Highly Reproducible Au-Decorated ZnO Nanorod Array on a Graphite Sensor for Classification of Human Aqueous Humors.
Wansun KimSoo Hyun LeeSang Hun KimJae-Chul LeeSang Woong MoonJae Su YuSam Jin ChoiPublished in: ACS applied materials & interfaces (2017)
Gold-decorated, vertically grown ZnO nanorods (NRs) on a flexible graphite sheet (Au/ZnONRs/G) were developed for surface-enhanced Raman scattering (SERS)-based biosensing to identify trace amounts of human aqueous humors. This Au/ZnONRs/G SERS-functionalized sensor was fabricated via two steps: hydrothermal synthesis-induced growth of ZnO NRs on graphite sheets for nanostructure fabrication, followed by e-beam evaporator-induced gold metallization on ZnONRs/G for SERS functionalization. The thickness of the Au layer and the height of the ZnO NRs for enhancing SERS performance were adjusted to maximize Raman intensity, and the optimized Au/ZnONRs/G nanostructures were verified by the electric finite element computational models to maximize the electric fields. The proposed Au/ZnONRs/G SERS sensor showed an enhancement factor of 2.3 × 106 via rhodamine 6G Raman probe and excellent reproducibility (relative standard deviation of <10%) via Raman mapping of a SERS active area with a square of 100 × 100 μm2. To evaluate the actual bioapplicability of point-of-care-testing (POCT) analysis in clinics, SERS data acquisition was performed with an integration time of 1 s from a 1 μL analytic droplet of the sample. The performance of this Au/ZnONRs/G sensor was evaluated using human aqueous humors with cataract and two oxidative stress-induced eye diseases, age-related macular degeneration, and diabetic macular edema. These three eye diseases could be identified without any labeling or modification using the Au/ZnONRs/G SERS sensor and the computational algorithm incorporating a support vector machine and multivariate statistical prediction. Therefore, these findings indicate that our label-free, highly reproducible and flexible Au/ZnONRs/G SERS-functionalized sensor supported by a multivariate statistics-derived bioclassification method has great potential in POCT applications for identifying eye diseases.
Keyphrases
- sensitive detection
- quantum dots
- reduced graphene oxide
- gold nanoparticles
- label free
- endothelial cells
- raman spectroscopy
- high glucose
- machine learning
- deep learning
- high resolution
- visible light
- induced pluripotent stem cells
- oxidative stress
- room temperature
- mass spectrometry
- diabetic rats
- climate change
- big data
- silver nanoparticles
- liquid chromatography
- stress induced