August, 196 11 OF NITROGEN I N FLUORINE-CONTAIKING COMPOUNDS Determination of Added Borates in Mixed 517 Fertilisers BY H. N. WILSON AND G. U. M. PELLEGRINI (Imperial Chemical Industries Ltd., Billingham Division, P.O. Box No. 6, Billingham, Co. Duvham) A method for determining borates in fertilisers in the range 0-1 to 1 per cent. of boron is described. Phosphate is removed by adding bismuth nitrate solution, and the “identical-pH” method is used to detect the end-point of the subsequent mannitol - boron titration. INTEREST in the manufacture of “compound” fertilisers to which borates have been added is increasing, and this has focused attention on the analytical problem of accurately deter- mining the borate present. The amount added is more than a “trace” and may generally be within the range 0.1 to 1 per cent.of boron. Current methods of determining borates are either photometric or volumetric. Most of the photometric methods are perhaps more suitable for “traces” than for larger amounts, and it is thought that a volumetric method will be preferable. These methods depend on the formation and titration of glyceroboric acid or, better, mannitoboric acid, and this necessitates prior separation of the borate from interfering substances, including phosphate. Distillation as methyl borate and hydrolysis of the distillate is well known, but it is time-consuming and difficult to carry out quantitatively. The American Association of Agricultural Chemists1 recommend two methods, one for acid- soluble and the other for water-soluble boron. In the first, phosphate is removed with lead nitrate, the excess of lead (and also calcium) being removed by sodium hydrogen carbonate.Finally, the boric acid is titrated by adjustment of the pH to 6.3, addition of mannitol and titration to the same pH value with sodium hydroxide solution (“identical-pH” method). The water-soluble boron is extracted by hot water, phosphate is removed with barium hydroxide solution, and ammonia is removed by boiling for at least 1 hour. After filtration and removal of carbon dioxide, the solution is made neutral to methyl red, mannitol is added, and the mannitoboric acid titrated to the phenolphthalein end-point.518 WILSON AND PELLEGRINI : DETERMINATION OF [Vol. 88 These methods can produce precipitates that are difficult to handle (even if “filter aids” are used), and in our hands the results were not as reproducible as one could wish.We therefore considered the possibility of using some other reagent to remove phosphate from the solution, Among the known precipitants for phosphate, bismuth nitrate in dilute acid solution seemed promising. As long ago as 1860, Chancelz determined phosphate by precipi- tation as bismuth phosphate and noticed that the crystalline precipitate settled and filtered well. Rathje3 had proposed to titrate phosphate with bismuth nitrate solution, locating the end-point by the orange colour of bismuth iodide, but pointed out that chlorides and sulphates interfere through formation of basic salts. It occurred to us that the use of bismuth would TABLE I BORATE FOUND I N COMMERCIAL FERTILISER Borate content found, as boron, by- Sample A No.Lnalyst A, analyst B, analyst E, % % % 0.171 0.173 0.177 0.169 0.174 0.172 0.176 0.172 0.177 0.160 0.157 0.159 0.158 0.161 0,162 l { also have the advantage that the excess could be readily removed as basic salts by increasing the pH and diluting the solution. A few experiments showed that a separation was possible, but also showed that an excess of bismuth was necessary over that equivalent to the phosphate present, presumably because of the formation of basic salts. EXPERIMENTAL Each of a series of 2.5-g portions of a commercial “compound” fertiliser containing 10 per cent. of P,05 (based on superphosphate) was dissolved in diluted nitric acid, and the resulting solutions were treated with different volumes of a 22 per cent.w/v solution of bismuth nitrate (see “Method”). The amount of phosphate remaining in solution was then determined spectrophotometrically ; the results were- Volume of bismuth nitrate solution used, ml. . 50 45 40 35 30 25 20 Proportion of original P,O, left in solution, yo 0.2 2.0 3.5 9.5 25.0 54-0 76.5 These results show that only with a volume of the bismuth nitrate solution equivalent to 5 ml for each 1 per cent. of P,05 in the fertiliser is removal of phosphate virtually complete. To obtain a crystalline precipitate that settles well, the bismuth nitrate solution must be added slowly to the hot acid solution of the fertiliser, with frequent agitation; after the addition is complete, the pH of the solution is about 1.7.It was found most convenient to allow the solution to cool, dilute it to a given volume and then filter an aliquot part rather than to wash the bismuth phosphate. That this does not lead to loss of borate is shownby the results on p. 519. Excess of bismuth in the aliquot is removed by addition of a small excess of sodium hydroxide. This precipitate (basic salts of bismuth) is washed, and the borate is then titrated in the filtrate by the “identical-pH” method after removal of carbon dioxide. For a fertiliser containing 0.15 to 0.2 per cent. of boron, the titre of 0.02 N sodium hydroxide is in the range 7 to 10 ml. There is a slight blank value; if possible, this should be determined by analysing a similar fertiliser to which no borate has been added.If this is not possible, an average blank value of 0.15 ml of 0-02 N sodium hydroxide can be used. I t is doubtful what significance would attach to water-soluble borate, and one or two preliminary experi- ments suggested that the method of extraction influenced the results. Although the “identical-pH” method has often been treated as empirical, our results justify the use of the stoicheiometric factor under our conditions. We have only concerned ourselves with total or acid-soluble borate. REPRODUCIBILITY OF RESULTS To two samples of fertiliser (A based on superphosphate and B based on ammoniumAugust, 19611 ADDED BORATES IN MIXED FERTILISERS 51 9 phosphate) known amounts of sodium tetraborate were added, and the boron contents were determined as described under “Procedure” ; the results were- Sample ... . A B && Boron added, yo . . 0.517 0.319 0.165 0.958 0.451 0.229 Boron found, yo . . 0-517 0.315 0.162 0-962 0.450 0.231 Two samples of a commercial borated fertiliser (based on superphosphate) were analysed by three analysts, the third of whom made only one determination on each sample and had no previous knowledge of the method; the results are shown in Table I. These results, together with those above, show satisfactory reproducibility. The time required for a determination in duplicate is about 3 hours. METHOD REAGENT- acid, with slight warming if necessary, and dilute to 100ml with water. PROCEDURE- Weigh out 2.5g of sample, transfer to a 400-ml beaker, add 2ml of nitric acid and 50 ml of water, stir, warm, and dilute to 100 ml with water.Warm the solution to 80” or 90” C (do not boil, as boric acid is volatile in steam), and slowly add from a burette, with continuous stirring, 5 ml of the bismuth nitrate solution for each 1 per cent. of P20, present in the fertiliser. Allow the precipitate to settle, cool, wash into a 250-ml calibrated flask, and dilute to the mark. Filter through a dry Whatman No. 40 filter-paper, rejecting the first few millilitres, and, by pipette, place 100 ml of the filtrate in a beaker. Add a few drops of bromothymol blue indicator solution and then a 10 per cent. solution of sodium hydroxide, with thorough stirring, until the indicator turns blue. Separate the precipitate on a Whatman No. 541 filter-paper, carefully wash it several times with cold water, and combine the washings with the filtrate (the total volume should be about 150 to 200 ml).Adjust the pH to about 5 by adding 5 per cent. nitric acid, heat to about 90” C (do not boil), and stir vigorously to remove carbon dioxide. Cool, place in the solution the electrodes of a suitable pH meter, and adjust the pH to 6.3; first use 10 per cent. sodium hydroxide solution, and finally bring the pH exactly to 6-3 with carbon dioxide-free 0.02 N sodium hydroxide. Add l o g of mannitol, and again bring the pH to 6.3 with the 0-02 N sodium hydroxide. Continue to add 10-g portions of mannitol and to re-adjust the pH to 6.3 until, after the final addition of mannitol, the pH remains constant at 6.3. (For samples containing up to 0-5 per cent. of boron, 20g of mannitol are usually sufficient.) The total amount of 0.02 N sodium hydroxide used after the additions of mannitol corresponds to the amount of boron present in the solution. Carry out a blank determination on a similar type of fertiliser to which borate has not been added, and subtract the blank titre from the titre previously obtained; if a “blank” sample is not available, deduct 0.15 ml from the titre of 0.02 N sodium hydroxide. Calculate the borate content, as boron, of the sample from the equation- Boron content, yo = (A - B) x 0.000216 x 100 in which A and B are the titres of 0.02 N sodium hydroxide in the sample and blank deter- minations, respectively. Bismuth nitrate solutim-Dissolve 22 g of Bi(N03)3.5H20 in S ml of concentrated nitric Keep the solution hot during precipitation. REFERENCES 1. 2. 3. Horwitz, W., Editor, “Official Methods of Analysis,” Eighth Edition, The Association of Official Chancel, G., Compt. Rend., 1860, 50, 416. Rathje, W., Angew. Chem., 1938, 257. Agricultural Chemists, Washington, D.C., 1955, p. 18. Received Februavy 24& 1961