Nov., 19581 IMPURITIES IN CARBON DIOXIDE BY GAS CHROMATOGRAPHY 609 The Determination of Molybdenum in Uranium and in Molybdenum - Uranium and Molybdenum - Niobium Mixtures BY C. 0. GRANGER (Research and Development Branch, U.K.A .E.A., Culcheth, n y . Warrington, Laws.) An improvement in the spectrophotometric determination of molybdenum by means of its complex with toluene-3 : 4-dithiol is described. This consists in forming the complex in an aqueous - solvent medium in which it is soluble, and therefore avoids the necessity for solvent extraction. When applied by difference, either in the presence or absence of uranium, the procedure has a coefficient of variation of k0.16 per cent. and can be applied to uranium - molybdenum alloys containing 0-2 per cent. or more of molybdenum.Alternatively, if a non-difference procedure is used, concentrations of molyb- denum in uranium down to 1 p.p.m. can be determined. With slight modification, the procedure is applicable to the determination of molybdenum in its alloys with niobium. The effects of impurities and of variations in temperature and acidity are discussed. A NEED arose for a method by which variations of about 1 per cent. in the molybdenum content of uranium - molybdenum alloys could be detected in samples containing only about 1 mg of molybdenum. A review of the literature indicated that established spectrophoto- metric methods, although not sufficiently precise, could be modified to suit the analytical requirements, especially if applied by difference. Toluene-3 : 4-dithiol (dithiol) , thiocyanates, phenylhydrazine and hydrogen peroxide were considered as possible developing agents, but the first-named seemed to be the most convenient, and experiments were confined to this compound.610 GRANGER: THE DETERMINATION OF MOLYBDENUM IN URAXIUM AND [Vol.83 EXPERIMENTAL Preliminary tests were made with a method in which the molybdenum - dithiol complex was precipitated, together with excess of reagent, in an aqueous hydrochloric acid medium and extracted by carbon tetrachloride, the optical density of the solvent extract being measured against the pure solvent. Recoveries by this method were not sufficiently con- sistent. Larger amounts of molybdenum were therefore taken, the optical density of the solvent extract being measured by difference against a similarly prepared molybdenum standard.Although this reduced the coefficient of variation from +la6 to jO.9 per cent., it was thought that a further reduction might result if the solvent-extraction stage could be eliminated. ELIMINATION OF SOLVENT EXTRACTION- Water-soluble solvents were added to the aqueous reaction mixture in an attempt to dissolve the molybdenum - dithiol complex, a!; well as any precipitated reagent, without separation of a second liquid phase. Ethanol and methanol did not yield clear solutions and acetone produced so much heat (presumably by reaction with hydrochloric acid) that its use was undesirable in view of the reputed instability of the reagent. With n-butyl alcohol, the resulting solution was clear and little heat wa.s evolved.n-Amy1 alcohol behaved similarly, but an extremely high concentration of hydrochloric acid and an inconveniently low water content were necessary to prevent separation of a second liquid phase. Ethyl methyl ketone, particularly at increased hydrochloric acid concentrations, yielded a single phase, but the yellow background colour of the solution was more intense than when the other solvents were used and appeared to be dependent on t'he concentrations of other reagents. When used in a difference technique, without solvent extraction, twenty 'determinations gave a coefficient of variation of k0.5 per cent. This procedure was, therefore, a significant improvement on the solvent- extraction procedure; it was also more rapid. n-Butyl alcohol was chosen as the most suitable solvent. SOLVENT AND ACID CONCENTRATIONS- Solutions of the complex ceased to obey the Beer - Lambert law at a concentration of about 60 pg per 60 ml, i.e., at an optical density of about 1.2 in 4-cm cells.I t seemed probable that, if this limit could be raised still further, improvements in precision would result. The concentration of hydrochloric acid was varied ; higher concentrations led to an increase in sensitivity, but not in reproducibility. When sulphuric acid was used in place of hydrochloric acid, the results were similar 'but less consistent. There was, however, at higher acidities than could be attained with hydrochloric acid, a decrease in the intensity of the green colour of the complex and even failure to form a colour, particularly when the solution was slightly warm.This suggested that dithiol is decomposed at high acidities and that temperatures should be kept low. When a solution that had failed to form a colour was diluted to an acidity at which development of the complex usually occurred, no colour was produced until a little further dithiol was added, thus confirming decomposition of dithiol at high acidity. The order in which the reagents were added was now changed, most of the 12-butyl alcohol being added before the dithiol. The concentration of the acid in contact with dithiol was thus greatly reduced, although the full final optimum acidity required for complete colour development was still maintained. This resulted in an increase in both degree and rate of colour formation.The effect of variation in acid concentration was again investigated, the new order of addition of reagents being used. TO 22-pg portions of molybdenum (as chloride) in 10ml of water were added different amounts of hydrochloric acid, sp.gr. 1.18, and the volumes were made up to 45 ml with n-butyl alcohol. The solutions were cooled to room temperature, and 4 ml of dithiol reagent solution were added to each. Each solution was diluted to 50 ml with n-butyl alcohol, and its optical density was measured in 2-cm cells with a Unicam spectrophotometer at 680 mp. The results were as follows- Amount of molybdenum present, pg . . . . Nil Nil 22 22 22 22 22 22 22 22 22 Amount of hydrochloric acid added, ml . . 10 26 10 12 14 16 18 20 22 24 26 Optical density . . . . 0.000 0.002 0.421 0.430 0.433 0.437 0.439 0.439 0.439 0.439 0.378Nov., 19581 611 From these results, it can be seen that the acid concentration should be kept between 16 and 24 ml; 20 ml were therefore used in all subsequent work.STABILITY OF THE COMPLEX- Development of the complex is complete in about 15 minutes and the optical density remains unchanged for at least 24 hours in the absence of uranium. The same is generally true in the presence of up to 0 6 g of uranium, but traces of opalescence have twice been observed on prolonged standing. However, neither opalescence nor change in optical density has been observed during a full working-day. IN MOLYBDENUM - URANIUM AND MOLYBDENUM - NIOBIUM MIXTURES EFFECT OF TEMPERATURE- When formed at room temperature, as in the final procedure, the complex is stable towards temperature changes.For example, when the temperature was raised to 50" C for 1 hour, no alteration in optical density was observed when re-measured at room temperature. The liquid medium in which the complex is formed has a thermal cubical expansion of 0.5 per cent. per "C, about five times that of water. Optical-density measurements at 680 mp, however, show a decrease of only 0.25 per cent. per "C when the temperature is raised, owing, presumably, to a reversible increase of 0.25 per cent. per "C in the absorption of the complex. OPTIMUM COXCEXTRATIOX OF MOLYBDEXUM FOR A DIFFERENCE PROCEDURE- When the final procedure was used, figures for the optical density of the complex over a wide range of concentrations were obtained by comparing the optical density at each level with that at the next lowest concentration, the lowest concentration of all being compared with a reagent blank solution.In this way, conformity to the Beer - Lambert law over the range 0 to 120 pg was demonstrated, with a deviation at 140 pg of 0.9 per cent. (see the first three columns of Table I). TABLE I COMPARISON OF OPTICAL DENSITY AND RELATIVE ACCURACY The optical-density measurements were made in 2-cm cells at 680 m p Mean Concentration concentration difference Relative Molybdenum Optical- of solutions between solutions accuracy concentration, Optical density compared (C), compared (AC), pg per 50 ml density difference (AA) p g per 50 ml pg per 50 ml (Z c) 0 25 50 75 100 120 140 160 180 200 - 0.250 0.501 0.752 1.001 1.199 1.389 1.525 1.642 1.750 0.250 0.251 0.251 0.249 0.198 1.190 0.136 0.117 0.108 12.5 37.5 62.5 87.5 110 130 150 170 190 25 25 25 25 20 20 20 20 20 0.125 0.377 0.628 0.871 1.089 1.235 1.020 0.995 1.026 To ascertain if an increase in precision was likely to result when the procedure was applied by difference at concentrations above 120 pg, values of a relative accuracy functionlp2 were calculated from these results.The values of this function- & C where AA is the increment of optical density for an increase AC in concentration C, are shown in Table I.612 [Vol. 83 For solutions that obey the Beer - Lambe:rt law, the relative accuracy is equal to the optical density. (The relative accuracy figures in Table I would have to be compared with the optical densities a t the mean concentrations of the solutions compared.) If, however, as the concentration increases, deviation from linearity occurs owing to a continuous decrease in absorption, the value of the function falls below that demanded by the Beer - Lambert law.However, the value of the function will continue to increase until the decrease in absorption nullifies any gain caused by increasing concentration ; beyond this, the value falls again. At this point of maximum accuracy the best precision is theoretically to be expected. In practice, consistent results could not be obtained a t concentrations above those at which the Beer - Lambert law was obeyed. For this reason it was decided that optical-density measurements would be made by difference between 100 and 11Opg of molybdenum per 50 ml, i.e., at optical densities of 2.0 to 2.2 in 4-cm cells.This range is unnecessarily narrow for most purposes, but was dictated by the analytical requirements, for which only the full precision of the method seemed likely to be satisfactory. EFFECT OF IMPURITIES- The effects of the major impurities normally to be found in uranium and uranium- molybdenum alloys were ascertained by applying the proposed method to known amounts of the impurities in the absence of both uranium and molybdenum. The results are shown in Table 11. GRANGER: THE DETERMINATION OF MOLYBDEYUM IN URANIUM AND TABLE I1 EFFECT OF 1,hlPURITIES The optical-density measurements were made in 4-cm cells a t 680 mp after the solutions had been set aside for 90 minutes Amount of Error (as Element element present, pg Optical density molybdenum), pg Iron .. . . . . .. 1120 Nil Xi1 Zinc . . .. . . * . 1000 0.027 1.4 Titanium . . * . .. 105 Nil Nil Vanadium . . . . .. 1000 0.020 1.0 Copper . . .. * . 1000 0.007 0.4 Tin . . . . * . . . 100 Xi1 Nil Cobalt . . . . . . 100 0.006 0.3 Nickel . . . . . . 500 0.004 0.2 Lead . . .. ,. 100 0.001 <0.1 Tungsten . . . . . . 200 0.028 1.4 Tungsten* . . .. . . 200 0.100 8.0 Tungsten* . . . . .. 200 0.2347 11.5 Tungsten plus 2 g of citric acid 200 0.010 0.5 * Not cooled t o room temperature before addition of dithiol reagent solution 7 Optical density measured after solution had been set aside overnight. R~ETIIOD REAGEXTS- beneath the surface of 500 ml of 1 per cent.w/v sodium hydroxide solution. with a glass rod and stir until dissolution is complete. 7 ml of 76 per cent. thioglycollic acid. a refrigerator. Dithiol reagent solution, 0.2 per cent. w/i~-Break a 1-g phial of toluene-3 : 4-dithiol Crush the solid Add slowly, with continuous stirring, Store the solution in small air-tight containers in Hydrochloric acid, s9.g. 1*18-Analytical-reagent grade. n-Butyl alcohol-Analytical-reagent grade. Molybdenum stock reference solutions-D issolve, separately, 15003 and 1.6503 g of analytical-reagent grade molybdenum trioxide, which has been freshly ignited at 500" to 525" C, in 2 M ammonia solution and gently evaporate the solutions t o small volume t o expel excess of ammonia. Cool, and dilute each solution to 1 litre. 1 ml = 1000 and 1100 pg of molybdenum, respectively.Molybdenum working Yeference solutions-Dilute aliquots of the stock reference solutions until they contain 10 and 11 pg of molybdenum per ml, respectively.Nov., 19581 PREPARATION OF SAMPLE SOLUTIONS- For samples containing less than 5 per cent. of molybdenum, take a weight of sample such that, when dissolved and conveniently diluted, a 10-ml aliquot of the solution contains 100 to 11Opg of molybdenum for the difference procedure or up to 30pg of molybdenum for comparison with water. The presence of at least 0.5g of uranium per 10ml can be tolerated in the difference procedure and 1 g in the alternative procedure. Digest the weighed sample with 25 per cent. v/v hydrochloric acid until all reaction has ceased. Add 100-volume hydrogen peroxide dropwise, with swirling, until the solution is clear.Evaporate, without boiling, just to dryness. To the residue, which is normallygreen or blue, add 10 ml of 25 per cent. v/v hydrochloric acid, and warm to ensure complete dissolutioii of molybdenum trioxide (water alone effects dissolution if the molybdenum content of the sample is less than 0.5 per cent.). Filter the solution through a 5 6 c m Whatman No. 40 filter-paper, and wash with 50 ml of 5 per cent. v/v hydrochloric acid. Use calibrated appara- tus to dilute the solution to a concentration of 100 to 110 pg of molybdenum per 10 ml for the difference procedure or up to 30 pg of molybdenum per 10 ml for the alternative procedure. Samples containing more than 5 per cent. of molybdenum are more difficult to dissolve than the lower alloys.At the 12 per cent. level, use 50 per cent. in preference to 25 per cent. hydrochloric acid both for the initial decomposition of the alloy and for dissolution of the residue after evaporation. Should the residue fail to dissolve completely, re-evaporate just to dryness, boil with water to dissolve uranyl chloride, add 5 g of sodium carbonate per 2 g of alloy, and warm. (In a few instances, some insoluble matter has remained after this treatment, but this has been shown not to contain significant amounts of molybdenum.) PROCEDURE- By pipette, place a 10-ml aliquot of the sample or standard molybdenum solution in a 50-ml calibrated flask. The aliquot should contain 100 to 110 pg of molybdenum if measurements are to be made by difference, but not more than 30 pg if measurements are to be made against water.Add 20 ml of hydrochloric acid, sp.gr. 1-18, and mix. (The total acid content should be between 19 and 21 ml of hydrochloric acid.) Add 15 ml of n-butyl alcohol from a measuring cylinder, and mix. Cool to room temperature to minimise inter- ference from tungsten. Add 4 m l of dithiol reagent solution, and mix. Adjust the tem- perature of the solution to 20" C by standing the flask in a constant-temperature bath for 30 minutes, and then dilute to the mark with n-butyl alcohol. If measurements are to be made by difference, compare the optical density of the solution in 4-cm cells a t 680 m p with that of a reference solution containing 100 pg of molybdenum, which has been prepared simultaneously in the same manner as the sample.Alternatively, measure the optical density in 4-cm cells at 680 mp against water or a reagent blank solution. If a Unicam spectrophotometer is used, the slit width must be such that a motion of the dial corresponding to 0.005 optical-density units produces a galvanometer deflection of 2 to 24- divisions. CALIBRATION- The high temperature coefficient of thermal expansion of solutions of the complex (0.5 per cent. per "C) and the (reversible) decrease in optical density with increase in tem- perature (0.25 per cent. per "C) make control of temperature important if the most precise results are to be obtained. Sample and reference solutions should be a t the same temperature a t both the final dilution and measurement stages.Either dilutions should be made a t a fixed temperature, e.g., 20" C as described under "Procedure," and optical densities measured a t as near to that temperature as possible or calibrations should be made at the same time as sample determinations. Provided that steady temperature conditions prevail, the latter alternative virtually overcomes both temperature errors and does not entail, a t least with the difference technique, much extra effort. The following procedure is favoured when measurements are made by difference- Prepare two 100-pg and two 110-pg molybdenum standards simultaneously with the samples. Use one of the 100-pg standards as the reference solution for both samples and all four standards ( i e . , compare the reference standard with itself also).From the average optical-density increment for the 10-pg difference in molybdenum, calculate the calibration factor. I N MOLYBDENUM - URAKIUM AND MOLYBDENUM - NIOBIUM MIXTURES 613 Do not bake.614 GRANGER: THE DETERMINATION OF MOLYBDENUM IN URANIUM AND [VOl. 83 With this procedure, the temperature of the water bath need not be specified and the risk of errors caused by using different slit widths for the sample and calibration measurements is avoided. RESULTS PRECISION EXPERIMENTS WITH THE DIFFERENCE PROCEDURE- Two sets of determinations were carried out under the conditions of the proposed method. In one set, which involved molybdenum only, the optical densities of 100 and 110-pg amounts of molybdenum were compared with that of a reference solution containing 100 pg of molyb- denum, The other set was similar, but uranium was present in most of the test solutions, the reference solution still containing 100 pg of molybdenum and being free from uranium.The results are shown in Table 111. TABLE I11 PRECISION OF THE DIFFERENCE PROCEDURE The optical-density measurements were made in 4-cm cells a t 680 m p Uranium absent Uranium present A I \ 7 7 Amount of Amount of Amount of molybdenum Optical-density molybdenum uranium Optical-density present, pg difference present, pg present, g difference loo* Nil 100' Nil Nil 100 - 0.002 100 Nil 0.005 100 0.003 100 Ni 1 0.007 110 0.198 100 0.5 0.010 110 0.196 100 0.5 0.012 110 0.197 110 0.5 0.207 110 0.200 110 0.5 0.208 110 0.199 110 0.5 0.206 110 0.199 110 0.5 0.207 110 0,199 110 0.5 0.204 110 0.195 110 0.5 0.207 110 0.200 110 0.5 0.209 110 0.197 110 0.5 0.209 - - 110 0.5 0.200 - - 110 0.5 0.208 * Reference :solution.For the figures in Table 111, the mean optical-density difference caused by the presence of 0.5 g of uranium corresponds exactly to the known molybdenum content of the uranium used (0.48 p.p.m.). Statistical analysis of the results indicates that, if sample determinations were to be made in duplicate and the optical-density increment caused by 10 pg of molybdenum were determined by using duplicate 100 and 110-pg standards, a coefficient of variation of 50.16 per cent. could be expected. for the average value of the molybdenum content of the sample. DETERMINATION OF TRACES OF MOLYBDENUM I:X VRANIUM- recovery experiments.of uranium. of 0.48 p.p.m. of molybdenum, was 0.010; that for the reagents was zero. were as follows- Uranium from the same source as in the previous experiment was used in a series of Amounts of molybdenum from 1 to 25 pg were added to 1-g amounts The blank value (optical density) for 1 g of uranium, caused by the presence The results Molybdenum added, pg . . * . .. 1 2 4 6 8 25 Optical density, corrected for blank value 0.019 0.039 0.080 0.120 0.162 0.500 Molybdenum found, pg .. * . . . 0.95 1.95 4.0 6.0 8.1 25.0 The proposed method was, after slight moclification, applied to niobium containing up to 10 per cent. of molybdenum. The alloy was, dissolved in a mixture of hydrofluoric and nitric acids, and the solution was evaporated to dryness.After the residue had been fused with potassium hydrogen sulphate, the melt was dissolved in a 50 per cent. w/v solution APPLICATION TO NIOBIUM ALLOYSNov., 19581 615 of citric acid, and an aliquot was treated by the difference procedure. Calibration factors and recoveries from synthetic mixtures of niobium and molybdenum corresponded closely to those for uranium alloys, CONCLUSIONS The proposed method has several advantages over the normal solvent-extraction pro- cedure for the spectrophotometric determination of molybdenum in uranium with toluene- 3 : 4-dithiol. It is more rapid, more precise and can be applied over a wider range of molybde- num contents. It is a robust method, in which neither acidity nor development time is critical. For the most precise results, however, the effects of temperature cannot be ignored, but errors from this source can be readily avoided. When the difference procedure is used, results as precise as those of a macro volumetric or gravimetric method can be obtained over a wide range. Below the level suitable for measurements by difference, good recoveries down to 1 p.p.m. can be made by using absolute measurements of optical density. Interferences are few, and it should be possible to adapt the method to determinations of molybdenum in a wide range of materials other than uranium. IN MOLYBDENUM - URANIUM AND MOLYBDENUM - NIOBIUM MIXTURES This paper is published by the kind permission of the Managing Director and the Director of Research and Development of the United Kingdom Atomic Energy Authority (Industrial Group). REFERENCES 1. 2. Hiskey, C. F., Anal. Chem., 1949, 21, 1440. Bacon, A., and Milner, G. W. C., Analyst, 1956, 81, 457 Received March 19th, 1958