In Situ Resistance Monitoring during Flame-Aerosol Deposition of Metal Oxide Gas Sensing Layers

Abstract

Introduction Chemo-resistive sensors are highly promising for the detection of trace-level compounds in medical diagnostics[1] or indoor air monitoring[2]. In particular, gas sensors made by direct flame-aerosol synthesis are attractive due to their highly porous and crack-free sensing films[3]. They are composed of high-surface area nanoparticles with diverse composition tuned for ppb-level sensitivity and high selectivity towards key analytes such as Si-doped WO3 for acetone[1] or Ti-doped ZnO for isoprene[4]. Fabrication development has also led to the ability to process flame-made sensors by techniques compatible with standard silicon-wafer micromachining[5]. This enables the preparation of low power sensors that are attractive for portable handheld devices. So far, the development of such sensors has been focused mostly on material composition and, to a lesser extent, on film morphology and design. However, optimizing the latter could improve the sensor sensitivity as well as response and recovery times. Measuring the resistance during flame-deposition of nanostructured films in situ offers fabrication control by providing immediate feedback when an interconnected film is created[6]. This ultimately allows the fabrication of sensing films of minimal thickness and optimal resistance in a reproducible and economical manner. In situ resistance monitoring This work focuses on in situ monitoring the resistance of nanostructured films fabricated by direct deposition of flame-made Sb-doped SnO2 nanoparticles[6]. The sensing film morphology was tuned by a substrate-impinging particle flame, fed by a precursor solution containing varying metal ion concentrations. In situ resistance monitoring enabled control over final film resistance as well as direct insight into network formation and nanoparticle growth, necking and coalescence. The performance of these films as ethanol vapor sensors was assessed and related to their in situ resistance and nanoparticle film characteristics. Method Gas sensing nanoparticle films were prepared by direct deposition of flame-made Sb-doped (12 at%) SnO2 onto alumina substrates with interdigitated platinum electrodes. Deposition was carried-out for up to 6 min. Substrate preheating, film deposition and cooling to room temperature was closely monitored by the in situ resistance (Ri) between the interdigitated electrodes with a multimeter (Tektronix DMM4050). Nanoparticle sensing films were imaged by scanning electron microscopy (Hitachi S-4800 FE-SEM) after particle deposition for 2 min. Results and Conclusions Figure 1a shows the formation of Sb-doped SnO2 nanoparticles in a flame, followed by direct deposition onto the sensor substrate and in situ film resistance (Ri) read out. Figure 1b shows the in situ film resistance as a function of deposition duration using 1 (red squares), 10 (blue circles) or 100 (green triangles) mM FSP precursor solution concentrations (PSC). At low PSC (1 mM), the Ri remains at first (up to ca. 1 min) rather indifferent to nanoparticle deposition as it takes some time to build the interconnected particle bridges. But as deposition is prolonged (t >1.5 min), the Ri starts to drop. This clearly indicates the formation of an interconnected network of nanoparticles. Also quite notable is the strong oscillation of Ri by more than an order of magnitude during network formation. This stems from the constant formation (by deposition) and break-up of necked nanoparticles (by coalescence) on the substrate (i.e., repeatedly from lace-like to cauliflower-like structures) during flame-deposition. Increasing the PSC from 1 to 10 mM or 100 mM leads naturally to a higher generation rate of nanoparticles and, therefore, faster drop in resistance. This is confirmed by SEM imaging after 2 minute deposition time for 1, 10 and 100 mM (Figure 2 a-c). As expected, nanoparticles are assembled more densely at higher PSC. As a result, the film formation and morphology of Sb-doped SnO2 can be monitored during direct flame-deposition. This monitored Ri is correlated to its evolving particle size and necking, enabling the ability to optimize material quantity, fabrication time and the final film resistance. This leads to the ability to rapidly synthesize thin flame-made gas sensors with improved sensitivity and reduced response/recovery times.