ANALYST, JANUARY 1987, VOL. 112 27 Multi-element Pre-concentration by Solvent Extraction Compatible with an Aqua Regia Digestion for Geochemical Exploration Samples Ivan Rubeska UNDP, P.O. Box 7285 ADC, Mia Road, Pasay City, Metro Manila, Philippines and Benilda Ebarvia, Edita Macalalad, Dahlia Ravis and Nenita Roque Bureau of Mines and Geosciences, North Avenue, Dilirnan, Quezon City, Philippines The interference of nitric acid in the organic solvent extraction of metal iodide complexes as ion associates with the trioctylmethyl ammonium ion was investigated. The results show that the interference is caused by nitric oxide extracted into the organic solvent in the form of nitrosyl compounds and that this can be eliminated by the addition of sulphamic acid or urea. A multi-element extraction procedure was applied to geochemical exploration samples digested with aqua regia.Ag, Cd, Se, Te and TI were determined by AAS in the extracts down to 100 p.p.b. levels. Keywords : Multi-element extraction; nitric acid interference; trace element determination; geochemical samples; trioctylmeth yl ammonium chloride Geochemical exploration often involves the determination of a number of trace elements, many of which are well below the detection limits of AAS or ICP-AES. For such elements a multi-element organic solvent extraction would be an attrac- tive solution1 if it could be made sufficiently simple and could cover all the elements of interest. Most of the trace elements of interest in geochemical exploration are easily solubilised by digesting the samples with boiling aqua regia and this is probably the most widely used decomposition procedure when looking for primary mineralisation.It may be performed on a large scale with very simple equipment, i.e., test tubes, an aluminium block and a hot-plate, and even fairly resistant minerals such as pyrite and cinnabar are attacked. For samples with a high iron content, one of the most promising multi-element extraction systems is the extraction of iodide or bromide complexes into 4-methylpentan-2-one (IBMK), possibly as ion associates with long chain aliphatic polyamines or polyphosphines. Two reagents of this kind have been extensively applied, namely, trioctylmethylammonium (TOMA) chloride2-5 and trioctylphosphine oxide (TOPO) .cg Unfortunately, the extraction of iodide or bromide com- plexes is incompatible with an aqua regia digestion of the samples unless the solutions are dried and then re-dissolved in, for example, HC1.This would make the procedure prohibi- tively time and labour consuming for geochemical explora- tion. In the course of establishing analytical procedures for the geochemical exploration of epithermal gold deposits, 16 trace elements had to be determined, possibly from a single sample decomposition. Seven of these (Cu, Pb, Zn, Co, Ni, Mn and Mo) had sufficient ratios of abundances to detection limits to provide meaningful analyses by direct flame AAS of the sample digest and four after hydride (As, Bi and Sb) or cold vapour (Hg) generation. For the five remaining (Ag, Cd, Se, Te and Tl), the possibility of applying the extraction of iodide complexes as ion associates with TOMA or TOPO were explored.Reactions causing interference were identified as being due to nitric oxide and an extraction procedure appropriate for the geochemical analyses of soil, stream sediment and rock-chip samples, where some precision may be sacrificed for greater speed, was established. Experimental Sample solutions were prepared by boiling 2 g samples in a test-tube with 8 ml of aqua regia for 30 min. After cooling, the volume was adjusted to 20 ml with distilled water. The final concentration was not critical and varied between 3 and 4 M HCl. After settling, 10 ml of this solution were drawn into a polypropylene syringe fitted with a 10 cm X 1.2 mm i.d. plastic capillary tube extension to the inlet, followed by 2 ml of ascorbic acid (30%), 2 ml of a 2 M KI solution and 3 ml of IBMK containing 5% V/V TOMA or 5% m/V TOPO.After shaking for about 1 min the syringes were left to stand, and after the separation of the phases the organic layer was transferred into glass vials, which were then capped. This solution was then used for measuring several different trace elements by AAS. A Varian-Techtron Model 1475 AA spectrometer was used, either with a microsampling flame attachement or a GTA 95 graphite furnace. For flame AAS, 50 pl of the extract were delivered to the nebuliser with a micropipette. A 3 s integration time was used, giving the operator enough time to keep the whole absorbance signal within the integrating period. Ag and Cd were determined down to 0.1 p.p.m.in the sample. Se, Te and T1, which have insufficient detection limits by flame AAS, were determined in a graphite furnace using pyrolytic graphite coated tubes for Se and Te and a pyrolytic graphite platform for T1. Te and T1 were determined down to 0.1 p.p.m. and Se down to 0.2 p.p.m. from integrated absorbance readings. A vessel containing IBMK was kept in the carousel of the automatic sampler to slow down the evaporation of the organic solvent from the sample extracts loaded in open vials. The carousel was always kept covered. Instrumental conditions are given in Table 1.10 Results and Discussion Preliminary experiments using extraction from 3 M HCl with the addition of up to 1 ml of nitric acid to the 10 ml of extracted Table 1.Conditions of measurement in graphite furnace AAS Element . . . . Se Te Tl Analytical line/nm . . 196.0 214.3 276.8 Atomisation surface Wall Wall Platform Samplevolume/pl . . 10 10 5 Modifiervolume/pl . . 3 3 2 Temperature programme/("C; sramP + shold) Drying . . . . 110;5 + 10 110;5 + 10 500;5 Pyrolysis . . . . 500;8 500; 8 600; 10 + 5 Atomisation . . 2800; 1 + 1.2 2600; 1 + 1.2 2600; 1 + 2 1200; 1 + 3 Modifier for Se and Te . . Modifier form . . . . 1% H2S04 + 0.5% Mg 1200; 1 + 3 10% HN03 + 0.5% Cu + 0.5% Mg28 ANALYST, JANUARY 1987, VOL. 112 solution showed that extraction with TOMA is less affected by the nitric acid than extraction with TOPO and the former was therefore used for further experiments. It soon became evident that the presence of nitric acid in the limited amount used did not affect the extraction itself, but rather the stability of the extracts.If determined immediately after separation, the correct amount of extracted metals was usually found. Metals in the extract showed different stabili- ties. Tellurium was very rapidly lost and was therefore used for checking the stability of the extracts. In sample extracts with a high metal content a yellow precipitate was formed within a few hours and the original brown colour became orange. Iodine, Pb, Cu, Zn and Ga were identified in the precipitate by XRF. Extracts with a lower metal content turned bright yellow, and even if no precipitate was visible the Te content fell markedly. If the extract was stirred just before introduction into the graphite furnace with an automatic sampler, highly scattered readings for Te were observed.As the determinations were repeated the higher values eventually disappeared. This behaviour indicated that Te was still present but was in the form of particles that were gradually settling out after stirring. If after separation the organic phase was stored undisturbed in a glass vial, the colour change from brown to yellow always began from the top surface. Discolouration was particularly frequent with samples high in iron. The extracts also emitted vapours, highly irritant to the eye. From these observations and some basic chemistry the following sequence of events may be seen: 1, Nitric acid oxidises iodide in the aqueous phase with the production of nitrous acid and nitric oxide.The latter forms nitrosyl compounds that are extracted into IBMK. 2, In the IBMK, nitric oxide catalyses the oxidation of iodide to iodine by atmospheric oxygen. 3, As iodide is lost, soluble metal iodide complexes turn into less soluble forms and precipitate. 4, Iodine reacts with IBMK in the enolic form, giving an iodinated ketone. Halogenated ketones are well known tear gas compounds. If this sequence of reactions is a correct description of the processes involved, the instability of the extracts is not due to the nitric acid itself but rather to the products of its reduction, i.e., nitrous acid and nitric oxide. These may be eliminated by the addition of urea, sulphamic acid or ammonia in general. The presence of the ammonium ion in TOMA explains why extracts with this reagent were more stable than with TOPO. Both urea and sulphamic acid were investigated.The addition of 1 ml of a 1% solution of these reagents was found to be sufficient to secure the stability of the extracts. A correlation between the instability of extracts and a high iron content in the samples pointed to an iron nitrosyl halide as the NO-containing species extracted. The amount of NO extracted, and presumably the stability of extracts, could therefore be assessed from the amount of iron in the organic phase. Iron extracted under different conditions, i . e . , 1 .o v) Q 0.5 A with and Fig. 1. Signal of iron extracted into IBMK - TOMA in presence of increasing amounts of NO2- from 2 M HCI, 0.266 M KI and: A, ascorbic acid (4%) + sulphamic acid (0.066%); B, sulphamic acid only; C, ascorbic acid only; and D, neither without the addition of ascorbic and/or sulphamic acids and with increasing concentrations of nitric oxide, was determined by microsampling flame AAS using the Fe line at 372 nm.Nitric oxide was added as sodium nitrite, which is reduced to nitric oxide by the excess iodide. The solutions contained 25 mg of iron added as FeC13. Other conditions of the extraction were identical with those used for the sample solutions, i.e., the final concentrations were 2 M HC1, 0.266 M KI and 4% ascorbic acid. The results show that in the presence of nitric oxide iron is extracted both in the absence and the presence of ascorbic acid. The addition of ascorbic acid reduces the amount extracted but less so than sulphamic acid.With an increase in the amount of nitrite, the amount of Fe extracted in the absence of ascorbic acid levels off at about 0.003 M NO*-, but increases further in its presence (Fig. 1). The addition of ascorbic acid alone thus cannot prevent the extraction of iron and nitric oxide into the organic phase. Only a combination of ascorbic and sulphamic acids reduces iron extraction to a negligible level. Whether nitric oxide is also extracted in some other form is difficult to assess. However, if so, the amount must be limited as no instability of extracts unconnected with a high iron content in the samples was ever observed. It is also difficult to assign any valence to the iron in the nitrosyl compounds extracted. Iron nitrosyl halides with metal valencies 1-111 are known.It is plausible that the nitrosyl compound extracted is a negatively-charged ferrous complex that forms an ion asso- ciate with TOMA or protonated IBMK. This is suggested by the observation that when carrying out the extraction under identical conditions but without the addition of sulphamic acid and using KBr instead of KI, sample solutions high in iron turned green on addition of ascorbic acid. This green compound is partially extracted into IBMK, as seen by the intense green of the organic layer. On standing, the green gradually fades and a red - brown precipitate forms at the contact with the aqueous layer. It is known that negatively charged nitrosyl halides of Fell are green, neutral halides are red and cationic halides are brown.11 A less intensive green colour is observed when adding ascorbic acid to the iodide solutions.It is, however, obscured by the brown of the extracts. The elements Ag, Cd, Se, Te and T1 were determined in the same extracts in order to check the expected correlation between the amount of iron extracted and the stability of the extract. Ag and Cd showed no systematic variation with the amount of nitrite added; the scatter of values was within 5%. I I 1 1 I I / I I I I 1 I 0 2 4 6 8 1 0 2 4 6 8 1 0 “ O Z ~ ] / ~ M Fig. 2. Signals of Se, Te and T1 measured from same solutions as in Fig. 1. (a) Solution A; (b) solution B; ( c ) solution C; and (d) solution D of Fig. 1. A, Se; B; Te; and C, Tl. Amounts extracted: Te and T1,l FLg; Se, 2.5 pgANALYST, JANUARY 1987, VOL.112 29 Signals of Se, Te and T1 determined in the graphite furnace are plotted in Fig. 2. Thallium, in a similar manner to Ag and Cd, does not show any systematic variation. Se and Te, at the highest N02- concentration and both with the addition of ascorbic acid only and without any addition, show significantly lower values indicating the instability of the extracts. The same two solutions also have the highest Fe content. The signal of Te in the extract without the addition of ascorbic acid, determined 30 min after extraction, was 0.166 A s and fell to 0.02 A s within 1 h. If the two readings for the unstable extracts are excluded there is no systematic variation. The scatter of readings has a relative standard deviation of about lo%, which is not statistically significant.Although 1 ml of 1% sulphamic acid removes all instability up to the concentration level of NO investigated, a 2% solution was used in the determinations. Out of more than 2000 samples analysed by this procedure, only about 10 showed any instability by decolourising. These were repeated using a smaller aliquot of the sample solution and dilution with 3 M HC1. The relative standard deviation of the determinations is about 10% for Ag and Cd, 15% for Se and Te and 20% for TI, which is adequate for geochemical exploration purposes. Conclusions The results reported in this paper indicate that it is possible to apply the extraction of iodide and/or bromide complexes of metals as ion associates with TOMA to the analysis of geochemical exploration samples, even in the presence of up to 10% V/V FINO3 if certain provisions are made.In particular, the time during which the nitric acid and iodide are in intimate contact must be short in order to minimise the amount of nitrous acid formed. The sample solution extracted should, therefore, always be clear as the presence of clay colloids slows down the separation of phases. If the sample solutions are clear the separation is very fast because of the high density of the aqueous layer. As the instability of the extracts is due to nitric oxide extracted in the form of nitrosyl compounds, it may be eliminated by reducing both nitrous acid and nitric oxide (which are in equilibrium) to nitrogen by the addition of sulphamic acid or urea to the extracted solutions.Ascorbic acid must be added to counteract the oxidation of iodide to iodine, which would make the separation of phases difficult. It is also needed to reduce some of the extracted elements to their lower valencies that form the iodide complexes. Extracts prepared under these conditions with the addition of sulphamic acid and kept in a refrigerator have been found to be stable for many days. As many elements of interest in geochemical exploration form relatively stable iodide and/or bromide complexes (e.g., Ag, Au, Bi, Cd, Cu, Ga, Hg, In, Pb, Sb, Se, Sn, Te, T1 and Zn) this extraction system has potentially very broad applications.3.10 The enrichment factor attainable is evidently limited by the presence of common base metals (Cu, Pb and Zn), which use up the reagents and may saturate the organic phase.This is not generally a limiting factor for geochemical exploration samples. The main advantage of the method is that the extraction may be applied to samples digested with aqua regia, which is a much simpler, and for geochemical exploration more widely used, oxidative decomposition procedure than HCl + KC103,2 HC1 + H2025 or fusion with potassium pyrosul- phate.12 All these have been used in order to make the extraction of iodide complexes with TOMA applicable to geochemical exploration samples. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Thompson, M., Analyst, 1985, 110, 443. Viets, J . G.. Anal. Chem., 1978, 50, 1097. Clark, J. R., and Viets. J. G., Anal. Chem., 1981, 53, 61. Motooka, T. M., Mosier, E. L., Sutley, E. T.. andViets, J . R., Appl. Spectrosc., 1979, 33, 456. O’Leary, R. M., and Viets, J. G., At. Spectrosc., 1986, 7, 4. Burke, K. E., Analyst, 1972, 97, 19. Burke, K. E., Tulanta, 1974, 21, 497. Bedrossian, M., Anal. Chem., 1978, 50, 1898. Janousek, I . , Coll. Czech. Chem. Commun., 1978, 43, 2136. Rubeska, I., Ebarvia, B., Macalalad, E., Ravis, D., and Roque, N., “Multielernent Extraction System for the Deter- mination of Trace Elements in Geochemical Exploration Samples,” PHI 85/001, Internal Technical Report GCR/86/3, 1986. Moeller, T., “Inorganic Chemistry,” Wiley, New York, 1952, Viets, J. G., O’Leary, R. M., and Clark, J. R., Analyst, 1984, 109, 1589. Paper A61229 Received July I7th, I986 Accepted August lath, 1986 p. 603.