This research addresses this issue by proposing a novel nanoscale FET sensor design that leverages the electron trapping effect in PdNPs to achieve superior signal transduction and sensitivity.

Existing H₂ sensors based on PdNPs and FETs have shown promise but suffer from sensitivity limitations due to weak capacitive coupling. Several approaches have been attempted to enhance this coupling, including the use of ultra-thin channels and nanowire channels with intimately attached side-gates. However, these modifications only partially address the underlying problem of indirect coupling caused by the floating nature of the PdNPs sensing layer. Previous work on other gas sensors (e.g., using carbon nanotubes, graphene, and other nanomaterials) have explored different sensing mechanisms and device architectures to improve sensitivity and response time, but none have addressed the issue of direct coupling in PdNP-FET H₂ sensors as effectively as the approach presented in this research.

The study utilizes a novel nanoscale FET sensor design, termed the SiNW-NAG FET, which employs a silicon nanowire (SiNW) channel passivated with a thin oxide layer and gated by two side-gates via nanoscale air gaps (NAGs). PdNPs are deposited near the SiNW sidewalls within the channel-gate loop, enabling direct coupling. The fabrication process involves standard silicon processing techniques, including SOI wafer processing, ion implantation, lithography, etching, and atomic layer deposition (ALD) for the oxide passivation layer. Electron beam evaporation is used for PdNP deposition. The key innovation is the use of an ultra-thin (2 nm) oxide layer to enable electron tunneling between the channel and PdNPs, establishing equilibrium. The H₂ sensing mechanism relies on the change in PdNP potential upon H₂ reaction, disrupting the equilibrium and causing electron transfer. The device's electrical characteristics, including transfer and output curves, are measured using standard semiconductor parameter analyzers and low-frequency noise analyzers. The H₂ sensing performance, including sensitivity, limit of detection (LOD), response time, and recovery time, are evaluated by exposing the devices to various H₂ concentrations and measuring the changes in channel current. The study also evaluates the selectivity of the sensor against other gases (CO, NO₂, NH₃). Device simulations are performed using a commercial TCAD simulator to understand the capacitive coupling and electron trapping effects.

The SiNW-NAG FET sensor with a 2 nm oxide layer exhibits a significantly higher sensitivity compared to a conventional back-gate FET with a 4 nm oxide layer. The 2 nm device shows a sensitivity of 3600%/ppm and an LOD of 4.4 ppb for H₂ detection at room temperature, representing a significant improvement over previous designs. Real-time measurements demonstrate fast response and recovery times, with response times comparable between the 2 nm and 4 nm oxide devices, indicating that the electron tunneling process is faster than the PdNP-H₂ reaction. The device achieves these results with an ultra-low power consumption of ~300 nW. The enhanced sensitivity is attributed to the direct coupling between the PdNPs and the channel facilitated by the electron trapping effect in PdNPs with a thin oxide layer. The electron trapping effect is confirmed by observing the subthreshold slope degradation and increased low-frequency noise with a thinner oxide layer. Analysis of the transfer curves in air and in the presence of H₂ reveals a H₂-induced energy shift in the PdNP trap distribution, providing further support for the electron trapping mechanism. Selectivity tests against CO, NO₂, and NH₃ show that the sensor exhibits good discrimination against these interfering gases. This superior performance surpasses that of other reported FET-based H₂ sensors.

The results demonstrate the efficacy of the electron trapping mechanism in achieving record-high sensitivity in H₂ gas sensing. The direct coupling between the gas reaction and the channel, enabled by electron tunneling through the thin oxide layer and subsequent trapping/de-trapping, is responsible for the significantly enhanced signal transduction. The high sensitivity, low power consumption, and good selectivity make this sensor highly suitable for various applications, such as H₂ safety monitoring in buildings and industries, and H₂ leakage detection in H₂-powered vehicles. The consistent response and recovery times across different oxide thicknesses indicate that the electron tunneling process is fast, and there are no significant performance tradeoffs associated with enabling the electron trapping effect. The observation that the sensitivity is directly related to the density of PdNP states near the Fermi level implies a potential pathway for further optimization by controlling PdNP size distribution.

This research demonstrates a highly sensitive nanotransistor-based gas sensor leveraging electron trapping in nanoparticles. The SiNW-NAG FET design with an ultra-thin oxide layer enables direct communication between the gas reaction and the channel, resulting in record-high H₂ sensitivity and ultra-low detection limits. The findings highlight the potential of this electron trapping-based signal transduction mechanism for achieving superior performance in gas sensing applications. Future work could explore optimizing PdNP size and distribution for further sensitivity enhancement, investigating the long-term stability of the sensor, and extending the mechanism to other gases by using different sensing nanoparticles.

While the study demonstrates excellent performance, certain limitations exist. Long-term stability under various environmental conditions (temperature, humidity) needs further investigation. The current study focuses on H₂ and a limited set of interfering gases; more extensive selectivity testing against a wider range of gases is needed. The study uses a specific fabrication process; scalability and cost-effectiveness of the process for mass production require further evaluation.