In the last few decades, significant improvements have been made in gas sensor technologies. Metal-oxide sensors have been used for low-cost detection of combustible and toxic gases. However, hurdles relating to sensitivity, stability and selectivity still remain. Recently, nanotechnology has helped tremendously through the introduction of nano-engineered materials like nanowires and nanoclusters. Nanowire sensors have much better sensitivity as compared with thin-film devices due to the larger detecting surface-to-volume ratio. But clearly, improvements are still needed.

For real-world applications, selectivity between different classes of compounds, such as combustible and toxic gases, is highly desirable. An ideal chemical sensor should distinguish between the individual analytes from a single class of compounds. For example, in detection of benzene or toluene, a good sensor will not be disturbed by other aromatic compounds present in the environment. This is a huge challenge for semiconductor based metal-oxide sensors, such as TiO2, SnO2, Fe2O3 and ZnO, which have inherent non-selective surface adsorption sites.

Recently, a new class of nanowire-nanocluster (NWNC) based gas sensors has gained interest. This type of sensor represents a new method of functionalizing the surface for selective adsorption and detection. The adjustable sensitivity can be achieved by tuning the density, size or composition of the nanoparticles that decorate the nanowires. These advantages make the NWNC sensors a good alternative to conventional thin-film sensors. So far, research into NWNC sensors has demonstrated the potential in sensing many important classes of compounds. However, most of these NWNC devices require elevated working temperatures. They also have long response/recovery times and must function in an inert atmosphere. All these limitation will be the obstacles in real-world usage for domestic, environmental or industrial applications. And finally, the sensors thus developed must be manufacturable. That is, they must be batch fabricated with high yield.

To remedy these problems, my thesis was divided into the following tasks,

1. Develop dry etching techniques to fabricate horizontally aligned GaN nanowires (NW), combining these techniques with wet etching treatment for surface damages removal. I call this a “top-down approach” using a subtractive process that fabricates NWs from thin-films and adding sensitive nanocrystals after the initial NW definition. This is to be compared to the additive “bottom-up” nanowire growth by MBE/HVPE/Sol-gel, in which NWs are grown, harvested from the growth surface and subsequently re-attached to a new surface. The top-down approach enhances the yield and homogeneity of the NW and it is mass-production oriented.

2. Study the metal-oxide nanoclusters (NCs) deposition method by physical vapor deposition (PVD) and rapid thermal annealing (RTA) for TiO2, SnO2, WO3, Fe2O3, etc. Develop the metal nanoparticle deposition method by PVD for Au, Ag, Pt, Pd, etc.

3. Study the crystalline phases and gas adsorption sites formed by the method and establish a database connecting metal-oxide bonding sites with different target chemicals.

4. Utilize Si doped n-type and unintentionally doped GaN nanowires functionalized with different metal-oxide and metal-oxide/metal composite nanoclusters to create a series of highly selective and sensitive gas sensing nanostructure devices.

5. Develop a low-cost micro-heater (MH) for local high temperature generation with low power consumption. This allows the rapid chemical desorption cycles as we anticipate frequently re-use or reset of the sensor. It also enables the use of these NWs in high temperature sensor applications.

6. Integrate the NW, NCs and MH into one working sensor, and integrate multiple types of gas sensors on a single chip. The chip can simultaneously sense many types of gases without interference.

In this study, the potential of multicomponent NWNC based sensors for developing the next-generation of ultra-sensitive and highly selective chemical sensors was explored. We have achieved uA and nA levels of baseline detector current and we have shown that low UV illumination enhances sensitivity for some cases. These sensors have low power consumption making them suitable for portable devices.