Dec., 19511 BUSH AND HIGGS 683 The Estimation of Boron in Boronised Metals BY G. H. BUSH AKD D. G. HIGGS (Presented at the meeting of the Society on Wednesday, May 2nd, 1951) The more general methods for the separation of boron from associated elements, and for its estimation, are briefly considered. Work carried out in adapting Tschischewski’s method to the estimation of boron in boronised iron, nickel, cobalt and copper is described, in which these metals are removed by mercury cathode electrolysis. In addition, a rapid method for the estimation of boron in boronised molybdenum is described in which the metal is separated as the sulphide. The boron after conversion to mannitoboric acid is finally determined by potentiometric titration with sodium hydroxide, which obviates the difficulties inherent in the use of colour-change indicators.The methods as described have an accuracy of about 1.4 per cent. of the amount of boron present over the range 0.27 to 4.30 per cent., which may be extended with slight modification. IN an investigation into the coating of iron, nickel and molybdenum with boron it was found that no approved method for the estimation of boron was available. The boronised samples were believed to contain amounts of boron varying between 1 and 10 per cent. Of the published methods for the separation of boron from associated elements that proposed by Kollock and Schmidt, referred to by Tschischewskil but without a detailed reference, appeared to be suitable to the problem in hand. Tschischewski separated boron from iron and nickel by mercury cathode electrolysis and this principle has now been extended to cover also the separation from cobalt and copper by the modification described below.In addition, an independent method has been devised for the separation of boron from molybdenum, for which the mercury cathode electrolysis is ineffective; in this the molybdenum is separated as sulphide. The object of this paper is not to discuss at length the various methods for the separation and estimation of boron, but it is appropriate to consider briefly the possible methods. Separation of boron by distillation as methyl borate-The distillation as methyl borate is perhaps the best known method for the separation of the element and it has been suggested as a method of separation from iron in It is however a tedious procedure and unsuited to the work under consideration. Separation by precipitatiotn of the associated elements--Wherry3 separated boron from iron by precipitation of the latter with calcium carbonate, and.other methods4 are described whereby iron and aluminium are precipitated with sodium hydroxide between pH 6 and pH 7.5. The large volume of the precipitates involved, with the danger of occlusion of the boron, renders these methods of doubtful value. Separation by means of mercury cathode electrolysis-The method of Kollock and Schmidt referred to by Tschischewski for the separation of boron from iron and nickel has the advantage that precipitation is avoided, the volume of solution required is small and the number of manipulative operations is reduced to a minimum.Estimations of boron-The hydrolysis of the distilled methyl borate with calcium oxide and the determination of the increase in weight due to the formation of calcium borate has been described by Gooch5 and later modified by Gooch and Jonesg The method, although capable of giving good results, is tedious. Colorimetric methods making use of curcumin, h ydroxyquinones, p-nitrobenzene-azo- chromotropic acid,7 carminic acid, etc., are mainly of use in the detection and estimation of boron in very small quantities, whilst methods proposed by Klinger,8 by Evans,g who proposed the estimation of boron in a boron alloy by the iodine liberated from potassium iodate in sulphuric acid solution, and by Glaze and Finn10 are considered less suited to the present investigation than the method to be described.684 BUSH AND HIGGS: THE ESTIMATION OF BORON PRELIMINARY EXPERIMENTAL WORK [Vol.76 ESTIMATIOK OF BOKOX IN BOKONISED IRON, NICKEL, COPPER AND COBALT- Electrolysis-An illustration of the cell used by Tschischewski is to be found in his published work, but the cell illustrated in Fig. 1 was considered to be more satisfactory. This piece of apparatus is a water-jacketed variety of the Cain cell containing an internal syphon that is closed, when the cell is held vertically, by the mercury level inside. By tilting the cell the liquid may be syphoned away through the tap in the stem. This avoids the use of an external syphon to remove the contents of the cell and facilitates washing. The cell is fitted with a piece of nickel - iron wire fused into the glass to make contact with the 50 ml of mercury used as the negative electrode.The anode is a cylinder of platinum foil capable of being rotated within the cell at about 100r.p.m. It was found that, with the exception of iron and molybdenum, all the metals deposit easily over a mercury cathode in a relatively wide range of acidity with a current of 4 amperes from a 12-volt supply. Iron, however, must be in nearly neutral solution before complete removal can be effected, and Fig. 2. Typical curve showing end-point of titration of boric acid with sodium hydroxide in presence of mannitol it is necessary to add small quantities of alkali, at frequent intervals, to neutralise the acid liberated during electrolysis.The temperature of the solution during electrolysis in an air- cooled cell may reach 65" C, but it has been found that this does not result in any appreciable loss of boron. Molybdenum cannot be deposited quantitatively at the mercury cathode. An alternative method of separation is described later in this paper. Titration of boric acid-Early titrations were carried out with p-nitrophenol as the indicator for the first stage of the titration and phenolphthalein as end-point indicator for the alkali titration of the monohydric acid formed by the addition of mannitol. It is assumed that the titration is stoicheiometric and that- The titration of boric acid with these indicators presented difficulty. The colour change was not always sharp and this could lead to appreciable personal error.The operation could be carried out with greater ease and accuracy by titrating the boric acid solution by means of a calomel - glass electrode system and a Cambridge pH meter. A boron solution was prepared by dissolving pure recrystallised borax in distilled water, free from carbon dioxide, and diluting to 1 litre in a calibrated flask to make a solution containing 0-001082 g of boron per ml. Aliquot portions of this solution were titrated with 0.10 N sulphuric acid by means of a calomel - glass electrode system and a Cambridge pH 1 ml of 0.1 N sodium hydroxide = 0.001082 g of boron.Fig. 1. Electrolytic cellDec. , 19511 I N BORONISED METALS 685 meter, Readings of pH against acid additions were recorded throughout the titration and the end-point was determined by plotting the differential and obtaining the point of inflection of the curve.This titration was used to standardist: the sulphuric acid. The pH value of the solution was then adjusted to that of the end-point, 5 g of mannitol were added and the solution titrated with 0.1 N sodium hydroxide, free from carbon dioxide, and readings of pH against alkali additions were recorded as before. The end-point was determined by noting the point of inflection of the curve. The amount of sodium hydroxide used is that required to neutralise the mannito-boric acid. A curve obtained for the end-point of this titration is shown in Fig. 2. A titration of the standardised sulphuric acid against sodium hydroxide solution was carried out by the potentiometric method and the concentration of the alkali found. The - r\ ' \ # \ I \ d \ l b - - f \ Titre equivalent to boron - I I I content of metal - '\ , -- \ .- - - , , I \ - 'I /- .d I' f ; i v0 -0 /* /* i /* i ,,I' &. ---- 1.0 0.8 0.6 0.4 D I 0.2 - - u - 2 amount of sodium hydroxide of this concentration required for the titration of the mannito- boric acid corresponded to a boron content of 0.02708 g, compared with the amount added, 0.02705 g . To apply the principle to the titration of the boric acid in the solution after removal of the associated metals by electrolysis, the pH at the beginning of the titration should be that at which the free mineral acid has been neutralised, which is determined by adjusting the solution to about pH 4.0, then titrating with sodium hydroxide and recording pH values against alkali additions.. The point at which the free mineral acid is neutralised is obtained by noting the point of inflection of the curve. It is unnecessary to adjust the pH of the solution to that of the end-point before adding mannitol and continuing the titration as described earlier. The standard sodium hydroxide corresponding to the boron present is that required to titrate the solution between the points of inflection of the two curves. A typical titration is shown in Fig. 3. Efect of glassuare-In the course of preliminary work a Pyrex beaker was inadvertently used for reducing the volume of the solution under mildly alkaline conditions and a serious error was found in the subsequent results. Boron added = 0.02040 g., Boron found = 0.02275 g.Boron added = 0.02754 g., Boron found = 0.02894 g. It is recommended that only boron-free glassware should be used. On the other hand, Pyrex and other boro-silicate glassware is reasonably resistant to acid attack and as none686 BUSH AND HIGGS: THE ESTIMATION OF BORON [Vol. 76 of the operations to be described requires the use of hot alkaline solutions, there is much less serious objection to its use. Under the conditions to be described no appreciable extrac- tion of boron takes place. Efwt of residual elements remaining after electrolysis-The presence of metals incompletely removed by electrolysis, e.g., iron, copper, nickel, cobalt, chromium and manganese, or not removed at all, e g . , aluminium, molybdenum, titanium and vanadium, may have serious effects on the titration of boric acid with sodium hydroxide; this depends on the pH range at which the corresponding metallic hydroxide - borate is formed.Residual manganese not deposited at the anode or cathode, but remaining in the solution as permanganic acid, can be removed by discharging the pink colour with sulphurous acid and boiling the slightly acid solution with potassium persulphate under a reflux condenser, when manganese will be precipitated as the dioxide. If the amount of manganese is small it is unnecessary to remove it from the solution, otherwise it may be removed by passing the cooled solution through a pulp filter. In alloys containing large amounts of aluminium the method described is inapplicable and an alternative separation is now being worked out.Blank due to reagents-The titration of the borax solution described earlier was conducted without dilution and when the same titration was carried out in a volume of 400m1, the sodium hydroxide required for neutralisation was 0.30 ml more. This amount represents the excess alkali required to change the pH of the increased volume of solution from one end of the titration range to the other. It is therefore necessary to perform a blank determination if the final volume of the solution is as much as 400ml. The blank appears to be due to (a) the volume of solution and @)-the reagents used, with possible boron pick-up from glassware. That due to (a) appears to be far the greater. The blank is determined by taking an amount of boron-free metal equal to that of the sample and treating it as follows.The acidity of the final solution is adjusted to about pH 4-0 and the solution titrated with 0.1 N sodium hydroxide to a pH slightly higher than the point of neutralisation of free acid in the sample, recording pH against titre throughout. Five grams of mannitol are then added and the titration continued to a pH just higher than that obtained for the end-point of the sample, recording pH against titre as before. In the absence of boron in the blank no depression of pH will occur on the addition of mannitol and the curve will be continuous. The amount of alkali required to titrate the blank solution over the pH range of the sample titration can be interpolated from a graphical representation of the results. The curve obtained before the addition of mannitol is used for the commencement of the titration and the curve after the mannitol addition for the end-point of the titration.In the present work the value of the blank was found to be 0.25 to 0.35 ml of 0-1 N sodium hydroxide and appeared to be independent of the amount and nature of the metal sample used. REAGENTS- METHOD All the reagents used complied with recognised analytical standards. Sulphuric acid (1 + 3)-Add 1 part of sulphuric acid to 3 parts of cold distilled water. Sulphuric acid, 0.1 N-Dilute 2.7 ml of concentrated sulphuric acid to 1 litre with distilled water, free from carbon dioxide. Standardise the solution against pure Na,B,O,. 10H,O. Sodium hydroxide, 0.1 N-Dissolve 4.00 g of sodium hydroxide in 1 litre of distilled water, free from carbon dioxide.Add 10ml of 10 per cent. barium hydroxide solution to the solution before making up to volume, and fit the bottle with a carbon dioxide trap. Standardise the solution against the standardised 0.1 N sulphuric acid reagent (above) or directly against standard borax solution as described earlier. Sodium hydroxide-A 20 per cent. solution. Sodium hydroxide-A 1 per cent. w/v solution in carbon dioxide-free water, to which add 10 ml of 10 per cent. barium hydroxide solution before making up to volume. Fit the container with a carbon dioxide trap. Sulphuric acid, 1 per cent .-Dissolve approximately 5.5 ml of concentrated sulphuric acid in 1 litre of distilled water, free from carbon dioxide.Dec., 19511 IN BORONISED METALS Hydrogen peroxide, 100-volume.Mercury-Pure redistilled metal. Mannitol-Pure reagent. 687 Distilled uater, free from carbon dioxide-Boil distilled water vigorously for 30 minutes and cool. Fit a carbon dioxide trap to the container. APPARATUS- Rejux unit-A 400-ml Erlenmeyer flask fitted to a straight tube water-cooled condenser was used in this work. Electrolytic cell-The modified Cain cell described earlier and illustrated in Fig. 1 was used. Titration unit-This consisted of a Cambridge pH meter, used in conjunction \yith a calomel positive electrode and a glass negative electrode suitable for use with solutions of pH 8 to 9. DETERMINATION OF BORON IN BORONISED IRON, NICKEL, COPPER AND COBALT- Weigh an appropriate amount of sample (not exceeding 2 g for iron samples) preferably in a fine state of division, into a 400-ml conical flask, attach the flask to the lower end of the water-cooled condenser and carefully add down the condenser column, 20 ml of diluted sulphuric acid (1 + 3) followed by 10 ml of 100-volume hydrogen peroxide, a little at a time.As the reaction may become violent, care must be taken in attacking boronised iron samples by this method. When the initial reaction has subsided, boil gently to complete the solution of the alloy and decompose the residual peroxide. The initial reaction with other alloys is not violent and heat may be applied immediately after the addition of the solvent. More peroxide may be added from time to time should the reaction become sluggish.When the sample is completely dissolved, should any residue remain, remove it by passing the cooled solution through a small filter, wash with distilled water and gently ignite the paper and contents in a platinum dish at a temperature just sufficient to destroy the organic matter. Fuse the residue with the minimum quantity of fusion mixture and dissolve out the melt in the original sulphuric acid solution. Cool the solution thoroughly and transfer it to any convenient electrolytic apparatus containing 50ml of mercury, or to the special cell described above and shown in Fig. 1. Reduce the acidity by adding 20 per cent. sodium hydroxide solution until the precipitate formed just re-dissolves. Electrolyse the solution with a current of 4 amperes from a 12-volt supply, until all trace of initial colour has disappeared.Frequent additions of sodium hydroxide will be necessary to neutralise the acid formed, more particularly when dealing with iron, otherwise complete deposition will not be effected. Continue to pass the current for 30 minutes after the solution has become colourless, or until a drop removed from the solution gives no reaction with the appropriate reagents. When electrolysis is complete, syphon off the electrolyte into a 1-litre conical flask and wash out the cell thoroughly with about 150 ml of distilled water, added in 30-mi quantities, so that the resultant volume is approximately 350ml. Add 1 per cent. sodium hydroxide solution until the electrolyte shows only a slight acid reaction to litmus paper, reconnect the flask to the reflux unit previously used and boil gently for 10 to 15 minutes to expel .all carbon dioxide, and then cool thoroughly. Transfer the liquid to a 600-ml squat-form beaker marked at a volume of 400 ml, adjust the volume up to the mark with distilled water, free from carbon dioxide, and measure the pH of the solution.Adjust the acidity to about pH 4-0 and then titrate with 0.1 N sodium hydroxide solution, recording the pH value against the titre throughout. Determine the point of neutralisation of the free mineral acid by plotting the differential and noting the point of inflection of the curve. Add 5 g of mannitol and continue the titration, recording pH against sodium hydroxide addition as previously. The end-point is at the point of inflection of the curve.The amount of 0.1 N sodium hydroxide solution required to titrate the test solution between the point of inflection of the two curves is equivalent to the amount of boron present. A blank on a sample of boron-free metal should be run under the conditions described.688 BUSH AND HIGGS: THE ESTIMATION OF BORON [Vol. 76 CALCULATION- [(mi of std. NaOH x factor of 0.1 N soln.) - ml of blank] x 0.001082 x 100 Weight of sample Boron, yo = REsurrs By the method described, the figures for recovery of known amounts of boron added as boric acid to iron, nickel, copper and cobalt were found and are shown in Tables I to IV. TABLE I RECOVERY OF ADDED .BORON FROM IRON Weight of iron taken, g 1.0 1.0 1.0 2-0 2.0 2.0 Weight of nickel taken, g 1.0 2.0 2.0 2.0 2.0 2-0 Weight of copper taken, g 1.0 1-0 1.0 2.0 2.0 2.0 Weight of cobalt taken, g 1.0 1.0 1.0 2-0 2.0 2.0 Boron added, g 0.00541 0.02705 0.03787 0.00541 0-01082 0.01623 Titration minus Blank 0-10N NaOH, ml 5.10 25.13 34.79 4.96 10.18 14.75 Boron Found, Added, 0.00552 0.541 0.02719 2.705 0.03764 3.787 0.00537 0-27 1 0.01 101 0.541 0.01596 0.812 A g % TABLE I1 RECOVERY OF ADDED BORON FROM NICKEL Boron added, g 0.03246 0.00541 0.01082 0.01623 0-02164 0.03030 Titration minus Blank.0.10 N NaOH, ml 29-70 5.08 10.36 15.64 19-84 27.87 Boron r Found, g 0.03214 0.00550 0*01121 0.01 692 0.02147 0.03016 Added, 3.246 0-27 1 0.541 0-812 1-082 1.515 % TABLE I11 RECOVERY OF ADDED BORON FROM COPPER 7 Found, Yo 0.552 2.719 3.764 0.269 0.551 0.798 7 Found, 3.214 0.275 0.561 0-846 1.074 1.608 Y O Boron added, g 0.01082 0.02164 0.03246 0.00541 0.0081 2 0-0 1623 fitration minus Blank.0.10N NaOH, ml 10.11 20.18 29.73 5.06 7.60 15.00 7 Found, g 0.0 1094 0.02 183 0.03217 0.00547 0.00822 0.01 623 Boron Added, Found, 1.082 1 -094 2.1 64 2.183 3.246 3-21 7 0.27 1 0.274 0.406 0.41 1 0.811 0.81 1 - % Yo TABLE I:V RECOVERY OF ADDED BORON FROM COBALT Boron added, g 0,01623 0.02 164 0.03030 0-00541 0.01 082 0.01623 Titration minus 0.10 N NaOH, ml 14-90 20.30 27.80 4.95 10.14 14.79 . Blank. f Boron Found, g 0.01612 0.02196 0.03008 0.00536 0.01097 0.01600 Added, 1.623 2-164 3-030 0.27 1 0.54 1 0.812 YO -b Found, Yo 1.612 2-1 96 3.008 0.268 0.549 0.800Dec., 19511 IN BORONISED METALS 689 Tables I to IV show that good agreement exists between the amount of boron added and that found, a mean value for the error over all the results being about 1-3 per cent.of the amount of boron present. The method described presents a fairly rapid and accurate method for the evaluation of the boron content of metals containing this element in amount between 0.27 and 3-76 per cent. There would, however, appear to be no reason why, with slight modifications, it should not be satisfactory for boron contents well outside this range. DETERMINATION OF BORON IN BORONISED MOLYBDENUM As molybdenum cannot be completely deposited on a mercury cathode, it was necessary to examine the precipitation methods for its separation. In addition to the occlusion of boron in bulky metal precipitates, the objection to this type of separation is the possibility of co-precipitation of metal borate taking place when the precipitation is made in alkaline solution.Precipitation in acid solution is not subject to this objection. The possibility was examined of precipitating molybdenum as sulphide and determining the boron, after elimination of the hydrogen sulphide, by titration with standard sodium hydroxide, as described earlier in this paper. The main problems investigated were solution of the sample, completeness of precipitation of molybdenum as sulphide and the blank due to the reagents used. Sohtion of the sample-The samples of boronised molybdenum received could be dissolved practically completely in aqua regia, any slight residue could be filtered, the paper and contents gently ashed at a temperature just sufficient to destroy the organic matter, the residue fused with a little fusion mixture and the melt leached out with dilute sulphuric acid.The solution so obtained could then be added to the original solution before proceeding with the separation. PRECIPITATION OF MOLYBDENUM AS SULPHIDE- Molybdenum can only with difficulty be completely precipitated by hydrogen sulphide in acid solution and it is necessary, after saturation with the gas, that the solution should be warmed under pressure. If, however, a solution of molybdenum in sodium hydroxide solution is saturated with hydrogen sulphide and then acidified with dilute sulphuric acid and heated, molybdenum trisulphide is completely precipitated in a flocculent form that settles readily. The red colour developed by the hydrogen sulpbide in alkaline solution is due to the formation of the thiomolybdate.The clear supernatant liquid from this separation was tested and found to be free from molybdenum. The thiomolybdate decomposes with sulphuric acid according to the reaction- Na2MoS4 + H2S04 = h’a,S04 + MoS, + H2S THE REAGENT BLANK- The remarks on this subject in the earlier part of this paper apply equally to this procedure and it is necessary to make a blank determination under the conditions to be described, carrying out the final titration as described earlier. METHOD REAGENTS- The following reagents are required in addition to those listed on p. 686. Bromine water-A saturated aqueous solution. Hydrogen sulphide-This should be washed by passing the gas successively through 10 per cent.hydrochloric acid and water. PROCEDURE FOR BORONISED MOLYBDENUM- Weigh 2 g of sample into a 400-ml conical flask, attach the flask to the lower end of a water-cooled condenser and carefully add 20 ml of aqua regia down the condenser column, a little at a time. When the initial reaction subsides, heat the flask and contents to effect complete solution, adding a little extra nitric acid from time to time if necessary. Should a small residue be left after the acid attack, cool the flask, dilute with 100 ml of distilled water and filter off the residue through a small pulp filter. Wash with warm distilled[Vol. 76 water and ignite gently in a platinum dish at a temperature just sufficient to destroy the organic matter. Fuse the residue with a small quantity of fusion mixture, dissolve the melt in a little dilute sulphuric acid and add it to the solution from the original acid attack.Neutralise the solution (or the combined solutions, when a fusion has been necessary) with 20 per cent. sodium hydroxide solution, adding 5 ml in excess and then saturate, in the cold, with hydrogen sulphide. When the solution has absorbed sufficient hydrogen sulphide it assumes a reddish colour owing to the formation of thiomolybdate. Neutralise the solution (in a fume cupboard) by the careful addition of diluted sulphuric acid (1 + 3). If the neutralisation is effected too rapidly, loss of sample may result from the rapid evolution of hydrogen sulphide. Add 5 ml of diluted sulphuric acid (1 + 3) in excess, fit the flask to the reflux condenser (see p.687) and bring the contents to boiling-point. Cool thoroughly, make up the contents of the flask to 500ml in a calibrated flask, return the contents to a 500-ml conical flask and leave the precipitate to settle. Decant 250ml of the supernatant liquid through a No. 41 Whatman filter-paper into a 250-ml calibrated flask. Transfer to a l-litre conical flask and add bromine water until a faint yellow colour persists. Connect to the reflux condenser and boil until the solution is colourless, to destroy any residual hydrogen sulphide and to expel bromine and carbon dioxide. Cool and transfer the solution to a 600-ml squat beaker marked at 400m1, measure the pH of the solution and adjust the acidity to about pH 4.0 by the addition of alkali, free from carbon dioxide.Dilute to 400ml with distilled water, free from carbon dioxide, and carry out the titration as described in the method for boron in iron, nickel, copper and cobalt. A blank on a sample of boron-free metal should be run under the conditions described. 690 BUSH AND HIGGS: THE ESTIMATION OF BORON CALCULATION- [(ml of std. NaOH x factor of 0.1 N soln.) - blank] x 2 x 0.001082 x 100 Weight of sample Boron, yo = RESULTS The results obtained using molybdenum with varying additions of boron as boric acid to the solution are shown in Table V. TABLE V RECOVERY OF ADDED BORON FROM MOLYBDENUM Weight of molybdenum taken, g 1.0 1.0 1.0 2.0 2.0 2.0 Boron added, g 0-00541 0.02164 0-04328 0.0054 1 0.00757 0.01623 Titration minus 0-10 N NaOH, ml 2.55 10.00 19.89 2.54 3.49 7.23 Blank.r ~~ Found, g 0.00552 0.02 164 0.04304 0.00550 0.00755 0.01564 Boron Added, 0.541 2- 164 4.328 0.27 1 0.379 0.812 - Yo v Found, % 0.551 2- 164 4.304 0.275 0.378 0*782* * The precipitation of molybdenum was incomplete in this experiment and a double treatment was necessary, hence slight loss of boron. The figures in Table V show good agreement between the amounts of boron added and those found and indicate that the analytical process described provides a satisfactory method for the determination of boron in boronised molybdenum. The method can be used without modification in the presence of small amounts of group I1 A metals, but in the presence of iron, nickel, cobalt and manganese it is necessary to remove these metals by mercury cathode electrolysis after elimination of the molybdenum and before titration.Thanks are due to the Director General of Scientific Research for permission to publish this paper.Dec., 19511 IN BORONISED METALS 691 1. 2. 3. 4. 5. 6. 7 8. 9. 10. REFERENCES Tschischewski, X., Ind. Eng. Chem., 1926, 18, 607. Lundell, G. E. F., Hoffman, J. I., and Bright, H. A., “Chemical Analysis of Iron and Steel,” Wherry, E. T., and Chapin, W. H., J . Amer. Chem. SOC., 1908, 30, 1687. Scott, W. W., and Furman, N. H., “Standard Methods of Chemical Analysis,” Fifth Edition, Volume 1, D. Van Nostrand Co., New York, and The Technical Press Ltd., London, 1939, p. 177. John Wiley & Sons Inc., New York, 1946, p. 390. Gooch, F. A., Anzer. Chem. J., 1887, 9, 23. Gooch, F. A., and Jones, L.C., “Methods in Chemical Analysis,” First Edition, John Wiley & Komarowski, A S , and Poluekstoff, N. S., MzRrochem., 1933-34, 14, 317. Klinger, P., quoted by Lundell, G. E. F., Hoffman, D. I., and Bright, H. A., “Chemical Analysis Evans, B. S., unpublished work. Glaze, F. W., and Finn, A. N., J . Res. Vat. Bur. Stand., 1941, 27, 33. Sons Inc., New York, 1912, p. 204. of Iron and Steel,” John Wiley & Sons Inc., New York, 1946, p. 396. MINISTRY OF SUPPLY ARMAMENT RESEARCH ESTABLISHMENT FORT HALSTEAD, SEVENOAKS, KENT DISCUSSION MR. G. R. BALL asked what was the useful working range of the method. He pointed out that the constant pH method of titration, in which the same pH was used for both end-points, could probably be conveniently applied to this method.This technique automatically eliminated interference from most radicals except phosphate and germanium. MR. BUSH replied that the method was not suited to the estimation of small amounts of boron in steel, but amounts of the order of 0.1 per cent. of boron could be determined readily with a titration figure about five times that of the blank. The higher range was limited by the initial weight of sample and possibly by the necessity for slight modification of the solution technique. He did not see any reason. why individual analysts should not use the method of fixed end-points if they wished. The procedure described was more fundamental and was preferred for that reason. MR. C. WHALLEY said that during the war he had had to make many analyses of boron, both in steels and in de-oxidising alloys of high silicon content.The boron content of the alloys was of the order of 2 to 3 per cent. and of the steels of the order of 0.001 per cent. To determine the lower concentrations of boron he had used the mercury-cathode separation followed by an absorptiometric measurement making use of quinalizarin. He had found no evidence of the loss of boron during the mercury-cathode separation, which conclusion the authors confirmed. He had checked his procedure from amorphous boron as distinct from solutions of boric acid and had found satisfactory results. For the de-oxidising alloys a sodium peroxide fusion had been followed by distillation of the boron as methyl borate and subsequent hydrolysis and titration. He felt that for these alloys the distillation pro- cedure was superior to the mercury-cathode separation owing to the complex nature of the materials.He asked the authors if they had had experience of the procedure of adding the mannitol in small increments during the titration rather than all a t once a t the beginning and he wondered if the results obtained by the two procedures were comparable. MR. BUSH agreed that the method used by Mr. Whalley for determining low boron contents in steel was probably one of the best for the purpose and was glad to hear that his results confirmed the authors’ experience that no appreciable loss of boron took place during mercury-cathode electrolysis. Before the introduction of the water-cooled cell, the temperature of the electrolyte often reached 60” to 70” C, but the results obtained were within the limits of experimental error. MR. HIGGS stated that during initial experiments he added mannitol in small portions, but later found that results were the same whether additions were made gradually or all at one time, providing the total addition was sufficient for the complete titration of the boric acid. Good quality mannitol was not found to give any appreciable acid or alkaline reaction. Were the boron content of the sample to exceed the limits quoted in the paper, i t might be necessary to increase the amount of mannitol used. MR. E. C. MILLS asked how far the results obtained would be representative of the whole consignment from which the sample was taken, considering the heterogeneous nature of the wire samples analysed, and also what degree of accuracy could be expected. MH. HIGGS replied that the question of a representative sample did not arise in connection with the work undertaken, since the whole sample, which weighed between 1 and 2 g, was used. If large supplies were available, the question of heterogeneity would have to be considered and would entail analyses of samples a t intervals along the specimen.