Anahst, October, 1968, Vol. 93, $$. 633-637 633 An Automatic Andy tical Procedure for Colorimetric Determination of Molybdenum the in Steel BY K. BRAITHWAITE AND J. D. HOBSON (Dunford Hadfields Ltd., East Hecla Works, Shefield S9 1TZ) A spectrophotometric method for the determination of molybdenum in steel, based on the colour reaction with thiocyanate in the presence of tin(I1) chloride and a titanium catalyst, has been adapted for the Technicon AutoAnalyzer. The method is suitable for most steels, but small corrections may be needed in the presence of high concentrations of chromium, cobalt and vanadium. Nickel, manganese and silicon a t the usual levels do not interfere. For low-alloy steels the initial solution is identical with that required for an existing automatic method for manganese and phosphorus, but is modified when required to deal with high-speed steels containing alloying additions of tungsten.Results on standard steels are presented, and a statistical survey of long-term reproducibility is given. AN important factor in the development of rapid steelmaking during recent years has been the application of analytical techniques based on physical methods, for example, automatic ultraviolet and X-ray fluorescence spectrometry. Such physical methods depend on the availability of standard specimens accurately analysed by chemical procedures, which are also needed for the analysis of specimens of non-standard size or metallurgical condition. It is also prudent and necessary to check a proportion of the physical analyses by chemical means in order to detect errors and drifts in both systems.The automation of chemical analysis offers advantages in the use of skilled manpower, economy in time or reagents and improved reproducibility. For these reasons apparatus for automatic chemical analysis was installed in our laboratories nearly 2 years ago as an auxiliary to conventional automatic emission-spectrographic analysis. Publications by Scholes and Thulbourne have described spectrophotometric analysis of steelmaking slags1 for total iron and the oxides of aluminium, manganese and phosphorus, and steels2 for phosphorus, manganese and silicon, with the Technicon AutoAnalyzer. More recently Gale, George and Williams3 have described a method of using the instrument for the rapid determination of soluble aluminium in steel.We have successfully used Scholes’ method for manganese and phosphorus in steel in our laboratories and have extended the procedure to include the determination of molybdenum. The method is based on the satisfactory, but now superseded, British Standard Method B.S. 1121 : Part 34 : 1955, in which a sulphuric acid solution of the steel is prepared, molybdenum and iron(II1) thiocyanates are formed, and the former is measured photo- metrically, after the latter has been selectively reduced by tin(I1) chloride, in aqueous ~olution.~ EXPERIMENTAL For economy in time and reagents it is desirable to use aliquots from the same solution of steel as prepared for the determination of phosphorus and manganese. The determination of phosphorus requires “reflux fuming” of the solution to convert phosphorus into orthophos- phoric acid, and chromium is at the same time oxidised to chromate.It is necessary to reduce 0 SAC and the authors.634 BRAITHWAITE AND HOBSON: AN AUTOMATIC ANALYTICAL PROCEDURE [Analyst, Vol. 93 this ion by a suitable reductant, and because iron is necessary for the formation of the molybdenum thiocyanate complex, the natural choice is an addition of iron(I1) sulphate during the mechanised manipulation. Preliminary “static” experiments with a Spekker absorptiometer showed that deter- minations made by the British standard procedure gave a stable colour after about 1 minute, although the procedure specifies a standing time of 15 minutes. This has an important practical bearing on the amount of delay required in the manifold designed for automatic analysis.The first manifold was designed to pump reagents at rates proportional to the additions used in the B.S. method and with a sample input of 0 4 m l per minute. The recording colorimeter gave straight plateaux and base-line under these conditions, but was over- sensitive, 0.5 per cent. of molybdenum giving an optical density of 1.3. Use of a smaller sampling tube to reduce sensitivity produced problems in the segmenta- tion of the reagents by air bubbles. Initially it was not possible to add enough air to maintain a satisfactory bubble pattern throughout the train but, when a second addition of air was made, interference of the two trains of bubbles produced unsatisfactory recorder traces.The problem was solved by adding a diluent acid at the point where the sample solution and air were mixed; the number of reagent additions was minimised by including the titanium catalyst in the diluent acid. However, an attempt to include the ammonium thiocyanate in the same reagent was unsuccessful because of the slow liberation of hydrogen sulphide during storage, followed by interaction with tin(I1) chloride and precipitation of tin(I1) sulphide in the AutoAnalyzer manifold. A further attempt to eliminate one mixing coil by combining ammonium thiocyanate and titanium sulphate solutions immediately prior to injection into the manifold was unsuccessful because base-lines and plateaux were not stable. Eventually considerable modifications had to be made to all the reagent concentrations to obtain a sample solution procedure compatible with the published methods for manganese and phosphorus in steel,2 covering the desired range of molybdenum content, and giving stable recorder traces.Much interaction was found between the reagent concentration variables, and the method that follows was reached after considerable experimentation. METHOD The manifold found to be satisfactory is illustrated in Fig. 1. It requires the following reagents . REAGENTS- Titanium sulfihate - acid mixture-Dissolve 1.5 g of titanium metal in 200 ml of sulphuric acid (1 + 7) and cool. To 200 ml of water add cautiously and with stirring 160 ml of per- chloric acid (sp.gr. 1.54) and 100ml of sulphuric acid (sp.gr. 1-84), and cool. Add the titanium sulphate solution and dilute to 1 litre. Iron solution-Dissolve 10 g of pure iron in hydrochloric acid (sp.gr.1-16), oxidise by dropwise addition of nitric acid (sp.gr. 1-42), add 130 ml of perchloric acid (sp.gr. 1.54) and evaporate to fumes of perchloric acid. Cover, and allow to reflux for 5 minutes. Cool, re- dissolve in water, filter if necessary, and dilute to 1 litre. Titanium - iron dilution solzltion-Mix 150 ml of the titanium sulphate - acid mixture, 100 ml of iron solution and 100 ml of water, then add 2.3 g of ammonium iron(I1) sulphate, shake to dissolve and dilute to 500 ml. Ammonium thiocyanate solution-Dissolve 200 g of ammonium thiocyanate in water and dilute to 1 litre. Tin(II) chloride solution-Transfer 40 g of tin(I1) chloride (SnC12.2H20) to a 400-ml beaker, add 10 ml of hydrochloric acid (sp.gr.1-16) and digest until a clear solution is obtained. Add 120 ml of water, cool, filter and dilute to 200 ml. Solvent acid-Add 100 ml of nitric acid (sp.gr. 1.42) to 80 ml of water, cool, add 800 ml of perchloric acid (sp.gr. 164) and dilute to 1 litre. Standard molybdenum solution-Dissolve 0-5 g of pure molybdenum in 15 ml of nitric acid (spgr. 1.20) and 10 ml of hydrochloric acid (sp.gr. 1-16>. Cool and dilute to exactly 500 ml.October, 19681 FOR THE COLORIMETRIC DETERMINATION OF MOLYBDENUM IN STEEL 635 Titanium -iron 20 per cent Ammonium thiocyanate 20 per cent. Tin(l I ) chloride 1 I Fig. 1. AutoAnalyzer flow system for the determination of molybdenum in steel CALIBRATION- Reserve one specimen as a blank, and to the remaining flasks add 1.0, 2.5, 3.5, 5-0, 6.5, 8.0 and 10.0 ml of standard molybdenum solution, equivalent to 0.10, 0.25, 0.35, 0.50, 0.65, 0.80 and 1-00 per cent.molybdenum, respectively. Follow the procedure described and use the peaks on the recorder trace to construct a calibration graph. Weigh 1-g portions of pure iron into eight 350-ml conical flasks. PROCEDURE FOR DISSOLVING THE SAMPLE- (a) Low alloy steels-Weigh 1 g of steel sample into a 350-ml conical flask and dissolve it in 15ml of solvent acid. Evaporate to fumes of perchloric acid and allow to reflux for 4 to 5 minutes. Cool, re-dissolve in 50 ml of water, transfer to a calibrated flask and dilute to 100ml. (b) Highly alloyed steels not containing tungsten-These can be dissolved in aqua regia and treated with 13 ml of perchloric acid (sp.gr. 1-54}, then fumed as before.The automated additions contain sufficient iron(I1) sulphate to deal with steels containing up to 25 per cent. of chromium after perchlorate oxidation. Aliquots from the solution of steel prepared as described in (a) and (b) are suitable for AutoAnalyzer determination of phosphorus and manganese by published methods,2 provided that the original step for arsenic separation, depending on volatilisation with hydrobromic acid, is re-inserted if phosphorus is to be determined. (c) Steels containing alloying amounts of tungsten-Dissolve 0-5 g of sample in hydro- chloric acid and digest until a clear solution is obtained, then oxidise by dropwise addition of nitric acid (spgr.1-42). Add 10 ml of sulphuric acid (1 + 1) and evaporate to fumes, fume gently for 2 minutes, cool, add 30 ml of water and boil until the salts have re-dissolved. Cool, add 10 ml of ammonium citrate solution (500 g per litre), 1 ml of saturated aqueous sulphur dioxide solution, and neutralise with ammonia solution (sp.gr. 0*91), adding 3 to 4 ml636 BRAITHWAITE AND HOBSON: AN AUTOMATIC ANALYTICAL PROCEDURE [Analyst, Vol. 93 excess to dissolve the tungstic acid, if necessary warming the solution. Neutralise the solution with sulphuric acid (1 + 1) with litmus paper as indicator, and add 25 ml in excess. Cool, transfer to a 100-ml calibrated flask, dilute to the mark and mix. (This solution is unsuitable for phosphorus and manganese determinations.) PROCEDURE FOR DETERMINING MOLYBDENUM- Assemble the manifold shown in Fig.1 and set the repetition rate at 40 samples per hour. If the steel solution is not clear, because of suspended particles of silica, filter a small portion through a dry filter-paper when filling the sample cups for the automatic sampler. Read the absorption of each sample from the recorder chart and convert molybdenum content into percentage by reference to a calibration graph prepared from calibration solutions analysed in a similar manner. INTERFERING ELEMENTS Elements occurring in steel that may cause interference include chromium, nickel, vanadium and cobalt. The effect of these and other elements has been investigated. Correct results for molybdenum at zero or 060 per cent. level were obtained from synthetic solutions containing additional reagents to simulate the presence of up to 30 per cent.of nickel, 30 per cent. of manganese or 3 per cent. of silicon. The interference from the background colour caused by tervalent chromium was small and can usually be neglected for low-alloy steels, 1 per cent. of chromium being equivalent to 0.001 per cent. of molybdenum. If a background compensation is needed, however, blank determinations can be obtained by substituting water for 20 per cent. ammonium thiocyanate solution a t the manifold, and repeating the passage of the solutions to be analysed. The text of B.S. 1121 : Part 34: 1955, as amended in 1961, states that a small correction is needed if cobalt is present (0.005 per cent. of molybdenum per 1 per cent.of cobalt). The following results in Table I show the effect of cobalt additions to a 0.48 per cent. standard molybdenum steel. The error corresponds to an apparent rise in molybdenum content of 0.002 per cent. per 1 per cent. of cobalt present. TABLE I EFFECT OF COBALT ADDITIONS ON APPARENT ANALYSIS OF 0-48 PER CENT. MOLYBDENUM STEEL 0-49 0.51, 0.54 : { "0:2 0.51 0.52 0.54 Cobalt addition, per cent. 0 10 20 30 Molybdenum per cent. Found, per cent. . . The British Standard method also states that vanadium causes interference in a rather complex manner depending on both the molybdenum and the vanadium contents of the steel. Our experiments gave the results shown in Table I1 when vanadium was added to two standard steels and to pure iron. TABLE I1 EFFECT OF VANADIUM ADDITIONS ON APPARENT ANALYSES OF MOLYBDENUM STEELS Vanadium added, per cent.0 0-20 0.40 0.80 1:oo Pure iron . . .. .. .. 0 0-005 0.008 0.010 0.013 B.C.S. 320 (0-22per cent. of molybdenum) . . 0.23 0.23 0.235 0.23 0.23 Dunford Hadfields Ltd. steel (0.48, per cent. of molybdenum) . . .. .. . . 0.49 0.50 0.50 0.50 0.52 The interference seems to be rather unpredictable, but barely significant when vanadium is less than 0.20 per cent., which it will be for most low-alloy steels. At higher levels the calibration solutions should be matched to the type of alloy to be analysed. RESULTS ON STANDARD STEELS- Table I11 shows the results of analyses for molybdenum with the AutoAnalyzer compared with the certificate value of several steels supplied by the Bureau of Analysed Standards Ltd.October, 19681 FOR THE COLORIMETRIC DETERMINATION OF MOLYBDENUM IN STEEL 637 TABLE I11 RESULTS OBTAINED FOR MOLYBDENUM IN MILD AND LOW-ALLOY STEELS WITH THE AUTOMATED PROCEDURE Molybdenum, per cent.Auto- Certificalte Steel No. Steel type Analyzer value B.C.S. 320 Mild steel 0.22 0.22 B.C.S. 321 with 0.07 0.068 B.C.S. 322 residuals 0.045 0.045 B.C.S. 324 series 0.10 0.10 B.C.S. 324 0.17 0.17 B.C.S. 325 0-16 0-16 B.C.S. 342 Ferritic 0.695 0.69 B.C.S. 254 Low alloy* 1.30 1.29 B.C.S. 241/1 Tungsten 0.52 0.52 B.C.S. 24 1 /2 high-speed 0-53 0.53 B.C.S. 482 chromium, 0.27 0-27 B.C.S. 483 cobalt and 0-18 0.17 B.C.S. 484 vanadium t 1.09 1.07 B.C.S. 485 0.68 0.67 * 0.5 g of steel plus 0.5 g of pure iron. t 0.5 g of steel dissolved by Procedure (c).stainless B.C.S. 481 steels with 0.22, 0.22 REPRODUCIBILITY- Results that have been obtained over a period of 7 months on four standard steels, which have accompanied batches of routine determinations of molybdenum by the Auto- Analyzer procedure, have been submitted to statistical examination and are summarised in Table IV. TABLE IV LONG TERM REPRODUCIBILITY OF THE AUTOMATED PROCEDURE Molybdenum, per cent. Number of Steel No. Steel type results B.C.S. 320 Mild steels with 48 B.C.S. 324 residuals 23 B.C.S. 342 Ferritic stainless 20 Dunford 1.5% 21 Hadfields 1.5% Cr, standard 0.7% Ni Range of results, minimum to Standard maximum deviation 0.19 to 0.23 0.006 0.16 to 0.19 0.006 0.67 to 0.70 0.014 0.48 to 0.51 0.01 1 Mean Accepted results value 0.219 0.22 0.17, 0.17 0.6S8 0-69 0 ~ 4 9 ~ 0.48, TypicaI coefficients of variation are found to be 2 to 3 per cent. of molybdenum content; these are similar to the variation from conventional “static” colorimetric methods. With the present method, 1.0 per cent. molybdenum gives an optical density of about 0-85, and the exponential response enables 0.01 per cent. molybdenum to be readily observed. How- ever, by appropriate changes in pumping rates or in sample weight or dilution, the manifold could be adapted to measure molybdenum content in steel over smaller or greater ranges. The authors thank the Directors of Dunford Hadfields Ltd. for permission to publish this work. REFERENCES 1. 2. -,- , Ibid., 1964, 89, 466. 3. 4. Scholes, P. H., and Thulbourne, C., Analyst, 1963, 88, 702. Gale, P., George, A. I., and Williams, W. D., Technicon European Symposium on Automation British Standards Institution, B.S. 1121 : Part 34 : 1955, and amendment published September Received March 25th, 1968 in Analytical Chemistry, Brighton, 13-1 5 November, 1967. llth, 1961.