Applications of laser absorption spectroscopy for trace gas detection are many and diverse, ranging from the environmental and atmospheric to the medical and industrial. The aim of creating a spectrometer which combines high sensitivities and selectivities (in order to measure small amounts of absorbers or species that are only weakly absorbing, in a complex background matrix) with a wide spectral coverage (to allow broadband absorbers or multi-component samples to be studied) can be realised by implementing three separate concepts: the exploitation of the strong, fundamental transitions of the mid-infrared; the use of sensitive spectroscopic techniques; and the selection of a widely tunable laser source. In this thesis, these ideas are investigated individually and in combination in order to achieve such a goal. Laser spectroscopic techniques based on optical cavities are used to build a high resolution spectrometer covering a large spectral range capable of selectively detecting low levels of gaseous compounds of interest, especially those of medical or environmental significance. Work in both the near- and mid-infrared is presented, including much of the initial, developmental work which was conducted in the former region. The thesis begins with an overview of both narrowband and broadband near-infrared radiation sources, with a particular emphasis on commonly available diode lasers (DLs). A novel laser source, the digital supermode distributed Bragg reector (DS-DBR) laser, is introduced as a useful laser source for spectroscopy, combining the usual benefits of telecom DLs with a wide tunability (1563 – 1613 nm). The laser can be operated in an internal or external ramping mode, allowing the output wavelength to be scanned or stepped across a desired region. The observation of mode-hopping during the application of the scanning methodology is examined and rationalised. The ability of the DS-DBR laser to perform high resolution spectroscopy over its entire spectral coverage is demonstrated by recording spectra of carbon dioxide (CO₂) over this range, covering transitions from two of the four Fermi resonance components of the 3ν₁ + ν₃ combination band. The results of conducting wavelength modulation spectroscopy on CO₂ are also reported. A system developed for performing cavity ring-down spectroscopy (CRDS), capable of the real-time retrieval of ring-down times (RDTs), is presented and discussed. The outcomes of initial tests performed with a conventional DL at 1557 nm, to study a calibrated mixture of CO₂ in air at various pressures, are given. In addition, the results of combining this system with the DS-DBR laser are discussed. The bandwidth of the DS-DBR laser was found to be larger than that of a standard DFB DL, resulting in the presence of noisy cavity modes. Despite this, the acquisition of reproducible RDTs is demonstrated, with single wavelength studies of an evacuated cavity at 1605.5 nm yielding a RDT of 24.54 ± 0.04 µs and Allan variance calculations signalling an attainable minimum detectable absorption coefficient, α_min, of 2.8 x 10^-10 cm^-1 over 20 s. The ability to perform CRDS across the whole DSDBR laser wavelength range without the need for cavity re-alignment is illustrated, and studies conducted on CO₂ in air, calibrated mixtures and breath are reported. Investigations are also described into the accurate determination of the ¹³C/¹²C ratio in exhaled CO₂ undertaken using CRDS and cavity enhanced absorption spectroscopy (CEAS) on CO₂ isotopologues, an approach which can be utilised as a diagnostic aid in determining Helicobacter pylori infection. The focus of the thesis then moves to the mid-infrared, to describe quasi phase matching difference frequency generation (QPM-DFG) and its use to generate laser light at 3 µm by optically mixing near-infrared DLs. The theory behind this non-linear optical interaction is outlined, and the construction of a free-space QPM-DFG system using periodically poled lithium niobate is detailed and characterised. This DL-based QPM-DFG arrangement has been coupled with the CRDS system developed to create a mid-infrared CRD spectrometer. The results of single wavelength studies indicate RDTs of ~ 6 µs and an achievable αmin of 2.9 x 10^-9 cm^-1 over 44 s for an evacuated cavity. Spectroscopic investigations carried out on methane (CH₄), acetone and deuterium are documented; for the latter species, Dicke narrowing of the electric quadrupole ν(1←0) Q(2) transition at 2987.29 cm^-1 is observed and the integrated absorption cross-section for the same transition measured as 2.29 ± 0.03 x 10^-27 cm²cm^-1molec^-1. The results of modifications made to the system, namely the use of a more powerful Nd:YAG laser as the pump radiation source, as well as a faster detector combined with a variable amplifier, are presented; these include the observation of an improved optimal α_min of 6.4 x 10^-10 cm^-1 over 151 s for an empty cavity. Finally, work utilising the DS-DBR laser as one of the near-infrared sources for the QPM-DFG set-up is presented. This configuration generates radiation covering a wide mid-infrared range (3130 – 3330 nm) and has been used to perform direct absorption and wavelength modulation spectroscopy on ro-vibrational transitions within the fundamental ν₃ (F₂) band of CH₄. The spectrum of methanethiol (CH₃SH) over this region has also been investigated, with preliminary studies identifying a feature at 3040 cm^-1 as a potential indicator for monitoring this biomarker in breath. The results of coupling this mid-infrared radiation with an optical cavity to perform CEAS combined with phase sensitive detection are subsequently reported. Studies were conducted on calibrated CH₄ mixtures and ambient air to examine two transitions of the fundamental ν₃ (F₂) band of CH₄ in order to characterise the system: effective path lengths of ~ 700 m and α_min of 6.2 x 10^-8 cm^-1 over 8 s were found. The ^RQ₄ CH₃SH absorption feature at 3040 cm^-1 was also further studied with this system using prepared samples of CH₃SH in N₂ at different concentrations, yielding a CH₃SH detection limit of 2.4 ppm at 19 Torr. The potential of such a cavity-based, DS-DBR sourced, QPM-DFG mid-infrared spectrometer for trace gas sensing having thus been demonstrated, possible improvements that could be implemented to increase the sensitivity of the system are then discussed.

The work described herein focuses on FT-IR and quantum cascade laser (QCL) based studies towards the development of compact and portable trace gas sensor for benzene, toluene, and xylenes (BTX). FT-IR broadband radiation was used to probe the mid-infrared fingerprint region for quantitatively detecting trace gas levels of BTX. Using direct absorption through a hollow waveguide, parts-per-million (ppm) detection limits for BTX with a response time of 39 seconds was demonstrated. Univariate calibration provided limits of detection (3σ) for benzene, toluene, and meta-xylene at 5, 17, and 11 ppm, respectively. Multivariate calibration using partial least squares regression algorithms were used to simulate real-world conditions with multiple analytes present within a complex sample. A calibration model was built with 110 training set standards enabled by using a customized gas mixing system. Furthermore, a preconcentration/thermal desorption (TD) step was added to the FT-IR HWG trace gas sensor enabling parts-per-billion detection of BTX. A univariate calibration was established in the laboratory with certified gas standards over a dynamic range of 1000 - 100 ppb for benzene, toluene, and the xylenes. The sensor was then taken to an industrial site during a field measurement campaign for the quantitative determination of BTX in field air samples. The laboratory calibration was used to predict unknown concentrations which were in close agreement with industrial hygiene standard techniques, and industrial prototype analyzers, that were simultaneously operated in the field environment. In addition to FT-IR, quantum cascade laser spectroscopy was also investigated due to enhanced spectral density and efforts to precisely overlap emission with analyte absorption. Particular efforts were dedicated on a novel principle for consistent and deliberate QCL emission wavelength selection by varying the QCL cavity length. These studies experimentally confirmed that using this straight-forward post-processing technique, emission wavelength tuning across a range of one hundred wavenumbers range may be achieved. This tuning range was experimentally demonstrated for a QCL emitting across an entire absorption feature of carbon dioxide by tailoring the length of the cavity. Additionally, using an external cavity (EC) - QCL combined with a HWG gas sensor module for the first time enabled the quantitative and simultaneous determination of ethyl chloride, trichloromethane, and dichloromethane within exponential dilution experiments at ppb limits of detection. Multianalyte detection was demonstrated utilizing partial least squares regression for quantitative discrimination of individual constituents within a mixture, yet applying a single broadly tunable QCL light source.

Articular cartilage helps in the motion of the diarthroidal joints by providing a near frictionless load bearing surface. Identification of changes in articular cartilage chemical and structural properties that arise from degeneration and suboptimal tissue repair have been the target of many studies. Current methods that evaluate these changes frequently involve destructive specimen preparation. Thus there is a need to develop a method to accurately evaluate changes in cartilage during disease or repair processes. Fourier transform infrared (FTIR) spectroscopy in the mid-infrared (MIR) spectral range is based on molecular vibrations and has been used to study the chemical and structural properties of biological tissues, including cartilage. However, this technique generally requires extensive sample preparation and modification of the intact tissue. An alternative approach is to use near-infrared spectroscopy (NIRS) which does not require sample preparation due to higher depth of penetration. This doctoral dissertation focuses on identification of NIR spectral features to evaluate the major components of cartilage. These NIR spectral features are then used to evaluate compositional changes in engineered and degraded cartilage, and the results validated with histological, biochemical, mechanical and MIR analysis of the same tissue. Together, these studies lay the groundwork for clinical and in situ applications of NIRS. / Bioengineering

In several fields, such as biology, chemistry, combustion and environmental science, laser absorption spectroscopy represents an invaluable tool for the detection and identification of a variety of molecular species in the gas phase. For this detection to be quantitative, it is of paramount importance to rely on accurate spectroscopic parameters for the involved absorption lines in terms of line strength, line center frequency, pressure broadening, and pressure shift coefficients. The mid-infrared region offers the most favorable conditions for sensitive and chemically selective detection. The sensitivity derives from the presence of intense fundamental ro-vibrational transitions of molecules, whereas chemical selectivity relates to the unique absorption spectrum that molecules possess in the mid-IR region, thereby known as the fingerprint region. In this thesis, we combine the accelerating technology of optical frequency combs (OFC), which are powerful tools for accurate optical frequency measurements, with the wide tunability and single line emission in the mid-IR of extended cavity quantum cascade lasers (EC-QCL), to perform highly resolved, accurate and sensitive measurements in the fingerprint region, from 7.25 to 8 μm. Specifically, we have been able to lock for the first time the optical frequency of an EC-QCL to an OFC by utilizing nonlinear optics in the form of sum frequency generation (SFG) (Lamperti, AlSaif et al., 2018) and have exploited this comb-locked EC-QCL for an accurate survey of the entire ν1 ro-vibrational band of one of the most important greenhouse gases, nitrous oxide (N2O). The developed spectrometer is able to operate over a wide region of ~ 100 cm-1, in a fully automated fashion, while affording a 63 kHz uncertainty on the retrieved line center frequencies. The measurement allowed us to determine very accurately rotational constants of both ground and excited states of the ν1 band of N2O through the measurements of tens of lines of the P and R branches (AlSaif et al., JQSRT 2018). The spectrometer was then upgraded with a more recent and narrower linewidth EC-QCL to perform sub-Doppler saturated spectroscopy on the same N2O sample at a spectral resolution below 1 MHz, the sharpest ever observed with this type of laser. Finally, we worked at adding high sensitivity to the apparatus by introducing the gas in a high-finesse passive resonator and by developing a system to measure the intra-cavity absorption with cavity ring-down spectroscopy (CRDS) together with comb calibration.

Des fibres optiques en verres de chalcogénure transmettant dans le moyen infrarouge (MIR) ont été développées par le laboratoire Verres et Céramiques de l’université de Rennes 1. Des travaux ont ensuite montré les potentialités de ces fibres comme outil diagnostic pour la spectroscopie MIR appliquée aux biofluides. Le spectre moyen infrarouge d’un échantillon complexe est le reflet de sa composition moléculaire qui, lorsque celui-ci est un biofluide, peut être assimilé à une image métabolique instantanée d’un individu. Cette technique constitue donc un outil intéressant pour le diagnostic médical. La société Diafir fut crée à la suite du programme ANR émergence FIR-MED pour développer les potentialités de la spectroscopie par fibre optique. L’objectif de l’entreprise est de développer un système composé d’un spectromètre spécifiquement adapté au capteur à fibre optique et d’un algorithme associé pour une réponse diagnostique ne nécessitant pas de connaissance particulière en spectroscopie infrarouge. Dans le cadre de ces travaux de thèse associant l’IRDL et la société Diafir, des projets variés d’applications médicales de la technologie ont été abordés avec pour objectif de développer une méthodologie d’analyse des spectres MIR. Pour cela, il a été nécessaire dans un premier temps de caractériser le signal obtenu à partir des capteurs et d’optimiser et valider des protocoles de mesure applicables aux différents biofluides de manière à réduire au mieux les sources de variabilité d’ordre physique et environnemental. La construction d’un diagnostic médical à partir de spectres infrarouge qui sont des données en grande dimension implique de sélectionner un petit nombre de variables spectrales explicatives. Des méthodes de sélection de variables ont été sélectionnées afin d’éviter l’effet « boite noire » en établissant le lien avec les fonctions biochimiques impactées par les pathologies. Cette approche spectrale infrarouge a notamment permis de mettre en évidence, au cours du développement de pathologies, l’existence de transitions a priori non détectées par les dosages biochimiques classiques. / Chalcogenide glass optical fibres exhibiting unique properties of mid-infrared (MIR) transparency have been developed by the Laboratoire Verres et Céramiques of Rennes 1 University. Our studies investigate the potential of such fibres as a tool for MIR spectroscopy applied to biofluids based diagnostic. The MIR spectrum of complex samples features its molecular composition which, when a biofluid is considered, can be assimilate to an instant metabolic imaging of an individual. The Diafir Company was created following the ANR emergence FIR-MED program to develop the potential of optical fibre biomedical spectroscopy. The company's goal is to develop a system that is composed of the optical fibre sensor, a spectrometer specifically designed to these sensors and an associated algorithm for a diagnostic response without specific knowledge of infrared spectroscopy. As part of this thesis work, linking the IRDL and the Diafir Company, various projects of biomedical applications were driven with the aim of developing a MIR spectra analysis methodology. Accordingly, it was necessary initially to characterize the signal obtained from the sensors and to optimize and validate robust measurement protocols for each biofluid tested with the aim to reduce the physical and environmental sources of variability. The medical diagnosis construction from infrared spectra, that are high dimensional data, involves selecting from the whole spectral data set a small number of explanatory variables. Purposely, particular variables selection algorithms were selected to avoid the effect of "black box" by establishing the link with the biochemical functions affected by the disease. This infrared spectral approach allowed identifying previously unrevealed transitions in the time course of pathologies which were not detected from conventional biochemical markers

We investigate the spectra of asteroids at near- and mid-infrared wavelengths. In 2010 and 2011 we reported the detection of 3 ?m and 3.2-3.6 ?m signatures on (24) Themis and (65) Cybele indicative of water-ice and complex organics [1] [2] [3]. We further probed other primitive asteroids in the Cybele dynamical group and Themis family, finding diversity in the shape of their 3 ?m [4] [5] [6] and 10 ?m spectral features [4]. These differences indicated mineralogical and compositional variations within these asteroid populations. Also in the mid-infrared region we studied a larger population of asteroids belonging to the Bus C, D, and S taxanomic classes to understand the relationship between any mineralogy and hydration inferred in the visible and near- infrared with the shape, strength, and slope of the 10 ?m emission. We have discovered that at least 3 of the main Bus taxanomic groups (Cs, Ds, and Ss as defined by their visible spectra) clearly cluster into 3 statistically distinct groups based on their 8-13 ?m spectra. Additionally we have attempted to simulate in a laboratory the possible organic compounds we have detected on two asteroids, using various mixtures containing aromatic and aliphatic hydrocarbons. We find that asteroid (24) Themis and (65) Cybele have ?CH2/?CH3 and NCH2/NCH3 ratios similar to our 3- methylpentane, propane, and hexane residues, suggesting that the organics on these asteroids may be short chained and/or highly branched. The ?CH2/?CH3 and NCH2/NCH3 for asteroid(24)Themis are most consistent with the DISM, and some carbonaceous chondrites. The band centers of the C-H stretch absorptions indicate that both asteroids may have aliphatic carriers chemically bonded to electronegative groups (i.e. aromatics), and some that are not. We also detect a 3.45 ?m feature in the spectra of both asteroids that is present in several dense molecular clouds. Our results suggest an interstellar origin for the organics on (24) Themis, and likely (65) Cybele. The differences in the organics of Themis and Cybele are likely related to variations in thermal processing, irradiation and/or formation region in the solar nebula.

No description available.

Este trabalho de pesquisa descreve três estudos de utilização de métodos quimiométricos para a classificação e caracterização de óleos comestíveis vegetais e seus parâmetros de qualidade através das técnicas de espectrometria de absorção molecular no infravermelho médio com transformada de Fourier e de espectrometria no infravermelho próximo, e o monitoramento da qualidade e estabilidade oxidativa do iogurte usando espectrometria de fluorescência molecular. O primeiro e segundo estudos visam à classificação e caracterização de parâmetros de qualidade de óleos comestíveis vegetais utilizando espectrometria no infravermelho médio com transformada de Fourier (FT-MIR) e no infravermelho próximo (NIR). O algoritmo de Kennard-Stone foi usado para a seleção do conjunto de validação após análise de componentes principais (PCA). A discriminação entre os óleos de canola, girassol, milho e soja foi investigada usando SVM-DA, SIMCA e PLS-DA. A predição dos parâmetros de qualidade, índice de refração e densidade relativa dos óleos, foi investigada usando os métodos de calibração multivariada dos mínimos quadrados parciais (PLS), iPLS e SVM para os dados de FT-MIR e NIR. Vários tipos de pré-processamentos, primeira derivada, correção do sinal multiplicativo (MSC), dados centrados na média, correção do sinal ortogonal (OSC) e variação normal padrão (SNV) foram utilizados, usando a raiz quadrada do erro médio quadrático de validação cruzada (RMSECV) e de predição (RMSEP) como parâmetros de avaliação. A metodologia desenvolvida para determinação de índice de refração e densidade relativa e classificação dos óleos vegetais é rápida e direta. O terceiro estudo visa à avaliação da estabilidade oxidativa e qualidade do iogurte armazenado a 4C submetido à luz direta e mantido no escuro, usando a análise dos fatores paralelos (PARAFAC) na luminescência exibida por três fluoróforos presentes no iogurte, onde pelo menos um deles está fortemente relacionado com as condições de armazenamento. O sinal fluorescente foi identificado pelo espectro de emissão e excitação das substâncias fluorescentes puras, que foram sugeridas serem vitamina A, triptofano e riboflavina. Modelos de regressão baseados nos escores do PARAFAC para a riboflavina foram desenvolvidos usando os escores obtidos no primeiro dia como variável dependente e os escores obtidos durante o armazenamento como variável independente. Foi visível o decaimento da curva analítica com o decurso do tempo da experimentação. Portanto, o teor de riboflavina pode ser considerado um bom indicador para a estabilidade do iogurte. Assim, é possível concluir que a espectroscopia de fluorescência combinada com métodos quimiométricos é um método rápido para monitorar a estabilidade oxidativa e a qualidade do iogurte / This research work describes three studies of chemometric methods employed for the classification and characterization of edible oils and its quality parameters through Fourier Transform mid infrared spectroscopy and near infrared spectroscopy, and for the monitoring the oxidative stability and quality of yogurt using fluorescence spectroscopy . The first and second studies aimed the classification and characterization of edible oil and its quality parameters using Fourier Transform mid infrared spectroscopy (FT-MIR) and near infrared spectroscopy (NIR) measurements, respectively. Kennard-Stone algorithm was used for selecting the training set, after a principal component analysis (PCA) was applied. The discrimination of canola oils from sunflower, corn and soybean was investigated using SVM-DA, SIMCA and PLS-DA. The quality parameters refraction index and relative density of edible oil was investigated using partial least squares (PLS), iPLS, LS- SVM multivariate calibration of FT-MIR and NIR data were evaluated. Several preprocessing alternatives (first derivative, multiplicative scatter correction, mean centering, orthogonal signal correction and standard normal variate) were investigated by using the root mean square error of validation cross validation (RMSECV) and prediction (RMSEP), as control parameters. In fact, the methodology developed is proposed for direct relative density and refraction index in edible oils and their classification, requiring a few minutes per sample without any previous treatment. The third study aimed to evaluate the oxidative stability and quality of yogurt stored at 40C with light or dark using the combined parallel factor (PARAFAC) analysis and fluorescence spectroscopy. PARAFAC analysis of the fluorescence landscapes exhibited three fluorophores present in the yogurt, where, at least one of them was strongly related to the storage conditions. The fluorescence signal was resolved into excitation and emission profiles of the pure fluorescent compounds, which are suggested to be vitamin A, tryptophan and riboflavin. Regression model based on PARAFAC scores for riboflavin were built using the scores obtained in the first day as dependent variable and the scores obtained during the storage as independent variable. It was clear demonstrated that the slope of the analytical curve has become smaller throughout the experiment. Therefore, riboflavin level could be considered a good indicator for the yogurt stability. Thus, it is concluded that fluorescence spectroscopy in combination with chemometrics has a potential as a fast method for monitoring the oxidative stability and quality of yogurt

Este trabalho de pesquisa descreve três estudos de utilização de métodos quimiométricos para a classificação e caracterização de óleos comestíveis vegetais e seus parâmetros de qualidade através das técnicas de espectrometria de absorção molecular no infravermelho médio com transformada de Fourier e de espectrometria no infravermelho próximo, e o monitoramento da qualidade e estabilidade oxidativa do iogurte usando espectrometria de fluorescência molecular. O primeiro e segundo estudos visam à classificação e caracterização de parâmetros de qualidade de óleos comestíveis vegetais utilizando espectrometria no infravermelho médio com transformada de Fourier (FT-MIR) e no infravermelho próximo (NIR). O algoritmo de Kennard-Stone foi usado para a seleção do conjunto de validação após análise de componentes principais (PCA). A discriminação entre os óleos de canola, girassol, milho e soja foi investigada usando SVM-DA, SIMCA e PLS-DA. A predição dos parâmetros de qualidade, índice de refração e densidade relativa dos óleos, foi investigada usando os métodos de calibração multivariada dos mínimos quadrados parciais (PLS), iPLS e SVM para os dados de FT-MIR e NIR. Vários tipos de pré-processamentos, primeira derivada, correção do sinal multiplicativo (MSC), dados centrados na média, correção do sinal ortogonal (OSC) e variação normal padrão (SNV) foram utilizados, usando a raiz quadrada do erro médio quadrático de validação cruzada (RMSECV) e de predição (RMSEP) como parâmetros de avaliação. A metodologia desenvolvida para determinação de índice de refração e densidade relativa e classificação dos óleos vegetais é rápida e direta. O terceiro estudo visa à avaliação da estabilidade oxidativa e qualidade do iogurte armazenado a 4C submetido à luz direta e mantido no escuro, usando a análise dos fatores paralelos (PARAFAC) na luminescência exibida por três fluoróforos presentes no iogurte, onde pelo menos um deles está fortemente relacionado com as condições de armazenamento. O sinal fluorescente foi identificado pelo espectro de emissão e excitação das substâncias fluorescentes puras, que foram sugeridas serem vitamina A, triptofano e riboflavina. Modelos de regressão baseados nos escores do PARAFAC para a riboflavina foram desenvolvidos usando os escores obtidos no primeiro dia como variável dependente e os escores obtidos durante o armazenamento como variável independente. Foi visível o decaimento da curva analítica com o decurso do tempo da experimentação. Portanto, o teor de riboflavina pode ser considerado um bom indicador para a estabilidade do iogurte. Assim, é possível concluir que a espectroscopia de fluorescência combinada com métodos quimiométricos é um método rápido para monitorar a estabilidade oxidativa e a qualidade do iogurte / This research work describes three studies of chemometric methods employed for the classification and characterization of edible oils and its quality parameters through Fourier Transform mid infrared spectroscopy and near infrared spectroscopy, and for the monitoring the oxidative stability and quality of yogurt using fluorescence spectroscopy . The first and second studies aimed the classification and characterization of edible oil and its quality parameters using Fourier Transform mid infrared spectroscopy (FT-MIR) and near infrared spectroscopy (NIR) measurements, respectively. Kennard-Stone algorithm was used for selecting the training set, after a principal component analysis (PCA) was applied. The discrimination of canola oils from sunflower, corn and soybean was investigated using SVM-DA, SIMCA and PLS-DA. The quality parameters refraction index and relative density of edible oil was investigated using partial least squares (PLS), iPLS, LS- SVM multivariate calibration of FT-MIR and NIR data were evaluated. Several preprocessing alternatives (first derivative, multiplicative scatter correction, mean centering, orthogonal signal correction and standard normal variate) were investigated by using the root mean square error of validation cross validation (RMSECV) and prediction (RMSEP), as control parameters. In fact, the methodology developed is proposed for direct relative density and refraction index in edible oils and their classification, requiring a few minutes per sample without any previous treatment. The third study aimed to evaluate the oxidative stability and quality of yogurt stored at 40C with light or dark using the combined parallel factor (PARAFAC) analysis and fluorescence spectroscopy. PARAFAC analysis of the fluorescence landscapes exhibited three fluorophores present in the yogurt, where, at least one of them was strongly related to the storage conditions. The fluorescence signal was resolved into excitation and emission profiles of the pure fluorescent compounds, which are suggested to be vitamin A, tryptophan and riboflavin. Regression model based on PARAFAC scores for riboflavin were built using the scores obtained in the first day as dependent variable and the scores obtained during the storage as independent variable. It was clear demonstrated that the slope of the analytical curve has become smaller throughout the experiment. Therefore, riboflavin level could be considered a good indicator for the yogurt stability. Thus, it is concluded that fluorescence spectroscopy in combination with chemometrics has a potential as a fast method for monitoring the oxidative stability and quality of yogurt

Instrumental disparities and soil moisture are two of the key limitations in implementing spectroscopic techniques in the field. This study sought to address these challenges through two objectives. The first objective was to assess Visible-near infrared (VisNIR) and mid-infrared (MIR) spectroscopic approaches and explore the feasibility of transferring calibration models between laboratory and portable spectrometers. The second objective addressed the challenge of soil moisture and its impact on spectra. The portable spectrometers demonstrated comparable performance to their laboratory-based counterparts in both regions. Spiking with extra-weight, was the most effective calibration transfer method eliminating disparities between instruments. The samples were rewetted under nine controlled conditions for the moisture study. Results showed that spiking with extra weights significantly outperformed other techniques and model enhancement was insensitive to the moisture contents. Findings of this study will be helpful for development and deployment of in situ sensors to enable field implementation of spectroscopy.