Applied Spectroscopy最新文献

In the post-fire stage, precipitation and superficial incorporation of ashes alter the chemical properties of the soil. This study evaluated the combined effects of spectral preprocessing methods, data partitioning strategies, and modeling approaches on soil pH prediction using a portable near-infrared (NIR) spectrometer in wildfire ash-enriched soil. A laboratory column experiment was conducted using disturbed sandy loam soil, in which wildfire ashes were incorporated. The experimental design considered five treatments (n = 3) of Eucalyptus globulus and Quillaja saponaria ash incorporations (C: no ash; T1: 2% ash at 2.5  cm; T2: 2% ash at 5  cm; T3: 4% ash at 2.5  cm; T4: 4% ash at 5  cm). After simulating a precipitation of 20  mm h^-1 for 6 hours, the soil columns were sampled at 5 depths (D1: 2-3  cm, D2: 7-8  cm, D3: 12-13  cm, D4: 16-17  cm, D5: 20-21  cm). The samples were analyzed using a NIR spectrometer (range: 1350-2550  nm), and the levels of pH (1:2.5) were determined in the laboratory. Eight preprocessing techniques (P0 to P7) were tested, including absorbance conversion, mean centering, trimming, smoothing, standard normal variate (SNV), moving window average (MWA), Savitzky-Golay filtering, and first derivative transformation. Using the Kennard-Stone method, 70% of the data was used for calibration (CAL) and 30% for validation (VAL), considering two partitioning approaches, the same partition by pseudo absorbance values (Scenario A) and different partitions by preprocessing method (Scenario B). Partial least square (PLS) and random forest (RF) models were applied, and performance was assessed using root mean square error (RMSE), coefficient of determination (r²), and ratio of performance to interquartile distance (RPIQ) analyses. The most accurate pH predictions were achieved with RF under Scenario B using trimming + standard normal variate (SNV) + moving weighted average (MWA) preprocessing, yielding r² values of 0.95 (CAL) and 0.91 (VAL), with RMSEs of 0.23 (CAL) and 0.57 (VAL), and RPIQs of 4.33 (CAL) and 4.61 (VAL). Overall, portable NIR spectroscopy demonstrated strong potential for soil pH prediction in ash-enriched soil, emphasizing the critical role of appropriate spectral preprocessing to avoid overfitting. These findings provide insights into applying portable NIR spectroscopy as a cost-effective tool for monitoring soil pH following wildfires.

Surface-enhanced Raman scattering (SERS) spectroscopy represents a powerful analytical platform that combines non-destructive, label-free molecular identification with exceptional sensitivity for trace-level detection. Its capacity to generate information-rich spectral fingerprints makes SERS particularly advantageous for simultaneous multi-analyte analysis across diverse sample matrices, including complex biological systems. This study addresses the analytical challenges associated with identifying and quantifying multiple molecular species in complex environments by integrating SERS with advanced machine learning methodologies. We developed a hierarchical analytical framework that leverages the complementary strengths of deep learning and regression techniques: A multi-label convolutional neural network (CNN) for discriminating structurally similar analytes from SERS spectral data, coupled with a support vector regression (SVR) model for semi-quantitative determination of relative concentration ratios among identified species. The methodology was systematically validated using binary mixtures of short-chain fatty acids (SCFAs) as representative biomolecular targets, with performance rigorously benchmarked against established multivariate statistical methods and conventional machine learning approaches. Experimental validation demonstrated robust classification accuracy for both analytes at physiologically relevant concentrations, maintaining consistent performance across simple aqueous media and complex cell culture environments. These results establish the viability of the integrated SERS-CNN-SVR approach for advanced mixture analysis applications where precise identification and quantification of multiple biomarkers is essential.

Clinical applications of Raman spectroscopy (RS) typically rely on fiber optic probes that directly interface with the tissue site. These devices are designed with small diameters, enabling them to navigate narrow body cavities and seamlessly integrate into routine medical instruments. However, the performance of conventional RS fiber probes suffers during noncontact operation due to decreased collection efficiency and a larger laser spot size that restricts spatial precision. To address these limitations, a novel RS probe design is presented here that can efficiently collect both fingerprint (FP) and high-wavenumber (HW) regions of the Raman spectrum at an offset from the target tissue using a miniature lens at the probe tip. The development process began with stochastic light propagation simulations that served as a foundation for the device's expected performance improvements compared to a standard RS probe design, which were then experimentally verified. Lenses were fabricated from various materials, including fused silica, quartz, sapphire, and calcium fluoride, to assess the impact of aberrant lens emissions on the analysis of tissue Raman features within the FP and HW spectral regions. Signal quality metrics are reported from in vivo tissue using each type of lens, demonstrating that crystalline lenses best preserve the weak Raman signal generated by tissues during dual-region RS analysis. Still, the ideal lens type will ultimately depend on material characteristics and which spectral region is required for tissue interrogation. This device demonstrated a 90% increase in signal intensity and a four-fold improvement in spatial selectivity compared to a conventional RS probe during noncontact operation. Finally, one embodiment of the noncontact probe is described to showcase a clinically compatible prototype, which incorporates a widefield camera module for positioning guidance during in vivo use.

We developed a sensor called the Submerged Plasma for Isotopic Detection and Elemental Resolution (SPIDER) probe, which uses an atmospheric pressure glow discharge below the surface of liquids to excite species in the liquid. Through emission spectroscopy of molten salts, liquid metals, and heavy water, we demonstrated the SPIDER probe's high resolution, accuracy, and versatility. We successfully identified trace concentrations of transition and rare-earth metals in molten salts and detected the isotopic shift of the H${}_{\beta }\to$ D${}_{\beta }$ emission line. Our analysis revealed unconventional spectral alkali line shapes, indicating two competing excitation modes: film explosion and droplet vaporization. The film explosion mode, characterized by dense plasma, exhibited self-reversal and broadband continuum emission, while the droplet vaporization mode, associated with diffusive plasma, produced narrow-line emissions. By analyzing circuit transients alongside individual plasma events, we observed that film explosions generate higher currents, likely due to a shorter plasma length as the current preferentially flows through the thin liquid layer. Altogether, our results highlight the SPIDER probe's efficacy and flexibility, making it well-suited for online material quantification of liquids in extreme environments.

High-spatial-resolution tip-enhanced Raman spectroscopy (TERS) measurements were carried out under ambient conditions on graphene nanobubbles with various associated structural features. The resulting signals were analyzed with consideration of the characteristic features inherent to high resolution TERS. Compared to flat graphene regions, nanobubbles and their associated nanoconvex pinning sites demonstrated enhanced TERS signals, attributed to the efficient coupling between the strong tip-enhanced electric field and out-of-plane deformations in graphene. Strong coupling with highly confined near-field light activates the D bands even in the absence of defects, with intensity depending on the degree of deformations. While the D band is observed across the nanobubbles, some local regions exhibit a weaker D band intensity compared to the surrounding areas. Given the finite number of hexagonal lattices within the area of highly confined near-field, this reduction in intensity is likely to result from defects that cause missing hexagonal lattices. These findings highlight the capability of near-field induced Raman signals in probing high resolution features of nanomaterials even under ambient conditions, providing deeper insights into their characteristics in situ.

As a preprocessing step of spectroscopic techniques such as Raman spectroscopy, infrared spectroscopy, electrophoresis, etc., the baseline correction is very important for improving the signal quality, thereby ensuring the reliability and accuracy of the data analysis. Methods such as polynomial fitting, wavelet transforms, and frequency-domain filtering are widely used for baseline correction, effectively reducing interference and enhancing the reliability of signal analysis. However, these methods have certain limitations: (i) Polynomial fitting faces challenges in determining the optimal order, which may affect the fitting quality, (ii) wavelet transforms are complex and require fine adjustments, and (iii) frequency-domain filtering may cause signal distortion. These shortcomings affect the implementation of the algorithm in spectral related industries. Therefore, finding an appropriate algorithm to optimize baseline removal is crucial for the development of automated spectral analysis equipment. Here, we propose a rolling ball baseline removal algorithm based on morphological operations. With its simple implementation and excellent baseline removal performance, this method effectively avoids the overfitting problems. It is suitable for baseline correction in not only Raman spectroscopy, but also various other types of spectral data. In all, this approach offers a convenient and efficient general solution for the processing of various spectral data.

This paper provides the first temporally resolved visualization of the formation and decay profiles of THermally-induced Optical Reflection of Sound (THORS) barriers in ambient air, revealing the spatiotemporal characteristics of these novel acoustic barriers. In this work, a 532 nm Nd:YAG coupled with an intensified charge coupled device (ICCD) is used to Raman image N₂ in ambient air, thereby allowing for the visualization of the spatial dynamics of the air density variations at these THORS barriers. Studies were conducted at various ambient temperatures and with air turbulence across the beam path revealing no change in barrier size or shape under typical environmental disturbance conditions. Raman images of a barrier formed by a repetitively pulsed CO laser reveal an abrupt barrier density change between the optically depleted region and the surrounding air, with the slope of the imaged barrier density increasing rapidly during the first 20 ms of barrier formation, indicative of the predicted increase in barrier abruptness associated with enhanced THORS efficiency. As seen in previous temporal studies of THORS barrier efficiencies, these images reveal that multiple laser pulses at an optimized optical frequency are capable of achieving maximum continuous suppression efficiencies through molecular depletion in the optically excited region. These imaging studies revealed that the maximum barrier efficiency required a minimum of eight laser pulses to achieve the desired barrier density change and depletion, agreeing with previous temporal studies that showed in maximum suppression efficiency after 16 ms with one ms excitation laser pulses. Furthermore, visualization of the barrier size revealed that thermal redistribution of the photothermally excited molecules resulted in a THORS barrier approximately 50% larger than the excitation beam width and that this barrier remains constant for as long as 15 ms after the final laser pulse and at laser powers between 50 and 250 W.

The increasing concern about the presence of pesticides in vegetable leaves has underscored an urgent need for real-time, nondestructive, and accurate detection methods. Traditional methods are reliable but laboratory-based, costly, and unsuitable for field monitoring. In this study, we propose an efficient learning model pipeline that uses hyperspectral reflectance signatures to detect pesticide residue in plant leaves. We extract a comprehensive set of 39 domain-specific features based on vegetation indices, red-edge metrics, spectral statistics, and derivative profiles. To enhance the performance, use a multilayer perceptron to extract more features. A feature fusion module is used to combine both domain-specific features and features extracted by a multilayer perceptron. Further refinement is achieved through a feed-forward attention scoring module that dynamically weights important features. The efficiency of the system is evaluated using an enhanced extra trees classifier, which shows superior classification performance and stability across different feature formats. With cross-validation, our model achieves an accuracy of 94.69%, significantly outperforming conventional classifiers such as convolutional neural networks, support vector machines, and ensemble models such as random forest and extra trees. This framework not only improves interpretability and performance but also provides a foundation for a real-time, on-site pesticide monitoring solution.

Attenuated total reflection infrared-far-infrared (ATR IR-FIR) spectra (4000-50 cm^-1) and Raman spectra (2000-50 cm^-1) were measured for twelve types of biogenic minerals (shells), including Corbicula sandai (aragonite), Corbicula fluminea (aragonite), Corbicula japonica (aragonite), Ruditapes philippinarum (aragonite), and Mytilus galloprovincialis (aragonite and calcite) from different origins and growing environments. In this study, we investigated the crystal structures of these biogenic minerals, the water contents and structure in them, and the differences in the crystal structures among the aragonite forms of these minerals. In the 4000-3000 cm^-1 region and around the 1650 cm^-1 band region in the IR spectra, the proportion of the IR absorption bands related to weak and strong hydrogen bonds was significantly different among the shellfish species investigated. Therefore, it has been found that IR spectroscopy is useful for discriminating among shells based on the content and structure of water such as hydrogen bonds. In the low-frequency region below 500 cm^-1, where bands corresponding to lattice vibrational modes are observed, we investigated the lattice vibration modes of aragonite of shells and discussed particularly the full width at half-maximum (FWHM) of the bands at around 267 cm^-1 in the FIR spectra and the intensity of the side band at around 140 cm^-1 in the Raman spectra. As a result, we demonstrated that using both IR and Raman spectroscopies including the low-frequency regions allows us to distinguish various biogenic minerals from different habitats and growing environments. Additionally, it suggests that both IR and Raman spectroscopies including low-frequency regions are useful for characterizing habitats of shellfish.

Arthroscopic procedures rely on qualitative methods for cartilage assessment, such as tissue visualization and mechanical probing. visible-near-infrared (Vis-NIR) spectroscopy offers the potential to include compositional tissue characterization which could improve surgical guidance. The primary objective of this study was to assess the feasibility of using a fiber optic Vis-NIR probe in environments typically experienced during arthroscopy. Given the geometric constraints of articulating joints, a probe was fabricated with a 90-degree bend at the tip to enable movement and access to tissues. Absorbances from arthroscopic irrigation fluid (saline) are prominent in the NIR spectral region and thus need to be minimized during spectral collection. The current study aims to identify spectral data where the probe was not in contact with the tissues and/or where environmental saline contributed to the spectra. Porcine patella tissues were used to model how spectra collection in various conditions (probe offset from tissue and presence of fluid) impact spectra. Spectra were collected from cartilage, bone, and osteochondral tissues (n = 6 each) in experimental configurations with and without tissue contact and/or saline. Additionally, arthroscopic spectra collection in an equine stifle joint was investigated. Spectra collected while the fiber optic probe was in contact with the tissues resulted in minimal impact of environmental saline. Principal component analysis of spectra resulted in the separation of groups based on experimental configuration, demonstrating the potential for the development of more advanced machine learning algorithms focused on exclusion of spectra without appropriate tissue contact and with saline interference.