Research On The Quantitative Analysis Of Complex Systems Using High-order Instruments Coupled With Chemical Multiway Calibration Methods

Two important features of modern analytical chemistry are the instrumentation of analytical tools and complication of analytical objects. It is because of the inherent interaction between the two major features that the modern analytical chemistry is being in an unprecedented “data tsunami” era. Fortunately, chemometrics can extract more useful information that analysts need from complex chemical measurement data, so as to provide a powerful tool for analytical chemistry to deal with the “data tsunami” challenge. Chemical multiway calibration, one of the research hotspots of chemometrics, in combination with modern high-order instruments for quantitative analysis of complex systems has the unique “second- or higher-order advantage”. By utilizing the green and smart “mathematical separation” to replace or enhance conventional “physical or chemical separation”, multiway calibration methods can avoid or simplify laborious and time-consuming sample pretreatment processes and exclude the impacts of background matrices and interfering signals, making the direct, simultaneous, rapid and accurate quantification of multiple components of interests even in the presence of unknown interferences possible. This analytical strategy has been increasingly recognized by the international researchers and widely used for the quantitative analysis of complex systems such as biological medicine, food and environment. Taking the development trends and application status of multiway calibration methods, together with the advantages and disadvantages of modern highorder instruments into account, this dissertation mainly focuses on the development and application of analytical methodologies based on high-order instruments coupling with multiway calibration, the simplification of high-order instruments as well as the development of some novel algorithms to handing non-multilinear multiway data, which is expected to be complementary, efficient, low-cost and environmentallyfriendly. The main contents of this dissertation are as follows:Part I: Excitation–emission matrix fluorescence coupled with three-way calibration methods for quantitative analysis of complex systems and its ability to deal with varying interfering patterns from different backgrounds(Chapter 2 and Chapter 3)Metoprolol(MET) has been widely used for treating cardiovascular disorders. It undergoes extensive metabolism in vivo and generates a main active metabolite named α-hydroxymetoprolol(α-OHM). Usually, it’s difficult to achieve complete chromatographic separation for these two compounds via traditional chromatographic method under simple elution program due to their extremely similar chemical structures and physical properties. In addition, because the fluorescence spectra of the two compounds overlap with each other seriously and there are a lot of unknown endogenous fluorescence interfering compounds in human plasma, conventional spectrophotometry cannot avoid the overlapped peaks and unknown interferences by selecting a special wavelength. In Chapter 2, a new and effective strategy that combines excitation–emission matrix fluorescence with second-order calibration methods based on parallel factor analysis(PARAFAC) algorithm and full-rank parallel factor analysis(FRA-PARAFAC) algorithm was developed for the simultaneous determination of MET and α-OHM in human plasma. The proposed strategy utilizes the “mathematical separation” of three-way calibration methods instead of traditional “physical and chemical separation” to avoid the pretreatment process and possesses characteristics of simple, fast, low-cost as well as environmentally-friendly. The experimental results indicated that the strategy is sensitive and accurate. This strategy is expected to be extended as a real-time, fast and simple technique for clinical drug quantitative analysis, which can provide reliable theoretical and technical basis for rational drug use and optimization of therapy.Alternating trilinear decomposition(ATLD) is one of the most commonly used second-order calibration methods. In its practical applications, it is found that ATLD not only holds the advantages of being insensitive to overestimated number of components and fast convergence, but also exhibits the ability to maintain the distinctive “second-order advantage” when handling excitation–emission matrix fluorescence data containing different backgrounds and interfering patterns. In other words, different backgrounds do have little effects on the quantitative results of ATLD as long as the measured data sets meet the trilinear component model. However, it is regrettable that this point has not yet been deeply investigated and elaborated. In Chapter 3, the property of ATLD to handling different backgrounds and interfering patterns was demonstrated by simulated and experimental multi-components’ excitation-emission matrix fluorescence data containing different backgrounds and interfering patterns. Moreover, for demonstrating the ability of ATLD to determine shared multi-components in multiple complex systems containing different backgrounds using one set of calibration samples, the predictive results obtained from individually and simultaneously analyzing this kind of samples by ATLD were compared. This study attempts to provide some valuable reference information to analysts who want to utilize ATLD algorithm to establish a simultaneous, fast and accurate method for the quantitative analysis of multiple components of interest in complex systems.Part II: Novel strategy for quantitative analysis based on chemometric mathematical separation-enhanced full scan mode of liquid chromatography-mass spectrometry and its application in complex systems(Chapter 4 and Chapter 5)Liquid chromatography-tandem mass spectrometry(LC–MS/MS) operated in multiple reaction monitoring(MRM) mode has been recognized as the “gold standard” approach for the analysis of target compounds in complex systems owing to its inherent characteristics of sensitive, fast-scanning and reproducibility with wide dynamic range. Nevertheless, the MRM technique needs expensive sophisticated apparatus and laborious chromatographic and tandem mass spectrometric conditions optimization. In Chapter 4, in order to overcome the above disadvantages, a novel strategy for quantitative analysis based on chemometric mathematical separationenhanced full scan mode of liquid chromatography–mass spectrometry(LC–MS) was firstly developed. This strategy skillfully utilizes the “mathematical separation” instead of the second quadrupole(Q2) and third quadrupole(Q2) to achieve the secondary ion selections, resulting in the equal or better quantitative ability as MRM method. Compared with traditional MRM method, the whole implementation of the proposed strategy is low-cost and no requirements of laborious chromatographic and tandem mass spectrometric conditions optimization. It has the advantages of simple, fast, high sensitivity and selectivity. In order to validate the feasibility of this strategy, it is applied to simultaneous determination of ten β-blockers in human urine and plasma samples, and satisfactory qualitative and quantitative results were obtained.Sulfonylurea-type oral antidiabetic agents(SOADs) are one of main type of oral antidiabetics for treating type II diabetes. Common adverse side-effects caused by the abuse of these drugs include life-threatening hypoglycemia, obesity, gastrointestinal upset, skin allergy and liver or kidney damage. Besides, this kind of drugs were listed as the prohibited substances in athletic competitions. Therefore, rapid, reliable, sensitive, and accurate methods are required for screening and determination of synthetic SOADs in “antidiabetic” health products and human plasma samples to avoid the harmful effects of these synthetic drugs. In Chapter 5, to further demonstrate the ability of the chemometrics-enhanced LC–MS strategy proposed in Chapter 4 for the quantitative analysis of complex systems, it was once again employed to determine six co-eluted SOADs in health teas and human plasma samples. Six SOADs were eluted completely within 5.5 min under an isocratic elution program(30% water: 70% acetonitrile, v/v). The strategy utilizes the distinctive “second-order advantage” of alternating trilinear decomposition(ATLD) method to achieve successful resolution and accurate quantification for six target analytes even in the presence of unresolved peaks and unknown interferences. The experimental results indicated that most of the degrees of fit(DOF) between resolved spectra and real spectra of six target analytes were close to 100% and their average spiked recoveries were between 81.6% and 110.1%, together with limit of detection within the range of 0.2-30 ng m L-1. It was demonstrated that the chemometrics-enhanced LC–MS strategy could be promisingly used as an efficient and green quantitative analysis method for determination of multianalytes of interest in complex samples while avoiding elaborate sample pretreatment steps and complicated experimental conditions as well as sophisticated high-cost instrumentations.Part III: New strategies for eliminating non-linear factors in the quantitative analysis of complex systems using high-order instruments coupled with multiway calibration methods(Chapter 6 and Chapter 7)Chemometrics-assisted full scan mode of liquid chromatography-mass spectrometry(LC–MS) uses the “mathematical separation” of multiway calibration methods to simplify the instrumental apparatus, which combines high sensitivity of single mass spectrometry and good selectivity of “mathematical separation” and thus can overcome the disadvantages of expensive apparatus and laborious conditions optimization of liquid chromatography-tandem mass spectrometry(LC–MS/MS). It exhibits great application potential in quantitative analysis of complex systems. However, the performances of chemometrics-assisted LC-MS depend on the operation conditions of the instrument seriously. For example, the signal of the instrument would be fluctuation due to the gradual fouling of the ion source, the vacuum instability, the temperature changes, and so on, resulting in the signal instability of LC–MS in different time. In this case, the resolution and afterwards quantification of target compounds will become inaccurate. In Chapter 6, a novel strategy that combines piecewise direct standardization and alternating trilinear decomposition method(PDS/ATLD) was developed to solve the signal instability of LC–MS and maintain the “second-order advantage”. The proposed strategy was validated by using both simulated and experimental data sets. Satisfactory results indicated that PDS/ATLD can replace the recalibration strategy to solve signal instability of LC-MS in different days, and can achieve simultaneous determination of multiple analyte(s) of interest in complex systems even in the presence of unknown interferences and seriously overlapped peaks.In the practical measurement of four-way data, some enviromental factors(e.g., temperature, pressure and time) or inevitable human factors usually affect the linearity of certain dimensions such as chromatography and kinetics, resulting in nonquadrilinear four-way data. In such cases, four-way calibration methods based on quadrilinear component model are not suitable for decomposing this type of data due to their strict requirements on the quadrilinear structure of the data. In Chapter 7, a novel augmented alternating trilinear decomposition(Augmented ATLD) algorithm was proposed to handing non-quadrilinear four-way data. The philosophy behind this algorithm is to unfold the non-quadrilinear four-way data array into a trilinear threeway augmented data array along the nonlinear dimension, and then decompose the trilinear three-way augmented data array into three underlying matrices by using alternating trilinear decomposition(ATLD) algorithm, which can eliminate the effects of nonlinear factors and show great advantages in handing non-quadrilinear four-way data. The performances of the Augmented ATLD algorithm were demonstrated by two sets of simulated data and two sets of experimental data. Results indicated that for decomposing the simulated and experiential non-quadrilinear four-way data, both the alternating quadrilinear decomposition(AQLD) and four-way parallel factor analysis(Four-way PARAFAC) algorithms cannot obtain satisfactory qualitative or quantitative results; while the proposed Augmented ATLD algorithm can obtain satisfactory qualitative and quantitative results.