ANALYST. FEBRUARY 1991, VOL. 116 14.5 Proposed Mechanism for the Action of Palladium and Nickel Modifiers in Electrothermal Atomic Absorption Spectrometry Anatoly Volynsky* V.I. Vernadsky Institute Qf Geochemistry and Analytical Chemistry, USSR Academ y of Sciences, 79 Kosygin Street, I 7 7975 Moscow, USSR Sergei Tikhomirov and Anatoly Elagin All-Union Research Institute of Organic Synthesis, 72 Radio Street, 107005 Moscow, USSR By the use of Fourier transform infrared spectrometry it was found that palladium chloride decreased the temperature of the reduction of PbO and Ga203 with graphite; nickel chloride only catalysed the reduction of Ga203. A hypothesis is proposed that nickel and palladium salts promote the reduction of compounds (in atomic absorption measurements) at relatively low ashing temperatures.The resultant free elements form intermetallic compounds or solid solutions with metallic Pd and Ni, thus reducing or almost eliminating loss of analyte due to sublimation of halides, oxides, dimers and other compounds. The high efficiency and universal action of Pd modifiers are because the Pd metal can be easily formed from its compounds and also by the unique catalytic properties of metallic Pd. Keywords: Electrothermal atomic absorption spectrometry; palladium modifier; nickel modifier; mechanism of action; Fourier transform infrared spectrometry The compounds of nickel' and palladium and platinum2 were introduced as modifiers in 197.5 and 1979, respectively. At present, nickel compounds, platinum group metals (PGMs) and various different mixtures based on these are widely used for the determination of a great number of elements in various Palladium and nickel are known to form inter- metallic compounds and solid solutions with the determined elements in the graphite furnaces.G9 The formation of such compounds causes an increase in the maximum permissible ashing temperature during the determination of elements of high- and mid-volatility.However, the mechanism of forma- tion of the intermetallic compounds and solid solutions in graphite atomizers, when using such modifiers, is still rather vague. I t is clear that Pd interacts in its metallic form. According to Morikawa rt al. , I 0 activated carbon partially reduces palladium chloride to the metal at room temperature. However, virtually all other elements detectable by elec- trothermal atomic absorption spectrometry (ETAAS) (with the exception of the noble metals), exist in the graphite atomizer in the form of oxides, chlorides" or other salts, within the temperature range 300-800"C.It is still not clear why Pd is the most efficient modifier in the majority of instances .s. 12 The supposition has been made,".l3 that Ni and PGM modifiers catalyse some processes that occur in graphite atomizers for ETAAS. It might be that in the first stage of the process of the formation of solid solutions and intermetallic compounds, the graphite of the atomizer catalytically reduces the oxide analytes at low temperatures.14 Nickel and PGMs Table 1 Initial (T,,,) and final ( Tf) temperatures of the reduction of Pb and Ga oxides with graphite PbO Gal03 Experimental conditions T,,I"C T,I"C T,,l0C T+l°C Without catalyst 430 740 735 990 present 520 780 600 840 NiCI2.6H20 PdClz present 340 500 360 720 * Present address: Laboratory of Organic Analysis. Department of Chemistry.Moscow State University, 119899 Moscow, USSR. are known as efficient catalysts for the reduction of the oxides of Mo, V, Cu, Sn, Re, Pb, W, Fe and Ni (other oxides have not yet been studied) with hydrogen, carbon monoxide and some hydrocarbons. By the use of X-ray photoelectron spectrometry,17 it has been found that lead chloride is thermostable in a graphite atomizer at up to 600°C; at higher temperatures it sublimes, whereas, in the presence of palladium chloride, metallic Pb is already apparent at 200°C.This paper describes the investiga- tion of the reduction of Pb and Ga oxides with graphite in the temperature range 100-1000 "C using Fourier transform infrared (FTIR) spectrometry. Experimental Apparatus For the identification of the gaseous products of the reactions, a Bruker FTIR spectrometer, Model IFS-113~. with a gas chromatographic interface was used. A JEOL pyrolyser, Model PL-722, was connected to the interface and the carrier gas flow (Ar, 30 ml min - 1 ) was controlled by a Carlo Erba Fractovap 4200 chromatographic block. The carrier gas was additionally purified, for traces of water vapour and free oxygen, with the aid of a Supelco high capacity gas purifier. The temperature of the pyrolytic oven was controlled by a thermocouple and registered on an analogue recorder with a 10 mV scale.Procedure Electrographite was pounded in a vibration mill. Just before the experiment, the graphite powder was heated in an Ar atmosphere for 40 min at 800°C to remove adsorbed gases. The pre-heated graphite powder was mixed with the oxides (PbO, 3.5 mg; Ga203, 1.6 mg) with a mass ratio of approximately 5 : 1. The catalyst (PdC12 or NiC12-6H20) was added to the reaction mixture in an atom ratio of carbon to metal of approximately 2 5 : 1 . The maximum mass of the reaction mixture (about 25 mg) was limited by the size of the quartz crucible. The oxide mass in the mixture resulted in absorbance values of up to 0.3 during the measurements. The identification of the gaseous products of the reaction was performed within the following spectral windows: 2200-2 100 cm- 1 for carbon monoxide; 2380-2300 cm- 1 for carbon146 ANALYST, FEBRUARY 1991, VOL.116 dioxide; 1700-1500 cm-1 for water; and 1860-1800 cm-1 for phosgene. The sensitivity for carbon dioxide is about 1.5-fold higher than that for carbon monoxide. Results As can be seen from Table 1, the reduction of lead monoxide in the absence of a catalyst starts at 430°C. Below this temperature carbon dioxide is formed [Fig. 1(a)] due to the decomposition of the traces of PbC03 (tdec = 315"C).18 The preliminary heating of PbO (10 min at 400°C in an Ar atmosphere) , although significantly reducing the area of this peak, fails to eliminate it completely. The interaction of Ga203 with graphite starts when the temperature is raised to 735 "C (Table 1).The erratic baseline shown in Fig. l(b) is due to the temperature fluctuation of the light-pipe.19 The effect appears to be pronounced in this particular instance because of the relatively high expansion of the ordinate axis. The addition of nickel chloride substantially decreases the temperature of the reduction of Ga203 with graphite (Table 1) and changes the mechanism of the process towards the formation of carbon dioxide (Fig. 2). Nickel chloride does not produce a catalytic effect on the reduction of lead monoxide with graphite (Table 1). No traces of phosgene or water vapour are registered within the temperature interval exam- ined (Fig. 3). Obviously, chlorine from nickel chloride evolves in the form of HC1,2" but absorption bands due to HCI (3000-2800 cm-1) lie beyond the limits of the interval examined.Water contained in NiCI2.6H20 apparently ad- sorbed on to the graphite starts to interact (at 80OoC) forming carbon dioxide (Fig. 2). Palladium chloride sharply decreases the temperature of the reduction of the metal oxides with the graphite (Table 1). It can be seen that the C02 evolution is entirely complete in 3 min [Fig. 4(a)]. Carbon monoxide formation during the reduction of Ga203 becomes negligible [Fig. 4(b)]. A special 'blank' experiment was performed by heating the powdered - 40 4- .- 8 3 ' 30 h 2 ; 20 2 2 10 a 4- .- 0 m 800 0 600 2 I a 400 $ F % 200 0 6 12 18 24 Time/mi n 0 8 16 24 32 Ti rne/m i n graphite with palladium chloride in the absence of both lead and gallium oxide and neither CO2 nor CO formation was detected.Discussion The results obtained corroborate the theory of the catalytic action of some modifiers on the reduction of the oxides of certain metals by graphite. It is necessary to note that these results have been obtained using electrographite powder as the reductant. Pyrolytic graphite is a much more inert material than electrographite, therefore, the reduction with pyrolytic graphite proceeds significantly slower. Hence, in the most popular graphite tubes with a pyrolytic graphite coating the catalytic effect should be more discernible. The direct experimental verification of this proposition is impossible by the procedure used. The main peculiarities of the pyrolytic graphite are the high degree of crystalline order and low concentration of active sites on its surface.21 After severe pounding these peculiarities are lost.Another difference of our experimental conditions from those used for ETAAS is the amount of lead and gallium oxides used, i.e. , mg instead of the pg and ng levels which are typical for ETAAS. The chemical and physical properties of clusters usually differ significantly from the properties of the bulk materials. This is a typical problem in the study of the processes that occur in the graphite tubes used in ETAAS, excluding the instances of usage of mass spectrometry or radioactive isotopic analysis. Thus, the temperatures of the reduction in the graphite tubes might differ from those obtained in our experiments. We considered the principal processes occurring in graphite atomizers in the presence of the PGM modifiers with palladium chloride as an example.When a reductant (e.g., the graphite of the atomizer) is present, palladium chloride 800 u B 600 2 h 4- 400 al I- 200 I I 1 I 8 16 24 32 Interaction of graphite with Ga203 in the presence of Timehi n Fig. 2 NiC12.6H20. 1, CO2: and 2, CO 0.12 800 0 B 2 20.08 2 600 2 Q) a m 400; 2 200 <0.04 0 2300 21 00 1900 1700 Wavenurn berlcrn-' Fig. 1 Interaction of graphite with ( a ) PbO and ( b ) GazO3. 1, CO,; and 2, CO Fig. 3 Ty ical IR spectrum of a mixture of Ga203, graphite and NiC12-6H28; T = 680 "CANALYST, FEBRUARY 1991. VOL. 116 147 Table 2 Thermal properties of the most thermostable chlorides and oxides of the PGMs and nickells v) C c .- al 0 L m 2 0 2o 'E z 0 6 12 18 24 Ti me/mi n A v) C 3 w .- 2 45 E a - k 30 c .- a, C m 15 5: 0 s 800 0 % 600 2 E 400 2 al I- 200 800 600 $ E 400 9 .: I- 200 8 16 24 32 Tim e/m i n Fig. 4 presence of PdCI,. 1, CO?; and 2, CO Interaction of graphite with ( a ) PbO and ( b ) Ga203 in the decomposes at the relatively low temperatures used during the ashing stage. The metallic Pd produced catalyses the reduction of the lead and gallium oxides and the lead chloride17 with the graphite. The reduced metals dissolve in the Pd forming intermetallic compounds7-9 or solid solutions.6 Such a process is promoted in the graphite tube due to the amount of the modifier being 100-1000-fold higher than the amount of analyte. The increase in the sensitivity of the determination in the presence of such modifiers is caused by the formation of compounds of low volatility at the relatively low ashing temperatures.This formation decreases or eliminates any losses of analyte due to the sublimation of volatile chlorides, 1722 oxides,7,'3 dimers,7 hydrides'4 and other com- pounds. Sometimes the catalyst not only decreases the temperature of the reduction for the compounds being determined, but also simultaneously changes the mechanism of the reduction. Without the catalyst, the products of Ga203 reduction are CO [Fig. l(h)] and possibly Ga202s or Ga0.26 In the presence of NiCI? and especially PdCI2 the main gaseous product of the reaction is C 0 2 [Figs. 2 and 4(b)]. The marked increase in the maximum ashing temperature and sensitivity for gallium26 suggests that in the presence of Pd and Ni modifiers, Ga203 is reduced to non-volatile products, e.g., to the free metal.This supposition is verified by the absence of GaO in the electrothermal atomizer gas phase at the ashing stage in the presence of Ni compounds.26 Nickel forms compounds that are more thermostable than those of the PGMs (Table 2). The lack of elemental Ni in the reaction mixture at 300400°C is probably the main reason for the absence of the catalytic effect of nickel chloride on the reduction of PbO. According to thermodynamic calculations, nickel chloride is stable in an argon atmosphere in the graphite atomizer in the temperature range 230-630"C.24 Free hydrogen can reduce it to the metal at these temperatures; however, below 1000°C the reaction between graphite and water, which produces free hydrogen, proceeds slowly with- out a catalyst.Nickel in the form of its compounds does not act as a catalyst for this p r o ~ e s s . ' ~ At 600°C thermohydrolysis results in transformation of the nickel chloride into the oxide .?O Chloride TI"C Oxide TPC PdCI? dec. * 500 PdO m.p. 870 PtCl, dec. 581 PtO dec. 550 dec. 1100 RuCl3 dec. >SO0 IrOz IrC12 dec. 773 Rhz03 dec. 1100-1150 m.p. 1984 RhCl3 dec. 450-500 NiO NiC12 subl. t 973-987 * Dec. = decomposes. ? Subl. = sublimes. Data from the literature, on the conditions for metallic Ni formation from the nitrate and oxide in the graphite atomizer, seem to be contradictory. According to thermodynamic calculations, nickel oxide is stable in the atomizer at tempera- tures between 230 and 630°C under a partial pressure of free oxygen of 1 x 10-6 bar.24 By the use of mass spectrometry it has been shown that metallic Ni is formed in the graphite atomizer in substantial amounts only at about 15OO0C.7.23 It has been found by thermal and X-ray diffraction analysis that graphite reduces Ni(N03)2.6H20 to Ni metal in an argon atmosphere at 1 15OoC.'8 Catalysts significantly decrease the reduction temperature (down to 630°C for CUCI).'~ At the same time less than 1% of the nickel nitrate is reduced to nickel dicarbide at 330 "C.23 Nickel nitrate is partially reduced to the free metal at relatively low ashing temperatures; thus it is necessary to introduce 1000 times as much Ni into the graphite atomizer as Pd in order to create a concentration sufficient for the effective catalysis of the reduction of the analyte.12 In our experiments it would appear that significant amounts of catalytically effective metallic Ni were formed from NiCl2.6H20 at temperatures between 520 and 600 "C (Table 1). Nickel metal catalyses the reduction of Ga203 with graphite to free Ga, and the oxidation of graphite with water vapour. It is known that graphite readily reacts with water at 800°C in the presence of metallic Ni.27 Nickel metal and nickel oxide are also efficient catalysts for the oxidation of carbon monoxide to carbon dioxide; hence, carbon dioxide is the final product of graphite oxidation (Fig. 2). Mixed oxide formation is a possible mechanism for the stabilization of highly volatile elements in the presence of nickel nitrate,.at ashing temperatures below 600°C. Such compounds are known for Ga, TI, Ge, As, Sb, Bi, Se and Te; ZnO and NiO form solid solutions.30 At elevated tempera- tures these compounds can be reduced to the corresponding intermetallic compounds. Evidently such a mechanism for stabilization is marginally possible using nickel chloride as the modifier. The high efficiency and universality of Pd modifiers can be explained not only by the ease of formation of Pd metal from its compounds but also by the unique catalytic properties of Pd. In particular the catalytic properties of Pd are only slightly dependent on the particle size, because Pd electron configura- tion depends weakly on cluster size.3' The electron configura- tion of Pd2 is 4d95s'. This configuration approaches 4d9.45~0." at eight to ten atoms where only minor changes take place with increasing particle size.The configuration determined from experimental measurements on bulk Pd is 4d9.65s0.4. Sample matrix and the parameters of determination (especially the temperature and heating rate of the atomizer during the drying stage) can significantly affect the size and the structure of the catalytically effective particles, thus restricting the field of efficient use of the PGM modifiers (except Pd). Excellent results have been reported by Dahl et a1.32 on the use of a mixture of Pd, Rh and Ir compounds for Sb determination in different matrices. They can probably be explained by the successful combination of the unique catalytic properties of Pd with the high melting points of Rh148 (m.p.= 1960°C) and Ir (m.p. = 2450°C). This prevents the loss of Sb and enables ashing at extremely high temperatures. The authors thank E. M. Sedykh and G. N. Takhtarova for their valuable assistance. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 References Ediger, R. D., At. Absorpt. Newsl., 1975, 14, 127. Shan, X.-q., and Ni, Z.-m., Huaxue Xuebao, 1979, 37, 261; Chem. Abstr., 1980, 92, 220474~. Voth-Beach, L. M., and Shrader, D. E., J. Anal. At. Spectrom., 1987, 2, 45. Schlemmer, G., and Welz, B., Spectrochim. Acta, Part B , 1986, 41, 1157. Ni, Z.-m. , and Shan, X.-q., Spectrochim. Acta, Part B, 1987,42, 937. Shan, X.-q., and Wang, D.-x., Anal. Chim. Acta, 1985, 173, 315. Styris, D. L., Fresenius Z. Anal. Chem., 1986, 323, 710.Teague-Nishimura, J. E., Tominaga, T., Katsura, T., and Matsumoto, K., Anal. Chem., 1987, 59, 1647. Wendl, W., and Miiller-Vogt, G., J. Anal. At. Spectrom., 1988, 3, 63. Morikawa, K.. Shirasaki, T., and Okada, M., in Advances in Catalysis and Related Subjects, ed. Eley, D. D . , Academic Press, New York, 1969, vol. 20, p. 98. Frech, W., Lundberg, E., and Cedergren, A., Prog. Anal. At. Spectrosc., 1985, 8, 257. Brzezinska-Paudyn, A , , and Van Loon, J. C., Fresenius Z . Anal. Chem., 1988, 331, 707. Rettberg, T. M., and Beach, L. M., J. Anal. At. Spectrom., 1989, 4, 427. Volynsky, A. B., XXVI Colloquium Spectroscopicurn Interna- tionale, Sofia, 1989, Abstracts Volume I, p. 95. Il’chenko, N. I., Usp. Khim., 1972, 41, 84. Charcosset, H., and Delmon, B., Ind. Chim.Belge., 1973, 38, 481. Sakurada, O., Takahashi, H., and Taga, M., Bunseki Kagaku, 1989, 38, 407. 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 ANALYST, FEBRUARY 1991, VOL. 116 CRC Handbook of Chemistry and Physics, ed. Weast, R. C., CRC Press, Boca Raton, 68th edn., 1987. White, R., ChromatographyIFourier Transform Infrared Spec- troscopy and its Applications, Marcel Dekker, New York, 1990, pp. 57 and 58. Welz, B., Akman, S . , and Schlemmer, G., Analyst, 1985, 110, 459. Huettner, W., and Busche, C., Fresenius Z . Anal. Chem., 1986. 323, 674. Sedykh, E . M., and Belyaev, Yu. I., Prog. Anal. At. Spectrosc., 1984, 7, 373. Droessler, M. S . , and Holcombe, J. A., Spectrochim. Acta, Part B, 1987, 42, 981. Dedina, J., Frech, W., Cedergren, A.. Lindberg, I., and Lundberg, E., J. Anal. At. Spectrom., 1987, 2, 435. McAllister, T., XXVI Colloquium Spectroscopicum Interna- tionale, Sofia, 1989, Abstracts Volume IV, p. 55. Shan, X.-q., Yuan, Z.-n., and Ni, Z.-m., Anal. Chem., 1985, 57, 857. McKee, D. W., in Chemistry and Physics of Carbon, eds. Walker, P. L., Jr., and Thrower, P. A., Marcel Dekker, New York, 1981, vol. 16, p. 1. Richardson, R. T., and Rowston, W. B., Proceedings of the Second European Symposium on Thermal Analysis, Aberdeen, 1981, p. 355. Pushkarev, V. A., in Physical Chemistry of Oxides, ed. Men, A. N., Nauka, Moscow, 1971, p. 87 (in Russian). Landolt-Bornstein. Zahlenwerte und Funktionen aus Natur- wissenschaften und Technik, Neue Serie, ed. Hellwege, K.-H., Springer-Verlag, Berlin, 1975, Gesamtherausgabe, Gruppe 111, Band 7, Teil a-f. Hamilton, J. F., and Baetzold, R. C., Science, 1979, 205, 1213. Dahl, K., Martinsen, I., Salbu, B., Radziuk, B., and Thomas- sen, Y., XXVI Colloquium Spectroscopicum Internationale, Sofia, 1989, Abstracts Volume I, p. 91. Paper 0101 009J Received March 6th, 1990 Accepted July 23rd, 1990