Single-model multi-tasks deep learning network for recognition and quantitation of surface-enhanced Raman spectroscopy

Surface-enhanced Raman scattering (SERS) spectroscopy analysis has long been the central task of nanoscience and nanotechnology to realize the ultrasensitive recognition/quantitation applications. Recently, the blooming of artificial intelligence algorithms provides an edge tool to revolutionize the spectroscopy analysis, especially for multiple substances analysis and large-scale data handling. In this study, a single-model multi-tasks deep learning network is proposed to simultaneously achieve the qualitative recognition and quantitative analysis of SERS spectroscopy. The SERS spectra of two kinds of hypoglycemic drugs (PHE, ROS) and the corresponding mixtures are collected, respectively, with the concentration grade from 10^-4 M to 10^-8 M. Based on the SERS spectroscopy dataset, the loss functions and hyperparameters of the multi-tasks classifications model are optimized, and the recognition accuracies are tested by simulation experiments. It is demonstrated that the accuracy rates of qualitative and quantitative analysis can reach up to 99.0% and 98.4%, respectively. Moreover, the practical feasibility of this multi-tasks model is demonstrated by using it to achieve qualitative and quantitative analysis of PHE and ROS in complex serum matrix. Overall, this single-model multi-tasks deep learning network shows significant potential for the recognition and quantitation of SERS spectroscopy, which provides the algorithmic and experimental basis for large-scale and multiple substances SERS spectra analysis.

Silicon-assisted surface enhanced fluorescence toward improved assay performances

A novel scheme of silicon-assisted surface enhanced fluorescence (SEF) is presented for SEF-based assays, where the blank signal suppression and the fluorescence signal enhancement is combined. The P-doped, (100) oriented silicon substrate is used to quench the fluorescence of Rose Bengal (RB) molecules attached to it, resulting in an effectively suppressed background signal, which is useful for a lower limit of detection (LOD). When a proper quantity of silver nanoparticles (AgNPs) is deposited on the RB-attached silicon substrate, a significant fluorescence enhancement of up to around 290 fold is obtained, which helps to improve the sensitivity in fluorescence-based assays. Besides, conventional gold nanoparticles (AuNPs) have also been demonstrated to exhibit excellent SEF effect using the presented scheme, providing improved stability and biocompatibility. The mechanism of the observed SEF effect has been investigated, and both the decreased apparent quantum yield and the silicon-induced electric field redistribution are considered to play important roles. The experimental results suggest that the presented scheme holds great potential in the SEF-based assays aiming at higher sensitivity and lower LOD.

A straightforward and sensitive “ON–OFF” fluorescence immunoassay based on silicon-assisted surface enhanced fluorescence

A straightforward immunoassay based on silicon-assisted surface enhanced fluorescence (SEF) has been demonstrated using a silicon-based fluorescent immune substrate and silver-antibody nanoconjugate (SANC). The P-doped, (100) oriented silicon wafers are used for both fluorophore attachment and antigen immobilization. The silicon substrate offers a very low blank signal in the “OFF” state, due to its fluorescence quenching effect. In the detection process, the capture of the SANCs by the surface-immobilized antigen leads to an effectively enhanced fluorescence to produce an “ON” state. The analytical performance of the presented scheme has been investigated and a limit of detection of 31.4 pg mL⁻¹ has been obtained. Besides the broadened application range compared with the conventional immunoassays, the presented scheme is straightforward, cost effective and sensitive, and is hence expected to find widespread applications in immunoassays as well as other fluorescence-based assays.

A straightforward immunoassay based on silicon-assisted surface enhanced fluorescence (SEF) has been demonstrated using a silicon-based fluorescent immune substrate and silver-antibody nanoconjugate (SANC).

Manipulating "Hot Spots" from Nanometer to Angstrom: Toward Understanding Integrated Contributions of Molecule Number and Gap Size for Ultrasensitive Surface-Enhanced Raman Scattering Detection

Narrower gaps between metal nanoparticles (so-called "hot spots") in surface-enhanced Raman scattering (SERS) substrates contribute to stronger electromagnetic (EM) enhancement; however, the accompanying steric effect hinders analyte molecules entering hot spots to access the benefit. To comprehensively understand integrated contributions of the gap size and molecule number accommodated in hot spots and then optimize design of SERS substrates, the thermal shrinking method was employed to manipulate hot spots and the "hottest zone" was defined to evaluate the integrated contributions to SERS intensity of the two factors. In the conventional shrink-adsorption mode, the contributions of the molecule number and gap size are competitive when the gap width is comparable with the target molecule size, which leads to oscillating behavior of SERS intensity versus gap size, and it is analyte molecule size dependent. This result suggests that engineering hot spots should be target molecule directed to achieve ultrasensitive detection. In the proposed adsorption-shrink mode, the contributions of the molecule number and gap size are synergistic, which makes the detection ability of the adsorption-shrink mode attains a single-molecule (SM) level. Excellent performance of the adsorption-shrink SERS strategy benefits detection of trace level pollutants in complex environments. Detection ranges for contaminants with different metal affinity, such as thiram, malachite green (MG), and formaldehyde, are as low as parts per billion, even down to parts per trillion.

Facing Challenges in Real-Life Application of Surface-Enhanced Raman Scattering: Design and Nanofabrication of Surface-Enhanced Raman Scattering Substrates for Rapid Field Test of Food Contaminants

Surface-enhanced Raman scattering (SERS) is capable of detecting a single molecule with high specificity and has become a promising technique for rapid chemical analysis of agricultural products and foods. With a deeper understanding of the SERS effect and advances in nanofabrication technology, SERS is now on the edge of going out of the laboratory and becoming a sophisticated analytical tool to fulfill various real-world tasks. This review focuses on the challenges that SERS has met in this progress, such as how to obtain a reliable SERS signal, improve the sensitivity and specificity in a complex sample matrix, develop simple and user-friendly practical sensing approach, reduce the running cost, etc. This review highlights the new thoughts on design and nanofabrication of SERS-active substrates for solving these challenges and introduces the recent advances of SERS applications in this area. We hope that our discussion will encourage more researches to address these challenges and eventually help to bring SERS technology out of the laboratory.