User:Fraserhof/FluorescenceExcitation-EmissionMatrix

Introduction
Excitation-emission matrix (EEM) fluorescence spectroscopy is a type of spectroscopy that produces a three-dimensional spectrum of fluorescence data. Measurement of fluorescence allows for the quantification of various properties on a sample, so EEM spectroscopy is used in analytical chemistry.

An excitation-emission matrix produces a multidimensional contour plot of excitation wavelength and emission wavelength against fluorescence intensity. This topographic spectrum produces much more information that a typical fluorescence spectrum.

Theory
In molecules, electrons only exist in quantized energy levels. These energy levels, known as singlet states (denoted by the letter S) can be further subdivided into vibrational energy states (denoted by the letter V) which are also quantized. Under normal conditions, most molecules occupy the lowest energy vibrational state of the lowest energy state. This is called the ground state (S0).

The first step of fluorescence is the excitation of an electron. An electron in the ground state absorbs energy from a photon and is promoted to a higher energy state.

As the electron relaxes back to the ground state it releases energy. The electron releases energy down to the lowest excited state through vibration, then returns to the ground state through fluorescence. The emission band for the electronic transition is shifted to a lower energy than absorption band due to the loss of energy from vibrational transitions, this is known as the Stokes shift. The Jablonski diagram represents the electronic transitions which are responsible for photoluminescence.

Fluorescence spectroscopy determines the concentration or identification of an analyte. Fluorescence intensity follows Beer’s Law, which means that the fluorescence intensity is proportional to the sample’s concentration. There are many naturally fluorescing molecules that can be analyzed, as well as synthesized fluorescent compounds can also be and used as tags in chemical reactions.

Fluorescence spectroscopy is a highly sensitive method of analysis. The concentration parameter for fluorescence is proportional to the source radiant power, whereas other spectrographic methods require analysis of source radiant power and detected radiant power. This means that fluorescence spectroscopy does not require comparison to a reference light source and is measured directly. This leads to a spectrometer with low detection limits that is highly sensitive.

Excitation-emission matrix spectroscopy shows luminescence signal as a function of excitation and emission wavelengths, producing a detailed three-dimensional fluorescence spectrum of the analyte. The signal produced from the fluorescence transition provides information about concentration, structure and properties of all fluorescent compounds in a sample.

Instrumentation
The classical set up for data collection of EEM fluorescence is as follows. A single light source shines through the first monochromator which filters the initial excitation wavelength of the sample. The emission monochromator is positioned at a 90-degree angle from the excitation axis to measure the sample's fluorescence emissions. A second monochromator filters emission wavelengths and directs them into a detector for the sample's fluorescence quantification.

Light Sources
Light sources in EEM fluorescence are used to excite the sample. The light source vertically illuminates the sample to allow for separations of the fluorescent light wavelengths. Light sources are often single line excitation sources such as; lasers, or they are continuous broadband light sources such as Xenon lamps and light-emitting diodes (LEDs).



Single line excitation sources such as lasers are a good excitation source. Lasers are high-power, which results in very distinct, resolved peaks. The disadvantage of using lasers is that they only offer one excitation wavelength. Therefore, producing the EEM fluorescence 3D spectra would require an array of different lasers tuned to different wavelengths, making this method time-consuming and expensive.

The use of LED-arrays is an excellent low-cost replacement to typical high powered fluorescence light sources. LED excitation on their own can result in wavelength overlap regions, creating broad regions in the spectrum making it difficult to decipher the 3D EEM fluorescence spectra. However, with an LED-diode, individual LEDs are tuned to different excitation wavelengths and intensities. The LED-diode allows for individual excitation of the analyte and provides the 3D information needed for the EEM analysis. LED-diodes provide sufficient power to allow for differentiation of even closely related analytes. The disadvantage of this low power light source is that the LEDs cannot emit low UV rays, limiting the excitation range of the analyte to long UV, VIS and NIR wavelengths.

Measurement Parameters
Measurements of samples are performed in fluorescence-quartz cuvettes. The sample must contain fluorophores, either from the molecular composition of the analyte or fluorescent dyes tagged to the molecule. To optimize the sample's fluorescence, the pH should be between 5-8, and the temperature must remain constant during measurement. Additionally, the sample must be filtered to remove any particle larger than 0.45 micrometres, as this can saturate the spectrum. Further, sample solutions need appropriate dilution to avoid the broadening of fluorescent emissions. Classical EEM fluorescence detectors are broadband photodetectors such as avalanche photodetectors, photomultiplier or array detectors. However, these detectors have low sensitivity, resulting in broad bands making it difficult to decipher between individual fluorophores. Increasing the dilution of the sample decreases the broadening of peaks in the EEM fluorescence spectra, but this does not eliminate peak broadening completely.

Another issue with classical EEM spectroscopy is the efficiency of collecting data. The two monochromator method of collecting EEM spectra is time-consuming, taking minutes to acquire a single EEM spectrum. Making the mapping of real-time reaction kinetics with EEM fluorescence impossible.

Spectrographs
Researchers and optical instrument companies have combated the issue of high acquisition time by replacing the emission monochromator with an emission imaging spectrometer or spectrographs. The use of spectrometers is a common modification of EEM fluorescence instrumentation and increases the speed of spectra generation. The spectrograph is a multichannel detector that simultaneously splits the incoming emission by its wavelength, and reduces the acquisition time for the EEM recording from minutes to seconds.

The fluorescence quenching and the inner-filter effect are two significant problems impacting accuracy when using a spectrograph in EEM fluorescence. Fluorescence quenching decreases the fluorescence intensity as molecules in the sample interact with the excited fluorophores. Resulting in a decrease in the excited fluorophores' lifetime and decreases the amount of fluorescence emitted. Fluorescence quenching is combated by adjusting the sample's parameters (pH, temperature and dilution) to reduce the interactions between fluorophores and molecules. The inner filter effect (IFE) affects the quantum yield of the fluorescence emissions in EEM fluorescence. IFE results from the loss of the optical density of the sample's emission signals. The fluorescence output is decreased by absorbance from another absorber in the sample solution. This absorber, often an analyte, decreases the emissions travelling towards the detector resulting in less fluorescence. The IFE will diminish by diluting the sample to limit the absorption of emission by the analyte. However, EEM fluorescence requires absorption from the analyte for the sample's excitation, so IFE is always present.

A-TEEM™
Horiba scientific pioneered an adaptation to combat IFE. They developed an absorbance transmittance fluorescence excitation-emission matrix (A-TEEM™) system that internally corrects IFE, eliminating spectral distortion. Absorbance spectra are collected simultaneously with full emission spectra at each excitation increment of the sample. With the A-TEEM™) normalization software, the emission spectra are corrected in-time from the absorbance spectra. This system increases EEM fluorescence measurement ease as there is no need to formulate specific IFE corrections equations . Additionally, A-TEEM™ decreases the broadening of peaks, which allows for better fingerprint analysis.

Applications
Fluorescence excitation-emission matrix spectroscopy has applications in a wide range of areas. The three-dimensional fluorescence fingerprint is especially useful when analyzing mixtures. In research as well as in various industries, it provides a wealth of information including chemical composition, concentration, hydrophobicity and molecular weight. The three-dimensional graph produced can also probe the chemical complexation and changes in chemical structure. Changing structure and bond connectivity alters fluorescence intensity and shifts the excitation and emission wavelengths. EEM has applications in environmental chemistry, the food industry and quality control in manufacturing.

The applications of fluorescence EEM are somewhat limited. The technique on its own cannot be used to determine reaction kinetics or perform real time analysis because the acquisition time is slow. On its own, its primary application is quantitatively determining concentration, or qualitatively determining presence of impurities or contaminants.

Dissolved Organic Matter
Dissolved organic matter (DOM) is a class of water-soluble compounds that contains primarily carbon, hydrogen and oxygen, but also can include nitrogen, phosphorous and sulfur. DOM is an important natural component of marine systems as well as in bodies of fresh water, but the composition must be monitored as they can affect ocean acidity and drinking water quality.

The most prevalent application for EEM spectroscopy is in environmental analytical chemistry for the analysis of dissolved organic matter (DOM) in various samples. Fluorescence excitation-emission matrix spectroscopy is highly sensitive and selective, as well as simple to operate with minimal sample preparation. This makes EEM spectroscopy effective for analysis of DOM’s because many samples contain a lot of complex organic matter which can be difficult to detect and to pre-treat.

Landfill leachate is a fluid produced in landfills that requires extensive treatment. Many components of landfill leachates contain fluorescent groups making them good samples for EEM fluorescence. The DOM from this liquid is analyzed to determine composition and concentration of fluorescent components in order to guide treatment protocols and assess environmental risk. Dissolved organic matter is analyzed in drinking water and in the ocean using fluorescence excitation-emission matrix spectroscopy. It has been used to analyze contaminants and pollutants, and to identify hydrocarbons dissolved in the ocean from crude oil spills and bilge water discharges through matching of spectrographic data. Researchers have chosen this technique over other spectroscopy methods because of the increased rate of data acquisition for higher sample throughput which is a valuable feature for many applications. This technique is also favoured for the large amounts of data it produces when compared to traditional fluorescence spectrometers.

Food Industry
Fluorescence excitation-emission matrix spectroscopy has been frequently used in the food industry. Many foods and drinks contain fluorescent components including proteins, vitamins, pigments and flavourings, which allow samples to be tested using fluorescence spectroscopy. EEM is an increasingly popular method for product quality testing because it is non-destructive, fast and environmentally safe.

Manufacturing
EEM spectroscopy has also been used for quality control in manufacturing. It has been adapted on several occasions for real-time data collection for quality assurance in the production process. A specific example used EEM for identification and quantification of fuels and oils, and to assess the oxidative stability of lubricant oil in real-time analysis.

Fluorescence region integration
To characterize the fluorescence distribution in EEM spectral regions, fluorescence region integration (FRI) is an analysis method. It integrates the fluorescence intensity of each region to calculate the regional for the total fluorescence. Excitation/Emission wavelength regions for the distribution of EEM fluorescence with respect to the composition of the DOM. FRI computes an EEM spectrum into five main regions : regions I and II (Ex<250 nm, Em<380 nm) is related to simple aromatic proteins such as tyrosine-like and tryptophan-like substances; region III (Ex<250 nm, Em>380 nm) corresponds to fulvic-like substances; region IV (Ex>250 nm, Em<380 nm) is related to soluble microbial byproduct-like substances; and region V (Ex >250 nm, Em >380 nm) represents humic-like substances. FRI and fluorescence peak identification have been widely used to track the transformation of DOM and investigate major contributors to membrane fouling.

Parallel factor analysis
Parallel factor analysis (PARAFAC) and parallel factor framework clustering analysis (PFFCA) have been used to decompose EEM spectra in order to produce quantifiable data. The model of an algorithm modifies the EEM data set into trilinear terms which is to estimate the EEM spectra and an array of residuals. There are F effective components, the fluorescence intensity of the i’th sample at the j’th emission wavelength and the k’th excitation wavelength modeled as a summed product of the component score (a), emission loading (b), and excitation loading (c), as shown as,

Xijk=∑_(f=1)^Faifbifckf+εijk

where εijk: residual term; a: relative concentration, b: the emission, c: the excitation spectra of the component.

PARAFAC is not suitable to handle non-trilinear EEM data and fails to get a unique solution when processing complex environmental samples. It generally treats the Rayleigh and Raman scatterings as interfering signals. All components are linearly independent assumed in PARAFAC, but this becomes invalid for complex cases. A single dissolved organic matter (DOM) molecule carries more than one fluorescent component simultaneously, these components will be completely correlated. In order to solve this point, PFFCA recombines the collinear PARAFAC components into physically meaningful components via clustering analysis. The PFFCA process consists of two steps: (a) decomposing the EEM data within the PARAFAC framework, as shown in the equation, and (b) clustering the components into appropriate groups. In the first step, the data matrix is changed into a set of three-line terms and residual arrays using the PARAFAC model.

F (the factor number) is determined by increasing F until the relative squared sum of the residual is less than a predefined threshold. Then, factors with a large variation in the scores are selected to form a data set G. Clustering analysis is performed on G with the final number of clusters equal to the number of principal components of G. Each cluster represents an effective fluorescent component in the samples. Compared with PARAFAC, PFFCA can process non-trilinear EEM data, the resultant components are more explicable and the obtained estimates are closer to the actual EEM spectra. Moreover, PFFCA can identify and separate Raman and Rayleigh scatterings so it is unnecessary to remove them prior to the analysis. Other methods are available to analyze EEM data, such as the self-organizing map (SOM), multivariate curve resolution (MCR), principal filter analysis (PFA), principal components regression (PCR), partial least squares regression (PLS), PLS discriminant analysis (PLS-DA), and soft independent modeling of class analogy (SIMCA).

EEMs with FRI, PARAFAC, and other analytical methods together have been effectively applied to DOM studying. Each peak of fluorescence and fluorescent component indicates the transfer and degradation of DOM components. The shifts of the fluorescence peaks reflect the changes in DOM structures along the process. It is related to the increase of carbonyl, hydroxyl, alkoxy, and amino groups in the functional group structure whereas other shifts are related to the decrease in characteristic functional groups, the degree of π-electron conjugation, and the number of aromatic rings conjugated bonds. It was investigated that the variations of DOM in the influent, anoxic phase, aerobic phase, and effluent of Membrane bioreactor (MBR) by FRI analysis.