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

The DOI matches its Crossref record; 41 of the 42 references it cites were found in a registry or an index; none of the 22 sources its claims cite is checked for support yet.

Bright and stable anti-counterfeiting devices with independent stochastic processes covering multiple length scales

Open access

Abstract

Physical unclonable functions (PUFs) are considered the most promising approach to address the global issue of counterfeiting. Current PUF devices are often based on a single stochastic process, which can be broken, especially since their practical encoding capacities can be significantly lower than the theoretical value. Here we present stochastic PUF devices with features across multiple length scales, which incorporate semiconducting polymer nanoparticles (SPNs) as fluorescent taggants. The SPNs exhibit high brightness, photostability and size tunability when compared to the current state-of-the-art taggants. As a result, they are easily detectable and highly resilient to UV radiation. By embedding SPNs in photoresists, we generate PUFs consisting of nanoscale (distribution of SPNs within microspots), microscale (fractal edges on microspots), and macroscale (random microspot array) designs. With the assistance of a deep-learning model, the resulting PUFs show both near-ideal performance and accessibility for general end users, offering a strategy for next-generation security devices.

Overview

CoChat Summary

TL;DR

The paper presents optical physical unclonable function (PUF) anti-counterfeiting devices that combine semiconducting polymer nanoparticles (SPNs) with photoresist patterning to create independent stochastic features at nanoscale, microscale, and macroscale. The devices show near-ideal PUF metrics, high photostability, and deep-learning-assisted authentication, with an encoding capacity of ca. 10^220322 from a 1 mm² area.

Key findings

  • SPNs, especially F8BT, outperformed carbon dots, CdSe:ZnS quantum dots, and CsPbBr3 perovskite nanocrystals in brightness and photostability under identical confocal conditions; Qdot intensity was two to ten times lower than SPNs of comparable size, and F8BT NPs showed superior photostability over 500 s of continuous laser scanning.
  • SPN size was tunable via surfactant/water/ethanol conditions: F8BT:PS-PEG at 1:1 maintained ~250 nm in 10-fold ethanol, while a 5:1 ratio precipitated into 1:5 THF:water aggregated to ~600 nm and gave clearer, larger spots detectable at 10% laser power.
  • A blend of blue, green, and red SPNs with emission maxima at 440, 535, and 635 nm enabled multichannel PUFs in SU-8; each microspot yielded 216 bits, and 50 microspots produced 10,800-bit keys with bit uniformity 0.4948 ± 0.0064, inter-device Hamming distance 0.5000 ± 0.0427, and intra-device Hamming distance 0.05568 ± 0.0181.
  • Additional stochastic processes—fractal edges and randomly delaminated microspot arrays—were incorporated; fluorescence channels alone gave an encoding capacity of ca. 10^220322 from 1 mm², and devices retained >90% reliability after one week at 40 °C, −20 °C, or 85% humidity, with no significant fluorescence decrease after 17 and 24 h UV exposure.

Methods

  • F8BT-red was synthesized by Suzuki–Miyaura coupling; SPNs were prepared by mixing F8BT, PFO, or F8BT-red with PS-g-FEO surfactant in THF, adding water, evaporating THF, and characterizing size by DLS.
  • Photolithography used SU-8 2002/2010 and AZ 4562 photoresists; standard processing included spin-coating, soft bake, 365 nm UV exposure, post-exposure bake, and development, while fractal edges were generated with SU-8 2002:AZ 4562 blends and microspot delamination was induced by sonication in developer.
  • Fluorescence imaging and photobleaching used confocal microscopy at 405 nm; PUF keys were generated by binarizing fluorescent images with 10 × 10 pixel units for F8BT/F8BT-red and 50 × 50 pixels for PFO, and identification used the LoFTR deep-learning model; stability tests included artificial sweat (0.9 g/L NaCl, 0.2 g/L KCl, 30 min), UV irradiation, and aging at 40 °C, −20 °C, and 85% humidity.

Limitations

  • The paper does not state an explicit limitations section; evident from the text, the encoding capacities of the microspot array and fractal edges could not compete with the SPN fluorescent patterns, though their readout was more straightforward and convenient.
  • Evident from the text, SPN aggregates grew larger four hours after ethanol dilution, requiring thin-film formation shortly after aggregate formation, and initial 10–100 nm particles were not ideal for practical readout due to low signal-to-noise ratios.
  • Evident from the text, PVA and PS films failed in artificial sweat, AZ-only microstructures were not stable and could be washed away, and some photoresist blends required too high an exposure dose or were unstable under lower UV exposure.

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

Integrity signals

Registry

Crossref · 10.1038/s41467-024-55646-4

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Design of the PUF platform by material and process engineering. a SPNs, with RGB emissions under the same wavelength excitation, are generated and aggregated in controlled sizes. The SPNs are mixed with photoresists and treated by a photolithography process. This cross-links the photoresists to improve the distribution stability of SPNs. b SPNs are randomly distributed in each microspot. Fractal edges and partially removed arrays are engineered during photolithography. The PUF devices can be applied as anticounterfeiting labels for different products. c Machine learning assisted identification process for end-user friendly security devices.
Fig. 1 Design of the PUF platform by material and process engineering. a SPNs, with RGB emissions under the same wavelength excitation, are generated and aggregated in controlled sizes. The SPNs are mixed with photoresists and treated by a photolithography process. This cross-links the photoresists to improve the distribution stability of SPNs. b SPNs are randomly distributed in each microspot. Fractal edges and partially removed arrays are engineered during photolithography. The PUF devices can be applied as anticounterfeiting labels for different products. c Machine learning assisted identification process for end-user friendly security devices.CC BY · PubMed Central
Fig. 2 Characterization and optimization of fluorescent NPs for PUF devices. Confocal images of a F8BT (poly(9,9-dioctylfluorene-alt-benzothiadiazole) NPs (Ex: 405 nm, laser power: 15%, gain: 150%), b perovskite NPs (CsPbBr3) (Ex: 405 nm, laser power: 15%, gain: 150%), and c carbon dots (Ex: 405 nm, laser power: 15%, gain: 500%), before and after photobleaching. The red square indicates in the area which was irradiated. d Photobleaching curves of nanoparticles imaged every 15 s under 80% laser power. Two shades of each colour indicate curves extracted from two separate regions of interest. e DLS analysis of F8BT NPs (n = 3, N = 3, S: surfactant, poly(styrene)-graft-poly(ethylene oxide)). Detailed results for each independent batch can be found in Fig. S8. The green arrows indicate nanoparticle batches which were cast as thin films, with their corresponding confocal images on the right (10% laser power and a gain value of 50% (Ex: 430 nm). Confocal images of f F8BT-PVA, g F8BT-PS, and h F8BT-SU-8 films before and after the incubation with artificial sweat (F8BT: Surfactant = 5:1, Ex: 430 nm, laser power: 10%, gain: 50%). Source data are provided as a Source Data file.
Fig. 2 Characterization and optimization of fluorescent NPs for PUF devices. Confocal images of a F8BT (poly(9,9-dioctylfluorene-alt-benzothiadiazole) NPs (Ex: 405 nm, laser power: 15%, gain: 150%), b perovskite NPs (CsPbBr3) (Ex: 405 nm, laser power: 15%, gain: 150%), and c carbon dots (Ex: 405 nm, laser power: 15%, gain: 500%), before and after photobleaching. The red square indicates in the area which was irradiated. d Photobleaching curves of nanoparticles imaged every 15 s under 80% laser power. Two shades of each colour indicate curves extracted from two separate regions of interest. e DLS analysis of F8BT NPs (n = 3, N = 3, S: surfactant, poly(styrene)-graft-poly(ethylene oxide)). Detailed results for each independent batch can be found in Fig. S8. The green arrows indicate nanoparticle batches which were cast as thin films, with their corresponding confocal images on the right (10% laser power and a gain value of 50% (Ex: 430 nm). Confocal images of f F8BT-PVA, g F8BT-PS, and h F8BT-SU-8 films before and after the incubation with artificial sweat (F8BT: Surfactant = 5:1, Ex: 430 nm, laser power: 10%, gain: 50%). Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 3 Encoding process and characterizations of multichannel PUF devices. a Absorption (solid line) and emission (dotted line) spectra of F8BT, F8BT-red, and PFO NPs in ethanol solutions. b Optical image of the printed spot arrays with different spot sizes (20 µm diameter on the top and 50 µm diameter on the bottom). c Four representative fluorescence images from the spots embedded with a blend of F8BT, F8BT-red, and PFO NPs (Blue channel: Ex 405 nm, Em 430–470 nm, 30% laser power, 150% gain. Green channel: Ex 405 nm, Em 470–550 nm, 20% laser power, 100% gain. Red channel: Ex 405 nm, Em 570–700 nm, 20% laser power, 100% gain). d Binarization of the area illustrated by the white square in c). e Digitized security keys with 10,800 bits are generated from 50 different microspots. f Inter- and intra-device Hamming distance of each PUF response. g Cumulative distribution functions, illustrating the probabilities authentication error (AE) and false authentication (FA) as a function of the decision threshold. h Heat map of intra-device (diagonal) and inter-device correlations obtained from 100 unique PUF patterns (generated using identical parameters). Source data are provided as a Source Data file.
Fig. 3 Encoding process and characterizations of multichannel PUF devices. a Absorption (solid line) and emission (dotted line) spectra of F8BT, F8BT-red, and PFO NPs in ethanol solutions. b Optical image of the printed spot arrays with different spot sizes (20 µm diameter on the top and 50 µm diameter on the bottom). c Four representative fluorescence images from the spots embedded with a blend of F8BT, F8BT-red, and PFO NPs (Blue channel: Ex 405 nm, Em 430–470 nm, 30% laser power, 150% gain. Green channel: Ex 405 nm, Em 470–550 nm, 20% laser power, 100% gain. Red channel: Ex 405 nm, Em 570–700 nm, 20% laser power, 100% gain). d Binarization of the area illustrated by the white square in c). e Digitized security keys with 10,800 bits are generated from 50 different microspots. f Inter- and intra-device Hamming distance of each PUF response. g Cumulative distribution functions, illustrating the probabilities authentication error (AE) and false authentication (FA) as a function of the decision threshold. h Heat map of intra-device (diagonal) and inter-device correlations obtained from 100 unique PUF patterns (generated using identical parameters). Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 4 Generation and characterization of stochastic processes across multiple length scales. (a) Real-time observation of the microstructures formed by the decomposition of positive photoresist AZ 4562. The exposure process was performed by hard contact under 365 nm UV irradiation with an intensity of 3 mW/cm2. b) Fractal structures with different complexity were generated by fine-tuning the lithography parameters. The ratios of SU 8 2002 to AZ 4562 and UV exposure time from left to right are 6:1, 120 s; 5:1, 120 s; 4:1, 120 s; 5:1, 80 s respectively. c) Stochastically distributed photoresist spot arrays can be binarized and encoded. d) The entropy along the x-axis and y-axis of the binarized bits, showing the randomness of the stochastic process. Source data are provided as a Source Data file.
Fig. 4 Generation and characterization of stochastic processes across multiple length scales. (a) Real-time observation of the microstructures formed by the decomposition of positive photoresist AZ 4562. The exposure process was performed by hard contact under 365 nm UV irradiation with an intensity of 3 mW/cm2. b) Fractal structures with different complexity were generated by fine-tuning the lithography parameters. The ratios of SU 8 2002 to AZ 4562 and UV exposure time from left to right are 6:1, 120 s; 5:1, 120 s; 4:1, 120 s; 5:1, 80 s respectively. c) Stochastically distributed photoresist spot arrays can be binarized and encoded. d) The entropy along the x-axis and y-axis of the binarized bits, showing the randomness of the stochastic process. Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 5 Authentication process, practical application and stability test. a) Workflow for the generation of challenge and response pairs, and machine learning assisted validation. Created in BioRender. Zhang, J. (2024) BioRender.com/i51y640. b) The optical PUF devices can be attached to different products including circuit boards, medicine packaging, and paper certificates for practical applications. c) Photostability test of the PUF devices. UV wavelength: 365 nm. UV intensity: 12 W/cm2. Green channel: Ex 405 nm, Em 470–550 nm, 20% laser power, 100% gain. d) Match ratios between UV irradiated PUF and the original PUF, based on the feature detecting results from the LoFTR algorithm. e) Bit error rate for each PUF under the storage conditions; f) reliability for each PUF in each storage condition. Values compared to images taken at day 0 and after one week. 100 PUFs were recorded per storage condition. Source data are provided as a Source Data file.
Fig. 5 Authentication process, practical application and stability test. a) Workflow for the generation of challenge and response pairs, and machine learning assisted validation. Created in BioRender. Zhang, J. (2024) BioRender.com/i51y640. b) The optical PUF devices can be attached to different products including circuit boards, medicine packaging, and paper certificates for practical applications. c) Photostability test of the PUF devices. UV wavelength: 365 nm. UV intensity: 12 W/cm2. Green channel: Ex 405 nm, Em 470–550 nm, 20% laser power, 100% gain. d) Match ratios between UV irradiated PUF and the original PUF, based on the feature detecting results from the LoFTR algorithm. e) Bit error rate for each PUF under the storage conditions; f) reliability for each PUF in each storage condition. Values compared to images taken at day 0 and after one week. 100 PUFs were recorded per storage condition. Source data are provided as a Source Data file.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 4 Not valid · 14 Not checked · 86 Own results and interpretations

56 more claims

Create a free account to see all 104 claims and to verify the 21 references still open.

See all claims — free account
  • It may state two separate claims as one. 33
  • Probably not a substantial claim. 14

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.

4 of 4

Citations

Cites

Show all 42 references

Research works printed but not linked: 4

  1. Becker, G. T. 535-555 (Springer Berlin Heidelberg).
  2. Sun, J., Shen, Z., Wang, Y., Bao, H. & Zhou, X. in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 8918-8927 (IEEE Computer Society, 2021).
  3. J. Sun, Z. S., Y. Wang, H. Bao, X. Zhou. in ArXiv. https://arxiv.org/abs/2104.00680 (2021). · arXiv
  4. Veith-Wolf, B., Witteck, R., Morlier, A., Schulte-Huxel, H. & Schmidt, J. in IEEE 43rd Photovolt. Spec. Conf. 1173-1178.

Cited by

See all 33

Metadata

Authors
Zhang, J., Creamer, A., Xie, K., Tang, J., Salter, L., Wojciechowski, J. P., & Stevens, M. M.
Affiliations
  • University of Oxford (GB)
  • Imperial College London (GB)
Corresponding author
Stevens, Molly M.
Venue
Nature Communications
Published
2025-01-08
Volume
16
Issue
1
Pages
502-502
ISSN
2041-1723 (ISSN-L)
OpenAlex
W4406170775
Cited by
33
References
42
Citations by year
2026: 22 · 2025: 11
FWCI
12.05
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • Oxford University Research Archive (ORA) (University of Oxford) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • Wellcome Trust (104931/Z/14/Z)
  • UK Research and Innovation (MR/R015651/1, BB/L015129/1, BB/T017929/1)
  • Cancer Research UK (100063)
  • Research Councils UK (EP/R00529X/1, EP/L016702/1)
  • Imperial College London (BB/T017929/1)
  • Rosetrees Trust
  • Medical Research Council (MR/R015651/1)
  • Engineering and Physical Sciences Research Council (EP/X027287/1)
  • Biotechnology and Biological Sciences Research Council (BB/L015129/1, L015129, BB/T017929/1)
Topics
  • Physical Sciences › Computer Science › Hardware and Architecture › Physical Unclonable Functions (PUFs) and Hardware Security
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Chaos-based Image/Signal Encryption
  • Physical Sciences › Computer Science › Information Systems › QR Code Applications and Technologies
Keywords
physical unclonable functions, anti-counterfeiting, semiconducting polymer nanoparticles, photostability, size tunability, deep learning, encoding capacity, UV resistance, security inks

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: Bright and stable anti-counterfeiting devices with independent stochastic processes covering multiple length scales. https://app.cochat.ai/papers/10.1038/s41467-024-55646-4
The paper
Zhang, J., Creamer, A., Xie, K., Tang, J., Salter, L., Wojciechowski, J. P., & Stevens, M. M. (2025). Bright and stable anti-counterfeiting devices with independent stochastic processes covering multiple length scales. Nature Communications. https://doi.org/10.1038/s41467-024-55646-4

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.