Journal of Raman Spectroscopy最新文献

Convolutional neural networks (CNNs) and support vector machines (SVMs) have seen numerous applications within Raman spectroscopy. However, the age-old question remains: Which is better? To shine some light on the matter, the stability of the two machine learning techniques was probed by intentionally introducing spectral artefacts to transmission low wavenumber Raman spectroscopic data. The data consisted of synthetic microcalcifications buried under various depths of chicken breast. We found that an SVM yielded the best model with an area under curve (AUC) of 0.989 compared to 0.979 for the CNN. However, generally, SVMs were more susceptible to the spectral artefacts than CNNs. Additionally, the performance of CNNs and SVMs was not dependent on the magnitude of the shifts and stretches. An example is the linear stretches, where the AUC remained at 0.977 and 0.969 for both 2 and 5 cm⁻¹ shifts for the CNN and SVM models, respectively.

The potential of time-domain Raman spectroscopy in space exploration is discussed. This work is motivated by the emergence of robust, space-qualified femtosecond lasers and by the fact that time-domain detection allows the design of very compact instruments. As is shown, time-domain Raman spectroscopy gives access to the same fingerprint spectrum of minerals as conventional Raman spectroscopy, while avoiding problems such as fluorescence or ambient light backgrounds.

In this work, aqueous nutrient solutions replicating bioreactor culture media for microalgae were analyzed using spontaneous Raman spectroscopy. Focusing on nitrate, sulfate, glucose, and phosphate, the study evaluated their potential for real-time monitoring in cell cultivations such as Chlorella vulgaris. Univariate analysis, based on Raman intensities of specific nutrient peaks, was conducted and compared to multivariate analysis results. Four multivariate calibration models were developed using partial least squares regression (PLSR), achieving high calibration and validation performance, with R² values above 0.99 and low RMSE_CV, indicating strong calibration accuracy. The study also examined the limit of detection (LOD) for each nutrient, finding that LODs for nitrate, sulfate, and glucose reached levels relevant for algae bioreactors even without the application of enhanced Raman techniques. To further validate the PLS models, independent real bioreactor samples were analyzed, showing strong predictive accuracy (R_P²: 0.9661–0.9892) and low RMSE_P values. Additional testing with five samples collected over a Chlorella vulgaris cultivation run (day 0 to day 9) confirmed the models' robust performance under real bioprocess conditions. Limitations in practical applications, such as phosphate's relatively high LOD, were also identified. The results suggest that Raman spectroscopy, combined with multivariate analysis, could deliver precise and reliable detection of critical nutrients and their concentrations in bioreactor culture media. This potential of the Raman technique, along with insights into nutrient LODs, PLS model accuracy, and practical application challenges, provides a solid foundation for future research and development in industrial bioprocess monitoring.

In this study, we utilized surface-enhanced Raman scattering (SERS) to examine the conformational changes of cysteamine on silver bumpy nanoshell (AgNS) in various solvents, including water, ethanol (EtOH), isopropyl alcohol (IPA), and acetonitrile (ACN). It was found that the interaction between cysteamine and solvent molecules could affect the conformational equilibrium of cysteamine on the AgNS surface, particularly the ratio between its trans and gauche conformations. Using the SERS bands at 630 and 720 cm⁻¹ assigned to S–C stretching modes of the gauche and trans conformations, respectively, and calculating the SERS intensity ratio of these bands (I_trans/I_gauche), we found that water stabilized the gauche conformation because of the strong hydrogen-bonding interactions with cysteamine, while solvents such as ACN favored the trans conformation. SERS measurements over time further indicated that cysteamine reached conformational equilibrium on AgNS slowly within hours in a single solvent environment. In mixed water/ACN solvents, the gauche conformation became increasingly dominant with higher water fractions. Conversely, introducing ACN into water-stabilized AgNS dispersion did not reverse the conformational shift, suggesting that the gauche to trans transition was limited by higher activation barrier due to irreversible stabilization of Ag–N bond in the gauche form and local solvent environment around surface-bound cysteamine. Conformational equilibrium of cysteamine on AgNS could be quantified using the degree of mixing (γ), a parameter reflecting the distribution and interaction of cysteamine in mixed solvent. This study shows that SERS is effective for monitoring solvent-mediated conformational changes in surface-bound cysteamine. This approach might help us optimize solvent selection for applications that leverage the reactive amine group of cysteamine on metal surfaces.

Pyrrhotite (Fe_1-xS; 0 < x < 0.125) is one of major iron sulfides found in sedimentary rocks. While it is sometimes considered merely a gangue for metal mining, it is geo/cosmochemically important mineral because its composition (x) and superstructure sensitively reflect an aqueous environment when it precipitated. Analyzing such properties presents difficulties, as pyrrhotite in natural rocks is often as small as micrometers and is embedded in silicate matrices. Micro-Raman spectroscopy is a powerful tool for analyzing the crystal structure with high spatial resolution. While several researchers had reported distinctly different Raman spectra for pyrrhotite, we recently revealed that the Raman bands of pyrrhotite are extremely weak and what appear to be intense peaks likely originate from its (hydro) oxides. However, at the time, we barely observed only one Raman band at ~120 cm⁻¹ so that even its fundamental properties such as its polarization dependence and relative intensity to other bands could not be obtained. In the present study, we optimized our lab-built Raman spectrometer to probe weak Raman bands in the low-wavenumber region (< 150 cm⁻¹) to provide a reliable dataset for the Raman spectra of pyrrhotite. As a result, we succeeded in finding lattice modes at ~68 ($�A_g$), ~87 ($�B_g$), ~117 ($�A_g$), and ~230 cm⁻¹ ($�A_g$) with a sufficiently low baseline level for the first time, enabling us to also determine polarization dependence of their intensity.