One Drop of Blood, Label-Free Clinical Direct-Bilirubin Diagnostics via SERS

One 10 µL drop, one dried chip, one laser read — and results across 12 patients matched hospital diazo-method accuracy in seconds, not hours.

Monitoring bilirubin is one of the most routine, and most consequential tests in medicine. It flags liver dysfunction, drives the diagnosis of jaundice, and guides treatment of neonatal hyperbilirubinemia. More recently, clinical studies have linked the direct-to-total bilirubin (D/T-BIL) ratio to survival outcomes in severe pneumonia, including COVID-19 pneumonia, giving the measurement new relevance beyond hepatology.

And yet, the tools used to measure it haven’t fundamentally changed in decades.

The Obstacle –
Fast, Cheap, and Accurate Don’t Usually Coexist

Clinical labs typically rely on the diazo method, enzymatic assays, or HPLC to quantify bilirubin. Each works — but each comes with a tradeoff. These methods generally require multi-step sample pretreatment, specific reagents or enzymes, and either benchtop instrumentation or trained technicians to run and interpret them. None of that supports a rapid test at the point of care, and none of it is cheap enough…

Researchers have looked to surface-enhanced Raman scattering (SERS) as a way around this: SERS reads a molecule’s vibrational “fingerprint” directly, with no labeling chemistry required. But SERS performance lives and dies by the substrate — the nanostructure has to generate enough plasmonic field enhancement, consistently, to pull a usable signal out of a complex biological matrix like blood serum.

The Innovation –
A “Woodpile” Nanochip That Reads Two Signals at Once

In a study published in Biosensors and Bioelectronics (Sahoo et al., 2022, DOI: 10.1016/j.bios.2022.114440), researchers at National Chung Hsing University, together with clinical collaborators at National Taiwan University Hospital and Changhua Show Chwan Memorial Hospital, designed a nanostructure that sidesteps this substrate problem by generating hotspots in three dimensions rather than a single plane.

Finite integration technique (FIT) simulations show why the geometry works: field enhancement concentrates sharply at nanowire junctions, and the 3D woodpile stacking multiplies the number of these junction hotspots compared to a flat, single-layer substrate.

The team confirmed this wasn’t just theoretical by photographing the hotspots directly with fluorescence microscopy: longer, higher-aspect-ratio nanoneedles concentrated hotspots at their crossing points, while shorter nanorods lit up at both ends instead — proof that nanowire geometry, not simulation alone, determines where a hotspot forms.

Their design allows the same chip to support two independent optical readouts from one sample:

• Surface-Enhanced Raman Scattering (SERS) – captures bilirubin’s molecular fingerprint directly. The researchers identified bilirubin’s characteristic SERS peaks at ~677 cm^-1 (C–O twisting in the COOH group), ~791 cm^-1 (ring deformation), ~970 cm^-1 (methylene bridge deformation), ~1097 cm^-1 (C–N stretching), ~1245 cm^-1 (CCH/CCC bending), and ~1441 cm^-1 (mixed C–C/C–N stretching), with the most sensitive quantification marker at ~1610 cm^-1, attributed to the lactam group.
• Surface Plasmon-Enhanced Fluorescence (SPEF) – the same silver nanowire network also amplifies bilirubin’s naturally weak fluorescence (bilirubin’s absorption and emission peaks sit at approximately 450 nm and 515 nm respectively). In benchmark testing with rhodamine 6G, the 3D-PHS structure delivered better than a 1000-fold fluorescence enhancement over the free dye in solid-state testing.

The dual-readout SERS/SPEF chip boosts fluorescence 1000-fold.

Bright-field and fluorescence images of rhodamine 6G with and without the 3D-PHS silver nanowire substrate — confirming the dual-readout SERS/SPEF chip’s fluorescence-enhancement capability before it was ever tested on bilirubin.

Because the two techniques probe the molecule through different physical mechanisms, one becomes a built-in cross-check for the other — a form of dual-verification that single-modality sensors can’t offer.

Key Results –
From Pure Chemical to Patient Serum

The research team validated the chip across an escalating series of real-world conditions rather than stopping at a clean chemical standard: a pure bilirubin standard set the baseline sensitivity, spiking bilirubin into fetal bovine serum tested performance in a real biological matrix, and blood serum from twelve patient volunteers put the chip head-to-head against the hospital’s own diagnostic reference method.

SERS Detected Concentrations down to 1 nM

SERS spectra of direct bilirubin across a range of concentrations show the same resolved peak pattern — including the diagnostic 1610 cm^-1 lactam band — with detection confirmed down to 1 nM (R² = 0.9997).

SPEF – Direct-Bilirubin Second Readout down to 0.1 µM

SPEF detection of direct bilirubin itself reached a limit of 0.1 µM — a second, independent quantitative readout from the same dual-readout SERS/SPEF chip, on the same sample.

Direct-Bilirubin Diagnostics Hold Up in Serum, Down to 10 nM

SERS spectra of direct bilirubin spiked into fetal bovine serum (FBS) show all characteristic peaks remaining clearly present, with a detection limit of 10 nM. The shift from the pure-chemical LOD reflects the added complexity of a real biological matrix, not a loss of selectivity.

Twelve Patient Samples Matched the Hospital’s Own Reference Method

The histograms above compare the direct bilirubin (D-BIL) concentration measurements of 12 clinical patient samples using three distinct methods: (1) Standard clinical Diazo assay (grey bar), (2) SERS (blue bar), and (3) SPEF (green bar).

The near-identical heights of the bars across all 12 patients prove that the dual-readout 3D-PHS nanochip matches the accuracy of standard hospital laboratory equipment, but delivers results in seconds rather than hours.

It also demonstrates the chip’s high resolution can reliably help differentiate healthy patients from those requiring immediate medical intervention, even in the case when concentration is below the critical clinical threshold of 25 µM (red dashed line).

All of this was achieved from just 10 µL of serum, drop-cast directly onto the chip and measured after drying — no chemical labeling, no reagents, and no purification step. The authors note this is the first time direct bilirubin has been measured in a clinically relevant range using a label-free, pretreatment-free SERS platform, and the first time bilirubin has been characterized using enhanced fluorescence imaging and spectra at all.

Note: This work was supported by Taiwan’s Ministry of Science and Technology, grants MOST 110-2113-M-005-022 and MOST 110-2634-F-005-006.

Why This Matters

For clinicians, a 10 µL, pretreatment-free test that matches hospital-grade accuracy in seconds rather than hours changes what’s operationally possible — from neonatal wards needing rapid hyperbilirubinemia screening, to resource-limited settings without access to HPLC or automated clinical analysers, to any point-of-care testing (POCT) setting where waiting for a lab turnaround has a real clinical cost.

For the broader biosensing field, this study validates a crucial design principle: engineering a nanostructure’s geometry to generate 3D hotspots. This density drives two independent optical readouts from a single label-free chip, a methodology that extends far beyond bilirubin detection.

Why ProTrusTech Is Featuring This Research

While this particular study predates our direct involvement, it’s part of the same research lineage we’re actively building on with National Chung Hsing University and Nanovie — the crossed-wire silver nanowire “woodpile” architecture behind this bilirubin chip is the same platform family we later helped extend to SARS-CoV-2 biomarker detection. See our feature on ≪ label-free SARS-CoV-2 detection ≫.

We follow this thread of research because it’s a rare case of a single substrate design generalizing cleanly across a small-molecule metabolite and a large viral protein — that versatility is exactly what we look for before committing engineering resources to a platform.

Engineering the Solution –
From Bench Science to a Manufacturable Chip

While the laboratory-scale chips in this study were constructed via manual solution-sprinkling, commercial production requires absolute layer consistency, high throughput, and reproducibility.

To bridge this gap, ProTrusTech partners with Nanovie to offer a fully automated, chemical-free fabrication pipeline:

• Nanovie Metal Xano produces high-purity metal nanoparticles in solutions directly from raw metal wires using an eco-friendly, chemical-free technology, generating up to 1,000 mg/hr per channel.
• Nanovie Xano Spray Coater (XSC) deposits uniform, multilayered nanostructures across areas up to 100 X 100 cm under ambient conditions with programmable parameters, making it possible to reproduce the exact 3D-PHS hotspot geometry optimized by the original researchers.

Paired with ProTrusTech’s SRflex portable Raman spectrometer series, the same dual-readout measurement demonstrated in this study becomes something a clinical lab can actually deploy at the point of care — not just something a research group can reproduce on an optical bench.

Featured Instruments:

What’s Next

• Batch-to-batch consistency at production scale. The manual solution-sprinkling used in this original study needs to become a repeatable spec — exactly the gap ProTrusTech and Nanovie’s Metal Xano + Xano Spray Coater pipeline is built to close.
• Integration with portable spectrometer workflows for non-specialist users. Pairing this dual-readout chip with ProTrusTech’s SRflex portable Raman spectrometer is what moves bilirubin testing out of the central lab and into neonatal wards, clinics, and point-of-care settings.
• Extending the same architecture to other clinically relevant analytes. This 3D hotspot design already generalized once, to SARS-CoV-2 biomarker detection — the open question is which clinical analyte it should tackle next.

Evaluating SERS/SPEF for a diagnostic or biosensing application? Get in touch with our team to talk through whether this chip architecture is the right fit for what you’re building.

Key Takeaways:

• A 3D silver-nanowire nanochip supports two independent optical readouts — SERS and SPEF — from a single drop of serum.
• No labeling chemistry, no reagents, and no sample pretreatment are required. Detection limits reached 1 nM in pure chemical and 10 nM in serum matrices.
• Across 12 patients, chip readings matched hospital diazo-method results, including around the clinically critical 25 µM threshold.
• ProTrusTech and Nanovie’s automated fabrication pipeline (Metal Xano + Xano Spray Coater) is built to reproduce this exact hotspot geometry at production scale.
• This is the same underlying platform later applied to ≪ SARS-CoV-2 biomarker detection ≫.

Featured Publications: