Self-heating in nanowires: are we reaching the limits of gas sensor integration and power efficiency?

Developing fully autonomous gas sensors for distributed detection and early alert applications is one of the most compelling reasons to lower the power consumption of these sensor systems. Besides the power needed to read, control and communicate with the sensors, additional power is necessary to heat the chemical sensing materials to the optimum operation temperatures (up to 400ºC). While the former requirement is being reduced by the continuous advances in ultralow power consumption integrated circuits and communication protocols, the latter need is being diminished on the basis of reducing the size of the sensors. Recently, individual nanowires have emerged as suitable materials for conductometric gas sensing essentially due their high surface-to-volume ratio (which maximizes their response to gases) but also due to their crystalline properties (the nanowires are well-faceted single-crystalline materials), which makes possible to achieve long-term stable and repeatable gas responses [1]. State-of-the-art works use microheaters that inefficiently heat the nanowires with power needs of tens of mW (ten times lower than the consumption of the best commercial sensors) [2]. In any case, these power values still hamper the development of autonomous gas sensors. It is well known that the current applied to individual nanowires can dissipate enough power (Joule effect) to self-heat the small mass of the nanowire [3]. To date, this self-heating effect was regarded as a negative factor that cause degradation and early failure in nanodevices (Fig.1). We recently demonstrated that, in a controlled way, this effect can be used to achieve optimum sensing conditions for the detection of manifold gas species, avoiding the need of external heaters [4]. Here, the response of self-heated individual SnO2 nanowires towards NO2 and CO is presented (Fig.2). These proof-of-concept systems exhibited performances identical to those obtained with external microheaters. Such a comparison also enabled an indirect experimental method to estimate the temperature reached by the nanowires. An important consequence of the integration of the heater in the sensing nanomaterial is the dramatic reduction of the heating volume and, consequently, of the power requirements. According to our measurements, these nanodevices can operate under optimal conditions for gas sensing with less than 20 ¿W to both bias and heat them, which is significantly lower than the power consumption of the microheaters. This achievement paves the way for the development of fully autonomous and distributed gas sensor networks and is a good example of how advancements in the knowledge and control of the nanomaterials properties make possible novel strategies that could not even be imagined few years ago. Refs: 1 F. Hernandez-Ramirez, et al., Adv.Funct.Mater. 18, 2990 (2008). 2 F. Hernandez-Ramirez, et al., Nanotechnol. 18, 495501 (2007). 3 F. Hernandez-Ramirez, et al., Phys. Rev. B 76, 085429 (2007). 4 J. D. Prades, et al., Appl. Phys. Lett. 93, 123110 (2008)