Sept., 19511 CLULEY 51 7 The Determination of Germanium The following three papers were presented at the Meeting of the Society on Wednesday, May 2nd, 1951. Part I. Titration of Mannito-Germanic Acid* BY H. J. CLULEY In aqueous solution germanium dioxide reacts with mannitol to form a strong, complex acid ; germanium can be determined volumetrically by titration with sodium hydroxide solution of the mannito-germanic acid so formed. A method has been evolved in which the volumetric procedure is applied after a preliminary separation of the germanium by precipitation. as sulphide. By a simple modification the interference of arsenic is readily obviated. GERMANIUM, long a neglected element, has in recent years achieved prominence owing to its extensive application as a semi-conductor in crystal rectifiers and similar devices.To meet the consequent demand germanium is now being produced in this country by extraction from flue dusts occurring in the producer systems of gasworks using certain coals in which germanium is a trace c0nstituent.l The analytical difficulties encountered during work on the extraction of germanium from flue dusts encouraged a search for new methods of deter- mination of the element and the results of this analytical investigation are reported in this and the following two papers (pp. 523 and 530). The majority of published procedures for the determination of germanium suffer from lack of selectivity and, in particular, from interference from arsenic, with which germanium is commonly associated. This is a serious objection, as a complete separation of germanium from arsenic is not readily achieved. The procedure most widely used for the determination of germanium is the tannin method, with which a separation from all elements other than tantalum, niobium and tungsten is stated to be theoretically possible.2 Initially, the tannin method was investigated, the conditions of precipitation recom- mended by Davies and Morgan2 being used.I t was found that at low concentrations the germanium - tannin precipitate tended to become colloidal, with consequent low results, and it was concluded from this investigation that the method was unlikely to be wholly satisfactory for the small amount of germanium it was required to determine. The method of precipitation employed by Davies and Morgan has subsequently been criticised by HolnessJ3 who advocated precipitation from an oxalic acid solution in preference to sulphuric acid solution.A further disadvantage of the tannin method is the tedious preliminary ignition at 600" C, with repeated nitric acid oxidation, that is necessary to ensure the absence of the volatile germanous oxide before the final ignition at 900" C. This paper describes a volumetric method applied to the determination of germanium ; the development of an absorptiometric method and its subsequent application to the determination of germanium in flue dust, coal and coke are described in the following two papers. Attention was then directed to volumetric and absorptiometric methods. EXISTING VOLUMETRIC METHODS There appear to be only two published procedures for the volumetric determination of germanium.Willard and Zeuhkle4 proposed a method based on the formation of a thio- germanate in an acetate-buffered solution. The thiogermanate was oxidised with standard iodine solution the excess of which was determined by titration with thiosulphate. The authors state that the method is of limited application. The other method is due to TchakirianJ5 who found that germanium dioxide in aqueous solution reacted with mannitol (and with other polyhydric alcohols) to form a strong complex acid. The mannito-germanic acid so formed could be titrated with sodium hydroxide solution * Taken from a Thesis submitted t o the University of London for the degree of M.Sc.518 CLULEY: THE DETERMINATION OF GERMANIUM [Vol.76 to the phenolphthalein end-point, one molecule of sodium hydroxide being equivalent to one atom of germanium. The mannito-germanic acid also reacted with an iodate - iodide mixture and the liberated iodine could be titrated with thiosulph’ate. Tchakirian stated that the alkalimetric procedure could not be applied to germanium solutions containing strong acids, and there is no evidence from his oraginal5 or subsequent6y7 papers that either of his methods had been used except for pure germanium solutions. The alkalimetric titration of boric acid in the presence of mannitol is a procedure universally used, and the corresponding method for the determination of germanium appeared worthy of investigation. In particular, the method for boron is applicable to solutions containing strong acids and this cast doubt on Tchakirian’s statement that the method for germanium was not applicable under these conditions.An investigation was therefore made in an attempt to verify Tchakirian’s observations and to extend the usefulness of his met hod. THE TITRATION OF MANNITO-GERMANIC ACID Preliminary experiments were performed on the formation of mannito-germanic acid, and pH curves were drawn for the neutralisation of the acid with sodium hydroxide. These curves gave inflection points at about pH 7.8 and confirmed that 1 molecule of sodium hydroxide was equivalent to 1 atom of germanium. Three different titration procedures were then examined, germanium solutions containing free sulphuric acid being used. (a) Calcium carbonate neutralisation method-The solution was neutralised with calcium carbonate and, after filtration, mannitol was added and the solution was titrated with sodium hydroxide solution to the phenolphthalein end-point.This procedure was based on the Wherry method for determination of boron.* (b) Double indicator method-The solution was adjusted with the sodium hydroxide solution to the P-nitrophenol end-point ; after addition of mannitol the solution was titrated to the phenolphthalein end-point. This type of procedure is commonly used for titration of boric acid.g (c) Fixed pH method-This was a modification of the foregoing method (6) in which the pH value of the solution before addition of the mannitol and the pH value at the end of the titration are the same.This procedure has also been used for the titration of boric acid. lo y1l It was found that all three methods were suitable for the determination of 1 to 10-mg quantities of germanium, approximately 0.02 N sodium hydroxide being used. Hence it was established that an alkalimetric procedure could be applied to the determination of germanium in solutions containing a strong acid. Methods (b) and (c) are more rapid as they avoid the necessity for filtration. The following theoretical comparison of methods (b} and (c) shows that the latter is likely to be less influenced by other substances present in solution. In the double indicator method (b) the pH value at the P-nitrophenol end-point is about 6.0, falling to about 4.0 on addition of mannitol and increasing to about 8.4 at the phenol- phthalein end-point.In this type of titration, therefore, the formation and subsequent neutralisation of the mannito-germanic acid are accompanied by a net change of pH value from 6.0 to 8.4. If substances exerting any buffering action are present the amount of sodium hydroxide required to effect this net pH change will be increased, with consequent error in the determination. In method (c) the neutralisation of the mannito-germanic acid is carried only to the stage where the pH value is the same as that before addition of mannitol, the net pH change is zero and the error due to any bufEer effect vanishes. In practice, buffering agents can in fact be tolerated provided that they ,are not present in sufficient concentration to render the titration end-point indefinite.In this type of titration the neutralisation of the mannito-germanic acid is obviously incomplete and an empirical relationship exists between the germanium and the sodium hydroxide solution. This relationship must be established by titration of known amounts of germanium under closely standardised conditions. The fixed pH method, (c), was therefore preferred as it appeared less susceptible to errors caused by the ionic environment of the germanium. A further advantage is the reduction in the size of the blank, a matter of some importance for the small titrations involved. The fixed pH value selected was 6.2, and the standardisation titrations were carried out in the following manner.Sept., 19511 PART I. TITRATION OF MANNITO-GERMANIC ACID 519 To the weakly acid germanium solutions, of volume about 80 ml, 7 drops of bromocresol purple indicator were added.Carbonate-free sodium hydroxide solution, 0.01 85 N , was then added until a pH of 6.2 was reached, as indicated by colour comparison with pH 6.2 buffer solution containing indicator. Then l o g of mannitol were added and the solutions were titrated with the sodium hydroxide solution until a pH of 6.2 was again attained. The usual correction for the blank was applied. TABLE I FIXED pH TITRATION METHOD : STANDARDISATION TITRATIONS Corrected titre of Germanium taken, sodium hydroxide, mg ml 0.0185 N 1 0.67, 0.70 3 2-19, 2-20 6 4-37, 4-42 10 7-35, 7.37 20 14.57, 14-59 . Weight of germanium equivalent to 1 ml of sodium hydroxide, mg 1.461 1.367 1.366 1-359 1-'371 NOTE-For complete neutralisation of mannito-germanic acid the theoretical relationship is 1 ml of 0.0185 N sodium hydroxide = 1.343 mg of germanium.From these titrations the germanium equivalence of the sodium hydroxide solution was calculated for each weight of germanium titrated, as shown in Table I. Theoretically the germanium equivalence of the titrant will vary slightly with the weight of germanium titrated, but it is clear that for the small amounts of germanium under consideration the use of a mean equivalence factor will occasion little error. It will be seen that the equivalence factor thus established empirically for this method of titration differs only slightly from the theoretical relationship obtaining for complete neutralisation of the mannito-germanic acid, so that the loss of sensitivity is very small.To illustrate this type of titration a pH curve was prepared of the titration of 10mg of germanium under the above conditions, and this is shown in Fig. 1. The dotted portion Fig. 1. Graph of a fixed pH titration of the first curve shows how the titration continues if mannitol is not added; the dotted portion of the second curve shows the completion of the neutralisation of the mannito- germanic acid after pH 6.2.520 CLULEY: THE DETERMINATION OF GERMANIUM [Vol. 76 TITRATION OF GERMANIUM AFTER: SEPARATION AS SULPHIDE Substances that would be expected to interfere with the titration of germanium fall (a) Buffering agents, if they are present in a sufficiently high concentration to render the titration end-points indefinite.(b) Bases precipitated by sodium hydroxide within the pH range of the titration. (This is a special case of buffering.) (c) Compounds which react similarly with mannitol, e.g., boric acid. For the general application of the volumetric method a preliminary separation of the germanium will normally be necessary. Precipitation as sulphide serves to separate germanium from boron and from the majority of bases of type (b). Experiments were therefore carried out on the application of the volumetric method after separation of germanium in this manner. Known amounts of germanium were precipitated as sulphide and the filtered precipitates were dissolved in ammonium hydroxide solution and oxidised with hydrogen peroxide.After boiling with sodium hydroxide solution to eliminate ammonia and hydrogen peroxide the solutions were just acidified with sulphuric acid and the fixed pH titration method applied. At first difficulties were encountered owing to the effect on the indicator of oxidising substances still present in solution. This trouble was overcome by boiling with an excess of sulphuric acid after the elimination of ammonia and hydrogen peroxide, then making the solution just acid and applying the volumetric method. The results obtained in this manner are shown in Table 11. into three classes- TABLE ICI TITRATION OF GERMANIUM AFTER PRECIPITATION AS SULPHIDE Germanium taken, Germanium found, mg mg 1 1.00, 0.98 3 2-98, 3-04 6 6.03, 6.01 10 9.96, 9.90, 9.96, 9.85 These results show that the volumetric procedure can successfully be applied after separation of germanium as sulphide from pure solutions.However, in practice, other elements of the analytical sub-group IIB will accompany germanium in this separation. Of these elements arsenic is commonly associated. with germanium and interferes with most methods for its determination. For these reasons the effect of arsenic on the volumetric method was examined in some detail. TITRATION OF GERMANIUM IN THE PRESENCE OF ARSENIC Titration by the fixed pH method of known amounts of germanium in the presence of arsenic established that as much as 100 mg of tervalent arsenic could be tolerated without detriment to the determination of 10 mg or less of germanium. Similar amounts of quin- quevalent arsenic were found to interfere owing to pronounced buffering of the solution.If the sulphide separation procedure previously examined were used, the co-precipitated arsenic would be oxidised by the hydrogen peroxide to the quinquevalent state and would therefore interfere with the final titration. The nm-interference of tervalent arsenic suggested that this difficulty could be resolved by reduction of the arsenic to the tervalent state before titration. For this purpose reduction with sulphur dioxide was used because the excess of reducing agent could readily be removed by boiling. This reduction was applied to arsenate solutions containing germanium and in the subsequent titrations no interference from buffering occurred. Finally, this modification was applied to the volumetric determination of germanium after precipitation as sulphide from solutions containing arsenic. The results of these experiments are shown in Table 111.Sept., 19511 PART I.TITRATION OF MANNITO-GERMANIC ACID 521 TABLE I11 DETERMINATION OF GERMANIUM IN THE PRESENCE OF 100mg OF ARSENIC Quinquevalent arsenic reduced to the tervalent state before titration Germanium added to 100 mg of arsenic, mg . . .. 1 3 5 5 6 Germanium found, mg . . .. .. .. . . 0.98 2.93 4.98 4.94 5.92 The results in Table I11 show that small amounts of germanium can be determined in the presence of substantial amounts of arsenic. The effect of other elements has not been extensively investigated. Traces of antimony and tin can be tolerated, but large amounts interfere owing to the formation of insoluble hydroxides or basic salts.RECOMMENDED METHOD FOR THE VOLUMETRIC DETERMINATION OF GERMANIUM APPARATUS- It is necessary to use beakers or flasks made from boron-free glass or platinum apparatus where strongly alkaline solutions are employed, as any boron dissolved from borosilicate glassware will be titrated with the germanium. SPECIAL REAGENTS- Sodium germanate solution-Transfer to a platinum crucible 1.4408 g of pure ignited germanium dioxide, fuse with 5 g of sodium carbonate and dissolve the cold melt in hot water; just acidify the solution with dilute sulphuric acid, boil to eliminate carbon dioxide, cool and dilute to 1000 ml. 1 ml of solution G 1 mg of germanium. Bufler solution, p H 6.2-Mix 33.9 ml of 0.1 M citric acid with 66.1 ml of 0.2 M di-sodium hydrogen phosphate. STANDARDISATION OF THE SODIUM HYDROXIDE SOLUTION- Transfer volumes of standard sodium germanate solution, covering the range of 1 to 20 mg of germanium, to 250-ml flasks.Dilute each volume to about 50 ml, boil for 5 minutes to eliminate any carbon dioxide and then cool. Add 7 drops of bromocresol purple indicator and add from a burette carbonate-free sodium hydroxide solution, approximately 0.02 N , until a pH value of 6-2 is reached as indicated by colour comparison with an equal volume of pH 6.2 buffer solution containing 7 drops of indicator. Add 10 g of mannitol and titrate with the sodium hydroxide solution until a pH value of 6.2 is again reached. Carry out a blank determination and correct the titration figures accordingly.From the corrected titration figures calculate the equivalence of the sodium hydroxide solution, in terms of mg of germanium per ml, for each weight of germanium titrated. It will be observed that the germanium equivalence of the sodium hydroxide solution differs slightly for each weight of germanium titrated, but for the small weights of germanium used it may be permissible to use a mean equivalence factor. PROCEDURE- The sample solution, of volume about 100 ml and containing 1 to 20 mg of germanium, should not contain large amounts of antimony or tin, and sulphuric acid should be the only acid present. Transfer the sample solution to a conical flask and add sufficient diluted sulphuric acid (1 + 1) to give a sulphuric acid concentration of about 5-5 N .Pass a rapid stream of hydrogen sulphide through the solution for 30 minutes, cork the flask and allow it to stand overnight. Filter the solution through a close filter-paper and wash the precipitate well with 5.5 N sulphuric acid saturated with hydrogen sulphide. Dissolve the precipitate through the paper into a boron-free flask (or platinum dish) with three successive 8-ml portions of diluted ammonium hydroxide solution (2 + 1) and wash the paper first with dilute ammonium522 CLULEY: THE DETERMINATION OF GERMANIUM [Vol. 76 hydroxide solution (1 + 9) and finally with hot water. Add 20ml of 6 per cent. w/v hydrogen peroxide and allow to stand for 10 minutes in the cold to ensure complete oxidation of the sulphides. Add 5 ml of freshly prepared 5 N sodium hydroxide and boil the solution vigorously until all ammonia and hydrogen peroxide has been evolved, adding hot water from time to time to maintain the volume of the solution.Add 8ml of dilute sulphuric acid (1 + 6 ) and boil for 10 minutes to remove any oxides of nitrogen formed from oxidation of the ammonia. If desired, the solution may be transferred to ordinary glassware at this stage. Pass a rapid stream of sulphur dioxide into the hot solution until it is cold; this takes about 30 minutes. Boil off the excess of sulphur dioxide, ensuring complete removal by continuing the boiling for 5 minutes after the smell of sulphur dioxide can no longer be detected. The treatment with sulphur dioxide may be omitted if arsenic is absent. Add 2 drops of bromocresol purple indicator and then add freshly prepared 5 N sodium hydroxide until the solution is just alkaline.Add N sulphuric acid until the indicator just turns yellow, dilute to about 80m1, boil for 5 minutes to eliminate any carbon dioxide and cool. Add 5 more drops of bromocresol purple indicator, adjust to pH 6.2, add 10 g of mannitol and titrate to pH 6-2 as in the standardisation titrations. Perform a blank determination, correct the sample titration figure and deduce the germanium content of the sample solution. NOTES- The presence of germanium in reagents is unlikely, but it is necessary to carry out blank determinations because of the possibility of deriving traces of boron from reagents or glassware. The titrations may be carried out with the aid of a pH meter if required, but it is advantageous to retain the use of the indicator so that readings need only be taken when the indicator colour shows proximity to pH 6-2.The method is equally satisfactory for large amounts of germanium, for which a stronger solution of sodium hydroxide should be used. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES Chirnside, R. C., Times Review of Industry, July, 1950. Davies, G. R., and Morgan, G. T., Analyst, 1938, 63, 388. Holness, H., Anal. Chim. Acta, 1948, 2, 254. Willard, H. H., and Zeuhlke, C. W., Ind. Eng. Chem., Anal. Ed., 1944, 16, 322. Tchakirian, A., Compt. Rend., 1928, 187, 229. -, Bull. SOC. Chim. France, 1932, [4] 51, 846. Wherry, E. T., J . Amer. Chem. SOC., 1908, 30, 1687. Hillebrand, W. F., and Lundell, G. E. F., “Applied Inorganic Analysis,” John Wiley and Sons Foote, F.J., Ind. Eng. Chem., Anal. Ed., 1932, 4, 39. Hollander, M., and Rieman, W., Ibid., 1946, 18, 788. -, Ibid., 1943, [5] 10, 98. Inc., New York, p. 611. RESEARCH LABORATORIES THE GENERAL ELECTRIC COMPANY LIMITED WEMBLEY, MIDDLESEX and SIR JOHN CASS COLLEGE JEWRY STREET LONDON, E.C.3Sept., 19511 PART 11. ABSORPTIOMETRIC DETERMINATIOX WITH PHENYLFLUORONE 523 Part 11. Absorptiometric Determination with Phenylfluorone* BY H. J. CLULEY An absorptiometric method is described for the determination of germanium by means of 2 : 3 : 7-tnhydroxy-9-pheny1-6-fluorone (phenyl- fluorone). The method is about four times as sensitive as the molybdenum blue method. By a simple distillation from hydrochloric acid solution the germanium is readily separated from the few elements that have been found to interfere with the phenylfluorone method.ABSORPTIOMETRIC methods for the determination of germanium have been investigated to only a slight extent. The only methods at present available are those based on the formation of the yellow germano-molybdic acid1 or on its subsequent reduction to molybdenum blue.2t3 These procedures have the disadvantage that similar reactions are given by silicon, phosphorus and arsenic, and a complete separation of germanium from arsenic in particular is not readily achieved. A search of the literature was carried out in an endeavour to find a more selective colour reaction for germanium that might form the basis of an absorptiometric method. Gillis, Hoste and Claeys? who have examined a number of derivatives of fluorone for their potential value as analytical reagents, have stated that 2 :3 :7-trihydroxy-9-phenyl- 6-fluorone (phenylfluorone, Fig.1) is a specific reagent for the detection of germanium. C H I d Fig. 1. Phenylfluorone Their spot test is carried out on a test paper pre-treated with an alcoholic solution of the reagent acidified with hydrochloric acid; a drop of a test solution containing quadrivalent germanium gives a pink colouration that does not disappear on treatment with 6 N nitric acid. The reaction of germanium with phenylfluorone has now been studied as the basis of an absorptiometric method. PRELIMINARY EXPERIMENTS Phenylfluorone was prepared from tri-acetylhydroxy-hydroquinone and benzaldehyde (see Procedure) by the method employed by Gillis, Hoste and Claey~.~ The reagent is only slightly soluble in alcohol and similar solvents, but its solubility in alcohol is increased by adding a small volume of dilute hydrochloric or sulphuric acid, which appears to convert phenylfluorone to the corresponding salt.For preliminary experiments a solution of 0-05 g of phenylfluorone in a mixture of 95 ml of alcohol and 5 ml of dilute sulphuric acid (1 + 6) was used, and this is subsequently referred to as a 0.05 per cent. solution. This reagent solution, yellow in colour, gave a colour reaction with quadrivalent ger- manium in dilute hydrochloric or sulphuric acid solution. The colour so formed was orange, presumably due to a combination of the yellow colour of the reagent and the pink colour of the germanium complex.On standing, the germanium compound tended to precipitate, but the use of gum arabic as a protective colloid was effective in stabilising the colour, and this was used throughout the subsequent experiments. Owing to the strong colour of phenylfluorone itself in acid solution, it was found necessary for absorptiometric measurements to use as a reference solution a blank containing the reagent but no germanium. Under these conditions of measurement the maximum absorption of * Taken in part from a Thesis submitted to the University of London for the degree of MSc.524 CLULEY: THE DETERMINATLOK OF GERMAKIUM [Vol. 76 the germanium colour was in the blue - green regAon. The initial absorptiometric measure- ments showed that the intensity of the germanium colour tended to increase with time.In their spot test, Gillis, Hoste and Claeys4 had found it necessary to treat the colour spot with 6 N nitric acid to destroy the colours produced by certain interfering elements. Similar treatment of germanium colours developed in dilute sulphuric acid solution was found to result in destruction of the reagent. 'This was possibly due to the presence of nitrous acid, but the use of nitric acid was not pursued. COLOUR DEVELOPMENT IN SULPHURIC ACID SOLUTION Systematic experiments were carried out to (establish conditions suitable for the rapid development of the germanium - phenylfluorone colour. For this work a sulphuric acid medium was used, as it was considered that in the preparation of a sample solution the use of hydrochloric acid should normally be avoided owing to the volatility of germanium tetrachloride. The concentration of sulphuric acid, reagenl and germanium were varied, but in all experiments 5 ml of 0.5 per cent.w/v gum arabic solution were added and the final volume was 50 ml. The absorptiometric measurements were carried out on a Spekker absorptiometer with 1-cm cells, the tungsten lamp and Ilford No. 603 blue - green filters, and as a reference solution a blank was used, prepared in the same manner as the test solution except that no germanium was added. The rates of colour development were found to increase with increasing reagent con- centrations and to decrease with increasing sulphuric acid concentrations. At low concen- trations of acid precipitation of phenylfluorone occurred.The use of 15 ml of a 0.02 per cent. reagent solution, and of 10 ml of dilute sulphuric acid (1 + 6), giving an acid concentration of about 1.05N in the final volume of 50m1, proved satisfactory and ensured complete colour development within 30 minutes. The relationship between absorption and germanium concentration with these conditions of colour development is shown in Table I. There is a slight deviation from Beer's law that becomes more marked with increasing germanium concentration. The sensitivity of the reaction is ,apparent. TABLE 1 EFFECT OF GERMANIUM CONCENTRATION ON COLOUR DEVELOPMENT IN Sulphuric acid concentration 1.05 N; 15 ml of 0.02 per cent. reagent solution used Germanium, pg .. . . * . . . 10 20 30 40 50 Drum reading, measured after 30 minutes 0.200 0.405 0.608 0.800 0.945 With the conditions outlined above, absorptiometric determinations of 25-pg quantities of germanium were carried out in the presence of a large number of individual elements to assess their effect on the method. Initially the weight of each element added was 25 mg (1000 times the weight of germanium) except for calcium and boron, where low solubility necessitated the use of smaller weights, and for sodium, potassium and ammonium, where 1-g quantities were used. This work showed that the following ions did not interfere: NH,', Na', K', Li', Cu", Ag', Be", Mg", Ca", Zn", Cd", Hg", Al"', Cr"', Mn", Fe", Co", Ni", B03"', PO4"' and Cl'. The following ions were found to cause interference, the numbers in brackets giving the maximum permissible concentration of the element expressed as a multiple of the germanium concentration: Ga"' (< l), Ti"" (2), Sn" (< l ) , Sn"" (< l), As"' (50), AsO,"' (loo), Sb"' (< I), Bi"' (lo), MOO," (< 1) and Fe"' (10).It will be observed that the interference of arsenic was very slight, but that that of gallium, tin, antimony and molybdenum was most marked. Strong oxidising agents such as dichromate and permanganate were also found to interfere by destroying the reagent. Under the conditions used the absorptiometric procedure was, therefore, selective but not specific for germanium. Modification of the conditions of colour development showed little promise of eradicating the interferences observed and it was clear that a preliminary separation of the germanium would be necessary for the general application of the method.The interference of arsenic, antimony, bismuth, molybdenum and tin precluded thc SULPHURIC ACID SOLUTIONSept., 19511 PART 11. ABSORPTIOMETRIC DETERMINATIOK WITH PHENYLFLUORONE 525 use of the sulphide method for separation of the germanium, but the alternative procedure of distillation of the tetrachloride appeared to be promising. Arsenic is the element that normally causes most difficulty in this separation, and the use of fractional distillation in a current of chlorine has been proposed as a method of avoiding the partial co-distillation of a r ~ e n i c . ~ However, the relatively low sensitivity of the reagent to arsenic suggested that a simple distillation without any such precautions might effect an adequate separation.The distillation is normally carried out from solutions containing at least half their volume of concentrated hydrochloric acid and similar concentrations of acid are present in the distillate. Before examining this method of separation it was therefore necessary to establish conditions suitable for the application of the absorptiometric procedure to hydro- chloric acid solutions. COLOUR DEVELOPMENT IX HYDROCHLORIC ACID SOLUTION The work on colour development in hydrochloric acid solution was carried out by the same general procedure as was used for the study of the reaction in sulphuric acid medium. The results led directly to the conditions finally adopted for the absorptiometric determination of germanium, and it is considered useful to record these results in some detail.EFFECT OF PHENYLFLUORONE CONCEKTRATION- With 25-pg quantities of germanium and a final hydrochloric acid concentration of 1.0 K , the effect of reagent concentration was studied for the range 2 to 10 ml of the 0.05 per cent. solution. In each experiment sufficient alcohol was added to bring the total volume of alcohol present to 15 ml. The results recorded in Table I1 again show the marked dependence of rate of colour development on reagent concentration and it will be seen that, for the acid concentration used, only the 8 and 10-ml quantities of the 0.05 per cent. phenylfluorone solution could effect complete development of the colour within 30 minutes.For subsequent work the intermediate quantity, 9 ml, was used, but to avoid separate addition of alcohol this was added as 15ml of a 0.03 per cent. solution. Owing to the significant absorption of light by the reagent solution alone, it is a disadvantage to use quantities of reagent greatly in excess of the amount required to give an adequate rate of colour development. TABLE I1 EFFECT OF PHENYLFLUORONE CONCENTRATION ON COLOUR DEVELOPMENT IN 25 pg of germanium used; final acid concentration 1.0 N HYDROCHLORIC ACID SOLUTION Drum readings, measured after 0.05% reagent, Alcohol, I -l ml ml 15 min. 30 min. 60 min. 90 min. 2 13 0.039 0.060 0.070 0.106 4 11 0.348 0.410 0,445 0.453 6 9 0.478 0.480 0.488 0.498 7 8 0.462 0.469 0.479 0.483 8 7 0.472 0.478 0.476 0,482 10 6 0.484 0.483 0.484 0.486 A EFFECT OF HYDROCHLORIC ACID COXCENTRATION- With 25-pg quantities of germanium and 15 ml of 0.03 per cent.phenylfluorone solution the effect of hydrochloric acid concentration was studied for the range 0.25 to 5.0 N. The results, recorded in Table 111, again show that the rate of colour development decreases with increasing acid concentration. For the quantity of reagent used, the acid range of 0.25 to 1.5 N is effective in producing rapid colour development and in giving almost identical drum readings for the fully developed colours. However, this does not mean that, within this range, the acid concentration need not be closely controlled. It was found that the absorption of the reagent solution itself vaned somewhat with the acid concentration and in the acid range 0.25 to 1.5 N consistent drum readings are obtained only if the acid con- centration in the blank or reference solution is essentially the same as that in the germanium solution,526 CLULEY: THE DETERMINATION OF GERMANIUM [Vol.76 For application of the absorptiometric method to highly acid distillates it was desirable to use the highest acid concentration consistent with the requirements of rapid colour Germanium, yg Fig. 2. Calibration graph for Determination of Gmnanium in Hydrochloric Acid Solution development. between 1.0 and 1.5 N . For this purpose it was decided to use an acid concentration intermediate TABLE I11 EFFECT OF HYDROCHLORIC ACID CONCENTRATION ON COLOUR DEVELOPMEST I N HYDROCHLORIC ACID SOLUTION 25 pg of germanium and 15 ml of 0.03 per cent. reagent solution used Drum xeadings measured after Hydrochloric acid I A \ concentration 15 min.30 min. 60 min. 90 min. 0.25 N 0.470 0.470 0.475 0.473 1.0 N 0.472 0.475 0.474 0.475 1.5 N 0.468 0.474 0.477 0.473 2.0 N 0,442 0.463 0.472 0.477 3.0 N 0.177 0.237 0.333 0,377 5.0 N 0.014 0.022 0.030 0.03 1 EFFECT OF GERMANIUM CONCENTRATION- With the quantity of phenylfluorone previously decided upon, viz., 15ml of 0.03 per cent. solution, colours were developed and measured for amounts of 10 to 50 pg of germanium. TABLE I V EFFECT OF GERMANIUM CONCENTRATION ON COLOUR DEVELOPMENT IN HYDROCHLORIC ACID SOLUTION Hydrochloric acid concentration 1.15 N ; 15 ml of 0.03 per cent. reagent used Drum reading Germanium, r A > rg Series 1 Series 2 Series 3 Series 4 Mean 10 0.208 0.204 0.209 0.208 0.207 20 0.402 0.403 0.387 0.397 0,397 30 0.577 0.582 0.566 0.575 0.575 40 0,739 0.741 0.725 0.742 0.737 50 0.866 0.871 0.872 0,880 0.872Sept., 19511 PART 11.ABSORPTIOMETRIC DETERMINATION WITH PHENYLFLUORONE 527 For convenience the volume of hydrochloric acid, sp.gr. 1.18, used was 5.0 ml, giving a con- centration of about 1.15 N in the final volume of 50 ml. The results of four such series of measurements, made at intervals over a period of about four months, are shown in Table IV. A calibration graph, prepared from the mean values, is shown in Fig. 2. The deviation from Beer’s law is more marked than was observed for the determination in sulphuric acid solution. STABILITY OF THE DEVELOPED COLOURS- The absorptiometric measurements recorded in Table IV were made after the solutions had stood in the cold for 30 minutes.Subsequent measurements confirmed that the colours were completely developed within this time and showed that their absorption remained constant for at least 14 hours. I t was usually observed that after 2 or 3 days precipitation of the germanium - phenylfluorone compound commenced. EFFECT OF OTHER ELEMENTS- A hydrochloric acid distillation would clearly separate germanium from most elements and it was therefore considered unnecessary to examine the effect of a large number of elements on the absorptiometric determination in hydrochloric acid solution. The effect of arsenic, which would be expected partly to co-distil with the germanium, and of some of the other elements previously found to interfere, was investigated.Under the conditions selected for the absorptiometric determination in hydrochloric acid solution it was found that arsenic did not interfere, that the interference of tin and titanium was diminished but that molybdenum and antimony still interfered strongly. DETERMINATION OF GERMANIUM AFTER SEPARATION BY HYDROCHLORIC ACID DISTILLATION Initially the simple distillation of germanium from pure solutions containing 50 per cent. by volume of hydrochloric acid was examined. This acid concentration, being very nearly the constant boiling composition, was convenient in that the acid concentration of the distillate would approximate to the same known strength throughout the distillation. With solutions of this acid concentration and of initial volume of 50 ml, the absorptio- metric procedure was applied to successive fractions of the distillate.It was established that not less than 95 per cent. of the germanium distilled over in the first 10m1, and that for complete recovery of the germanium it was necessary to distil only 20 ml. In this manner a number of determinations was made on pure germanium solutions, using an initial volume of 50 ml and applying the absorptiometric procedure to an appropriate aliquot of the 20 ml of distillate collected, with the results shown in Table V. A rate of distillation of about 2 ml per minute was used, so that a single distillation was complete in about 15 minutes. TABLE V ABSORPTIOMETRIC DETERMINATION OF GERMANIUM AFTER DISTILLATION FROM HYDROCHLORIC ACID SOLUTION Aliquot of distillate Germanium taken, used Germanium found, Pg PLg 10 50 _ _ 150 500 1000 25/60 25/50 10/50 20/250 20/500 i n - _ 50 148 510 1006 The same procedure was then applied to binary mixtures of germanium with other elements.These included arsenic, titanium and stannic tin, all of which form anhydrous chlorides with relatively low boiling-points, although it was considered unlikely that the last two elements would co-distil under the conditions used. The separations from molyb- denum and antimony were also examined individually as these two elements were known to exert the greatest interference with the absorptiometric determination. The results in528 CLULEY: THE DETERMINATIOii OF GERMANIUM [Vol. 76 Table VI show that none of these five elements interferes with the procedure even when initially present to the extent of 200 times the weight of germanium. TABLE VI ABSORPTIOMETRIC DETERMINATION OF GERMANIUN AFTER DISTILLATION FROM SOLUTIONS CONTAINING OTHER ELEMENTS Weight of Germanium mg /% Arsenic, as As,O, .. .. . . . . .. 10 50 Tin, as SnC14.6H,0 . . . . . . . . . . 10 60 Titanium, as Ti(SO,), . . .. .. . . . . 10 50 Antimony, as SbCl, .. . . 10 50 Molybdenum, as (NH4j~Mo,O,;.4H,d ' . . . . 10 60 Element added element, taken, Germanium found, Pg 60 50 51 49 49 The separation of germanium from other individual elements by this method of distillation was not examined extensively as, apart from the elements discussed above, none of the common elements forms chlorides likely to distil under the conditions employed.However, as a final test, the procedure was applied to known amounts of germanium in the presence of a complex mixture prepared to simulate a flue dust and containing the following ions: Na', K', Cu", Ag', Mg", Zn", BO,"', Al"', Ga"', SO,", Sn"", Pb", Ti"", PO4"', As ", Sb"', VO,', Cr"', Moo4'', Mn", Fe"' and Ni". Each ion was present in amount equivalent to 10 mg of the corresponding oxide except sodium, for which the equivalent of 1 g of the carbonate was added. The results in Table VII show good recovery of the germanium and no evidence of interference. These experiments were designed to assess the possibilities of the method for the determination of germanium in flue dusts, but they serve to show that the method should be applicable to a wide range of materials.TABLE YII DETERMINATION OF GERMANIUM IN SYNTHETIC FLUE DUST MIXTURES Elements present in the mixture Germanium added, Germanium found, Pg Na, K, Cu, Ag, Mg, Zn, B, 20 Al, Ga, Si, Sn, Pb, Ti, P, As, Sb, V, Cr, Mo, Mn, Fe, Ni 20 79 293 RECOMMENDED METHOD FOR THE ABSORPTIOMETRIC DETERMINATION OF GERMANIUM WITH PHENYLFLUORONE PREPARATION OF PHENYLFLUORONE- Dissolve 25 g of tri-acetylhydroxyhydroquinone by warming with a mixture of 150 ml of alcohol, 130 ml of water and 40 ml of diluted sulphuric acid (1 + 1). Add 25 g (about 24 ml) of benzaldehyde and allow the mixture to stand for 8 days, with occasional stirring. Filter by suction the yellow precipitate of phenylfluorone sulphate so obtained and wash the precipitate with the solution mixture.Suspend the precipitate in about 300ml of water, add sufficient sodium hydroxide solution to give a pH value of about 4 to ensure complete hydrolysis to phenylfluorone, stir and allow the mixture to stand overnight. Filter the precipitate of phenylfluorone by suction, and then wash first with water and finally three times with alcohol to remove the last traces of benzaldehyde. Dry the precipitate in a vacuum desiccator. SPECIAL REAGENTS- Phenyl$uorone solution-Dissolve 0.030 g of phenylfluorone by warming with a mixture of 85 ml of alcohol and 5 ml of dilute sulphuric acid (1 + 6), cool and dilute to 100 ml with alcohol. Standard sodium germanate solutiort (stock solution)-Transfer 14408 g of pure ignited germanium dioxide to a platinum crucible, fuse with 5 g of sodium carbonate and dissolveSept., 19511 PART 11.ABSORPTIOMETRIC DETERMINATION WITH PHENYLFLUORONE 529 the cold melt in hot water; just acidify the solution with dilute sulphuric acid, boil to eliminate carbon dioxide, cool and dilute to 1000 ml. 1 ml of stock solution = 1 mg of germanium. Standard sodium germanate solution (working solution)-Prepare freshly when required by 100-fold dilution of the stock solution. 1 ml of working solution = 10 pg of germanium. Gum arabic solution-Dissolve 1.0 g of gum arabic in 200 ml of hot water and cool. APPARATUS- The distillation apparatus consists essentially of a 100-ml distilling flask, a water-cooled condenser and a 25-ml measuring cylinder as receiver. The use of components with ground- glass joints facilitates the assembly and avoids trouble due to attack by hydrochloric acid on rubber bungs.PREPARATION OF CALIBRATION GRAPH- To 50-ml graduated flasks add amounts of 0 to 5 ml of the dilute (working) germanate solution, to cover the range 0 to 50 pg of germanium. Add sufficient water to give a volume of 20m1, then add 5ml of the gum arabic solution and 5.0ml of hydrochloric acid, sp.gr. 1.18, and mix. Cool to about 20" C and add from a pipette 15 ml of the phenylfluorone solution, dilute to 50 ml and mix well. Allow the solutions to stand at about 20" C for 30 minutes. Measure the germanium solutions on the Spekker absorptiometer or similar instrument with 1-cm cells, the tungsten lamp and Ilford No. 603 blue - green filters, using as a reference solution the solution containing no added germanium.Prepare a calibration graph from the results. PROCEDURE- The sample solution should contain 10 to 1000 pg of germanium and should be free from nitrates and other oxidants likely to liberate chlorine from strong hydrochloric acid solutions. Transfer the neutral or alkaline sample solution, or a suitable aliquot, to the 100-ml distillation flask, neutralise if necessary with hydrochloric acid and dilute to 25 ml as indicated by an appropriate line previously marked on the flask. Add 25 ml of hydrochloric acid, sp.gr. 1.18, swirl once and immediately connect the flask to the distillation apparatus. Heat the solution to boiling over a period of about 5 minutes and then distil at a rate of about 2 ml per minute.The distillate should be quite cold. Collect exactly 20 ml of distillate, remove the receiver and stop the distillation. Dilute the distillate to a known volume and transfer to a 50-ml graduated flask an aliquot that should not exceed half the distillate and is expected to contain 5 to 50 pg of germanium. With the same quantities of reagents carry out a blank distillation in the same manner and to a second 50-ml graduated flask transfer the corresponding aliquot of the blank distillate. Treat both the test and blank aliquots as follows, mixing after each addition. Add 5ml of the gum arabic solution followed by sufficient hydrochloric acid, sp.gr. 1.18, to make the total acid in the solution equivalent to 5.0 ml of hydrochloric acid. For this purpose it may be assumed that the 20 ml of distillate contained the equivalent of 10 ml of the concentrated acid.Add sufficient water to give a total volume of about 30 ml and adjust the temperature to about 20" C. Add with a pipette 15 ml of the phenylfluorone solution, dilute to 50 ml and allow to stand at about 20" C for 30 minutes. With the blank as a reference solution, measure the absorption of the test solution under the conditions used in the preparation of the calibration graph. From the graph deduce the germanium content of the measured solution and calculate the germanium content of the sample. DISCUSSION OF METHOD The experimental work had established that the distillation provided a satisfactory separation of germanium from any common element likely to interfere by colour reaction530 CLULEY : THE DETERMINATION OF GERMANIUM [Vol.76 with phenylfluorone. The reagent is, however, destroyed by strong oxidising agents, and for this reason the sample solution should not contain any substance that might liberate chlorine from the hydrochloric acid during the distillation. The practice of carrying out a blank distillation was adopted mainly to ensure that the acid concentrations in the blank and test aliquots were virtually identical. Any interfering elements present in reagents used prior to the distillation are also removed in this manner. Hitherto the absorptiometric determination of germanium has been based on the forma- tion of the yellow germano-molybdic acid or on its subsequent reduction to molybdenum blue, the latter method having the greater sensitivity and selectivity. Comparison with the published data on the molybdenum blue method2y3 shows that the phenylfluorone procedure, as applied in hydrochloric acid solution, is of similar selectivity and is about four times as sensitive.The main advantages of the phenylfluorone procedure are that it can be applied directly to a hydrochloric acid distillate with the minimum of manipulation, and that the distillation can be simply performed without the precautions necessary to attain complete separation from arsenic. After preparation of the sample solutions, the distillation and absorptiometric deter- mination with phenylfluorone can be completed in 2 to 24 hours for duplicate determinations. Hence this procedure should effect a considerable saving of time over such gravimetric methods as determination with tannin.This YS illustrated in the following paper dealing with the application of the phenylfluorone procedure to the determination of germanium in flue dusts. REFERENCES 1. 2. 3. Kitson, R. E., and Mellon, M. G., Ind. Eng. Chem., Anal. Ed., 1944, 16, 128. Hybbinette, A. G., and Sandell, E. B., Ibid., 1942, 14, 715. Boltz, D. F., and Mellon, hl. G., Anal. Chem., 1987, 19, 873. 4. 5. THE GENERAL ELECTRIC COMPANY LIMITED Gillis, J., Hoste, J., and Claeys, A,, Anal. Chim. Acta, 1947, 1 , 302. Dennis, L. M., and Johnson, E. B., J . Amer. Chem. Soc., 1923, 45, 1380. RESEARCH LABORATORIES WEMBLEY, MIDDLESEX and SIR JOHN CASS COLLEGE JEWRY STREET LONDON, E.C.3 Part 111.Determination in Flue Dust, Coal and Coke BY H. J. CLULEY The author's absorptiometric method (p. 523) for the determination of germanium with phenylfluorone has been applied to the analysis of flue dusts. The high sensitivity of the absorptiometric procedure permits the use of samples weighing 0.1 g or less and this greatly facilitates the decom- position of the sample by sodium carbonate fusion. Duplicate determinations can be completed in 3 to 3+ hours. The absorptiometric procedure has also been applied to the determination of germanium in coal and coke. In one method the sample is decomposed by a procedure similar to the Eschka method for the determination of sulphur in coal. I n a second method the sample is decomposed by combustion in a bomb calorimeter.Samples of coal and coke examined by these methods were found to contain 7 to 12 parts of germanium per million. THE available information on the determination of germanium in flue dusts and similar materials is somewhat meagre. One method for flue dusts that has been used at the Chemical Research Laboratory1 consists in fusion of the sample with sodium hydroxide, separation of the germanium from the majority of constituents, by precipitation as sulphide and finally gravimetric determination with tannin. Apart from being time-consuming, one apparentSept., 19511 PART III. DETERMISATION IN FLUE DUST, COAL AKD COKE 531 disadvantage of this procedure appears to be that for low germanium contents it may be desirable to carry out the fusion on several grams of sample.Alimarin and his co-workers have published a number of procedures for the similar problem of determining germanium in coal ash. These methods involve decomposition of the sample by treatment with hydrofluoric and sulphuric acid^,^$^ by fusion with sodium peroxide2p4 or by fusion with sodium carbonate and sulphur5; the germanium is then either separated by distillation of the tetrachloride and determined by precipitation as sulphide,2J or separated as sulphide and determined with tannin.415 Of these methods those involving a hydrofluoric acid attack appear open to criticism owing to the risk of loss of volatile germanium tetrafluoride. The methods involving fusion with an alkaline flux and final determination with tannin are similar to the method outlined above.The fact that Alimarin and his colleagues have within a short time proposed four successive methods might suggest that the earlier procedures were not wholly satisfactory. The previous work of the author on the absorptiometric determination of germanium with phenylfluorone (p. 523) indicated that this procedure might be applied with advantage to flue dusts. It had already been established that the procedure could readily be applied to germanium in the presence of substantial proportions of the large number of elements normally encountered in flue dusts. The high sensitivity of the method showed that a maximum sample weight of 0.1 g would be required and the use of such small weights should greatly facilitate the initial decomposition by fusion. The speed of the absorptiometric procedure also gave promise of a rapid method.This paper describes the application of the absorptiometric determination of germanium with phenylfluorone to flue dusts and subsequently to coal and coke. THE DETERMINATION OF GERMANIUM IN FLUE DUST Two flue dusts, samples 90 and 44, of different origins and of contrasting types, were selected for the initial experiments. In sample 44 most of the germanium was in a readily accessible form, as shown by the relatively high yield of germanium resulting from direct treatment with hydrochloric acid, whereas sample 90 had yielded only a small proportion of its germanium with this treatment. The germanium contents of these two samples were determined in the following manner. A 0.1-g portion of the ground sample was fused with 1 g of sodium carbonate in a platinum crucible, the melt being finally heated at 1000" C for 15 minutes. The melt was disintegrated with water and the mixture was transferred to a 100-ml distillation flask, neutralised with hydrochloric acid and diluted to 25 ml.A volume of 25 ml of hydrochloric acid was then added, and the distillation and the subsequent absorptiometric determination on an aliquot of the distillate were carried out as previously described.6 To investigate the possibility of loss of germanium by volatilisation during the fusion, further determinations were carried out in a similar manner but fusing at 1200" C, a treatment calculated to aggravate any such loss. TABLE I DETERMIXATIONS OF GERMANIUM IN FLUE DUSTS : COMPARISON OF PHENYLFLUORONE METHOD WITH OTHER METHODS Spectrographic Germanium estimate of Sample Method* found, Mean, germanium, Y O % Y O 0.69, 0.72 0.70, 0.70 B 0.65, 0.67 -4, 0.92, 0.94 44 A2 0.93, 0.94 B 0.85, 0.87 4 90 A2 0.71 0.70 0.8 0.66 0.93 0.94 1.0 0.86 * The methods of determination mere as follows- Method A,-Sodium carbonate fusion a t 1000" C, separation of germanium by hydrochloric acid Method A,-As A,, but fusion a t 1200" C.Method B-Sodium hydroxide fusion, separation of germanium as sulphide, final gravimetric deter- distillation, final absorptiometric determination with phenylfluorone. mination with tannin.532 CLULEY: THE DETERMINATION OF GERITASIUM [voi. 76 Complete solution was always obtained prior to the distillation; this showed that the methods of carbonate fusion as used resulted in complete decomposition of the samples, The results, shown in Table I, methods A, and ,Az, indicated good agreement between the individual determinations on each sample, and there was no evidence of loss of germanium during the fusion.As an independent chemical check on the results obtained absorptiometrically, deter- minations were also made by a procedure similar to the Chemical Research Laboratory method previously outlined. The results so obtained are also shown in Table I, method B. In addition, spectrographic estimations, based 011 comparison with synthetically prepared standards, were made on the two samples. Taking into consideration the difficulties of determining a minor constituent in such a complex material as flue dust, the agreement between the two chemical methods is reasonable. The results of both chemical methods are in approximate agreement with the spectrographic data, for which the expected accuracy would be 20.1 per cent.of germanium. The over-all mean of the chemical results for the two samples, 90 and 44, are 0.68 and 0.91 per cent. respectively, and when the completely independent chemical nature of the two chemical methods is borne in mind, it must be considered Ihat these values are very near to the true germanium contents of the two samples. Further trials of the method were made with samples of flue dust chosen to cover a range of germanium contents and to represent a number of different sources of the material. For these determinations fusion at the full heat of a Meker burner was used and the complete solution attained in all determinations proved thzt this method of decomposition was satis- factory. The results of the further trials are recorded in Table 11, which, for completeness, also contains the results obtained earlier for samples 90 and 44. It will be seen that the precision of the method is of a high order and that for each sample the value for the germanium content is consistent with the spectrographic result based on comparison with synthetic standards.TABLE I1 ABSORPTIOMETRIC DETERMINATIONS OF GERMANIUM IN FLUE DUSTS Sample 61 38 60 65 90 55 44 43 Germanium found, % 0.021, 0.021 0.14, 0.145 0.235, 0.23 0.30, 0.305 0.69, 0.72, 0.70, 0.70 0.825, 0.81 0.93, 0.94, 0.92, 0.94 1.15, 1-15 Mean, % 0,021 0.14 0.23 0.30 0.70 0.82 0.93 1.15 Spectrographic estimate of germanium, negligible % 0.1 0.2 0.3 0.8 0.7 1.0 > 1.0 Comparison with the tannin method shows the absorptiometric procedure to have a number of advantages.These include the ability to use small weights of sample, which greatly facilitates the decomposition by fusion an'd permits a more elegant procedure. The high sensitivity of the phenylfluorone reaction renders the absorptiometric method particularly suitable for samples containing small amounts of germanium; for example, 0.01 per cent. of germanium can readily be determined on 0.1 g of sample. The time required for duplicate determinations is 3 to 34 hours, compared with 13 to 2 days for the tannin method. The length of this latter method is largely due to the necessity for allowing the sulphide precipitate to stand overnight and to the lengthy low temperature ignition and nitric acid treatment of the tannin precipitate.RECOMMENDED METHOD FOR THE DETERMINATION OF GERMANIUM I N FLUE DUST APPARATUS AND REAGENTS- The distillation apparatus and the special reagents required for the final absorptiometric determination are described on p. 528 of the preceding paper on the absorptiometric deter- mination of germanium with phenylfluorone.Sept., 19511 PART 111. DETERhIINATIOh' 1 s FLUE DUST, COAL AKD COKE PROCEDURE- 533 Weigh 0.01 to 0.1 g of the ground sample, depending on expected germanium content, into a platinum crucible. Mix the sample intimately in the crucible with 0.5g of sodium carbonate and cover the charge with a further 0 6 g of sodium carbonate.Heat the crucible gently over a Meker burner, increasing the heat gradually over a period of about 5 minutes until a temperature of about 1000" C is reached. Continue heating in this manner, with occasional swirling of the melt, for a further 15 minutes. To the cool melt add about 10 ml of hot water and allow the crucible to stand on the edge of a hot-plate until the melt is completely disintegrated. Cool the crucible in running water and then transfer the contents to the 100-ml distilling flask. Add to the crucible 4 ml of diluted hydrochloric acid (1 $- l), swirl to dissolve any solid particles still adhering to the crucible and then transfer the acid solution to the main solution contained in the flask.Swirl the flask to promote the liberation of carbon dioxide and dilute the solution to 25 ml as indicated by an appropriate line previously marked on the flask. Add 25 ml of hydrochloric acid, sp.gr. 1.18, and carry out the distillation and the subsequent absorptio- metric determination on a suitable aliquot of distillate as described on p. 529 of the preceding paper,6 using a blank taken through the method as the reference solution. XOTE- that the prepared sample is representative and homogeneous. owing to the small weight of the sample. Owing to the heterogeneous nature of flue dusts, precautions must be taken to ensure This is particularly important THE DETERMINATIOK OF GERMANIUM IN C09L AND COKE The facility with which germanium in flue dust can be determined by the above method encouraged an attempt to apply the phenylfluorone procedure to the determination of ger- manium in coal and coke.The germanium content of so-called "germaniferous" coal appears to be of the order of 10 parts per million, and other workers7~* have ashed the coal to effect a concentration of the germanium before determination. I t has been stated, however, that unless considerable care is taken to maintain strongly oxidising conditions during the ashing, substantial loss of germanium by volatilisation may It was considered that the sensitivity of the phenylfluorone reaction might permit a direct determination on the coal or coke without recourse to preliminary ashing. Initial experiments were carried out on a sample of Boldon coal and on a sample of coke from this coal.The procedure used for the decomposition of the sample was similar to that used in the Eschka method for determination of sulphur. Sodium carbonate alone was used in place of the usual admixture with magnesium oxide, as it was required to fuse the residue remaining after destruction of the carbonaceous matter. This destruction was carried out at 600" C instead of the customary 800" C in order to prevent loss by volatilisation of germanous oxide, which is stated to be volatile at 700" C. The determinations were completed by distillation and absorptiometric measurement as used for the determination of germanium in flue dusts. As a check on the results of these determinations a different method of decomposition was used. Samples were decomposed in a bomb calorimeter by precisely the same method a s used for sulphur determinations, and the contents of the bomb were washed into a platinum dish.Sodium carbonate was added to neutralise the acids formed during the combustion and the solution was evaporated to dryness. The residue was then fused with a further quantity of sodium carbonate and the determination completed by distillation and absorptio- metric measurement. The results of determinations by these two procedures are shown in Table 111, which also includes results by the sodium carbonate - ignition method on further samples of coal and coke. Although the combustion in the bomb should reduce the risk of loss of germanium by volatilisation to a minimum, it will be observed that the results so obtained are slightly lower than those of the sodium carbonate - ignition method.This difference is attributed to some loss of ash incurred by fusion into the silica crucible used to contain the sample for combustion. I t is likely that this loss could be avoided by using a platinum container534 CLULEY: THE DETERMINATION O F GERMAXIUM [l'ol 76 if work of higher accuracy were desired. It is interesting to note that spectrographic examina- tion of the four samples, made before the chemical determinations, showed that the most sensitive lines of germanium could only just be tdetected, and from this it was concluded that the germanium content of the samples was approximately of the order of 10 parts per million. TABLE I11 DETERMINlTIONS OF GERMANIlJM I N COAL AND COKE Sodium carbonate - ignition method Bomb method r-y---------7 7 Germanium found, Mean, Germanium found, Mean, p.p.m.P P.m. p.p.m. p.p.m. Boldon coal . . .. . . 7.0, 8.0 7.5 6.6, 6.6 6.5 Coke from Boldon coal. . . . 7.0, 8.0 1.5 6.1, 6.4 6 Harton coal .. . . . . 7.2, 9.2 8 Coke from Harton coal . . 11.0, 12.5 12 ,4 comparison of the two procedures shows that the ignition method requires considerably less manipulation and has the advantage that a number of samples can be decomposed simultaneously. However, the bomb method has a greater sensitivity as it uses a 1-g sample in contrast to a maximum of 0.5 g for the ignition method. The limitation of the size of the sample in the latter method is associated with the quantity of sodium carbonate required, as large concentrations of sodium salts result in precipitation of sodium chloride from the strong hydrochloric acid solution used for the distillation. During these experiments it was found that the effective sensitivity of both methods could be enhanced by collecting a smaller volume of distillate and applying the absorptio- metric procedure directly to this without subdivision. The loss of germanium resulting from this modification was found to be negligible for the very small amounts of germanium involved.In this manner it should be possible to determine germanium down to 4 p.p.m. with the ignition method and 2 p.p.m. with the bomb method. Although complete data are not available for an accurate assessment of the efficiency with which germanium is deposited in flue dusts, it may be of interest to record the information resulting from the above determinations.This shows that on conversion of the Boldon coal to coke, some 60 per cent. of the germanium was retained by the coke; on use of this coke in producers, a much smaller proportion of germanium was deposited in the flue dusts in the waste heat system, although these dusts contained sufficient to justify extraction of the element. - - - - THE TWO METHODS FOR THE DETERMINATION OF GERMANIUM I N COAL AKD COKE APPARATUS AND SPECIAL REAGENTS- The distillation apparatus and the special reagents required for the final absorptiometric determination are described in the preceding paper on the absorptiometric determination of germanium with phenylfluorone.6 The bomb calorimeter required for the bomb method is of the conventional type used for the determinations of calorific value and sulphur in coal and coke.SODIUM CARBONATE - IGNITION METHOD- Weigh into a platinum crucible 0.5 g of the sample previously ground to pass a 120 B.S. screen. Mix the sample intimately in the crucible with 1.5 g of sodium carbonate (do not tap down) and cover the charge with a further 0.5 g of sodium carbonate. Transfer the uncovered crucible to a muffle furnace and, allowing access of air, raise the temperature to 600" C over a period of about an hour. Maintain this temperature for 14 hours, or longer if carbonaceous matter is still visible. Remove the crucible from the furnace, stir the contents and return the crucible to the furnace for a further hour to destroy any last traces of carbonaceous matter.Fuse the contents of the crucible, disintegrate the melt and prepare the solution for distillation in the same manner as used in the method for flue dusts, except that the volume of diluted hydrochloric acid (1 + 1) used should be 8 ml to neutralise the larger weight of sodium carbonate used.Sept., 19511 PART 111. DETERMINATIOK I S FLUE DUST, COAL ASD COKE 535 Distil at the rate of about 2 ml per minute, collecting only the first 10 ml of distillate. Apply the absorptiometric procedure directly to this distillate without subdivision and to a blank distillate similarly prepared as described in the preceding paper.6 The 10ml of distillate contains the appropriate amount of hydrochloric acid for the absorptiometric determination and no further addition of acid is required.BOMB METHOD- Place 10 ml of water in the bomb and carry out the combustion of the pellet in the normal manner, with an oxygen pressure of 25 atmospheres. After the combustion allow the bomb to stand for 30 minutes to permit the acid mist to settle and then slowly release the pressure. Wash the solution and ash into a platinum dish, add 0.2g of sodium carbonate to make the solution alkaline and then evaporate to dryness. Add 1 g of sodium carbonate, fuse, disintegrate the melt and prepare the solution for distillation exactly as in the method for flue dusts. Perform the distillation and the final absorptiometric determination exactly as in the ignition method. REFERENCES Make about 1 g of the ground sample into a pellet and weigh it.1. 2. 3. 4. Alimarin, I. P., and Aleksewa, 0. A., Ibid., 1940, 13, 1393. 5. 6. 7. 8. THE GENERAL ELECTRIC COMPANY LIMITED “Chemical Research, 1938-46,” H.M. Stationery Office, London, p. 30. Alimarin, I. P., Ivanov-Emin, B. N., Medvedeva, 0. A., and Yanovskaya, C. Y., Zauod. Lab., Alimarin, I. P., and Ivanov-Emin, B. N., J . A@$. Chem., U.S.S.R., 1940, 13, 951. Alimarin, I. P., Trudy Vsesoyuz Konferentsii Anal. Khim., 1943, 2, 371. Cluley, H. J,, Analyst, 1951, 76, 523. Goldschmidt, V. M., J . Chem. SOC., 1937, 655. Morgan, G. T., and Davies, G. R., Chem. and Ind., 1937, 16, 717. 1940, 9, 271. RESEARCH LABORATORIES WEMBLEY, MIDDLESEX and SIR JOHN CASS COLLEGE JEWRY STREET LONDON, E.C.3 DISCUSSIOK ON THE ABOVE THREE PAPERS MR. W.H. BENNETT asked whether fluorine interfered with the phenylfluorone colour reaction, and if it did, had the author considered the possibility of the interference of fluorine in the determination of germanium in coals,. some of which contained small amounts of fluorine. MR. CLULEY replied that he had no specific information on the effect of fluorine on the colour reaction, (Later work has shown that as much as 2.5 mg of fluorine may be present as fluoride in the final solution without effect on the colour reaction, although substantially greater amounts of fluorine cause interference by reducing the intensity of the germanium colour, presumably owing to partial complexing of the germanium. When the maximum sample weight of 1 g of coal was used, the amount of fluorine in the coal would not exceed 0.2 mg, so that no interference from this source would be expected.) MR.R. F. MILTON drew attention to the author’s statement that reduction of germanomolybdate to molybdenum blue was not satisfactory in the presence of arsenic, phosphorus or silicon. In point of fact, each one of this group of elements could be estimated in the presence qf the others, if cognisance were taken of the conditions under which the complex molybdate formation occurred. Thus, silicomolybdate formed in 0.1 N sulphuric acid and the reduction occurred in 2 N sulphuric acid with stannous chloride. The molybdenum blue could be extracted with butyl solvent and estimated. If the residual solution were neutralised to 1.2 N sulphuric acid, phosphomolybdate would reduce and the colour could then be extracted and estimated. The reduction of acidity to 0.9 N would allow germanomolybdate to reduce and to be estimated. If arsenic, which had been maintained in the tervalent state, were then oxidised, it would form the arsenomolybdate, which could then be estimated after reduction with stannous chloride. He had two questions to put to the author. First, he asked if it was necessary in the determination of germanium in flue dust to make a fusion of the material before the hydrochloric acid distillation. In his experience, when arsenic and germanium were present together, the germanium usually distilled very readily with the arsenic when heated with hydrochloric acid. Was it possible to distil the flue dust directly with hydrochloric acid? Secondly, he asked whether the author could give any idea of the sensitivity of the colour reaction DR. J. H. HAMENCE said he had listened t o the papers with great interest.DETERMIXATION OF THE GRADE STRENGTH OF PECTINS [Vol. 76 536 between phenylfluorone and molybdenum. Detection of traces of molybdenum had now assumed con- siderable importance in agricultural work, and new sensitive tests for this element were always welcome. Finally, Dr. Hamence thanked the author for having shown in a most conclusive manner that the distillation process for the separation of arsenic from antimony and tin was in fact a most efficient method, and that neither of these elements distilled over with the arsenic. In the past this had always been assumed to be a fact, but to the best of his knowledge it had not previously been confirmed experimentally. MR. CLULEY replied that although some flue dusts gave an almost quantitative yield of the germanium on direct distillation with hydrochloric acid, other dusts afforded only a small proportion of their germanium with this treatment. In the latter type of dust the germanium was presumably mainly present in a chemical form in which it was not readily soluble in hydrochloric acid. Decomposition of the sample by fusion was therefore necessary to ensure complete recovery of the germanium in the determination. The sensitivity of the colour reaction of phenylfluorone with molybdenum was of the same high order as the germanium reaction, and phenylfluorone could probably be used for the detection of molybdenum if required. Gillis, Claeys and Hostel have stated that the associated reagent o-hydroxyphenylfluorone is, under appropriate conditions, a specific reagent for the detection of molybdenum, REFERENCE TO DISCUSSION 1. Gillis, J , , Claeys, A,, and Hoste, J., Anal. Chim. Acta, 1947, 1, 421.