analytical

CONTROL OF ANALYTICAL INTERFERENCES THE FLAME PROCESS Atornic absorption is known as a very specific technique with few interferences. The ultimate analytical method which is absolutely free of any interferences from the nature ofthe sample will probably never exist. The next best thing to not having interferences is to know what the interferences are and how to eliminate them or compensate for them. The interferences in atomic absorption are well-defined, as are the means for dealing with them. In order to understand these interferences we will examine what goes on in the flame atomization process of atomic absorption in order to get the atomic absorption process to occur; we must produce individual atoms from our sample which starts out as a solution of ions. This process is diagrammcd in Figure 3-1. First, by the process of nebulization, we aspirate the sample into the burner chamber ,where it mixes as a fine aerosol with the fuel and oxidant gases. At this point , the metals are still in solution in the fine aerosol droplets As these tiny droplets pass into the heat of the flame, the process of evaporation or desolvation removes the solvent and leaves tiny solid particles of sample material. As more heat is applied, liquefaction will take place, and additional heat will vaporize the sample. At this point the metal of interest, called the analyte, is still bound up with some anion to form a molecule which does not exhibit the atomic absorption phenomenon we wish to measure. By applying still more heat ener~y, this molecule is dissociated into the individual atoms which make it up. (Solution) ) Nebulization (Aerosol) 2)Desolvation (Solid) 3) Liquefaction (Liquid) 4) Vaporization (Gas) 5) Atomization (Gas) 6) Excitation (Gas) 7) Ionization (Gas) . 3-2 Since the thermal energy from the flame is responsible for producing the absorbing species, flame temperature is an important parameter governing the flame process. Temperatures for some flames that have been used in atomic absorption are listed in Table 3-1. Cooler flames are subject to more interference problems resulting from insufficient energy for complete atomization. The two premix flames now used almost exclusively for atomic absorption are air-acetylene and nitrous oxide-acetylene. While the air-acetylene flame is satisfactory for the majority ofelements determined by atomic absorption, the hotter nitrous oxide-acetylene flame is required for many refractory-forming elements. The nitrous oxide-acetylene flame is also effective in the control of some types of interference. Table 3-1 Temperatures of Premix Flames Oxidant-Fuel Temperature Air-Methane 1850-1900 Air-Natural Gas 1700-1900 Air-Hydrogen 2000-2050 Air-Acetylene 2125-2400 N20-Acetylene 2600-2800 The number of ground state metal atoms formed in step 5 of the flame process (Figure 3-1) will determine the amount of light absorbed. Concentration is determined by comparing the absorbance ofthe sample to that of a known standard concentration. The relationship between the number of atoms in the flame and the concentration of analyte in solution is governed by the flame process. If any constituent ofthe sample alters one or more steps ofthis process from the performance observed for a standard, an interference will exist, and an erroneous concentration measurement will result ifthe interference is not recognized and corrected or cornpensated. NONSPECTRAL INTERFERENCES Interferences in atomic absorption can be divided into two general categories., spectral and nonspectral. Nonspectral interferences are those which affect the formation of analyte atoms Matrix Interference first place in the flame atomization process subject to interference is the very first step, the nebulization. If the sample is more viscous or has considerably differert surface tension characteristics than the standard, the sample uptake rate or nebuIization efficiency may be different between sample and standard. If samples and standar(is are not introduced into the process at the same rate, it is obvious tat the numher of atoms in the light beam and, therefore, the absorbance, will not correl ate between the two. Thus, a matrix interference will exist. An example of this type of interference is the effect of acid concentration on absorbanec.. From Figure 3-2, it can be seen that as phosphoric acid concentration increases(and the sample viscosity increases), the sample introduction rate and the sample absorbanee decrease. Increased acid or dissolved solids concentration normally will lead to anegative error if not recognized and corrected. Matrix interferences can also cause positive error The presence of an organic solvent in a sample will produce an enhanced nebulization efficiency, resulting in an increased absorption. One way of compensating for this type of interference is to match as closely as possible the major matrix components of the standard to those ofthe sample. Any acid or other reagent added to the sample during preparation should also be added to the standards and blank in similar concentrations. Method of Standard Additions There is a useful technique which may make it possible to work in the presence of a matrix interference without eliminating the interference itself, and still make an accurate determination of analyte concentration. The technique is called the mc.tod of standard additions. Accurate determinations are made without eliminating interferences by making the concentration calibration in the presence of the matrix interference. Aliquots of a standard are added to portions of the sample, thereby allowing any interferent present in the sample to also affect the standard similarly. The standard additions technique is illustrated in Figure 3-3. The solid line passing through the origin represents a typical calibration line for a set of aqueous standards. Zero absorbance is defined with a water blank, and, as the concentration of analyte increases, a linear increase in absorbance is observed.Let us now take equal aliquots ofthe sample. Nothing is added to the first aliquot; a measured amount of standard is added to the second; arid a larger measured amount is added to the third. The first volume of added standard is usually selected to approximate the analyte concentration in the sample, and the second volrune is nonnally twice t'lie first volume. However, for the method of standard additions to be used accurately, the absorbances for all of the solutions must fall within the linear portion of the working curve.diluted to the same vol-Finally, all portions are ume so that the final concentrations of tbe original sample constituents are the same in each case. Only the amount of added analyte differs, and then by a known amount.If no interference were present in this sample, a plot of measured absorbance versus the concentration of added standard would be parallel to the aqueous standard calibration, and offset by an absorbance value resulting from the analyte present in the unspiked sample. If some material is present in the sample which causes a matrix interference, the number of ground state atoms producing atomic absorption will be affected, as will be the absorbance from the analyte in the unspiked sample. However, the absorbance increase from added standard will also be changed by the same proportional amount since the concentration of interferent is tile same in each solution. Therefore, a straight line will still result, but because of tile interference, its slope will be different from that observed for the aqueous standards. In this situation, if the absorbance of the unspiked sample were to be compared dircctiy to the aqueous calibration, an error would result. If however, the slope dotormined by the standard additions to our sample is used as the calibration slope, an accurate determination of the sample concentration can still be made. By continuing tile concentration calibration on the abscissa backward from zero and cxtrapolating the calibration line backward until it intercepts the concentration axis,the concentration responsible for the absorbance of the unspiked sample is indicated. An accurate determination has been made by calibrating in the presence of the interference. ProperIy used, the method of standard additions is a valuable tool in atomic absotptiorn. The presence of an interference can be confirmed by observing the slope of the spiked sample calibration and determining whether or not it is parallel to .ilc aqueous standard line. If it is not, an interference is present. If an interference is present the method of standard additions may allow an accurate determination or t~AC unknown concentration by using the standard additions slope for the calihration Caution should be used with the technique, however, as it can fail to give correct answers with other types of interference. The methodofstandardadditions wil1 not compensate for background absorption or other types ofspectral intere, and normally will not compensate for chemical or ionization types of interference. Chemical Interference A second place where interference can enter into the flame process is in step numher 5 of Figure 3-1, the atomization process. In this step, sufficient energy must be available to dissociate the molecular form of the analyte to create free atoms. If tile sample contains a component which forms a thermally stable compound with the analyte that is not completely decomposed by the energy available in the flame a cilemical interference will exist. iThe effect of phosphate on calcium, illustrated in Figure 3-4, is an example of a chemical interference. Calcium phosphate does not totally dissociate in an air-acetylene flame. Therefore, as phosphate concentration is increased, the absor-bance due to calcium atoms decreases. 3-6 There are two means of dealing with this problem. One is to eliminate the interference by adding an excess of another element or compound which will also form athermally stable compound with the interferent. In the case of calcium, lanthanum is added to tie up the phosphate and allow the calcium to be atomized, making the calcium absorbance independent of the amount of phosphate. There is a second approach to solving the chemical interference problem. Since the problem arises because of insufficient energy to decompose athermally stable analy'te compound, the problem can be eliminated by increasing the amotint ofenergy; that is, by using ahotter flame. The nitrous oxide-acetylene flame is considerablyhotter than air-acetylene and can often be used to minimize chemical interferences for elements generally determined with air-acetylene. The phosphate interference on calcium, for instance, is not observed with a nitrous oxide-acetylene flame, eliminating the need for the addition of lanthanum. Ionization Interference There is a third major interference, however, which is often encountered in hot flames. As illustrated in Figure 3-1, the dissociation process does not necessaril~ stop at the ground state atom. If additional energy is applied, the ground state atoni can be thermally raised to the excited state or an electron may be totally removed from the atom, creating an ion. As these electronic rearrangements deplete the number of ground state atoms available for light absorption, atomic absorption at the resonance wavelength is reduced. When an excess ofenergy reduces the popu lation of ground state atoms, an ionization interference exists. Ionization interferences are most common with the hotter nitrous oxide-acetylene flame. In an air-acetylene flame, ionization interferences are normally encoun Control of Analytical Interferences 3-7 tcmd only with the more easily ionized elements, notably the alkali metals and alkaline earths Ionization interference can be eliminated by adding an excess ofan clement which is very easily ionized, creating a large number of free electrons in the flame and suppressing the ionization of the analyte. Potassium, rubidium, and cesium salts arc commonly used as ionization suppressants. Figure 3-5 shows ionization suppression for the determination of barium in a nitrous oxide-acetylene flame The increase in absorption at the barium resonance ine, and the corresponding decrease in absorption at the barium ion line as a function of added potassium, illustrate the enhancement of the ground state species as the ion form is suppressed. By adding 1000 mg/L to 5000 mg/L potassium to all blanks,standards and samples, the effects of ionization can usually be eliminated. SPECTRAL iNTERFERENCES spectral interferences are those in which the measured light absorption is erroneously high due to absorption by a species other than the analyte element. The most common type of spectral interference in atomic absorption is '' background absorption ''. Background Absorption Background absorption arises from the fact that not all of the matrix materials in a sample are necessarily 100% atomized. Since atoms have extremely narrow absorption lines, there arc few problems involving interferences where one clement absorbs at the wavelength of another. Even when an absorbing wavelength ofanother element falls within the spectral bandwidth used, no absorption can occur unless the light source produces light at that wavelength, i.e., that element is also present in the light source. However, undissociated molecular forms ofmatrix materials may have broadband absorption spectra, and tiny solid particles in the flame may scafler light over a wide wavelength region. When this type of nonspecific absorption overlaps the atomic absorption wavelength of the anal ytc, backgroL'nd absorption occurs. To compensate for this problem, the background absorption must be measured and subtracted from the total measured absorption to determine the true atomic absorption component. While now virtually obsolete, an early method of manual background correction illustrates clearly the nature of the problem. With the "two line method", background absorption, which usually varies gradually with wavelength, was independently measured by using a nonabsorbing emission line very close to the atomic line for the analyte element, but far enough away so that atomic absorption was not observed, as illustrated in Figure 3-6. By subtracting the absorbance measured at the nonabsorbing line from the absorbanec at the atomic line, the net atomic absorption was calculated. Nearby, nonabsorbing lines arc not always readily available, however, and inaccuracies in background correction will result ifthc wavelength for background measurement is not extremely close to the resonance line. Therefore, for accuracy, as well as convenience, a different method was needed Continuum Source Background Correction Continuum source background correction is a technique for automatically measuring and compensating for any background component which might be present in an atomic absorption measurement. This method incorporates a continuum light source in a modified optical system, illustrated in Figure 3-7. The broad band continuum ("white" light) source differs from the primary (atomic line) source in that it emits light over a broad spectrum of wavelengths 3-10 instead of at specific lines. From Figure 3-8, it can be seen that atomic absorption, which occurs only at very discrete wavelengths, will not measurably attenuate the emission from the continuum source. However, background absorption which has very broad absorption spectra will absorb the continuum emission as well as flie line emission. As shown in Figure 3-7, light from both the primary and continuum lamps is combined and follows a coincident path through the sample, through the monochromator, and to the detector. Thc two lamps are obscrved by the detector alternately in time, and as illustrated in Figure 3-9, instrument electronics separate the signals and compare the absorbance from both sources. An absorbance will be displayed only where the absorbance of the two lamps differs. Since background ah-sorption absorbs both sources equally, it is ignored. True atomic absorption, which absorbs the primary source emission and negligibly absorbs the broad band continuum source emission, is still measured and displayed as usual. Figure 3-10 shows how background absorption can be automatically eliminated from the measured signal usmg continuum source background correction. In the cx- Control of Analytical lnterferences 3-11 ample, a lead determination is shown without background correction (A) and with background correction (B). Both determinations were performed at the Pb 283.3 om wavelength with 15x scale expansion and a 10-second integration time. Continuum source background correction is widely applied, and except in some ~ unusual circumstances, is fully adequate for all flame AA applications. There arc some limitations to continuum source background correction, however, which especially impact graphite furnace atomic absorption, to be discussed in later eb~pters. These limitations are summarized in Table 3-2. Table 3-2 Limitations of Continuum Source Background Correction 1 Requires additional continuum light source(s) and electronics. 2. Requires the intensities of the primary and continuum sources to be similar. 3. Two continuum sources are required to cover the full wavelength range. 4. Requires critical alignment of the continuum and pnmary sources for accurate correction. 5. May be inaccurate for structured background absorption. The fact that continuum source background correction requires two sources (primary and continuum) imposes convenience, economic, and performance consideration on the use of the technique. The convenience and economic factors come from tb fact that the continuum source has a finite lifetime and must be replaced on a periodic basis. The performance factor originates from the fact that the background component of the absorption signal is measured from one source, while tbe total uncorrected signal is measured with another. This leads to the possibility ofinaccurate compensation if the two sources do not view exactly the same portion ofthe atom cloud, especially at higher background absorption levels. Finally, since tbe two sources are spectrally different, background absorption exhibiting fine spectral structure may attenuate the two source lamps to different degrees, leading to inaccuracies in background correction for such cases. Introduction to Zeeman Background Correction For those applications where the limitations of the continuum source approach are signiFeant to the analysis, the Zeeman background correction system may be preferable. Zeeman background correction uses the principle that the electronic energy levels of an atom placed in a strong magnetic field are changed, thereby changing the atomic spectra which are a measure of these energy levels. When an atom is placed in a magnetic field and its atomic absorption profile observed with polarized light, the normal single-line atomic absorption profile is split into two or more components symmetrically displaced about the normal position, as illustrated in Figure 3-liThe spectral nature of background absorption, on the other hand, is usually unaffected by a magnetic field. By placing the poles of an electromagnet around the atomizer and making alternatmg absorption measurements with the magnet off and then on, the uncorrected total absorbance (magnet off) and "background only" absorbance (magnet on) can be made, as in Figure 3-12. The automatic comparison made by the instrument to compensate for background correction is similar to that for the continuum source technique, except that only the one atomic line source is used. As a result, there are no potential problems with matching source intensities or coincident alignment of optical paths. Also, background correction is made at the anal yte wavelength rather than across the entire spectral bandwidth, as occurs with continuum source background correction. With Zeeman background correction, the emission profile of the line source identical for both AA and background measurements. As a result, most complex structured background situations can be accurately corrected with Zeeman background correction. This can be seen in Figure 3-13, where background absorption due to the presence of aluminum in a graphite furnace determination of arsenic is cornpletely compensated using Zeeman correction but produces erroneously high resuIts with continuum source background correction. Table 3-3 summarizes the advantages of Zeeman effect background correction. Table 3-3 Advantages of Zeeman Effect Background Correction 1. Corrects for high levels of background absorption. 2. Provides accurate correction in the presence of structured background. 3. Provides tme double-beam operation. 4. Requires only a single, standard light source. 5. Does not require intensity matching or coincident alignment of multiple sources. examples used above to illustrate Zeeman effect background correction are based on the use of a transverse AC Zeeman system, the type most commonly used 3 with commercial AA instrumentation. However, there are three types of Zeeman effect background correction systems available on commercial atomic absorption instruments: DC Zeeman, transverse AC Zeeman and longitudinal AC Zeeman. These systems differ in the way the magnetic field is applied and by the means used to measure the combined (atomic absorption plus background absorption) and background absorption only signals. DC Zeeman systems use a permanent magnet and a rotating or vibrating polarizer to separate the combined and background only signals. AC Systems use an electromagnet, and measure the combined and background only signals by turning the magnetic field on and off. The difference between transverse (magnetic field applied across the light path) and longitudinal (magnetic field applied along the light path) AC Zeeman Systems is that transverse systems uses a fixed polarizer, while the longitudinal system does not require apolarizer. The advantages and limitations ofeach type ofZeeman system are summarized in Tables 3-4 and 3-5 on the following page. Other Spectral Interferences If the atomic absorption profile for an element overlaps the emission line ofanother, a spectral interference is said to exist. As has already been mentioned, ~is is an infrequent occurrence, because of the very' wavelength-specific nature of atomic absorption. Even if an absorption line for an element other than the analyte but also present in the sample falls within the spectral bandpass, an interference will occur only if an emission line of precisely the same wavelength is present in the source. As the typical emission line width may be only 0.002 nanometers, actual overlap is extremely rare. The chances for spectral interference increase when multi-element lamps are used, where the source may contain close emission lines for several elements. Particular care should be exercised when secondary' analytical wavelengths are being used in a multi-element lamp. Procedures for circumventing spectral interference include narrowing the monochromator slit width or using an alternate wavelength. INTERFERENCE SUMMARY The major interferences in atomic absorption include: (1) matrix interference, (2) chemical interference, (3) ionization interference, and (4) background absorption. For the first type, special considerations in sample preparation or the use of tlie method of standard additions may compensate for the problems generated. For the second and third, addition of an appropriate releasing agent or ionization buffer or changing the flame type used will normally remove the interference. For the fourth, background absorption, an instrumental correction technique will automatically compensate for the biasing effects. Application ofthe techniques described bcrc will make possible accurate atomic absorption determinations in very cornplcx samples. Table 3A DC Zeeman Systems Advantages: Less expensive to operate (lower power consumption) Disadvantages: Has poorer sensitivity and analytical working range relative to AC Zeeman systems. The polarizer reduces light throughput by as much as 50%, affecting analytical performance. A mechanical assembly is required to rotate or vibrate the polarizer. Table 3~ AC Zeeman Systems Advantages: Offers better sensitivity and expanded analytical working ranges relative to DC Zeeman systems. No polarizer is required, so it provides higher light throughput and improved analytical performance. (Longitudinal AC Zeeman systems only) Requires no additional mechanical devices. Disadvantages: Requires more electrical power than DC Zeeman systems, so has higher operating expenses. The polarizer causes reduced light throughput by as much as 50%, affecting analytical performance. (Transverse AC Zeeman Systems only.)