254 BRADFIELD : AN IMPROVED FORMALDOXIME METHOD FOR THE Fol. 82 An Improved Formaldoxime Method for the Determination of Manganese in Plant Material BY E. G. BRADFIELD A method is described for the determination of manganese in plant material in which no preliminary separation of interfering elements is needed. Interference from iron and copper is eliminated by warming the solution, which results in decomposition of the formaldoxime complexes of these two metals, and precipitation of metal phosphates is prevented by using the seques- tering agent N-hydroxyethylethylenediaminetriacetic acid. The method is as rapid as the normal periodate procedure and its sensitivity is about 6 times as great. The coefficient of variation is 1.2 per cent. for high concentrations of manganese and 2.3 per cent.for lower concentrations. THE manganese content of plant material is usually determined colorimetrically by oxidation with potassium periodate to form permanganic acid.1 Compared with most colorimetric methods the sensitivity is poor and, when used for material of low manganese content, a sample weight of 5 to 10 g is required. This weight of sample may not be available if the amount of material is limited. Formaldoxime, which forms an orange-red colour with manganese in alkaline solutions (Mellan2), has been used for the determination of manganese in plant material by Sideris3pQ and in textiles by Hamlir~.~ Both authors note interference by phosphates, which are precipitated as alkaline-earth phosphates when the solution is made alkaline, and by iron, which forms an intense violet-red colour with formaldoxime.Sideris4 recommends the precipitation of iron as ferric phosphate and phosphates as lead phosphate in acetic acid solution, while Hamlin5 removes iron by extracting its cupferron complex with methylene dichloride and prevents precipitation of alkaline-earth phosphates by the addition of ammonium salts and sodium pyrophosphate. Both procedures involve prior separation of iron and this increases the time required for the determination. ELIMINATION OF INTERFERENCE BY IRON- of a complex with sodium cyanide. EXPERIMENTAL Sideris3 has previously shown that interference by iron can be avoided by the formation It is, however, necessary to compensate for the greenish-April, 19571 DETERMINATION OF MANGANESE I N PLANT MATERIAL 255 yellow colour of the cyanide complex by adding iron to the standards in amounts roughly equivalent to the content of the test solutions, a difficult procedure when the iron content of the test materials is variable. A search was therefore made for alternative agents that would selectively complex iron and prevent its reaction with formaldoxime.Five sequestering agents (i) ethylenediaminetetra-acetic acid (EDTA), (ii) N-hydroxy- ethylethylenediaminetriacetic acid (HEEDTA), (iii) NN'-dihydroxyethylethylenediamine- diacetic acid (HEEDDA), (iv) diethylenetriaminepenta-acetic acid (DTPA) and (v) 1 : 2- diaminocyclohexanetetra-acetic acid (CDTA) were tried. Of these, only CDTA prevented iron from forming a coloured complex with formaldoxime; unfortunately, it also prevented the formation of the manganese complex.During these experiments, it was noticed that the iron - formaldoxime complex decomposed on prolonged standing, ferric hydroxide being precipitated from solutions containing EDTA and DTPA, but not from solutions containing HEEDTA and HEEDDA. It was found that the rate of decomposition was increased by warming; the reaction was therefore investigated quantitatively. Portions of solutions of manganese and iron containing 40 pg and 300 pg, respectively, were placed in six 50-ml calibrated flasks and diluted to about 30 ml with water. To each, 5 ml of a 10 per cent. w/v solution of HEEDTA, 1 ml of diluted formaldoxime reagent (prepared from hydroxylamine sulphate and paraformaldehyde) and 2 ml of a 10 per cent, w/v solution of sodium hydroxide were added.The flasks were placed in a water bath at 65" C for 0, 10, 20, 30, 60, 120 and 240 minutes and then removed and cooled; the contents of each were diluted to the mark and the absorptions of the solutions were measured. The experiment was repeated with, in one series, 40 pg of manganese and no iron; in another 300 pg of iron and no manganese; and in a third no manganese or iron. The results are C F;I I a Fig. 1. Effect of heating iron and man- ganese with formaldoxime reaction mixture at 65" C: curve A, 40 pg of manganese + 300 pg of iron; curve B, 40pg of manganese; curve C, 300pg of iron; curve D, blank shown in Fig. 1. It will be seen that, although the iron - formaldoxime complex is com- pletely decomposed after 1 hour, the colour due to the manganese complex, after an initial256 BRADFIELD: AN IMPROVED FORMALDOXIME METHOD FOR THE [Vol.82 fall in intensity, is stable over the period-of 1 to 3 hours. This initial fall is probably due to contamination of the reagents by iron, since the blank has a high initial absorption, which subsequently decreases on warming. The initial colour of the blank was pink, similar to that of the iron - formaldoxime complex. It is therefore possible to overcome interference by iron by warming the reaction mixture for 2 hours at 65" C before measuring the absorption. EFFECT OF HEAT ON INTERFERING IONS- interfering metal ions was investigated, the results being as follows- In view of these results, the effect of warming the formaldoxime complexes of other Co$+er-Copper reacts with formaldoxime to give a violet coloured solution ; interference can be overcome by addition of ~yanide.~ When a solution containing 100 pg of copper was warmed, the formaldoxime complex was rapidly decomposed and the solution became colourless.Nickel-The golden brown solution produced by 100 pg of nickel with formaldoxime decomposed slowly on warming. After 2 hours at 65" C the absorption was equivalent to that of 2 pg of manganese. Cobalt-The pale yellow-green colour of 100 pg of cobalt with formaldoxime slowly faded on warming: after 2 hours the absorption was equivalent to that of 2 p g of manganese. Other elements-There was no colour with 100-pg amounts of zinc, cadmium, chrom- ium, lead, molybdenum or titanium or 500 pg of aluminium, and consequently there was no interference in the determination of manganese.EFFECT OF SEQUESTERING AGENTS ON THE PRECIPITATION OF ALKALINE-EARTH PHOSPHATES- In plant material, calcium and magnesium are the predominant elements that form insoluble phosphates in alkaline solution and precipitation of these can be prevented by adding sufficient of a suitable chelating agent to sequester the metal ions. Experiments showed HEEDTA to be the most effective agent; 5 ml of a 10 per cent. w/v solution were adequate for the types of plant material analysed. METHOD REAGENTS- in water and dilute to 100 ml. sulphate in boiling water and dilute to 100 ml. in water and dilute to 100 ml. HEEDTA solution-Dissolve 10 g of trisodium N-hydroxyethylethylenediaminetriacetate Formaldoxime reagent-Dissolve 20 g of paraformaldehyde and 55 g of hydroxylamine Sodium hydroxide solution-Dissolve 10 g of analytical-reagent grade sodium hydroxide For use, dilute this solution 10 times.PROCEDURE- Weigh 1 g of plant material, previously ground in a mortar and dried at 105" C, into a 250-ml tall Pyrex-glass beaker. Add 25 ml of concentrated nitric acid, cover with a watch- glass, and digest at low heat until no solid material remains. Then add 2.5 ml of 60 per cent. perchloric acid and continue digestion until fumes of perchloric acid are given off. If at this stage the digest is still coloured, add a further 5 ml of nitric acid and again heat until fumes of perchloric acid are given off. Uncover the beaker slightly and continue heating until nearly all the perchloric acid has been removed.Cool, add 25 ml of water, boil and filter the solution through a Whatman No. 540 filter-paper into a 100-ml Pyrex-glass calibrated flask. By pipette place an aliquot, containing 10 to 50 pg of manganese, in a 50-ml Pyrex- glass calibrated flask, dilute to about 30 ml and add 5 ml of HEEDTA solution. Add sodium hydroxide solution dropwise until any free perchloric acid is neutralised, using a wide-range pH paper as external indicator. Then add 1 ml of diluted formaldoxime reagent, followed im- mediately by 2 ml of sodium hydroxide solution. It is necessary to add the sodium hydroxide solution within 2 minutes of adding the reagent, since formaldoxime decomposes rapidly in dilute solutions.Remove the stopper from the flask and place it in a water bath at 65" C for 2 hours. At the end of this time remove the flask from the bath, cool the contents and dilute to 50 ml. When cool, dilute to volume.April, 19571 DETERMINATION OF MANGANESE I N PLANT MATERIAL 257 Measure the absorption of the solution at a wavelength of 450 mp on a Unicam SP600 spectrophotometer, using 4-cm cells, and calculate the manganese content from a calibration graph, prepared by plotting absorption readings against the manganese contents of solutions containing 0 to 50 pg of manganese, added as a standard solution of manganese sulphate. Develop the colour of the standards in exactly the same manner as that of the samples. Determine the blank by carrying out this procedure on the reagents only.RESULTS The recovery of manganese was tested by the addition of 20 pg and 40 pg of manganese to 10-ml aliquots of a solution containing potassium, calcium, phosphate, magnesium, iron, copper, boron and zinc in such amounts that these aliquots were roughly equivalent in mineral content to an acid digest of 1.0 g of dried plant material. The results of triplicate determina- tions gave a mean recovery of 20.6 pg of manganese from a recovery range of 20.4 to 20.7 pg when 20 pg had been added to the aliquot. For the addition of 40 pg of manganese the mean recovery was 40.2 pg and the recovery range 39.8 to 40.6 pg. The precision of the method was determined by carrying out a series of 10 determinations of manganese on a sample of apple leaves that had a low manganese content and a similar series on a sample of blackcurrant leaves that had a fairly high manganese content.For the apple leaves a mean content of 20.7 p.p.m. of manganese was found with a range of 20.2 to 21.6 p.p.m., the standard deviation being 0-5 and the coefficient of variation 2.3 per cent. For blackcurrant leaves the mean content was 132.1 p.p.m., with a range of 130.4 to 135.2 p.p.m., the standard deviation being 1.6 and the coefficient of variation 1.2 per cent. I t will be seen that the coefficient of variation is low. The manganese contents of a number of plant materials were then determined by the method described and compared with the values obtained by the periodate oxidation method given by Piper.l It will be seen from Table I, in which each result is the mean of three determinations, that there is good agreement between the two methods.For the deter- mination of the small amounts of manganese present in apple and cauliflower leaves, only 1 g of plant material was required for the formaldoxime method whereas, for comparable accuracy, 5 g were required for the periodate method. TABLE I MANGANESE CONTENT OF PLANT MATERIALS BY FORMALDOXIME AND PERIODATE METHODS Material Blackcurrant leaves Blackcurrant leaves Apple leaves . . Cauliflower leaves . . Kale leaves .. Wheat straw . . Pasture herbage . . Broccoli leaves . . Strawberry leaves , . Cacao leaves . . .. . . .. . . . . . . . . . . . . . . . . . . .. . . .. . . . . * . .. . . Manganese found by Manganese found by formaldoxime method, periodate method, p.p.m. p.p.m. 129 104 21 19 39 49 106 56 325 564 130 107 20 18 38 51 109 55 318 582 REFERENCES 1. 2. 3. 4. 5. Piper, C. S., “Soil and Plant Analysis,” The University of Adelaide, Adelaide, 1942, p. 346. Mellan, I., “Organic Reagents in Inorganic Analysis,” The Blakiston Company, Philadelphia, 1941, pp. 106 and 450. Sideris, C. P., Ind. Eng. Chem., Anal. Ed., 1937, 9, 445. __ , Ibid., 1940, 12, 307. Hamlin, A. G., J. Text. Inst., 1956, 47, T 445. RESEARCH STATION LONG ASHTON BRISTOL November 22nd, 1956