Frequently asked questions

The NZGS 2 is a selective metal-oxide (MOx) gas sensor from NANOZ. Your contact is TSR Messtechnik AG in Schaffhausen, Switzerland. Four sensing elements and two heaters sit on a die of 1.15 × 1.15 mm, housed in a QFN-8 package of 2.8 × 2.8 × 1.6 mm. Unlike ordinary MOx sensors it does not only register that a gas is present – it identifies which one. It is the basis of MoxAi V5.0 from TSR Messtechnik AG, which adds the read-out electronics and the trained classification model.

NZGS 2 combines two different sensitive layers – SnO₂ for volatile organic compounds and WO₃ for ozone, CO and NO₂ – across four sensing elements, driven by a triangular heater modulation between 1.8 and 2.2 V over a 6-second period. Each gas produces a distinct response pattern across the four sensors. AI classification of those patterns identifies the gas rather than only registering that a gas is present.

Seven features are extracted from each of the four sensor responses: triangle amplitude, area under the curve, rising and falling slope, peak-to-peak value, relative abundance, kurtosis and skewness. A deep learning model classifies the gas type and estimates concentration from them. Principal component analysis (PCA) separates the gas clusters visibly in feature space.

Heater voltage V_H sweeps triangularly between 1.8 and 2.2 V with a period of exactly 6 seconds; sensor voltage V_S sits between 0.7 and 1.0 V, with an optimum of 0.8 V. Period and voltage swing are not guidance values: if the modulation deviates, the response patterns no longer match the trained model and selectivity is lost.

Yes. It builds on patent US 2016/0238548 A1 – Heated sensitive layer gas sensor with multiple supply points for extended lifetime. The patent is held by CNRS and Aix-Marseille University; NANOZ holds an exclusive licence.

The main parameters at a glance:

The datasheet gives warranted limits of detection for five gases; the lowest is 25 ppb for acetone:

Beyond those, Hydrogen (H₂), Ammonia (NH₃), Methane (CH₄), Hydrogen sulphide (H₂S), Nitrogen dioxide (NO₂), Ozone (O₃), Nitrous oxide (N₂O), BTEX, Electrolyte vapours EMC, DMC, DEC are also detected. The datasheet gives no warranted limits for these – measured values and test conditions are provided per project on request.

Determined under controlled laboratory conditions: 24 h stabilisation, 25 °C, 45 % RH. The datasheet notes that these are typical characteristics which vary from chip to chip and depend heavily on the specific application.

Chip-to-chip calibration curves exist for the gases listed in the datasheet. The datasheet also states that these are typical characteristics which vary from chip to chip and depend heavily on the application. For an alarm threshold what counts is therefore not the catalogue figure but the reproducibility measured in your own gas matrix – which is what the evaluation kit is for.

In seven steps:

1. 24 hours of pre-heating on first power-up.
2. Record the baseline in clean reference air, at least 30 minutes.
3. Apply known gas concentrations using certified calibration gases.
4. Record the response of all four sensors at each concentration.
5. Build the calibration curve: sensor resistance against gas concentration.
6. Validate using PCA cluster separation.
7. Recalibrate every 6 to 12 months depending on the environment.

24 hours – and specifically at 2.3 V heater and 0.8 V sensor voltage, a higher heater voltage than in later operation. That is followed by at least 30 minutes of baseline stabilisation in clean reference air. Cutting the pre-heating short leaves the baseline drifting and makes classification unreliable.

At −40 to +125 °C and 10 to 95 % relative humidity, non-condensing. NZGS 2 is an ESD-sensitive component supplied on tape and reel – use ESD protective equipment when handling.

Silicone compounds. They poison the SnO₂ layer irreversibly – the sensor loses sensitivity for good and cannot be restored. Silicone-based sealants, adhesives, lubricants and release agents therefore have no place near the sensor, neither in operation nor during assembly. Condensation must also be avoided; operating humidity is 15 to 90 % RH, non-condensing.

The most common causes and how to check them:

• No response to any gas: check the heater supply (1.8–2.2 V) and confirm the 24-hour pre-heat completed.
• Baseline drifting upward: humidity change or contamination – allow 30 minutes to stabilise, check RH is within 15–90 %.
• Reduced sensitivity: ageing or poisoning – check operating hours, rule out silicone contact.
• Gases are confused with each other: check the modulation – the triangular sweep must be exactly 1.8–2.2 V at a 6 s period.
• High signal noise: inspect the QFN-8 solder joints; V_S must sit between 0.7 and 1.0 V.

Yes. An evaluation kit with read-out electronics and software is available, so you can measure the response in your own application before planning an integration. Enquiries to sales@nanoz.ai.

Because it buys warning time before a cell vents. Lithium-ion thermal runaway progresses through four stages:

• Stage 1 – electrolysis: hydrogen forms at cell level.
• Stage 2 – electrolyte vapour: EMC, DMC and DEC escape and are detectable by NZGS 2.
• Stage 3 – first venting: CO, H₂ and electrolyte gases spike.
• Stage 4 – thermal runaway: massive gas release, too late for safe intervention.

Hydrogen-only sensors typically alarm at stage 3. NZGS 2 detects stage 2 – the difference between an early warning and a damage report.

A hydrogen sensor measures a single quantity: H₂. Thermal-conductivity devices therefore only pick up the stage 3 venting event during thermal runaway. NZGS 2 evaluates a pattern across four sensing elements and covers 14 gases, including the stage 2 electrolyte vapours. At around 70 mW it also draws roughly one tenth of the power of heated single-gas sensors with comparable sensitivity.

NZGS 2 measures VOC signatures in exhaled breath. Relationships investigated in research include:

• Diabetes and hypoglycaemia: acetone – above 1.8 ppm in diabetic breath versus 0.3 to 0.9 ppm in healthy subjects.
• Lung cancer: hexanal, heptanal, octanal, nonanal, benzaldehyde.
• Breast cancer: alkanes C3–C12, benzene derivatives, ketones.
• COVID-19: octanal, nonanal, heptanal.
• Cystic fibrosis: carbonyl sulphide, dimethyl sulphide, acetoin.

These figures describe the state of research into VOC biomarkers and serve to frame possible applications. NZGS 2 is a sensor component, not an approved medical device, and does not replace a diagnosis.

Across five fields:

• Medical – real-time breath analysis, personal protection at the workplace.
• Automotive, aviation, greentech – thermal runaway early detection on EV batteries from vehicle to container, in-car air quality.
• Industry 5.0 – hydrogen safety on fuel cells, electrolysers and boilers, ammonia leak detection, H₂S monitoring in oil and gas.
• Air quality – indoors formaldehyde, BTEX and CO from building materials and combustion; outdoors NO₂, O₃ and CO, from which a combined air quality index can be calculated.
• Food and agriculture – ethylene for fruit ripeness, spoilage detection via shifts in the VOC signature, cold chain integrity monitoring.

The website nanoz.ai is operated by TSR Messtechnik AG in Schaffhausen, Switzerland. Address: Ebnatstrasse 164, CH-8207 Schaffhausen. Phone +41 41 510 23 05, e-mail sales@nanoz.ai.

On request by e-mail to sales@nanoz.ai. The full datasheet covers pinout, electrical specifications, test conditions and calibration curves. Investors can request the pitch deck via the investors page.