Analyst, December, 1966, Vol. 91, $9. 763-790 763 The Determination of Iron(I1) Oxide in Silicate and Refractory Materials Part 11. A Semi-micro Titrimetric Method for Determining Iron(I1) Qxide in Silicate Materials BY H . N. S. SCHAFER (Division of Coal Research, CSIRO, P.O. Box 175, Chatswood, New South Wales, Australia) A description is given of an apparatus of simple construction that can be made from plastic, in which the decomposition of silicate materials by hydrofluoric acid can be carried out. The procedure for the titration of iron(I1) with dichromate under conditions that eliminate the risk of oxidation is outlined. A comparison has been made between dichromate titres obtained potentiometrically and with diphenylamine sulphonate as indicator. Errors associated with the latter in the titration of small amounts of iron(I1) are discussed. The method of sample decomposition and titration of iron(I1) has been assessed by determining the iron(I1) oxide content of the diabase W-1, and has been applied to the determination of iron(I1) oxide in slags from a boiler fired with pulverised fuel.IN Part 1 of this paper methods are reviewed for determining iron(I1) oxide in refractory materials. A simplified analytical procedure comprising dissolution of the sample followed by direct titration of iron(I1) is now considered. Its application is restricted to silicate materials that are decomposed by hydrofluoric acid. The method described here was developed after consideration had been given to the methods of Wilson1 and Reichen and Fahey,2 both of which involve sample decomposition in the presence of an oxidant. The former method,l in which vanadate is used as added oxidant with subsequent colorimetric determination of iron( 11), gives satisfactory results when complete decomposition of the sample can be achieved at room temperature.However, most of the samples studied decomposed only very slowly under these conditions. In the method proposed by Reichen and Fahey,2 in which dichromate is used as oxidant with subse- quent titration of the excess, reproducible results could not be obtained for the titration of dichromate in the blank determination and the method was, therefore, not examined further. DEVELOPMENT OF THE METHOD APPARATUS- The apparatus for sample decomposition was required to be of simple construction, to have provision for maintaining an inert atmosphere, to be suited to operation at temperatures above room temperature and to be capable of being used as the titration vessel.The con- struction of an all-plastic apparatus satisfying most of these requirements is fully described under Experimental, and is illustrated in Fig. 1. This apparatus is similar in application to that devised by Sarver3 and further described by S~hollenberger,~ who used a platinum crucible, which enabled the solution to be boiled, for sample decomposition ; an inert atmosphere was maintained by carbon dioxide introduced above the solution through a tube in a bakelite lid. DICHROMATE TITRATION OF IRON(II)- The titration of iron(T1) with dicliromate was examined in the presence of phosphoric acid, of hydrofluoric acid after addition of boric acid and phosphoric acid, and of hydrofluoric acid.In all titrations sulphuric acid was present and sodium diphenylamine sulphonate was used as indicator, the titrations being carried out in an atmosphere of nitrogen. The same dichromate titre was obtained each time. As the titration in hydrofluoric acid solution gave sharp and distinct end-points, removal of hydrofluoric acid by forming complexes with boric acid appeared to be unnecessary. Accordingly all subsequent titrations were performed764 SCHAFER: DETERMINATION OF IRON(II) OXIDE IN [Analyst, Vol. 91 Stopper Funnel (neck from bottle) Cap (base from bott I e) Body of 100-ml bottle Fig. 1. Sample decomposition apparatus (poly- thene) directly in solutions containing hydrofluoric acid.This acid acts in the same way as phosphoric acid in lowering the redox potential of the iron(I1) - iron(II1) system. These observations confirm those of Schollenbergel"' who examined the titration of iron(I1) with dichromate, with diphenylamine as indicator. He noted that the presence of hydrofluoric acid sharpened the end-point ; and maintained that in the determination of iron(I1) oxide in minerals it was preferable to perform a direct titration with dichromate, and that the addition of boric acid to inactivate the hydrofluoric acid used in the decomposition of a silicate should be omitted. TITRATION OF DIFFERENT VOLUMES OF IRON(II) SOLUTION- The next aspect studied was the effect of varying the amount of iron(I1) titrated.Meyrowitz5 observed that the reverse titration was disproportional. There was an apparent change in the normality of an iron(I1) solution as the volume of dichromate was changed; the normality gradually increased as the volume of dichromate titrated increased. This disproportion was found to be more pronounced in large volumes (300 ml) than in small volumes (100 ml). To overcome this effect in an iron(I1) oxide determination, a sample weight was chosen that required about the same volume of dichromate as was used in the standardisation of ammonium iron( 11) sulphate. TABLE I TITRATION OF APPROXIMATELY 0-02 S IRON(I1) SULPHATE WITH 0.02 N POTASSIUM DICHROMATE UNDER NITROGEN IN A HYDROFLUORIC - SULPHURIC ACID SOLUTION (10 mi of 1 + 3 H,SO, and 5 ml of HF diluted to 25 ml) Iron( 11) solution, ml 1.00 2.00 3.00 4.00 5.00 6-00 8-00 10.00 Potentiometric titration Potassium Apparent dichromate, iron(I1) ml normality 0.97 0.01940 1.95 0.0 1950 2.93 0.01953 3.91 0.01955 4.88 0.01 952 5.85 0.01950 7.80 0.01950 9.76 0.01952 Y Diphenylamine sulphonate Potassium - s t dichromate, iron(I1) ml normality 1.01 0.02020 2.00 0~02000 2.95 0-01967 3-93 0.01965 4-89 0.01956 5.86 0.01953 7.80 0-01950 9.77 0.01954 Ferroin Potassium? - Apparent dichromatc, iron (I I) ml normality 0.98 0.0 1960 1-96 0.01960 2.94 0.01960 3.93 0.01965 4.90 0.01960 5.88 0.01960 7.84 0.01960 9-80 0.01960 * Corrected for blank of 0-02 ml of 0-02 N potassium dichromate.t Corrected for blank of 0.08 ml of 0.02 N potassium dichromate.December, 19661 SILICATE AND REFRACTORY MATERIALS. PART 11 7 65 A similar effect has been reported by Rodden6 and De Sesa7 in the determination of small amounts of uranium, in which quadrivalent uranium is reacted with iron(II1) ions, and the iron(I1) ions formed are titrated in sulphuric acid - phosphoric acid solution with dichromate.No satisfactory explanation was offered for this disproportion. Toni,8 in a study of the disproportion in uranium determinations, showed that the indicator, diphenylamine sul- phonate, was responsible, and overcame the problem by determining the end-point potentio- metrically. Earlier, Kolthoff and Sarver9 had discussed the possibility of side reactions occurring when diphenylamine was used as indicator in dichromate titrations. In the present study the titration of iron(I1) solutions in the presence of hydrofluoric acid with dichromate, with diphenylamine sulphonate as indicator, also proved to be dis- proportional for small amounts of iron(I1). In view of the findings of Toni, comparisons were made between titres obtained when the end-point was determined potentiometrically, with diphenylamine sulphonate, and with ferroin.If the disproportion observed in titrations in which diphenylamine sulphonate was used arose from side reactions occurring on oxidation of this indicator, it was thought that this would not occur with ferroin with which the colour change on oxidation is caused by the oxidation of iron(I1) to iron(II1) within the o-phenan- throline complex. The apparent iron( 11) normality, as well as the dichromate titres, obtained for different volumes of iron(I1) solution are shown in Table I.The disproportion observed in the titration of small amounts of iron( 11) with dichromate, with diphenylamine sulphonate as indicator, is not apparent when the end-point is detected potentiometrically or when ferroin is used as indicator. In the latter case an apparently higher normality for the iron(I1) solution was obtained. I t would appear that, as the potential of the iron(I1) - iron(II1) - o-phenanthroline system is about 1-06 volts, the indicator ferroin is not really suitable for use with dichromate as oxidant at the acidity used (3-6 N) when the chromium(II1) - dichromate potential is about 1-11 volts (Smith and KichterlO). Because of this the solutions were possibly over-titrated in reaching the point of colour change of the indicator, which would account for the apparently higher normality of the iron(I1) solution.Disadvantages of ferroin as indicator are the large indicator blank and the gradual colour change from the orange of the iron(I1) form through a virtually colourless stage to the blue of the iron(II1) form. The change in potential during a potentiometric titration of iron(I1) with dichromate in the presence of hydrofluoric acid is particularly large and abrupt. A typical titration graph is shown in Fig. 2. End-points were calculated from the second derivative. 0 I 2 Millilitres of 0.02 N potassi um dich romate yig. 2. Typical potentiometric titration curve766 SCHAFER: DETERMINATION OF IRON(II) OXIDE IN [Analyst, Vol.91 Plotting dichromate titres, obtained by using potentiometric end-point detection, against millilitres of iron(I1) solution (about 0.02 N) gives a linear relationship. A similar procedure in which dichromate titres, obtained with diphenylamine sulphonate as indicator, are used shows a change of slope occurring at approximately 5 ml of iron(I1) solution. By using the titres shown in Table I the following equations may be derived- P = 0.9758 M - 0.001 P = 0.9780 M - 0.006 P = 0.9761 M - 0-004 D = 0.9690 M + 0.049 D = 0-9756 M + 0.007 . . . . . . . . . . where P = dichromate titres obtained potentiometrically ; D = dichromate titres obtained with diphenylamine sulphonate; and M = millilitres of about 0.02 N iron(I1) solution. (1) (2) (3) (4) (5) (For the range 0 to 10 ml of iron(1J) solution.) (For the range 0 to 5 ml.) (For the range 5 to 10 ml.) (For the range 0 to 5 ml.) (For the range 5 to 10 ml.) From equations (2) and (4) the relationship between potentiometric and indicator titres for volumes of iron(I1) solution in the range up to 5ml is given by- Similarly, from equations (3) and (5) for volumes of iron(I1) solution in the range 5 to 10 ml, the relationship is- Thus in this range titres obtained with the indicator diphenylamine sulphonate tend to be about 0.01 ml higher than those obtained potentiometrically.This is probably because the potentiometric end-point is more easily detected than the indicator end-point, slightly more titrant being required with the latter to develop a perceptible colour change.However, P = 1.0093 D - 0.055 . . . . . . - * (6) P = 1.0005 D - 0.011 . . . . .. * * (7) TABLE I1 IRON(II) OXIDE DETERMINED IN W-1 500 ml minute per Sample Nitrogen weight, Time, flow-ra te mg minutes 100 ml 75.95 60 per 72-75 50 minute 74.75 50 75.00 60 71-22 60 69.10 60 70.10 30 74.85 60 7 1.05 60 72.15 30 77.75 60 79.95 30 80.80 30 79.25 30 69.30 60 68-70 15 70.18 60 68.00 70 70.50 60 69.80 60 67-22 120 73-40 60 71-91 30 71-76 30 73.57 30 72.84 30 73.43 30 Standard deviation . . Average . . . . .. 0.02 N Potassium dichromate, ml 4.44 4.30 4.36 4.55 4.30 4.06 4.19 4.5 1 4-28 4-36 4.69 4.89 4.93 4-83 4-18 4-19 4.28 4.14 4.28 4.25 4.06 4.47 4.36 4.34 4.48 4-43 4.44 Iron (11) oxide, per cent. 8-40 8-49 8.42 8.72 8-68 8.44 8-59 8-66 8.66 8.68 8.67 8.79 8.77 8.76 8.67 8.76 8-76 8.75 8.72 8-75 8.68 8-75 8.71 8.69 8.75 8-74 8.69 .. .... 8.72 .. . . .. 0.04 95 per cent. confidence limits for average . . 8.72 f 0.02 Iron( 11) * oxide, per cent. 8-38 8.47 8-40 8.70 8.66 8.40 8.55 8-64 8.64 8.66 8-65 8.77 8.75 8.74 8.63 8-72 8.74 8.7 1 8470 8.73 8.64 8.73 8.69 8.67 8.73 8.72 8-67 8.70 0.04 8.70 f 0.02 * Calculated from the relationship between potentiometric and diphenylamine sulphonate titres according to equation (6).December, 19661 SILICATE AND REFRACTORY MATERIALS. PART 11 767 within the limits of experimental error it is considered that for titrating volumes of iron(I1) solution in the range 5 to 10 ml, the difference in dichromate titres between the two methods is not significant. Equation (6) shows that for up to 5-ml volumes of iron(I1) solution the difference between the dichromate titres is significant.Accepting potentiometric titration as the reference method, equation (6) may be used to correct titres obtained with diphenylamine sulphonate as indicator and thus overcome the disproportion. In practice, the error in determining an iron(I1) oxide content will only be significant for low titres and low sample weights. In Table 11, in which iron(I1) oxide determined in W-1 is shown, titres were in the range of 4 to 5 ml for the sample weights taken. These titres, corrected on the basis of equation (6), are on the average 0-01 ml less than un- corrected titres, and accordingly the percentage of iron(I1) oxide is lower by 0.02. There- fore, the correction at this level is no greater than the probable experimental error.For practical purposes titres greater than 4 ml need not be corrected, but to ensure wider applica- tion of the method it is desirable to establish the relationship between potentiometric and diphenylamine sulphonate titres. EXPERIMENTAL APPARATUS- Sample decomposition apparatzcs-This was constructed from 100-In1 polythene bottles, and its structure may be clearly seen in Fig. 1. For potentiometric titrations the cap shown was replaced by one with holes for the insertion of the platinum electrode and salt bridge, and for the addition of nitrogen. Platinum electrode-This consisted of thin-sheet platinum, 15 x 5 mm, welded to a platinum wire and sealed into a glass tube, which was covered with thin-walled plastic tubing to protect it from hydrofluoric acid vapours (see Fig.3). Salt bridge-A length of polythene tubing, to one end of which a plug of tightly rolled filter-paper had been fitted, was filled with saturated potassium chloride solution. A dip-type calomel electrode could be inserted tightly into the other end of the tube (see Fig. 3). Potentiometer-Potentiomctric measurements were made with a pH meter (supplied by W. G. Pye and Co., England) in conjunction with a dip-type calomel reference electrode. Burette-A 10-ml Class A burette graduated at 0.02 ml intervals. Copper / wire V Platinum Plastic sheath D i p-type Saturated Filter-paper SGI ut ion calomel electrode potassium chloride Fig. 3. Platinum electrode and “salt bridge” REAGENTS- Hydrojuoric acid, analytical-reagent grade, 40 per cent.Sulphztric acid, 1 + 3-One volume of concentrated sulphuric acid poured into three Potassium dichromata, 0.02 N-Prepare by powdering analytical-reagent grade potassium Weigh 0.9807 g, Diphenylamiize sulflhonate solution, 0.3 per cent .-Dissolve 0.3 g of sodium diphenylamine Ferroin indicator, 0.025 M-Dissolve 1-485 g of o-phenanthroline monohydrate in 100 ml volumes of distilled water. dichromate and drying it at 120” C for 2 hours; cool it in a desiccator. dissolve it in distilled water and dilute the solution to 1 litre. sulphonate in 100ml of distilled water. of a solution containing 0.695 g of iron(T1) sulphate, I;eS0,.7H,O.768 SCHAFEII: DETERMINATION OF IRON(II) OXIDE IN [Analyst, Vol. 91 PROCEDURE- Weigh, to the nearest 0.01 mg, a 20 to 100-mg sample depending on the expected iron(I1) oxide content.Transfer to the decomposition vessel, assemble the apparatus and flush with nitrogen at a rate of about 100 ml per minute. Carefully add 10 ml of freshly boiled, 1 + 3, sulphuric acid solution, and increase the nitrogen flow to 500ml per minute. Make sure that the nitrogen inlet is below the surface of the acid. Swirl the apparatus to disperse the sample, then add 5 ml of hydrofluoric acid from a small plastic vial. Place the stopper loosely in the funnel and immerse the apparatus to about half its depth in a water-bath maintained at 80" C. During sample dissolution swirl the flask occasionally. When dissolution is complete (generally in about 30 minutes) remove the apparatus from the hot water-bath and place it in a cold water-bath.Remove the funnel, and wash down the inner surface of the vessel with freshly boiled and cooled distilled water until the volume is about 25 ml. Allow the apparatus to cool for 10 to 15 minutes. Remove it from the cooling bath, add one drop of diphenylamine sulphonate indicator and immediately titrate with dichromate until a definite purple colour is reached. Correct the volume of dichromate used by subtracting the titre obtained in a blank determination. The percentage of iron(I1) oxide in the sample is given by- Dichromate (ml) x 143.7 Sample weight (mg) For dichromate titrations of less than 4 ml, correct the titre on the basis of equation (6). If a potentiometric titration is desired, the cap fitted with platinum electrode and salt bridge is used in place of the normal cap.The electrode and bridge are withdrawn to the top of the cap during sample dissolution and lowered into the solution before titration. TABLE I11 EFFECT OF SAMPLE WEIGHT ox THE DETERMINATIOK OF IRON(II) OXIDE IN W-1 0.02 N Sample Potassium Iron(I1) oxide, End-point weight, dichromate, per cent. detection mg rnl Potentiometric 19.62 1-19 8.72 7 38-14 2.30 8.67 8-70 73.73 4.47 8.71 } Average 78.47 4.74 8.68 J Ferroin 20.29 38.58 74.59 79-30 Diphen ylamine 20.20 sulphonate 37.63 77.17 1.24 8.78 2.33 8.68 '1 8.71 4.50 ::3; J Average 4.82 1.26 8.96 8-68* 2-31 8.82 8.71* 4-88 8-71 8*70* * Corrected on basis of equation (6). RESULTS AND DISCUSSION INFLUENCE OE- KITROGEN FLOW-RATE- To prevent oxidation of iron( 11) during decomposition of a silicate material, "oxygen- free" nitrogen was used to displace air from the apparatus.On testing the method for the determination o f iron(I1) oxide in the diabase U-1, it was found that the variation in results initially obtained occurred because the nitrogen flow-rate of 100 ml per minute was insufficient, especially when, during the course of sample dissolution, the apparatus was swirled to ensure dispersion of the sample. Consistent results were obtained when the flow-rate was increased to 500 ml per minute, which ensured a good flushing of the apparatus and in addition served to keep the sample dispersed. Passing the nitrogen through a saturated iron( 11) sulphate solution and then through a bubbler containing reduced anthraquinone p-sulphonic acid did not significantly change the results as compared with those obtained by using nitrogen directly from the cylinder.Tests on an iron(I1) sulphate solution treated in the apparatus in the same way as a sample for periods up to 1 hour gave dichromate titres averaging about 99.5 per cent. of those obtained by immediate titration. The results obtained for the iron(I1) oxide content of W-1 under numerous different conditions are shown in Table 11.December, 19661 SILICATE AND REFRACTORY MATERIALS. PART 11 769 VARIATION IN SAMPLE WEIGHT- In view of the disproportion evident in the titration of iron(I1) in amounts equivalent to less than 4 ml of 0.02 N dichromate when diphenylamine sulphonate is used as indicator, various weights of W-1 were taken for the determination of the iron(I1) oxide content, with three methods of end-point detection.The results of these determinations are shown in Table 111. The iron(I1) oxide contents of 8.96 and 8.82 per cent. obtained for sample weights of approximately 20 and 38 mg, respectively, with diphenylamine sulphonate are significantly higher than the average value of 8.70 per cent. obtained for 20 determinations with sample weights averaging about 72 mg. However, if the dichromate titres found in these determina- tions with the lower sample weights are corrected on the basis of equation (B), iron(I1) oxide contents become 8.68 and 8.71 per cent., respectively, in agreement with the average value. In the determination of iron(I1) oxide in a sample with diphenylamine sulphonate as indicator, it is necessary to take a sample weight which will require a dichromate titre of not less than 4 ml of 0.02 N solution, or alternatively to correct the titre on the basis of the relationship between potentiometric and diphenylamine sulphonate titres.The use of ferroin for end-point detection does not suffer from this limitation, but the tendency towards over- titration and the difficulty of observing the exact colour change are disadvantageous. Potentio- metric titration is by far the most reliable method of end-point detection, as it is not subject to the disproportion effect when small volumes of iron(I1) are titrated, and because the potential break is so distinct. APPLICATION OF THE METHOD The method of sample decomposition and subsequent direct titration of iron( 11) with dichromate, with diphenylamine sulphonate as indicator, has been applied to the determination of iron(I1) oxide in a number of slags from a boiler fired with pulverised fuel.The results of the determinations shown in Table IV are compared with iron(I1) oxide contents obtained by Wilson’s colorimetric method. In general, the titration procedure gave slightly higher contents than the colorimetric method. Two samples allowed to stand at room temperature for 4 hours did not give complete decomposition, and a large increase in iron(I1) oxide contents was obtained in repeat determinations left overnight, CONCLUSION The determination of iron(I1) oxide in silicate materials decomposed by hydrofluoric acid can be easily performed in an inexpensive apparatus under conditions that prevent oxidation. Difficulties associated with the titration of small amounts of iron( 11) with dichromate, with diphenylamine sulphonate as indicator, can be avoided by using adequate sample weights or, if this is not possible, by application of a correction where necessary, based on the relationship between potentiometric and diphenylarnine sulphonate titres of an iron( 11) solution.For the most precise determinations potentiometric titration is recom- mended. In the determination of iron(l1) oxide in W-1, an average of 8.70 per cent, was TABLE I V DETERMINATION OF IRON(II) OXIDE IN SLAGS FROM PORT AUGUSTA POWER STATIOK, SOUTH AUSTRALIA Proposed Sample method A 5*17* 5*26* B 9-48 9-45 C 9.74 9.67 D 10-99 10-92 E 7.78 7.85 Wilson’s colorimetric method r 4 hours 16 hours - - - - - - 9.25 8.74 7-07 6-67 5.01 5.01 9.12 9.18 9.62 9.47 10.69 10.10 7.89 7.60 * Corrccted on basis of equation (6).770 SCHAFER found in 20 determinations, with a standard deviation of +0-04 per cent.The 95 per cent. confidence limits for the average are 8-70 Ifr 0.02 per cent. The preferred value for the iron(I1) oxide content of W-1 as given by Fleischer and Stevensll is 8.74 per cent., whereas Ingamells and Suhrl2 give a preferred value of 8-71 per cent. It should be possible to extend the application of the method by using an apparatus constructed of PTFE, to enable the acid to be boiled during decomposition. The author thanks Dr D. J. Swaine for helpful discussions, and Mr. R. J. Cosstick for providing results of iron( 11) oxide determinations by Wilson’s colorimetric method. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. REFERENCES Wilson, A. D., Analyst, 1960, 85, 823. Reichen, L. E., and Fahey, J . J., Bull. U.S. Geol. Surv., 1962, 1144-B. Saxver, L., J . Amer. Chem. Soc., 1927, 49, 1472. Schollenberger, C. J., Ibid., 1931, 53, 88. Meyrowitz, R., Amer. Miner., 1963, 48, 340. Rodden, C. J , , in Susano, C. D., House, 33. S., and Marler, M. -4., Editors, “First Conference on Analytical Chemistry in Nuclear Reactor Technology,” Report TlD-7555, U.S. Atomic Energy Commission, p. 25. De Scsa, M. A., in Susano, C. D., lIousc, H. S., and Marler, M. A., Editors, op. cit., p. 58. Toni, J . E. -4., Analyt. Chem., 1962, 34, 99. Kolthoff, I . , and Sarvcr, L., J . Amer. Chem. Soc., 1931, 53, 2902. Smith, G. F., and Richter, F. P., “Phenanthroline and Substituted Phcnanthroline Indicators,” G. Frederick Smith Chemical Co., Columbus, Ohio, 1944. Fleischer, M., and Stevens, K. E., Georhim. Cosmochim. Acta, 1962, 26, 525. Ingamells, C. O., and Suhr, N. H., Ibid., 1963, 27, 897. Received October loth, 1965