Researchers are now one step closer to larger-scale image-based detection of materials thanks to artificial intelligence. A team from Ecole Polytechnique Fédérale de Lausanne (EPFL) created a way to detect and analyze molecules without massive equipment thanks to a nanotech sensor.
Small sensors called metapixels sit on the surface of the device, and the system produces a bar code unique to each particle that lands on the surface. The bar codes can be then analyzed on a large scale and arranged using an AI pattern recognition system and artificial neural networks.
Organic molecules have specific vibrations in their chemical bonds, and those unique vibrational modes impact the way particles absorb light. Previously, one of the only ways to identify those unique signatures was through infrared spectroscopy. Spectroscopy identifies if a given molecule is available in a sample by if the sample assimilates light rays at a molecules frequencies. While that technology was accurate, it required large lab instruments with a high price tag.
The nanotech sensor shrinks the power and efficiency of a massive spectrometer down to a hypersmall scale.
"The molecules we want to detect are nanometric in scale, so bridging this size gap is an essential step," said Hatice Altug, head of EPFL's BioNanoPhotonic Systems Laboratory and a co-author of the study.
The EPFL's metapixels adjust and adapt to how a molecule interacts with light when it comes into contact with the sensor's surface.
"Importantly, the metapixels are arranged in such a way that different vibrational frequencies are mapped to different areas on the surface," said Andreas Tittl, lead author of the study.
The EPFL team has already used the metapixels to detect pesticides and organic compounds within a larger, more complex sample.
"Thanks to our sensors' unique optical properties, we can generate bar codes even with broadband light sources and detectors."
"Thanks to our sensors' unique optical properties, we can generate bar codes even with broadband light sources and detectors," said Aleksandrs Leitis, a coauthor of the study.
The unique development serves as a cost-effective solution for research in physics, nanotechnology, and even big data, AI and machine learning. The applications of the nanotechnology could be used beyond the traditional scope of even what the EFPL researchers anticipated.
"For instance, it could be used to make portable medical testing devices that generate bar codes for each of the biomarkers found in a blood sample," said Dragomir Neshev, another coauthor of the study.
Beyond just that, AI could be paired with the nanotechnology in order to process an entire library of molecular barcodes specific to any given compound. This could help classify proteins and DNA to other polymers and dietary chemicals. This would give biologists and chemical engineers a tool to quickly and succinctly spot even trace amounts of certain compounds within large or complex samples.
The research was published in full in Science.