Applied Spectroscopy最新文献

Sediment contamination by trace elements (TE) is a major environmental issue. In particular, TE speciation is of great importance because the form of the TE determines their mobility, bioavailability, and consequently their potential toxicity. Characterizing the chemical speciation of TEs can be complex and costly with current analytical methods. Non-destructive spectroscopic methods, which require limited sample preparation, are therefore useful tools for characterizing and possibly quantifying TEs in complex sedimentary matrices. Thus, this study explores the potential of visible and near-infrared hyperspectral imaging (HSI) to estimate the speciation of some TEs in sediments based on their spectral properties. Standard ranges of sixteen chemical species of six TEs, i.e., arsenic (As), cadmium (Cd), copper (Cu), nickel (Ni), lead (Pb), and zinc (Zn), were produced using three model sediment matrices (clay, silt, and organic matter). The results obtained show specific absorptions for each of the TE species, and nine of them could be quantified with detection limits of around 1 g/kg in the visible range and around 10 g/kg in the short-wave infrared range. This approach enables a more accurate and rapid assessment of environmental risk using HSI, in addition to conventional analytical methods.

Raman spectroscopy (RS) is a label-free, non-destructive optical modality that provides a detailed profile of the molecular composition of a sample. There is growing interest in the clinical application of RS to characterize biomolecular signatures associated with radiotherapy response in tumor cells and tissues. A critical step before analyzing Raman data consists of performing spectral pre-processing to increase the quality of the measurements. Spectral pre-processing comprises baseline subtraction, signal smoothing, cosmic ray (CR) correction, and removal of poor-quality measurements. Herein, we present a convolutional autoencoder (AE) for single-step, automated pre-processing of Raman spectra obtained from tumor cells and tumor tissue. We trained two separate models using the same proposed architecture, one for eliminating spectral artifacts from preclinical single-cell line and xenografted tissue spectra exposed to single-fraction radiation, and the other for correcting clinical prostate tumor biopsy spectra collected from patients receiving high-dose-rate brachytherapy (HDR-BT). The autoencoder demonstrated fast, excellent performance in removing baseline, noise, and CRs. For the preclinical data, the model obtained a root mean squared error (RMSE), and a percentage root mean squared difference (PRD) of 7.1 × 10^-5 and 3.1%, respectively, between the AE-corrected spectra and their corresponding target data (pre-processed by our current baseline-removal algorithm). Also, the autoencoder successfully removed 94.0% of CRs from the spectra. For the clinical biopsy data, the AE achieved an RMSE and a PRD of 8.1 × 10^-5 and 3.7%, respectively, and a CR removal rate of 90.2%. Overall, the AE corrected approximately 11 000 spectra within 2.4 s without the need of a GPU. Furthermore, comparative supervised learning-based post-processing data analyses were performed separately on the spectra pre-processed by the autoencoder versus the target data, and we show consistency in the biochemical radiation response profiles extracted. Finally, the AE architecture was leveraged to train a reconstruction AE to facilitate semi-automated identification of poor-quality prostate biopsy spectra, and we demonstrate 96.4% agreement between AE and manually removed outliers. These results support the development of a deep learning framework for efficient, automated pre-processing of tumor cell and tissue Raman spectra collected for radiation response monitoring studies.

Hair fluorescence spectroscopy was evaluated as a novel, non-invasive biomarker for the diagnosis and disease monitoring of systemic lupus erythematosus (SLE). Hair samples were collected from 47 female patients with SLE and 49 age-matched healthy controls (HC), with patients stratified into three clinical groups: active flare, remission of six months to three years (R6M-3Y group), and remission of more than three years (R>3Y group). Fluorescence emission spectra of hair strands were recorded under ultraviolet excitation and analyzed using multivariate statistical methods, including principal component analysis and hierarchical clustering, to assess group discrimination. The fluorescence profiles differed significantly between SLE patients and the HC group, and within the SLE cohort, spectral signatures varied according to disease activity, enabling discrimination between flare and remission (low disease activity) states. Patients in long-term remission (R>3Y) showed partial convergence toward the HC group, suggesting progressive normalization over time. Overall, hair fluorescence spectroscopy emerges as a non-invasive, inexpensive, and stable biomarker reflecting both disease presence and remission dynamics in SLE, with potential to complement existing clinical and laboratory indices and to provide rheumatologists with a novel tool for longitudinal disease monitoring.

Identifying characteristic bands, those exhibiting the most distinct features (i.e., minimal correlation), is a critical step in two-dimensional correlation spectroscopy (2D-COS) analysis. This process is essential for establishing effective correlation filters to simplify congested spectral datasets. Historically, such bands were selected using subjective methods, primarily the visual inspection of correlation cross-peaks. We now propose a more systematic and objective procedure based on the sequential multiplication of horizontal slices from a 2D discrimination spectrum. This unsupervised, automatic method is potentially integrable into model-free 2D-COS analyses, making it compatible with automated, machine-based interpretation.

Optical phantoms are widely used to characterize diffuse optical setups and data analysis methods for in-vivo/ex-vivo measurements. Coconut oil is an interesting compound to use in phantoms, because it could be used to model lipidic tissues, such as the one in breast tissue. In this paper, we measure the absorption and scattering spectra of coconut oil from 600 to 1600 nm, encompassing the so-called "therapeutic window". To cover the entire range, we exploit a supercontinuum pulsed laser and a superconducting nanowire single photon detector operating in the time domain. Finally, we demonstrate the use of a homogeneous coconut oil phantom to characterize a hyperspectral continuous-wave (CW) setup.

Bamboo craftsmanship is highly valued for its aesthetics and cultural significance. The classification of bamboo craftsmanship plays a key role in preserving its heritage and ensuring its quality. However, the surface characteristics of bamboo exhibit substantial variation due to environmental factors. This study proposes a novel method using laser-induced breakdown spectroscopy (LIBS) combined with spectral data fusion to enhance the identification accuracy of bamboo age ranges. By fusing spectra from different bamboo parts, a broader range of elemental compositions can be captured while minimizing the influence of regional variations. A total of 50 bamboo craftsmanship samples of five different age ranges were prepared, and their internal and external surface LIBS spectra were collected for data analysis. Experimental results demonstrate that the peak selection-linear discriminant analysis model presents the highest classification accuracy of 99.0% before spectral data fusion. After fusion, the accuracy can be further improved to 99.9%. Additionally, a comparison of various data fusion methods reveals that the Concat method, which increases the dimensionality of the feature space and provides richer data representation, exhibits the best compatibility with LIBS spectral characteristics and classification models. In conclusion, the combination of LIBS and data fusion methods proves to be an effective approach for accurately identifying bamboo age ranges.

Femtosecond, broadband stimulated Raman spectroscopy is a popular approach to measuring molecular dynamics with excellent signal-to-noise and spectral resolution. We present a new method for broadband stimulated Raman spectroscopy that employs Kerr instability amplification to amplify the supercontinuum spectrum from sapphire and create a highly tunable Raman probe spectrum spanning from 530 to 1000  nm (-6000 to 2800 cm^-1). Our method, called Kerr instability amplification for broadband-stimulated Raman spectroscopy (KAB-SRS) provides an alternative to optical parametric amplifiers by producing a broader and more tunable spectrum at a significantly reduced cost to OPA implementations. We demonstrate the effectiveness of KAB-SRS by measuring the stimulated Raman loss spectrum of 1-decanol.

Inverse least squares (ILS) regression is an advancement of classical least squares (CLS) regression, enabling the calculation of concentrations without requiring prior knowledge of the number of components in a mixture. Complex-valued ILS further enhances the performance of ILS by incorporating the complex refractive index function, as demonstrated in the thermodynamically ideal mixtures of benzene-toluene and benzene-cyclohexane. In both systems, the mean absolute error can be reduced by over 50% using the leave-one-out cross-validation (LVOOCV) scheme with complex-valued ILS. Additional error reduction is achievable by leveraging correlations between the errors and the imaginary components of the concentrations or volume fractions. Since the complex refractive index function can be conveniently determined using conventional infrared spectroscopy through the Kramers-Kronig relations, we believe that complex-valued machine learning has the potential to significantly advance analytical applications.

Raman spectroscopy was applied to monitor polymerization, verify quality control, and analyze additive distribution in styrene-divinylbenzene (Sty-DVB) based proppants. The method relies on C=C vibrational markers to follow monomer consumption, crosslinking, and additive incorporation. Quality control included quantifying vinyl, cis, and trans C=C in polybutadiene (PB) modifiers and the ethyl-vinyl benzene (EVB) content in DVB crosslinkers. EVB content by Raman showed excellent agreement with independent carbon-13 nuclear magnetic resonance (¹³C-NMR) measurements. Styrene copolymerization with DVB was tracked in real time using a fiber-optic Raman probe in a temperature-controlled microreactor. DVB accelerates styrene consumption due to its higher reactivity and radical stabilization. PB additives do not affect overall polymerization kinetics. In terms of additives, Raman calibration confirms that PB double bonds remain largely unreacted, consistent with limited copolymerization and phase separation. Polyphenylene oxide (PPO) slows down Sty polymerization while Raman mapping demonstrates its homogeneous dispersion within the matrix, validating its incorporation and expected impact on material properties. Overall, Raman spectroscopy provides a direct, non-invasive, and scalable approach to monitor polymerization and verify additive distribution, establishing it as a practical tool for process optimization in Sty-DVB proppant formulations.