Search

accuracy of forecasts is influenced by both the quality of past data and the method selected to forecast the future. This paper shows a method to accurately predict the time series signal through a combination of decomposition methods and Echo State Network (ESN). Wavelet and Multiwavelet transforms are used to decompose highly nonlinear time series into several stationary time series components. Thereby, they are used to reduce the degree of nonlinear time series and make the issue easy to analyze and predict. These components are fed to an ESN, which predicts the signal. As an illustration for proposed pattern, one of time series signals of the neural network competition (NNC 2010) is used, without knowing its properties. The performances of all the methods used in this work have been evaluated by computer using MATLAB 7.9.0.287 (R2009b) language and RCToolbox version 2.1. Finally, comparison between these two transforms was done in terms of mean square error (MSE). The simulation results showed the effectiveness and significant improvement of the MWT-ESN model compared with DWT-ESN.

No Abstract.

The common analyte internal standardization (CAIS) chemometric technique is extended to correct for drift in signal intensity in inductively coupled plasma-mass spectrometry. The CAIS technique is used because, unlike the conventional internal reference method, it allows for the analyte to behave differently from the internal reference under the influence of drift. Thus, only one internal reference element is sufficient to correct for drift effect for all the analytes irrespective of the difference of their mass and ionization potential from that of the internal reference element. Experimental test with 15 analytes, as a representative of the whole mass range, in different geological matrices, using only indium as internal reference element, demonstrates that the developed drift correction method is efficient in correcting up to 25% drift error and is easy to use. Further, a CAIS scheme was developed to correct for both drift and non-spectroscopic matrix effects simultaneously in a single run, using standard containing no matrix. This scheme is based on the use of two internal reference elements for simultaneous monitoring and correction of drift and matrix effects for all the analytes irrespective of their difference in mass and ionization potential from the two internal reference elements. The developed scheme was validated using river water samples containing appreciable matrix concentration and was measured under the influence of drift. The results of the experimental validation indicate that the proposed scheme is capable of removing error as large as 43% that arises from drift and non-spectroscopic matrix effects.

The Istituto di Geoscienze e Georisorse (IGG), on behalf and with the support of the International Atomic Energy Agency (IAEA), prepared eight geological materials (three natural waters and five rocks and minerals), intended for a blind interlaboratory comparison of measurements of boron isotopic composition and concentration. The materials were distributed to twenty seven laboratories - virtually all those performing geochemical boron isotope analyses in the world -which agreed to participate in the intercomparison exercise. Only fifteen laboratories, however, ultimately submitted the isotopic and/or concentration results they obtained on the intercomparison materials. The results demonstrate that interlaboratory reproducibility is not well reflected by the precision values reported by the individual laboratories and this observation holds true for both boron concentration and isotopic composition. The reasons for the discrepancies include fractionations due to the chemical matrix of materials, relative shift of the zero position on the δ11B scale and a lack of well characterized materials for calibrating absolute boron content measurements. The intercomparison materials are now available at the IAEA (solid materials) and IGG (waters) for future distribution. L'Istituto di Geoscienze e Georisorse (IGG), pour le compte et avec le soutien de l'Agence Internationale de l'Energie Atomique (AIEA), a prepare huit materiaux geologiques (trois eaux naturelles et cinq roches et mineraux) destines a la conduite d'une comparaison en aveugle des mesures de composition isotopique et de concentration du bore entre differents laboratoires. Vingt-sept laboratoires -pratiquement tous ceux qui travaillent sur les isotopes du bore en geochimie - ont accepte de participer a cette intercomparaison, mais quinze seulement ont finalement envoye leurs donnees de composition isotopique et/ou de concentration du bore. Les resultats demontrent que la precision des valeurs indiquee par les laboratoires reflete mal la reproductibilite entre laboratoires des mesures de composition isotopique comme de concentration du bore. Les raisons de ces discordances peuvent comprendre l'existence de fractionnements isotopiques dus aux matrices chimiques des materiaux, un deplacement relatif du zero de l'echelle de δ11B, ainsi que le manque de materiaux bien caracterises pour calibrer les mesures de concentration du bore. Les materiaux utilises pour l'intercomparaison des mesures sont maintenant a disposition a l'AIEA (materiaux solides) et l'IGG (eaux) pour etre distribues.

An on-channel sample preconcentration-matrix removal arrangement, based on coupling field-flow fractionation (FFF) to inductively coupled plasma mass spectrometry (ICP-MS), has been constructed for on-line sample pretreatment ICP-MS trace element determination. A commercial FFF system is modified to incorporate an on-channel preconcentration procedure allowing injection of up to 50 ml of sample, which could be preconcentrated by 50–1400 fold. A high molecular weight complexing agent added to the sample forms strong complexes with the measured trace analytes but not with the sample matrix. When the sample-complexing agent mixture is introduced to the FFF unit, the uncomplexed matrix element is removed by permeation through a membrane that separates the FFF sample compartment. The trace analytes remain in the FFF channel, because their high molecular weight complexes do not permeate through the membrane. Preconcentration and matrix elimination occur simultaneously. The matrix-free, preconcentrated sample is introduced directly to the ICP-MS nebulizer. The method was tested using 10-ml sample aliquots that contain As, Cd, Cu, Mo, Pb, Re, Sn, Te, Tl, Y, Zn and Zr analytes and 5000 mg l−1 Ca or Na matrices and ethylene imine polymer complexing agent. Copper and Re isotopic ratio values in reference standards also were determined after preconcentration and matrix element removal.

Thorium and iodine memory effects have been characterized experimentally for inductively-coupled plasma mass spectrometry by adding ammonia gas directly to the spray chamber and nebulizing aqueous ammonia sample solutions to assess analyte memory retention sites. Thorium memory effect originates from the tendency of an unidentified thorium compound to volatilize from the spray-chamber walls, and not from Th compound adsorption to nebulizer tubing. The mass spectrometer skimmer and sampler cones, ion optics, quadrupole, and other components are not responsible for the memory effect. Unlike that of thorium the iodine memory effect originates from adsorption of iodine compounds on nebulizer tubing surfaces and from volatilization of HI and I2 from the spray-chamber walls. Addition of ammonia sample solutions or ammonia gas directly to the spray chamber eliminated the Th and I memory effects in practical analyses, and blank levels were achieved after 2 min wash-outs. Quantitative recoveries were obtained for Th and I in reference materials.

The common analyte internal standardization (CAIS) chemometric technique is extended to improve isotope ratio precision obtained by inductively coupled plasma-quadrupole mass spectrometry (ICP-QMS). A correction function for isotope ratio measurement is derived from a relationship of mass sensitivity as a function of element mass after considering the degree of ionization and isotopic abundance. An analyte isotope ratio correction factor can be determined by simultaneously measuring several internal reference isotopic ratios and the analyte. The CAIS approach corrects for isotopic mass bias and isotope ratio variation with sample matrix concentration. Isotope ratio precision is improved from about 0.4 to 0.05%, because long-counting times allow correcting mass bias and isotope ratio drift during extended data-acquisition periods. Isotope ratio accuracy is improved from 0.2 to 0.02%.

A method was developed to improve the precision of inductively coupled plasma quadrupole mass spectrometry (ICP-QMS) for the determination of boron isotope ratios (11B/10B) in various environmental materials including seawater. This approach is based on the common analyte internal standardization (CAIS) chemometric algorithm. The sample solution obtained after digestion is spiked with lithium, and both 7Li/6Li and 11B/10B values are measured using long-counting periods (20 min). The CAIS algorithm corrects the measured 11B/10B values for (a) statistical fluctuations resulting from short-term noise; (b) drift in 11B-to-10B ratio as a result of long-term deviation in instrumental parameters likely to occur during long counting times; (c) change in 11B-to-10B ratio caused by variation in matrix elements concentrations; and (d) drift in mass bias correction factor. Comparing boron isotopic ratios in seawater measured by conventional and the new isotope ratio methods validates the procedure. A synthetic isotopic mixture of boron SRM 951 and enriched 10B SRM 952 also was examined. The CAIS method provided a measured boron isotopic ratio precision of 0.05% R.S.D. while eliminating 5.1% matrix concentration error and 0.25% instrumental drift error.

Injection of 10–20 ml/min of ammonia gas into an inductively coupled plasma-mass spectrometry (ICP-MS) spray chamber during boron determination eliminates the memory effect of a 1 μg/ml B solution within a 2-min washing time. Ammonia gas injection also reduces the boron blank by a factor of four and enhances the sensitivity by 33–90%. Boron detection limits are improved from 12 and 14 to 3 and 4 ng/ml, respectively, for two ICP-MS instruments. Trace boron concentrations in certified reference materials agree well using ammonia gas injection.

Two sample treatment methods are evaluated to provide accurate boron determination at low concentrations in biological and botanical samples. The first approach is a hot 1M nitric acid extraction of boron from the sample. The second technique uses wet digestion with concentrated sulfuric acid. The accuracy of the procedures was demonstrated with botanical and animal reference materials (Corn Bran RM 8433 and Whole Egg Powder RM 8415). At least three results are given for each reference material. Two are direct measurements of B using independent 10B and 11B calibration curves with a Be internal reference, and the third is obtained by isotope dilution mass spectrometry (IDMS). The 10B and 11B values are consistent for both acid treatment procedures. The IDMS results also are consistent. Overall results for Whole Egg Powder and Corn Bran RM's match the best-estimate values within their confidence intervals. These results demonstrate the ability to measure B accurately at the 0.3 µg/g concentration range. Thus, low-level B samples can be analyzed with accuracy and precision by the two approaches. These methods introduce very little dissolved organic carbon (DOC) in the final solution and allow the use of large (2 g) sample aliquots. Direct introduction of biological fluids including whole blood serum also was evaluated critically for the determination of B.

A novel supervised classification algorithm, direct clustering in n-dimensional space (DCNS), was developed for difficult data sets where conventional methods of supervised clustering are expected to fail. The method is based, when applied on >3-dimensional spaces, on an algorithm that performs special treatment on the measurement space, so that the treated space can allow a computer-aided clustering methodology similar to that used by human vision. However, unlike other techniques that reduce the dimensionality of the space, the proposed method preserves the original dimensions while performing a computer-simulated human vision clustering in the original n-dimensional space. Thus the overlap between clusters that results from the dimensionality reduction is eliminated. The proposed method was applied to two real data sets. The results are compared with those obtained using principal component analysis (PCA), an artificial neural network (ANN), and the k-nearest-neighbor (KNN) technique. On one data set containing only two clusters, the DCNS algorithm gives better cluster separation than the other three methods. However, when all four methods were applied on the second data set, containing eight different clusters, PCA, ANN and KNN were unable to give useful cluster separation, while the DCNS method was able to separate all clusters and classify the unknown points successfully with their corresponding clusters. The DCNS technique is able to perform other important cluster analysis tasks, such as testing the discriminatory power of a variable, selecting one variable from many, and conducting preliminary unsupervised clustering.

The mechanism by which carbon matrix species cause non-spectroscopic matrix interference on boron (B) and beryllium (Be) during the determination of B in biological samples is investigated. The study indicates that, unlike other matrix species such as Ca and Na that cause non-spectroscopic matrix interference only through space charge effect, carbon species manifest non-spectroscopic interference by two mechanisms. The minor non-spectroscopic interference of carbon is through space charge effect. However, the major non-spectroscopic effect of carbon is by a charge transfer mechanism from C + -species to B and Be atoms in the central channel of the plasma discharge. The large difference in the magnitude of the carbon charge transfer non-spectroscopic matrix interference between Be and B makes Be unsuitable as an internal reference for B in solutions containing more than 1500 μg/ml dissolved organic carbon (DOC). This DOC content is approximately half the DOC usually present in the final sample solution for B determination in biological samples. However, Be still acts as a perfect internal reference to B in solution containing matrix elements that exert their non-spectroscopic interference effect through space charge mechanism (such as Na, K, Ca, etc.).

The Common Analyte Internal Standardization (CAIS) chemometric technique is extended to correct for non-spectroscopic matrix effects in inductively coupled plasma-mass spectrometry (ICP-MS). The approach is based on using an internal reference element to correct for the matrix effect. Unlike the conventional internal reference method, the CAIS technique allows for the analyte to behave differently from the internal reference under the influence of the matrix. With the CAIS technique a single internal reference element is sufficient to correct for all the analytes. Experimental tests with 13 analytes in four different matrices using different ICP-MS instruments demonstrate that the CAIS is efficient and general for matrix effect correction. Not only is the corrected concentration more accurate, but the precision is significantly better. The capability of CAIS to correct for the effect of a mixture of two matrices was also established experimentally, and 20–30% matrix suppression was eliminated. Furthermore, the developed technique was used as a simple diagnostic quality assurance procedure to evaluate the performance of the mass spectrometer.

Boron memory effect in inductively coupled plasma-mass spectrometry is characterized theoretically and verified experimentally. The boron memory effect originates from the tendency of boron to volatilize as boric acid from the sample solution layer that covers the inside surface of the spray chamber. The skimmer, the sampler, the ion lenses, the quadrupole and other components of the mass spectrometer are not responsible for the memory effect at trace boron concentration levels. A technique to eliminate the memory effect is proposed and tested. Ammonia gas is used to convert the boron from the volatile boric acid form to the non-volatile ammonium borate. Ammonia is introduced by injecting a small volume of ammonia solution simultaneously with the sample solution. Addition of ammonia is effective in completely eliminating the boron memory effect and reducing boron signals to blank levels instantaneously (within 20 s) in a conventional concentric nebulizer-double pass spray chamber arrangement.

Inelastic collisional deactivation of the analyte excited state is demonstrated as a dominant cause for non-spectroscopic matrix interference in inductively coupled plasma atomic emission spectrometry (ICP-AES) for commonly used plasma operating conditions in routine analysis. A mathematical simulation of the inelastic collisional model was examined. Comparison between the theoretical model and experimental results using atomic and ionic lines of the analytes Zn, Ba, Mg, Mn and Sr validates the inelastic collisional deactivation model as a dominant cause for non-spectroscopic matrix effect. Matrices evaluated were NH 4 Cl, NH 4 SCN, (NH 4 ) 2 SO 4 , and H 2 SO 4 to represent difficult-to-ionize matrices (DIE) and NaCl and CaCl 2 to represent easy-to-ionize element matrices (EIE).

A novel technique was developed to correct for the error in inductively coupled plasma atomic emission and mass spectrometric measurements arising from a difference in volatility between the sample analyte compounds and standards. The technique is based on the measurement of the analyte signal at two spray-chamber temperatures. A volatility correction factor is then estimated from a linear correlation between the reciprocal of a correction factor and the relative change in intensity resulting from measurements at two spray-chamber temperatures. Tests with organosilicon and organochlorine compounds demonstrate a significant decrease (from 2 to 30 times) in error after correction. The technique requires no prior knowledge of the chemical structure of the analyte.