Spectral Band Interferences

The detection of a specific gas (10) is accompHshed by comparing the signal of the detector that is constrained to the preselected spectral band pass with a reference detector having all conditions the same except that its preselected spectral band is not affected by the presence of the gas to be detected. Possible interference by other gases must be taken into account. It may be necessary to have multiple channels or spectral discrimination over an extended Spectral region to make identification highly probable. Except for covert surveillance most detection scenarios are highly controlled and identification is not too difficult. 

Previous experience in arc and spark emission spectroscopy has revealed numerous spectral overlap problems. Wavelength tables exist that tabulate spectral emission lines and relative intensities for the purpose of facilitating wavelength selection. Although the spectral interference information available from arc and spark spectroscopy is extremely useful, the information is not sufficient to avoid all ICP spectral interferences. ICP spectra differ from arc and spark emission spectra because the line intensities are not directly comparable. As of yet, there is no atlas of ICP emission line intensity data, that would facilitate line selection based upon element concentrations, intensity ratios and spectral band pass. This is indeed unfortunate because the ICP instrumentation is now capable of precise and easily duplicated intensity measurements. 

Sodium is still often determined by flame photometry, measuring the emission intensity of the doublet at around 589 nm, but care is necessary to make sure that excess calcium does not cause spectral interference (from molecular emission). This is unlikely to be a problem if AES is used, with a narrow spectral band-pass, and the intensity of emission at 589.0 nm from an air-acetylene flame is measured. However, at low determinant concentrations it is then advisable to add 2-5 mg ml 1 potassium or caesium as an ionization buffer. This is even more true if a nitrous oxide-acetylene flame is used for FES, although its use is rarely justified in environmental analyses because the additional sensitivity gained is rarely necessary. 

Under these conditions spectral interferences become predictable, as the wavelength ranges are well documented [10], in which the few diatomic molecules that have been observed in ET AAS exhibit their electron excitation spectra. An example of an interference-free spectral environment is shown in Figure 4.17 for the determination of Cr in whole egg powder at 357.869 nm. The second line within the spectral window at 358.119 nm is due to Fe, and obviously causes no interference in HR-CS AAS. It also causes no interference in LS AAS unless deuterium BC is used with a spectral band pass >0.5 nm. 

Adjust the instrumental parameters of wavelength, spectral band pass and lamp current in accordance with manufacturer s recommended conditions. A fuel rich nitrous oxide—acetylene flame should be used for the determination of vanadium and may be useful for other elements if interference effects are suspected. The air—acetylene flame may be used for the determination of Pb, Na, Ni, and Zn. For other elements consult the manufacturer s handbook. Establish the flame conditions while aspirating the blank solvent. It will normally be necessary to reduce the acetylene flow to compensate for the solvent contribution. In some cases the use of auxiliary air in the air-acetylene flame may prove advantageous. 

Fig. 2b is an idealized illustration of a single, uncomplicated Cotton effect. In reality, the occurrence of a complete curve in the electronic spectrum is rare. Complete dispersions are more likely to be observed in the vibrational spectral range because of the increased spectral resolution. However, even there, dispersions are too often complicated by extensive band overlap. The same is true for electronic spectra where hidden absorption bands coupled vibronic excitations and interferences from bands associated with other chiral chromophores contribute to producing anomalous ORD curves that are so complex they have little utility in quantitative analytical applications (Fig. 3). 

The equipment required to record the electron absorption at these low doses can be much simpler than was used in the present experiments. It is possible to use a 1 cm. absorption cell and a photocell instead of a photomultiplier if the monochromator is replaced by a combination of color or interference filters. This scheme uses a greater spectral band width and thus greatly improves the signal to the noise ratio. As the e aq absorption band is very broad, the increased band width will cause only a small change in the apparent e aq extinction coefficient, although the e aq absorbance would have to be recalibrated for the particular optical system used. With this type of system the major expense is in the oscilloscope and recording camera. 

Additional approaches using spectrophotometric detection techniques (UV and FL) in line with MS can help identify TPs in the environment. UV-spectra have been used for qualitative purposes for a long time. However, compared to other identification techniques (such as NMR and MS) the broad spectral bands and the relatively small frequency range in UV-spectrometry are not particularly informative. Beside compound identification, the UV trace can be used for quantification of interference-free analytes. The use of a diode array detector (DAD) allows one to collect full UV spectra of the analytes eluting from the LC column, and both identification and quantification can be done. DAD instruments are suitable as detectors in HPLC. DAD spectra are solvent-sensitive comparing spectra from reversed phase runs in acetonitrile to reference spectra collected in methanol may not match due to a l-2nm shift for some compounds [11]. Spectra collected during solvent gradients may also suffer from this problem, and an even worse problem oc-... 

Interference filters provide the best wavelength selectivity of any filters available. It is not possible to provide the necessary resolving power required for more complex spectral isolation. The filters therefore are primarily useful for simple systems where passage of a spectral band will meet spectral isolation requirements. 

The most common spectral band interference in qualitative spectral analysis is that produced by cyanogen. Cyanogen produces a number of bands, and three of them, with band heads at 4216.0, 3883.4, and 3590.4 A,... 

Copper electrodes are sometimes used if copper is not an interferent. Cyanogen bands are eliminated but spectral line sensitivity is also decreased. Graphite electrodes reach higher temperatures than copper electrodes during arcing and thus provide a higher excitation energy. 

Another type of spectral interference can occur between the analytical spectral line and spectral band systems produced by molecules or radicals in the flame. This type of interference is encountered more often than atomic line interference. The band emission may be from various metal oxides or hydroxides, or may be produced by species generated from the fuel and oxidant system. If organic solvents are employed, these substances also can contribute species that produce band spectra. Examples of this type of interference include the copper line at 3274.0 A and the OH band at 3274.2 A, and the aluminum line at 3961.5 A and the CH band series starting at 3872 A. 

Spectral band interference also can be troublesome. At low dispersion a spectral band appears as a single, broad band. With high dispersion the fine structure of the band appears. Frequently encountered bands in flame emission spectroscopy include those produced by OH, C2, CH, metal oxides, and metal hydroxides. 

Several approaches are possible to minimize the effects of spectral band-spectral line overlap. One method is to use a high resolution spectrometer with narrow slit widths. This method frequently resolves the band into its separate components, thus permitting better separation of spectral line and spectral band components. Another method is to determine if another line is available for use in a different spectral region. For example, there is less OH band interference with the copper 3274.0 A line than with the copper line at 3247.0 A. Cobalt at 2873.1 A is in a region of strong CH band interference, while the cobalt line at 3453.5 A is not. 

Spectral interferences occur whenever any radiation overlaps that of the analyte element. The interfering radiation may be an emission line of another element, radical, or molecule, unresolved band spectra, or general background radiation from the flame, solvent, or analytical sample. If the spectral interference does not coincide or overlap the analyte element, spectral interference may still occur if the resolving power and spectral band pass of the monochromator permit the undesired radiation to reach the photoreceptor. 

Spectral line interference is less critical in atomic absorption than it is in flame emission. This is due to the fact that absorption is usually concerned with one spectral line only for each element and that, by proper modulation of the source signal, extraneous spectral lines that do not actually overlap the desired line are not detected. It is wise, however, to use as narrow a slit width as possible to keep the spectral band pass of the monochromator to a minimum. Actual spectral line overlap cannot be corrected by this means. If spectral line overlap occurs, as might happen, e.g., with palladium at 3404.6 A and cobalt at 3405.1 A, the only solutions are (1) to use another spectral line of the element or (2) remove the offending element from the analytical sample. 

Spectral bands or a general continuum can be troublesome for example, the copper line at 3274.0 A and the OH band with a band head at 3274.2 A mutually interfere. The best system for control of the OH band interference is to use a narrow slit and read the magnitude of the OH band contribution... 

---