A sensor that requires no nanoscale patterning detects single molecules of a breast cancer protein at concentrations 250 times lower than previous devices of its type.
(Nanowerk Spotlight) A single drop of blood contains trillions upon trillions of molecules. Among them, the earliest signals of cancer, protein biomarkers shed by nascent tumors, may number in the single digits. Detecting one molecule among 10²⁴ others defines the frontier of biosensing, a threshold measured in attomolar concentrations. Most diagnostic technologies cannot approach it. Such sensitivity could transform cancer care: breast cancers caught before metastasis show five-year survival rates of 99%, compared with around 30% for metastatic disease, according to data from the National Cancer Institute’s SEER program.
Plasmonic sensors exploit interactions between light and metallic nanostructures. They have pushed detection limits into the femtomolar range, roughly a thousand times less sensitive than what early cancer diagnosis demands. Their fabrication also typically requires expensive lithographic patterning of nanoscale features.
Physicists have explored an alternative rooted in non-Hermitian systems, a class of physical systems that exchange energy with their environment rather than conserving it internally. These systems can exhibit exceptional points, conditions where two eigenvalues collapse into a single degenerate value. Eigenvalues are mathematical descriptors governing how a system responds to stimuli; when they merge, the system becomes exquisitely sensitive. Small perturbations near an exceptional point trigger disproportionately large responses that scale with the square root of the disturbance rather than linearly.
A 2020 study published in Nature Physics (“Symmetry-breaking-induced plasmonic exceptional points and nanoscale sensing”) exploited this effect using double-layered plasmonic structures and detected anti-IgG protein at 50 aM, pushing into attomolar territory. Yet this sensitivity still falls short for cancers whose earliest biomarkers hover near or below one attomole per liter.
A team at the University of Science and Technology Beijing and Tsinghua University has now extended this approach into single-molecule territory. Their study, published in Advanced Functional Materials (“Topological Engineering of Exceptional Points for Label‐Free Single‐Molecule Detection”), demonstrates what they call topological engineering of exceptional points. Rather than simply operating near an exceptional point, this method reshapes the mathematical landscape governing the sensor’s response, steepening it to amplify sensitivity 50 to 250 times beyond previous exceptional-point biosensors.
Topological engineering of exceptional points (EPs) for single-molecule detection. a) Schematic of the gold-polyimide-gold three-layered EP structure. b) Schematic of a multilayered EP structure. c) The eigenvalue topology of the three-layered EP structure. d) The engineered topology of scattering EPs for enhanced spectral response with multilayer design. e) Schematics of the functionalization of the EP structure and single-molecule detection of biomarker. (Image: Reproduced with permission from Wiley-VCH Verlag)
Fabricating the device requires no nanoscale patterning. The researchers deposited alternating thin-films of gold and polyimide, a common polymer, onto a silica substrate using magnetron sputtering and spin coating. They built structures with three, five, seven, and nine layers, tuning each to achieve forward reflectionlessness at around 1426 nm in the near-infrared spectrum. All fabrication and measurements took place in a Class 1000 cleanroom maintained at 25.0 °C.
Unpolarized light striking one of these multilayer stacks produces almost no reflection from the front surface, yet substantial reflection returns from the back. This pronounced front-to-back asymmetry creates an exceptional point in the system’s scattering matrix, the mathematical framework governing how light enters and exits the structure.
Biomolecules landing on the sensor disturb the local optical environment and break this delicate degeneracy. The resulting signal grows as the square root of the perturbation, a nonlinear amplification that makes vanishingly small disturbances visible. Increasing the number of layers heightens the reflection asymmetry and steepens the response curve. Among the configurations tested, the seven-layer design produced the steepest topology and the greatest sensitivity.
To validate these predictions, the researchers functionalized the seven-layer sensor with biotin and exposed it to streptavidin, a protein that binds biotin with high specificity. Plotting the spectral shift against streptavidin concentration on logarithmic axes yielded a slope of approximately 0.5, confirming the expected square-root behavior. Detection reached 0.9 aM, corresponding to about five molecules per square millimeter. With an illuminated area of roughly 0.03 mm², the sensor registered individual molecules.
The team then targeted ErbB2, also known as HER2, a protein whose elevated levels correlate with aggressive breast cancers. Early-stage tumors produce this biomarker at attomolar concentrations that elude conventional tests. For comparison, hyperbolic metamaterial sensors achieve detection limits around 10⁴ aM, and nanophotonic cavities reach approximately <10⁶ aM. After functionalizing the exceptional-point sensor with ErbB2 antibodies, the researchers achieved a detection limit of 0.2 aM, equivalent to one molecule per square millimeter.
This 250-fold improvement over the previous exceptional-point record places the sensor among the most sensitive label-free biosensors ever demonstrated.
Three practical advantages distinguish this approach. Thin-film construction sidesteps lithography, reducing fabrication complexity and cost. Operating in the near-infrared permits use of standard spectroscopic equipment. Topological engineering also provides a systematic design framework: adjusting layer count and thickness tunes the response steepness to match specific detection targets.
Challenges remain before clinical deployment. Blood and other biological fluids contain protein mixtures that could foul the sensor surface or introduce noise, and microfluidic integration would be necessary to deliver samples reproducibly. The authors acknowledge these hurdles but emphasize that the underlying physics is robust and the fabrication scalable.
Manipulating the topology of exceptional points, rather than merely exploiting their existence, unlocks a new sensitivity regime. Single-molecule detection of cancer biomarkers in clinical samples has remained beyond the reach of label-free biosensors. This research demonstrates it is now achievable.
For authors and communications departmentsclick to open
Lay summary
Prefilled posts
ORCID information
Yang Bai (University of Science and Technology Beijing)
, 0000-0002-6917-256X corresponding author
Nanowerk Newsletter
Get our Nanotechnology Spotlight updates to your inbox!
Thank you!
You have successfully joined our subscriber list.
Become a Spotlight guest author! Join our large and growing group of guest contributors. Have you just published a scientific paper or have other exciting developments to share with the nanotechnology community? Here is how to publish on nanowerk.com.
