150 STOCKDALE : THE ESTIMATION OF IRON BY DICHROMATE Pol. 75 The Estimation of Iron by Dichromate BY D. STOCKDALE SYNOPSIS-The relative merits of diphenylamine, diphenylbenzidine, barium diphenylamine sulphonate, ferrous phenanthroline and potassium ferricyanide as indicators in the estimation of iron by dichromate have been examined. The source of the iron was either ferrous ammonium sulphate or an ore, and the titrations were made in both the presence and in the absence of ortho- phosphoric acid. All these indicators proved efficient when the conditions were suitable, Barium diphenylamine sulphonate was found to be the best, because its colour change is marked and because the end-point that it gives is the least affected by the conditions of the titration. NUMEROUS indicators, both internal and external, and a wide range of conditions have been recommended for this estimation.Because of certain discrepancies in the literature and occasional failures in the laboratory, it seemed desirable to make a systematic examination of the position. The indicators used were as follows- Diphenylamine (DA) . . . . Diphenylbenzidine (DB) . . . . Barium diphenylamine sulphonate Ferrous phenanthroline (FeP) . . Potassium ferricyanide . . . . Dilute solutions, used externally. ELECTROCHEMISTRY Since the pioneering work of J. H. Hildebrandl the electrochemistry of the oxidation of iron by dichromate and of the indicators has been much explored, notably by I. M. Kolthoff2J94 and G. F. Smith6 and their collaborators. Only the principal points need be recapitulated here.When a bright platinum wire is placed in a solution of ferrous and ferric ions, the half-cell potential at 25" C. is given by 0.04 ml. of a 1 per cent. solution in concentrated 0-04 ml. of a 1 per cent. solution in concentrated sulphuric acid. sixlphuric acid. (BaDS) . . .. .. . . 1.0 ml. of a 0.2 per cent. solution in water. 0.05 ml. of an aqueous solution 0-025 M with respect to both ferrous sulphate and 1 : 10 phenanthroline monohydrate. [Fe" ] [Fe"'] E = Eo - 0.059 loglo--* E,,, the standard electrode potential at infinite dilution, is -0.771 volt.6 However, as the ferric ion is very liable to form complex ions, the formal electrode potential in a solution of significant concentration, such as would be obtained by dissolving ferrous and ferric sulphates in the ratio of their equivalent weights in N sulphuric acid, is substantially less than Eo.It is about -0.69 volt when the ferrous and ferric salts are each about 0.05 N in either N sulphuric or N hydrochloric acid. In general, the lower the pH of the solution, the lower this potential, and it can be depressed further by the addition of some compound, such as orthophosphoric acid, which forms complexes with the ferric ion even more readily than do the common strong acids. The electrode potential for the reduction of the dichromate ion, CrzO," + 14H' + 6e -+ 2Cr"' + 7H,O, is similarly given by 0.059 [Cr"']2 E = E o - - 6 loglo[Cr,O,"] [H]14 * The standard electrode potential of this reaction is -1.36 volts, but more usually the activity coefficients of the ions are far from unity, and the potential obtained by adding 50ml.ofMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 151 0.1 N dichromate to 25 ml. of 0.1 N ferrous solution is about -1.10 volts when the final concentration of strong acid is formally normal. This potential is much influenced by the pH, as the equation suggests, and will be increased by increasing the hydrogen ion concentra- tion of the solution in which the reduction is taking place. These points are illustrated by curves which have been obtained during the present investigation (Fig. 1). These curves show the results obtained when 25 ml. of 0.1 N ferrous salt in a solution of acid of the nature and concentration indicated were titrated with 50 ml. of 0.1 N potassium dichromate containing such an excess of the acid that the formal con- centration of acid remained constant at that shown for each curve until the equivalence-point .was reached, The volume at the equivalence-point was 50ml., except with phosphoric acid present, when it was 60 ml., since 10 ml.of 50 per cent. by volume of orthophosphoric acid had been added initially. The horizontal lines near the curves mark the potentials of first colour change for the diphenylamine group of indicators and the formal electrode potential of ferrous phenanthroline in N sulphuric acid. 10 3Q Milfllitrer of 0.1 N K,Cr,O, Fig. 1. Millilitres of potassium dichromate plotted against 25 ml. of 0.1 N solution of Fe". Curve (a) 0.1 N hydrochloric acid; curve (b) N sulphuric acid ; curve (c) N sulphuric + orthophosphoric acids The rate of change of potential per unit volume of dichromate was found to be greatest a t -0.92 volt in N acid in the absence of phosphoric acid, and this was accepted as the equivalence-point potential.Fig. 1 shows that the first colour change of DA and DB should occur substantially before, and that of BaDS slightly before this point. It is to be expected that the maximum intensity of colour would be reached at potentials about 0.1 volt more than those of the first colour change. If this were so, the end-points of these indicators assessed on maximum intensity, not on first colour change, would coincide substantially with the equivalence-point. These potentials were found to be -0.86 volt for DA, -0.91 volt for BaDS and -0.94 volt for DB, the first two results being in accordance with expectation.The result for DB, which should be the same as that for DA because DB is the first oxidation product of diphenylamine, is high, presumably because DB is so insoluble (0.06 mg. per litre in water at 25" C.2). Most of this indicator is precipitated when a drop of it is added to the ferrous solution. When that remaining in the solution has been oxidised, more will slowly enter into solution and be oxidised in turn. The maximum colour intensity, therefore, can be obtained only slowly, and an estimation of the oxidation potential made during a titration carried out a t a practicable rate will necessarily be high.152 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE [Vol. 75 Fig. 1 shows that FeP is an unsuitable indicator when in a normal solution of a strong acid.However, as Smith and Richter5 have shown, its oxidation potential is decreased by increasing the concentration of acid. The oxidation potential of the dichromate system is increased by the same means. It should therefore be possible to lower the line representing the potential of FeP relatively until it cuts the vertical part of the titration curve (Fig. l ) , and under this condition the indicator would prove efficient. In an experiment using 2 N sulphuric acid, FeP was fully red at -0.91 volt. The addition of 0.05 ml. of 0.1 N dichromate at 25 ml. raised the potential to -1.05 volts and caused an immediate and noticeable change in the intensity of the colour. The red tinge faded out completely in about 30 seconds.In 10 N sulphuric acid the indicator was fully red at -0.87 volt, and a further addition of 1 drop of dichromate raised the potential to -1.00 volt, causing the indicator to change colour rapidly and completely. It would seem that 2 N is approximately the minimum concentration of sulphuric acid in which it is possible to make this titration. A final con- centration oi '2.5 N was later successfully adopted as standard. Results at the lower concentrations of hydrochloric acid were similar. With 1.5 N acid a titration was just possible, provided it was made slowly. With 2.5 N acid, 1 drop changed the potential from -0.84 volt (red) to -11.06 volts (colourless), and the oxidation of the indicator took place in about 10 seconds.In 6 N solution, on the other hand, it was almost impossible to obtain a result, because the greater part of the indicator was precipitated on the electrode tubes and on the sides of the beaker, and because the yellow colour of the ferric chloride ion made the final colour change difficult to see. In these experiments, drops of the solutions were withdrawn from time to time and tested with potassium ferricyanide. It was found that in a normal solution of a strong acid the critical potential for this indicator was near -0.85 volt, a potential very slightly lower than that required Tor DA and definitely lower than those needed for the other indicators. It is also lower than the value of -0.92 volt taken as the equivalence-point potential. These results suggest that potassium ferricyanide ceased to give a blue colour when about i part per thousand of the bivalent ion remained unoxidised.This corresponds to a concentration of approximately 3 mg. per litre. This conclusion is confirmed by later work (see Tables I and 11). There is, therefore, an appreciable indicator error with potassium ferricyanide, and it can be used in exact estimations only when a solution of dichromate is used to connect a known with an unknown quantity of iron, the two ferrous solutions being of approximately the same concentration. EXPERIMENTS WITH FERROUS AMMONIUM SULPHATE Portions from 1.2 to 1.4 g. of the salt were weighed into conical flasks, dissolved in 100 ml. of approximately N sulphuric acid, and titrated with 0.1 N potassium dichromate (exact, assuming the salt to be pure).The final volume was 150ml. and the final concentration of sulphuric acid was approximately 0-65 N. In a parallel set of experiments, 10 ml. of 1 : 1 by volume orthophosphoric acid was added in each titration, the final volume again being 150ml., and the weight of sulphuric acid being the same as before. The above conditions apply for all the indicators except FeP, when 100 ml. of 4 N sulphuric acid was used to give, a final concentration of 2.5 iV with respect to this acid in a final volume of 150 ml. The results, each the arithmetical mean of four titrations, are given in Table I. TABLE I Without H,PO,, With H,PO,, rnl. ml. 25.50 f 0-01 Diphenylamine . . ,. .. .. 25-49 rt 0.015 Diphenylbenzidine .. .. .. 25.48 * 0.02 Barium diphenylamine sulphonate .. 25.49 f 0.01 25.50 rt 0.01 No result possible Ferrous phenanthroline . . .. .. 25-48 f 0.01 - Potassium ferricyanide . . .. .. 25.44 f 0.01 - Table I gives the volume in ml. of 0.1 N potassium dichromate found to be equivalent to 1 g. of ferrous ammonium sulphate. The ferrous ammonium sulphate used was referred by potassium permanganate to pure sodium oxalate. Its equivalent weight was found to be 392.9 (for pure salt, 392.1). One gram of the salt should be oxidised by 25.45 ml. of 0.1 N potassium dichromate. It seems reasonable to assume that the sample of potassium dichromate was reasonably pure and that the end-points as shown by the various indicatorsMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 163 (Table I) are all close to the true equivalence-point.The standard deviations are given in the table; they mean only that if the groups of experiments were repeated under the same conditions the probability that the arithmetical means would fall within the range indicated is about 60 per cent. It was hoped that a deviation would serve as a measure of the value of the indicator, but, surprisingly, this was far from being the case. All the deviations are small, and the spread of results in any group can be entirely accounted for by the errors inherent in volumetric analysis with the type of apparatus used. The numerical results show that all the indicators worked perfectly; personal impressions of their behaviour were quite different. Some gave clear colours, with sharp changes; other gave muddy colours with the intensity of the colour changing noticeably with time after the addition of the oxidant, in such a way that considerable personal judgment seermd to be required in making an assessment.I t seemed probable that such judgment could not be free from considerable error. It would appear, however, that satisfactory results can be obtained in volumetric analysis, even when using what seems to be an indifferent indicator and with the observer in doubt about his gauging of the end-point. NOTES ON INDICATORS Diphenylamine is first oxidised irreversibly to DB, which is then oxidised reversibly to diphenylbenzidine violet. The behaviour of BaDS is similar. The end-points given in Table I are those of maximum intensity. I t is probably advisable to assess them without reference to a blank, partly because of the difficulty of obtaining equal concentrations of indicator when such small volumes of indicator solution are used, and partly because the diphenylbenzidine violet is itself further oxidised by a small excess of dichromate to com- pounds of less intense colour. This further oxidation under the experimental conditions was slow, usually taking not less than an hour for the removal of the violet colour, and it is unlikely that it interfered with the assessment of the point of maximum intensity, but the colour of a spent and oxidised solution is often not permanent enough to be reliable as a standard in a subsequent titration.The results given in Table I show that when an end-point is obtainable reasonably quickly with one of these three indicators it makes no difference to the numerical result which indicator is used, or whether orthophosphoric acid is present or not.DIPHENYLAMINE-This indicator behaved somewhat erratically, particularly in the earlier stages of its oxidation in the absence of phosphoric acid. In two of the titrations the violet colour appeared early and deepened over a range of about 0.6 ml. in a titration of some 30 ml. Usually a dirty muddiness appeared in the solution, often as much as 1 ml. before the end-point. Despite these variations the end-point was consistent. In the presence of phosphoric acid the solution remained clear, the colours were bright, and the range was approximately 0.04 ml. In some cases, however, the colour at the end-point was blue, not violet. Titrations without phosphoric ' acid are perhaps to be preferred, because of the warning given of the approach to the end-point.DIPHENYLBENZIDINE-The colours, in the absence of phosphoric acid, were much brighter than with diphenylamine. Blue, not violet, predominated, and the blue colour changed to violet when the solution was allowed to stand after the titration. The range was approximately 0.7 ml. The titration must be carried out slowly with this indicator, because the colour deepens only slowly after the addition of the dichromate. This is perhaps due to its marked insolubility. In all cases a precipitate was produced when the indicator was added initially. Conversely, it also reacted slowly when ferrous iron was added to a solution of potassium dichromate.When phosphoric acid was present there was a further slowing of the oxidation of the indicator to such an extent that to obtain consistent end-points was impracticable. BARIUM DIPHENYLAMINE SULPHONATE-without phosphoric acid the approach to the end-point was marked by the appearance of a slight muddiness in the solution. The full rose-violet colour was developed after its first appearance by about 0.05 ml. of dichromate, and the range from first appearance of muddiness to maximum colour was about 0.1 ml. With phosphoric acid the colours were clear and the full colour was developed by about 0.03 ml. of dichromate. The indicator proved pleasant to use by either method, the element of doubt in judging the end-point being almost entirely absent.FERROUS PHENANTHROLINE-The oxidation of this indicator, and indeed of the other oxidation - reduction indicators discus~ed,~ appears to be complex, with ferrous iron playing In two others, the range was only about 0.25 ml.154 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE [Vol. 75 some part in the mechanism. It is possible to prepare a solution of FeP containing an appreciable excess of dichromate, the indicator being obviously in the reduced form, and to discharge the red colour by the addition of a small quantity of a solution containing ferrous iron. It follows that FeP is a more satisfactory indicator when the ferrous solution is run into the dichromate. This will usually entail the use of a standard iron solution in a titration, the stages being the oxidation of the iron for estimation with excess dichromate, followed by titration of this excess with standard iron. With DA and BaDS this process would seem to be unnecessary, because so little trouble arises in the direct titration, and, indeed, the direct titration with FeP is quite practicable provided a generous concentration of sulphuric acid is present.EXPERIMENTS WITH IRON ORE These experiments were made to find out whether the elements likely to be present in a mineral or the compounds introduced by its reduction with stannous chloride interfered in any.way with the indicators. The material used was Iron Ore A supplied by British Chemical Standards, and contained 58.20 per cent. of iron, together with very small quantities of sulphur, phosphorus, arsenic, copper and titanium, in addition to lime, magnesia, alumina and silica.About log. of the mixed and dried ore were digested with concentrated hydrochloric acid, potassium chlorate being added as a saturated solution from time to time in the later stages of the attack, a total of about 1 g. of the solid being used. The excess of hydrochloric acid was removed by evaporation, and the contents of the dish were baked a t 110" C. for 1 hour. The ferric chloride was extracted with 1 : 1 hydrochloric acid, and the residue was fused in a platinum crucible with 3 6 g . of sodium carbonate. After treatment with hydrochloric acid, the residue was evaporated to dryness and baked again at 110" C. for 1 hour. This second extraction in hydrochloric acid was filtered into the main solutiop, which was diluted to 1 litre.For each titration 25 ml. of this solution was used. The iron was reduced with approximately 0.4 M stannous chloride in hydrochloric acid, and the small excess of this reagent was oxidised by 0.25 g. of mercuric chloride added as a solution. The final volume was about 120 ml., and was approximately normal with respect to hydrochloric acid, except that twice this quantity of acid was used for FeP. When phosphoric acid was present the volume added was 10 ml. of 1 : 1 acid in a final volume of 120ml. Titrations were made with the sample of potassium dichromate used previously in 0.1 N solution. The results are given in Table 11. TABLE I1 DICHROMATE REQUIRED FOR 2 5 ~ ~ . OF SOLUTION OF IRON ORE Without H,PO,, With H,PO,.Diphenylamine . . .. .. .. 26.67 rt 0.005 26-61 f 0-005 Diphenylbenzidine . . .. .. . . 26.66 3z 0-005 No result possible Barium diphenylamine sulphonate .. 26-66 f 0.01 26.65 f 0.015 26.69 rt 0.015 ml. ml. - Ferrous phenanthroline . . .. .. Potassium ferricyanide . . .. .. 26-57 rt 0.03 - COMMENTS ON INDICATORS The various substances present in the more complex solution made little difference to the behaviour of the indicators. The diphenylamine group always gave a violet colour at the end-point, whereas a blue was sometimes obtained with pure ferrous ammonium sulphate. The colours faded more rapidly in the presence of a small excess of the oxidising agent. DIPHENYLAMINE-PhOSphOriC acid reduced the indicator range from about 0-2 to about 0.03 ml.It also improved the quality of the colour change, which was, however, moderately good in its absence. The reduction by the phosphoric acid of the volume of dichromate equivalent to the ore remains unexplained. The effect was checked by titrating a volume of the solution to just short of the end-point. It was then divided, and phosphoric acid was added to one half. One drop of dichromate caused this portion to change colour, but 2 drops were required to develop the full colour in the other. DIPHENYLBENZIDINE-AS before, the full violet colour was slow to develop, and therefore the final additions of the dichromate had to be made slowly. Phosphoric acid again delayedMarch, 19501 STOCKDALE: THE ESTIMATION OF IRON BY DICHROMATE 155 the oxidation to such an extent that the use of this indicator in conjunction with this acid was not possible.BARIUM DIPHEXYLAMINE SULPHONATE-The end-point in the absence of phosphoric acid appeared to be somewhat indefinite and the colours obtained were again muddy. The indicator range was about 0-25ml. In all titrations made in the presence of phosphoric acid the colours were clear and the change at the end-point was marked. The indicator range was only about 0-03 ml. in this case. FERROUS PHENANTHROLINE-The colour of the FeCL" ion masked the red colour of the indicator to such an extent that it was almost impossible to obtain an end-point. Even when phosphoric acid was present the colour change was not particularly obvious, and it was found advisable to compare the colour of the solution under titration with the colour of a similar solution containing a small quantity of ferrous iron. The end-point was taken a t the first sign of lightening.POTASSIUM FERRICYANIDE-The end-point with this particular sample of ore proved somewhat troublesome. It will be noticed that the standard deviation for this indicator, though still small, is larger than those for the others. PERCENTAGE OF IRON I N THE ORE With 25.49 ml. of potassium dichromate equivalent to 1 g. of ferrous ammonium sulphate of equivalent weight 392-9 (Table I), and 26-66 ml. of the dichromate equivalent to 25 ml. of the solution of the ore at 10.244 g. per litre (Table II), the percentage of iron in the ore is 58.05. The average value returned by the fourteen analysts co-operating with British Chemical Standards is 58.20 & 0.02.The organiser's own value is 58.09 per cent. Better agreement is not to be expected, in view of the length of the standardisation chain-sodium oxalate, potassium permanganate, ferrous ammonium sulphate, potassium dichromate, ore. Interference by the foreign substances present during the titration of the solution of the haematite can, therefore, only have been small. CONCLUSION s Of the indicators examined, barium diphenylamine sulphonate is the best for use in the estimation of iron with dichromate. Kolthoff and Sandell7 have previously arrived at the same conclusion. The indicator gives a satisfactory result when the solution is at least 0.5 N with respect to hydrochloric or sulphuric acid. If orthophosphoric acid is present the colours are clear and the range of colour change is small. In the absence of this acid, titrations must be continued until the violet colour of the indicator is fully developed. Without phosphoric acid the colours may be somewhat muddy, but the colour change takes place over a rather larger range of dichromate, so giving some warning of the approach of the end- point. Both methods give satisfactory results, and which is selected seems largely to be a matter of personal choice. REFERENCES 1. Hildebrand, J . H., J. Amer. Chem. Soc., 1913, 35, 847. 2.. Kolthoff, I. M., and Sarver, L. A., Ibid., 1930, 52, 4179. 3. Sarver, L. A., and Kolthoff, I. M., Ibid., 1931, 53, 2902. 4. Hume, D. N., and Kolthoff, I. M., Ibid., 1943, 65, 1895. 6. Smith, G. F., and Richter, F. P., Ind. Eng. Chem., Anal. Ed., 1944, 16, 580. 6. Latimer, W. M., Oxidation Potentials, Prentice-Hall, New York, 1938. 7. Kolthoff, I. M., and Sandell, E. B., Textbook of Quantitative Inorganic AnaZysis, Macmillan, New York, 1946. THE UNIVERSITY CHEMICAL LABORATORY CAMBRIDGE June, 1949