ANALYST, JANUARY 1984, VOL. 109 39 Direct Complexometric Determination of Aluminium and Moderate to Low Amounts of Titanium and Iron Using Tartaric Acid as a De-masking Agent Samarendra Dasgupta and Birendra Chandra Sinha Analytical Chemistry Division, Central Glass and Ceramic Research Institute, Calcutta-700032, India Narendra Singh Rawat Department of Chemistry, Fuel and Mineral Engineering, Indian School of Mines, Dhanbad-826004, India A critical study has been made on the masking of titanium and de-masking of the Ti0 - EDTA (ethylenedi- aminetetraacetic acid) complex with tartaric acid. It reveals that tartaric acid de-masks Ti0 - EDTA selectively and quantitatively with the release of EDTA at pH 5.3 from a solution at 60-70 "C containing Ti0 - EDTA, Al - EDTA and Fe - EDTA complexes.The release of EDTA from Al - EDTA is, however, effected by the usual practice of de-masking with sodium fluoride or ammonium fluoride. Based on these observations, procedures are given for the direct complexometric determination of aluminium, titanium and iron present in the same analyte by stepwise indirect titration with EDTA, which does not involve prior separation of the metals. The method is, therefore, very simple, rapid and accurate and has been applied successfully to determine aluminium and moderate to low amounts of titanium and iron present in common aluminosilicate materials. Results for the determinations compare favourably with both the certified values and values obtained by other standard methods. Keywords: Aluminium titanium and iron determination; complexometry; eth ylenediaminetetraacetic acid; tartaric acid; aluminosilicate In common aluminosilicate materials, A1 is invariably asso- ciated with moderate to low amounts of Ti and Fe.In such a system combined A1 and Ti(A1 + Ti) and Fe are complexome- trically determined by a back-titration with ethylenediamine- tetraacetic acid (EDTA)lJ at pH 5.3. The A1 value is obtained indirectly from the total of (A1 +Ti) after deducting the Ti (equivalent to Al), determined separately by colorimetric or other methods. The limitations of these methods are as follows: the indirect determination of A1 from the total (A1 + Ti); the adverse effect of the intense yellow Fe(II1) - EDTA complex on the end-point of the titration when Fe is present in significant amounts; and the sluggish end-point caused by large amounts of Ti, although some workers3 have applied this method to the determination of Ti and A1 present in large amounts.Under such conditions, the separation of Ti and Fe is imperative and is fulfilled by a cupferron - chloroform extrac- tion4.5 Again, the complexometric determination of a large amount of Ti by an indirect titration with EDTA in the presence of H2026 requires the removal of both A1 and Fe from the system. Fe, however, offers no difficulty as it can be determined in such a mixture by direct titration with EDTA at pH 2-3 in the presence of sulphosalicylic acid.7 Pribil and Veselyg devised a method for the complexometric determina- tion of Ti, Fe and A1 when present together. Ti is separated as the hydroxide in the presence of triethanolamine, which masks the Fe and A1.Ti(OH)4 is then dissolved and Ti determined as Ti - H202 - EDTA at pH 1-2.9 From the filtrate A1 and Fe are determined by the usual back-titration procedure.However, this method has the disadvantage that Fe is also coprecipitated when present in large amounts. The method of Voinovitch et al. 10 for the complexometric determination of A1 in the presence of Ti and Fe with an excess of EDTA, using ZnC12 as the back-titrant (in the presence of tartaric acid, diammonium phosphate and fluoride) at pH 5-6 and with a dithizone indicator appears attractive. However, the absorption of the dithizone indicator, which occurs because of the precipitation of titanium phosphate (formed after phosphate addition) interferes with the end-point determination.Interference is also caused by tartaric acid, which tends to complex zinc, the back-titrant, and disturbs the end-point. Recently Tikhonovll recommended a procedure using only tartaric acid and zinc chloride in an acetate buffer (pH 5 . 9 , which was also tried initially in this work but was found to be completely unsatisfactory owing to similar interference from tartaric acid. Therefore, owing to the lack of suitable methods presented in the literature, this work has been aimed at providing a direct determination, by a stepwise indirect titration with EDTA for Ti, Fe and A1 when present together. The method developed does not involve any lengthy separation techniques. Experimental All reagents were of analytical-reagent grade.Standard EDTA solution, 0.025 M. Dissolve 9.30 g of the disodium salt of EDTA in water and dilute to 11. Standardise the solution by direct titration against standard zinc chloride solution (0.025 M) in a hexamine buffer solution (pH 5.3) with xylenol orange indicator. Lead nitrate solution, 0.025 M. Dissolve 8.28 g of lead nitrate [Pb(NO&] in water, acidified with a few drops of 8 N HN03 and dilute to 11 in a calibrated flask. Check the molarity of the solution by titration with standard EDTA solution in a hexamine buffer solution (pH 5.3) using xylenol orange as indicator. Iron(III) ammonium sulphate solution, 0.025 M. Dissolve, with warming, 6.0 g of iron(II1) ammonium sulphate [NH4Fe- (S04)2.12H20] in water containing 10 ml of 18 N HzS04.Cool and dilute to 500 ml in a calibrated flask. Check the molarity by titrating a hot (40-50 "C) solution against the standard EDTA solution at pH 2-3 using sulphosalicylic acid7 as the indicator. Titanium solution, 0.025 M. Weigh 1.0 g of titanium dioxide (Ti02) in a platinum basin. Add 20ml of 40% HF, 10 ml of 18 N H2S04, heat and evaporate the contents of the basin until copious fumes of SO3 are evolved. Cool, dissolve the titanium sulphate in 400 ml of 2 N H2SO4 on a steam-bath. Dilute the solution to 500ml in a calibrated flask with 2~ H2S04 and standardise the solution at pH 5.3 and at 30 k 5 "C in the40 ANALYST, JANUARY 1984, VOL. 109 presence of H2026 by back-titrating an added excess of EDTA with standard lead nitrate solution, using xylenol orange as indicator.Aluminium sulphate solution, 0.025 M. Weigh 5.93 g of potash alum [KAl(SO&. 12H20], dissolve in water, acidify with a few drops of 1 2 ~ HCl and dilute to 500ml in a calibrated flask. Standardise the solution at pH 5.3 by back-titrating an added excess of EDTA with standard lead nitrate solution using xylenol orange as indicator. Hexamine buffer solution, pH 5.3. Dissolve 20 g of hexam- ine in 100 ml of water and adjust the pH to 5.3 by adding 12 N HCl and checking with a pH meter. Tartrate solution. Dissolve 10 g of tartaric acid in 100 ml of water. Adjust the pH of the solution to 5.3 by adding 2 N NaOH solution and checking with a pH meter. Hydrogen peroxide, 30 volume. Hydrofluoric acid, 40%. Sulphuric acid, 18 N. Hydrochloric acid, 6 and 12 N.Nitric acid, 8 N. Potassium hydrogen sulphate, solid. Xylenol orange solution (slightly acidified) 0.2%. Ammonium fluoride, solid. Potassium hydroxide solution, 2 and 5 N. Preparation of Sample Solution Weigh 0.5g of a well ground (80-90pm) and dried (105- 110°C) sample in a platinum basin. Moisten with water and add 10 ml of 40% HF and 2 ml of 18 N H2SO4. Evaporate the contents on a sand-bath until copious fumes of SO3 are evolved and ultimately to dryness. Fuse the contents with 5 g of KHS04. Dissolve the melt with heating in water that contains 5ml of 1 2 ~ HCl. Cool the solution and dilute to 250 ml in a calibrated flask. Determination Procedure Pipette an aliquot (25-30ml) of the solution into a 250-ml conical flask and pipette an excess of EDTA (in at least a 5-ml excess over the stoicheiometric requirement for quantitative formation of the A1 - EDTA, T i 0 - EDTA and Fe - EDTA complexes) into the flask.Neutralise the solution with 2~ KOH solution in the presence of xylenol orange indicator when the solution turns slightly red. Acidify with a few drops of 1 2 ~ HC1 until the solution turns yellow. Add 20ml of hexamine buffer solution (pH 5.3), dilute to 100ml and boil the solution for 5 min. Cool to room temperature (30 k 5 "C), add 5 ml of hexamine buffer solution and titrate the standard lead nitrate solution to a sharp, red end-point. Add 20ml of tartrate solution (the colour of the solution turns yellow), boil for 5 min and titrate again with standard lead nitrate solution at 60-70 "C, the end-point being indicated by a sharp change of colour from yellow to red.Again add 2g of NH4F, boil the solution for 5min, cool to room temperature and similarly titrate the solution with lead nitrate solution. Perform a blank titration for the excess of EDTA added at the start against lead nitrate at the same pH and room temperature in the presence of xylenol orange indicator. The second and third titres correspond to Ti and Al, respectively, while the difference between the last and the sum of the first, second and third titres is a measure of Fe. by step from Al-EDTA and TiO-EDTA complexes. If, however, a complexing agent, which can preferentially de-mask either A1 or Ti from their EDTA complexes before the addition of fluoride, the problem is solved. Preliminary experiments with citric acid, oxalic acid, lactic acid and tartaric acid as de-masking agents have indicated tartaric acid to be the most promising.12 Experiments have been carried out by adding a known excess of EDTA to different concentrations of Ti solutions containing tartaric acid, boiling and back-titrating the excess of EDTA with lead nitrate solution at pH 5.3 using xylenol orange indicator and hexamine.Lead nitrate solution was selected as a back-titrant instead of zinc because the tartrate present in the solution complexes with zinc but not lead. The results for the lead nitrate solution back-titration (Table 1) correspond to the actual amount of EDTA added, indicating the complete masking of Ti with tartaric acid. Under the same conditions, experiments have also been performed to determine whether tartaric acid can de-mask Ti from its EDTA complex.A known excess of EDTA has been added in solutions containing different amounts of titanium. The solution is brought to pH 5.3 with hexamine, boiled and titrated at room temperature (30 k 5 "C) against lead nitrate solution using xylenol orange as indicator. Tartrate is then added, the solution is boiled and titrated with lead nitrate solution at the same pH and temperature to measure the release of EDTA. The lead nitrate solution titres for the second titration (Table 2) indicate a quantitative release of EDTA from the Table 1. Masking of titanium by tartrate against EDTA titration Lead nitrate (0.025 M) Ti (0.025 M) taken/ EDTA (0.025 M) for back-titration/ ml added/ml ml - 10.10 10.10 1.20 10.10 10.10 2.40 10.10 10.10 3.60 10.10 10.10 4.80 10.10 10.10 Table 2.De-masking of titanium from Ti0 - EDTA by tartrate EDTA (0.025 M) Ti (0.025 M) added in taken/ml excess/ml 1.20 10.10 2.40 10.10 3.60 10.10 4.80 10.10 Lead nitrate (0.025 M) for back- titration/ ml 8.88 7.65 6.50 5.33 Ti (0.025 M) obtained by difference/ ml 1.22 2.45 3.60 4.77 Ti (0.025 M) obtained from release of EDTA by tartrate/ ml 1.25 2.38 3.62 4.80 Results and Discussion In developing the method for the complexometric determina- tion of Al, Ti and Fe in aluminosilicates using a stepwise indirect titration with EDTA, the main problem has been how to release EDTA step by step from a mixture of A1 - EDTA, Ti0 - EDTA and Fe - EDTA. The usual addition of fluoride to release EDTA from A1 - EDTA does not release EDTA step c .- C 3.00 I I I I I I I (0 20 30 40 50 60 70 80 90 -I Titration temperature/"(= Fig.1. Effect of temperature on the quantitative release of EDTA by tartrate from Ti0 - EDTA in the presence of A1 - EDTA indicated by the lead nitrate titration of the released EDTA. The broken line represents the theoretical value for titaniumANALYST, JANUARY 1984, VOL. 109 41 Table 3. Determination of Al, Ti and Fe in synthetic solutions A1 taken/ mg 16.75 13.40 6.70 3.85 0.67 A1 found/ mg 16.73 13.42 6.70 3.84 0.69 Ti Devia- taken/ -0.02 1.44 +0.02 5.76 - 2.88 -0.01 4.32 +0.02 5.76 tion/mg mg Ti found/ mg 1.48 5.71 2.85 4.29 5.80 Fe Devia- taken/ +0.04 1.35 -0.05 4.05 -0.03 1.35 -0.03 4.05 +0.04 4.05 tiodmg mg Fe found/ Devia- mg tionlmg 1.32 -0.03 4.00 -0.05 1.40 +0.05 4.03 -0.02 4.05 - Table 4.Determination of A1203, Ti02 and Fe203 in some aluminosilicate materials Found, YO Mean, ‘/O Certified value, “/o Sample A1203 Bauxite . . . . 58.05 58.17 58.20 Flint clay (NBS 97) 38.76 38.70 38.82 Firebrick . . . . 41.15 41.13 41.22 Clay . . . . . . 34.50 34.58 34.45 Ti02 9.95 9.86 9.90 2.45 2.37 2.35 3.90 3.95 3.85 3.19 3.26 3.15 Fez03 A1203 Ti02 Fe203 A1203 Ti02 Fe203 1.90 58.14 9.90 1.89 58.25* 10.03? 1.87$ 1.82 1.95 1.05 38.76 2.39 0.97 38.808 2.409 0.980 0.90 0.95 3.60 41.17 3.90 3.65 41.05* 4.001 3.66$ 3.70 3.65 0.40 34.51 3.20 0.39 34.44* 3.12t 0.40$ 0.35 0.42 * After deducting TiOz equivalent to AI2O3 from combined (AI2O3 + Ti02) determined with EDTA. t Spectrophotometric determination with H202. $ Spectrophotometric determination with orthophenanthroline.0 Certified values from NBS. Ti0 - EDTA complex establishing quantitative de-masking of Ti. However, when A1 is present along with Ti a very interesting observation is the drifting of the end-point of the second titration at room temperature (30 f 5 “C) and the titre is less than the theoretical value indicating partial release of EDTA by tartrate from the Ti0 - EDTA complex. However, this irregularity has been overcome with a solution temperat- ure of 60-70°C in the second titration after the addition of tartaric acid. Fig. 1 shows the effect of temperature during the second titration on the stability of the Ti -tartrate complex or in other words the release of EDTA from the TiO-EDTA complex with 5.76 and 13.4 mg of Ti and Al, respectively.It is evident from Fig. 1 that the optimum temperature of the solution for titration is 60-70°C at which the end-point is stable and the correct value for Ti taken (equivalent to EDTA released on addition of tartrate) is obtained; at temperatures between 25 and 50°C, results are lower and drifting of the end-point is considerable; at temperatures >70 “C, the results approximate to the correct values but the end-point of titration is not sharp and is unsatisfactory. The release of EDTA from the A1 - EDTA complex is, however, effected by the addition of NaF or NH4F. Based on the above observations, a stepwise complexome- tric method for the direct determination of A1 and moderate to low amounts of Ti and Fe in common aluminosilicates has been developed.However, the application of the method is limited by the presence of large amounts of Ti and Fe. Mn, Ni, Co, Zn and Pb interfere by increasing the Fe value but their presence is unexpected in common aluminosilicates. Trace amounts of phosphorus, which are likely to contaminate refractory bricks, do not interfere. Table 3 shows the results of the determination of Al, Ti and Fe in synthetic solutions and Table 4 those of clays, firebricks, bauxites, etc. The results are comparable to the actual amounts taken, to the certified values or to the values obtained by other standard methods. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Brady, G. W. E., and Gwilb, J . , J . Appl. Chem., 1962,12,75. Grechko, L. I . , and Fartushnaya, L. H., Zuvod. Lab., 1980, 46, 565; Anal. Abstr., 1981,40, 2B206. Chasar, A. G., and Flork, G. R., Analysr, 1969, 94, 695. Bennett, H., Hawley, W. G., and Eardley, R. P., Trans. Br. Ceram. SOC., 1962, 61, 201. Bhargava, 0. M. P., Tulantu, 1979, 26, 146. Sinha, B. C., and Roy, S. K., J . Inst. Chem., India, 1974, 46, 19. West, T. S., “Complexometry with EDTA and Related Reagents,” Third Edition, BDH Chemicals Ltd., Poole, 1969, p. 187. Pribil, R., and Vesley, V., Talanta, 1963, 10, 383. Bieber, B., and Vecera, Z . , Collect. Czech. Chem. Commun., 1961, 26, 2081. Voinovitch, I. A., Guedon, D., and Louvrier, J . , “The Analysis of Silicates,’’ Israel Programme for Scientific Trans- lation Ltd., Jerusalem, 1966, p. 250. Tikhonov, V. N., Zh. Anal. Khim., l982,37,435;Anal. Abstr., 1982,43,6B83. Chen, Tsun-Chai, and Wei, Chung-Mei, Fen Hsi Hua Hsueh, 1979, 7, 327; Anal. Abstr., 1980, 40, 2B55. Paper A31143 Received May 19th, 1983 Accepted August 2nd, 1983