How nanotechnology can be used to intensify our sense of smell

Augmenting our olfactory senses will help us determine the concentration of hazardous gases and carry out the exhaled breath analysis for possible disease diagnosis.

Human nose has about 6 million sensors and capability to detect nearly one trillion odours | Pixabay

Bengaluru: Imagine the possibility of a hand-held gadget, perhaps your mobile phone, augmenting your senses — smell, taste, touch, vision and audition. Let us consider a specific example of smell sensation.

Human nose has about 6 million sensors and capability to detect nearly one trillion odours. The canine nose is even more sophisticated and powerful with 300 million sensors. What if we can create a chip with a few billion nanosensors (olfactory receptors) and embed it in the cell phone! Then we can augment our nose: Determine the concentration of hazardous gases in the environment, or even carry out the exhaled breath analysis for possible disease diagnosis. The possibilities would be mind-boggling.

The next frontier for nanotechnology, actually, lies in revolutionising the way humans interact with the environment, through the sensory perceptions. While this might sound extremely difficult, akin to science fiction, we can draw inspiration from what we have achieved in semiconductor technology over the last few decades for micro/nanoelectronic chips.

Chip technology key to creating olfaction chips

Enabled by the miniaturisation of silicon transistors, the chip revolution has given us the capabilities to create a few billion transistors on a chip to yield an unprecedented performance for computation, storage and communication applications. Perhaps, we can use similar technology to create olfaction chips with a few billion nanosensors.

I surmise that the architecture of such a chip should necessarily be three dimensional (3D) with a few billion olfactory sensors created on top of a powerful silicon chip fabric with computation, storage and communication capabilities. This would facilitate the efficient processing of signals coming from the nanosensors without being influenced by external noise sources.

This architecture mimics the biological pathway of smell-sensing, analogous to the front-end sensors in the nose, followed by the powerful computation and pattern recognition engine at the back end, in the brain.

I believe that the semiconductor nanosensors are the most suitable receptors for this architecture. They work on the principle that the electrical resistance of the sensor changes, when a chemical (odour) molecule attaches to the sensor.

Using artificial intelligence

The magnitude of the electrical resistance can be calibrated to estimate the concentration of a particular molecule. In order to achieve the complex electronic nose, the sensors have to be created using a very large variety of nanomaterials to detect a wide range of characteristic features of an odour.

The material set can include different types of metal oxide semiconductors, organic polymers, 2D layered materials such as transition metal dichalcogenides. It is also essential that appropriate low temperature processing techniques such as printing, dispensing or other additive-manufacturing techniques should be developed with nanoscale resolution to deposit these materials on top of the silicon chip, so that the underlying computation fabric does not get affected while the nanosensors are integrated on top.

Since the interpretation of a specific odour is a complex pattern recognition problem, such a chip should have a powerful artificial intelligence (AI) algorithm/engine, preferably on hardware. In order to facilitate this, the underlying silicon compute engine can be complemented using AI accelerator engine, constructed either using digital architecture or with neuromorphic architecture.

To summarise, nanosensors enabling the ambient intelligence, through highly-augmented sensory perception, will be one of the biggest frontiers for nanotechnology in the next few decades.

This can be achieved through a combination of multiple breakthroughs, including nanomaterials processing technology, massively parallel nanosensor array architecture, heterogeneous 3D integration of sensors with computation-storage-communication engines and the capability to handle big data obtained from nanosensors through AI algorithms.

Navakanta Bhat is a professor at Indian Institute of Science, Bengaluru, and Chairperson, Centre for Nano Science and Engineering, IISc. Bhat won the Infosys Prize 2018 in Engineering and Computer Science. 

This article is a part of the ‘Nano is the New Micro’ article series from Infosys Science Foundation.