Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 58 of the 58 references it cites were found in a registry or an index; none of the 48 sources its claims cite is checked for support yet.

Gap-enhanced Raman tags for physically unclonable anticounterfeiting labels

Open access
Gu, Yuqing, He, Chang, Zhang, Yuqing, Lin, Li, Thackray, Benjamin David, Ye, Jian · 2020 · Nature Communications

Abstract

Anticounterfeiting labels based on physical unclonable functions (PUFs), as one of the powerful tools against counterfeiting, are easy to generate but difficult to duplicate due to inherent randomness. Gap-enhanced Raman tags (GERTs) with embedded Raman reporters show strong intensity enhancement and ultra-high photostability suitable for fast and repeated readout of PUF labels. Herein, we demonstrate a PUF label fabricated by drop-casting aqueous GERTs, high-speed read using a confocal Raman system, digitized through coarse-grained coding methods, and authenticated via pixel-by-pixel comparison. A three-dimensional encoding capacity of over 3 × 10 15051 can be achieved for the labels composed of ten types of GERTs with a mapping resolution of 2500 pixels and quaternary encoding of Raman intensity levels at each pixel. Authentication experiments have ensured the robustness and security of the PUF system, and the practical viability is demonstrated. Such PUF labels could provide a potential platform to realize unbreakable anticounterfeiting.

Overview

CoChat Summary

TL;DR

The paper demonstrates SERS-based physical unclonable function labels made from drop-cast gap-enhanced Raman tags (GERTs), read by confocal Raman mapping and digitized for authentication. Ten GERT types with 50 × 50-pixel quaternary encoding yield a theoretical encoding capacity of 3.2 × 10^15051, with authentication separating same-label from different-label readouts.

Key findings

  • PUF labels were fabricated by drop-casting aqueous GERTs, read using a confocal Raman system, digitized through coarse-grained coding, and authenticated via pixel-by-pixel comparison.
  • The labels use three-dimensional encoding: pixel position, GERT type, and Raman intensity level. For ten GERT types at 50 × 50 pixels (2500 pixels) with quaternary encoding, the theoretical capacity is (4^10)^2500 = 3.2 × 10^15051; binary encoding gives (2^10)^2500 = 5.6 × 10^7525. A one-type 4-NBT GERT label at 50 × 50 pixels gives 4^2500 = 1.4 × 10^1505 for quaternary and 2^2500 = 3.8 × 10^752 for binary.
  • Authentication used 100 PUF labels, each measured three times. Same-label similarity indexes were around 94% for binary and 84% for quaternary encoding; different-label indexes dropped to around 57% for binary and 30% for quaternary. Suggested error margins were 85% for binary and 70% for quaternary, with distribution gaps of around 20% and 30%, respectively.
  • High-speed readout in DuoScan/SWIFT mode acquired a 50 × 50-pixel map in 6 s with about 0.7 ms exposure per pixel, whereas STAGE mode required 20 min for the same resolution. A practical label on Scotch tape (1.8 × 1.8 mm²) was read at 50 × 50 pixels with 10 ms per pixel.

Methods

  • Core-shell GERTs with Au core/shell and Raman reporters embedded in a ~1 nm nanogap were synthesized using ten thiolated aromatic reporter molecules. Labels were made by drop-casting 2 μL of GERT solution onto a silica substrate or Scotch tape, using one, three, or ten GERT types.
  • Readout used a confocal Raman system (LabRAM XploRA INV, Horiba) with a 785 nm laser, 3 × 10^5 W cm−2, 60× objective, and 10- or 50-ms exposure per pixel. Mapping was performed in STAGE or DuoScan/SWIFT modes at resolutions of 2 × 2, 10 × 10, and 50 × 50 pixels.
  • Analysis included Savitzky–Golay smoothing, baseline correction, Z-score standardization, non-negative least squares (NNLS) demultiplexing for ten GERT types, global search threshold determination, and coarse-grained coding to binary {0,1} or quaternary {0,1,2,3}. Authentication used similarity index I by pixel-by-pixel comparison; 100 labels were split into 70 training and 30 test labels, each measured three times.

Limitations

  • Stated in the paper: readout speed still lags behind some other PUF labels, such as silicon PUFs, and is not yet enough for manufacturer’s registration; a protection layer and physical robustness need future work; digitization thresholds and error margins are preliminary and could be optimized with more sample data; a possible vulnerability is replacement of digitized signals uploaded to a cloud server, which the authors propose addressing with algorithmic encoding.
  • Stated in the paper: real encoding capacity shrinks when an error margin is introduced, and no fabrication method accesses all states with equal probability; small readout variations can significantly reduce real capacity. NNLS demultiplexing caused all GERT types to appear strongest at the same pixel, likely due to spectral overlap and signal bleed from exceptionally bright particles, which could be avoided with more spectrally distinct reporters or an improved algorithm.
  • Evident from the text: the 100-label authentication study used only 4-NBT GERT labels at 50 × 50 pixels, so the reported same-label/different-label similarity statistics and error margins are not demonstrated for the ten-type GERT labels.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41467-019-14070-9

Cites

58 of 58 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Fabrication and encoding of Raman PUF labels. a A Raman signal-based physical unclonable function (PUF) label can be fabricated in four steps. First, surface-enhanced Raman scattering (SERS) nanoparticles (NPs) functionalized with ten different Raman reporter molecules are synthesized by a wet chemistry method. Each type of SERS NPs emits a unique vibrational Raman spectrum. Second, SERS NPs are deposited on a substrate and stochastically form a two-dimensional (2D) pattern after drying. The PUF label can then be read using a confocal Raman system by mapping and recording the signals at different points (e.g., 8 × 8 pixels in this example). The various SERS NPs at different locations on the substrate produce Raman spectra with specific bands and intensities. A PUF label is finally produced by digitizing Raman mapping signals at each pixel in terms of the type of SERS NPs and the level of Raman intensities with the combination of the physical position of each pixel. b Calculation of the encoding capacity of a Raman PUF label. The formula (left) and the graph (right) show that the encoding capacity of a PUF label increases exponentially with the number of pixels (m) and as a power function of the number of possible responses for each pixel (R) in the recorded pattern. Furthermore, the number of responses of each pixel is a power function of the number of Raman intensity levels (L) and an exponential function of the type of SERS NPs (n). For example, if n = 10, L = 4, and m = 2500, the encoding capacity will be 410×2500 (3.2 × 1015051). This figure composition refers to the ref. 7.
Fig. 1 Fabrication and encoding of Raman PUF labels. a A Raman signal-based physical unclonable function (PUF) label can be fabricated in four steps. First, surface-enhanced Raman scattering (SERS) nanoparticles (NPs) functionalized with ten different Raman reporter molecules are synthesized by a wet chemistry method. Each type of SERS NPs emits a unique vibrational Raman spectrum. Second, SERS NPs are deposited on a substrate and stochastically form a two-dimensional (2D) pattern after drying. The PUF label can then be read using a confocal Raman system by mapping and recording the signals at different points (e.g., 8 × 8 pixels in this example). The various SERS NPs at different locations on the substrate produce Raman spectra with specific bands and intensities. A PUF label is finally produced by digitizing Raman mapping signals at each pixel in terms of the type of SERS NPs and the level of Raman intensities with the combination of the physical position of each pixel. b Calculation of the encoding capacity of a Raman PUF label. The formula (left) and the graph (right) show that the encoding capacity of a PUF label increases exponentially with the number of pixels (m) and as a power function of the number of possible responses for each pixel (R) in the recorded pattern. Furthermore, the number of responses of each pixel is a power function of the number of Raman intensity levels (L) and an exponential function of the type of SERS NPs (n). For example, if n = 10, L = 4, and m = 2500, the encoding capacity will be 410×2500 (3.2 × 1015051). This figure composition refers to the ref. 7.CC BY · PubMed Central
Fig. 2 SERS NPs with embedded Raman reporters, termed as gap-enhanced Raman tags (GERTs), used for fabricating Raman PUF labels. a Schematic representation of a GERT. b Molecular structures, c transmission electron microscopic (TEM) images, and d SERS spectra of ten GERTs with different embedded reporter molecules. Insets in c are the corresponding photos of various GERT solutions. All scale bars are 20 nm.
Fig. 2 SERS NPs with embedded Raman reporters, termed as gap-enhanced Raman tags (GERTs), used for fabricating Raman PUF labels. a Schematic representation of a GERT. b Molecular structures, c transmission electron microscopic (TEM) images, and d SERS spectra of ten GERTs with different embedded reporter molecules. Insets in c are the corresponding photos of various GERT solutions. All scale bars are 20 nm.CC BY · PubMed Central
Fig. 3 Fabrication and digitization of a PUF label composed of 4-NBT GERTs. a A bright-field image and b a scanning electron microscopy image (corresponds to the dashed square area in a) of a PUF label composed of 4-NBT GERTs in an area of 100 × 100 μm2 on a SiO2 substrate. Scale bars are 20 (left) and 1 μm (right). c Readout of the PUF label by the Raman mapping (plotted using the band at 1078 cm−1) and the corresponding digitization of e binary encoding and f, g quaternary encoding of Raman intensity levels at each pixel with a resolution of 2 × 2 (left), 10 × 10 (middle), and 50 × 50 (right) pixels. Panel d shows the SERS spectra at the points (1–6) indicated in c. Panels f and g show the two-dimensional (2D) and three-dimensional (3D) plots of digitization of the PUF label based on quaternary encoding of Raman intensity levels, respectively.
Fig. 3 Fabrication and digitization of a PUF label composed of 4-NBT GERTs. a A bright-field image and b a scanning electron microscopy image (corresponds to the dashed square area in a) of a PUF label composed of 4-NBT GERTs in an area of 100 × 100 μm2 on a SiO2 substrate. Scale bars are 20 (left) and 1 μm (right). c Readout of the PUF label by the Raman mapping (plotted using the band at 1078 cm−1) and the corresponding digitization of e binary encoding and f, g quaternary encoding of Raman intensity levels at each pixel with a resolution of 2 × 2 (left), 10 × 10 (middle), and 50 × 50 (right) pixels. Panel d shows the SERS spectra at the points (1–6) indicated in c. Panels f and g show the two-dimensional (2D) and three-dimensional (3D) plots of digitization of the PUF label based on quaternary encoding of Raman intensity levels, respectively.CC BY · PubMed Central
Fig. 4 Fabrication and digitization of a PUF label composed of ten types of GERTs. a A bright-field image and b the scanning electron microscopy image (corresponds to the dashed square area in a of a PUF label fabricated by ten types of GERTs in an area of 100 × 100 μm2 on a SiO2 substrate. Scale bars are 20 (left) and 1 μm (right). c Spectral demultiplexing to quantify the abundance of ten types of GERTS at the position indicated by a green dashed box in panel d. The best-fit spectrum (black) is obtained by fitting the measured Raman spectrum (red) with ten pure reference Raman spectra using the non-negative least squares (NNLS) method. d 2D and e 3D plot for the readout of the PUF label by the Raman mapping (with a resolution of 10 × 10 pixels) of multiplexed measured Raman signals using the integrated areas of Raman bands from 802 to 1600 cm−1 (as indicated in the gray area in panel c). f 2D plot for the readout of the PUF label by the Raman mapping (with a resolution of 10 × 10 pixels) of demultiplexed Raman signals and 2D plot of the corresponding digitizations using g binary encoding and h quaternary encoding of Raman intensity levels at each pixel.
Fig. 4 Fabrication and digitization of a PUF label composed of ten types of GERTs. a A bright-field image and b the scanning electron microscopy image (corresponds to the dashed square area in a of a PUF label fabricated by ten types of GERTs in an area of 100 × 100 μm2 on a SiO2 substrate. Scale bars are 20 (left) and 1 μm (right). c Spectral demultiplexing to quantify the abundance of ten types of GERTS at the position indicated by a green dashed box in panel d. The best-fit spectrum (black) is obtained by fitting the measured Raman spectrum (red) with ten pure reference Raman spectra using the non-negative least squares (NNLS) method. d 2D and e 3D plot for the readout of the PUF label by the Raman mapping (with a resolution of 10 × 10 pixels) of multiplexed measured Raman signals using the integrated areas of Raman bands from 802 to 1600 cm−1 (as indicated in the gray area in panel c). f 2D plot for the readout of the PUF label by the Raman mapping (with a resolution of 10 × 10 pixels) of demultiplexed Raman signals and 2D plot of the corresponding digitizations using g binary encoding and h quaternary encoding of Raman intensity levels at each pixel.CC BY · PubMed Central
Fig. 5 Reproducibility verification of PUF labels. A PUF label composed of 4-NBT GERTs in a 100 × 100 μm2 area is read with a resolution of 50 × 50 pixels and is digitized with a binary and b quaternary encoding of Raman intensity levels at each pixel for the first (left), second (middle), and third (right) measurement. Digitization of additional three labels numbered 2 (left), 3 (middle) and 4 (right) for c binary and d quaternary encoding of Raman intensity levels at each pixel with a resolution of 50 × 50 pixels. Pairwise match of ten PUF labels with e binary and f quaternary encoding of Raman intensity levels at each pixel. The x-axis and the y-axis represent the first and the second measurement of the labels, respectively, and the color bar shows the similarity index. Distribution of the similarity indexes (I) for the same PUF labels (red and blue bars) and those for different labels (orange and purple bars) of the training set (top) and the test set (bottom) in terms of g binary and h quaternary encoding of Raman intensity levels at each pixel. Figure e and f are plotted by referring to the ref. 17. Source data are provided as a Source Data file.
Fig. 5 Reproducibility verification of PUF labels. A PUF label composed of 4-NBT GERTs in a 100 × 100 μm2 area is read with a resolution of 50 × 50 pixels and is digitized with a binary and b quaternary encoding of Raman intensity levels at each pixel for the first (left), second (middle), and third (right) measurement. Digitization of additional three labels numbered 2 (left), 3 (middle) and 4 (right) for c binary and d quaternary encoding of Raman intensity levels at each pixel with a resolution of 50 × 50 pixels. Pairwise match of ten PUF labels with e binary and f quaternary encoding of Raman intensity levels at each pixel. The x-axis and the y-axis represent the first and the second measurement of the labels, respectively, and the color bar shows the similarity index. Distribution of the similarity indexes (I) for the same PUF labels (red and blue bars) and those for different labels (orange and purple bars) of the training set (top) and the test set (bottom) in terms of g binary and h quaternary encoding of Raman intensity levels at each pixel. Figure e and f are plotted by referring to the ref. 17. Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 6 High-speed readout of PUF labels. a Schematic of configurations of Raman system used for high-speed scanning. Under the DuoScan mode, two galvo mirrors rotate along two orthogonal axes in the mirror plane, controlling movements of the laser spot along the x- and the y-axis, respectively. b Two-dimensional (2D) (top) and 3D (bottom) plot for a PUF label (100 × 100 μm2) read in DuoScan mode in 6 s with a resolution of 50 × 50 pixels.
Fig. 6 High-speed readout of PUF labels. a Schematic of configurations of Raman system used for high-speed scanning. Under the DuoScan mode, two galvo mirrors rotate along two orthogonal axes in the mirror plane, controlling movements of the laser spot along the x- and the y-axis, respectively. b Two-dimensional (2D) (top) and 3D (bottom) plot for a PUF label (100 × 100 μm2) read in DuoScan mode in 6 s with a resolution of 50 × 50 pixels.CC BY · PubMed Central
Fig. 7 Demonstration of Raman PUF labels and their utilization in an anticounterfeiting system. a A PUF label fabricated using 4-NBT GETRs on Scotch tape and b the corresponding 2D and 3D plots of the readout by high-resolution (50 × 50 pixels) Raman mapping. The background (Scotch tape and the printing paper) exhibits a negligible Raman signal (e.g., point 1) and the PUF label emits strong signals from its 4-NBT GERTs (e.g., point 2). Scale bar in panel a is 1 mm. c Schematic illustration of Raman PUF labels applied in the supply chain with an anticounterfeiting system. The composition of figure (c) refers to the ref. 16.
Fig. 7 Demonstration of Raman PUF labels and their utilization in an anticounterfeiting system. a A PUF label fabricated using 4-NBT GETRs on Scotch tape and b the corresponding 2D and 3D plots of the readout by high-resolution (50 × 50 pixels) Raman mapping. The background (Scotch tape and the printing paper) exhibits a negligible Raman signal (e.g., point 1) and the PUF label emits strong signals from its 4-NBT GERTs (e.g., point 2). Scale bar in panel a is 1 mm. c Schematic illustration of Raman PUF labels applied in the supply chain with an anticounterfeiting system. The composition of figure (c) refers to the ref. 16.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 6 Not valid · 24 Not checked · 58 Own results and interpretations

30 more claims

Create a free account to see all 88 claims and to verify the 47 references still open.

See all claims — free account
  • It may state two separate claims as one. 38
  • Probably not a substantial claim. 18
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 58 references

Research works printed but not linked: 6

  1. Jung, L., Hayward, J. A., Liang, M. B. & Berrada, A., Vol. 9, 266, 370 (Google Patents, 2016).
  2. Sleat, R. & Van Lint, G., Vol. 8, 783, 194 (Google Patents, 2014).
  3. Kydd, P.H., Vol. 4, 441, 943 (Google Patents, 1984).
  4. Rührmair, U. et al. Revisiting optical physical unclonable functions. IACR Cryptol. ePrint Arch. 2013, 215 (2013).
  5. Kreyszig, E. Advanced Engineering Mathematics. 4th edn (John Wiley & Sons, Inc, 1979).
  6. Gassend, B., Clarke, D., Van Dijk, M. & Devadas, S. in Proceedings of the 9th ACM Conference on Computer and Communications Security. 148–160 (ACM, 2002).

Cited by

See all 268

Metadata

Authors
Gu, Y., He, C., Zhang, Y., Lin, L., Thackray, B. D., & Ye, J.
Venue
Nature Communications
Published
2020-01-24
Volume
11
Issue
1
Pages
516-516
ISSN
2041-1723 (ISSN-L)
OpenAlex
W3002885741
Cited by
268
References
58
Citations by year
2026: 36 · 2025: 51 · 2024: 56 · 2023: 52 · 2022: 41 · 2021: 23 · 2020: 9
FWCI
24.53
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders
  • National Natural Science Foundation of China (81871401, 81622026, 81571763)
  • Shanghai Municipal Education Commission (ZXWF082101, No. ZXWF082101)
  • Science and Technology Commission of Shanghai Municipality (19441905300)
  • Shanghai Jiao Tong University (YG2019QNA28, YG2016MS51, YG2017MS54)
  • Ruijin Hospital
  • Shanghai Key Laboratory of Gynecologic Oncology
Topics
  • Physical Sciences › Computer Science › Hardware and Architecture › Physical Unclonable Functions (PUFs) and Hardware Security
  • Physical Sciences › Materials Science › Electronic, Optical and Magnetic Materials › Gold and Silver Nanoparticles Synthesis and Applications
  • Physical Sciences › Computer Science › Information Systems › QR Code Applications and Technologies
Keywords
physical unclonable functions, Raman reporters, photostability, encoding capacity

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-19, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-19 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Gap-enhanced Raman tags for physically unclonable anticounterfeiting labels. https://app.cochat.ai/papers/10.1038/s41467-019-14070-9
The paper
Gu, Y., He, C., Zhang, Y., Lin, L., Thackray, B. D., & Ye, J. (2020). Gap-enhanced Raman tags for physically unclonable anticounterfeiting labels. Nature Communications. https://doi.org/10.1038/s41467-019-14070-9

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.