Smart facemask for wireless CO₂ monitoring

Facemask wearing has been widely adopted to mitigate COVID-19 transmission, though concerns persist regarding physiological effects such as CO₂ rebreathing in the dead space volume (DSV) of masks. Existing studies report mixed findings on physiological impacts, but agree on increased breathing resistance and DSV extension with masks. Current CO₂ monitoring solutions (capnography, portable monitors) are often bulky, power-demanding, and suited to laboratory conditions rather than wearable, real-life use. The research question addressed is whether a compact, battery-free, near-field powered, flexible sensor platform can provide accurate, real-time CO₂ measurements inside FFP2 facemasks under daily-life conditions. The purpose is to design, develop, and evaluate an opto-chemical CO₂ sensor integrated with an NFC-enabled flexible tag to deliver reliable measurements, compensate for temperature effects, assess independence from humidity, and operate under varying ambient light, thereby enabling non-invasive health monitoring and potential clinical applications.

CO₂ sensing techniques include infrared absorptiometry, catalytic bead, electrochemical, and optical sensors. Infrared sensors tend to be larger, costlier, and more susceptible to interference; catalytic bead sensors require combustion; electrochemical sensors may suffer cross-sensitivity. Optical luminescence-based sensors offer miniaturization, electrical isolation, low power, and reduced noise sensitivity, making them suitable for wearable NFC-powered platforms. Prior approaches to measuring CO₂ inside mask DSV include home-built and commercial instruments that are reliable but not wearable due to size and power requirements. Literature reports on mask-related physiology are mixed: some report negligible physiological effects from elevated DSV CO₂, others find reductions in SpO₂ and increases in heart and respiratory rates and discomfort. There is consensus on increased breathing resistance and DSV with masks. Wearable facemask sensing efforts have targeted gases, particulates, temperature, strain, humidity, respiration, and vital signs, often requiring batteries and rigid electronics. The present work differentiates itself by providing a battery-less, flexible PET NFC platform with an optical chemical CO₂ sensor integrated into FFP2 masks and smartphone-based powering and readout.

System architecture: A fully passive NFC-based flexible tag mounted on the inner face of an FFP2 mask powers a UV LED and reads luminescence from an opto-chemical CO₂ sensor using a digital color detector. Energy is harvested from a smartphone NFC field; an ultra-low-power MCU processes RGB data, computes the R coordinate and converts it to CO₂ concentration via a calibration curve. A temperature sensor enables temperature compensation. A custom Android app powers the tag, communicates via NFC, processes data, displays results, manages alerts, and allows sharing. Opto-chemical sensor design: The sensor uses La₂O₂S:Eu phosphor (bright emission at 625 nm, long decay) modulated by the pH indicator α-naphtholphthalein stabilized with TMAOH in a hydroxypropyl methylcellulose (HPMC) membrane, with EMIM BF₄ and Tween 20 to optimize interfacial properties. The UV LED (peak 367 nm) excites the phosphor; the indicator’s basic form absorbs at ~638 nm overlapping the phosphor emission, producing an inner-filter effect that decreases as CO₂ protonates the indicator, thereby increasing luminescence. Luminescence lifetime was ~0.33 ms across 0–100% CO₂, consistent with inner-filter mechanism. Membrane preparation: Cocktail: 1 mL HPMC 1% w/w in water + 450 µL ethanol solution with 6.6 mg α-naphtholphthalein + 5 µL Tween 20 + 2 µL EMIM BF₄ + 50 µL TMAOH 6.04×10⁻² M + 3 mg La₂O₂S:Eu. Spin coat on 125 µm PET (80 rpm 60 s, then 180 rpm to dry). Optical characterization with luminescence spectrometer; analytical relationship verified with Nakamura–Amao model. Flexible tag fabrication: Printed on 125 µm PET using Voltera V-One with flexible Ag ink; cured at 160 °C for 15 min; solder paste applied; components placed manually; reflow: soak 140 °C/45 s, peak 160 °C/30 s. NFC antenna planar coil: 5 turns, 43×58 mm, 400 µm track width, 600 µm spacing, optimized via ADS and COMSOL after Grover-based initial estimate. Final tag size 60×45 mm²; thickness limited by MCU to ~1.75 mm. Power consumption ~4.45 mW during operation (≈20% of 22.5 mW available from NFC IC). Characterizations: - Coupling and range: Using a Samsung Galaxy S7 NFC reader, maximum separation before power-off 19 mm; normalized coupling decreases with distance; >25% of max coupling required to power tag. Bending tests with cylinders (R=100, 50, 30 mm) along X and Y axes showed correct operation with reading distance >10 mm up to ~110° bending. - Calibration: The digital color sensor provides RGB; R coordinate is most sensitive. Calibration performed 500–48,000 ppm CO₂, 5 replicas at 25 °C. Fitted quadratic: [CO₂] (ppm) = 1.76093×10⁵·R² − 1.54898×10⁵·R + 2.75×10⁴ (R normalized; r²≈0.999). Resolution derived from fit: 103 ppm in the range of interest (worst case 221 ppm at 4.8% CO₂). LOD 140 ppm. Response/recovery below 1 s in this DSV-relevant range; for 0–100% CO₂ steps, response ~9 s and recovery ~30 s. - Temperature and humidity: Temperature drift of R compensated via on-tag sensor with linear coefficient a = (−15.6 ± 0.7)×10⁻³ K⁻¹. RH dependence tested 10–90% at 30 °C showed independence of R from humidity (no RH compensation required). - Ambient light: Tests from ~200 lx to 67 klx (indoor to direct sunlight) with tag inside white vs black FFP2 masks. Relative errors <3% indoors with white masks, <1% with black masks; outdoors, black masks required to keep errors <8% even in direct sun. - Sensor lifetime: In climate chamber (30 °C, 70% RH, 2% CO₂), readings every 30 min for 12 h. Relative error <3% up to ~5.5 h; increases afterwards; ~13% at 8 h (considered maximum allowable), establishing ~8 h sensor lifetime. Electronics reusable; membrane replaceable after disinfection (e.g., UV). Performance tests: - Short-term static test: Sitting, monitoring every 30 s; comparison with MultiRAE Lite reference (range up to 50,000 ppm, 100 ppm resolution). Good agreement; slight delay initially due to DSV filling; breath-hold apnoeas visible in both systems. - Long-term static test: Worn for 2.5 h; readings every 15 min; temperature logged. CO₂ around ~3% except during apnoeas at minutes 80 and 110; during a 2-min apnoea, high-resolution readings every 10 s showed clear CO₂ decrease and subsequent increase. - Graded cycling exercise: Instrumented facemask readings compared with cycling power and heart rate (HR). CO₂ in breathing zone increased by ~1.9% during exercise; CO₂ variations showed time-delayed correlation with power and HR; HR lagged CO₂ by ~3 min. Ethics approval obtained; single healthy male subject; non-clinical pilot demonstration. Cost and usability: Estimated large-scale production cost of tag <€5 (ICs dominate); sensing membrane cost <€0.01 (without mass production). Tag unobtrusive inside mask; PET presence increases DSV CO₂ by ~5% relative to tagless mask. Smartphone app developed in Android Studio (API 18+).

- Developed a flexible, battery-less NFC-powered opto-chemical CO₂ sensor integrated into FFP2 masks; smartphone app for powering, data processing, alerts, display, and sharing. - Analytical performance: Limit of detection 140 ppm; resolution 103 ppm in the range of interest (worst-case 221 ppm at 4.8% CO₂); calibration across 500–48,000 ppm with quadratic fit (r²≈0.999). Response/recovery <1 s in typical mask DSV CO₂ range; for 0–100% CO₂ steps, response ~9 s, recovery ~30 s. - Temperature compensation with linear coefficient −15.6×10⁻³ K⁻¹; independence from RH (10–90% at 30 °C). - Ambient light robustness: Indoors, relative errors <3% (white masks) and <1% (black masks); outdoors, black masks keep errors <8% even in direct sun. - Lifetime: Stable (<3% relative error) up to ~5.5 h; acceptable up to ~8 h (relative error ~13%), aligning with FFP2 NR recommended usage time. - Hardware: Tag size 60×45 mm², thickness ~1.75 mm; power consumption ~4.45 mW; maximum NFC read/power separation ~19 mm; reliable operation under bending up to ~110°; production cost <€5 per tag, membrane <€0.01. - Performance vs reference: Good agreement with a MultiRAE Lite CO₂ monitor during short-term static tests; in long-term static use (~2.5 h), mask DSV CO₂ ~3% with clear apnoea detection. - Exercise test: CO₂ in breathing zone increased by ~1.9% during graded cycling, consistent with literature; CO₂ changes correlated with power and HR, with HR lagging CO₂ by ~3 min.

The work demonstrates that accurate, real-time CO₂ monitoring inside mask DSV can be achieved with a compact, fully passive NFC-powered optical sensor, addressing the limitations of bulky, power-hungry traditional instruments. The selection of an inner-filter-based luminescent membrane provides high stability, low power operation, and immunity to humidity, while an on-board temperature sensor enables compensation of temperature drift. System-level design (ultra-low-power electronics, optimized NFC antenna, flexible PET substrate) yields a thin, comfortable tag with adequate read range and robustness to bending, suitable for real-world wear. Calibration and characterization indicate sensitivity and dynamic range appropriate for mask DSV CO₂ levels (typically 0.5–5%), with fast response in the relevant range. Ambient light studies clarify operational constraints (preference for black masks outdoors) ensuring reliable measurements in diverse environments. Pilot tests confirm agreement with a reference instrument and capture physiologically meaningful dynamics during rest and exercise (e.g., CO₂ elevation during apnoea and exercise; correlation with power and HR), illustrating utility for non-invasive health monitoring and potential preclinical/diagnostic applications. Compared to prior facemask sensors, the simultaneous combination of battery-less operation, flexible integration, and smartphone-based powering/communications is distinctive and facilitates scalability and user adoption.

A flexible, battery-less NFC-enabled opto-chemical platform was developed for real-time CO₂ monitoring inside FFP2 masks. The system achieves low LOD (140 ppm), high resolution (103 ppm), rapid response in the relevant concentration range, temperature-compensated accuracy, humidity independence, and practical robustness to ambient light with proper mask selection. The tag is thin, low-power, comfortable, and low-cost, with a sensor lifetime (~8 h) matched to FFP2 usage time. Pilot studies during daily activity and exercise demonstrate feasibility and alignment with reference measurements, supporting applications in non-invasive health monitoring, preclinical research, and diagnostics. Future work could include: formal clinical validation with larger cohorts; integration of multiple optical chemical sensors (e.g., O₂, other gases) on a single passive NFC tag; improved optical shielding to further mitigate ambient light; extended-lifetime membranes; and enhanced app features for exposure-time risk assessment and personalized alerts.

- Monitoring is not continuous without the smartphone present to power the NFC tag; measurements occur on demand when the phone is brought near the tag. - Ambient light can interfere outdoors; black masks are recommended to limit errors (<8% in direct sunlight). White masks are more suitable indoors. - Chemical sensor aging limits membrane lifetime to ~8 h; requires replacement for longer use (though electronics are reusable). - Preliminary evaluation with very small sample size (n=1 subject in performance tests); not a formal clinical trial. - Mechanical integrity could be affected by excessive bending/twisting due to rigid ICs on a flexible substrate. - Initial delays may occur due to DSV filling and membrane absorption/desorption kinetics, and temperature drift requires compensation. - Limited NFC power/read range (~19 mm threshold) constrains user-phone positioning during readout. - Slight increase (~5%) in DSV CO₂ caused by PET tag presence compared to tagless mask.