Assessment of techniques for DOSY NMR data processing

Diffusion-ordered spectroscopy (DOSY) NMR is based on a pulse-field gradient spin-echo NMR experiment, in which components experience diffusion. Consequently, the signal of each component decays with different diffusion rates as the gradient strength increases, constructing a bilinear NMR data set of a mixture. By calculating the diffusion coefficient for each component, it is possible to obtain a two-dimensional NMR spectrum: one dimension is for the conventional chemical shift and the other for the diffusion coefficient. The most interesting point is that this two-dimensional NMR allows non-invasive “chromatography” to obtain the pure spectrum for each component, providing a possible alternative for LC-NMR that is more expensive and time-consuming. Potential applications of DOSY NMR include identification of the components and impurities in complex mixtures, such as body fluids, or reaction mixtures, and technical or commercial products, e.g. comprising polymers or surfactants.

Data processing is the most important step to interpret DOSY NMR. Single channel methods and multivariate methods have been proposed for the data processing but all of them have difficulties when applied to real-world cases. The big challenge appears when dealing with more complex samples, e.g. components with small differences in diffusion coefficients, or severely overlapping in the chemical shift dimension. Two single channel methods, including SPLMOD and continuous diffusion coefficient (CONTIN), and two multivariate methods, called direct exponential curve resolution algorithm (DECRA) and multivariate curve resolution (MCR), are critically evaluated by simulated and real DOSY data sets. The assessments in this paper indicate the possible improvement of the DOSY data processing by applying iterative principal component analysis (IPCA) followed by MCR-alternating least square (MCR-ALS).     

Characterization of the measurement error structure in 1D H NMR data for metabolomics studies

NMR-based metabolomics is characterized by high throughput measurements of the signal intensities of complex mixtures of metabolites in biological samples by assaying, typically, bio-fluids or tissue homogenates. The ultimate goal is to obtain relevant biological information regarding the dissimilarity in patho-physiological conditions that the samples experience. For a long time now, this information has been obtained through the analysis of measured NMR signals via multivariate statistics.NMR data are quite complex and the use of such multivariate statistical methods as principal components analysis (PCA) for their analysis assumes that the data are multivariate normal with errors that are identical, independent and normally distributed (i.e. iid normal). There is a consensus that these assumptions are not always true for these data and, thus, several methods have been devised to transform the data or weight them prior to analysis by PCA. The structure of NMR measurement noise, or the extent to which violations of error homoscedasticity affect PCA results have neither been characterized nor investigated.A comprehensive characterization of measurement uncertainties in NMR based metabolomics was achieved in this work using an experiment designed to capture contributions of several sources of error to the total variance in the measurements. The noise structure was found to be heteroscedastic and highly correlated with spectral characteristics that are similar to the mean of the spectra and their standard deviation. A model was subsequently developed that potentially allows errors in NMR measurements to be accurately estimated without the need for extensive replication.     

Peak alignment of NMR signals by means of a genetic algorithm

Nuclear magnetic resonance (NMR) analysis of complex samples, such as biofluid samples is accompanied by variations in peak position and peak shape not directly related to the sample. This is due to variations in the background matrix of the sample and to instrumental instabilities. These variations complicate and limit the interpretation and analysis of NMR data by multivariate methods. Alignment of the NMR signals may circumvent these limitations and is an important preprocessing step prior to multivariate analysis. Previous aligning methods reduce the spectral resolution, are very computer-intensive for this kind of data (65k data points in one spectrum), or rely on peak detection. The method presented in this work requires neither data reduction nor preprocessing, e.g. peak detection. The alignment is achieved by taking each segment of the spectrum individually, shifting it sidewise, and linearly interpolating it to stretch or shrink until the best correlation with a corresponding reference spectrum segment is obtained. The segments are automatically picked out with a routine, which avoids cutting in a peak, and the optimization process is accomplished by means of a genetic algorithm (GA). The peak alignment routine is applied to NMR metabonomic data.¹     

Application of an extended DLVO theory for the calculation of the interactions between emulsified oil droplets in alcohol solutions

The stability of diluted emulsions (0.1% v/v) of n-dodecane in 1 M methanol, ethanol or propanol was studied. The effective diameter and zeta potential were determined by dynamic light scattering. The parameters were measured 5, 30, 60, 120 min and after 1 day after preparation of the emulsions by mechanical mixing at 10 000 r.p.m. for 3 min. Calculations of the free energy interactions between dodecane droplets were conducted applying van Oss et al.’s extended DLVO theory, in which acid–base interactions involving electron donor and electron acceptor parameters are also accounted for. For this purpose the interfacial tensions of oil–alcohol solutions were taken from the literature. The acid–base interactions were evaluated considering two variants. In the first we assumed a close-packed monolayer of alcohol molecules on the droplet surface, interacting by hydrogen bonds with water as well with alcohol molecules. In the second variant, it was considered that in these electrolyte-free systems (pH close to neutral) the measured zeta potentials were due to the oriented alcohol dipoles on the droplet surface. This would mean that the slipping plane is very close to the droplet surface. Both variants lead to the same conclusion that in these system the dominant role is played by attractive acid–base interactions, which is much bigger than the equally attractive apolar Lifshitz–van der Waals interaction. Repulsive electrostatic interactions play only a minor role.     

Phase transitions and azeotropic properties of acetic acid–n-propanol–water–n-propyl acetate system

The brief review of the data on VLE and LLE in acetic acid–n-propanol–water–n-propyl acetate system is presented. The azeotropic properties and the topological structure of the residue curve map at 313.15 K are discussed. This system is one of the few reacting systems with an extensive set of data on binary and ternary subsystems, in chemically nonequilibrium states. The main aim of the paper is to present the set of combined data that could be helpful for the development of thermodynamics of the systems with chemical reactions, and for modeling of coupled phase and reactive equilibria.     

Structural evolution of calcium-exchanged (NH4)2SiF6-dealuminated Y zeolite after various chemical treatments

The structural evolution of Y zeolite (Si/Al 2.17) weakly dealuminated by hexafluorosilicate (Si/Al 3.13), denoted Y_D, and exchanged with calcium (CaY_D), has been studied after acid–base treatments at 80 °C close to the cation exchange conditions. The stability of the samples was followed by X-ray diffraction and solid-state NMR of ²⁹Si and ²⁷Al; Y_D zeolite was completely destroyed by treatment with acid pH 2.5 and suffered serious degradation on treatment with alkali at pH 11.8. The introduction of calcium improved the stability of the zeolite in acid and base. In acid CaY_D was not destroyed until pH 1. At pH 2, silicon and aluminium were extracted and an amorphous phase was formed. Base treatment at pH 13 did not affect the calcium-exchanged zeolite.     

Role of acid–base interactions on the adhesion of oral streptococci and actinomyces to hexadecane and chloroform—influence of divalent cations and comparison between free energies of partitioning and free energies obtained by extended DLVO analysis

Calcium is an abundantly present, divalent cation in the oral cavity and plays a crucial role in the adhesion of oral microorganisms to tooth surfaces as well as in coaggregation and coadhesion among the oral microflora. The aim of this study was to determine the effects of divalent cation (Ca²⁺, Mg²⁺, Ba²⁺) adsorption on the adhesion of two actinomyces and two streptococcal strains to hexadecane (MATH) and chloroform (MATS) in order to detect changes in acid–base character of the cell surfaces. Initial removal rates of the organisms by hexadecane, lacking an acid–base interaction with the organisms, were always smaller than those by chloroform. Furthermore, adsorption of divalent cations generally increased the initial removal rates of the microorganisms, but no statistically significant differences among different cations were observed. Gibbs energies of partitioning calculated from the stationary end-point adhesion of the organisms ranged from −2 to −4 kT for adhesion to hexadecane and were about twofold more negative for adhesion to chloroform. Contact angles on lawns of microorganisms with and without adsorbed divalent cations were similar. Zeta potentials of all microorganisms were slightly negative under the conditions of MATH and MATS and became only 4 mV more positive upon divalent cation adsorption. Hexadecane had a zeta potentials of −21 mV in the potassium phosphate solution used, which became 13 mV less negative upon Ca²⁺ adsorption. An extended DLVO approach of microbial adhesion to hexadecane, based on microbial contact angles and zeta potentials, taking into account Lifshitz–van der Waals, acid–base and electrostatic interactions did not show any potential energy barrier and demonstrated a deep primary interaction minimum at close approach due to acid–base attraction. As the Gibbs energy of partioning was only −2 to −4 kT, it is concluded that for the collection of organisms studied here, the final contactable surface area is small and structural features on the cell surfaces like fibrils and fimbriae, maintain a distance of ca. 10–15 nm between the hexadecane and the overall cell surface and therewith prevent acid–base interactions to become operative to a significant extend. Furthermore, from the lack of influence of divalent cations on macroscopic cell surface contact angles and zeta potentials, it is suggested that cation adsorption is minor and localized to the fibrils and fimbriae.     

Spectrophotometric determination of acidity and tautomeric constants and hydrogen bonding strength for a new Schiff base using hard modeling and multivariate curve resolution alternative least squares methods

The equilibria of a new Schiff base derived from 3,6-bis((aminoethyl)thio)pyridazine were studied spectrophotometrically with the aid of factor-analytical methods. Hard modeling program was used for determination of the acidity constants of the Schiff base in dimethylformamide (DMF)/water mixture (30:70 v/v). In this method acidity constant equations act as hard models and the score vectors obtained by decomposing of absorbance data matrix will be linear combinations of equilibrium concentrations of species that exist in the absorption matrix. Two rank annihilation factor analysis (TRAFA) was used as a standard method to investigate the accuracy of the method. The tautomerization constant, K_t, of the Schiff base solution in various DMF/water mixtures has also been determined using spectral variations of the Schiff base solutions in various volume ratios of water with the aid of evolving factor analysis (EFA) and multivariate curve resolution alternative least squares (MCR-ALS) methods. In addition the intramolecular hydrogen bonding strength and its related thermodynamic parameters have been determined using MCR-ALS and spectral variation of the Schiff base solutions in different temperatures.     

Comparison of methods for outlier detection and their effects on the classification results for a particular data base

One of the drawbacks for using linear discriminant analysis (LDA) is the presence of outliers. Some methods of detecting outliers are compared and applied to a particular data base. When multivariate methods (multinormal distribution procedure and Hawkins' procedure) were applied, the two subsets produced did not differ greatly. Assumptions needed for the application of LDA were evaluated for each subset. Classification ability, feature selection and prediction ability were considered for each subset. Results for each subset were quite different. Hawkins' procedure seems the better method for detecting outliers.     

The utility of Bacillus subtilis as a bioflocculant for fine coal

The application of Bacillus subtilis as a flocculant for fine coal has been reported here. Zeta-potential measurements showed that both the coal and bacteria had similar surface charge as a function of pH. Surface free energy calculations showed that the coal was hydrophobic while the bacterium was hydrophilic. The adhesion of the bacteria to coal and subsequent settling was studied in detail. Adhesion of bacteria to coal surface and subsequent settling of coal was found to be quick. Both adhesion and settling were found to be independent of pH, which makes the process very attractive for field applications. The presence of an electrolyte along with the bacterium was found to not only enhance adhesion of bacteria, but also produce a clear supernatant. Further, the settled fraction was more compact than with bacteria alone. Interaction energy calculations using the extended DLVO theory showed that the electrical forces along with the acid–base interaction energy play a dominant role in the lower pH range. Above pH 7, the acid–base interaction energy is the predominant attractive force and is sufficient enough to overcome the repulsive forces due to electrical charges to bring about adhesion and thus settling of fine coal. With increase in electrolyte concentration, the change in total interaction energy with pH is minimal which probably leads to better adhesion and hence settling.     

31P nuclear magnetic resonance titrations: Simultaneous evaluation of all pH-dependent resonance signals

The simultaneous evaluation of ali pH-dependent resonance signals (or multiplets) of an NMR titration offers a substantially increased accuracy and significance. The number of linearly independent titration equilibria is determined by graphical matrix rank analysis. The chemical shifts of all pH-dependent resonance lines are plotted against each other (chemical shift or CS diagrams) indicating whether a single or more titration equilibria are NMR spectrometrically observable and how far they overlap with each other. An iterative curve-fitting program allowing the simultaneous evaluation of all (pH) curves is available, from which pK values and chemical shifts of all species can be calculated. The starting pK values for the iteration need only be estimated very approximately (accuracy ±1–2 units). The titration end-points do not have to be experimentally accessible. The different methods for the simultaneous evaluation of all pH-dependent NMR signals are exemplified in the ³¹P NMR titration of thiamine pyrophosphate. In this case either the observed resonance lines (two doublets in a broad band proton decoupled spectrum) or the calculated chemical shifts for this AB system can be evaluated. A titration of sodium pyrophosphate was performed and evaluated for comparison.     

Sulfonated polysulfone with 1,3-1H-dibenzimidazole-benzene additive as a membrane for direct methanol fuel cells

1,3-1H-Dibenzimidazole-benzene (DBImBenzene) has been synthesized using phosphorus pentoxide-methanesulfonic acid (PPMA) as a solvent and dehydration agent and investigated as an additive (up to 2.0 wt.%) in sulfonated polysulfone (SPSf) membranes to promote proton conduction via acid–base interactions. The SPSf/DBImBenzene blend membranes with various DBImBenzene contents (0–2.0 wt.%) have been prepared and characterized by proton conductivity measurement and electrochemical polarization and methanol crossover measurements in direct methanol fuel cells (DMFCs). The blend membranes with DBImBenzene content of 0.5 and 1.0 wt.% show higher proton conductivities (3.4 and 2.9 × 10⁻⁴ S/cm, respectively) than plain SPSf (2.4 × 10⁻⁴ S/cm) even though the blend membranes have lower ion exchange capacity (0.81 and 0.75 mequiv./g, respectively) than plain SPSf (0.86 mequiv./g). The blend membranes exhibit better electrochemical performance in DMFC than plain SPSf membrane due to an enhancement in proton conductivity through acid–base interactions and lower methanol crossover.     

A solid-phase extraction procedure coupled to 1H NMR, with chemometric analysis, to seek reliable markers of the botanical origin of honey

The aim of this work was to establish an analytical method for identifying the botanical origin of honey, as an alternative to conventional melissopalynological, organoleptic and instrumental methods (gas-chromatography coupled to mass spectrometry (GC–MS), high-performance liquid chromatography HPLC). The procedure is based on the ¹H nuclear magnetic resonance (NMR) profile coupled, when necessary, with electrospray ionisation-mass spectrometry (ESI-MS) and two-dimensional NMR analyses of solid-phase extraction (SPE)-purified honey samples, followed by chemometric analyses. Extracts of 44 commercial Italian honeys from 20 different botanical sources were analyzed.Honeydew, chestnut and linden honeys showed constant, specific, well-resolved resonances, suitable for use as markers of origin. Honeydew honey contained the typical resonances of an aliphatic component, very likely deriving from the plant phloem sap or excreted into it by sap-sucking aphids. Chestnut honey contained the typical signals of kynurenic acid and some structurally related metabolite.In linden honey the ¹H NMR profile gave strong signals attributable to the mono-terpene derivative cyclohexa-1,3-diene-1-carboxylic acid (CDCA) and to its 1-O-β-gentiobiosyl ester (CDCA-GBE). These markers were not detectable in the other honeys, except for the less common nectar honey from rosa mosqueta. We compared and analyzed the data by multivariate techniques. Principal component analysis found different clusters of honeys based on the presence of these specific markers.The results, although obviously only preliminary, suggest that the ¹H NMR profile (with HPLC–MS analysis when necessary) can be used as a reference framework for identifying the botanical origin of honey.     

Tautomeric polymorphism in salicylideneamine derivatives: an X-ray diffraction and solid-state NMR study

The crystal structure of the N-(3-hydroxysalicylidene)-4-methoxyaniline has been studied by single-crystal X-ray diffraction and solid-state NMR spectroscopy. This is the first example of a Schiff base derived from 3-hydroxysalicylaldehyde which displays in the asymmetric unit, four distinct molecules linked together in the crystal lattice by two types of intermolecular O–HO hydrogen bonds and formed by two independent tetramers. The ¹³C CPMAS NMR study corroborates the above results; the presence of different tautomeric equilibria in the same crystal structure is demonstrated and a qualitative estimation of the equilibrium mixture composition is given.     

Methodology of chemometric modeling of spectrometric signals in the analysis of complex samples

A generalized algorithm of the multivariate simulation of spectrometric data is considered for solving typical analytical problems, like the determination of the concentration of a particular analyte and the assignment of a sample to one of predefined classes. In particular, we considered preliminary data processing, exploratory analysis, optimization of a chemometric model, calculation of performance characteristics, transfer of the model to other spectrometers, and automation of chemometric processing for the routine analysis of samples. To illustrate the potential of the method, we selected a system of bovine and porcine heparin, mixtures of soy and sunflower lecithin, and a set of red and white wine samples as test samples. Partial least squares and discriminant analysis were used as chemometric methods. We used proton nuclear magnetic resonance (1H NMR) to record signals. Using the MATLAB environment, chemometric programs were developed for automated data processing in the context of problems under consideration and for the transfer of multivariate models to other spectrometers. Based on the results obtained, a methodology is proposed for the multivariate analysis of spectrometric data, which can be used in the analysis of various types of matrices and spectrometric signals.     

Independent component analysis (ICA) algorithms for improved spectral deconvolution of overlapped signals in 1H NMR analysis: application to foods and related products

The major challenge facing NMR spectroscopic mixture analysis is the overlapping of signals and the arising impossibility to easily recover the structures for identification of the individual components and to integrate separated signals for quantification. In this paper, various independent component analysis (ICA) algorithms [mutual information least dependent component analysis (MILCA); stochastic non‐negative ICA (SNICA); joint approximate diagonalization of eigenmatrices (JADE); and robust, accurate, direct ICA algorithm (RADICAL)] as well as deconvolution methods [simple‐to‐use‐interactive self‐modeling mixture analysis (SIMPLISMA) and multivariate curve resolution‐alternating least squares (MCR‐ALS)] are applied for simultaneous ¹H NMR spectroscopic determination of organic substances in complex mixtures. Among others, we studied constituents of the following matrices: honey, soft drinks, and liquids used in electronic cigarettes. Good quality spectral resolution of up to eight‐component mixtures was achieved (correlation coefficients between resolved and experimental spectra were not less than 0.90). In general, the relative errors in the recovered concentrations were below 12%. SIMPLISMA and MILCA algorithms were found to be preferable for NMR spectra deconvolution and showed similar performance. The proposed method was used for analysis of authentic samples. The resolved ICA concentrations match well with the results of reference gas chromatography–mass spectrometry as well as the MCR‐ALS algorithm used for comparison. ICA deconvolution considerably improves the application range of direct NMR spectroscopy for analysis of complex mixtures. Copyright © 2014 John Wiley & Sons, Ltd.     

High resolution magic angle spinning NMR for analyzing small molecules attached to solid support

Solid phase synthesis has become a routine technique in combinatorial chemistry. The need in analytical methods to characterize nondestructively resin bound molecules has been fulfilled by the introduction of High Resolution Magic Angle Spinning (HR MAS) NMR of solvent swollen beads. HR MAS NMR can give solution like proton NMR spectra and one- and two-dimensional NMR techniques are amenable, allowing detailed structure analysis. Recent developments are the application of a diffusion filter to suppress solvent signals and dipolar recoupling techniques to gain spatial information. HR MAS NMR has been applied to monitor reactions and elucidate reaction products.     

Application of H-point standard additions method to spectrophotometric and spectrofluorimetric determinations of glafenine and glafenic acid in mixtures

H-point standard additions method (HPSAM), based on spectrophotometric and spectrofluorimetric measurements, was proposed for simultaneous determination of glafenine (G) and glafenic acid (GA). A study of the absorption spectra of G and GA in various pH media has been carried out. Reasonably resolved UV-absorption spectra were obtained with a solution adjusted at pH 4.5 with citric acid–phosphate buffer. Additionally, the fluorescence properties in aqueous micellar systems of anionic, cationic and non-ionic surfactants were investigated. Well resolved fluorescence spectra were established in aqueous Triton X-100 solution at pH 7.8 (citric acid–phosphate buffer). As a comparative method, UV-derivative spectrophotometry (based on zero-crossing technique) was suggested. First-derivative value at 352 nm (¹D₃₅₂) and second-derivative value at 366 nm (²D₃₆₆) were selected for the quantification of G and GA, respectively. The relative standard deviations of the proposed methods approximate 2%. The proposed methods were evaluated through the analysis of commercial tablets. The results were accurate and precise.     

Chemical and physicochemical characteristics of humic acids extracted from compost, soil and amended soil

In order to gain understanding about how “HA-like substances” from organic amendments may change some properties in the soil solution, the knowledge of chemical and physicochemical characteristics (charge development, acid–base behavior and heterogeneity) should be known.

The aim of this research were (i) to study the elemental and functional composition, (ii) to determine charge behavior, acid–base properties (apparent dissociation constant and buffer capacity) and (iii) to evaluate heterogeneity of humic acids (HA) isolated from municipal solid waste compost (MWC) and from the corresponding MWC-amended soil, in comparison to those of the unamended soil HA using potentiometric titration and differential scanning potentiometry (DSP).

Potentiometric titration and the first derivative of −Q versus pH (negative charge development versus pH) curves could be used to determine proton-affinity distribution and the chemical heterogeneity of the HA as well as the average pK_app and buffer capacity in a wide range of pH.

Differential scanning potentiometry allows determination of the pK_app values in a narrower range of pH than potentiometric titrations and is another simple methodology to study acid–base behavior of HA.

DSP allows us to determine seven different pK_app values for HA-S and HA-E and four different pK_app for HA-C. Each one of these values corresponds to known acidic groups that can be present in the macromolecule of HA.     

Evaluation of complex spectral-pH three-way arrays by modified bilinear least-squares: determination of four different dyes in interfering systems

This article reports on the first application of a modified version of the bilinear least-squares model to absorbance-pH second-order data recorded for complex samples. The latter are composed of fruit drink powders, where four different analytes and additional background components occur. The analytes are the common juice colorants tartrazine, yellow sunset, allura red and indigo carmine. The data have been measured after generating a double pH gradient within a flow injection system. The selected chemometric methodology adequately exploits the second-order advantage, needed to take into account the background interferents present in real samples. Due to severe spectral overlapping between the acid and basic forms of each of the colorants in the working pH range, other second-order multivariate calibration methods such as parallel factor analysis and multivariate curve resolution-alternating least-squares could not be successfully applied to the presently studied samples. Recoveries of 94.8, 104.7, 109.3 and 105.3% were obtained for yellow sunset, indigo carmine, allura red and tartrazine respectively in the real test samples.