Determination of antimony in concentrates, ores and non-ferrous materials by atomic-absorption spectrophotometry after iron-lanthanum collection, or by the iodide method after further xanthate extraction

Tuluntu. Vol. 26. pp. 999 lo 1010 Pergamon Press Ltd 1979. Printed in Great Britain

DETERMINATION OF ANTIMONY IN CONCENTRATES, ORES AND NON-FERROUS MATERIALS BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY AFTER IRON-LANTHANUM COLLECTION, OR BY THE IODIDE METHOD AFTER FURTHER XANTHATE EXTRACTION ELSIE M. DONALDSON Mineral Sciences Laboratories, Canada Centre for Mineral and Energy Technology, Department of Energy, Mines and Resources, Ottawa, Canada

(Receioed 13 February 1979. Accepted 22 March 1979) Summary-Methods for determining trace and moderate amounts of antimony in copper, nickel, molybdenum, lead and zinc concentrates and in ores are described. Following sample decomposition, antimony is oxidized to antimony(V) with aqua regia, then reduced to antimony(III) with sodium metabisulphite in 6M hydrochloric acid medium and separated from most of the matrix elements by co-precipitation with hydrous ferric and lanthanum oxides. Antimony (2 100 PgJg) can subsequently be determined by atomic-absorption spectrophotometry, at 217.6 nm, after dissolution of the precipitate in 3M hydrochloric acid. Alternatively, for the determination of antimony at levels of 1 pg/g or more, the precipitate is dissolved in 5M hydrochloric acid containing stannous chloride as a reductant for iron(W), and thiourea as a complexing agent for copper. Then tin is complexed with hydrofluoric

acid, and antimony is separated from iron, tin, lead and other co-precipitated elements, including lanthanum, by chloroform extraction of its xanthate. It is then determined spectrophotometrically, at 331 or 425 ntn, as the iodide. Interference from co-extracted bismuth is eliminated by washing the extract with hydrochloric acid of the same acid concentration as the medium used for extraction. Interference from co-extracted molybdenum, which causes high results at 331 nm, is avoided by measuring the absorbance at 425 nm. The proposed methods are also applicable to high-purity copper metal and copper- and lead-base alloys. In the spectrophotometric iodide method, the importance of the preliminary oxidation of all of the antimony to antimony(V), to avoid the formation of an unreactive species, is shown.

For use in the Canadian Certified Reference Materials Project, a method was required for the determination of antimony in concentrations as low as 1 PgJg. Recent work by the author on the extraction of metal ethyl xanthate complexes from hydrochloric acid media’ has resulted in methods for determining tellurium2 and arsenic3 in copper, lead, zinc, nickel and molybdenum concentrates. These methods are based on the co-precipitation of tellurium(V1) and arsenic(V) with hydrous ferric oxide from an ammoniacal medium, followed by their separation from iron by chloroform extraction as the xanthates from >, 11M hydrochloric acid media. It is known that antimony can also be separated by the same co-precipitation procedure4s5 and that antimony(II1) can be quantitatively extracted as the xanthate from acid media.‘q6 Therefore, it was considered that a similar method might be developed for antimony. Probably the most common methods currently used for determining antimony involve atomicabsorption finishes, often after prior separation of antimony by co-precipitation’-” or extraction techniques. “*” However, these methods are not sensitive Crown Copyrights reserved. 999

enough for the determination of pg-quantities of antimony. More sensitive atomic-absorption methods, based on hydride-evolution techniques, have been

reported but these are subject to numerous iriterferences.13*‘4 The most widely used spectrophotometric methods involve the formation and extraction of the ion-association complexes formed between the chloro-complex of antimony(V) and xanthene or basic triphenylmethane dyes15*‘6 such as Rhodamine B”*‘* or Brilliant Green,19v2’ respectively, followed by direct measurement of the absorbance of the extract. Although these methods are sensitive (molar absorptivities up to 4 1 x lo4 l.mole-‘.mm-‘), they are not very specific because other elements that form chloro-complexes react in a similar manner. “.I6 Fur: thermore, reproducibility is generally poor because the methods are based on the reaction of antimony(V), which rapidly hydrolyses to form unreactive compounds in the strongly acidic chloride medium required for the oxidation of antimony and for the extraction of the complex. ’ 1-23 Reasonably reproducible results can be obtained only if the time interval after the oxidation step, and before the extraction step, is rigidly controlled.‘7**9*20*23 Recently, the author reported the determination of

1000

ELSIEM.

DONALDSON

bismuth at the &g-level by the iodide method, after its separation by diethyldithiocarbamate and xanthate extractions.24 In this work, the sensitivity of the method was increased about threefold by measuring the absorbance of the complex at the wavelength of maximum absorption (337 nm) in the near ultraviolet. Because antimony(III) [and antimony(V) which is reduced by iodide] forms a similar stable iodide complex in dilute sulphuric acid media, and because the sensitivity of the method can be increased about sevenfold, i.e., to 3.11 x lo3 l.mole-’ .mm-I, by measuring the absorbance at 331 nm,*’ the wavelength of maximum absorption, this simple and sensitive spectrophotometric finish was also chosen for use in the present work. This paper describes both a spectrophotometric and an atomic-absorption method for the determination of antimony in copper, lead, zinc, nickel and molybdenum concentrates, and in ores, high-purity copper metal and copper and lead-base alloys. Both methods involve the preliminary separation of antimony(II1) from most of the matrix elements by coprecipitation with hydrous ferric and lanthanum oxides. The spectrophotometric method, which is based on the ultimate formation of antimony(III) iodide, involves further separation of antimony by chloroform extraction as the xanthate from a 5M hydrochloric acid medium containing tartaric and hydrofluoric acids, stannous chloride and thiourea. Antimony is subsequently separated from coextracted bismuth by washing the extract with hydrochloric acid of the same acid concentration as the medium used for extraction. An advantage of the proposed spectrophotometric method over those based on separations involving the extraction of the antimony(V) chloro-complex,“*‘* or similar coloured ion-association dye complexes, is that the xanthate extraction step involves antimony(II1) which is stable in acidic chloride media.

EXPERIMENTAL

Apparatus All results by atomic-absorption spectropbotometry were obtained with a Varian Techtron model AA6 spectrophotometer equipped with a lo-cm laminar-flow air-acetylene burner. For maximum efficiency, the gas-dispersion tubes used for aeration were bent to hook over the side of the beakers, with the fritted glass tips parallel to the bottoms of the beakers. Reagents Standard antimony solutions, 5, 50 and lOOpt~/rnl. Dissolve 0.2669 g of pure potassium antimony tartrate (dried at 105” for 1 hr) in water and dilute to 1 litre with water. Transfer a 5-ml aliquot of this 100 &ml stock solution to a lOO-ml standard flask, and a 50-ml aliquot to another lOO-ml flask. Dilute each solution to volume with water. Prepare the diluted 5 and 50 &ml solutions fresh as required. Potassium iodide-ascorbic acid solution. Components 35% and 2.5% w/v, respectively. Prepare fresh as required.

Iron(lll) sulphate solution (1 ml = 1Omg of iron). Dissolve 25 g of ferric sulphate monohydrate in hot water containing 20ml of 50% v/v sulphuric acid, cool and dilute to 5OOml with water. Lanthanum chloride solution (1 ml = 10 mg oflanthanum). Dissolve 12.5g of lanthanum chloride hexahydrate in water and dilute to 500ml with water. Hydrochloric acid-stannous chloride-tartaric acidthiourea solution. Prepare a sufficient volume of solution

just before use., the composition being 43 ml of concentrated hydrochloric acid, 0.5 g of stannous chloride dihydrate, 2 g of tartaric acid and 0.5 g of thiourea per 100 ml. Hydrochloric acid-tartaric acid wash solution. Dissolve 4g of tartaric acid in water, add 430ml of concentrated hydrochloric acid and dilute to 1 litre with water. Potassium ethyl xanthate, 20% w/v solution. Prepare fresh as required. Thiourea, 5% w/v solution. Prepare fresh as required. Aqua regia. Mix 3 parts of concentrated hydrochloric acid with 1 part of concentrated nitric acid. Prepare fresh as required. Potassium hydroxide, 10% w/v solution. Store in a plastic bottle. Tartaric acid, 5% w/v solution. Ammonia, 10% v/v solution. Sulphuric acid, 50% v/v. Hydrochloric acid, 25 and 50% v/v. Nitric acid, 50% v/v. Chloroform. Analytical-reagent grade. Determination of anthnony by the spectrophotometric iodide method Calibration. Add 4ml of 50% sulphuric acid and 1 ml of 5% tartaric acid solution to each of fifteen 25-ml standard flasks; then, from a burette, add to the first five flasks 1, 2, 3, 4 and 5 ml respectively, of standard 5-&ml antimony soiution. Add to the next nine flasks 1, 1.5, 2, 3, 4, 6, 8, 10 and 12 ml respectively, of standard 50-&ml antimony solution. The last flask contains the blank. If necessary, dilute each solution to approximately 15 ml with water and cool’to room temperature in a water-bath. Add 5 ml of freshly prepared 35% potassium iodide-2.5% ascorbic acid solution to each flask, dilute to volume with water and mix. Allow the solutions to stand for about 30min to complete the complex formation, then determine the absorbance, at 331 nm, of the blank and each of the five solutions in the first series against water as the reference solution, using 40-mm cells. Determine the absorbance of the blank, the last solution in the first series, and each of the first four solutions in the second series in a similar manner, at 425 nm, using 40-mm cells. Determine that of the blank and each of the last seven solutions in the second series, at 425nm, using lO-mm cells. Correct the absorbance value obtained for each antimony iodide solution by subtracting the corresponding blank value. Plot pg of antimony vs. absorbance for each series of measurements. Copper, nickel, zinc and lead concentrates and ores. Transfer 0.05-0.5g of sample (see Notes l-3), containing up to 2 mg of antimony, to a 400-ml beaker. Add 25 ml of freshly prepared aqua regia, cover and heat gently until all or most of the sample is decomposed. Add 25 ml of 50% sulphuric acid, heat until the evolution of oxides of nitrogen ceases, then remove the cover, wash down the sides of the beaker with water, and carefully evaporate the solution to dryness. Cool, add 50ml of 50% hydrochloric acid, cover and, if necessary, heat gently to dissolve the salts (particularly lead* sulphate). Cool the resulting solution to room temperature, add 3 g of sodium metabisulpliite, mix, and allow the solution to stand for = 5 min. Boil the solution (covered) for _ 10min to remove the excess of sulphur dioxide, then add 25 ml of water. Place a gas-dispersion tube in the beaker and pass air through

Determination

of antimony in concentrates

the solution at a fairly rapid rate for - 10 min to reoxidize any iron(H) present. Remove the tube after washing it thoroughly with water. If necessary, add sufficient iron(II1) sulphate solution to the resulting solution so that at least 50mg of iron are present. Add 5 ml of lanthanum chloride solution and sufficient concentrated ammonia solution to precipitate iron as the hydrous oxide, then add 75 ml in excess and heat the solution to the boiling point to coagulate the precipitate. Allow it to settle and, using a short-stemmed funnel, filter the hot solution (Whatman No. 40 paper). If molybdenum and/or more than - 75 mg of copper or nickel are absent, wash the beaker twice and the paper and precipitate three times with 10% ammonia solution (Note 4). Discard the filtrate and washings. If molybdenum and/or more than -75 mg of copper or nickel are present, wash the beaker, the paper and precipitate once each with 10% ammonia solution. Place the original beaker under the funnel and add 25ml of 25% hydrochloric acid to the funnel to dissolve the precipitate. Wash the paper three times with 25% hydrochloric acid added from a plastic wash-bottle, then wash down the sides of the beaker with the acid. Reprecipitate the iron and lanthanum, and filter off and wash the precipitate as described above (Note 4). Discard the filtrate. Carry a blank, with - 50 mg of iron(II1) added, through the whole procedure (if some samples have high antimony content and some have low, more than one blank will be needed, see below). Transfer the funnels containing the blank and sample precipitates to 250-ml separatory funnels, marked at 100 ml. Wash down the sides of each precipitation beaker with 25 ml of freshly prepared hydrochloric acid-stannous chloride tartaric acid-thiourea solution (Note 5). Add each of the resulting solution to the funnel containing the corresponding precipitate and wash the beaker three times with the same acid solution added from a plastic wash-bottle. Wash the paper three times with the same acid solution, then discard the paper. Dilute each solution to the 100-m] mark with the acid each solution (Note 6) add 2 ml of concentrated hydrofluoric acid and mix thoroughly (Note 7). Add 10 ml of chloroform and then 1 ml of freshly prepared 20% potassium ethyl xanthate solution (Note 8). Stopper and shake for 1 min. Allow several min for the layers to separate, then drain the chloroform phase into a 125-ml separatory funnel (Note 9). Extract the aqueous phase twice more, in a similar manner, with lO- and 5-ml portions of chloroform and 1 and 0.5 ml of xanthate solution, respectively then wash the aqueous phase by shaking it for -30 set with 3 ml of chloroform. Add 30 ml of hydrochloric-tartaric acid wash solution and 1 ml of 5% thiourea solution (Note 5) to the combined extracts, stopper and shake for 1 min. After the layers have separated, drain the chloroform phase into a 100-ml Teflon beaker (Note 10). Add 5 ml of chloroform and 0.5 ml of xanthate solution to the aqueous phase and shake for 1 min. Allow the layers to separate and drain the chloroform phase into the beaker containing the initial extract. Wash the aqueous phase by shaking it for - 30 set with 5 ml of chloroform. then add 8 ml of 50% nitric acid to the combined extracts and heat in a hot water-bath to remove the chloroform. Add 1 ml of concentrated perchloric acid and 0.5 ml of 50% sulphuric acid, cover the beaker and heat until the evolution of oxides of nitrogen ceases. Remove the cover, wash down the sides of the beaker with water and evaporate the solution to fumes of perchloric acid. Cool to room temperature, add 5 drops of freshly prepared aqua regia and mix thoroughly. Wash down the sides of the beaker with water and evaporate the solution until the diameter of the drop remaining in the bottom of the beaker is 3-4 mm. Cool the beaker in a water-bath, then wash down the sides with 5 ml of 10% potassium hydroxide solution added from a pipette, and heat the solution gently for -5 min.

1001

Cool slightly and add 1 ml of 5% tartaric acid solution and 4.5 ml (Note I I) of 50% sulphuric acid. Heat the solution gently again for - 5 min, then add - 5 ml of water and cool the resulting solution to room temperature in a waterbath (Note 12). If the sample contains 6OOpg or less of antimony, transfer the solution to a 25-ml standard flask containing 5 ml of 35% potassium iodide-2.5% ascorbic acid solution (Note 13). Treat the blank similarly. Dilute to volume with water, mix and proceed with the subsequent determination of antimony as described above (Note 14) using either lOor 40-mm cells and a wavelength of 425 or 331 nm as required (Note 15). If the sample contains more than 600 pg of antimony, transfer the solution to standard a flask of appropriate size (25 or 50ml). Add sufficient additional 5% tartaric acid solution for 1 ml to be present for each 10 ml of final solution and dilute to volume with water. Transfer a lO-ml aliquot to a 25-ml standard flask and add sufficient 50% sulphuric acid for a total of 4ml to be present. Mix and cool the resulting solution to room temperature, then proceed with the addition of potassium iodide-ascorbic acid solution and the subsequent determination of antimony as described for the calibration, using lo- or 40-mm cells as required and a wavelength of 425 nm. Treat the blank similarly. Molybdenum ores and concentrates. Transfer up to 0.5 g of sample (see Notes l-3) to a 40&m] beaker. Add 1.5 g of potassium chlorate, moisten with a few ml of water, cover and carefully add 20ml of concentrated nitric acid. Heat gently until the sample is decomposed, then add 25 ml of 50% sulphuric acid and heat until the evolution of oxides of nitrogen ceases. Remove the cover, wash down the sides of the beaker with water and carefully evaporate the solution to fumes of sulphur trioxide. Cool to room temperature, add - 10 ml of water, cover and add 15 ml of freshly prepared aqua regia. Heat gently for 5-lOmin, then remove the cover and evaporate the solution to dryness (Note 16). Proceed with the dissolution of the salts and the reduction and separation of antimony(I11) by coprecipitation with hydrous ferric and lanthanum oxides as described above. Dissolve the precipitate, and reprecipitate iron and lanthanum as described, then proceed with the separation of antimony by chloroform extraction of its xanthate complex and its subsequent determination as described above, using a wavelength of 425 nm (Note 15). Copper metal and-copper- and-lead-base alloys. Decomuose a suitable weight of samnle (0.1-0.5 a) (Notes 1 and 17) containing not more than 2 mg‘of antimony, and determine antimony by the method described for copper, nickel, zinc and lead concentrates. Determination of antimony by atomic-absorption spectrophotometry Calibration solutions. Add 2 ml of 5% tartaric acid solution, 15 ml of concentrated hydrochloric acid and 5 ml each of iron(II1) sulphate and lanthanum chloride solutions to each of eight 100-m] standard flasks; then, from a burette, add to the first seven flasks 1, 3, 5. 7.5, 10, 15 and 20ml of standard IOO-ng/ml antimony solution. The last flask contains the zero calibration solution. Dilute each solution to volume with water and mix. General procedure. Following the separation of 250 pg of antimony by a single co-precipitation with hydrous ferric and lanthanum oxides as described above (Note 18) wash the beaker twice and the paper and precipitate three times with 10% ammonia solution. Discard the filtrate and washings and under the funnel, place a lOO-ml standard flask containing 2ml of 5% tartaric acid solution. Wash down the sides of the beaker with 45 ml of 25% hydrochloric acid and add the resulting solution to the funnel containing the paper and precipitate. Wash the beaker twice with - 5-ml portions of water and add the washings

1002

ELSIE M. JXJNALDSON

to the funnel. Wash the paper three times with -S-ml portions of 25% hydrochloric acid added from a plastic wash-bottle, then wash it twice with water. Discard the paper. Dilute the resulting solution to volume with water, mix, and measure the absorbance of the solution, at 217.6nm, in an oxidizing air-acetylene flame. Determine the antimony content of the sample solution by relating the resulting value to those obtained concurrently for calibration solutions of slightly higher and lower antimony concentrations. Notes

1. The use of samples containing more than - 150 mg of iron is not recommended because the mixed hydrous oxide filtration step becomes unduly slow. Low results, caused by incomplete co-precipitation of antimony, will be obtained at the 2-mg level if more than - 25 mg of either, aluminium or tin, or more than - 10 mg of each, are present. Up to 50mg of either element, when present separately, will not interfere in the co-precipitation of < 500 pg of antimony. 2. If the sample contains an appreciable amount of silica, use a Teflon beaker and add 2-3ml of concentrated hydrofluoric acid after the cover has been removed. Evaporate the solution to fumes of sulphur trioxide, cool to room temperature, add - 15 ml of water and heat to dissolve the salts. Transfer the solution to a glass beaker, evaporate it to dryness and proceed as described. Low results will be obtained if the excess of sulphuric acid is not removed by evaporation. 3. If the sample contains an appreciable amount of acidinsoluble material or the presence of an insoluble antimony compound is suspected, it can be decomposed by fusion as follows. Mix a suitable weight of sample, contained in a 60-ml nickel (not zirconium or iron) crucible, with 3g of sodium peroxide and cautiously fuse the mixture over an open flame. Allow the melt to cool, then transfer the crucible to a covered 400-ml beaker (Note 2) containing 50 ml of water and 25 ml of 50% sulphuric acid. Remove the crucible immediately after the melt has dissolved and add 10ml of 50% nitric acid to prevent (if chloride is present) the loss of antimony by volatilization as the chloride.26 Evaporate the covered solution to - 30m1, then remove the cover and evaporate the solution to fumes of sulphur trioxide. Cool to room temperature, add 15ml of freshly prepared aqua regia, cover and heat until the evolution of oxides of nitrogen ceases. Remove the cover and evaporate the solution to dryness., Dissolve the salts in 75ml of 50% hydrochloric acid and proceed with the reduction and a double co-precipitation of antimony as described. 4. If the subsequent xanthate extraction cannot be completed the same day, allow the precipitate to stand overnight (or longer) at this point. 5. Thiourea can be omitted if it is known that the sample contains little or no copper. 6. Unless molybdenum is present (pale yellow solution), the solution should be colourless at this point. Sufficient stannous chloride is present to reduce up to -240mg of iron(II1). 7. To minimize the attack of hydrofluoric acid on glass, the antimony(II1) xanthate extraction should be done immediately after the hydrofluoric acid has been added. Similarly, the funnel should be washed immediately after the extraction has been completed. 8. The xanthate solution should be added by pipette with use of a suction bulb (not by mouth) or by using a graduated or marked medicine dropper, and the-extra& tion should be carried out in a fume hood. Prolonged exposure to xanthate vapour can produce an allergic re&tion. 9.. A . _reddish or purple extract indicates the presence of molybdenum.

10. Glass beakers should not be employed, because the potassium hydroxide solution subsequently used may leach antimony or lead from the glass. Teflon beakers may become partly discoloured (i.e., yellow-brown or black) inside because of the subsequent use of aqua regia to oxidixe antimony to antimony(V). Before the beakers are used again, this discolouration should be removed by heating perchloric acid to dense fumes in the covered beaker. 11. The additional 0.5 ml of 50% sulphuric acid added to the sample solution-as compared to the calibration solutions--is required to react with the potassium hydroxide. 12. Salts may crystallize from the solution on standing but these will redissolve when the solution is ultimately diluted and mixed thoroughly. 13. The presence of arsenic is signified by a deep yellow or orange colour due to iodine liberated during the reduction of arsenic(V) by potassium iodide. The iodine is subsequently reduced by ascorbic acid when the solution is mixed. 14. If the solution is slightly opalescent, filter it through a dry Whatman No. 42 filter paper before the spectrophotometric measurement. 15. If molybdenum was co-extracted as the xanthate (Note 9). absorbance measurements should be ma& at 425 nm, after the solution has been stood overnight to ensure complete complex formation. 16. If the residue becomes deep blue (molybdenum blue) on cooling, add -5-10 ml of water and 2 ml of conccntrated perchloric acid and evaporate the solution to dryness again. 17. If the expected antimony content is low, up to at least 1 g of sample can be used for high-purity copper metal and for copper-base alloys of low aluminium and tin content (Note 1). 18. The use of samples containing more than - 100 mg of lead is not recommended because the resultant lead chloride is not completely soluble in 15% hydrochloric acid. RESULTS Spectrophotometric determination of antimony by the iodide method

In initial tests, low and erratic results were usually obtained by the iodide methodz5 after treatment of the antimony(II1) xanthate extracts [or pure potassium antimony(M) tartrate solutions] with nitric, perchloric and sulphuric acids and evaporation of the solution to fumes of sulphur trioxide before complex formation. This was considered to be due to the formation of an insoluble basic antimony compound.” In subsequent work, bromide was found to interfere at low levels of antimony when antimony(II1) xanthate in the extract was oxidized with brominecarbon tetrachloride solution,’ followed by back-extraction of antimony(V) into 10% sulphuric acid and removal of the excess of bromine by evaporation. The

amount of bromide formed did not interfere at high levels of antimony (2mg) when an aliquot of the resultant solution was taken for analysis. Hydrogen peroxide was found to be completely ineffective as an oxidant for antimony(II1) xanthate. Further tests, involving oxidation of antimony(II1) xanthate with acids as described above, also usually yielded slightly low results when the solution was ultimatelv evaborated to drvness in Teflon beakers and .

Determination of antimony in concentrates the salts were dissolved in 10% potassium hydroxide solution. The final sulphuric-tartaric acid solution obtained was also not stable on standing. Furthermore, the Teflon beakers became contaminated with ~gquantities of antimony, suggesting the presence of a compound that is insoluble in both potassium hydroxide solution and dilute sulphuric acid. Ultimately-as a result of later work involving the coprecipitation of antimony(II1) with iron(II1) and lanthanum-it was found that complete recovery of antimony and a stable solution are obtained if aqua regiu is added just before the evaporation of the solution to dryness and dissolution of the salts in potassium hydroxide solution. Low results were still obtained if the potassium hydroxide treatment was omitted. A probable explanation of this anomalous behaviour of antimony is given in the discussion below. Separation of anfimony(ZZZ) by co-precipitation with hydrous ferric and lanthanum oxides

Previous investigators have recommended the separation of antimony by co-precipitation with hydrous ferric4*5 or lanthanum oxides”10*14*2s but the oxidation state of antimony is usually not clearly indicated. In the present work, tests carried out to determine the required oxidation state, and also the effectiveness of iron and lanthanum individually as collectors, showed that antimony should be in the tervalent state and that lanthanum is a better collector than iron(III). Potassium antimony tartrate solutions (2OOOpg of antimony) were used in these tests and potassium permanganate was employed in a hot, fairly concentrated sulphuric acid medium to oxidize antimony(II1) to antimony(V) and to destroy tartrate which interferes in the co-precipitation step. In the tests involving collection of antimony(III), the resultant antimony(V) was subsequently reduced with sodium metabisulphite as described in the proposed method, and reduced iron, present in tests with iron(III), was oxidized by passing air through the solution by means of a gas-dispersion tube fitted with a fritted glass disc. Co-precipitation, at -pH 9, with iron(II1) (100 mg) was carried out as described previously for tellurium* and arsenic.’ Co-precipitation, at -pH 10, with lanthanum (1OOmg) was performed as described in the proposed method. Antimony was ultimately extracted as the xanthate and the extract was treated with 5 ml of 20% bromine-carbon tetrachloride solution as described above. Interference from bromide, in the spectrophotometric iodide finish, was avoided by taking one-fifth of the final solution for the determination of antimony. Complete recovery of antimony was obtained only in the test involving the collection of antimony(II1) with lanthanum (Table 1, test 1). In the tests involving collection of antimony(V), it was found necessary to dissolve the hydrous oxide precipitate with SM hydrochloric acid containing stannous chloride so that antimony(V) was immediately reduced to anti-

1003

mony(II1) (for the xanthate extraction) during the dissolution step. Much lower results were obtained if SM hydrochloric acid alone was used and antimony and iron were subsequently reduced with stannous chloride. No doubt, this is due to the formation of unreactive antimony(V) hydrolysis compounds in the hydrochloric acid medium.*‘*** Subsequent work involving collection of antimony(II1) with lanthanum (100 mg) (see Table l), followed by xanthate extraction and the iodide finish, invariably yielded low results when pure antimony(II1) solutions were treated with nitric acid or nitric and perchloric acids. However, complete recovery of antimony was obtained when an atomicabsorption finish was used after dissolution of the lanthanum precipitate. This indicated that the low results obtained in the xanthate-iodide scheme were due to incomplete formation of antimony(II1) xanthate. The results of tests 4 and 5 in Table 1 suggested that the pretreatment of the solutions with nitric acid was, in some way, responsible for this behaviour. The fact that complete recovery of antimony was obtained in test 3, in which it was present as antimony(V) before the addition of nitric acid, suggested that the use of nitric acid probably results in the formation of an unreactive oxidized species that is not reduced by either sodium metabisulphite or hydrazine sulphate, or with stannous chloride during the dissolution of the lanthanum precipitate. The formation, during oxidation of antimony with nitric acid or nitric and sulphuric acids, of an unreactive species which is not easily oxidized has been described by Maren. However, although Maren found that this species, referred to as antimony(IV), can be oxidized with perchloric acid and easily reduced with sulphite, these properties do not agree with those found in the present work (see tests 4 and 5, Table 1). On the basis of these findings, it was considered that complete recovery of antimony by use of the xanthateiodide scheme could only be obtained if it was completely oxidized to antimony(V) during the decomposition procedure. In this state, it is readily reduced to antimony(II1). The results of subsequent tests, in which a certified, high-antimony, reference ore, CD-l,” was decomposed (in a zirconium crucible) by an oxidizing fusion with sodium peroxide, are given in Table 2. In these tests, hydrazine sulphate was used as reductant (see Table 1, test 3 for conditions) and iron(H) was reoxidized as described in the proposed method. Both iron(III) and lanthanum (50 mg of each) were used for co-precipitation because this mixture produces a more readily filterable precipitate than lanthanum alone. The use of nitric acid after fusion was tested because it is required in the presence of chloride to prevent loss of antimony by volatilization as the chloride26 during subsequent evapor,ation of the solution. Tests l-4 show that antimony is not completely oxidized to antimony(V) during fusion with sodium peroxide unless an auxiliary oxidizing compound such as potassium nitrate (see

1004

ELITEM.

DONALDSON

Table 1. Effect of preliminary acid treatment on the determination of antimony by the iodide method after lanthanum collection and xanthate extraction Preliminary treatment of pure Sb(III) solutions*

Test 1

KMnO* oxidation in H2S04 medium and

2

evaporation to dryness KMn04 oxidation in HzSOo medium

3

4 5

KMnO, oxidation in H2SOL medium + HNOj and evaporation to fumes

Method of reduction used Na&O,

(3 g) in 6M HCl

Sb found, w 2005

NzHI. H$O, (1 g) and evaporation to dryness + salts dissolved in 6M HCl As in test 2

2012,1992

As in test 2

1969

As in test 1

1869

of SOS HNO, + H,SO, and evaporation to fumes of SO, HNO, + HClO., + H,S04 and evaporation to dryness

1988

* Sb(II1) taken, 2000pg. t Required to destroy the excess of reductant. tests 6 and 7) is present. This also suggests that an unreactive species is formed during fusion with sodium peroxide, and that it is not reduced with hydrazine sulphate or sodium metabisulphite (test 81 and not oxidized with potassium permanganate (test 3), or with potassium chlorate (test 4) unless nitric acid is absent (test 5). Although it was ultimately found that zirconium, derived from the crucible, can interfere in the co-precipitation of antimony with lanthanum, the results of test 8 in which nickel crucibles were used, confirm these conclusions (cJ: test 14). A mixture of sodium peroxide and potassium nitrate cannot be used for the decomposition of sulphide-containing materials because of the violence of the reaction. Consequently, the effectiveness of

various oxidizing agents was tested in conjunction with the decomposition of CD-1 with acids. The results of these tests, given in Table 2, ultimately showed that a mixture of potassium chlorate and perchloric acid (test 12).is effective in the absence of nitric acid, but that saqua regia (test 13) is a superior oxidizing agent for antimony. It can be used after fusion with sodium peroxide (test 14) and, because of the nitric acid content, antimony is not lost as the volatile chloride. It was also found that the use of aqua regia eliminated the problem of low and erratic results which was encountered in initial work involving the treatment .of the xanthate extracts with acids. Preliminary experiments, in which synthetic concentrates [2ooOpg of antimony(III) added] were

Table 2. Effect of method of decomposition on the determination of antimony in CD-l* by the iodide method after iron-lanthanum collection and xanthate extraction Test 1 2 3 4 5 6 7 8 9 10 11 12 13 14

.

Method of decomposition used

Na,02 fusion (Zr crucible)--melt dissolved in dilute HISO,--HNOs + HF added and evaporation to fumes of SO9 As for test 1 without the addition of HNO, NazOz fusion (Zr crucible)-melt dissolved in water-boiled to remove peroxidesHzSO, + KMn04 + HNOs f HF added and evaporation to fumes of SO, As for test 1 except KCIOp added just before evaporation to fumes of SO, As for test 4 without the addition of HNO, NazO, + KNOI$ fusion (Zr cru@ble)--melt dissolved in dilute HISO,-HNOJ + HF added and evaporation to fumes of SO, As for test 6 without the addition of HNO, NazOz fusion (Ni crucible)--as for test 1 except solution evaporated to dryness? HzSO,, + HF + HClO, and evaporation to fumes of SO1 H2S04 + HNO, + HF f KClOs and evaporation to fumes of SO, As for test 9 without the addition of HNO, H,S04 + HF + HC104 + KC103 and evaporation to fumes of SOB Aqua regia + HzSOo + HF and evaporation to fumes of SOa NazOz fusion (Ni crucible)--as for test 1 except nqw regia added after evaporation to fumes of SOa followed by evaporation to dryness?

* 50 mg samples taken; certified value 3.57% Sb.“O $200 mg added. t Antimony reduced with sodium metabisulphite in 50% hydrochloric acid. 11 Mean of 4 values, viz. 3.38, 3.46, 3.49 and 3.36%. T Mean of 2 values, viz. 3.55 and 3.61%.

Sb found, % 3.45 3.35 3.41 3.38 3.58 3.59 3.62 3.421) 3.47 3.41 3.49 3.58 3.57 3.58s

Determination of antimony in concentrates decomposed with aqua regia, and then, except for the use of hydrazine sulphate as reductant, were treated as described in the proposed spectrophotometric method, gave recoveries of 90-95% for copper and molybdenum concentrates, and complete recovery for lead and zinc concentrates. It was found that this was due to the use of hydrazine sulphate as reductant. Copper and molybdenum are partially reduced with this reagent and cause co-oxidation of antimony(II1) during the air-oxidation step.31*32Complete recovery was obtained for copper and molybdenum concentrates when antimony was reduced with sodium metabisulphite in + 50% hydrochloric acid.‘* Under these conditions, minimal reduction of copper and molybdenum occurs.

1005

tion was recovered by adding xanthate solution and re-extracting the antimony. It was found that up to at least 2 mg of antimony(II1) as the xanthate can be quantitatively extracted into chloroform in three successive extractions. In tests with iron(II1) and thallium(III), slightly high results were obtained at 331 nm when stannous chloride was used as a reductant for both elements. This was due to the co-extraction, as the xanthate, of a small amount of tin that was not completely removed by the washing step. This was obviated by complexing tin with hydrofluoric acid before the extraction of antimony(III) xanthate. Effect of diwrse ions

I

Selenium(IV), tellurium(IV), arsenic(III) and palladium(I1) are completely extracted into chloroform as xanthate complexes from 5M hydrochloric acid, and copper( platinum(W), gold(III), and molybPrevious work by the author’ showed that antidenum(V1) are partly extracted. Selenium and tellurmony(II1) xanthate can be quantitatively extracted ium will not interfere in the proposed iodide method into chloroform from O.l-5M hydrochloric acid, and because they are reduced to the elemental state with that iron( and lead, which forms an insoluble stannous chloride before the extraction of antiiodide,* 5 are not co-extracted from SM acid media. mony(I11) xanthate. Palladium, platinum and gold are Furthermore, although thallium(III), which also forms almost completely separated from antimony by the an insoluble coloured iodide, is appreciably extracted co-precipitation step. Up to at least 1 mg of each will from SM hydrochloric acid, thallium(I) is not not interfere after a single ammonia separation. Up extracted from 2 OSM acid media.’ Subsequent work to at least 1Omg of arsenic will not interfere, either by the author resulted in the development of methods in the extraction of up to 2 mg of antimony or in for tellurium* and arsenic3 in ore concentrates. These the determination of small amounts at either 331 or are based on the separation of tellurium and arsenic from iron and lead by xanthate extraction from 425 nm. The co-extraction of copper that is retained’ in the mixed hydrous oxide precipitate can be largely 2 11M hydrochloric acid media, after their prelimiprevented or inhibited by complexing it with nary separation from copper, zinc, nickel and molybthiourea.’ Molybdenum, which is also retained in the denum matrices by co-precipitation with iron(II1). precipitate and subsequently co-extracted as the xanAntimony can also be separated from these elements by similar co-precipitation techniques.4*5*~10*14*2* thate, causes high results for small amounts of antimony at 331 nm but not at 425 nm. It also slightly Consequently, the development of a similar method inhibits complex formation when + 1OOpg or more for small amounts of antimony, based on xanthate extraction from a 5M hydrochloric acid medium in of antimony are present, but this effect can be eliminated or minimized by allowing the solution to stand the presence of a suitable reductant for iron and thalfor _ 24 hr before measuring the absorbance. lium, was investigated. Up to at least 50 mg of manganese will not interfere In a recent method for determining bismuth,24 also based on a xanthate-iodide scheme, interference from in either the proposed iodide or atomic-absorption small amounts of lead that are co-extracted as the methods. However, the same amount of vanadium causes low results by both methods, probably because xanthate is eliminated by washing the chloroform extract with a solution of the same hydrochloric acid it is reduced to vanadium(IV) during the initial reducconcentration as the medium used for extraction. In tion step. This is probably partly oxidized during the air-oxidation step and causes some co-oxidation of the present work, similar tests showed that washing the xanthate extract with 5M hydrochloric acid effec- antimony(III). Large amounts of chromium (50mg) interfere because an insoluble compound is formed. tively removes bismuth, which is partly extracted during the co-precipitation step. Small amounts of from 5M acid and would interfere in the determination of antimony as the iodide.25 No interference was these elements will not interfere. It has been reported that other elements that form observed in tests involving 15 fig of antimony(II1) and hydrous oxides, notably iron and aluminium, do not 5 mg of bismuth when absorbance measurements were made at 331 nm. Up to 1Omg of bismuth can interfere in the co-precipitation of antimony with lanthanum. However, tests involving an atomic-absorpbe present during the extraction step without producing significant error. Similarly, after the washing step, tion finish, after dissolution of the iron-lanthanum no interference was observed at 331 nm when 300 mg precipitate, showed that large amounts (50 mg) of aluof lead were present during the extraction step. In minium, zirconium and tin interfere at the 2-mg level these tests, any antimony that entered the wash solu- when 50 mg of lanthanum are used. This is probably Separation of antimony by extraction of its ethyl xanthate complex

1006

ELSIE M. DONALDSON

because they preferentially form similar compounds with lanthanum or partly soluble compounds with antimony_ Up to _ 25 mg of each, when present separately, or w 10 mg each of aluminium and tin, will not interfere in the co-precipitation of up to 2 mg of antimony; 50 mg of either aluminium or tin will not affect the co-precipitation of ~500 pg. Larger amounts can be tolerated if more lanthanum is used but this results in a bulkier precipitate that takes longer to filter. Furthermore, when the precipitate contains 50 mg or more of aluminium, the resultant solution passes very slowly through the filter paper. In the absence of these and other hydrous oxides, 50mg of lanthanum will be sufficient for the co-precipitation of up to at least 10mg of antimony. Zirconium and iron crucibles are not recommended for fusion purposes because zirconium interferes and because too much iron would be introduced into the resultant sample solution. Tests, which were carried out to determine the effects of iron and other co-precipitated elements on the determination of 5 &ml of antimony by atomicabsorption spectrophotometry, showed that up to at least 500 &ml of copper, tin, aluminium, nickel, molybdenum, manganese and zinc, 1000 pg/ml of lead and sodium, 15OO&ml of iron, and 2OOIrglml of arsenic will not interfere. More than + loo0 fig/ml of lead will cause slightly high results and may result in the precipitation of lead chloride in the solution. Applications

The proposed spectrophotometric method was applied to the analysis of a series of four synthetic concentrates (ground to 200-mesh) in which antimony, added as antimony(III), was varied from 0.0005 to 1%. Because a lead sulphide concentrate of low antimony content could not be obtained, synthetic mixtures of lead sulphate and other elements usually found in commercial concentrates were used. Both of the proposed methods were applied to a variety of certified reference materials, and several commercial-purity copper rodsJ3 were analysed by the iodide method. The results of these analyses are given in Tables 3 and 4.

DISCUSSION

Table 3 shows that the results obtained for the synthetic concentrates by the spectrophotometric iodide method agree favourably with the calculated values. Those obtained for the synthetic nickel concentrate are also in reasonably good agreement with the calculated values, although it was found that the wncentrate was not homogeneous with respect to antimony; the mean value (reported in the footnote to Table 3) of seven results was used to calculate the total amount present. The results (Table 4) obtained for

CD-l, CPB-1 and CZN-1 by the spectrophotometric iodide method, after sample decomposition both by fusion and with acids, are in excellent agreement with those obtained by the atomic-absorption method. The results by both methods are also in good agreement with the corresponding recommended mean values. The results for the National Bureau of Standards and British Chemical Standards non-ferrous alloys by both methods are in good agreement and, in most instances, agree well with the certified values. Those obtained for the copper rods by the iodide method agree with the recommended values. This investigation has shown (Tables 1 and 2) that the nature of the decomposition method employed, before the separations of antimony by co-precipitation with lanthanum followed by xanthate extraction, can cause a significant difference in the results obtained for antimony. This is because of the formation of an unreactive species when the decomposition procedure initially involves the use of nitric and/or perchloric acids or fusion with sodium peroxide. Although tests involving an atomic-absorption finish have shown that this species is completely co-precipitated with lanthanum apparently it is not reduced with tin(I1) during the subsequent dissolution of the precipitate with hydrochloric acid containing stannous chloride. In early work, this apparent and variable loss of antimony during acid digestions in the presence of oxidizing acids such as those listed above or hydrogen peroxide34*35 was attributed to volatilization of antimony.34 Later, Marenz9 showed that it was caused by partial oxidation to an unreactive state, which he called antimony(IV) because it was known that nitric acid oxidation of antimony results in partial formation of a tetroxide (Sb,O,). This has a definite composition and is still used as a weighing form for the gravimetric determination of antimony.’ and recent Miissbauer However, titrimetricJ6 studies3’ have shown that the tetroxide, which is presumably the dehydrated form of the unreactive compound mentioned above, is a compound containing antimony(II1) and antimony(V) in a 1 :l ratio (i.e., Sb40s = Sb203 + Sb205). Maren” found that this compound, hereafter referred to as antimony (III + V), is not readily oxidized [with cerium(IV)] to antimony(V), which is required for the formation of the Rhodamine B complex, unless it is first reduced to antimony(II1) with sulphite. However, he found that it can be oxidized with perchloric acid. In the present work, it was found that the antimony(II1 + V) species is not readily oxidized with perchloric acid, potassium permanganate or potassium chlorate, but that it is completely oxidized to antimony(V) with aqua regia. It was also found, contrary to Maren’s findings, that it is not reduced by either sulphite or hydrazine sulphate. However, tests wlth reference ore CD-1 involving direct formation of the iodide complex, after decomposition under conditions that favour the formation of the unreactive species, showed that this species is reduced by potassium iodide.

Determination

1007

of antimony in concentrates

Table 3. Recovery of antimony from synthetic concentrate samples

Matrix and nominal composition, % Cu concentrate (24.7 Cu, 30.7 Fe, 35.6 S, 3.2 Zn, 1.2 Si)

MO concentrate (95.9 MO&)

Ni concentrate (33.3 Ni, 30.2 Fe, 32.1 S, 4.5 Cu, 0.2 As, 0.02 Bi)

Zn concentrate (50.6 Zn, 33.5 S, 10.1 Fe)

Pb concentrate (68 Pb, 10 Fe, 0.2 Bi, 0.4 As, 2 Cu)

Total Sb present, % 0.00050 o.OO1oo o.00500 0.0100 0.0500 0.1000 0.500 1.000 o.00190 OsX124~ 0.0064, 0.0114 0.0514 0.1014 0.501 1.001 O.OO234 0.00284 0.0068., 0.0118 0.0518 0.1018 0.502 1.002 o.00096 0.0014, 0.0054, 0.0105 0.0505 0.1005 0.501 1.001 0.0005, 0.0010, 0.0050, 0.0101 0.0501 0.1001 0.500 1.000

Sb found, % Iodide method 0.0006, 0.00115 o.00520 0.010, 0.050, 0.100, 0.499 1.01, 0.00210 o.00220 o.00620 O.OlOg o.0500 0.099, 0.49, 0.99, 0.00272 0.0034, o.00670 0.0119 0.050, o.lOOo 0.49, 0.99, 0.0010, 0.0013, 0.0054, O.OlOg 0.049, 0.098~ 0.50, 0.98, 0.00tX2 0.0010, 0.00520 0.010, 0.0510 o.lOOo 0.50, 0.99,

Duplicate determinations of antimony in the Cu, MO,Zn and Pb concentrates by the proposed iodide method gave none detected and none detected, 0.0014, and 0.0014s, 0.0005, and 0.0004,, and O.OOO1,~O and none detected, respectively. Seven determinations of antimony in the nickel concentrate-ranging from 0.0011s to 0.00279%-gave a mean value of 0.0018&.

On the basis of the findings above regarding the inertness of the antimony(III + V) species to both oxidation and reduction, it is emphasized that, decomposition procedures that result in the formation of this compound, or the use of oxidants or reductants that do not completely convert it into the desired oxidation state, may ultimately cause low results for antimony if the methods used involve complexation reactions for separation purposes and/or for spectrophotometric finishes. Similarly, low results may also be obtained by such methods when antimony(V) is present in hydrochloric acid or sulphuric acid-chloride media because of the rapid formation of hydralysed species.21-23,38 Neither the formation of the the hydrolysis antimony(II1 + V) species nor products of antimony(V) in acidic chloride media affect the determination of antimony by atomic-

absorption spectrophotometry. This suggests that these species are present in a true or colloidal solution.38 It is considered that the low and erratic results usually obtained in initial tests involving the treatment of the xanthate extract with nitric, perchloric and sulphuric acids were due to the formation of both basic antimony(V) and antimony(II1 + V) compounds [or to basic antimony(V) compounds alone if aqua regia was added before evaporation to fumes of sulphur trioxide] that are not completely soluble in dilute sulphuric acid.27 Although antimony(V) compounds are appreciably soluble in potassium hydroxide solution, probably the antimony(II1 + V) compound is not, because it has been reported that the tetroxide is not very soluble in sodium hydroxide.36 This would explain why low results were often

=.

_

_

0.52 (0.500.54)

_

O.Sli, 0.510

0.005a (0.005- cO.01)

_

0.011, 0.011

0.100s

80.1 Cu, 18.1 Zn, 1.5 Al, 0.09 As 57.4 Cu, 38.0 Zn, 1.3 Mn, 1.0 Al, 1.0 Sn, 0.8 Fe 80.5 Cu, 9.4 Pb, 9.0 Sn, 0.2 P, 0.02 As

_. -, -

c>

.,A

.

.

_.

a

_

I

-

-

_i~

__,

____-_____.---__

* 95% confidence limits of the recommended mean value. t Mean of 2 values after sample decomposition with acids. $ Mean of 2 values after sample decomposition by fusion (described in Note 3). ((Mean of 10 values after sample decomposition with acids. T N.B.S. provisional result. a Certified value based on the 2 results shown in brackets. b Direot determination of antimony, i.e., without co-precipitation.

SSC-2-Copper rod SSC4-Copper rod

BCS-207 Bronze “C”

BCS-18313 Leaded gunmetal BCS-207/2 Gunmetal

BC.%183/1 Bronze

30.0 Sn, - 70 Pb, 0.13 As, 0.04 Bi 84.8 Cu, 5.0 Sn, 5.2 Zn, 3.5 Pb, 0.5 P, 0.14 As 84.5 Cu, 6.7 Sn, 3.3 Zn, 3.4 Pb, 1.5 Ni, 0.15 As 87.3 Cu, 9.7 Sn, 1.6 Zn, 0.7 Pb, 0.07 As, 0.04 Bi 86.8 Cu. 9.8 Sn, 2.5 Zn, 0.05 As -1oocu -1OOcu

0.092, 0.091

0.005

72.9 Cu. 27.1 Zn

_

_

_.

_

-

-

_

_

_

0.0005, 0.0005 0.0013, 0.0013

0.0006 0.0011

_

0.045, 0.044

0.04 (0.03-0.05)

_

-

0.093, 0.094

0.10 (0.093-0.11)

-

0.043

0.252, 0.254

0.25 (0.240.27)

_

=-

0.098

0.261

0.240 0.234, 0.239

0.24 (0.230.24)

--___*

0.797, 0.7921,

0.513

0.009

0.095

0.0036

0.011, 0.011

0.054

0.367

0.791

0.0049

0.012

0.05411,0.055$

0.370/l, 0.3711

3.587, 3.623

Sb found, % Atomic-absorption method

alloys and in commercial-purity

0.79 (0.78-0.80)

0.052 (0.050-0.055)*

0.36 (0.340.39)*

3.57?, 3.581

0.012

NBS-Cl 101 Cartridge brass B NBS-Cl 102 Cartridge brass C NBSC1120 Aluminum brass C NBS-62b Manganese bronze NBS&c Phosphor bronze bearing metal NBS-127A Solder

CZN-l-Zinc concentrate

CPB-l-Lead concentrate

3.57 (3.533.60)*

32.9 Si, 0.7 As, 5.5 Al, 1.4 Ca, 2.8 Fe 64.6 Pb, 4.4 Zn, 8.5 Fe, 17.8 S, 0.3 Cu. 0.3 Si. 0.6 Ca, 0.06 A& 0.06 As, 0.02 Bi, 0.2 Al 44.5 Zn, 7.4 Pb, 11.0 Fe, 30.2 S, 0.1 Cu, 0.5 Si, 0.2 Mn, 0.03 As, 0.1 Al 69.5 Cu, 30.3 Zn

ore

CD-l-Antimony

Iodide method

CZN-1, in N.B.S. and B.C.S. non-ferrous

Certified value and range, % Sb

CD-l, CPl3o\;s;d

Nominal composition, %

of antimony in reference ores and concentrates,

Sample

Table 4. Determimation

_-,

copper

=

.-

1009

Determination of antimony in concentrates obtained, and why the Teflon beakers used became contaminated with small amounts of antimony, when these acid mixtures were evaporated to dryness and the salts were dissolved in potassium hydroxide solution before the formation of the iodide complex. This view is supported by the fact that complete recovery of antimony is obtained when aqua regia is added after the removal of nitric acid by evaporation as described in the proposed method. Under these conditions, the antimony(II1 + V) compound is completely oxidized to antimony(V) which, on evaporation of the resultant solution to dryness, forms salts that are completely soluble in potassium hydroxide solution. Although Marenz9 suggests that decomposition. with sulphuric acid alone produces only antimony(III), this was not confirmed later” or in the present work. Low results were obtained by the’ iodide method, in tests with solutions prepared by dissolving antimony metal in hot concentrated sulphuric acid, when antimony was subsequently extracted as the xanthate in the absence of stannous chloride as reductant. This shows that some oxidized antimony was present, as otherwise the extraction would have been quantitative. Complete extraction was obtained in the presence of stannous chloride. In the present work, it is recommended that any iron that is reduced to iron(I1) during the initial reduction of antimony should be reoxidized to iron(III), before the co-precipitation step, by passing air through the solution. Antimony(III) is not oxidized to either antimony(II1 + V) or antimony(V) under these conditions. Reoxidation of the reduced iron by adding nitric acid and boiling the solution before the co-precipitation of antimony(II1) with hydrous ferric oxide has been recommended5 However, in initial work it was found that this often produces low and erratic results for antimony. It is reasonable to assume that some oxidation of antimony(II1) to both antimony(II1 + V) and antimony(V) will occur under these conditions. From the results obtained in the present investigation, it is recommended that antimony should be present as antimony(II1) before its co-precipitation with a mixture of iron(II1) and lanthanum. However, trace amounts (pg-quantities), present as antimony(V), can probably be separated in a single co-precipitation with iron and lanthanum without causing significant error in the result. Large amounts of antimony(V) should not be separated by this method because the results obtained will be low. Four tests with CD-l (100 mg), involving single co-precipitations of antimony(V) and its subsequent determination by atomic-absorption spectrophotometry, yielded results ranging from 3.33 to 3.45%. Lower results are obtained if a double precipitation is performed. The prop&d methods are suitable for samples containing 5 O.OOO1°/Oor moreof antimony. The atomicabsorption method, which involves only a single coprecipitation step, is recommended for samples containing -0.01% or more because it is considerably

simpler and much faster than the spectrophotometric iodide method. The accuracy with which antimony can be determined at the l-&g level depends primarily on the antimony content of the concentrated ammonia solution used for precipitation and neutralization purposes. This was found to vary considerably from bottle to bottle. In this work, the reagent blank obtained after single and double ammonia separations contained 5 2-3 and 3-5 vg of antimony, respectively. Acknowledgement-The author thanks P. E. Moloughney for performing residual gold, platinum and palladium analyses.

REFERENCES

1. 2. 3. 4.

E. M. Donaldson, Talanta, 1976, 23, 411. Idem., ibid., 1976, 23, 823. Idem ibid.. 1977. 24. 105. W. g. Hiliebrand, d. E. F. Lundell, H. A. Bright and J. I. Hoffman, Applied Inorganic Analysis, 2nd Ed., p. 279. Wiley, New York, 1953. 5. American Society for Testing and Materials, Chemical Analysis of Metals; Sampling and Analysis of Metal Bearina Ores. Part 12. D. 73. 1975. 6. T. Chikrab&ty and k: K. &, Z. Anal. Chem., 1968, 242, 152. 7. K. E. Burke, Anal. Chem., 1970, 42, 1536. 8. W. Reichel and B. G. Bleakley, ibid., 1974, 46, 59. 9. E. A. Jones, Natl. Inst. Met& Repub. S. Afi. Rept. No. 1787, 1976. 10. R. V. D. Robert, ibid., No. 1838, 1976. 11. M. Josephson and K. Dixon, ibid., No. 1736, 1975. 12. W. K. NQ, Anal. Chim. Acta. 1973. 64. 292. 13. S. TerasGma, Bull. Geol. S&v. Jaian,‘l975, 26, 53. 14. M. Thompson, B. Pahlavanpour and S. J. Walton, Analyst, 1978, 103, 705. 15. A. G. Fogg, C. Burgess and D. T. Burns. Talanta, 1971, 18, 1175. 16. P. P. Kish and Yu. K. Onishchenko, Zh. Analit. Khim, 1974, 29, 102; Chem. Abstr., 1974, 80, 140792m. 17. C. L. Luke, Anal. Chem., 1953, 25, 674. 18. C. L. Luke and M. E. Camubell. ibid.. 1953. 25. 1588. 19. R. W. Burke and 0. Menis,‘ibid.,’ 1966; 38, i719: 20. 4. G. Fogg, J. Jillings, D. R. Marriott and D. T. Burns, Analyst, 1969, 94, 768. 21. H. M. Neumann, J. Am. Chem. SOL, 1954, 76, 2611. 22. H. M. Neumann and R. W. Ramette, ibid., 1956, 78, 1848. 23. E. B. Sandell, Calorimetric Determination of Traces of Metals, 3rd Ed., pp. 258-265. Interscience, New York, 1959. 24. E. M. Donaldson, Talanta, 1978, 25, 131. 25. A. Elkind, K. H. Goyer and D. F. Boltz, Anal. Chem., 1953, 25, 1744. 26. Reference 23, p. 268. 27. Reference 4, p. 274. 28. I. Adamiec and Z. Marczenko, Chem. Analit. (Warsaw), 1975, 20, 985. 29. T. H. Maren, Anal. Chem., 1947, 19,487. 30. G. H. Faye, W. S. Bowman and S. Sutarno, CANMET Rept, 77-63, Department of Energy, Mines and Resources Ottawa 1977. 31. C. L. Luke, Anal. Gem., 1943, 15, 626. 32. ldem, ibid., 1944, 16,448. 33. G. L. Mason, Rept. No. MRP/MSL 74-149 (TR), Department of Energy, Mines and Resources, Ottawa, .--_ lY75.

1010

ELUE M.

34. M. B. Jacobs, Analytical Chemistry of Industrial Poisons, p. 209. Interscience, New York, 1944. 35. T. T. Gorsuch, The Destruction of Organic Matter, pp. 110411. Pergamon, New York, 1970. 36. N. Konopik and J. Zwiau&, Monatsh., 1952, 83, 189; Chem. Abstr., 1952, 45, 6025f.

DONALDSON

37. D. J. Stewart, 0. Knop and C. Ayasse, Can. J. Chem., 1972,50,690. 38. A. A. Al-Sibaai and A. G. Fog& Analyst, 1973, 98, 732.