Biodegradable, flexible silicon nanomembrane-based NOx gas sensor system with record-high performance for transient environmental monitors and medical implants

The study targets the development of biologically benign and mechanically flexible electronics that can be fully bioresorbed after use, addressing needs in E-skin and biomedical applications. Nitric oxide (NO) is a key biomarker in human physiology (vascular, nervous, respiratory systems) and, upon conversion to NO₂, a significant environmental pollutant contributing to acid rain and ozone formation. Existing NOx sensors based on graphene, carbon nanotubes, polymers, and metal oxides face challenges such as rigidity, low sensitivity/selectivity, high power consumption, and nonbiodegradability. The research question is whether a flexible, bioresorbable single-crystal silicon nanomembrane (SC-Si NM) platform can deliver record-high sensitivity and selectivity for NOx at room temperature while maintaining mechanical robustness and operability in humid/wet conditions suitable for environmental monitoring and implantable medical devices.

Prior approaches for NOx detection have leveraged high-performance channels like graphene, carbon nanotubes, polymers, and metal oxides, often incorporating structural modifications to increase surface-to-volume ratios. Despite advances, limitations remain for wearable/implantable systems: rigidity, insufficient sensitivity and selectivity, high power consumption, and lack of biodegradability. The present work positions SC-Si NMs as an alternative to overcome these shortcomings.

Fabrication: SC-Si NMs were thinned from SOI wafers (top Si ~300 nm) via repeated thermal oxidation at 1100 °C and HF etching to ~100 nm. Phosphorus doping (SOD, 950 °C) formed highly doped regions for electrical contacts. BOX undercut released the thinned SC-Si NMs for transfer printing onto PMMA/diluted PI temporary substrates. Active sensor areas (NOx gas, humidity, temperature) were defined by SF6-based RIE. Interdigitated electrodes (Mg ~300 nm) were deposited by e-beam evaporation. Patterned dielectric layers (SiO₂ ~100 nm) and contact pads were formed by BOE wet etching and e-beam evaporation. Mesh-type bridges facilitated lift-off and transfer. After removing the bottom PI, devices were transfer-printed onto biodegradable substrates (e.g., PLGA or PCL). For humid/wet operation, a semipermeable PDMS membrane (~20 µm) was spin-cast as encapsulation. Electrical characterization: Conducted with a Keithley 2636B electrometer at 0.01 V DC bias in a temperature-adjustable chamber (RT–50 °C) with 1000 sccm flow (MFC-controlled). Responses defined as Ro/Rgas for reducing gases and Rgas/Ro for oxidizing gases, where Ro is baseline in dry N₂/air and Rgas is saturated resistance after 500 s exposure. Selectivity tests: Assessed responses to NO₂, NO (oxidizing) and NH₃, acetone (CH₃COCH₃), ethanol (C₂H₅OH), CO₂, CO, H₂S (reducing/other gases) at defined concentrations (e.g., 5 ppm NO₂) at room temperature. Temperature and humidity sensing: Integrated resistive temperature sensor characterized for conductance vs. temperature, deriving sensitivity (~9.6 nS/°C) and temperature coefficient (~0.006/°C). Humidity sensor capacitance measured across relative humidity range (@100 Hz) to correct gas readings. Bioresorption tests: Visualized dissolution of systems on PLGA in PBS (pH 7.4) at 37 °C over 0, 1, 4, 8 h; tracked dissolution of Mg electrodes (Mg + 2H₂O → Mg(OH)₂ + H₂), SC-Si NMs (Si + 4H₂O → Si(OH)₄ + 2H₂), SiO₂ (SiO₂ + 2H₂O → Si(OH)₄), and PLGA to lactic/glycolic acids + H₂. Cell studies: RAW 264.7 macrophages cultured in DMEM + 10% FBS, antibiotics; M1 activation via 100 ng/ml LPS + 20 ng/ml IFN-γ for 24 h. DAF-FM assay used for NO measurement after incubation in L-arginine-free DMEM; 10 µM DAF-FM DA ±150 µM L-arginine for 30 min; fluorescence read by Cytation 3. Immunofluorescence: fixation, permeabilization, blocking, primary anti-iNOS (1:500), secondary, DAPI and phalloidin staining; imaged by confocal microscopy. NOx detection from M1-activated cells performed in flasks with 2 ml L-arginine-free DMEM; sensors inserted, L-arginine delivered via needles; sensors connected to electrometer and logged via LabVIEW. Mechanical testing and modeling: Large-scale arrays (5×5, 1.7 cm × 1.9 cm) on PCL substrate/encapsulant with serpentine interconnects evaluated under bending (flat, 5 mm, 3 mm radii) and uniaxial tensile strain up to 40%; cyclic tests at 5 mm bend radius and 35% stretch up to 1000 cycles. 3D finite element analysis (ABAQUS, C3D8R) simulated deformation of a 2×2 array with PCL substrate (E=0.25 GPa, ν=0.44) and PDMS encapsulation (E=2.61 MPa, ν=0.49) for bending; PDMS substrate for tensile. Thin layers (Si NM E=188 GPa, Mg E=45 GPa, SiO₂ E=66.3 GPa) modeled as skin layers.

- Record-high room-temperature NO₂ sensitivity: response Rgas/Ro ≈ 136 (13,600%) at 5 ppm with fast response (~30 s) and recovery (~60 s). - Detection limit ~20 ppb (linear least squares from 0.1–5 ppm data at RT). - High selectivity: responses to NO and NO₂ at least 100-fold higher than to NH₃, acetone, ethanol, CO₂, CO, and H₂S without additional surface functionalization. - Crystallinity dependence: SC-Si NMs outperform polycrystalline Si and amorphous Si with identical dimensions (~100 nm thick, 1.45 mm × 1.2 mm). - Temperature dependence: response inversely related to temperature from RT to 50 °C, highlighting the need for compensation. - Integrated temperature sensor: sensitivity ~9.6 nS/°C and temperature coefficient ~0.006/°C; humidity sensor provides well-resolved capacitance changes across RH. - Mechanical robustness: array maintained ≥90% of response under bending (down to 3–5 mm radius) and ≥95% under uniaxial stretch up to 40%; negligible degradation after 1000 cycles of 5 mm bending or 35% stretching. - Bioresorption: complete dissolution of constituent materials in PBS (pH 7.4) at 37 °C over hours; Mg electrodes dissolve fastest; SC-Si NM, SiO₂, and PLGA degrade to benign products. - In vitro operation: sensors function stably in humid/wet environments with semipermeable PDMS membranes and can detect NOx produced by M1-activated macrophages.

The findings demonstrate that SC-Si nanomembranes afford exceptional surface reactivity and electronic properties for NOx detection at room temperature, achieving sensitivity and selectivity surpassing prior graphene, CNT, polymer, and metal oxide sensors while addressing biodegradability. The device architecture (thin SC-Si NM channels, Mg IDEs, SiO₂ dielectrics) and operation at low bias (0.01 V) enable low-power, high-performance sensing. Integration of temperature and humidity sensors provides pathways to compensate ambient effects, crucial given the inverse temperature dependence observed. Mechanical designs employing device islands and serpentine interconnects ensure stable sensing under bending and stretching and over repeated cycles, validated by experimental results and FEA. The semipermeable PDMS membrane allows reliable operation in wet/biological conditions, and in vitro tests with activated macrophages indicate feasibility for biomedical monitoring of NO-related physiology. The transient, bioresorbable construction supports applications in temporary implants and disposable environmental monitors without retrieval.

This work introduces a biodegradable, flexible SC-Si nanomembrane-based NOx sensing system that achieves record-high sensitivity (R/R0 ~136 at 5 ppm), rapid response/recovery, and strong selectivity at room temperature. The system integrates temperature and humidity sensing for environmental compensation, exhibits excellent mechanical robustness in large-area arrays, and fully bioresorbs under physiological conditions. In vitro demonstrations highlight potential for transient biomedical implants and environmental monitoring. Future directions include in vivo validation of implantable operation, extended long-term stability studies in physiological conditions, system-level wireless/readout integration, and broader analyte detection via multiplexed or functionalized SC-Si NM platforms.

Demonstrations are limited to in vitro evaluations and controlled chamber tests (RT–50 °C, defined gas flows); in vivo performance and long-term physiological stability were not reported. Temperature sensitivity requires compensation for accurate readings. Detailed selectivity and interference studies in complex biological matrices beyond the tested gases are not provided. Full author affiliation details (for some coauthors) and certain system-level aspects (power/wireless) are outside the scope of the presented data.