Identification of microplastics using a convolutional neural network based on micro-Raman spectroscopy

A double-branch convolutional neural network model for species identification based on multi-modal data

For species identification analysis, methods based on deep learning are becoming prevalent due to their data-driven and task-oriented nature. The most commonly used convolutional neural network (CNN) model has been well applied in Raman spectra recognition. However, when faced with similar molecules or functional groups, the features of overlapping peaks and weak peaks may not be fully extracted using the CNN model, which can potentially hinder accurate species identification. Based on these practical challenges, the fusion of multi-modal data can effectively meet the comprehensive and accurate analysis of actual samples when compared with single-modal data. In this study, we propose a double-branch CNN model by integrating Raman and image multi-modal data, named SI-DBNet. In addition, we have developed a one-dimensional convolutional neural network combining dilated convolutions and efficient channel attention mechanisms for spectral branching. The effectiveness of the model has been demonstrated using the Grad-CAM method to visualize the key regions concerned by the model. When compared to single-modal and multi-modal classification methods, our SI-DBNet model achieved superior performance with a classification accuracy of 98.8%. The proposed method provided a new reference for species identification based on multi-modal data fusion.

Pretreatment-free SERS sensing of microplastics using a self-attention-based neural network on hierarchically porous Ag foams

Low-cost detection systems are needed for the identification of microplastics (MPs) in environmental samples. However, their rapid identification is hindered by the need for complex isolation and pre-treatment methods. This study describes a comprehensive sensing platform to identify MPs in environmental samples without requiring independent separation or pre-treatment protocols. It leverages the physicochemical properties of macroporous-mesoporous silver (Ag) substrates templated with self-assembled polymeric micelles to concurrently separate and analyze multiple MP targets using surface-enhanced Raman spectroscopy (SERS). The hydrophobic layer on Ag aids in stabilizing the nanostructures in the environment and mitigates biofouling. To monitor complex samples with multiple MPs and to demultiplex numerous overlapping patterns, we develop a neural network (NN) algorithm called SpecATNet that employs a self-attention mechanism to resolve the complex dependencies and patterns in SERS data to identify six common types of MPs: polystyrene, polyethylene, polymethylmethacrylate, polytetrafluoroethylene, nylon, and polyethylene terephthalate. SpecATNet uses multi-label classification to analyze multi-component mixtures even in the presence of various interference agents. The combination of macroporous-mesoporous Ag substrates and self-attention-based NN technology holds potential to enable field monitoring of MPs by generating rich datasets that machines can interpret and analyze.

Tracking organic matrix in the seashell by elemental mapping under laser-induced breakdown spectroscopy

As a biogenic calcium carbonate, the seashell plays a crucial role in marine environmental studies. In these studies, it is essential to investigate the composition of the seashell. In this study, we used laser-induced breakdown spectroscopy (LIBS) to analyze the elemental composition of cultured scallop-shell (Patinopecten yessoensis), with a specific focus on examining the organic elements (C, N, O, H) to track the shell organic matrix (SOM). Our findings indicate that the seashell organic layer can be accurately identified by referencing the strong emission of nitrogen or the low signal of calcium. To further confirm the presence of this layer, we employed fluorescence spectroscopy, Raman spectroscopy and FTIR spectroscopy. Correlation analysis revealed a strong connection between LIBS emissions (H, O, CC) and seashell organics, as well as demonstrated the presence of organics in metallic emissions (Si, Ba). However, when we conducted elemental mapping on the shell cross-section, the distribution similarity was observed between the elements N, Ba, and Sr. Based on the correlation of organics and the distribution similarity, it is concluded that barium is an element associated with the SOM. These results highlight the potential of LIBS for organic analysis, which can complement traditional seashell analysis.

Fast Detection and Classification of Microplastics below 10 μm Using CNN with Raman Spectroscopy

In light of the growing awareness regarding the ubiquitous presence of microplastics (MPs) in our environment, recent efforts have been made to integrate Artificial Intelligence (AI) technology into MP detection. Among spectroscopic techniques, Raman spectroscopy is preferred for the detection of MP particles measuring less than 10 μm, as it overcomes the diffraction limitations encountered in Fourier transform infrared (FTIR). However, Raman spectroscopy's inherent limitation is its low scattering cross section, which often results in prolonged data collection times during practical sample measurements. In this study, we implemented a convolutional neural network (CNN) model alongside a tailored data interpolation strategy to expedite data collection for MP particles within the 1-10 μm range. Remarkably, we achieved the classification of plastic types for individual particles with a mere 0.4 s of exposure time, reaching an approximate confidence level of 85.47(±5.00)%. We postulate that the result significantly accelerates the aggregation of microplastic distribution data in diverse scenarios, contributing to the development of a comprehensive global microplastic map.

Deep metric learning framework combined with Gramian angular difference field image generation for Raman spectra classification based on a handheld Raman spectrometer

Rapid identification of unknown material samples using portable or handheld Raman spectroscopy detection equipment is becoming a common analytical tool. However, the design and implementation of a set of Raman spectroscopy-based devices for substance identification must include spectral sampling of standard reference substance samples, resolution matching between different devices, and the training process of the corresponding classification models. The process of selecting a suitable classification model is frequently time-consuming, and when the number of classes of substances to be recognised increases dramatically, recognition accuracy decreases dramatically. In this paper, we propose a fast classification method for Raman spectra based on deep metric learning networks combined with the Gramian angular difference field (GADF) image generation approach. First, we uniformly convert Raman spectra acquired at different resolutions into GADF images of the same resolution, addressing spectral dimension disparities induced by resolution differences in different Raman spectroscopy detection devices. Second, a network capable of implementing nonlinear distance measurements between GADF images of different classes of substances is designed based on a deep metric learning approach. The Raman spectra of 450 different mineral classes obtained from the RRUFF database were converted into GADF images and used to train this deep metric learning network. Finally, the trained network can be installed on an embedded computing platform and used in conjunction with portable or handheld Raman spectroscopic detection sensors to perform material identification tasks at various scales. A series of experiments demonstrate that our trained deep metric learning network outperforms existing mainstream machine learning models on classification tasks of different sizes. For the two tasks of Raman spectral classification of natural minerals of 260 classes and Raman spectral classification of pathogenic bacteria of 8 classes with significant noise, our suggested model achieved 98.05% and 90.13% classification accuracy, respectively. Finally, we also deployed the model in a handheld Raman spectrometer and conducted identification experiments on 350 samples of chemical substances attributed to 32 classes, achieving a classification accuracy of 99.14%. These results demonstrate that our method can greatly improve the efficiency of developing Raman spectroscopy-based substance detection devices and can be widely used in tasks of unknown substance identification.

Component identification for the SERS spectra of microplastics mixture with convolutional neural network

With the increasing interest in microplastics (MPs) pollutants, relevant detection technologies are also developing. In MPs analysis, vibrational spectroscopy represented by surface-enhanced Raman spectroscopy (SERS) is widely used because they can provide unique fingerprint characteristics of chemical components. However, it is still a challenge to separate various chemical components from the SERS spectra of MPs mixture. In this study, it is innovatively proposed to combine the convolutional neural networks (CNN) model to simultaneously identify and analyze each component in the SERS spectra of six common MPs mixture. Different from the traditional method, which requires a series of spectral preprocessing such as baseline correction, smoothing and filtering, the average identification accuracy of MP components is as high as 99.54 % after the unpreprocessed spectral data is trained by CNN, which is better than other classical algorithms such as support vector machine (SVM), principal component analysis linear discriminant analysis (PCA-LDA), partial least squares discriminant analysis (PLS-DA), Random Forest (RF), and K Near Neighbor (KNN), with or without spectral preprocessing. The high accuracy shows that CNN can be used to quickly identify MPs mixture with unpreprocessed SERS spectra data.