Optimizing Pd/SnO₂ Nanomaterial Synthesis for High‑Performance Hydrogen Gas Sensors

Background

Hydrogen is a key feedstock in chemical synthesis and a clean energy carrier, yet its flammability demands stringent monitoring during production, transport, and storage [1–3]. Metal‑oxide sensors offer a promising platform for real‑time H₂ detection [4–6].

Nanostructured SnO₂ possesses desirable attributes—chemical inertness, thermal stability, and robust O₂ chemisorption—making it an attractive material for hydrogen sensing [7–9]. Sensitivity can be further amplified by incorporating catalytic Pd, a well‑known facilitator of H₂ oxidation [6,10].

The physicochemical properties of the sensing layer depend heavily on the synthesis route, composition, and sintering conditions, which govern particle size and morphology—critical factors for sensor efficacy [11–18].

While high‑temperature sintering is required for mechanical integrity and electrical conductivity, it tends to promote particle growth. Optimizing sintering schedules can suppress coarsening while maintaining robust nanoparticle contacts [18].

Our study focuses on how varying Pd loading and sintering protocols affect Pd/SnO₂ nanomaterial characteristics and, consequently, hydrogen sensor performance.

Methods

Synthesis of Nanosized Tin Dioxide

SnCl₄·5H₂O (1.5 g) was dissolved in 15 mL ethylene glycol, evaporated at 110–120 °C to yield a dark brown gel, then dried at 150 °C to form a xerogel. Thermal decomposition in air produced ~8 nm SnO₂ particles. The resulting nanomaterial, carboxymethyl cellulose, and PdCl₂ were blended to prepare the sensing layer.

Preparation of Adsorption‑Semiconductor Sensors

A 3 μL paste (SnO₂ + 3 wt% CMC) was deposited on a ceramic substrate with integrated contacts and heater [19]. The paste yielded a ~70 μm thick layer (SEM data). Sensors were dried at 90 °C for 1 h. Palladium impregnation was performed with PdCl₂ solutions (0.05 × 10⁻²–0.15 M). Post‑impregnation, sensors were sintered using two distinct temperature profiles (Fig. 1). Sensors sintered via Mode 1 were labeled S₁; those via Mode 2 were labeled S₂.

Schemes of temperature heating of the sensors based on SnO₂. a Mode 1. b Mode 2

Methods of Measurement

Sensors were tested in 400 mL/min gas streams within a dedicated chamber. Heater voltage controlled operating temperature, verified by an Optris Laser Sight pyrometer. Prior to measurement, sensors were aged at 400 °C for 1 week, with periodic exposure to 1000 ppm H₂.

Sensor response was defined as R₀/R_H₂, where R₀ is resistance in air and R_H₂ is resistance in 40 ppm H₂. Response time (t₀.₉) and recovery time (τ₀.₁) were measured by the time to reach 90 % of the equilibrium signal upon gas exchange.

Gas mixtures containing H₂, CO, CH₄, and their combinations were prepared at the Ukrainian Centre of Certification and Metrology to assess selectivity.

Long‑term stability was evaluated by measuring the 40 ppm H₂ response over 6 months for sensors S₂ (S‑67, S‑69).

Specific surface area was measured by BET. Palladium content was quantified via atomic absorption (AAS1N Carl Zeiss) using an acetylene‑air flame. Phase composition was characterized by XRD (Bruker D & Advance, Cu Kα). TEM (SELMI PEM‑125 K, 100 kV) and FESEM (JEOL JSM‑6700F) provided morphological details; HRTEM (JEM‑2100F) resolved interparticle boundaries. Layer thickness was determined by SEM (JEOL JSM‑6060LA).

Results and Discussion

The synthesized ~8 nm SnO₂ nanomaterial served as the base for all sensors. Comparison of sintering protocols revealed that Mode 1 (590–620 °C, 180 min) produced particles averaging 17 nm (Fig. 2a) with a sensor response of 6.7. Shortening the dwell time to 80 min at 590 °C drastically increased resistance (>500 MΩ) and reduced response (~2), indicating insufficient interparticle contact.

TEM images of the gas‑sensitive nanomaterials a S₁ and b S₂

Mode 2 introduced a prolonged low‑temperature phase (280 °C and 410 °C) to foster particle contact formation, followed by a 620 °C step. This yielded smaller grains (average 11 nm) and a higher BET surface area (47 m²/g vs. 39 m²/g for S₁). Pd loading increased linearly with the precursor concentration (0.001–0.193 wt % for 0.05–0.15 M PdCl₂).

XRD confirmed a cassiterite structure (a = 0.4738 nm, b = c = 0.3187 nm) across all samples. FESEM and HRTEM images (Fig. 3) revealed well‑defined SnO₂ grains and dispersed Pd nanoparticles, with clear interparticle boundaries.

a FESEM and b, c HRTEM images of the sensor Pd/SnO₂ nanomaterial

Electrical resistance in air displayed a complex dependence on Pd content: a low‑Pd minimum and a high‑Pd maximum were observed for both sintering modes (Fig. 4). The initial resistance drop at ≤0.05 % Pd likely reflects metallic Pd’s conductivity. At higher Pd loadings, reduced oxygen chemisorption at the Pd–SnO₂ interface leads to increased resistance; beyond a threshold, Pd aggregation shortens interfacial boundaries, lowering resistance again.

Dependence of R₀ values of sensors a S₁ and b S₂ on Pd content at various temperatures (410–225 °C).

Higher resistance in air correlates with enhanced H₂ response (Fig. 5). Sensors S₂ outperformed S₁, achieving a response of 19.5 at 40 ppm H₂ and 261 °C. Excessive Pd (≥0.15 %) diminishes response due to cluster aggregation covering SnO₂ surface sites.

Dependence of the sensor response to 40 ppm H₂ for sensors a S₁ and b S₂ on Pd content at various temperatures.

The optimally doped sensor (0.09 % Pd/SnO₂) exhibited linear conductivity over 2–1000 ppm H₂ at 327 °C and 382 °C, with a detection limit of 2 ppm (R₀/R_H₂ = 2.1 at 261 °C). Response and recovery times were 3 s and 12 s, respectively—remarkably fast compared to prior reports (120 s/15 min) [31].

Conductivity of the sensor S₂ (0.09 % Pd/SnO₂) versus H₂ concentration at 261, 327, and 382 °C.

Dynamic response of the optimal Pd‑doped sensor (0.09 % Pd/SnO₂) at 261 °C.

Selectivity tests (500 ppm CO/CH₄) demonstrated a markedly higher response to H₂ compared to CO and CH₄, confirming minimal cross‑interference at 261 °C. The lower optimal temperature for H₂ relative to CH₄ (382 °C) and CO (327 °C) reflects the superior catalytic activity of H₂ on Pd/SnO₂ surfaces.

Response of sensor S₂ (0.09 % Pd) to various gas mixtures at 261 °C.

Long‑term stability over 6 months showed no drift or loss of sensitivity for S₂ (S‑67, S‑69) (Fig. 9), underscoring the robustness of the sintering strategy.

Stability of the 40 ppm H₂ response for sensors S₂ (S‑67 and S‑69) over 6 months at 261 °C.

Conclusions

Tailoring the high‑temperature sintering regime for Pd/SnO₂ sensors yields sub‑10 nm grains that dramatically enhance hydrogen response (R₀/R_H₂ = 19.5) at 261 °C. These sensors detect H₂ from 2 to 1089 ppm with a 2 ppm detection limit, exhibit 3 s response and 12 s recovery, and maintain stability over extended use.