A research team at the Shanghai Institute of Microsystem and Information Technology (SIMIT), part of the Chinese Academy of Sciences, has developed a flexible gas sensor that combines wood-derived cellulose nanopaper with the organic semiconductor F8BT. The sensor responds within 2 seconds to diethyl chlorophosphate (DCP), a simulant for nerve agents such as sarin. Its theoretical detection limit is 0.65 ppb, and the lowest concentration at which a concentration-dependent response was actually confirmed was 1.68 ppb. The team also succeeded in reading the detection signal contactlessly, without a battery, by combining the sensor with a commercial near-field communication (NFC) tag.
The results were published online in Advanced Functional Materials on November 13, 2026. SIMIT also presented the research on November 19. Research on repurposing commercial NFC tags as gas sensors has developed from work such as the technology MIT announced in 2014. In this study, a paper-like substrate serves as part of the detection mechanism, combining a highly sensitive chemiresistive sensor with passive wireless communication.
How the F8BT–nanopaper interface changes resistance
The institute explains that detecting nerve agents has traditionally required large analytical instruments and complex sample preparation, making it poorly suited to continuous on-site monitoring. Chemiresistive sensors, by contrast, rely on the change in electrical resistance that occurs when a target substance is adsorbed, so they are simple in structure and consume little power. Challenges remain, however, in selectively detecting specific molecules, in the efficiency of charge transfer between materials, and in stability against changes in temperature and humidity.
The sensor the team developed consists of three main layers. The base is a thin, flexible sheet called wood-derived cellulose nanopaper (WCN). On top of it, interdigitated gold electrodes, with two sets of comb teeth arranged alternately, are formed, and a sensing layer of the conjugated polymer F8BT is laid over them. According to the paper's abstract, the WCN provides a lightweight, flexible substrate and a stable interfacial environment, while F8BT handles molecular recognition and charge transfer. The heterointerface where the two materials meet greatly improves detection performance for DCP, the abstract says.
The institute also describes the roles of the materials at this interface. Nitrogen and sulfur atoms in F8BT act as active sites that capture DCP molecules, while oxygen-containing functional groups and sodium ions in the WCN regulate charge transfer at the interface. In other words, the cellulose nanopaper is not merely a support but a functional material directly involved in the sensor's electrical response.
When DCP molecules reach the F8BT sensing layer from the air, their adsorption alters charge transfer at the interface between F8BT and WCN. This changes the electrical resistance between the interdigitated electrodes, so the presence of DCP can be detected as an electrical signal.
Theoretical detection limit of 0.65 ppb versus a measured range of 1.68 ppb to 132 ppm
In the team's measurements, exposure to 132 ppm of DCP produced a resistance change of about 74% with a response time of 2 seconds. These values are consistent between the paper's abstract and the institute's announcement.
The theoretical detection limit of 0.65 ppb must be distinguished from the 1.68 ppb lower bound at which concentration dependence was actually confirmed. A detection limit generally refers to the minimum concentration at which a measured signal can be distinguished from noise. The institute explicitly calls 0.65 ppb a "theoretical detection limit," but neither the announcement nor the abstract describes how it was calculated. The range over which the sensor's response was checked while varying concentration runs from 1.68 ppb to 132 ppm, and a result actually measured at 0.65 ppb cannot be confirmed from these materials.
This measured range is very wide. 132 ppm equals 132,000 ppb, about 79,000 times the lower bound of 1.68 ppb. The team therefore confirmed a DCP concentration-dependent response across roughly five orders of magnitude. However, the abstract only describes "good concentration dependence," so it is unclear whether concentration and resistance change are linearly related across the whole range.
The institute has also published several tests of environmental durability. The sensor reportedly maintained a stable response in environments with 40–90% relative humidity, through 500 bending tests, and over a five-week long-term test. The abstract likewise reports stability against mechanical deformation, humidity changes, and long-term storage, but gives no specific test conditions. The team also says selectivity and reproducibility were good, but the types of interfering gases and the test concentrations cannot be confirmed from the institute's announcement or the abstract.
Reading resistance changes wirelessly with a battery-free NFC tag
The team also built a prototype system in which the sensor is connected to the resonant circuit of a commercial NFC tag so the detection signal can be read wirelessly. NFC is a short-range communication technology that uses electromagnetic induction. When DCP changes the sensor's electrical resistance, the circuit characteristics of the NFC tag change too, and that change can be read as a high-frequency reflected signal.
The system uses a passive wireless approach with no onboard battery. The tag operates on electromagnetic waves sent from the reader and returns the detection signal. From the reader's perspective, the tag's resonant circuit weakens the reflected signal near a particular frequency. The dip that appears when reflection strength is plotted against frequency is called the resonance dip.
When the sensor's resistance changes, the depth of this resonance dip changes. According to the institute, after exposure to DCP the resonance dip in the reflection coefficient (S_{11}) became clearly deeper, while the resonance frequency barely changed. The resistance change caused by DCP can therefore be read by tracking the depth of the dip rather than a shift in resonance frequency.
The abstract positions this result as a proof-of-concept demonstration that integration with wireless systems is possible. While the resistance change caused by DCP was actually read without contact, neither the institute's announcement nor the abstract gives the specifications of the reader used or the reading distance. It is also not stated whether a typical smartphone could read the same signal.
Wireless gas sensor research since MIT's 2014 work
Repurposing commercial NFC tags as gas sensors did not originate with this study. In December 2014, Joseph M. Azzarelli, Timothy M. Swager, and colleagues at MIT reported in Proceedings of the National Academy of Sciences (PNAS) a technology that modifies commercial NFC tags into chemical sensors so a smartphone can detect gases. The authors described it as the first application of NFC to chemical sensing of gases. That work targeted ammonia, hydrogen peroxide, cyclohexanone, and water, among others, and did not include DCP.
In 2016, Swager's group published two further studies on wireless sensors for nerve agent simulants. In a paper in the Journal of the American Chemical Society (JACS) published online in June, they developed a dosimeter-type sensor that detects trace amounts of DCP and similar substances, using a mechanism in which a supramolecular polymer wrapped around carbon nanotubes dissociates upon reaction with the target. They also demonstrated wireless power delivery via NFC and readout by smartphone. The authors included, alongside the MIT researchers, Shinsuke Ishihara of the International Center for Materials Nanoarchitectonics (MANA) at the National Institute for Materials Science (NIMS).
In July of the same year, MIT's Rong Zhu, Azzarelli, and Swager reported a disposable wireless badge for detecting DCP in Angewandte Chemie International Edition. It was a chemical dosimeter-type sensor using carbon nanotubes and an ionic liquid, with a DCP detection limit of 28 ppb. In this approach, exposure to DCP changes the electrical resistance irreversibly, recording the cumulative amount of exposure. The nature of its detection signal differs from that of the SIMIT sensor.
Highly sensitive sensors for nerve agent simulants are also being developed elsewhere in China. In September 2022, a team including Northeast Forestry University reported a fluorescent sensor using a thin film of conjugated microporous polymer in Nature Communications. Its DCP detection limit was 2.5 ppt, and the team also built a prototype portable detector with wireless communication.
In January 2024, a team including the Institute of Acoustics of the Chinese Academy of Sciences reported a passive wireless surface acoustic wave (SAW) sensor operating in the 433 MHz band in Microsystems & Nanoengineering. It targets not DCP but dimethyl methylphosphonate (DMMP), another nerve agent simulant. Using a SAW approach rather than NFC, it measured concentration-dependent signals between antennas 20 cm apart.
The references cited in the SIMIT paper include MIT's 2016 wireless badge and the 2022 Northeast Forestry University study. Comparing these earlier studies with the new result shows that, even for the same target DCP, the sensor type, the nature of the response, and the definition of the detection limit differ.
| Study (publication date) | Detection method | Nature of response | DCP detection limit | Definition / measurement conditions |
|---|---|---|---|---|
| SIMIT (Nov. 2026) | Chemiresistive, F8BT and WCN | Reported as repeatable | 0.65 ppb | Theoretical detection limit; measured concentration-dependence range 1.68 ppb–132 ppm |
| MIT (Jul. 2016) | Chemical dosimeter, carbon nanotubes | Irreversible, cumulative detection | 28 ppb | Value stated in abstract |
| Northeast Forestry University (Sep. 2022) | Fluorescence, conjugated microporous polymer | Not stated in abstract | 2.5 ppt | Value stated in abstract |
Note: All values are detection limits for DCP vapor. Because detection methods and calculation methods differ, performance cannot be ranked by the numbers alone.
Converted to the same unit, SIMIT's theoretical value is 0.65 ppb, MIT's 2016 wireless badge is 28 ppb, and Northeast Forestry University's fluorescent sensor is 0.0025 ppb. However, MIT's 2016 sensor is a dosimeter that irreversibly records exposure, so its use differs from that of the new chemiresistive sensor. In addition, SIMIT's 0.65 ppb is a theoretical detection limit, and the lowest concentration at which dependence was measured is 1.68 ppb. All three results are for the simulant DCP, not values obtained with actual nerve agents such as sarin.
Do simulant results apply to actual nerve agents?
How far experiments with simulants can predict performance against real nerve agents is an important question in sensor research.
In a study by Valentin A. Bureau and colleagues published online in The Analyst in April 2025, the researchers compared actual sarin with representative simulants under the same conditions and examined the responses of colorimetric sensors. They concluded that reactivity to simulants cannot reliably predict reactivity to actual sarin and that evaluation with the real agents is necessary. That study addressed colorimetric methods that rely on color change, so its results cannot be applied directly to the new chemiresistive sensor. Still, it serves as a reference showing that high sensitivity to DCP does not mean equivalent performance against sarin.
The only detection target confirmed in SIMIT's announcement and the paper's abstract is DCP, and no test results using actual sarin or other agents are given.
In Japan, too, research on nerve agent detection is advancing. In June 2026, a team from the National Research Institute of Police Science and Yokohama City University announced a new mass spectrometry method for detecting nerve agents such as VX and Novichok. The study developed a way to detect trace amounts of nerve agents in samples wiped from object surfaces. While sarin volatilizes readily and spreads as vapor, VX and Novichok are less volatile and tend to remain on surfaces. The sensor characteristics required differ depending on the target and the setting.
SIMIT positions the result as a new design approach for monitoring hazardous chemicals with portable devices and sounding on-site alarms. Its distinguishing features are the use of a flexible, wood-derived substrate as part of the detection mechanism and its combination with a battery-free NFC tag. Once the sensitivity and selectivity against actual nerve agents, the wireless reading distance, and the conditions used to calculate the theoretical detection limit are clarified, its performance as a field detection device can be evaluated more concretely.
