gas sensing

Gas sensing is the detection and measurement of gaseous molecules in the surrounding environment, and it underpins everything from the carbon monoxide alarm on your ceiling to experimental breath tests that may one day screen for cancer. The field has grown rapidly in the past decade, driven by advances in nanomaterials, shrinking electronics, and machine learning. What makes gas sensing both fascinating and tricky is that the challenge is rarely just “can we detect this gas?” but rather “can we detect it reliably, cheaply, selectively, and fast enough to be useful in the real world?”

How Gas Sensors Actually Work

Most gas sensors fall into a handful of broad families, each exploiting a different physical or chemical principle. The most widely used type is the chemiresistive sensor, which relies on a thin layer of semiconducting metal oxide. When gas molecules land on the oxide surface, they trigger chemical reactions that change the material’s electrical resistance. Measure that resistance shift, and you know something about the gas. The sensitivity of these sensors depends heavily on the chemical makeup of the sensing layer, its microstructure, the operating temperature, and even the humidity in the air.

1MDPI Sensors / Europe PMC. Metal oxide gas sensors: sensitivity and influencing factors

Optical sensors take a completely different approach. Instead of measuring electrical changes, they shine infrared light through a gas sample and measure how much light gets absorbed at specific wavelengths. This is the principle behind nondispersive infrared (NDIR) sensors, which are workhorses for carbon dioxide monitoring in buildings, greenhouses, and industrial settings. Recent engineering has compressed NDIR sensors to remarkably small packages. One design measures just 24 mm long and can detect CO₂ concentrations from 0.5% to 20% across a working temperature range of −20 °C to 60 °C, at roughly a third the size of conventional sensors with similar accuracy.

2ScienceDirect (Infrared Physics & Technology). Development of a compact NDIR CO2 gas sensor for harsh environments

Electrochemical sensors use a different trick: gas molecules diffuse through a membrane into an electrolyte, where they undergo a chemical reaction at an electrode, producing a tiny electrical current proportional to the gas concentration. These are the sensors inside many portable detectors used by firefighters and industrial workers for toxic gases like carbon monoxide and hydrogen sulfide. A fourth family, acoustic sensors, measures tiny mass changes on a vibrating surface. When gas molecules stick to a coated quartz crystal, the added weight shifts the crystal’s resonant frequency by a measurable amount, allowing detection of very small quantities of gas.

Why Nanomaterials Changed Everything

Traditional metal oxide sensors work well enough for many applications, but they typically need to be heated to a few hundred degrees Celsius to function, which makes them power-hungry and limits where you can put them. Nanomaterials have opened the door to sensors that operate at or near room temperature, which is a game-changer for portable and wearable devices.

Graphene and MXenes (a class of two-dimensional materials made from layered metal carbides) are among the most promising candidates. Hybrid fibers combining MXene with graphene have demonstrated strong ammonia sensitivity at room temperature, with a response about an order of magnitude better than either material alone.

3PubMed. Room-Temperature, Highly Durable Ti3C2Tx MXene/Graphene Hybrid Fibers for NH3 Gas Sensing

Researchers have also used laser processing to create crumpled graphene/MXene nanocomposite foams that serve as both stretchable gas sensors and self-charging power units in a single integrated system. These sensors achieved detection limits down to about 5 parts per billion for nitrogen dioxide and ammonia, which is sensitive enough to monitor exhaled breath or local air quality in real time.

4PubMed. Laser Processing of Crumpled Porous Graphene/MXene Nanocomposites for a Standalone Gas Sensing System

Another class of materials gaining traction is metal-organic frameworks, or MOFs. These are porous crystalline structures that can be tuned at the molecular level to selectively trap specific gases while ignoring others. Think of them as molecular sieves with customizable pore sizes. MOF-based sensors have shown elevated sensitivity and selectivity for various target gases, thanks to their ability to preconcentrate gas molecules inside those pores.

5PubMed Central. Metal-Organic Framework Based Gas Sensors

For example, a MOF-based capacitive sensor built on nanofibers demonstrated detection of trace sulfur dioxide down to 0.1 parts per million, enabled by abundant hydroxyl groups on the material’s surface that improve SOâ‚‚ adsorption.

6Applied Surface Science. MOFs/nanofiber-based capacitive gas sensors for the highly selective and sensitive sensing of trace SO2

The Selectivity Problem

Sensitivity gets the headlines, but selectivity is often the harder engineering challenge. A sensor that responds strongly to a target gas is not very useful if it also responds to three other gases in the environment. Most metal oxide sensors, for instance, respond to a broad range of reducing or oxidizing gases. If you are trying to detect a trace amount of formaldehyde in a room that also contains ethanol, carbon monoxide, and varying humidity, your sensor needs to tell all of those apart.

Humidity is a particularly stubborn interferer. Water molecules interact with the same oxide surfaces that target gases use, which can shift the baseline reading or mask a genuine detection event. One approach to mitigating this involves placing a desiccant like calcium chloride in the gas flow path upstream of the sensor. Researchers showed that this simple addition pushed detection accuracy above 90% even when relative humidity varied widely, without needing to train the system across every humidity condition.

7Sensors and Actuators B: Chemical. Reducing the destructive effect of ambient humidity variations on gas detection capability of a temperature modulated gas sensor by calcium chloride

The more ambitious solution to cross-sensitivity is to stop asking a single sensor to do all the work and instead deploy an array of sensors, each with slightly different responses. The combined response pattern across the array creates a kind of chemical fingerprint. Pattern recognition algorithms then learn to distinguish gas species and concentrations from these fingerprints, much like a human nose uses hundreds of different receptor types to identify thousands of odors.

8PubMed Central. Overcoming the Limits of Cross-Sensitivity: Pattern Recognition Methods for Chemiresistive Gas Sensor Array

Electronic Noses and the Role of Machine Learning

The sensor-array-plus-algorithm approach has a catchy name: the electronic nose, or e-nose. It has been a concept since the 1980s, but recent progress in deep learning has given it new teeth. A modern e-nose typically pairs a handful of chemically distinct sensors with a trained classification model. In one study optimizing a sensor array using multiple deep learning architectures, a convolutional neural network achieved a classification accuracy of 0.986 and a concentration-prediction R² of 0.979. Even when the array was cut from eight sensors to four, the accuracy stayed above 0.9 for classification and 0.97 for concentration prediction, suggesting that smart algorithms can compensate for hardware simplicity.

9Sensors and Actuators B: Chemical. Sensor array optimization for the electronic nose via different deep learning methods

Researchers have also built e-noses from electrospun metal oxide nanofibers, using arrays of four different semiconductor oxides (tin oxide, copper oxide, indium oxide, and zinc oxide) to discriminate five different volatile organic compounds relevant to human health. After extracting features from the sensor response curves and applying dimensionality-reduction algorithms, the system successfully separated all five gases.

10Frontiers in Sensors. Electronic nose based on multiple electrospinning nanofibers sensor array and application in gas classification

A more biologically inspired approach programs sensor surfaces with DNA base sequences to mimic the molecular recognition of biological olfactory receptors. One group built a biosensor array on MXene surfaces, using different DNA sequences to create varied binding affinities, and then applied machine learning for pattern recognition aimed at noninvasive cancer detection from breath.

11PubMed. DNA-Mediated Bioinspired MXene Gas Sensor Array with Machine Learning for Noninvasive Cancer Recognition

Breath Analysis and Medical Diagnostics

Your exhaled breath contains hundreds of volatile organic compounds and inorganic gases in trace amounts, and changes in this chemical cocktail can reflect what is happening inside your body. The idea of breath-based disease detection is ancient (physicians have noted the fruity breath of uncontrolled diabetes for centuries), but modern gas sensors are making it quantitative. Gas biomedical sensors are being developed as noninvasive, point-of-care alternatives to blood draws and urine tests, which are uncomfortable and require lab processing.

12PubMed Central. Metal oxide-based gas sensors for the detection of exhaled breath markers

Wearable breath sensors take this further, aiming for continuous, real-time monitoring of disease-related gas biomarkers. The vision is a small device worn near the mouth or integrated into a mask that tracks volatile compound levels over hours or days, potentially catching disease flare-ups in conditions like asthma, chronic obstructive pulmonary disease, or metabolic disorders before symptoms worsen.

13Advanced Materials Technologies. Wearable Breath Gas Sensors for Chronic Disease Management: Challenges and Perspectives

The gap between laboratory proof-of-concept and a clinically validated device remains wide, however. Breath composition varies with diet, medications, time of day, and hydration. A sensor that can detect acetone in a controlled lab chamber does not necessarily perform well when someone has just eaten garlic bread and drunk coffee. Controlling for these confounders in clinical trials is one of the field’s major hurdles, and the regulatory pathway for breath-based diagnostic devices is still being defined.

Food Safety and Agriculture

When food spoils, it emits characteristic gases. Decomposing protein releases biogenic amines and sulfur compounds. Ripening fruit produces ethylene and various esters. Gas sensors can detect these markers in real time, offering a faster and cheaper alternative to laboratory microbiological testing. The appeal is obvious for the food supply chain: a sensor embedded in packaging or held near a crate of fish at a distribution center could flag contamination before it reaches your kitchen.

14Journal of Agriculture and Food Research. Advances in food spoilage detection through gas sensing: From material insights to prospects for commercialisation

E-nose systems have already demonstrated striking results in food quality testing. One study built a sensor array to detect tuna contaminated with Pseudomonas aeruginosa, a common spoilage bacterium. Using a support vector machine classifier, the system distinguished healthy from contaminated fish with 99% accuracy, capturing 95% of the variance in the gas patterns with just the first two principal components.

15PubMed Central. Gas Array Sensors based on Electronic Nose for Detection of Tuna (Euthynnus Affinis) Contaminated by Pseudomonas Aeruginosa

In agriculture more broadly, gas sensors can monitor greenhouse COâ‚‚ levels to optimize plant growth, detect ethylene buildup in fruit storage facilities to manage ripening, or sniff out early signs of grain mold in silos. The technology is not yet ubiquitous in these settings, mainly because of cost and durability concerns, but the direction of travel is clear.

Hydrogen Safety in the Energy Transition

As hydrogen gains traction as a clean fuel, the need for reliable hydrogen leak detection grows with it. Hydrogen is odorless, colorless, and explosive at concentrations above about 4% in air. Palladium-based chemiresistive sensors are among the most promising detection systems for hydrogen, valued for their process simplicity and low cost.

16International Journal of Hydrogen Energy. Hydrogen sensing towards palladium-based nanocomposites: A review

Palladium absorbs hydrogen atoms into its crystal lattice, swelling slightly and changing its electrical resistance. This gives it a natural affinity for the job. The engineering challenge is making palladium-based sensors fast enough and durable enough for real deployment. Nanostructured palladium designs, which maximize the surface area available for hydrogen contact, have been identified as a common denominator in the best-performing research prototypes, and design rules for these nanostructures are becoming clearer.

17PubMed Central. High-Performance Nanostructured Palladium-Based Hydrogen Sensors-Current Limitations and Strategies for Their Mitigation

Speed matters especially here. A hydrogen leak in a fueling station or pipeline needs to be caught in seconds, not minutes. The most stringent performance targets demand response times under one second and detection at concentrations well below the explosive threshold. Meeting all of these targets simultaneously, fast response, low detection limit, long-term stability, and resistance to poisoning by other gases, remains an active area of research.

Wearable and Flexible Sensors

The push toward wearable electronics has driven significant interest in gas sensors that can bend, stretch, and conform to the body or clothing. Flexible transparent sensors have been built from semiconducting oxides, carbon materials, and conducting polymers deposited on plastic or fabric substrates. The potential applications range from personal air-quality monitors to health-tracking patches that detect disease biomarkers in sweat vapor or breath.

18PubMed. Flexible Transparent Electronic Gas Sensors

For wearable devices, power consumption becomes critical. You cannot strap a sensor that needs hundreds of milliwatts to someone’s wrist and expect the battery to last a day. One approach integrates gas-sensitive metal oxide nanowires directly onto micro light-emitting diodes, using light rather than heat to activate the sensing surface. The smallest version of this design had an active area of just 30 by 30 micrometers and operated on roughly 184 microwatts of power, yet showed strong sensitivity to nitrogen dioxide at concentrations as low as 1 part per million.

19PubMed. Monolithic Micro Light-Emitting Diode/Metal Oxide Nanowire Gas Sensor with Microwatt-Level Power Consumption

That level of power consumption is low enough to run off a small coin cell or even a harvested energy source, which makes truly autonomous wearable gas sensors plausible rather than aspirational.

Scaling Up and Fitting In

Making a sensor work in a lab is one thing; manufacturing millions of them at reasonable cost is another. The semiconductor industry already has a mature fabrication infrastructure built around CMOS processes, and there is strong interest in making gas sensors compatible with that existing pipeline. Conventional metal oxide sensors complicate this because they require high temperatures and integrated microheaters, which adds complexity and cost to the fabrication.

20PubMed Central. Application of Two-Dimensional Materials towards CMOS-Integrated Gas Sensors – Section: Abstract

Two-dimensional materials like graphene, MXenes, and transition metal dichalcogenides are attractive partly because many of them function at or near room temperature, potentially eliminating the need for microheaters altogether. If these materials can be deposited reliably on standard silicon wafers using processes compatible with existing chip fabrication, gas sensors could ride the same cost curve that made microprocessors cheap. That is still an “if,” though. Uniformity of 2D material deposition over large wafer areas, batch-to-batch consistency, and long-term stability under real-world conditions all need more work before mass production is routine.

Calibration and Drift in the Real World

A freshly calibrated gas sensor can be impressively accurate. The harder question is whether it stays accurate over months and years of continuous exposure to real air. Sensor drift, the gradual shift in baseline readings caused by aging, contamination, and environmental changes, is a persistent headache. Low-cost sensors for air quality monitoring have been evaluated against laboratory reference instruments, and the results are encouraging but qualified. In one evaluation, low-cost CO, ozone, and SOâ‚‚ sensors showed good linearity, precision above 95%, response times under two minutes, and accuracy above 89% after calibration, with no significant baseline drift over the evaluation period.

21Elsevier / Atmospheric Pollution Research. System for performance evaluation and calibration of low-cost gas sensors applied to air quality monitoring

The catch is that “after calibration” does a lot of heavy lifting in that sentence. Calibration requires access to known reference gas concentrations or co-location with expensive reference instruments, and it may need to be repeated periodically. For consumer devices like smart home air quality monitors, the calibration step is often hidden from the user or done at the factory, with limited ability to correct for drift over the product’s lifespan. This is one reason air quality readings from low-cost consumer monitors should be treated as approximate indicators rather than regulatory-grade measurements.

Where Gas Sensing Is Headed

Several threads are converging to push gas sensing into places it has never been before. Sensors are getting smaller and cheaper, nanomaterials are enabling room-temperature operation, machine learning is compensating for the imperfect selectivity of individual sensors, and wearable form factors are moving detection out of fixed installations and onto people. The combination of self-powered sensor systems, such as the stretchable graphene/MXene devices that integrate gas sensing with energy harvesting, hints at a future where environmental and health monitoring runs continuously without plugging in or changing batteries.

Food packaging with embedded freshness sensors, personal air quality badges for urban commuters, breath-analysis patches for chronic disease management, and distributed hydrogen leak networks across the emerging hydrogen economy are all plausible near-term applications. The remaining barriers are mostly about engineering maturity rather than fundamental science: making these sensors last years instead of months, ensuring they stay calibrated in unpredictable conditions, manufacturing them at a price point that allows widespread deployment, and navigating the regulatory frameworks that govern medical and environmental monitoring devices. The sensing materials themselves have largely proven they can do the job; the rest is about making them do it reliably, at scale, outside the controlled comfort of a laboratory.