Inductively Coupled Plasma in Fluorescence Spectrometry Source and Atomhon Reservoir* Journal of Analytical Atomic Spectrometry L I Invited Lecture STANLEY GREENFIELD Department of Chemistry Loughborough University of Technology Loughborough Leicestershire LEI 1 3 TU UK Over the last decade three systems involving the use of inductively coupled plasmas in atomic fluorescence spectrometry have been subjected to similar thorough investigations which have resulted in similar findings. This similarity is not altogether surprising as all the systems employ a low-power plasma as an atomizer and differ only in the source employed namely another low-power plasma a high-power plasma and hollow-cathode lamps. The system employing a high-power plasma is taken as the role model and the investigations and findings on that system are described.Possible ways of improving on the results from all three systems are described and suggestions are made on the manner in which the technique may advance. Keywords Inductively coupled plasma; atomic fluorescence spectrometry In the early 1980s work started independently on three atomic fluorescence spectrometric (AFS) systems incorporating an inductively coupled plasma (ICP). Kosinski Uchida and Winefordner' described a system which they called ICP-ICP- AFS where the radiation from excited species in the tail flame of a low-power plasma was used to excite fluorescence in atomic species in the tail flame of a similar low-power plasma. Greenfield2 described a system incorporating a high-power ICP as the source and a low-power plasma as the atomizer; this system was given the acronym ASIA representing atom- izer source ICPs in AFS.Demers and Allemande3 described a system incorporating an ICP as an atomizer and hollow- cathode lamps (HCLs) as sources. This they called HCL-ICP- AFS and it was the precursor to the commercial instrument manufactured by Baird.4-6 Work on these three systems followed much the same pattern with similar areas of investigation in each instance and with similar results. The first experiments with the ASIA system' were to produce power emission excitation and fluorescence curves of growth. These experiments gave the range and operating parameters of the system and confirmed the fact that an ICP is optically thin is a line source and in addition is an efficient atomizer.There are far fewer spectral interferences in AFS than there are in atomic emission spectrometry (AES) or even atomic absorption spectrometry (AAS) undoubtedly because certain stringent conditions which could lead to interference are not often met with in AFS. First the emission profile of the source must overlap with the absorption profile of the interfering element in the atomizer. Second the population of the interfer- ing element in the correct energy level must be significantly high. Third the amount of energy absorbed by the interfering * Presented at the Seventh Biennial National Atomic Spectrometry Symposium (BNASS) Hull UK July 20-22 1994. element to that emitted as fluorescence radiation must be significant. This freedom from spectral interference was investigated using ASIA.In early experiments the classical work of Larson and Fassel' on background shifts due to radiative recombina- tion continua and interference due to collisional broadening were repeated. Because ASIA employs a.c. coupled electronics and background shifts are essentially d.c. shifts this type of interference was shown to be absent.' In emission there is considerable interference on the emission lines of A1 in the 393-396nm region when 1OOOppm of calcium is added to a 1 ppm solution of aluminium. However there is no such interference in atomic fluorescence with ASIA when a similar amount of calcium is added to a 1 pprn solution of Al. This lack of spectral interference is remarkable. Thus although the major zinc resonance line is only 0.003 nm away from a copper non-resonance line trace amounts of zinc can be determined in pure copper.' Many more instances of spectral interference which occur in emission spectrometry were shown to be absent in A F S .~ ~ Studies were also made" of the non-resonance transitions of lead using filters between the source and atomizer plasmas to isolate the required wavelengths. All the fluorescence mech- anisms resonance Stokes direct line anti-Stokes direct line stepwise line and thermally excited fluorescence were observed and verified. This type of non-resonance fluorescence would not only make spectral interference virtually impossible but would also avoid any problems of scattered light and enable multi-pass optics to be used. Studies of chemical and ionization interferences were made." These included the well known depressive effect of PO4 and A1 on Ca I emission and the enhancement due to alkali metal elements.It was found that when the ASIA operating param- eters had been optimized for the best limit of detection (LOD) by the alternating variable search (AVS) method,12 PO4 had no effect on the fluorescence signal of Ca I up to a PO4 concentration of lo4 ppm the Ca concentration being 1 ppm. On the other hand A1 showed a marked depression of the signal. Na and K gave first an enhancement and then a depression of the signal. The effect of A1 on the Ca was judged to be that of stable compound formation. The effect of the alkali metals was thought to be an ionization effect followed by a quenching effect.On re-optimization for the figure of merit minimum inter- ference the results obtained were different; PO4 had no effect A1 could be tolerated up to a concentration of 900ppm and Na and K up to concentrations of 1000 and 100 ppm respec- tively before a depression could be seen in the fluorescence signal. These experiments confirmed the efficacy of the AVS optimization and the spatial dependence of the reactions in the plasma. Optimization for minimum interference will of necessity degrade the LOD obtained as it does in emission. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 183When the experiments were repeated with a 1 ppm concen- tration of Ca but with measurement of the Ca I1 line it was found that when optimized for minimum LOD PO4 had no effect but Al Na and K had the expected depressive and enhancement effects.However re-optimization for minimum interference showed no interference from PO and Na and K could be tolerated up to about 300ppm but the depressive effect of A1 could not be reduced. As stated at the beginning similar investigations were carried out by other workers on the other two systems i.e. ICP-ICP- AFS'. l3-I6 and HCL-ICP-AFS.'5-17 The results obtained were in broad agreement with those obtained by ASIA. The selec- tivity was excellent. The linear dynamic range was 5-6 orders of magnitude." Chemical and ionization interferences were no greater than in ICP-AES itself regarded as relatively free from this type of interference. Comparison of the LODs obtained by the three systems is difficult rather like comparing apples with oranges as can be seen from Table 1.All else being equal one might expect ASIA to yield lower LODs than the low-power system ICP-ICP-AFS as the fluor- escence radiation produced in the atomizer is initially proportional to the source radiation. This in turn will be proportional initially to the concentration of excited species in the tail flame which in turn is related to the power in the plasma and the concentrations of the appropriate element in the solution nebulized into the plasma. All of these are much greater in the high-power plasma of ASIA than in the low- power systems. Referring to Table 1 and the non-refractory elements this expectation would appear to be fulfilled. This does not appear to be so when one considers the refractory elements.However it should be noted that the efficiency of transfer of energy from the source to the atomizer is 1-2% in ASIA and 13% in the low-power system owing in the latter instance to superior optics. From experience it can be said that a five-fold improve- ment in LOD would be a modest figure to assume for an increase in transfer efficiency from 1 to 13% source to atomizer in the ASIA system. The energy transfer efficiency from source to atomizer for the Baird instrument is not known. However the transfer efficiency of fluorescence radiation from the atomizer plasma to the detector is likely to be higher than in the other two systems as an interference filter is used to isolate the required wavelength and not a monochromator.It is known that an ultrasonic nebulizer when used in conjunction with the source plasma on ASIA will give an increased signal by a factor of 3-4. Further it is also knownI7 that equal signals can be obtained from two solutions the elemental content of one being 60-70 times lower than the Table 1 LODs (ng m1-l) (3 x sb) obtained by ICP-ICP-AFS,l6 HCL- ICP-AFS'*and ASIA" compared with LODs obtained by ICP-AES19 ICP-ICP-AFS Element ( 13 YO)* Non-refractory elements- Ca 0.6 c o Cr 15.0 c u 0.6 Fe Na 1.5 Zn 3.0 Refractory elements- A1 15.0 B 15.0 Ba 1.4 Mo - Si 10.5 - - ASIA ( 1 Yo)* 0.2 9.5 3 .O 0.4 5.0 0.1 2.0 20.5 28.0 3.5 63.0 54.5 HCL-ICP-AFS (?)* co.1 0.5 0.6 0.1 0.5 <0.1 <0.1 5.0 60.0 25.0 8.0 40.0 ICP-AES 0.15 3.0 3 .O 1.5 1.5 6.0 1.5 6.0 3.0 0.15 7.5 5.0 ~~ ~~ * Efficiency of light transfer source to atomizer.other if the aerosol from the weaker solution is heated and de- solvated before it is passed through the ICP. There are problems associated with the de-solvation of heated aerosols which cannot be discussed here but these problems may not occur with cryogenic membrane de-solvation systems. Adopting the same approach to the atomizer plasma would undoubtedly raise its temperature which would necessitate lowering the power or observing further along the tail flame. If this proved possible then a combination of ultrasonic nebulizer with de-solvation should produce a higher signal. However the greatest advantage to be gained from de- solvation on the atomizer plasma is likely to be the removal of oxygen which should improve the LOD of the refractory elements.The point of this discussion is that an improvement in optical transfer from source to atomizer and from atomizer to the monochromator or filter and detector together with the use of ultrasonic nebulizers and de-solvation should result in LODs at the very least equal to those in ICP-AES for the refractory elements and much better for the non-refractory elements. This is a good place to discuss the future of the ICP in AFS (excluding the laser-ICP systems as being beyond the scope of this paper). The Baird instrument (which has been withdrawn from the market for as far as is known commercial and not scientific reasons) may have been before its time. It was a sophisticated machine somewhat expensive and exceeded the requirements of many laboratories and encouraged a simplistic approach to a technique that requires some understanding.That the technique of ICP-AFS has merit is an established fact but further progress will only come about if the scientific community can be encouraged to follow the example of the pioneers of ICP-AES who were prepared to retrofit an ICP to existing monochromators or polychromators in order to find out for themselves what such a system would do for them. In a similar manner use can be made of the ICP in an existing ICP-AES instrument as a source and as an atomizer in an AFS mode. Also simple instruments along the lines of early AAS machines can be assembled to investigate the use of the technique of AFS. Such an instrumental set-up has been put together at Loughborough University of Technology (LUT) and is shown in Fig.1. Recorder -. -/x I torch Ta i If I a m e' Hollow cathode '0 lamD Fig. 1 Schematic arrangement for BDHCL-ICP-AFS with axial viewing of the ICP 184 Journal of Analytical Atomic Spectrometry March 1995 Vol. 10Basically it consists of a rearrangement of a very old Shandon AAS machine together with a plasma in an end-on configuration. The low-power generator used to drive the plasma was built in-house. The lamp turret and pulsed power supply are retained as is the monochromator and detector which is turned through 90" and placed on its side so that its slit is parallel to the tail flame of the plasma. A lock-in amplifier integrator and recorder complete the set-up. The sources are HCL and boosted discharge hollow cathode lamp (BDHCL) the latter requiring a boosted power supply. The end-on configuration was chosen in order to increase the size of the fluorescence cell.The HCL source can be replaced by the end of a fibre-optic bundle conveying radiation from an ICP in an existing ICP-AES instrument if it is so desired. To facilitate this use of an existing plasma torch the device shown in Fig. 2 was constructed at LUT. The 'top-hat' that is used to prevent 'arc-over on many ICP torches is replaced by a ground silica flange joint. The joint fits into a cylindrical block of machinable ceramic that has a hole bored through the middle to allow the passage of the plasma and tail flame. Another hole is drilled into the block at right-angles to the central hole to allow the insertion of an optical fibre bundle.If the light from the bundle is required to be modulated it is passed through condensing lenses between which is placed a mechanical chopper. Omenetto2' described an ICP-AFS system in which light from an ICP in a commercial polychromator was used via a lens system with an optical chopper to excite fluorescence using a graphite rod atomizer with spectacular LODs being obtained as can be seen in Table 2. A fibre optical bundle Ceramic blocks A m - 1 Collimating beam probes Fig. 2 Plasma radiation transference device and optical chopper Table 2 LODs (ng ml-l) (3 x sb) obtained by ICP-electrothermal atomization ( ETA)-AFSZo compared with LODs obtained by graphite furnace (GF) AASI9 and by ICP-MS19 Element ICP-ETA- AFS GFAAS ICP-MS Ag 0.04 0.05 0.003 Cd 0.01 0.02 0.003 c u 0.02 0.25 0.003 Mg 0.002 0.01 0.007 Zn 0.003 0.3 0.003 could be used to transfer the radiation from the source to the graphite rod more conveniently.Another possible use of the optical bundle light transference device is in the field of molecular fluorescence where light from an existing plasma instrument could be directed into a cuvette containing a solution of a compound of interest which emitted fluorescence when irradiated by light of a suitable wavelength. This fluorescence radiation after passing through a filter or monochromator could be detected in a variety of ways. Tallant" reported about 15 years ago that an ICP has a comparable performance to an arc lamp in molecular fluor- escence.Surprisingly no further work seems to have been reported following this early paper; yet as Tallant pointed out good reasons exist for further work to be done. By introducing a suitable element into an ICP a desired spectral region may be selected as an excitation source. This ability will be of greatest value for regions of the spectrum in which arc sources have relatively weak output. Prominent plasma lines are available through the ultraviolet region with many elements emitting their strongest line below 250 nm and a few below 200nm. Thus for certain applications the ICP may be more desirable as a source for molecular fluorescence than the commonly used arc lamps such as xenon which have relatively weak UV output or like Hg lamps have holes in their intensity - wavelength profiles.All such lamps can show considerable instability in their radiant fluxes. So far the discussion has been of the use of simple instrumen- tation in order to further the use of ICP-AFS and suggestions for the retrofitting of the ICP in existing commercial instru- ments as a source in fluorescence applications. There is also the possibility of replacing the conventional torch in existing instruments with a long-sleeve torch and using it as an atomizer with an HCL or BDHCL as a source. In the latter instance light transfer would be by lens and/or fibre-optic bundle depending on the geometry of the torch box. There is every possibility that the monochromators in the ICP instru- ment could be used to isolate the required wavelength and with the addition of a lock-in amplifier the existing data collection system could be utilized.In conclusion it can be said that the use of the ICP in AFS has been well researched (over 100 papers have been published on this topic in the last 10 years) and the technique has been found to have many of the attributes of AES and AAS and can be virtually free from spectral interference. It has been demonstrated that the LODs obtained by the technique good as they are can be improved by giving attention to optical performance nebulization and desolvation. Although no com- mercial instruments for ICP-AFS are available it has been shown how simple instruments can be built and the plasma in existing ICP-AES instruments can be utilized to perform both atomic and molecular fluorescence and thus further the use of the technique.REFERENCES 1 2 3 4 5 6 7 8 9 10 Kosinski M. A. Uchida H. and Winefordner J. D. Anal. Chem. 1983 55 688. Greenfield S. Anal. Proc. 1984 21 61. Demers D. R. and Allemand C. D. Anal. Chem. 1981,53 1915. Demers D. R. and Allemand C. D. paper presented at the 1981 Pittsburgh Conference Atlantic City NJ paper No. 122. Demers D. R. and Allemand C. D. paper presented at the 1981 Pittsburgh Conference Atlantic City NJ paper No. 123. Demers D. R. Busch D. A. and Allemand C. D. Znt. Lab. 1982 14 40. Greenfield S. and Thomsen M. Spectrochim. Acta Part B 1985 40 1369. Larson G. F. and Fassel V. A. Appl. Spectrosc. 1979 33 592. Greenfield S. and Thomsen M. Anal. Proc. 1987 24 22. Greenfield S. Malcolm F. M. and Thomsen M.J . Anal. At. Spectrom. 1987 2 711. Journal of Analytical Atomic Spectrometry March 1995 Vol. 10 18511 12 13 14 15 16 17 18 Greenfield S. Salmon M. S. and Tyson J. F. Spectrochim. Acta Part B 1988 43 1087. Greenfield S. Salmon S. M. Thomsen M. and Tyson J. F. J. Anal. At. Spectrom. 1989 4 55. Long G. L. and Winefordner J. D. Appl. Spectrosc. 1984,38,563. Long G. L. Voigtman E. G. Kosinski M. A. and Winefordner J. D. Anal. Chem. 1983 55 1432. Walters P. E. Long G. L. and Winefordner J. D. Spectrochim. Acta Part B 1984 39 69. Kruppa R. J. Long G. L. and Winefordner J. D. Spectrochim. Acta Part B 1985 40 1485. Greenfield S. Durrani T. M. and Tyson J. F. ICP I$. Newsl. 1990 16 159. Demers D. R. and Montaser A. in Inductively Coupled Plasmas in Analytical Atomic Spectrometry ed. Montaser A. and Golightly D. W. VCH New York 2nd edn. 1992 p. 553. 19 The Guide to Techniques and Applications of Atomic Spectroscopy Perkin-Elmer Norwalk CT 1993. 20 Omenetto N. paper presented at Analytiktreffen Atmospektrosk Fortschr. Anal. Ammend Haup vortr. meeting held at Karl Marx University Leipzig Germany 1982. Tallant D. R. ICP In$ Newsl. 1979 5 171. 21 Paper 4/04 755 I Received August 2 1994 Accepted August 24 1994 186 Journal of Analytical Atomic Spectrometry March 199.5 Vol. 10