February, 19581 ROONEY 83 The Determination of Trace Amounts of Lead and Bismuth in Cast Iron BY R. C. ROONEY (The British Cast Iron Research Association, Bordesley Hall, Alvechurch, Birmingham) Trace amounts of lead and bismuth are separated from cast iron by extraction of the iron with isobutyl acetate from hydrochloric acid solution, and then by extraction of the lead and bismuth as their diethyldithiocarba- mate complexes with chloroform from ammoniacal tartrate - cyanide solution. The lead and bismuth are then determined simultaneously with a cathode-ray polarograph, an acidified tartrate base electrolyte being used. The purification of the reagents is described. BOTH lead and bismuth are rarely encountered in cast iron in other than trace amounts, almost invariably less than 0.1 per cent.and frequently less than 0.01 per cent. being present. At these concentrations the classical methods of separation, e.g., of lead as sulphide,l*2,3 are not satisfactory and at concentrations of less than 0-001 per cent. are completely inapplicable, even in the presence of carriers.4~~ The direct polarographic procedure6 for lead in unalloyed steel and cast iron can be used for as little as 0-001 per cent. if a modern polarograph, such as the cathode-ray polaro- graph, is available, but it is open to interference by copper and tin. Most ,cast irons contain sufficient copper to interfere. For bismuth the direct iodide colorimetric procedure7 is satisfactory down to about 0.01 per cent., but below this a preliminary separation becomes necessary. Morrogh* and DawsonQ have recently shown that very small amounts of various elements (titanium, lead, bismuth and antimony) have a markedly deleterious effect on the structure and physical properties of nodular-graphite cast iron.Their work has emphasised the need for accurate methods of determining these “subversive elements” at concentrations below 0.001 per cent. The work described in this paper has been carried out as part of the British Cast Iron Research Association’s research programme in this field. EXPERIMENTAL It has been statedlo that, in the presence of cyanide and tartrate at a pH of approxi- mately 11, sodium diethyldithiocarbamate reacts only with lead, bismuth, thallium and cadmium. This statement was investigated with reference to the extraction of microgram amounts of both lead and bismuth from pure solutions.Final determination of the two elements was to be carried out simultaneously with the cathode-ray polarograph. POLAROGRAPHY- When 0.1 N hydrochloric acid was used as base electrolyte, it was not possible to deter- mine less than 10 p.p.m. of bismuth in solution, owing to the severe distortion of the trace that occurs as the sweep passes through zero potential. The bismuth wave occurs at -0.15 volt* and is found on the steeply sloping part of the trace, which makes measurement impos- sible at low concentrations. In 0-1 N nitric acid, the wave is shifted to -0.35 volt, whereas the lead wave occurs at -0.76 volt. This bismuth wave, however, is closely followed by a large rounded wave, which has been attributed to the electro-capillary maximum.11 According to Lingane,12 the most satisfactory base electrolyte for the simultaneous determination of lead and bismuth is a solution of sodium tartrate of pH 4 to 5.When a solution that was 0.1 N in nitric acid and approximately M in sodium tartrate (pH 4-5) was used, it was found that bismuth gave a very well defined wave at -0.45 volt and lead at -0.72 volt. Minor variations of pH resulted in slight shifting of these peak potentials, but had no apparent effect on the values of the peak current. By using this base electrolyte, it was found possible to determine bismuth and lead down to as little as 0.02 p.p.m. When a 1-g sample and a final volume of solution of 5 ml are used, this corresponds to O-ooOOl per cent., which was considered a satisfactory extension * All potentials are given against the mercury-pool anode.84 ROONEY: THE DETERMINATION OF TRAC.E AMOUNTS OF [Vol.83 of the lower limit of determination for all samples, with the possible exception of very pure iron. EXTRACTION- Pure solutions of lead and bismuth were used and extraction was carried out from a solution containing 2 g of sodium tartrate, 1 g of potassium cyanide and 0.01 g of sodium diethyldithiocarbamate. The pH was adjusted to approximately 10 with ammonium hydroxide and the added lead and bismuth were extracted with chloroform. It was found to be most satisfactory to wash three times with chloroform, first with 15 ml and then with 10 and 5ml. The chloroform extracts were transferred to a 50-ml beaker and evaporated to dryness, the organic residues were destroyed.with nitric and perchloric acids and the excess of perchloric acid was removed by evaporating to dryness. The residues were dis- solved in 1.0 ml of 5 per cent. nitric acid, 4.0 ml of sodium tartrate solution (200 g per litre) were added and the solution was transferred to the polarograph cell. With the lower range of concentrations, it was found to be necessary to de-oxygenate for at least 20 minutes. The oxygen step is never completely removed, but if it is sufficiently reduced the bismuth wave can be measured. The results of this series of extractions were as follows- Lead and bismuth added, pg of each 1000 100 10.0 1.0 0.1 0.05 Lead found, pg . . .. . . 1000 100 10-5* 1*4* 0*51* 0*60* Bismuth found, pg * .. . 1000 100 9-8 0-97 0.11 0-04 * The blank value for lead in the reagents was 0.4B pg, but bismuth was not detected. The reagents used for this series of determinations were purified as described later, p. 86. As the results were satisfactory and since, under these conditions, the extraction is said to be nearly specific for lead and bismuth, a direct extraction procedure from a solution of a sample of iron was investigated. A l-g sample of pure iron and some lead and bismuth were dissolved in dilute nitric acid, the iron was complexed with tartrate and, after the pH had been adjusted, traces of other metals, e.g., copper and nickel, were complexed with potassium cyanide. Sodium diethyldithiocarbamiate was then added and the extraction was carried out as before.The results were extremely unsatisfactory; approximately 20 per cent. of both elements was recovered from each addition, the range of additions being the same as that used in the previous series of determinations. In order to complex all the iron, it had been necessary to increase the amount of tartrate and the increased buffering action had necessitated the use of more ammonium hydroxide to attain the desired pH. A systematic investigation into the effect of the concentrations of tartrate, cyanide and diethyldithiocarbamate and the pH showed that none of the com- plexing agents had any deleterious effect when present in very large amounts. The effect of up to 20 g of tartrate and 5 g of cyanide in 100 ml of solution were investigated. Between pH 7 and pH 11 the recovery was quantitative.It appears that ferric iron interferes in the extraction in some manner. A further series of additions and extractions was made, in which the iron was dissolved in hydrochloric acid and any ferric iron was reduced to ferrous iron with hydrazine dihydrochloride immediately before the extraction. The results were still very low. The next series of additions was made to iron dissolved in aqua regia, and the bulk of the iron was removed by extraction with isobiutyl acetate. Results were still low, and it was noticed that, after the addition of sodium diethyldithimarbamate, the solution acquired a blue tinge. It was thought that this was due to ferro-ferricyanide formed by the reduction of some of the ferricyanide present by the diethyldithiocarbamate.This possibility was investigated further. A further series of determinations was carried out with 4-g samples of pure iron, in which the residual iron present after the extraction with isobutyl acetate was reduced before extraction of the lead and bismuth. The results were very satisfactory and are shown in Table I. It is apparent that reduction of the iron is necessary for complete extraction. When this is done, however, the extraction of bismuth under the conditions used is quantitative at least down to 0.1 pg, and the extraction of lead at least down to 3 pg. It has been shown in our laboratory that, by using pure solutions of lead, the lead can be extracted quantitatively down to our lower limit of determination, i.e., 0.05 pg.February, 19581 LEAD AND BISMUTH IN CAST IRON TABLE I RECOVERY OF LEAD AND BISMUTH FROM PURE IRON Lead and bismuth added to 4 g of pure iron, Pg Nil 0.1 1-0 10.0 100 1000 Lead found, Pg 3.7 3.7 4.7 13.5 117 1030 Lead Bismuth recovered, found, Plz Pg - 0.18 - 0.29 1.0 1.2 9-8 10.2 113 98 1026 997 Bismuth recovered, I*g 0.11 1.02 - 10.0 98 997 The effect of ferric iron on this extraction does not appear to have been reported previously in the literature. It is suggested that the effect is due to oxidation of the sodium diethyl- dithiocarbamate by ferric iron, when it is reasonable to assume that sulphide-type compounds are formed.These sulphide-type compounds would then react with lead and bismuth ions to form unionised compounds insoluble in chloroform, thereby removing them from the solution and preventing the extraction.Some support for this theory is given by the fact that the ferro-ferricyanide colour appears when sodium diethyldithiocarbamate is added to an ammoniacal tartrate - cyanide solution containing a small amount of ferric iron, but no ferrous iron. Further, on addition of the reagent a white turbidity similar to colloidal sulphur has been observed. Finally, solutions of sodium diethyldithiocarbamate on ageing for 24 hours deposit sulphur, and in this condition, as in the presence of ferric iron, the results are low. By using the recommended procedure, a number of samples of iron were examined, and the results are shown in Table 11. Corrections for the blank values of the reagents have been applied and the results show satisfactory reproducibility, even at the lowest levels investigated, TABLE I1 Sample Mild steel .. . . Pure iron , . .. Spectrographically Grey cast iron . . pure iron . . .. REPRODUCIBILITY OF RESULTS Weight of sample taken, Lead found, g % 0.0025, 0.0025 0.0028, 0.0028 0.0073, 0.0075, 0.0070 0.0049, 0.0051, 0.0048 0.0067, 0,0067, 0.0067 1.0 { 5.0 0-00019, 0~00019 5.0 0-00009, 0~00009 0.00014, 0.00015 4.0 { 0*00013, 0.00014 Bismuth found, % 0.00004, 0-00005 0~00002, 0~00002 0.0007 1, 0-00076 0*00002, 0.00003, 0.00003 0~00001, 0~00002, 0-00002 0*000008, 0.000007 0.000004, 0.000004 0.000006, 0.000007 0.00015, 0*00015 The extraction of cadmium and thallium was also investigated, but under the conditions used the recoveries were very low, probably owing to the high concentration of ammonium ions present.Recoveries varied from 10 to 60 per cent. of the amounts added. As thallium and lead will give superimposed polarographic waves in acidified tartrate media, it is suggested that, if the presence of thallium is suspected, the lead should bedetermined in a sodium hydroxide base electrolyte. The cathode-ray polarograph will indicate the presence of thallium by giving a wave with a slight kink in the face of the peak, and will give a flat, or sometimes even a double-peak when the derivative circuit is used. Generally speaking, however, if thallium is present, it will have been added to the metal and its presence will therefore be suspected. None of the cast irons examined in our laboratory has been found to contain thallium, and cadmium has only been found at concentrations much lower than 0.0001 per cent.86 ROONEY: THE DETERMINATION OF TRACE AMOUNTS OF [Vol.83 INTERFERENCE- The two major interfering elements in the direct polarographic determination of lead in iron are copper and tin. To check the effect of these elements, 0.05 g of copper and tin, and various amounts of lead and bismuth were added to a series of 1-g samples of pure iron. The lead and bismuth were determined and corrections were applied for the blank values of the reagents, the results being as follows- Lead and bismuth added, pg of each 1000 100 10 1.0 Lead found, pg . . .. .. 1000 99 10 1.0 Bismuth found, pg .. .. 997 98 12 0.8 From these results it can be seen that there is little or no interference from these two elements.No work was done on other alloying elements, but if the concentration of other elements is sufficiently low for them to be masked by the tartrate and cyanide, there should be no interference. It has been found that, with manganese contents greater than 1 per cent., a small amount of manganese is sometimes extracted. This has no effect on the polarographic determination, unless it is present as manganese dioxide, when it colours the final solution pale yellow to brown. Under these conditions, a. wave from zero potential is obtained, but it is easily suppressed by adding 1 or 2 mg of hydrazine dihydrochloride or hydroxylamine hydrochloride. METHO:D REAGENTS- As the processing of up to 5 g of iron requires a considerable amount of reagents, the blank values are frequently so high that they render the results useless.The use of AnalaR reagents when available gave blank values of 25 pg of lead and 5 pg of bismuth. Steps were therefore taken to obtain reagents of a higher degree of purity. Nitric, hydrochloric and perchloric acids and ammonium hydroxide were originally distilled in order to purify them, but the reagents supplied as "lead free, for foodstuffs analysis'' were found to be satisfactory, having lead contents of below 0.005 p.p.m. Nitric acid, s$.gr. 1*42-"Lead free, for foo4dstuffs analysis." Nitric acid, 5 per cent.-Dilute 50 ml of the nitric acid, sp.gr. 1.42, to 1 litre with water. Hydrochloric acid, s$.gr. l*18-"Lead free, for foodstuffs analysis." Perctzloric acid, sp.gr.1.54--"Lead free, for foodstuffs analysis." Ammonium hydroxide, s$.gr. 0.880-"Lead free, for foodstuffs analysis." isoBlatyZ acetate-Carry out a blank test on each bottle by shaking 100 ml of the reagent with 10.0 ml of 2 N nitric acid and determining the lead polarographically. When necessary, purify the reagent by shaking it with 2 N nitric acid and then with 5 N hydrochloric acid to remove most of the nitric acid and leave the reagent ready to extract ferric chloride. Sodium tartrate solution, 200 g per litre-Dissolve 200 g of neutral sodium tartrate in 700 to 800 ml of water and shake with 50-ml portions of a 0.05 per cent. solution of dithizone in chloroform until there is no further colour change in the chloroform layer. Then shake the aqueous phase with chloroform until the extracts are colourless and remove the excess of chloroform by boiling.Cool and dilute the solution to 1 litre. Potassizcm cyanide solution, 200 g per litre-Dissolve 200 g of potassium cyanide in 700 to 800 ml of water and add 10 ml of a 0.1 per cent. solution of sodium diethyldithiocarbamate. Extract the solution with three 50-ml portions of chloroform and then dilute to 1 litre. This solution must be freshly prepared. Sodium diethyldithiocarbamate solzction, 0.1 per cent.--Dissolve 0.1 g of the pure salt in 100 ml of water, add 2 or 3 drops of ammonium hydroxide, sp.gr. 0480, and extract with two 10-ml portions of chloroform. This solution must be freshly prepared. Chtloroform-AnalaR chloroform has not been found to contain lead or bismuth.If the presence of these elements is suspected, shake the reagent with 2 N nitric acid and then wash it with water. GENERAL REMARKS ON REAGENTS AND APPARATUS- For lead contents down to approximately 0.O001 per cent. and bismuth contents down to approximately 0*00005 per cent., AnalaR acidls and ammonium hydroxide will give satis- factory blank figures. The normal reagent grade isobutyl acetate rarely requires purification and AnalaR chloroform has never been found to be contaminated.February, 19581 LEAD AND BISMUTH IN CAST IRON 87 For all lead contents below 0.01 per cent., however, it is necessary to purify the tartrate and cyanide solutions as described. Distilled water is satisfactory in conjunction with AnalaR acids and ammonium hydroxide. For lead contents below 0.0o01 per cent.and bismuth below 040005 per cent., it is necessary to use the highest purity reagents as already described. Distilled water can be further purified by passing it through a mixed-bed de-ionising column. De-ionised water cannot, however, be used for the final solutions, as its use has been reported to cause dif€iculty with sensitive polarographs,13 and our own experience confirms this. The sodium tartrate solution and the 5 per cent. nitric acid must be made up with distilled water and the polarograph cells must be rinsed with distilled water. Apparatus must be scrupulously clean when used for the lower levels of lead and bismuth. Pyrex glass or a similar glass should be used for all apparatus, including cover- glasses. The use of soda-glass cover-glasses led to high and variable blank values.Beakers must be covered and 2 N nitric acid boiled in them for 5 minutes before use, and they should then be thoroughly rinsed with distilled or de-ionised water. If apparatus can be reserved only for determinations of low concentrations of lead and bismuth, it will be found to give lower blank values after the first few determinations. Removal of graphite and silica residues by filtration, with subsequent evaporation of the filtrates before extraction of the iron, was found to give high blank values. Filter-papers often contain microgram amounts of lead, which are difficult to remove by washing before use. For this reason, the silica and graphite residues in cast iron are removed after centrifugation. MODIFICATIONS TO PROCEDURE- The procedure should be modified to suit the iron under consideration, as folIows- (a) For lead and bismuth contents down to 0.0001 per cent.-Use a 1-g sample and AnalaR reagents.If the contents are greater than 0.01 per cent., a base electrolyte of 0.1 N nitric acid can be used, and the solution can be made up to 25 ml in a calibrated flask. However, a blank determination on all the reagents must be carried out. (b) For lead and bismuth contents below 04001 per cent. in normal irons and steels-Use a 4-g sample and purified reagents. (c) For pure iron with low concentrations of manganese and other constituents-Omit centrifugation and use the entire 5-g sample. Transfer the dissolved residue directly to the separating funnel, hydrochloric acid being used to assist the transfer.The volume of tartrate solution can be reduced to 5ml and the cyanide solution to between 1 and Zml, and the solution made just ammoniacal before extraction of lead and bismuth. These modifications help to give a lower blank value. ( d ) For materials with low concentrations of silicon that do not contaifi graphite or other insoluble residzces-Omit centrifugation. PROCEDURE- Weigh 5 g of sample in a 400-ml squat beaker and dissolve it in 35 ml of hydrochloric acid and 10 ml of nitric acid, sp.gr. 1-42. When dissolved, evaporate to dryness, but do not allow the residue to bake. Dissolve the residue in 20 to 30ml of hydrochloric acid, with gentle warming if necessary, and cool. By using hydrochloric acid contained in a polythene squeeze wash bottle as wash liquid, transfer the solution to a 100-ml calibrated flask (or direct to a separating funnel, see modification (c)).Dilute to the mark with hydrochloric acid and mix well. Transfer the solution to a clean dry centrifuge tube and spin in a centrifuge at 3000 r.p.m. and 10-cm radius for 3 to 5 minutes. By means of a pipette, put either 20 ml (for the 1-g sample) or 80 ml (for the 4-g sample) into a pear-shaped separating funnel of suitable size and add isobutyl acetate (50 ml for the 1-g sample or 150 ml for the 4 or 5-g sample). Shake well, allow to separate and run off the acid phase into a 150-ml beaker. Evaporate just to dryness. Dissolve the residue in 10 drops of hydrochloric acid and add 15 ml of water. Heat to between 80" and 90" C and add approximately 0-1 g of hydrazine dihydrochloride; maintain at 80" to 90" C for about 3 minutes in order to reduce all the iron, and then cool.Transfer the solution to a 150-ml pear-shaped separating funnel, using water to assist the transfer. Add 10 ml of sodium tartrate solution, 30 ml of ammonium hydroxide, 10 ml of potassium cyanide solution and 10 ml of sodium diethyldithiocarbamate solution, shaking the separating funnel after each addition.88 SAINT: A MICRO PROCEDURE FOR THE ELECTROLYTIC [Vol. 83 Run off the chloroform layer into a 50-ml beaker. !Shake the aqueous layer with a 10-ml and then a 5-ml portion of chloroform and add these extracts to the contents of the beaker. If the lead or bismuth content is in excess of 0.011 per cent., use two 15-ml portions before the 10-ml and 5-ml portions and evaporate the earlier portions before adding the next.Evaporate the combined chloroform extracts to dryness, remove the beaker from the hot-plate and add 2.0 ml of nitric acid, sp.gr. 1-4-2, and 2-01 ml of perchloric acid. With the beaker covered, evaporate gently to fumes of perchloric acid and continue until all the organic material is destroyed. Cool and dissolve the residue in 1.0rnl of 5 per cent. nitric acid, added from either a l-ml pipette or a semi-micro burette, and add 4.0ml of sodium tartrate solution from a semi-micro burette. Transfer the solution to a polarograph cell and pass nitrogen through it for between 3 and 20 minutes depending upon the amount of lead and bismuth present. Record polarograms for bismuth, peak potential at -0.45 volt, and lead, peak potential at -0.72 volt, against the mercury-pool anode. For the lower range of concentrations a cathode-ray polarograph will be required. I thank the Director and Council of the €Wish Cast Iron Research Association for Add 15 ml of chloroform, shake for 1 minute and allow the layers to separate. Remove the cover and evaporate to dryness. permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. REFERENCES Pigott, E. C., “Ferrous Analysis. Modern Practice and Theory,” Second Edition, Chapman & Hall Ltd., London, 1953, p. 234. British Standard 1121 : Part 1C: 1943. United Steel Companies Ltd., “Standard Methods of Analysis of Iron, Steel and Ferro-alloys,” Rosenquist, I. T., Amer. J . Sci., 1942, 240, 356; see especially p. 359. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Second Edition, Interscience United Steel Companies Ltd., op. cit., p. 138. Westwood, W., and Mayer, A., “Chemical Analysis of Cast Iron and Foundry Materials,” Allen & Unwin Ltd., London, 1951, p. 76. Morrogh, H., B.C.I.R.A. J . Res. and Dev., 1952, 4, 292. Dawson, J. V., Ibid., 1956, 6, 180. Bode, H., 2. anal. Chem., 1955, 144, 165. Davis, T. R., personal communication. Lingane, J. J., Ind. Eng. Chem., Anal. Ed., 1943, 15, 582. Ferrett, D. J., Milner, G. W. C., and Smales, A. A., Analyst, 1954, 79, 731. Sheffield, 1951, p. 37. Publishers Inc., New York and London, 1950, p. 390. Received June 3rd, 1957