In March 2019, chemist Jean-François Masson and his colleagues were at a Canadian sugar shack in Quebec testing one of the province’s biggest exports: maple syrup. The researchers were there to assess a new type of quality-control device capable of detecting flavour-distorting molecules that typically go unnoticed by the untrained tongue.
Masson and his team at the University of Montreal were testing a portable nanophotonic biosensor they had developed — a device made of nanometre-sized components that harnesses the properties of light to identify specific biological compounds. This particular sensor was essentially a small test tube containing a solution of gold particles that measure in the order of one-billionth of a metre in diameter. Flavour-distorting amino acids and amines bind to the gold particles and cause them to aggregate, altering how light interacts with the solution. This shifts the solution from red to blue — a change that can be seen in minutes with the naked eye. In a validation test of 1,818 syrup samples, the sensor was able to pinpoint 98% of those that were off-flavour1. The device was simple to use, with non-technical staff conducting the tests with relative ease. The team has since extended its workflow to maple sap — from which the syrup is made — and is working with syrup producers to make the devices available for wider use.
Masson’s maple-syrup sensor is among the simplest types of nanophotonic device, he says. But it demonstrates the power of nanophotonic biosensors as point-of-need devices, which can be deployed directly at the site where results are required. (In health-care applications, these are known as point-of-care tests.) He and other researchers are working on developing the tests for a wide range of uses, from the early detection of cancer and neurodegenerative diseases to monitoring environmental pollution2.
Interest in using nanophotonics in field-deployable biosensors has exploded because of the COVID-19 pandemic. Before that, it was difficult to convince companies to invest in this technology, but the huge demand for tests during the outbreak has made it clear that “we need point-of-care biosensing for almost everything, not only for COVID-19”, says nanotechnologist Laura Lechuga, a group leader at the Catalan Institute of Nanoscience and Nanotechnology in Barcelona, Spain. Rapid tests have been an invaluable tool for mitigating the spread of the coronavirus SARS-CoV-2. But when cases surged, overcrowded hospitals lacked the capacity to admit patients who had other diseases, meaning that many faced long waits to receive diagnoses. For conditions such as cancer, that can sometimes mean the difference between life and death, Lechuga adds.
Despite growing interest from industry, there is a long way to go before nanophotonic sensors more complex than Masson’s maple-syrup sensor can be rolled out on site. Hatice Altug, an applied physicist at the Swiss Federal Institute of Technology in Lausanne (EPFL), says that although researchers initially concentrated on building up the technology underlying these sensors, the focus has expanded in recent years to proof-of-concept demonstrations in various applications. But issues such as cost and the complexity of biological samples remain, she says.
Still, Altug is excited about the potential impact. After she and her colleagues published a review of nanophotonic biosensor technology in January3, they received numerous enquiries from people across commercial sectors, including diagnostics, food and hardware, who wanted to know whether and where these instruments were commercially available, she says. “It’s quite nice that the end-user is there.”
Generally speaking, biosensors use biological components such as antibodies, enzymes or nucleic acids to capture and quantify a substance in a biological sample. Nanophotonic biosensors are a subset of these tools that detect the interaction between recognition elements and target molecules, by applying the evanescent field principle.
An evanescent field is a rapidly decaying electromagnetic field. In a class of sensors known as nanoplasmonics, light briefly interacts with the materials in these sensors, such as nanometre-sized gold particles, to create a plasmonic evanescent wave — a group of charged particles that oscillate together at a certain frequency (or resonance frequency). This is influenced by factors such as the particles’ shape, size, composition and external environment.
When a molecular interaction occurs, the resonance of this plasmonic wave changes, shifting the properties of the reflected light. (A similar phenomenon is at play in stained glass windows: light reflecting off metal nanoparticles creates many hues.) This leads to a visible shift in the light’s colour that can be detectable with the naked eye, although an instrument for measurement is sometimes needed. The phenomenon, known as surface plasmon resonance, is commonly used in biosensors.
Picture ‘singing’ wine glasses. Rubbing a wet finger around the rim of a part-filled wine glass generates sound at a specific frequency, which varies according to the volume of liquid in the glass. Nanophotonics works by a similar principle, says Ryan Bailey, a chemist at the University of Michigan in Ann Arbor. “When biomolecules bind to the surface of a sensor, it’s essentially changing the amount of water in the glass.”
Other nanophotonic biosensors exploit the evanescent field principle and resonance in different ways. For instance, dielectric devices4, made of materials including silicon, operate using techniques such as interferometry. This analyses the patterns generated by superimposing light waves.
Whatever their precise mechanism, Altug describes the nanophotonic units that underlie these sensors as light-attracting antennas — and the smaller the structure, the more they concentrate the light, bringing smaller substances into the detection window. That’s what makes these sensors so sensitive, she says. This also lets researchers tune biosensors to the size of the molecule they are trying to detect. “The evanescent waves make it possible to confine the electromagnetic fields to the dimensions that are comparable to those of the molecules,” says Jiří Homola, a physicist at the Czech Academy of Science’s Institute of Photonics and Electronics in Prague.
According to Lechuga, nanophotonic sensors have several advantages over other types of biosensor: they can provide real-time information without requiring labels such as fluorescent tags, and are extremely sensitive. Researchers have shown, for example, that a nanophotonic sensor can detect proteins at attomolar concentrations (10–18 moles) — the equivalent of about 6 molecules in a 10 microlitre sample5. The relative simplicity of these sensors makes them well-suited to point-of-care applications, Lechuga says. And, in contrast to other point-of-care diagnostics, they can reveal not only the presence or absence of a molecule, but also its abundance — a measurement that is usually possible only using techniques such as the polymerase chain reaction.
For some researchers, such as Masson, nanoplasmonics are attractive because metal nanoparticles can be synthesized relatively easily, thanks to the effort that has gone into their development over the past 10–15 years. Their production has gone from an art form to reproducible science, says Masson, and now undergraduate students can create them. “In the past, that would have been very, very difficult.” Masson notes that gold is one of the easier materials to work with. Silicon fabrication, by contrast, requires a specialized clean room, meaning silicon-based biosensors are less accessible to researchers.
Yet Altug notes that, when it comes to large-scale manufacturing, silicon-based nanoparticles have the advantage over metallic ones. This is because they are produced through the same process as silicon-based electronics, so can, in theory, be manufactured at a large scale in existing factories. However, alternative nanoplasmonics materials, such as aluminium, can also be created in this way. Researchers are working to develop these, so mass production of nanoplasmonics should become feasible in the near future, according to Altug.
Today’s nanophotonic biosensors can detect a wide range of molecules — it’s often just a matter of selecting the correct capture reagent. However, Masson says, these sensors are more effective at identifying larger molecules, such as antibodies and enzymes, than at detecting small molecules such as metabolites. Still, this variety supports a wide range of potential applications. “If you talk to 20 different researchers in the field, they will give you 20 different answers of what is the best application,” Masson adds.
Photonic biosensors are already in use in clinics and research labs — surface-plasmonic resonators have been around since the 1990s. The Biacore platform from diagnostics firm Cytiva (formerly GE Healthcare Life Sciences), for example, is a popular option for quantifying intermolecular interactions in biochemistry labs. And, in 2007, Bailey helped to launch a company, Genalyte in San Diego, California, that now has photonic biosensors on the market. These include a silicon-based photonic device that can conduct more than 20 simultaneous screens for signs of autoimmune diseases.
Other groups are developing such devices for applications including disease diagnosis, food assessments and environmental monitoring.
Health care is an area of particular interest. One focus of Lechuga’s team, for example, is infectious disease. Last December, her group reported the development of a portable nanophotonic antibody test for COVID-19. In a clinical validation study6 using samples from 100 patients diagnosed with COVID-19 and from 20 COVID-19-negative individuals (collected before the pandemic), the test showed 99% sensitivity and 100% specificity. It also provided information about the quantity of antibodies present.
Altug and her colleagues are developing sensors for neurodegenerative diseases. For example, by using an array of gold nanoantenna arrays to detect proteins such as α-synuclein that misfold and clump together into toxic aggregates in Parkinson’s disease7, they hope to pinpoint the biochemical changes in the brain that happen years before disease onset. Detecting these conformational changes might open the door to early intervention, Altug says. Using such tools to investigate how protein misfolding occurs might also help researchers to identify new therapeutics that could stop the pathological process in its tracks, she adds.
Several groups are working on screening tools for the early detection of cancer, or for monitoring treatment efficacy by making nanophotonic biosensors capable of pinpointing specific biomarkers, DNA, proteins and cytokines. In May 2021, for example, Altug and her team reported8 that one of their biosensors could detect a subset of extracellular vesicles called exosomes — nanometre-sized sacs ejected from cells that contain components such as DNA and proteins9. Altug’s team tracked exosomes that had been released by breast cancer cells in real time, with a limit of detection of 267 nanograms per millilitre. That concentration, the authors write, is “clinically relevant for the detection of cancer-related [extracellular vesicles]”8.
But when it comes to point-of-need devices, obstacles remain. Chief among them is economics — manufacturing costs remain high, making it difficult to convince companies to produce the devices in large enough quantities to make them cost-effective, Altug says. Another challenge is dealing with the complexity of biological materials such as blood, which can vary significantly both between and within patients. Dealing with this variability while keeping costs low remains difficult, she adds.
“There are a lot of optical sensors out there on the market, but they’re not portable”, and most need to be used in a lab setting, says Eleni Makarona, a physicist at the Institute of Nanoscience and Nanotechnology in Athens, part of the Greek National Center for Scientific Research ‘Demokritos’. Her team has developed a customizable nanophotonic platform that can be used for various applications, including pinpointing harmful substances in food and identifying disease-related biomarkers. “The question is how to take these things and make them fit into a purse or a small suitcase to carry it with you, without losing the analytical capability or making it way too expensive?”
Although still a work in progress, many researchers are already looking ahead to the future of point-of-care nanophotonics.
Some are exploring wearable and implantable nanophotonic devices that could provide continuous monitoring of important biomolecules, such as glucose10. But significant challenges remain. Masson says that the biggest issues include the difficulty of delivering light inside the body, and the body’s tendency to reject foreign objects.
Biomedical engineer Eden Morales-Narváez at the Center for Research in Optics in León, Mexico, is leading a team that is developing wearable sensors. In one approach, they are embedding nanophotonic materials such as gold into nanopaper, a transparent and flexible material made from nanocellulose, a substance produced by bacteria11. His team has already used this approach to develop a tattoo-like wearable nanophotonic patch that changes colour when a user is exposed to harmful levels of UV radiation12. The patch design is relatively simple, because there is no biomolecule involved — but Morales-Narváez’s team is working on integrating biosensing capabilities into its nanophotonic devices. Doing so is a challenge, Morales-Narváez says, because it involves adding complex components.
Before nanophotonic biosensors can become broadly available, researchers must demonstrate them as a reliable technology that can be used in the clinic, pharmacy, home or anywhere else, Lechuga says. Eventually, Lechuga imagines a kind of diagnostic ‘Nespresso machine’: a single device with different cartridges carrying nanophotonic sensors for use in various clinical applications. She and her colleagues are starting a new company, called EROICA Diagnostics, to commercialize this approach.
“Compared to ten years ago, we’ve made really good progress,” Altug says. “In another ten years, maybe we will be comparing products.”