Label-Free SARS-CoV-2 Detection: Direct SERS Antigen and Antibody Biosensing
Skip the chemical reporters and statistical software. Learn how a single SERS chip achieves label-free viral protein detection through distinct, raw spectral fingerprints.
Since SARS-CoV-2 — the virus behind the COVID-19 pandemic — emerged, rapid tests have carried the weight of screening for viral exposure at both the individual and population level: workplaces, clinics, and airports all leaned on fast antigen strips because PCR turnaround was too slow for real-time decisions. But most of those rapid tests still depend on antibody-based lateral-flow chemistry, and label-free point-of-care testing (POCT) that reads a viral protein’s own molecular fingerprint has largely stayed confined to the research bench.
Most SERS-based biosensors for viral or protein targets lean on a workaround: attach an antibody to the nanostructured metal, or tag the target with a Raman-active reporter molecule, and let that label do the work of amplifying an otherwise weak, ambiguous signal. It’s a reliable strategy — and it’s also exactly the kind of complexity that drives up fabrication cost, adds processing steps, and limits how fast a chip can move from lab bench to point-of-care device.
The Obstacle –
2D Substrates and Labeled Chemistry Cap
A study published in Applied Physics Reviews (Impact Factor: 19.0; Sahoo et al., 2023, DOI: 10.1063/5.0155256) frames the problem plainly: most SERS chips built for biomolecular detection require antibodies conjugated onto the nanostructured metal, or a reporter connected to the target, just to get a usable signal. That conjugation chemistry is exactly what makes fabrication complicated and expensive. Compounding the problem, most SERS substrates in the literature are effectively single-layer, two-dimensional structures — which caps the number of plasmonic hotspots available and limits sensitivity from the outset.
The research team, based at National Chung Hsing University with direct collaboration from ProTrusTech and Nanovie, set out to see how much of that complexity could be removed rather than engineered around.
The Innovation –
One Material, Stacked in 3D, With Nothing Attached
The result is a multiscale, three-dimensional SERS substrate (M3D-SERS) built from nothing but silver nanowires (AgNWs) — a single component, with no secondary materials, no antibody conjugation, and no reporter chemistry. The nanowires are sprayed onto a substrate in successive layers, from a single layer (1L) up to six layers (6L), building genuine depth rather than just surface density.
To understand why this geometry works, the team ran finite integration technique (FIT) simulations across a wide parameter space — nanowire length and diameter, interwire gap distance, end-to-end gaps, and crossing angle — and compared the results against real SERS measurements. A few findings stand out for anyone thinking about substrate design generally, not just this chip specifically:
- Two intersecting nanowires generate the largest area of enhanced electric field — larger than parallel or end-to-end configurations — which is the geometric reason random crossed-wire stacking outperforms more “orderly” 1D or 2D nanowire arrangements.
- Layer depth has a sweet spot, not a “more is better” relationship. SERS signal intensity increased through the first four layers, then declined — the deeper layers’ hotspots exist, but the top layer increasingly blocks laser penetration to the layers beneath it. The optimal construction, confirmed both by simulation and by benchmarking against bovine serum albumin (BSA), landed at four layers (4L).
- The random, imperfect stacking is a feature, not a limitation. Unlike precisely ordered nanowire arrays reported elsewhere, this randomly cross-stacked architecture creates gaps of varying size — which the researchers describe as being able to “trap” analytes across a range of sizes, from small organic molecules to DNA, proteins, and viruses, rather than being tuned to one specific target dimension.
Key Results –
From a Protein Benchmark to a Live Viral Target
The team validated the M3D-SERS chip in progressively harder stages. A benchmark protein set the baseline sensitivity, the actual viral target and its antibody were then detected at working concentrations directly enabling SARS-CoV-2 antibody detection without any labeling chemistry. Across three different laser wavelengths, the two were told apart from nothing but their raw spectral fingerprints, that same signal held up once the antibody was mixed into realistic human serum, and the chip kept working after months of ordinary storage.
Bovine Serum Albumin (BSA), as a Benchmark Protein
The optimized 4-layer chip achieved a limit of detection of approximately 10 pM in a 10 µL sample volume.
Spike Protein and Antibody Both Detected at 0.1 nM, Across Three Wavelengths
The spike RBD protein and its antibody (1A9) were both detected at 0.1 nM, consistently across three laser excitation wavelengths (532, 633, and 785 nm), with the 4-layer construction again proving optimal.
Distinguishing the Virus Protein From Its Antibody — Without Help
The two produced distinct, well-resolved Raman peaks, with the researchers tracing the difference primarily to tryptophan content (peaks near 737, 887, and 1579 cm-1) and secondary structure variation (α-helix-associated bands near 1646 and 1288 cm-1). Notably, the authors point out that signals reported elsewhere in the literature for SARS-CoV-2 SERS detection are often weak enough to require statistical or machine-learning post-processing just to differentiate biomolecules — where this substrate’s raw signal was intense and clear enough that no such processing was needed.
Two more crucial real-life application characteristics to note:
- Robustness in a realistic mixture: the antibody signal remained identifiable even when mixed at varying ratios into human serum albumin (HSA), a step toward relevance for real serum conditions rather than a purified buffer.
- Stability: the chip’s SERS signal held with less than a 10% decrease after being stored in ambient atmosphere for over 90 days.
Together, these results carry the chip from a proof-of-concept protein assay to a realistic diagnostic scenario: a single label-free substrate that detects a specific viral target, tells it apart from its own antibody using nothing but raw spectral shape, and keeps that performance intact for months on the shelf.
All of this — importantly — was measured on ProTrusTech’s own MRID Dual Lasers system, integrated with a thermoelectrically-cooled CCD. This wasn’t a proof-of-concept run on generic lab equipment; it’s peer-reviewed, published data generated on the same instrument platform ProTrusTech builds and sells.
Note: This research was supported by Taiwan’s Ministry of Science and Technology and the Ministry of Education’s Higher Education Sprout Project (Advanced Plant Biotechnology Center).
Why ProTrusTech Is Featuring This Research
This one is close to home. Co-author Chun-Ta Huang is part of ProTrusTech’s own R&D team, and co-author Kunju Tsai represents Nanovie — meaning this isn’t research we’re summarizing from the outside; it’s a study we helped generate, using our own Raman instrumentation, alongside our long-standing academic and manufacturing partners. It also builds directly on the same crossed-wire nanowire “woodpile” architecture behind our other feature on ≪ label-free direct-bilirubin detection ≫ — same underlying platform, applied here to a viral protein and antibody instead of a small-molecule metabolite. Seeing one substrate architecture generalize across such different biomarker classes, on our own instrument, is exactly the kind of evidence we look for before we invest further in a technology direction.
Engineering the Solution –
Precision Layering Is a Manufacturing Problem
The study’s own layer-by-layer spray method — building up from 1 to 6 layers and finding the four-layer optimum — is precise enough to work on a lab bench, repeated by hand. Turning “spray four layers, no more, no less” into a specification that holds across thousands of chips is a different problem, and it’s the one ProTrusTech and Nanovie are positioned to solve together:
- Nanovie Metal Xano produces high-purity metal nanoparticles in solutions directly from raw metal wires through a chemical-free process, avoiding the batch-to-batch variability that comes from wet-chemical synthesis routes.
- Nanovie Xano SC (Spray Coater) replaces manual, hand-held spraying with programmable, repeatable deposition — the kind of layer-count precision this study shows really matters, since going from three to four layers versus four to five layers measurably changes chip performance in either direction.
Paired with ProTrusTech’s MRID and SRflex Raman platforms, this gives a path from “four layers is optimal in this paper” to “every chip that ships is built to that same specification.”
Featured Instruments:
Why This Matters for SARS-CoV-2 Antibody Detection
Beyond SARS-CoV-2 specifically, this study is a useful proof point for label-free SERS more broadly: a single-component, unmodified nanostructure can distinguish a protein from its own antibody using nothing but the intrinsic Raman signal — no conjugation chemistry, no reporter, and in this case, no computational signal amplification. That’s a meaningfully lower bar for building new biosensor chips targeting other proteins or pathogens, since each new target doesn’t require redesigning the chip’s surface chemistry — only characterizing its native Raman fingerprint.
What’s Next
The researchers point to the same random-stacking architecture as a platform intended for variable-sized analytes — small organic molecules, DNA, proteins, and viruses alike — rather than a chip engineered around one target. For ProTrusTech and Nanovie, the open work is what it usually is at this stage: pushing fabrication consistency from lab-bench spraying to production-line repeatability, and testing this same substrate family against additional biomarkers beyond bilirubin and SARS-CoV-2.
If you’re evaluating label-free SERS platforms for viral, protein, or antibody detection — or want to see this substrate architecture demonstrated on our Raman systems — get in touch with our team.
Key Takeaways
- A single-component, multilayer silver-nanowire SERS substrate (M3D-SERS) detects biomolecules without antibody conjugation or reporter chemistry.
- Random, crossed-wire 3D stacking outperforms ordered 1D/2D nanowire arrays by generating a greater density of plasmonic hotspots at nanowire intersections.
- Four stacked layers (4L) is the experimentally and computationally confirmed optimum.
The chip detected SARS-CoV-2 spike protein and its antibody at 0.1 nM, with signal clear enough to distinguish the two without statistical or machine-learning processing. - All measurements were performed on ProTrusTech’s own MRID-Raman system.
This is the same nanowire-woodpile platform behind ProTrusTech’s ≪ direct-bilirubin detection research ≫ — evidence the architecture generalizes across biomarker classes.
