Predicting and evaluating separation quality of micellar electrokinetic capillary chromatography by artificial neural networks

Computer-aided optimization of micellar electrokinetic capillary chromatography (MEKC) separations was demonstrated by artificial neural networks (ANNs) using a Levenberg-Marquardt algorithm and an orthogonal experimental design. A novel criterion, named Q, for evaluating the separation quality of MEKC was firstly presented, which considered both separation selectivity and analysis time. MEKC separation conditions of seven plant hormones were then simulated and optimized using ANNs based on this novel criterion. The result was further compared to that obtained using ANNs based on a traditionally used criterion of overall normalization resolution (named r). Finally, the separation under optimum conditions predicted by ANNs using the criterion Q was compared to, and proved to be better than that obtained by empirical step-by-step optimization procedures. This method may also be adapted to other separation methods due to its generality.     

Application of Artificial Neural Networks in Multifactor Optimization of Selectivity in Capillary Electrophoresis

ABSTRACT

A methodology based on the coupling of experimental design and artificial neural networks (ANNs) was proposed in the optimization of selectivity in capillary electrophoresis. The effect of the buffer composition, concentration, SDS concentration, ethanol percentage and the applied voltage on the separation of six choice solutes was examined by using orthogonal design. Feedforward-type neural networks with faster back propagation (BP) algorithm were applied to model the separation process, and then optimization of the experimental conditions was carried out in the modeled neural network with 5-7-1 structure, which had been confirmed to be able to provide the maximum performance. It was demonstrated that by combining ANN modeling with experimental design, the number of experiments necessary to search and find optimal separation conditions can be reduced significantly. Because of its general validity, the new proposed approach can also be applied in other separation conditions.     

Artificial neural networks for quantification in unresolved capillary electrophoresis peaks

The application of the combination of experimental design (ED) and artificial neural networks (ANNs) for the quantification of overlapped peaks in capillary zone electrophoresis is described. When the total separation cannot be achieved by separation techniques, the use of ED-ANN can be a suitable approach. The unstability of EOF causes peak shift that has to be corrected in order to apply ED-ANN methods. In this work, normalization procedure of electropherograms with consequent application of ANNs for quantification purpose was developed. Both, spectra and electropherograms can be used as multivariate data. In general, both kinds of data were found to be suitable for unresolved peaks quantification by ED-ANN approach.     

Artificial neural networks for modeling electrophoretic mobilities of inorganic cations and organic cationic oximes used as antidote contra nerve paralytic chemical weapons

Electrophoretic mobility of various analytes can be modeled and thus also predicted using artificial neural networks (ANNs) evaluating experiments done according to a suitable experimental design. In contrast to response surfaces modeling which can be used to predict optimal separation conditions, ANNs combined with experimental design were shown to be efficient for modeling and prediction of optimal separation conditions, while no explicit model and any knowledge of the physicochemical constants is needed. Methodology has been developed and demonstrated on separation of inorganic cations and organic oximes while various additives (methanol, complexation agent), pH or buffer concentration were followed. In our approach proposed the number of experiments necessary to find optimal separation conditions can be reduced significantly.     

Data processing in metabolic fingerprinting by CE-UV: application to urine samples from autistic children

Metabolic fingerprinting of biofluids such as urine can be used to detect and analyse differences between individuals. However, before pattern recognition methods can be utilised for classification, preprocessing techniques for the denoising, baseline removal, normalisation and alignment of electropherograms must be applied. Here a MEKC method using diode array detection has been used for high-resolution separation of both charged and neutral metabolites. Novel and generic algorithms have been developed for use prior to multivariate data analysis. Alignment is achieved by combining the use of reference peaks with a method that uses information from multiple wavelengths to align electropherograms to a reference signal. This metabolic fingerprinting approach by MEKC has been applied for the first time to urine samples from autistic and control children in a nontargeted and unbiased search for markers for autism. Although no biomarkers for autism could be determined using MEKC data here, the general approach presented could also be applied to the processing of other data collected by CE with UV-Vis detection.     

Optimisation of the chlorthalidone chiral separation by capillary electrochromatography using an achiral stationary phase and cyclodextrin in the mobile phase

The separation of chlorthalidone enantiomers in capillary electrochromatography on an achiral stationary phase when adding a chiral selector, hydroxypropyl-β-cyclodextrin, to the mobile phase, was optimised. The goal was to investigate the feasibility of modelling retention times and resolution when during the optimisation procedure regular replacement of columns is required due to their fragility. Therefore, it is essential that the packing procedure delivers reproducible columns. The optimisation of an existing chlorthalidone separation was chosen as case study. The influence of two factors, chiral selector concentration and organic modifier content, on the responses was modelled. The experiments performed prior to modelling were defined by a central composite design. Results on different columns, obtained under identical experimental conditions, were found comparable and thus modelling was possible in situations where several columns were required to complete a design. A second-order polynomial model was built for both responses. Optimal separations were also predicted using Derringer’s desirability functions. The optimum was found at 33 mM cyclodextrin and 16% (v/v) acetonitrile on two types of columns (with different packing times) leading to a strong reduction in analysis time for an equally good separation compared to the initial conditions. Measured and predicted responses were found comparable, indicating that acceptable models were obtained.     

Micellar electrokinetic chromatography - From theoretical concepts to real samples (Review)

Micellar electrokinetic chromatography (MEKC) is an alternative to liquid chromatographic separations. It is a highly efficient separation technique that is performed with the same experimental set-up as is used for capillary electrophoresis (CE), thus extending the applicability of CE to neutral solutes. MEKC can be regarded as a separation technique with a similar scope to that of reversed-phase high-performance liquid chromatography (RP-HPLC), having advantages over HPLC with regard to the efficiency of the separation system, separation speed, cost, and tolerance to matrix constituents. This paper discusses the applicability of MEKC to real samples and also addresses developments widening the scope of this emerging technique: on-line concentration by stacking or sweeping and sensitive detection schemes.     

Automated off-line optimisation of programmed elutions in reversed-phase high-performance liquid chromatography using ternary solvent mixtures

A new system for the unattended optimisation of gradient elutions in reversed-phase high-performance liquid chromatography has been developed. The system is based on the simulation of retention times under conditions of ternary solvent mixtures and ternary gradient programmes. This model is constructed departing from a few experimental data obtained in isocratic elutions and validated against experimental gradient elutions. Once validated, this retention model can be used for the unattended search of an optimum separation. The optimisation process is driven by an evolutionary algorithm (EA), specially developed to map the ternary gradient elutions problem. The development and characteristics of the retention modelling as well as those of the EA and its particularities are discussed and some real world examples of separation presented.     

Artificial Neural Networks in water analysis: Theory and applications

Artificial Neural Networks (ANNs) have seen an explosion of interest over the last two decades and have been successfully applied in all fields of chemistry and particularly in analytical chemistry. Inspired from biological systems and originated from the perceptron, i.e. a program unit that learns concepts, ANNs are capable of gradual learning over time and modelling extremely complex functions. In addition to the traditional multivariate chemometric techniques, ANNs are often applied for prediction, clustering, classification, modelling of a property, process control, procedural optimisation and/or regression of the obtained data. This paper aims at presenting the most common network architectures such as Multi-layer Perceptrons (MLPs), Radial Basis Function (RBF) and Kohonen's self-organisations maps (SOM). Moreover, back-propagation (BP), the most widespread algorithm used today and its modifications, such as quick-propagation (QP) and Delta-bar-Delta, are also discussed. All architectures correlate input variables to output variables through non-linear, weighted, parameterised functions, called neurons. In addition, various training algorithms have been developed in order to minimise the prediction error made by the network. The applications of ANNs in water analysis and water quality assessment are also reviewed. Most of the ANNs works are focused on modelling and parameters prediction. In the case of water quality assessment, extended predictive models are constructed and optimised, while variables correlation and significance is usually estimated in the framework of the predictive or classifier models. On the contrary, ANNs models are not frequently used for clustering/classification purposes, although they seem to be an effective tool. ANNs proved to be a powerful, yet often complementary, tool for water quality assessment, prediction and classification.     

Separation and quantification of 9-(alkylthio)acridines by capillary micellar electrokinetic chromatography and capillary liquid chromatography

Various thioacridine derivatives are potential chemotherapeutics against various diseases which are intensively synthesized, characterized, and investigated by many research groups. Efficient, fast, and reliable separation and quantification methods for their analysis are still to be developed. MEKC and capillary LC (CLC) were applied for the separation and quantification of five highly hydrophobic, weakly basic, and structurally similar 9-(alkylthio)acridines. Since the common anionic and cationic surfactants failed to separate the strongly hydrophobic thioacridines by MEKC, sodium cholate was used in an alkaline BGE and successfully employed for their fast separation. In CLC, the weakly basic nature of the thioacridines necessitated use of LiChrosorb RP-select B sorbent as the stationary phase, which combined with a very simple mobile phase methanol/water yielded an efficient chromatographic separation system. Both, the MEKC and CLC optimized separation methods were then applied to quantify the thioacridines within a concentration range of 1.0 x 10(-5)-1.0 x 10(-3) mol/L and the obtained experimental results were critically compared. In practical terms, the MEKC analytical method can quantify the analytes much faster but with a lower reliability while the CLC method performs slower analysis with a higher repeatability of the experimental results.     

Optimization of the separation of a group of triazine herbicides by micellar capillary electrophoresis using experimental design and artificial neural networks

The micellar electrokinetic chromatography separation of a group of triazine compounds was optimized using a combination of experimental design (ED) and artificial neural network (ANN). Different variables affecting separation were selected and used as input in the ANN. A chromatographic exponential function (CEF) combining resolution and separation time was used as output to obtain optimal separation conditions. An optimized buffer (19.3 mM sodium borate, 15.4 mM disodium hydrogen phosphate, 28.4 mM SDS, pH 9.45, and 7.5% 1-propanol) provides the best separation with regard to resolution and separation time. Besides, an analysis of variance (ANOVA) approach of the MEKC separation, using the same variables, was developed, and the best capability of the combination of ED-ANN for the optimization of the analytical methodology was demonstrated by comparing the results obtained from both approaches. In order to validate the proposed method, the different analytical parameters as repeatability and day-to-day precision were calculated. Finally, the optimized method was applied to the determination of these compounds in spiked and nonspiked ground water samples.     

Application of a multivariate approach for analyte focusing by micelle collapse‐micellar electrokinetic chromatography for analyzing sunscreen agents in cosmetics

The operating parameters that affect the performance of the online preconcentration technique “analyte focusing by micelle collapse‐MEKC (AFMC‐MEKC)” were examined using a multivariate approach involving experimental design to determine the sunscreen agents in cosmetics. Compared to the single‐variable approach, the advantage of the multivariate approach was that many factors could be investigated simultaneously to obtain the best separation condition. A fractional factorial design was used to identify the fewest significant factors in the central composite design (cCD). The cCD was adopted for evaluating the location of the minimum or maximum response in this study. The influences of the experimental variables on the response were investigated by applying a chromatographic exponential function. The optimized condition and the relationship between the experimental variables were acquired using the JMP software. The ANOVA analysis indicated that the Tris pH value, SDS concentration, and ethanol percentage influenced the separation quality and significantly contributed to the model. The optimized condition of the running buffer was 10 mM Tris buffer (pH 9.5) containing 60 mM SDS, 7 mM γ‐CD, and 20% v/v ethanol. The sample was prepared in 100 mM Tris buffer (pH 9.0) containing 7.5 mM SDS and 20% v/v ethanol. The SDS concentration in the sample matrix was slightly greater than the CMC value that makes the micelle be easily collapsed and the analytes be accumulated in the capillary. In addition, sunscreen agents in cosmetics after 1000‐fold dilution were successfully determined by AFMC‐MEKC.     

Separation and analysis of cis-diol-containing compounds by boronate affinity-assisted micellar electrokinetic chromatography

Cis-diol-containing compounds (CDCCs) are usually highly hydrophilic compounds and are therefore difficult to separate by conventional reversed-phase-based micellar electrokinetic chromatography (MEKC) due to poor selectivity. Here, we report a new method, called boronate affinity-assisted micellar electrokinetic chromatography (BAA-MEKC), to solve this issue. A boronic acid with a hydrophobic alkyl chain was added to the background electrolyte, which acted as a modifier to adjust the selectivity. CDCCs can covalently react with the boronic acid to form negatively charged surfactant-like complexes, which can partition into micelles formed with a cationic surfactant. Thus, CDCCs can be separated according to the differential partition constants of their boronic acid complexes between the micellar phase and the surrounding aqueous phase. To verify this method, eight nucleosides were employed as the test compounds and their separation confirmed that the combination of boronate affinity interaction with MEKC can effectively enhance the separation of CDCCs. The effects of experimental conditions on the separation were investigated. Finally, the BAA-MEKC method was applied to the separation and analysis of nucleosides extracted from human urine. BAA-MEKC exhibited better selectivity and improved separation as compared with conventional MEKC and CZE. Successful quantitative analysis of urinary nucleosides by BAA-MEKC was demonstrated.     

Theoretical models of separation selectivity for charged compounds in micellar electrokinetic chromatography

Equations and theoretical models for MEKC separation selectivity (α(MEKC) ) were established to explain a change in separation and electrophoretic mobility order of fully charged analytes, in which α(MEKC) is related to the dimensionless values of mobility selectivity in CZE (α(CZE)) and retention selectivity (α(k)) in MEKC, and where α(CZE) and α(k) are defined as the ratio of electrophoretic mobility in CZE and the ratio of retention factor (k) in MEKC for two charged analytes, respectively. Using four alkylparabens as test analytes, excellent agreement was found between the observed α(MEKC) and the proposed α(MEKC) models of test analytes in MEKC over a wide range of SDS concentrations and values of k. For example, in comparison with CZE separation of charged analytes, MEKC separation can enhance separation selectivity up to the maximum value when the selectivity ratio (ρ) is greater than 1.0 (ρ=α(k)/α(CZE)), while lower separation selectivity is obtained with ρ<1.0 (α(CZE) >α(k) >1).     

Pharmaceutical applications of micelles in chromatography and electrophoresis

This review surveys the use of micelles as separation media in chromatography and electrophoresis. Applications to pharmaceuticals whose molecular masses are relatively small are focused on in this review. In high-performance liquid chromatography (HPLC), chromatography using micelles and reversed-phase stationary phases such as octadecylsilylized silica gel (ODS) columns is known as micellar liquid chromatography (MLC). The main application of MLC to pharmaceutical analysis is the same as in ion-pair chromatography using alkylsulfonate or tetraalkylammonium. In most cases, selectivity is much improved compared with other short alkyl chain ion-pairing agents such as pentanesulfonate or octanesulfonate. Direct plasma/serum injection can be successful in MLC. Separation of small ions is also successful by using gel filtration columns and micellar solutions. In electrophoresis, especially capillary electrophoresis (CE), micelles are used as pseudo-stationary phases in capillary zone electrophoresis (CZE). This mode is called micellar electrokinetic chromatography (MEKC). Most of the drug analysis can be performed by using the MEKC mode because of its wide applicability. Enantiomer separation, separation of amino acids and closely related peptides, separation of very complex mixtures, determination of drugs in biological samples etc. as well as separation of electrically neutral drugs can be successfully achieved by MEKC. Microemulsion electrokinetic chromatography (MEEKC), in which surfactants are also used in forming the microemulsion, is successful for the separation of electrically neutral drugs as in MEKC. This review mainly describes the typical applications of MLC and MEKC for the analysis of pharmaceuticals.     

Improvement of the sensitivity of 2D LC-MEKC separation of phenolic acids and flavonoids natural antioxidants using the on-line preconcentration step

A 2D method was developed for separation of phenolic acids and flavonoids natural antioxidants combining LC with MEKC. The in-capillary preconcentration step was applied for the improvement of the sensitivity of 2D method before the second dimension MEKC analysis. The influence of first dimension LC mobile phase composition on migration times in the second MEKC dimension was evaluated. When gradient elution is applied in the first dimension of 2D LC-MEKC system, increasing concentration of organic solvent in the mobile phase and in fractions transferred from LC influences the electroosmotic flow, partitioning equilibria of samples in micelles and properties of the micelles, which results in shifts of migration times during the consecutive runs in the second MEKC separation dimension. The shifts of migration times caused by the influence of increasing concentration of ACN on MEKC separation in second dimension of 2D LC-MEKC system were compensated by aligning the time axis using electroosmotic flow and micellar marker migration times. The optimized LC-MEKC method was applied on the separation of natural antioxidants in the plant extracts samples.     

人工神经元网络法提高毛细管电泳药物控制分析准确度的研究

研究了人工神经元网络法在毛细管电泳定量测定memantine中提高测定准确度 的可行性。在毛细管电泳法定量测定memantine的过程中，其浓度与峰高或峰面积 以及与二者和内标的比值均没有良好的线性关系。人工神经元网络具有很强的非线 性校正能力，其最大优点是无须对分离体系及组分的迁移行为预先予以了解。人工 神经元网络的输为memantine的峰高和峰面积，输出为memantine的浓度。通过实验 确定的网络结构为2：1：1型。由于人工神经元网络的通用性，该法也可用于毛细 管电泳在其他药物控制分析中改善定量分析的准确度。     

Comparison of microemulsion electrokinetic chromatography and micellar electrokinetic chromatography as methods for the analysis of ten benzophenones

Microemulsion electrokinetic chromatography (MEEKC) and micellar electrokinetic chromatograpy (MEKC) were compared for their abilities to separate and detect ten similar benzophenones, which are commonly used as UV filters in various plastic and cosmetic products. Sodium dodecyl sulfate (SDS) concentration and column temperature rarely affected separation resolution for MEEKC, but separation of benzophenones could be improved by changing the SDS concentration and column temperature for MEKC. Buffer pH and ethanol (organic modifier) were found to markedly influence the separation selectivity for both MEEKC and MEKC systems. In addition, a higher electric voltage improved the separation efficiency without a noticeable reduction in separation resolution for MEEKC, whereas it caused a poor separation resolution for the MEKC system.     

Stacking and separation of enantiomers by acetonitrile-salt mixtures in micellar electrokinetic chromatography

The feasibility of employing the "acetonitrile stacking" method in micellar electrokinetic chromatography (MEKC) for the on-line preconcentration and separation of enantiomers is demonstrated for the first time. The effects of various experimental parameters on the stacking and separation of three different pairs of optical isomers, i.e., two substituted naphthyl enantiomers and one dansylated-DL-amino acid, were examined. In particular, the effectiveness of the addition of acetonitrile and salt in the sample matrix to induce narrowing of the analyte bands was investigated in the presence of sodium cholate as the chiral surfactant micelle in the separation buffer. For example, it was found that the presence of both acetonitrile and 1% NaCl in the sample matrix (volume ratio = 2:1) led to a significant improvement of the peak height and resolution for the MEKC separation of a pair of R(-)/S(+)-1,1'-binaphthyl diyl hydrogen phosphate enantiomers when the injection sample size was relatively large (e.g., 12% capillary volume). Furthermore, the feasibility of combining salting-out solvent extraction (off-line) and acetonitrile stacking (on-line) as a novel approach for sample preconcentration in capillary electrophoresis was also demonstrated.     

Multivariate approach for the enantioselective analysis in MEKC‐MS: II. Optimization of 1,1′‐binaphthyl‐2,2′‐diamine in positive ion mode

Enantiomeric separation and detection of 1,1′‐binaphthyl‐2,2′‐diamine (BNA) has been successfully optimized by MEKC‐ESI‐MS using a polymeric surfactant polysodium N‐undecenoxycarbonyl‐L‐leucinate (poly‐L‐SUCL) as a pseudostationary phase. In the first step, MEKC conditions were optimized by a five‐factor three‐level central composite design (CCD) of experiment. All five MEKC factors (buffer pH, percentage of ACN in the running buffer, concentration of surfactant, concentration of ammonium acetate (NH₄OAc), and voltage) were found significant to the responses (measured as the chiral resolution and analysis time). The interactions between MEKC factors were further evaluated using a quadratic model equation which allowed the generation of 3‐D response surface image to reach the optimum conditions. To obtain the best S/N, sheath liquid composition and spray chamber parameters were successfully optimized using the same strategy. Baseline enantiomeric resolution in less than 20 min and optimum MS signal of BNA enantiomers (S/N = 45 at 0.4 mg/mL) were ultimately achieved at the optimized conditions. The adequacy of the model was validated by experimental runs at the optimal predicted conditions. The predicted results were found to be in good agreement with the experimental data.