One of the most serious problems that can occur when a multivariate model is used for the compositional analysis of an unknown mixture is the presence of an unexpectd constituent, not modeled in the calibration phase. The interferent will almost certajnly influence the predicted concentrations of the modeled constituents, which leads to erroneous and, more seriously, misleading results. Usually, a rdbration-building a new caliiration model in which the interferent is included+ be necessary. However, in many applications of muttiVariate calibration, recalibration will be possible only if an unambiguous identification of the interferent can be made. In this paper, we describe how spectral residuals resulting from a multivariate prediction can be used to detect and identify unknown interferents. The identification ofthe interferent is performed by matching the residual spectrum with a library of residual spectra. "his library was built by processing the members of a regular spectral library by the calibration model and storing the resulting residual spectra. After successful identification, a straightforward procedure can be used to correct the concentrations of the modeled constituents, without the need for a recalibration. l'he methods are demonstrated using a relatively simple principal component calibration model to predict the concentrations of organic vapors and gases in ambient air with ET-IR spectroscopy. In addition, the influence of Merent interferents on the predicted concentrations of the modeled constituents is described.Multivariate calibration can be regarded as a general tool to solve selectivity problems in multichannel instruments, e.g., overlapping spectra in mixture samples. In the last decade, the number of applications using multivariate calibration has grown rapidly, especially in the field of spectroscopy. An example of this growth is the use of near-infrared (near-IR) to predict the octane number of gasoline samples' or glucose in blood? Besides near-IR many applications exist in which multivariate calibration is used with Fourier transform infrared (FT-IR) and UV/vis spectroscopy. The use of multivariate calibration in spectroscopy (1) Kelly, J. J.; Barlow, C. H.; Jinguji, T. M.; Callis, J. B. Anal. Chem. 1989, 61, (2) Haaland, D. M.; Robinson, M. R; Koepp, G. W.; Thomas E. V.; Eaton, R P. 313.offers a combination of little or no sample preparation with highspeed quantitative analysis, enabling real-time monitoring of chemical processes. However, monitoring and controlling processes on the basis of the results of these measurements requires that the multivariate model gives precise and reliable predictions for future incoming samples. In addition, during the prediction of these unknown samples, procedures must be available that are able to detect the occurrence of outlying samples or, more generally, situations not appropriate for the model.Building a model in multivariate calibration consists of different steps. First, and probably most critical, is the collection of representative calibr...

No abstract

Thermal relaxation processes in high-pressure Hg and Na arc discharges have been studied by measuring the frequency dependence of the optogalvanic effect up to 5000 Hz. The experimental results are interpreted in terms of two linear relaxation processes, using a simple model for the energy flow in the arc. For the Hg arc (400 W dc) the two characteristic times are found to be 4.0×10−4 s and ?4×10−6 s. For the Na arc (180 W dc) only one characteristic time of 5.3×10−5 s is obtained.

Articles you may be interested inMass spectrometric ion analysis in the sputtering of oxide targets