Traditional strategies for fabricating SMO-based gas sensors often rely on separate steps for material synthesis, followed by deposition onto device substrates. Wet-chemical methods like hydrothermal and sol-gel techniques, while able to produce high-performance nanomaterials, struggle with spatial selectivity and reproducibility necessary for wafer-level sensor arrays.
Physical and chemical vapor deposition methods offer good control of film thickness and composition but generally yield dense films with low porosity, which limit gas interaction and reduce sensitivity. To address these challenges, there is a critical need for a scalable, spatially selective deposition approach that produces porous and uniform metal oxide films ideal for gas sensing.
Spark-Ablation Printing Technique
The researchers developed a spark-ablation printing technique, which combines high-voltage electric discharges and aerosol deposition to form metal oxide nanoparticle films directly on MEMS-based microheater substrates. This solvent-free method vaporizes metal target electrodes (such as Sn, Zn, Ni, and Au) by spark discharges in an oxygen-argon atmosphere, creating atomic and cluster species. These are rapidly quenched and diluted, leading to nucleation and growth of nanoparticles that aggregate into nanoporous films.
Using this approach, the team fabricated SMO films including pristine SnO2, ZnO, NiO, and noble-metal-functionalized SnO2 (e.g., Au-decorated). The films were deposited onto MEMS microheater chips under controlled conditions, with parameters such as spark current, gas composition, and printing speed optimized to achieve porous films with sizes in the 1–20 nm range. Sensors were electrically characterized for resistance and gas sensitivity. Their sensing performance was evaluated under various target gases (NO2, H2S, NH3, H2) across concentration ranges down to parts-per-billion (ppb) levels.
Sensor Performance & Machine Learning Integration
The spark-ablation-printed SnO2 sensors demonstrated excellent uniformity across arrays, with baseline resistances showing low variation (coefficient of variation ~0.13), confirming the high reproducibility achievable by this method. These sensors exhibited high sensitivity and selectivity to nitrogen dioxide (NO2), with response magnitudes increasing almost linearly with concentration (R2 ≈ 0.99) in the range of 0.5–5 ppm.
Notably, Au decoration on SnO2 films introduced a significant improvement: a sensor with 7.8 wt% Au showed an 11-fold increase in response to 2 ppm NO2 compared to pristine SnO2, along with enhanced selectivity against interfering gases such as H2S, NH3, and H2.
In terms of detection limits, the Au-functionalized SnO2 sensor reached an ultra-low limit of detection (LOD) of 0.11 ppb for NO2, markedly lower than 1 ppb for bare SnO2 sensors, illustrating the catalytic and electronic benefits of Au nanoparticles, facilitating charge transfer and oxygen activation at the interface. Additionally, Au decoration dramatically shortened recovery times (25 s vs. 546 s), beneficial for practical rapid sensing applications. These improvements highlight the role of noble metals in enhancing SMO sensor dynamics.
The versatility of spark-ablation printing was further demonstrated by fabricating sensors from ZnO (n-type) and NiO (p-type). Both exhibited sensitive responses to hydrogen sulfide (H2S), a toxic gas, across a wide range of concentrations down to 10 ppb. ZnO sensors showed concentration-dependent Langmuir adsorption behavior (R2=0.999) and fast response times (~43 s), while NiO sensors exhibited strong linear response dependence (R2=0.98) with response times near 99 s.
Download the PDF of this page here
Crucially, the researchers exploited the capability to print multi-material oxide sensor arrays by combining SnO2, ZnO, and NiO sensors on a chip. Employing machine-learning algorithms, specifically random forest classifiers, they demonstrated near-perfect identification accuracy (>99%) for four gases: NO2, H2S, NH3, and H2.
Data visualization using t-SNE and radar plots revealed well-separated clusters for each gas, confirming that complementary sensor responses improve selective recognition. Minimal misclassifications occurred mostly between H2 and NH3, reflecting known cross-sensitivity issues but still indicating excellent performance overall.
Scalable Fabrication for Intelligent Gas Sensing
This study presents a novel spark-ablation printing technique that unifies material synthesis and device fabrication into a streamlined, one-step process for producing high-performance SMO gas sensors. The method enables direct printing of porous, nanostructured metal oxide films - including noble-metal-decorated heterostructures - with high uniformity and reproducibility onto MEMS microheater chips. The fabricated sensors exhibit ultra-low detection limits at the ppb level, fast response and recovery times, and strong selectivity toward target gases such as NO2 and H2S.
Moreover, multi-material sensor arrays printed via this approach, when combined with robust machine-learning models, deliver near-perfect gas classification across multiple analytes, showcasing potential for intelligent, portable gas-sensing platforms. Overall, spark-ablation printing constitutes a scalable, versatile route for fabricating next-generation electronic noses and chemiresistive sensors with improved reliability, making it highly promising for widespread environmental monitoring, industrial safety, and healthcare diagnostics.
Journal Reference
Fu W., Tang Z., et al. (2026). Scalable fabrication of gas sensors via spark-ablation printing of semiconductive metal oxide nanoparticles and heterostructures. Microsystems & Nanoengineering 12, 141 (2026). DOI: 10.1038/s41378-026-01208-1, https://www.nature.com/articles/s41378-026-01208-1