Spectroscopy-11-01-2020

The advantages of machine-learning methods have been widely explored in Raman spectroscopy analysis. In this study, a lightweight network model for mineral analysis based on Raman spectral feature visualization is proposed. The model, called the fire module convolutional neural network (FMCNN), was based on a convolutional neural network, and a fire-module was introduced to increase the width of the network, while also ensuring fewer trainable parameters in the network and reducing the model’s computational complexity. The visualization process is based on a deconvolution network, which maps the features of the middle layer back to the feature space. While fully exploring the features of the Raman spectral data, it also transparently displays the neural network feature extraction results. Experiments show that the classification accuracy of the model reaches 0.988. This method can accurately classify Raman spectra of minerals with less reliance on human participation. Combined with the analysis of the results of feature visualization, our method has high reliability and good application prospects in mineral classification.

Articles in this column have addressed five main areas: theory, functional groups containing the C-H bond, those containing the C-O bond, those with the C=O bond, and those with organic nitrogen compounds. Here, we review the concepts.

Low frequency Raman scattering measurements can be used to predict physical properties of polymers and the crystalline polymorphic form of active pharmaceutical ingredients (APIs). These measurements are made by recording the Stokes and anti-Stokes side of the laser line with the laser centered on the detector. Spectra of polyethylene and linear alkanes were recorded down to 4 cm-1.

Two recent warning letters show that the US FDA is substantially increasing the amount of remediation work it requires for companies to correct data integrity noncompliance. That work can be very expensive—far exceeding the cost of ensuring compliance in the first place.

Test firing a firearm is frequently used for forensic firearms and bullet identification. Airborne lead-containing particles are emitted when a firearm is tested, leading to lead building up on surfaces, exposing employees to potential lead-related health risks. Prior to cleaning, lead surface concentrations in the firing range at the National Forensic Laboratory Services in Ottawa were found to be higher than the Environmental Abatement Council of Ontario (EACO) post-abatement limit, with the highest level 56 times the limit. Inductively coupled plasma–mass spectrometry (ICP-MS), along with internal standardization, revealed that wiping surfaces with either a commercial decontamination product containing ethylene glycol butyl ether (EGBE) or alcohol cleaning pads satisfied the EACO standard by removing over 90% of lead from test surfaces whereas an external cleaning company only removed 36% of lead from the same surfaces. Fortunately, lead cross-contamination was minimal outside the firearms section and well below the residential EACO limit.

Fungal infections and mycotoxin contamination in food products pose a major threat to the world population. Mycotoxins contaminate approximately 25% of the world’s food products and cause severe health problems through the utilization of affected food products. The major mycotoxins in different foods are aflatoxins, ochratoxins, fumonisins, zearalenone, trichothecenes, and deoxynivalenol. Today, various conventional and nondestructive techniques are available for the detection of mycotoxins across multiple food products. Conventional methods are time-consuming, require chemical reagents, and include many laborious steps. Therefore, nondestructive techniques like near-infrared (NIR) spectroscopy, Fourier transform infrared (FT-IR) spectroscopy, hyperspectral imaging, and the electronic nose are a priority for online detection of fungal and mycotoxin problems in different food products. In this article, we discuss recent improvements and utilization of different nondestructive techniques for the early detection of fungal and mycotoxin infections in various food products.