Fluorescence Excitation-emission Matrices Research Articles

Sensors support machine learning

Sensors support machine learning

Fluorescence coupled with chemometrics for simultaneous monitoring of cell concentration, cell viability and medium nitrate during production of carotenoid-rich Dunaliella salina

Two-dimensional (2D) fluorescence spectroscopy was investigated as a monitoring tool for cultivation, harvesting, and effluent treatment of Dunaliella salina with high carotenoid concentration; aiming to improve the production process and minimise costs. Chemometric analysis, namely Principal Component Analysis (PCA) and Projection to Latent Structures (PLS), were used to build models for estimation of cellular concentration, cellular viability, and nitrate concentration in media. The estimations were based on fluorescence excitation-emission matrices (EEMs) acquired directly from algal suspensions. Cell concentration during cultivation and harvesting can be predicted by a single model capturing 92.0% of the variance, and with R2 of 0.92 and 0.97, for training and validation, respectively. Cell viability during harvesting by ultrafiltration was modelled with 79% of variance and R2 of 0.79 for training and 0.73 for validation. Nitrate concentration was successfully predicted during cultivation and permeate treatment using a single model with 81.8% of variance and R2 of 0.82 for training and 0.80 for validation. Therefore, this work demonstrates the strong potential of combining 2D fluorescence and chemometrics for monitoring different processes during microalgae production.

Characterizing the interactions between copper ions and dissolved organic matter using fluorescence excitation–emission matrices with two-dimensional Savitzky–Golay second-order differentiation

Monitoring dissolved organic matter (DOM) content in aquatic environments is crucial for not only understanding the dynamics of heavy metals but also predicting their bioavailability. Fluorescence spectroscopy is typically employed to characterise DOM. Here, the interaction between DOM and trace metals was investigated by combining excitation–emission matrix (EEM) quenching with two-dimensional Savitzky–Golay second-order differentiation (2D-SG-2nd-df) analysis. The 2D-SG-2nd-df analysis decomposed the EEM spectra of commercial humic acid (HA) samples into six separate fluorescence peaks, which agreed with the results obtained through conventional parallel factor (PARAFAC) analysis. Compared with PARAFAC modeling, the 2D-SG-2nd-df approach provided more valid and reliable results when the dataset contained distinct samples. Moreover, since the results obtained from 2D-SG-2nd-df for each sample are independent, shifts in the peak wavelength can be reproduced more efficiently using this method. Triplicate titration experiments showed clear differences in HA–copper interactions for samples with different HA composition and molecular weight. The binding strength between copper and low-molecular-weight DOM in water was weaker than that observed for high-molecular-weight DOM. The results obtained in this study will serve as a basis for applying 2D-SG-2nd-df not only to DOM but also to other samples studied using EEM measurements.

Determination of relative concentrations in two-component mixtures with unknown composition by laser-induced fluorescence spectra

The paper is devoted to the development of a new methodology of statistical analysis of laser-induced fluorescence excitation-emission matrices. The method allows one to calculate the number of fluorescent components, their excitation and fluorescence spectra in samples under study without using any prior information about their nature. The algorithm is based on the principal component method and is modified for analysis of excitation-emission matrices. Its features were tested on two-component optical phantoms, namely, optically thin mixtures of fluorescent dyes (pyridoxine and fluorescein). The calculated excitation and fluorescence spectra of components coincided with the initial ones with good accuracy (deviation less than 5% in intensity). The weight coefficients proportional to concentrations were calculated by the new algorithm. The comparison of its ratios with those calculated from known concentrations also showed good agreement (deviation of 5%).

Effects of biochar amendment on the availability of trace elements and the properties of dissolved organic matter in contaminated soils

Abstract The effects of biochar on the amount and characteristics of dissolved organic matter (DOM) and the relationship between the DOM and the availability of trace elements in soils are poorly understood. An incubation experiment was used to investigate the effects of biochar on DOM properties and the mobility and availability of As, Cd, and Pb in two types of contaminated soils (vegetable plot vs. paddy). The availability of trace elements in the soil was assessed by diffusive gradients using the thin film (DGT) technique. The results showed that mixing biochar at a 10 wt% dose increased the amount of dissolved organic carbon (DOC) from 12.8 to 42.9 mg L−1 in soils, and this corresponded to an increase in soil pH. Fluorescence excitation–emission matrices (EEMs) indicated that rice straw biochar (RSBC) amendment to soil changed the DOM composition towards higher aromaticity. The DGT-extractable Cd and Pb concentrations were significantly decreased with biochar amendments, while DGT-extractable As was significantly increased RSBC treatment decreased acid-soluble Cd and Pb fractions but increased the acid-soluble As fraction in soil solutions compared to control samples. RSBC amendment can effectively adjust soil metal availability, as well as DOM amount and composition, and that the amendment conditions must be tailored according to the specific trace elements and soil types.

Unraveling long-term changes in lake color based on optical properties of lake sediment

A number of boreal surface waters have become browner over the last two decades. Recovery from acid rain is regarded as an important driver for this lake color increase, indicating a general browner lake color in preindustrial times. However, the lack of long-term monitoring data makes it challenging to unravel historical changes in lake color. In this study, we estimated long-term development in lake color (1800 to 2015) based on the optical properties of alkaline extractable dissolved organic matter (DOM) from sediment using UV–Vis and fluorescence spectroscopy. We found that the present lake color (2015) was significantly browner (four times higher in absorption coefficient) than for the period from 1800 to 1915 when lake color was at a lower and more stable level. Fluorescence excitation–emission matrices combined with parallel factor analysis (EEM-PARAFAC) indicate that terrestrially derived DOM was the main source of sediment DOM. However, the importance of in-lake source of DOM has significantly increased with time. The long-term trend in DOM burial was not consistent with the anthropogenic sulfur (S) deposition pattern. However, along with the increased sediment DOM, there has been increased precipitation, temperature and forest growth with time, which affect the production and degradation of DOM. Even though S deposition might have delayed the runoff of terrestrial DOM for a certain period, it comes in addition to other color-regulating factors. Thus, there is no single driver for the observed lake browning, but rather an interplay between different drivers varying in strength over time, such as afforestation, changes in areal use, declined S deposition, and increased temperature and precipitation.

The Parallel Factor Analysis of Beer Fluorescence.

Fluorescence excitation-emission matrices were measured for 111 samples of different types of beer and studied by the parallel factor analysis (PARAFAC). The 5-component PARAFAC model was found to suitably describes the beer fluorescence, accounting for 99.4% of the fluorescence variance in the measured set of samples, and providing the completely resolved excitation and emission spectra of each component. The model was chosen based on a model's core consistency and split-half analysis. It is shown that beer fluorescence is the sum of fluorescence of aromatic amino acids (tryptophan, tyrosine, and phenylalanine), different forms of vitamin B, and phenolic compounds. Obtained PARAFAC model of beer fluorescence demonstrated the potential for the quantification and quality analysis of beer fluorophores and classification of different beer types.

Fusing data of different orders for environmental monitoring

In the present work a novel application of data fusion to an environmental monitoring study is proposed. This paper involves the joint analysis of zeroth-, first- and second-order data measured on a particular environmental system. The main advantage of this methodology is the possibility of analyzing the relationships of the different order data provided by several analytical techniques. This approach enables to achieve new knowledge, in a way that would be not accessible if considering the information individually.Environmental monitoring databases usually generate large amount of data. Multivariate statistical techniques are necessary to process all this information and obtain a correct interpretation. The Ludueña Stream located in Argentina was chosen as the study system. Samples from different sites of the basin were taken periodically. Conductivity and pH (zeroth-order data) were fused with near-infrared (NIR) spectra of suspended particulate material (first-order data) and with fluorescence emission-excitation matrices of dissolved organic matter (second-order data). Different chemometric algorithms made it possible to extract and merge all the information in a new database, enabling its later analysis as a whole.This methodology allowed to successfully studying the behavior of dissolved organic matter together with suspended particulate material and other specific variables, showing links between them. Their distributions along the basin and their evolutions over time were possible to obtain. Therefore, a simpler interpretation to evaluate the system status was achieved. This model allowed differentiating the variables affected by anthropic activities from those with a natural origin.

An investigation of changes in water quality throughout the drinking water production/distribution chain using toxicological and fluorescence analyses

Changes in water quality from source water to finished water and tap water at two conventional drinking water treatment plants (DWTPs) were monitored. Beside the routine water quality testing, Caenorhabditis elegans-based toxicity assays and the fluorescence excitation–emission matrices technique were also applied. Both DWTPs supplied drinking water that met government standards. Under current test conditions, both the investigated finished water and tap water samples exhibited stronger lethal, genotoxic and reprotoxic potential than the relative source water sample, and the tap water sample was more lethal but tended to be less genotoxic than the corresponding finished water sample. Meanwhile, the nearly complete removal of tryptophan-like substances and newly generated tyrosine-like substances were observed after the treatment of drinking water, and humic-like substances were identified in the tap water. Based on these findings, toxic pollutants, including genotoxic/reproductive toxicants, are produced in the drinking water treatment and/or distribution processes. Moreover, further studies are needed to clarify the potentially important roles of tyrosine-like and humic-like substances in mediating drinking water toxicity and to identify the potential sources of these contaminants. Additionally, tryptophan-like fluorescence may be adopted as a useful parameter to monitor the treatment performance of DWTPs. Our observations provided insights into the importance of utilizing biotoxicity assays and fluorescence spectroscopy as tools to complement the routine evaluation of drinking water.

Identifying the Stoichiometry of Metal/Ligand Complex by Coupling Spectroscopy and Modelling: a Comprehensive Study on Two Fluorescent Molecules Specific to Lead.

Two new chemosensors for lead (II) were synthesized based on 5-((anthracen-9-ylmethylene) amino)quinolin-10-ol (ANQ). ANQ was modified in the para position of the imine group via a methoxy link either with methylmethacrylate (ANQ-MMA) or styrene (ANQ-ST). Complexation of those molecules with Pb2+ was studied at room temperature using UV-Visible absorption and fluorescence spectroscopies. Thanks to the UV-visible absorption spectroscopy, it appeared that ANQ-MMA formed 1:1 and 1:2 complexes with lead (II) and ANQ-ST only 1:1 complex. For both molecules, the fluorescence excitation-emission matrices (EEM) signal intensity increased from 0 to 100μmol.L-1 of lead (II) followed by a saturation for higher concentrations. The decomposition of the obtained EEMs gave a set of empiric fluorescent components that have been directly linked to the distribution of lead complexes obtained with the UV-visible absorption spectroscopy study. This correlation allowed to evidence metal/ligand complex stoichiometry and emerge as a new method to identify empiric components. Moreover, the two ligands showed a promising selectivity for Pb2+, turning them interesting probes for this hazardous metal.

Pesticide residues in fruits and vegetables: High-order calibration based on spectrofluorimetric/pH data

The aim of the present work was to quantify simultaneously four pesticides, thiabendazole, fuberidazole, carbaryl and naphthyl acetic acid, in vegetables and fruits, measuring excitation–emission fluorescence matrices in a flow injection system with double pH–gradient. The generated third–order data were processed by applying multivariate calibration procedure unfolded partial least squares coupled to residual trilinearization (U–PLS/RTL), permitting overcome the severe spectra overlapping among the analytes and also with the background of real samples. The vegetable tissues were fortified at concentration levels taking into account the maximum residue level, acquiring good recoveries for all analytes, with values between 80 and 115%. The developed method enabled determining the analytes in complex samples (lettuce, pear, orange and mushroom) with a simple extraction step with ethyl acetate, in an appropriate analysis time and without sophisticated instruments.

Chemical Transfers Occurring Through Oenococcus oeni Biofilm in Different Enological Conditions.

Chardonnay wine malolactic fermentations were carried out to evaluate the chemical transfers occurring at the wood/wine interface in the presence of two different bacterial lifestyles. To do this, Oenococcus oeni was inoculated into must and wine in its planktonic and biofilm lifestyles, whether adhering or not to oak chips, leading to three distinct enological conditions: (i) post-alcoholic fermentation inoculation in wine in the absence of oak chips, (ii) post-alcoholic fermentation inoculation in wine in the presence of oak chips, and (iii) co-inoculation of both Saccharomyces cerevisiae and O. oeni directly in Chardonnay musts in the presence of oak chips. Classical microbiological and physico-chemical parameters analyzed during the fermentation processes confirmed that alcoholic fermentation was completed identically regardless of the enological conditions, and that once O. oeni had acquired a biofilm lifestyle in the presence or absence of oak, malolactic fermentation occurred faster and with better reproducibility compared to planktonic lifestyles. Analyses of volatile components (higher alcohols and wood aromas) and non-volatile components (Chardonnay grape polyphenols) carried out in the resulting wines revealed chemical differences, particularly when bacterial biofilms were present at the wood interface. This study revealed the non-specific trapping activity of biofilm networks in the presence of wood and grape compounds regardless of the enological conditions. Changes of concentrations in higher alcohols reflected the fermentation bioactivity of bacterial biofilms on wood surfaces. These chemical transfers were statistically validated by an untargeted approach using Excitation Emission Matrices of Fluorescence combined with multivariate analysis to discriminate innovative enological practices during winemaking and to provide winemakers with an optical tool for validating the biological and chemical differentiations occurring in wine that result from their decisions.

Investigating trivalent chromium biosorption-driven extracellular polymeric substances changes of Synechocystis sp. PCC 7806 by parallel factor analysis (PARAFAC) analysis

In this study, experiments were conducted to investigate the biosorption effect of the cyanobacteria Synechocystis sp. PCC 7806 on trivalent chromium (Cr (III)) and the fate of extracellular polymeric substances (EPS) during the biosorption process. Results indicated that Synechocystis sp. PCC 7806 could effectively adsorb Cr (III) but this adsorption capacity was inhibited by high amount Cr (III). The biosorption process decreased the growth rate of Synechocystis sp. PCC 7806. Furthermore, the characteristics of Synechocystis sp. PCC 7806 EPS were evaluated by parallel factor analysis (PARAFAC) combined with fluorescence excitation–emission matrices (EEMs). The results showed that there were four main components of EPS in Synechocystis sp. PCC 7806. It indicated that the fluorescent characterization of EPS was obvious affected by adsorption process of Cr (III). The tryptophan-like substances, amino acid substances and fulvic acids-like substances largely participated in the binding and biosorption of Cr (III).

Unsupervised classification of PSII with and without water-oxidizing complex samples by PARAFAC resolution of excitation-emission fluorescence images

The potential of excitation-emission fluorescence spectroscopy combined with three-way analysis was investigated for discriminating the photosystem II (PSII) (with the water-oxidizing complex) and without the water-oxidizing complex (wPSII) using unsupervised classification methods. The water-oxidizing complex within PSII carry out the reaction of water splitting which is as a vital process on the earth. Therefore, discriminating the presence of the water-oxidizing complex in protein samples is crucial. Low cost and accurate spectroscopic determination of the amount of clusters inside PSII or any other protein containing species are important when investigating the inclusion and exclusion of such clusters into and from species. Fluorescence data of samples were similar, and we showed the potential usefulness of multivariate methods, such as parallel factor analysis (PARAFAC) and principal component analysis (PCA) for recognition of the two types of samples. Both techniques were applied to the excitation-emission fluorescence matrices (EEM) of solutions at two of different pH values (2.0 and 12.0). Three fluorescent components were found for all samples that are related to tyrosine (Tyr), tryptophan (Trp) and phenylalanine (Phe) amino acids. These three amino acids are representative of all datasets and indicate their similarities and differences. We then found the effectual wavelengths for separation of samples in a specific acidity, including the excitation wavelengths of 220 and 230 nm and the emission wavelengths of 300 and 305 nm. The acidity of the solutions has various influences on the conformation of proteins. In PSII and PSII the without water-oxidizing complex samples conformational changes can change their spectra which was applied for discrimination purpose. This separation was better in pH = 12.0. We also showed the effect of time on small conformational changes within datasets were higher in pH = 2.0. In the end, for indicating the high distribution of spectral data from proteins which is the result of conformational changes, we compared the distribution of measured spectral data with that from a simple organic molecule, fluorescein. Altogether, we could distinguish between the two groups of protein samples properly at pH = 12.0 using low-cost EEM spectral images and PARAFAC.

Investigation of the fouling behaviors correlating to water characteristics during the ultrafiltration with ozone treatment

This study investigated the fouling behavior and mechanism of ozone treatment correlating to water characteristics for micro-polluted water during ultrafiltration (UF). The results indicated that pre-ozonation efficiently mitigated membrane fouling of natural organic matter (NOM). The higher ozone doses were, the more the performance transmembrane pressures (TMPs) decreased. Ozone mainly converted macro molecule organics into low molecule organics. Macro molecular biopolymers (BP) can be removed up to 35.5% with an ozone treatment of 9 mg/L, while low molecular weight building blocks of acids and humics (BB) and neutrals (LMWN) increased 7.25% and 14.62%, respectively, with an ozone treatment of 9 mg/L. Analysis of fluorescence excitation emission matrices (EEMs) coupled with parallel factor analysis (PARAFAC) indicated that ozone mainly removed soluble microbial organics and fulvic-like and humic-like organics but not tyrosine organics. Hydrophobic organics (HPO) were reduced with an increase of ozone doses, especially macro molecular BP and humic substances (HS), and the neutral hydrophilic fraction (N-HPI) was enhanced. Ozone treatment helped to reduce the interception of BP and HS in HPO and improved the interception of BP and HS in N-HPI, as well as BB and LMWN, in both fractions. Principal component analysis suggested that BP, as well as UV254, had high correlations with a membrane fouling index, which can be used as the fouling indicator during ozone treatment.

Comparing electron donating/accepting capacities (EDC/EAC) between crop residue-derived dissolved black carbon and standard humic substances

Although dissolved black carbon (DBC) is an important component of dissolved organic matter (DOM), little is known about the electron transfer properties of DBC and their correlation with the structural properties. This study determined the electron donating/accepting capacities (EDC/EAC) of six DBC collected from water extracts of black carbon derived by pyrolyzing crop residues (soybean, wheat, rice, sorghum, peanut, and corn) using an electrochemical approach, along with their chemical and spectroscopic properties. The EDC of the tested DBC ranged from 2.42 to 7.10 mmole-(gC)−1and was generally much higher than the EDC (1.31–3.78 mmole-(gC)−1) of the four standard dissolved humic substances (DHS). In contrast, with the exception of the sorghum DBC, the EAC of all DBC (0.40–0.81 mmole-(gC)−1) was apparently lower than the EAC of DHS (0.87–1.68 mmole-(gC)−1). For the whole pool of DBC and DHS, a strong positive correlation (r = 0.94, P < 0.05) existed between the EDC and the phenolic group content quantified by base titration, suggesting that phenolic groups were the major electron-donating groups. The EAC and EDC of the tested DOM pool were further correlated with the components identified from the fluorescence excitation emission matrices and parallel factor (EEM-PARAFAC) analysis. This is the first study to show that DBC has significantly different electron transfer properties from humic substances and thus likely behaviors differently in many geochemical and environmental aspects.Capsule: Dissolved black carbon (DBC) has significantly different electron-donating/accepting capacities (EDC/EAC) from humic substances.

Fluorescence determination of cochineal in strawberry jam in the presence of carmoisine as a quencher by means of four-way PARAFAC decomposition

The determination of cochineal (E-120) in strawberry jam was carried out in the presence of carmoisine (E-122) using the four-way PARAFAC decomposition and excitation-emission fluorescence matrices. In the measured conditions, there was no fluorescence signal for carmoisine due to a strong quenching effect and this colorant also led to a decrease of the fluorescence signal of cochineal. The European Union has fixed a maximum residue level, MRL, for cochineal in jam (100 mg kg−1). Therefore, the addition of other food colorant (carmoisine) in the jam could lead to false compliant decisions. The four-way PARAFAC decomposition avoided false compliant decisions caused by the quenching effect.Cochineal was unequivocally identified. Detection capability (CCβ) was 0.72 mg L−1 for probabilities of false positive and false negative fixed at 0.05. Cochineal was detected in the jam (104.63 mg kg−1) above the MRL. This amount was compared with the one obtained using a HPLC/DAD method.

The influence of tillage on dissolved organic matter dynamics in a Mid-Atlantic agroecosystem

Dissolved organic matter (DOM) represents a fraction of soil organic matter essential to many biogeochemical processes in surface soils and subsoils. Given its dynamic nature, DOM is known to be sensitive to land management practices. This study investigates the influence of tillage on DOM dynamics, including amount and composition of DOM, through a soil profile up to 72 cm depth from moldboard plowed (MB) and no-till (NT) plots in a Mid-Atlantic agroecosystem. Composition of DOM is probed using UV/Visible spectrophotometry (UV/Vis), Fourier transform infrared spectroscopy (FTIR) and fluorescence excitation-emission matrices (EEMs). Tillage decreases extractable DOM at the topmost depth interval studied (0–12 cm), with similar amounts extracted at depths below this point from MB and NT soils. DOM content of MB and NT soils decreases with depth, following a gradual decline in NT soils and a sharp decrease below the plow layer (20 cm) in MB soils. In both soils, chemical composition of DOM becomes less aromatic and less polysaccharide-like in character with depth accompanied by a relative enrichment in carboxylate groups. EEMs demonstrate that species within DOM follow a corresponding transition from conjugated fluorophores to moieties attributable to simple phenol, protein, amino acid and nucleic acid-like DOM constituents. These depth trends in functional group composition follow a distribution similar to DOM content, with observed transitions more abrupt in the MB than NT soils and marked by the plow layer. The decrease in DOM content with depth and the aforementioned shifts in DOM characteristics point to sorptive fractionation of DOM by the soil matrix and/or a conversion from more plant-derived DOM near the surface to more microbially processed DOM in subsoils. Overall, this work further highlights the impact of tillage on DOM behavior below the zone of physical disruption, suggesting that tillage also influences downstream DOM dynamics.

Combination of fluorescence excitation emission matrices in polar and non-polar solvents to obtain three- and four- way arrays for classification of Tempranillo grapes according to maturation stage and hydric status

The applicability of front-face excitation-emission fluorescence spectroscopy to compare grape water extracts of two consecutive sampling dates, corresponding with two maturation stages, and subjected to full irrigation and non-irrigation, was carried out. The decomposition of the obtained three way grape samples was initially analyzed by means of parallel factor analysis (PARAFAC). A tentative identification of fluorophores was done by matching PARAFAC score values with HPLC measurements. It was found that the first PARAFAC component was highly correlated with the sum of concentrations of catechin and epicatechin. The decomposition of the three way fluorescence data by linear discriminant analysis (LDA)-PARAFAC and LDA-unfolding partial least squares (U-PLS) allowed to discriminate between the first and the second maturation stage. The incorporation of an additional mode to the data, achieved by a diethyl ether extraction, giving rise to a four-way excitation-emission-solvent-samples data set, allowed to differentiate between irrigated and non-irrigated samples with the same assayed algorithms. As far as we know, the use of four-way data arrays for classification issues has been reported for the first time.

Tracking of Dissolved Effluent Organic Matter (dEfOM) in wastewater treatment plant by using fluorescence method

Discharge of wastewater into water body can change the quantity and quality of natural organic matter (NOM) in source water, because wastewater may contain dissolved effluent organic matter (dEfOM). dEfOM consists of a heterogenous mixture of organic compounds with various structures, such as dissolved organic matter, microbial by-products, endocrine disrupting compound, etc. The objectives of this study is to track the fate of dEfOM by using fluorescence excitation emission matrices (FEEMs) in order to apply of reclaimed water from industrial wastewater. Tracking of NOM was observed in industrial park of wastewater treatment, which is consist of equalization tank, coagulation, flocculation, sedimentation, activated sludge tank and clarifier tank. FEEMs with parallel factor analysis (PARAFAC) analysis decomposed the organic matter in wastewater into two components, namely component C1 at (Ex/Em): 330(235)/410 is humic acid like and fulvic, component C2 at (Ex/Em): 280(230)/350 is soluble microbial products and aromatic protein. Characteristic of dEfOM depends on the wastewater characteristic and its treatment processes.