The current commercial landscape for gas sensing is dominated by metal oxide (MOx) based sensors due to their balance of performance and cost as well as the opportunity for simple device integration, potential for miniaturisation and incorporation into multi-analyte detection systems. However, MOx sensors have significant shortcomings which include poor stability and selectivity, high power consumption/operating temperatures, long recovery and response times.

I investigated the limitations of industry leading gas sensors and addressed a number of issues plaguing traditional sensors, with a fully inkjet-printed metal oxide(MOx)/graphene-based intelligent sensor capable of predictive gas sensing in mixed environments. Through optimisation of new synthesis strategies of the active sensing material, exploitation of the synergistic advantages of a combined MOx/2D nanocomposite material, the use of inkjet printing as a scalable device manufacturing process and a combination of data analysis and machine learning (ML), I could successfully distinguish 5-different gas classes with 98.1% accuracy at very low concentrations whilst also compensating for humidity interference using a single sensor. My sensors allow detection at room temperature, offering opportunities for low power sensing networks enabled by Internet of Things (IoT) revolution. Beyond detection, my proposed strategy also represents the first ever demonstration of a predictive gas sensor that can successfully detect gases at untrained concentrations in a mixed gas environment, representing a revolutionary step towards robust, highly sensitive, scalable, low cost and low power sensors for indoor air quality monitoring.

Building upon my previous work, I synthesize 5 rGO/MOx (rGO/Co₃O₄, rGO/Cu₂O, rGO/WO₃, rGO/CuOₓ/CoOₓ and rGO/CoOₓ/WOₓ) inks and present a fully inkjet printed 4th order Hilbert-Piano fractal sensor array for room temperature CH₂O detection. I demonstrate the highest room temperature gas sensing response for 1 ppm formaldehyde (CH₂O) in literature using my rGO/CuOₓ/CoOₓ material, overcoming the difficulty with its detection and providing evidence for increased sensitivity through the implementation of a heating platform to my fabricated sensors. My findings offer a cost effective, sensitive VOC sensing framework, capable of selective detection of formaldehyde at the concentrations where adverse health effects start to occur.