In HG-AAS, the graphite furnace is utilized as both the hydride trapping site and the atomization cell. Early studies employed bare graphite furnaces as the trapping medium. A small amount of the pyrolytic-graphite coating was removed from the surface of the furnace to activate the surface for the trapping of As and Sb hydrides.51, 52 In general, poor trapping efficiencies for Se, As and Ge hydrides were obtained when pyrolytic-graphite coated furnaces were used rather than electrographite furnaces.40, 42, 53 Electrographite is a porous material that allows diffusion of analyte species into the material and is relatively reactive towards metals.54 The porosity and reactivity of electrographite may induce peak tailing, memory effects, or incomplete atomization of the sample. Pyrolytic-graphite coated furnaces are produced by heating an electrographite furnace in 5 % methane/argon.54 The pyrolytic- graphite forms a dense and essentially pore-less layer on the electrographite base material with a high oxidation resistance and an extremely low permeability.55 The pyrolytic-graphite reduces both diffusion of analyte into the graphite and the chemical reactivity of the graphite.54
There are a number of disadvantages for the trapping of hydrides on bare graphite furnaces; these include the need for a well-developed porous graphite structure for optimum trapping efficiency, high accumulation temperature for some elements (As and Sb), and little chance of multi-element determinations due to differences in thermal trapping parameters between elements.44 These problems can be overcome with the use of modifiers such as Pd, which act as a substrate onto which the analyte hydride can be adsorbed.44 Where used in this thesis, Pd implies Pd(s) formed by thermal decomposition of a PdII salt. The use of Pd leads to an improvement in the thermal stability of the hydride species.56 Significant improvements in sensitivity and precision for Pd treated graphite furnaces compared to bare graphite furnaces have been reported.57 Liang et al.41 observed greater than a 6-fold increase in sensitivity when
a Pd coated furnace was used rather than a bare pyrolytic-graphite coated furnace to trap AsH3. Walcerz et al.40 reported that the use of Pd resulted in greater than a 4-fold increase in
sensitivity for electrographite furnaces, and greater than a 40-fold increase in sensitivity for pyrolytic-graphite coated graphite furnaces when determining As. Walcerz et al.40 also found better sensitivity on pyrolytic-graphite coated furnaces for Sb determinations compared to electrographite furnaces.
This significant improvement in sensitivity for the Pd-based system (and other modifiers) is due to the stabilizing effect of the metal on the analyte at high temperatures. This reduces loss of volatile precursor species which leads to an increase in sensitivity,44 and higher ashing temperatures may be used which can be advantageous in reducing interferences. Furthermore, because volatilization of the analyte species is occurring at a higher temperature, there is more efficient dissociation of any molecular species. In addition, due to the increase in sensitivity observed in the presence of modifier, sample dilution can be used to decrease interferent concentrations.43 Lastly, trapping of the hydride within the graphite furnace prior to atomization eliminates problems associated with variable rates of hydride formation and evolution.43
Many different modifiers have been used for the trapping of hydrides. These include, but are not limited to, Pd,33, 38, 40-44, 53, 56, 58-61 Zr,22, 42 Ir,22, 30, 38, 48, 56, 62, 63 Pt,44, 53, 59 Rh38, 44, 56, 58, 62 Ru,38, 44, 58, 59, 62 W,22 Ni,53 Ag,59 and Ce.59.
3.1.2.1 Permanent and non-permanent modifiers
Modifiers can be generally classified into two groups: permanent and non-permanent. Non- permanent modifiers must be deposited into the graphite furnace after each atomization step. Permanent modifiers generally can be used for numerous furnace firings without deterioration of analytical performance. The elements most commonly used as permanent modifiers are Pt-group metals with high melting points (Rh, Ru, Ir).
Palladium is an example of a non-permanent modifier; its melting and boiling points are 1555
o
C and 2964 oC, respectively.64 At temperatures normally used for atomization (> 2200 oC), some Pd would be volatilized from the furnace. However, for permanent modifiers such as Ir (melting and boiling points are 2447 oC and 4428 oC, respectively64), the modifier is retained within the furnace, at temperatures normally required for atomization, because of its low
volatility. Therefore Ir is not required to be deposited after each atomization step. In addition, very stable carbide forming elements (W, Ta, Zr) can also be used as permanent modifiers.55 Zr, Ir and W permanent coatings have been used for up to 500 accumulation/atomization cycles without the need to re-deposit modifier.22 However, Garbos et al.42 reported that for a Zr modifier, only 80 cycles could be carried out before deterioration of the analytical signal was observed, presumably due to loss of Zr modifier. The lifetime of permanent modifiers depends on the type of matrix and acids used, the pyrolysis, atomization and cleaning temperatures, and the corresponding times of the individual steps.65 The quality of the graphite furnace surface may also have an effect on the lifetime of permanent modifiers.
The use of permanent modifiers offers advantages over non-permanent modifiers. These can include; longer tube lifetime, improved detection limits for some modifiers, better long-term signal stability, simpler and quicker heating programmes, higher sample throughput, and lower analytical costs. 62
Mixed modifiers have also been used.33, 48, 62 Lima et al.62 compared the use of single permanent modifiers (Rh, Ir, Ru) with mixed permanent modifiers (W-Rh, W-Ir, W-Ru), in which W was deposited first. There was no advantage in using mixed permanent modifiers for the determination of As in a water reference material. The maximum pyrolysis temperature was 1400 oC in each case. However, for the analysis of more complex solid reference materials, the maximum pyrolysis temperatures obtained for mixed-modifiers were 50 to 150 oC greater than their single modifier analogues. Furthermore, improved reproducibility and better analyte recoveries, of certified values in reference materials, were obtained when using W-Rh, W-Ir and W-Ru mixed modifiers, when compared with single Rh, Ru and Ir modifiers. Yang and Zhang33 reported a 40 % increase in sensitivity when a mixed Pd-Zr modifier was used in comparison to just Pd. In addition, the use of the mixed Pd-Zr modifier increased the pyrolysis temperature from 1300 to 1500 oC.
3.1.2.2 Pd modifiers
Palladium is a very effective modifier that can be used to stabilize many elements to several hundred degrees higher than the temperature possible when no modifier is used.60 An increase in pyrolysis temperature between 400 and 800 oC can be achieved depending on the analyte of interest.61 Palladium is still extensively used even though it is required to be
deposited after each atomization step; in some instances Pd has been reported to give better sensitivities than permanent modifiers.42, 44, 58, 59
Palladium can be used to stabilize hydride species such as AsH3 (as in the work in this thesis)
or to stabilize analytes that are injected directly into the graphite furnace from a solution. In the absence of a modifier, Volynsky and Wennrich58 found that the maximum pyrolysis temperature for As, Se and In was 500, 200, and 600 oC, respectively. However, in the presence of Pd, the pyrolysis temperature for As, Se, and In could be increased to 1300, 1200 and 1400 oC, respectively, without the lost of analyte from the furnace. Pyrolysis temperatures for As in the presence of Pd have been reported to be as high as 1500 oC.56 The temperature at which the hydride is accumulated (i.e. accumulation temperature) on the Pd surface can also be important. A variety of optimum accumulation temperatures have been reported; these range from 16041 to 600 oC.43
The change in appearance temperature (the temperature at which the analyte signal appears) when Pd is used, is believed to be due to the formation of some sort of Pd-As species.60 It is generally accepted that Pd metal acts as the modifier.61 Pre-treatment of the Pd modifier by heating to 1000 oC33 or using reducing agents such as ascorbic acid, hydroxylamine hydrochloride, and hydrogen33, 61 can reduce the metal salt to its metallic form. Voth-Beath and Shrader61 found considerable differences in the performance of the Pd modifier depending on the reduction method used. Scanning electron micrographs of the various surfaces illustrated that the size and distribution of the Pd particles on the graphite surface varied with the reduction method used.
Even though the addition of reducing agents offers some advantages, thermal reduction (or pre-treatment at high temperatures) of Pd seems to be the most common way of forming metallic Pd on the graphite furnace. Reported pre-treatment temperatures for the Pd modifier vary from as low as 100 oC60 up to 1200 oC,38, 40 with many reported different temperatures within this range.33, 43, 44, 53, 56, 58, 59, 61
Often a surfactant is added along with the modifier solution.40, 42, 53 In some instances this can improve reproducibility and sensitivity;53 presumably by promoting a more even distribution of Pd over the graphite surface.
3.1.2.3 Pd modifier mechanism
The exact mechanism of action of the Pd modifier is not very well understood. In a recent review by Ortner et al.,55 they comment that the literature is full of very different and often contradictory proposals for the mechanism of action of Pd modifiers. The main reason for these contradictory explanations is the differences in experimental conditions, and the methods used to investigate transformations of solid analyte compounds. 55, 66 Volynsky and de Loos-Vollebregt66 comment that differences in the amounts of analyte used in ETAAS and in model studies may reach 6 orders of magnitude. This may result in various reaction paths for the same analyte-modifier-graphite systems.66 Even experiments with realistic analyte to modifier mass ratios, but with an unrealistic relation to the mass of the graphite furnace system, may lead to misleading conclusions.55
It is suggested by some workers that the hydride forms an inter-metallic species or alloy with metallic Pd on the graphite surface.53, 60 However, Ortner et al.55 disagree that analyte stabilization occurs by formation of intermetallic compounds or thermally stabilized alloys. They make the comment that the modifier to analyte ratio is 1000:1 to 100000:1 in all practical cases. However, all inter-metallic compounds or thermally stable alloys are only formed at mass ratios from 1:1 to 100:1. Therefore, they conclude that such compounds cannot be formed and those that found such compounds used unrealistic analyte to modifier mass ratios. Instead, they proposed that Pd forms intercalation compounds with the graphite. The intercalated Pd metal forms strong covalent bonds with easily volatile elements, leading to their stabilization at high temperatures. The action of the modifier is not an effect of the modifier present in particles on the surface, but of activated modifier atoms in the near surface region (to a depth of approximately 10 µm).
Rettberg and Beach67 examined the effect of Pd on the absorbance signals for As, Cd, Cr, Pb, Sn and Tl. They found that when the mass of modifier was increased, the analyte appearance temperature increased to higher values. The authors conclude that given the large excess of Pd relative to analyte, this observation cannot be explained by more complete conversion of the analyte to the inter-metallic form, and therefore bulk effects may be contributing to the stabilizing action of Pd.
Sturgeon et al.44 showed that the release of the bulk of Pd from the graphite surface was approximately 500 oC earlier than release of As from the Pd surface. In addition, comparisons
of absorbance profiles for Pd in the presence and absence of a large amount of As showed that the release of Pd is delayed by 160 oC in the presence of As; this presumably reflects the formation of a less volatile Pd-As species.