Sprctrochimica Acta, Vol.41B, No. II, pp.1157-l 165, 1986. Printed inGreat Britain.

0584-8547/66 s3.00 +o.oo 0 1986 Pergamon Joumair Ltd.

Palladium and magnesium nitrates, a more universal modifier for graphite furnace atomic absorption spectrometry
GERHARD SCHLEMMER

and

BERNHARD WE=*

Department of Applied Research, Bodenseewerk Perkin-Elmer & Co. GmbH, D-7770 oberlingen, F.R.G. (Receioed 5 June 1986; in revisedjorm 12 June 1986) Abstract-A mixture of palladium and magnesium nitrates was found to be a very powerful modifier for the determination of As, Bi, In, Pb, Sb, Se, Sn, Te and Tl in graphite furnace atomic absorption spectrometry. Thermal pretreatment temperatures of 90&1400C can be used with the proposed modifier. This is in most cases substantially more than what can be applied with the modifiers recommended up to now, so that separation of the analyte from the concomitants should be easier. This is shown to be true for the determination of lead in sea water and of selenium in biological materials. Optimum atomization temperatures are more uniform and typically around 2OOOYfor the investigated elements when the palladium and magnesium nitrates mixed modifier is used. This modilier therefore allows the use of common conditions for all the investigated elements with a minimum sacrifice in sensitivity, an important pre-requirement for multi-element furnace techniques. The proposed mixed modifier also minimizes the risk of contamination because palladium as well as magnesium nitrate can be obtained in high purity, and both elements are infrequently determined in the graphite furnace.

1.

INTR~OUCTI~N

MATRIX OR analyte modification is the addition of a reagent-the modifier-in excess to the solution to be measured in graphite furnace atomic absorption spectrometry (GF AAS) to convert the analyte element into a phase of higher thermostability and/or to increase the volatility of concomitants. The aim of analyte and matrix modification in GF AAS is to permit high enough pyrolysis temperatures to remove the bulk of concomitants during thermal pretreatment of the sample without losing any analyte element prior to the atomization stage. Effective analyte or matrix modification greatly reduces a number of spectral interferences due to absorption of source radiation by overlapping molecular or atomic lines of concomitants and scattering of source radiation by non-volatilized particles formed by the concomitants. It also reduces or removes volatilization and vapor-phase interferences and is an important integral part of the stabilized temperature platform furnace (STPF) concept [l] for interference-free GF AAS. The concept and the term matrix modification was introduced into GF AAS by EDIGER [2] more than a decade ago. Among the reagents, he proposed the addition of nickel to decrease the volatility of arsenic and selenium, and the addition of ammonium phosphate to stabilize cadmium. Ediger also proposed to add ammonium nitrate to sea water samples to allow removal of sodium chloride at lower pyrolysis temperatures at which the analyte element is not yet volatile. Other modifiers proposed in later years were molybdenum [3], lanthanum [4], phosphoric acid [S, 63, magnesium nitrate [7], potassium dichromate

*To whom correspondence should be addressed.

[l] [Z] [3] [4] [S] [6] [7]

W. SLAVIN,D. C. MANNING and G. R. CARNRICK, At. Spectrosc. 2, 137 (1981). R. D. EDIGER, At. Absorpt. Newslett. 14, 127 (1975). E. L. HENN, Anal. Chem. 47,428 (1975). G. MACHATAand R. BINDER,Z. Rechtsmed. 73,29 (1973). K. G. BRODIEand J. P. MATOIJSEK, Anal. Chim. Acta 69, 200 (1974). E. J. CZOBIKand J. P. MATOUSEK, Talanta 24, 573 (1977). W. SLAVIN,G. R. CARNRKKand D. C. MANNING, Chem. 54, (1982). 1157

1158

GERHARD SCHLEMMER and BERNHARD WELZ

[8] and several others. Summaries of modifiers recommended for various elements may be found in several publications [ 1, 9-l 11. To make the application of modifiers generally acceptable in GF AAS, a number of requirements are expected to be met particularly for routine work. Among those are: (i) The analyte element should be stabilized to as high a pyrolysis temperature as possible, hopefully at least lOOOC,to allow volatilization of the bulk of concomitants. (ii) The modifier should be applicable to as many elements as possible for simplicity reasons. (iii) The modifier should be available in high purity, and not contain the analyte element(s) in measurable concentrations. (iv) The modifier should not contain an element at high concentration which has to be determined at trace levels in the furnace at later time. (v) The modifier should not markedly reduce the lifetime of graphite tubes. (vi) The modifier should not produce excessive background attenuation around the wavelength of the analyte element. Unfortunately, most of the proposed modifiers do not conform with one or more of these expectations. Nickel for example, which is recommended for As, Au, Bi, Sb, Se and Te [lO-121 has frequently to be determined at very low levels so that contamination of the furnace with this modifier is not desirable. The same is true for copper which has been proposed as an alternative to nickel [ 131. Ammonium phosphate which is the recommended modifier for Ag, Cd, Pb and Sn is typically not available in high enough purity so that it has to be purified before use [14]. Lanthanum has a very corrosive effect on graphite tubes and platforms [15], and the same is true to a lesser extent also for dichromate [16] so that the application of these modifiers cannot be generally recommended. In 1981 two groups independently proposed palladium as a passible modifier in GF AAS. WEIBUST et al. [ 171 reported that, in the presence of palladium or platinum, inorganic as well as organically-bound tellurium could be heated to 1050C without losses whereas other modifiers could stabilize only one form or the other. JIN and NI [ 183 used palladium as the
modifier for the determination a pyrolysis temperature of bismuth in waste water, sea water and urine and found that

of up to 1200C could be applied. In the following years there were reports of the use of palladium as the modifier for the determination of arsenic in soil, coal fly ash, biological samples [ 191 and environmental samples [20], of thallium in waste water [Zl], of mercury [22,23] and of indium in minerals, river sediments and coal fly ash [24]. Since the application of palladium as a modifier had already been published for six elements, the question arose if it could be used as a generally applicable modifier for the elements of Groups III A to VI A. Palladium is certainly not a frequently determined element, it can be obtained in high purity, it does not reduce tube lifetime and does not produce any excessive background attenuation. It would therefore fulfill all expectations if stabilization to

[8] G. F. KIRKBRIGHT, S. HSIAO-CHUAN and R. D. SNOOK,At. Spectrosc. 1, 85 (1980). [9] B. WEU, Atomic Absorption Spectrometry, 2nd Edn. VCH, Weinheim (1985). [lo] W. SLAVIN, G. R. CARNRICK, D. C. MANNING and E. PRUSZKOWSKA, At. Spectrosc. 4, 69 (1983).

[ll] [12] [13] [la] [15] [16]

[17]

[18] [19] [20] [21] SHANXIAO-QUAN, NI ZHE-MING and ZHANGLI, Talanta 31, 150 (1984). [22] Z. GROBENSKI, W. ERLER and U. VOELLKOPF, At. Spectrosc. 6, 91 (1985). [23] L. PING,K. FUWAand K. MATSUMOTO, Anal. Chim. Acta 171, 279 (1985). [24] SHANXIAO-QUAN, NI ZHE-MING and YVAN ZHI-NENG,Anal. Chim. Acta 171, 269 (1985).

Techniques in graphite furnace atomic absorption spectrophotometry, Perkin-Elmer Part No. 0993-8150, April (1985). J. BAUSLAUGH, B. RADZIUK, K. SAEEDand Y. THOMASSEN, Anal. Chim. Acta 165, 149 (1984). B. WELZ, G. SCHLEMMER and U. VOELLKOPF,Spectrochim. Acta XIB, 501 (1984). B. WELZand G. SCHLEMMER, J. Anal. At. Spectrom. 1, 119 (1986). B. WELZ,A. J.CURTIUS,G.SCHLEMMER, H. M. ORTNER and W. BIRZER, Spectrochim. Acta 41B, 1175 (1986). B. WELZ, G. SCHLEMMER and H. M. ORTNER, Spectrochim. Acta 41B, 567 (1986). G. WEIBUST, F. J. LANGMYHR and Y. THOMASSEN, Anal. Chim. Acta 128, 23 (1981). JIN LONG-ZHUand NI ZHE-MING,Can. J. Spectrosc. 26, 219 (1981). SHANXIAO-QUAN, NI ZHEMINGand ZHANGLI, Anal. Chim. Acta 151, 179 (1983). SHANXIAO-QUAN, NI ZHE-MING and ZHANGLI, At. Spectrosc. 5, 1 (1984).

Pd modifier GF AAS

1159

high enough pyrolysis temperatures can be obtained, as indicated by preliminary results. The aim of this work was to investigate the extent to which palladium could be recommended as a generally applicable modifier for a larger group of elements. It has been shown recently that mixed modifiers with magnesium nitrate are superior to those without magnesium, at least when it comes to the analysis of biological materials [25-271, so we also investigated mixtures of palladium and magnesium nitrates. 2. EXPERIMENTAL
A Perkin-Elmer Zeeman/3030 Atomic Absorption Spectrometer with HGA-600 Zeeman graphite furnace and an AS-60 furnace autosampler was used throughout this work. Electrodeless discharge lamps were used for all elements investigated in this work. Wavelength, slit width and lamp power were set according to the manufacturers recommended conditions [ 111. Time resolved absorption pulses were plotted with a Perkin-Elmer PR-100 printer. Integrated absorbance values computed by the AA instrument were used exclusively for signal evaluation. The temperatures reported were read from an Ircon Modline II i.r. pyrometer. Pyrolytic graphite coated tubes (Perkin-Elmer Part No. B 010-9322) with inserted pyrolytic graphite platforms (Lvov Platform, Perkin-Elmer Part No. B 010-9324) were used for all experiments. A typical temperature program used for most of the determinations is given in Table 1. The optimum pyrolysis (step 3) and atomization (step 4) temperatures recommended for the individual elements are given in Section 3. All reagents used were of the highest purity available and at least of analytical reagent grade. Nitric acid was further purified by subboiling distillation in a quartz subboiling still (Kuerner Analysentechnik, Rosenheim F.R.G.). Standard stock solutions were prepared from Titrisol concentrates (Merck, Darmstadt, F.R.G.) according to the manufacturers recommendations. Reference solutions were prepared daily by further dilution with 0.2 y0 (w/v) nitric acid. Palladium modifier was prepared by dissolving palladium metal powder (Alpha Products) in the minimum amount of concentrated nitric acid and further dilution with deionized water. Magnesium nitrate modifier was prepared by dissolving an appropriate amount of Mg(N03b. 6 HzO(Merck, suprapure) in deionized water. The mixed palladium and magnesium nitrates modifier was prepared by mixing equal volumes of a solution containing 3000 mg l- of Pd and 2000 mg l- of Mg(NOsh. Ten ~1 of this solution were typically added to the solution of measurement, corresponding to masses of 15 pg of Pd (or 32.5 pg Pd(NOJb and 10 pg of Mg(N03)*. The following reference materials were used in this work: NASS-1 Seawater Reference Material for Trace Elements (National Research Council Canada, Ottawa); Oyster Tissue, NBS Standard Reference Material No. 1566 (US Department of Commerce, National Bureau of Standards, Washington DC); Seronorm Trace Elements Human Serum Batch 105 (Nyegaard, Oslo, Norway). About 0.5 g of the oyster tissue were decomposed with 10 ml of nitric acid (1 + 1) for 3 h at 140C under pressure in a Perkin-Elmer Autoclave-3. The lyophilized human blood serum was dissolved in water according to the recommendations, and diluted 1 + 3 with 0.05 y0 (w/v) Triton X-100 solution.

Table 1. Temperature program used for the determination of all elements with the Pd/Mg(NOa)r modifier Internal gas flow (ml min- ) 300 300

Step No. 1 2 3

Furnace temp. (C) 90 120

VU

Time (s) ramp hold 1 10 10 0 1 1 9 30

Read -

30
4 6 5

300
0 300 300

* -

4
5 6

var 2650 20

[25] E. PRUSZKOWSKA, G. R. CARNRICK and W. SLAVIN, At. Spectrosc. 4, 59 (1983). [26] E. PRUS~KOWSKA, D. C. MANNING,G. R. CARNRICK and W. SLAVIN, At. Spectrosc. 4, 87 (1983). [27] G. R. CARNRICK, D. C. MANNING and W. SLAVIN, Analyst 108, 1297 (1983).

1160

GERHARD~HLEMMER

and BERNHARD WELZ

3. RESULTSANDDISCUSSION 3.1. General parameters The thermal pretreatment and atomization curves for the nine elements investigated in this study, using the proposed palladium and magnesium nitrates mixed modifier, and the curves without a modifier, are depicted in Fig. 1. The data have been blank corrected. It becomes apparent from these curves that the proposed modifier has a quite substantial stabilizing effect on all of the investigated elements. This does not only include a higher thermal pretreatment temperature but also a shift in the appearance and in the optimum atomization temperature. More detailed information on the maximum applicable pyrolysis temperatures for a lossfree thermal pretreatment is given in Table 2. Generally higher thermal pretreatment temperatures can be used with palladium and magnesium nitrates than with the typically recommended individual modifiers.
1 Arsenic

I-

Selenium

0.2

Indium d ;Tp

Antinwny

I 0.2 1 0.1 0 300 900

Lead

Bismuth

[Cl Fig. 1. Thermal pretreatment and atomization curves for nine elements without the use of a modifier (curves a and b) and with the palladium and magnesium nitrates modifier (curves c and d).

Table 2. Maximum pyrolysis temperatures for loss-free thermal pretreatment of the analyte element with and without modifiers Maximum pyrolysis temperature (C) no modifier 200 600 700 500 900 200 800 700 600 Ni* Ni* individual modifier [ 1l] 1300 900 900 1100 900 800 1000 600 Pd(NWz only 1200 [20] 1200 [18] 1200 [24] Pd(NW, + Mg(N% )z 1200 1200 1200 1000 1300 1100 1400 1200 900

Element As Bi In Pb Sb Se Sn Te Tl

PO, + Mg+
Ni* Ni + Mgt

PO4+ Mg+
Ni* H2S04

1050 [17] 1000 [21]

* 0.02 mg Ni as the nitrate. TO.2 mg NH4H2P04 or (NH&HP04 + 0.01 mg Mg(NO,b $0.02 mg Ni as the nitrate + 0.025 mg Mg(NOsb. 61% (v/v) H,SO,.

Pd modifier GF AAS With 0.02 mg

1161

of nickel as the nitrate, arsenic can be stabilized very well so that pyrolysis of 1300 to 1500C [Z, 10, 111 can be used. The same modifier is typically recommended also for the other elements of Group V A, but the stabilizing effect decreases substantially with increasing atomic weight of the analyte. There is only a small gain in the pyrolysis temperature for bismuth in pure solutions when nickel is used. A stabilizing effect is obtained for this modifier, however, ~bismuth has to be determined in complex matrices. The palladium and magnesium nitrates mixed modifier, in contrast, stabilizes all three elements essentially to the same extent so that pyrolysis temperatures of 1200-1300C can be used without a problem. Nickel was also recommended as a modifier for the elements of Group VI A, selenium and tellurium [2, lo], but a mixture of copper and magnesium nitrates was found to be superior for selenium [133. Using these modiiiers, pyrolysis temperatures of 900-1OOOYJcan be applied for both elements. An even better sibilation is obtained with the stadium and magnesium nitrates mixed modifier so that approximately 200C higher thermal pretreatment temperatures can be used. A primary or secondary ammonium phosphate and magnesium nitrate mixed modifier is typically recommended for the Group IV A elements tin and lead [ 10,25,26], but the stabilizing power of this modifier is not very great. Particularly the 800C maximum pyrolysis temperature for tin is not sufficient for many applications. With the palladium and magnesium nitrates mixed modifier, however, up to 1400C can be applied for thermal pretreatment of this element, and at least 1000C for lead. Essentially no useful modifier has been found up to now for the heavier elements of Group III A, indium and thallium, and their relatively high volatility has frequently hampered their successful determination with GF AAS. The palladium and magnesium nitrates mixed modifier permits the use of substantially higher pyrolysis temperatures of 1200C for indium and 900-1000C for thallium. Besides the stabilization to higher thermal pretreatment temperatures the application of a modifier usually permits the use of higher atom~tion temperatures. The effect of the palladium and magnesium nitrates mixed modifier on atomization temperature in comparison with the recommended individual modifiers [ 1l] is not uniformly in one direction, as can be seen in Table 3. The optimum atomization temperatures for arsenic and antimony, which are fairly high if nickel is used as the modifier, can be lowered to 2100C. Those for indium and thallium, which are the lowest amongst the elements investigated, are raised by several hundred degrees with the application of the palladium and magnesium nitrates mixed modifier. It should be added that the optimum atomization temperatures given in Table 3 do not necessarily reflect the temperature at which the highest integrated absorbance is obtained. These temperatures are also optimized with respect to the shape of the atomization pulse
temperatures

Table 3. Optimum atomhtion

temperatures using different modifiers

Atomization temperature (Y)

IlO

Element AS Bi In Pb Sb Se Sn Te Tl

modifier 1900 1400 1600 1400 1800 1900 2000 1600 1300

individual modifier [ 1l] Ni* Ni* PO,+Mgt Ni* Ni+ Mgt P04fMgt Ni* HLS04 2300 1900 1400 1800 2400 2100 2100 2ooo 1400

Pd(NO3 )2+ Mg(NO& modifier 2100 2ooo 2100 1900 2100 2100 2100 1900 1600

For footnotes see Table 2.

1162

GERHARD SCHLEMMER and

BERNHARD WELZ

0.5

zok.o
Element As Bi In Pb Sb Se Sn Te Tl individual modifier [ 111 Ni* Ni* PO1+Mgt Ni* Ni+ Mgt POd+Mgt Ni* t-W%

Time

( s)

Fig. 2. Optimization of the atomization temperature for antimony using time resolved signal display. (a) 2000C; integrated absorbance 0.186 As (b) 2100C; integrated absorbance 0.176 As.

Table 4. Characteristic mass values obtained with different modifiers Characteristic mass m. (pg) (0.0044 As)- Pd(NG& + Mg(NWz modifier 14 32 14 12 25 25 26 12 15

17 24 12 12 38 30 23 1.5 17

For foot notes see Table 2.

which should be fairly symmetric with minimum tailing, and it should return to the baseline within 2 to 4 s after its appearance. This is shown in Fig. 2 for antimony for which a slightly higher integrated absorbance is obtained at 2OOOC,but the tailing is substantially reduced at 2100C which was then accepted to be the optimum temperature. Table 4 shows a comparison of the characteristic mass data published for the nine elements with their individual modifiers and the values obtained with the palladium and magnesium nitrates mixed modifier. For the majority of the elements the characteristic mass values appear to be largely inde~ndent from the modifier used. Only for antimony does the palladium and magn~ium nitrates mixed modifier result in a clearly better. value for mo, whereas a somewhat higher value is obtained for bismuth. The reason for the latter effect is most probably in the shift to higher atomization temperatures effected by the palladium modifier. It can be seen in Fig. 1 that all atomization curves exhibit a more or less pronounced maximum. The drop to lower integrated absorbance values at higher temperatures is most likely caused by increasing diffusional losses of analyte atoms under these conditions.

3.2. Application examples If a new modifier is r~ommended, it is particularly important that its stabilizing power for the analyte element persists also in the presence of a complex matrix and of high

Pd modifier GF AAS

1163

total salt concentrations. The palladium and magnesium nitrates modifier was therefore applied to a number of samples that had to be analyzed in this laboratory. The determination of lead in seawater by GF AAS causes considerable trouble and is typically carried out via extraction. One reason for that is the low concentration of lead which is usually found in unpolluted offshore waters. Another reason is, however, the lack of a suitable modifier which can stabilize the analyte element to a pyrolysis temperature at which sodium chloride is sufficiently volatile. HALLIDAY et al. [28] used the ammonium nitrate modifier proposed by EDIGER et al. [29] to remove sodium chloride at low temperatures. Ammonium nitrate, however, cannot stabilize lead so that a maximum pyrolysis temperature of only 500C can be used. Ammonium phosphate mixed with magnesium nitrate, the recommended modifier for lead [l l] allows pyrolysis temperatures of up to 900C which is, however, too low to remove sodium chloride effectively. The palladium and magnesium nitrates modifier permits the use of a pyrolysis temperature of at least 1OOOC for lead which should be enough to volatilize sodium chloride without losing the analyte element. Because of the very low lead content of open ocean water samples, a sea water reference material (0.039 f 0.006 pg 1-l Pb) was spiked with 1, 5 and 10 pg 1-l lead. The standard furnace program given in Table 1 was used with a pyrolysis temperature of 1OOOC and an atomization temperature of 19OOC, i.e. the optimum conditions worked out for reference solutions. The sensitivity obtained for lead in sea water was not significantly lower than in the aqueous reference solutions which indicated the absence of major interferences. A typical signal for lead and the simultaneously recorded background attenuation are shown in Fig. 3. The detection limit (30) for lead was calculated to be around 1 pg l- l in sea water. This may not be enough to determine the low concentrations of this element in the open ocean. It is certainly good enough, however, for screening purposes and the analysis of polluted shoreline waters. Selenium is analytically one of the most difficult elements in GF AAS. Different oxidation states of selenium can for example exhibit substantially different thermal stability, depending upon the modifier used [13]. This may be an explanation for the remaining problems in the determination of this element in biological materials observed by CARNRICK et al. [27] even with the most up-to-date graphite furnace technology. On the other hand there is the rapidly increasing importance that this element has gained in biochemistry and in the environment which is accompanied by the need for sensitive methods and reliable methods of its determination in a variety of materials. The effectiveness of the palladium and magnesium mixed modifier was investigated for the determination of selenium in human blood serum and NBS oyster tissue standard reference material. If matrix-free solutions are investigated, palladium alone has essentially the same stabilizing power as the mixture with magnesium nitrate (compare Table 2). In real samples, particularly when biological materials are investigated, the addition of magnesium nitrate becomes essential to obtain the same sensitivity (characteristic mass) in sample and reference solutions. In spite of the magnesium nitrate, slight selenium losses were observed in biological materials if pyrolysis temperatures above 900C were used. A modified temperature program,

AA0.2 BG2.0

Time

(s)

Fig. 3. Direct determination of lead in sea water. NASS-1 standard reference material, spiked with 10 Fgl- Pb. Dotted line is background attenuation. [28] M. C. HALLIDAY, C. [29] R. D. EDIGER, G. E.
HOUGHTON and J. M. OTTAWAY, Anal. PETERSON and J.

D.

KERBER, At.

Chim. Acta 119, 67 (1980). Absorpt. Newsleft. 13,61 (1974).

GERHARD SCHLEMMER and BERNHARD

WELZ

Table 5. Temperature program for the determination of selenium in

biological materials using the Pd/Mg(N03)*

Internal gas flow (ml min-)

modifier

Step No. 1

Furnace temp. (C) 100

Time (s) ramp hold 1 40

Read

2 3 4 5 6

180 850 2100 2650 20

10 10 0 1 1

30 30 4 6 5

300 300 300 0 300 300

* -

Fig. 4. Analytical curves for selenium in aqueous reference solutions and added to blood serum. (0) solutions prepared from selenite; ( x ) solutions prepared from selenomethionine.

which is given in Table 5, was therefore used for the determination of selenium in biological materials. The analysis of the NBS oyster tissue standard reference material, which was decomposed with nitric acid under pressure, was straightforward and without problems. The characteristic mass for selenium was the same within experimental error for the matrix-free reference solutions and for selenium added to the sample solution. The result obtained for selenium, 1.82kO.06 egg-, is well within the certified range of 2.1 f 0.5 pgg- i. Human blood serum was not externally decomposed but only diluted 1 + 3 with 0.05 % (w/v) Triton X-100 solution, injected into the graphite tube and the modifier solution was added. As can be seen in Fig. 4 there was typically a 10 y0 difference in the slope of the curves, i.e. the characteristic mass for selenium in matrix-free reference solutions and in the serum samples, respectively. This difference was independent from the form in which selenium was added, as inorganic selenite or as organically bound selenomethionine. It has been shown that the difference between these two selenium compounds can be very pronounced for other modifiers [13]. If the small difference in the characteristic mass is taken into consideration, a selenium content of 88 pg l- 1 is obtained in the blood serum sample. This is in good agreement with the accepted value of 90.7 + 6.0 pg l- 1 obtained in an interlaboratory trial [30]. Substantially lower, wrong results were found in earlier work [31] for the same material but with other modifiers.
[30] M. IHNAT,M. S. WOLYNETZ,Y. THOMASSEN and M. VERLINDEN, Pure Appl. Chem. in press (1986). [31] B. WELZ, M. MELCHERand G. SCHLEMMER, Fresenius 2. Anal. Chem. 316, 271 (1983).

Pd modifier GF AAS

1165

4. CONCLUSION A mixture of ~lladium and magnesium nitrates has been shown to be a powerful modifier for nine elements of Group III A through Group VI A of the periodic table. Preliminary results obtained in our laboratory, and reported by others, show good promise that this modifier could be used for even more elements, including those of Group I B and Group II B. This would mean that the palladium and magnesium nitrates modifier could possibly replace almost all the other modifiers recommended up to now except perhaps for the magnesium nitrate modifier proposed for several transition elements [7]. This would certainly simplify graphite furnace AAS in routine application. The palladium and magn~ium nitrates modifier permits use of thermal pretreatment temperatures of at least 900-1000C for all investigated elements. For most elements this modifier enables the application of substantially higher pyrolysis temperatures than the modifier recommended previously [ 111. These higher pyrolysis temperatures allow eff~tivecharring of biological matrices and removal of most inorganic concomitants prior to analyte element volatilization. The palladium and magnesium nitrates modifier has a substantial equalizing effect on the atomization temperature of the nine elements investigated. The optimum atomization temperature for all but one element (thallium} is between 1900 and 2100C. This means that all elements can be determined at a %ompromize atomization temperature of 2100C with a minimum sacrifice in sensitivity (compare Fig. 1). Such uniform conditions for as many eiements as possible are of vital importance if simultaneous multielement furnace techniques are envisaged, such as the SIMAAC [32] or the FANES [33 J system. But also in conventional GF AAS uniform conditions for a number of elements can greatly facilitate and simplify daily routine analysis. The mechanism of stabilization of the palladium and magnesium nitrates modifier has not been investigate in detail. It is known, however, that p~ladium nitrate decomposes via the oxide to the metal at 870C which melts at 1552C [343. The appearance temperature for palladium in a graphite furnace is around 1250C [35]. As most of the investigated elements are stabitized to temperatures around 1200C it can be assumed that the modifier acts by inbedding the analyte into the palladia matrix or even by the forming a kind of an alloy with the analyte. More work has to be done, however, before definite claims can be made on the mechanisms involved with this modifier. It has only recently come to our attention that Shan and Ni used Pd as a modifier in the determination of mercury [36] in a paper published in 1979, and in a later publication [37] for antimony, arsenic, bismuth, selenium, silver and tellurium.

[323 J. M. HARNLY, N. J. MILLER-~~~1and T. C. OHAVER,S~c~~b~. Actu 398,305 (1984). [33] H. FALK, E. HOFFMANNand CH. LODKE, Spectrochim. Acta 39B,283(1984). [34] R. C. WEASTand M. J. ASTLE, Eds, Handbook of Chemistry and Physics, 61st Edn. CRC Press, Boca Raton
(1980). [353 G. SCHLEMMER and B. WELZ, Fresenius 2. Anal. Gem. 232,703 (1986). [36] SHANXIAO-QUAN and Nr ZHE-MING,Acta Chin. Sinica 37,261 (1979). [373 SHANXIAO.QUAN and NI ZHE-MING,Acta Wm. Sinica 39, 575, (1981).