Emerging from the background

Background interference in atomic absorption spectrophotometry can be reduced - Vince Phelan tells us how

ATOMIC absorption spectrophotometry is an extremely well-established analytical technique capable of detecting elemental concentrations down to part per billion (ppb) levels. A sample is atomised either in a flame or furnace arrangement and the atoms are irradiated with light. An optical detection system allows the amount of light absorbed to be measured to give the concentration of the element in the sample. The traditional instrumental configuration uses an element-specific light source, usually a hollow cathode lamp (HCL).  More recently, a completely new approach to AAS (HR-CS AAS) has been introduced, where a high resolution continuum source (a specially designed xenon short-arc lamp) is used in conjunction with a linear CCD array detector. For conventional AAS, HCLs are available for most elements, so the appropriate lamp can be chosen for each analysis and for multi-element analyses, lamps can be mounted on a rotating turret so that different elements can be analysed sequentially. In the HR-CS AAS configuration, the single lamp produces emission lines for all elements and the optical configuration allows the appropriate emission line to be selected for extremely fast sequential analysis.

One major problem encountered in atomic absorption spectrophotometry (AAS)

Figure 1: Atomic absorption calculated from the difference of two measured intensities

  is the non-specific absorption of radiation by free molecules from matrix components in the sample which have not dissociated and light scattering by particles of matrix substances which have not completely evaporated, resulting in an excessively high background signal. The recorded signal consists of the analyte-specific absorption which we want and the non-specific absorption of the background which we don’t.

For conventional AAS, a number of techniques have been developed to separatethe contribution from the background signal to leave the required analyte signal. The most common ones are: Deuterium background correction, Zeeman-technique with different variations and the Smith-Hieftje-background correction, which, for patent reasons, has a number of different names. For HR-CS AAS, however, background correction becomes a natural consequence of the way the optical system is configured.

The deuterium background correction offers a

Figure 2: Magnetic field causes the energy levels of an atom, and consequently the spectral lines of the atom to be split into three components, s-, p and s-

simple and inexpensiveapproach.In this method, a continuous light source - usually a D2-lamp - is added to the optical system and light from the HCL and the D2-lamp pass alternately through the atomisation cell at a frequency dependent on the particular instrument. In the presence of the appropriate element, the intensity of the HCL is reduced by absorption corresponding to the elemental concentration, but the reduction in the intensity of the D2-lamp is negligible. Any broad-band background absorption, however, causes an equal reduction in intensity for both light sources. Since a general background causes an equal decrease in intensity of both light sources, but absorption from the element of interest is only significant for the HCL, the atomic absorption is calculated from the difference of the two measured intensities (Figure 1). There is no loss of sensitivity using the deuterium background correction method and it is easily retro-fittable. However it is only generally effective for elements whose absorption lines lie in the 190 to 360nm region of the spectrum and it cannot be used to correct for a structured background.  

The Zeeman-background correction uses theprinciple that free atoms show a

Figure 3: During high current mode dense atom clouds are scattered by the hollow-cathode causing the resonance line to be strongly broadened, losing its Gaussian shape, with a decreased intensity in the middle

Zeeman-splitting in a magnetic field but molecules and liquid or solid particles show either no or negligible Zeeman-effect and so advantage can be taken of the polarization properties of the light. The magnetic field causes the energy levels of an atom, and consequently the spectral lines of the atom to be split into three components, s-, p and s- (Figure 2). The p component is  linearly polarised parallel to the magnetic field while the s- components are circularly polarised perpendicular to the magnetic field. Molecules, particles and any other background materials which do not exhibit the Zeeman Effect will absorb or scatter equally light of either polarisation. There are a variety of configurations used for Zeeman correction, with the magnetic field being applied either transversely or longitudinally. In general, the magnetic field is pulsed with the polariser applied while the magnetic field is turned on to eliminate the p

component, so absorbance of the background only is measured with the field turned on whilst the background plus analyte is measured with the field off. The difference between the two signals gives the analyte absorbance.

The Smith-Hieftje-Background Correctionmethod uses an alternating highcurrent

Figure 4: In HR-CS AAS the radiation emitted by the continuum radiation source is attenuated by the sample in the atomisation unit

pulse to self-reverse the emission from the HCL. During the normal current mode the lamp emission shows the normal sharp spectral line. During the high current mode dense atom clouds are scattered by the hollow-cathode which absorbs light that has already been emitted by other atoms producing a dip or self reversal peak. This causes the resonance line to be strongly broadened, losing its Gaussian shape, with a decreased intensity in the middle (Figure 3a-c). Both the atomic and the background absorption are measured during the normal current mode. Measurements during the high current mode mainly give the background because of the line broadening and the self-reversal of the line. The difference gives the pure atomic absorption. Special HCLs have to be used because normal lamps cannot be operated with such high current pulses and not all elements have lamps that can be configured for self-reversal. In addition, losses of sensitivity up to 50% can occur because even at that high current there is still an appreciable contribution to the specific atomic absorption. A structured background may cause problems in the same way as the D2-background correction.

High Resolution Continuum Source Correction (HR-CS) AAS is based on ISAS

Figure 5: detector readout of the analytical line for iron at 248.814nm

(Institute for Analytical Sciences, Berlin) technological principles. The radiation emitted by the continuum radiation source is attenuated by the sample in the atomisation unit (flame or furnace), then enters a double monochromator (Figure 4). The intermediate slit separates the part of the spectrum that contains the analytical line which then enters the second monochromator, where it is highly resolved and then transmitted to the detector. This is a linear CCD array that not only detects the analytical line, but also its spectral environment at high resolution. The lamp emission intensity is on average a factor of 100 higher than that of conventional HCL over the entire spectral range. Although the radiation intensity has no influence on sensitivity in AAS, it has an influence on the signal-to-noise ratio. As a result of this, detection limits are on average about a factor of 5 better in HR-CS AAS, compared to conventional AAS. The resolution of the double monochromator is in the range of 140 000, which corresponds to a spectral bandpass of 1.6pm at 200nm – a value that is about a factor of 1000 better than the resolution of classical AAS monochromators. The detector is a linear CCD array with typically 512 pixels (picture elements), 200 of which are used for analytical purposes. Each individual pixel is evaluated independently, so that the equipment basically works with 200 independent detectors. All 200 pixels are illuminated simultaneously (for 1-10 ms) and read out simultaneously. The next illumination is already being carried out during signal evaluation, allowing a very rapid measurement frequency. This fast sequential analysis has particular benefits in flame applications, since multi-element analysis can be carried out using a single sample aspiration, whereas conventional flame AAS would require a separate sample aspiration for each element. This not only saves time, but also reduces sample consumption.

Figure 5 shows the detector readout of the analytical line for iron at 248.814nm.The absorption line is essentially covered by just a few pixels, while the other pixels only exhibit the statistical variation of the baseline, i.e. the noise. As only a few pixels are used in most cases to measure atomic absorption, the others may be used for background correction purposes. Since all pixels are illuminated and read out simultaneously, any variation of the intensity that is independent of wavelength, such as variations in lamp emission, can be detected and eliminated through the use of correction pixels. This results in an extremely stable system with low noise level and significantly improved signal-to-noise ratios. The ultra-high resolution of the detection system also allows the correction for structured backgrounds since it is possible to recognise and avoid spectral interferences much more easily.

First a reference spectrum of the background is recorded and then it is subtracted from the sample spectrum. A good example is the detection of 0.05mg/L Zn in 1% HNO3. During atomisation, NO is produced which gives rise to a spectrum with a severe overlap with the Zn line at 213.857nm. Figure 6a shows the NO spectrum produced from a 1% HNO3 blank, and Figure 6b shows the spectrum with of the Zn + HNO3. Figure 6c shows the resulting spectrum when the background due to NO is subtracted, clearly resolving the Zn line. The fast sequential analysis capability adds a third dimension to the analysis, since we can also resolve the spectra in terms of time (Figure 7). Some background effects occur as a result of effects in the atomisation process that take a finite time to manifest themselves. In conventional AAS, these effects are averaged over the analysis period, but the time resolution offered by HR-CS AAS means that we can see the spectral interference and therefore avoid it.

Figure 6a	Figure 6b
Figure 6c	Figure 7

Figure 6a-c: Figure 6a shows the NO spectrum produced from a 1% HNO3 blank, and Figure 6b shows the spectrum with of the Zn + HNO3. Figure 6c shows the resulting spectrum when the background due to NO is subtracted, clearly resolving the Zn line.

Figure 7: The fast sequential analysis capability adds a third dimension to the analysis - time

The choice of instrument type depends very much on the application and, of course, the price of the instrument. Deuterium lamp correction works well for simple samples in flame AAS, but is less effective for furnace AAS. Zeeman works well on more complex samples and particularly in furnace AAS. HR-CS AAS is clearly the most versatile of the techniques, yet has the traditionally low running costs, maintenance cost and the simplicity associated with an AA instrument. It offers better detection limits on all elements by factors between 3 and 7. In many cases this allows the use of the flame for analysis rather than having to use a graphite furnace. This saves time with the analysis by requiring only a single sample aspiration and economises on sample consumption. For samples that definitely require a furnace system, the HR-CS AAS background correction method is the most effective for all types of samples. The added benefit of the 3D data offered by HR-CS AAS allows the background complexities to be visualised, allowing an easier diagnosis of background issues. Indeed studies are now being reported in the literature where samples with complex backgrounds are being successfully revisited using HR-CS AAS where deuterium lamp or Zeeman correction methods have led to erroneous results in the past.