QD 
 
 M654 
 
 
 UC-NRLF 
 
 
 
 1 
 
 
 $B SD 776 
 
 mm 
 
 O 
 O 
 CO 
 
 o 
 
LIBRARY 
 
 OF THE 
 
 University of California. 
 
 RECEIVED BY EXCHANGE 
 
 Class 
 
The Electrolytic Determination 
 of Manganese and Its Sepa- 
 ration from Zinc and Iron. 
 
 THESIS. 
 
 Presented to the Faculty ot the Department 
 of Philosophy ot the University ot Penn- 
 sylvania in Partial Fulfilment of the 
 Requirements for the Degree ot 
 Doctor of Philosophy, 
 
 BY 
 
 George Philipp Scholl, 
 Philadelphia. 
 
 Eastox. Pa. : 
 
 The Chemical Publishing Company. 
 
 1903. 
 
The Electrolytic Determination 
 of Manganese and Its Sepa- 
 ration from Zinc and Iron, 
 
 THESIS. 
 
 Presented to the Faculty of the Department 
 of Philosophy ot the University of Penn- 
 sylvania in Partial Fulfilment of the 
 Requirements for the Degree of 
 Doctor of Philosophy, 
 
 BY 
 
 George Philipp Scholl, 
 
 n 
 
 Philadelphia. 
 
 OF THE * 
 
 UNIVERSITY 
 
 OF 
 
 Easton, Pa. : 
 
 The Chemical Publishing Company. 
 
 1903. 
 

THE 
 
 UNIVERSITY 
 
 THE ELECTROLYTIC DETERMINATION OF MANGA- 
 NESE AND ITS SEPARATION FROM ZINC AND IRON. 
 
 The deposition of manganese during electrolysis takes place in 
 a different manner from that of the other metals, with the excep- 
 tion of lead, inasmuch as under ordinary conditions the metal is 
 not deposited as such on the kathode, but comes down as peroxide 
 on the anode. Under certain conditions it is, however, possible to 
 deposit the metal as such on the kathode, 1 but its quantitative esti- 
 mation has been found impossible on account of the very rapid 
 oxidation of the metal during the process of drying. This method 
 therefore cannot be used for analytical purposes. 
 
 The deposition of manganese as peroxide for the purpose of its 
 quantitative chemical estimation by electrolysis has been the subject 
 of a good deal of investigation, 2 but the results have not been very 
 satisfactory. The stumbling block in the way of a successful pre- 
 cipitation has been the fact, that it is extremely difficult to obtain 
 the deposit on the anode in an adherent form, it showing a tend- 
 ency to become loose and to scale off. This even takes place 
 when the peroxide is deposited on a platinum dish, the inner sur- 
 face of which has been roughened by subjecting it to the action of 
 a sand-blast. Owing to this tendency, manganese is probably the 
 most difficult metal to deal with, as far as electrolysis is concerned. 
 
 Various electrolytes have been proposed for the quantitative 
 estimation of manganese by electrolysis, but in 1894 Classen 3 states, 
 that none of the methods heretofore proposed had been found by 
 him to yield satisfactory results. He advises the use of acetic acid 
 as an electrolyte and states that it has been found by him to be the 
 most suitable of all the organic acids. This statement is repeated 
 in the last (4th edition) of his "Quantitative Chemical Analysis by 
 Electrolysis," translated by Herrick and Boltwood, 1898, p. 149, 
 where it is further said, that "the precipitation of large quantities, 
 
 1 Smith and Prankel : Journal Analyt. Chem., 3, 3S5; Chemical News, 60, 262. 
 
 2 literature given by Smith, Electrochemical Analysis, 3rd ed., 1902; Riche : Annales 
 de Chimie et de Physique [5th series], 13, 50S ; Iyuckow : Zeitsch. f. anal. Chem., 19, 17 ; 
 Schucht; Zeitsch. f. anal. Chem., 33, 493; Classen and von Reiss : Berichte deutsch. 
 Chem. Ges., 14, 1622 ; Brant: Zeitsch. f. anal. Chem., 38, 581; Riidorff ; Zeitsch. f. angew. 
 Chem., 15, 6 ; Classen : Berichte deutsch chem. Ges., 37, 2060 ; Engels : Zeitsch. f. Elec- 
 trochemie, 3, 413; 3, 286; Groeger: Zeitsch. f. angew. Chem., 1895, p. 253; Koeppel: Zeitsch. 
 f. anorgan. Chem., 16, 268. 
 
 i Berichte d. deutsch. Chem. Ges., 37, 2060. 
 
 173285 
 
even when roughened dishes are used, cannot be successfully car- 
 ried out, since it is impossible to obtain firmly adherent precipi- 
 tates." Evidently no improvement in this method had been made 
 up to 1901, for in Classen's large handbook on "Selected Methods 
 of Analytical Chemistry," which appeared in that year, the follow- 
 ing statement is found on page 368 : "In acetic acid solution it is 
 only possible to deposit quantities of about 0.08 gram of manga- 
 nese in the form of peroxide, even on roughened dishes." 
 
 A few years ago, Engels 1 proposed a method, which is said to 
 give firmly adherent deposits up to 0.25 gram of manganese, and 
 which consisted in adding chrome alum in quantities of from 1.5 
 to 2 grams to the solution of manganese sulphate. This method is 
 open to objection, inasmuch as the addition of chrome alum intro- 
 duces a complex metallic compound into the electrolyte, which 
 compound by its own decomposition under the influence of the cur- 
 rent may give rise to complications. Further, as stated by 
 Kaeppel, 2 an extended washing of the deposit, on account of the 
 chromic acid adhering to it, is absolutely necessary, and a source 
 of error may easily be introduced from the fact, that chromic acid, 
 which is carried down by the deposit, is quite firmly held by the 
 peroxide of maganese. 
 
 The last work done on the subject in recent years has been by 
 Kaeppel, who in 1898 published a method 2 by which, according to 
 him, very adherent deposits of manganese peroxide are obtained. 
 It consists in the addition of acetone, in quantities varying from 
 2.5 to 10 cc, to a solution of manganese sulphate. He gives one 
 experiment, in which he obtained 1.61 grams of peroxide of man- 
 ganese in a firmly adherent form. Classen 3 states, however, that 
 the carrying out of Kaeppel's method in his laboratory has not led 
 to any satisfactory result. 
 
 The question of the deposition of manganese was therefore, as 
 is apparent from the above, in quite an unsatisfactory state. 
 Engels's method was open to the objections mentioned, and 
 Kaeppel's method was described by him in such a way in his publi- 
 cation, that it was impossible to repeat it. As has been justly 
 criticized by Classen, 4 he omitted to state the precise conditions, 
 under which his experiments were performed. He gives no data 
 
 1 Zeitsch. f. Electrochemie, 2, 413 ; 3, 286. 
 J Zeitsch. f. anorgan. Chemie. 16, 268. 
 
 3 Selected Methods of Analytical Chemistry, 1901, p. 370. 
 
 4 Ausgewahlte Methoden der analytischen Chemie (selected methods, etc.), p. 370. 
 
whatever in regard to the size and form of his electrodes, the cur- 
 rent density at the anode and kathode or the general arrangement 
 for carrying out the electrolysis. It is quite apparent to any one 
 conversant with electrolytic operations in general, that these are 
 questions, which largely influence the result, and on which the 
 success of the operation may depend and usually does depend to no 
 small degree. 1 Such being the case, the question presented itself, 
 if it was not possible to find another electrolyte, which would lend 
 itself to a more satisfactory deposition of the peroxide of manga- 
 nese, and to the separation of manganese from other metals, prin- 
 cipally iron. Organc acids appeared most suitable for this pur- 
 pose. Classen 2 maintains, that strong inorganic acids interfere 
 with the deposit and even make it impossible. The same author's 
 statement, that acetic acid is the best of all organic acids, has been 
 referred to above. No further experimental details are given in 
 his publications to substantiate that statement and there is also no 
 mention made as to which organic acids have been tried and under 
 what conditions. Kaeppel, in his paper, makes some remarks 
 about formic acid. He says: "I then carried out experiments 
 with formic acid. By the use of a few drops of it, an adherent 
 deposit could also be obtained, but it was shown still more, that a 
 quantitative deposition could not be effected. The addition of 
 more than 5 drops of the acid prevented the deposition of peroxide 
 altogether. This statement is also unaccompanied by any experi- 
 mental evidence as to the strength of acid used, the manner of 
 carrying out the electrolysis and the general conditions under 
 which it took place. 
 
 The above statements to the contrary notwithstanding, it was 
 considered, that formic acid, on account of its simple constitution, 
 would be an extremely desirable substance to be used as electro- 
 lyte, and it was thought, that under the right conditions it might 
 be made to yield satisfactory results. It was therefore determined 
 to try it with the addition of sodium formate. 
 
 The apparatus used for carrying out the experiments, consisted 
 of the usual stand for electrolysis, provided with a glass rod 
 carrying a ring wih 3 contact points, for the purpose of supporting 
 the platinum dishes in which the deposition took place, and the 
 
 1 See in this connection the pertinent remarks of Smith :Electrochemical Analysis 
 3rd ed., p. 33-36. 
 
 2 Ibid., p. 368. Quantitative Analysis by Electrolysis, p. 149. 
 
usual clamp for fastening the other electrode. The current was 
 supplied by storage batteries, which were constantly kept in a 
 charged condition. The current was regulated by means of a disk 
 rheostat, and a Weston voltmeter and amperemeter were kept 
 constantly in the circuit. The platinum dishes used for the deposi- 
 tion of the peroxide of manganese were of the usual shape and 
 had a diameter of 9 cm. Their interior surface was roughened. 
 The kathode consisted of a spiral of platinum wire of 1^2 mm. 
 diameter, having a surface of 6 square cm. As usual, the dishes 
 were covered by cover glasses. 
 
 A solution of manganese sulphate was made up, calculated to 
 contain about 10.60 grams of manganese in the liter. The formic 
 acid used in the experiments was the chemically-pure article of 
 1.06 specific gravity, made by Merck & Co. The electrolysis was 
 conducted in the following manner: 10 cc. of the manganese 
 sulphate solution were introduced into the platinum dish and 
 5 cc. of the formic acid solution of 1.06 specific gravity added to it. 
 A solution of .10 cc. of sodium formate, containing 1 gram of the 
 salt, was then added, and the whole volume of the solution made 
 up with distilled water to about 130 cc, so that it covered, as 
 accurately as could be determined, 100 sq. cm. of the surface of 
 the dish. The dish was connected as anode and the spiral kathode 
 introduced into the solution, care being taken to arrange it as 
 nearly as possible equidistant from the sides and the bottom of 
 the dish, in order to obtain as uniform a distribution of the cur- 
 rent as possible. It was considered best to use a high current 
 density, as most suitable for the deposition of the peroxide. The 
 current was therefore regulated to give a strength of 1.4 amperes, 
 which was maintained during electrolysis. The pressure between 
 the electrodes was 12 volts at the beginning, decreasing to 8.6 
 volts at the end. Almost immediately after the circuit had been 
 closed, a deep black, shining deposit appeared at the anode. The 
 current was allowed to act for one and one-half hours, when the 
 level of the liquid was raised to ascertain whether any more 
 peroxide would precipitate. As no further precipitation took place, 
 the liquid was siphoned off and replaced by water, without in- 
 terrupting the current. The deposit showed a deep black, shining 
 color and was firmly adherent to the dish, showing no tendency 
 to come off during washing. It was washed carefully with water 
 
and dried. Smith 1 states that it is useless to try to obtain a 
 definite compound by drying the deposit, as it is so extremely 
 hygroscopic, that ignition alone to the protosesquioxide will give 
 definite and concordant results. This was found to be true. 
 Though the precipitate looked very well and had quite a metallic 
 appearance, it was not possible to get a concordant weight, as it 
 attracted moisture very rapidly. Following the directions given 
 by Smith (I.e.) the deposit was ignited over the Bunsen burner 
 and the blast lamp, and thus changed to the protosesquioxide, 
 when it could be weighed without any trouble. The solution, 
 when tested, failed to show the presence of manganese. The 
 following table shows the results of four determinations made in 
 exactly the same manner : 
 
 N.D.iqr Mn found. 
 
 Electrolyte. Time. Temperature. at anode. Voltage. Grams. 
 
 10 cc. MnS0 4 Ordinary temp. 
 
 + 5CC.HCOOH at start, heated 12.0 volts at °'"°° 
 
 1.06 sp. gr. i 1 /, hrs. up by action of 1.4 start to 8.6 " 1I0 5 
 
 -f- 10 cc.= 1 gm. current to 40 - volts at end. * II0 7 
 
 sodium formate. 50 at the end. 
 
 It was thus established that Kaeppel's statement, that more 
 than five drops of formic acid entirely prevented the formation 
 of peroxide of manganese, did not hold true under the conditions 
 under which the experiments were run. It was further proved 
 that it was very well possible to deposit the whole of the man- 
 ganese in the presence of formic acid, even if the acid was present 
 in amounts considerably above that given by Kaeppel as the 
 maximum. According to Classen, 2 it is necessary in the carrying 
 out of Engels's method, to weigh the platinum dish after the 
 deposit of peroxide has been removed from it, as the dish suffers 
 a diminution of weight during electrolysis, amounting to about 
 1 mg. With the electrolyte used in the experiments under con- 
 sideration this does not take place, the dish losing either no 
 weight at all or at the most only 1 / 10 mg. 
 
 Though it was evident from the results tabulated above that 
 concordant results could be obtained in this way by exactly dupli- 
 cating the experiments, yet the results obtained were too high. It 
 was also observed that even a prolonged ignition over theblast lamp 
 failed to remove the hygroscopic character of the deposit entirely. 
 Though the deposit could be weighed without any difficulty, it 
 
 1 Electrochemical Analysis, 3rded., p. 96. 
 
 2 Ausgewahlte Methoden, p. 369. 
 
8 
 
 continued to slowly attract moisture, and when standing over 
 night, had gained perceptibly in weight. 
 
 Experiments were then conducted with a low voltage, so as to 
 study the effect of a lower current and consequently lower current 
 density on the deposit. The character of the deposit, however, 
 remained unchanged, the lower currents, of course, requiring a 
 proportionately longer time to effect the complete deposition of 
 the manganese. Heating up of the electrolyte, to about 50 or 
 55 , had no accelerating effect, but it was in so far harmful, as 
 it tended to make the deposit blistery and was inclined to scale off. 
 It was then tried to see whether larger quantities of sodium 
 formate would have a beneficial effect. It was, however, ascer- 
 tained that larger quantities of that salt had a decidedly retardent 
 effect. If it was present in quantities of 2*/2 grams the deposition 
 was not complete in less than two and one-half hours. An in- 
 teresting fact was revealed by the results, which are tabulated 
 below . 
 
 Manganese 
 sulphate. 
 
 Formic 
 
 acid. 
 1.06 sp.g. 
 
 Sodium 
 
 formate. 
 
 Grams. 
 
 Time. 
 
 Temp. 
 
 N.D 100 . 
 
 Voltage. 
 
 Mn found. 
 
 IO 
 
 5 
 
 i over night 
 
 ordinary 
 
 O.I8 
 
 3.0 
 
 O.II31 
 
 IO 
 
 5 
 
 1 
 
 4 
 
 " 
 
 0.2I 
 
 4.0 
 
 O.II2I 
 
 IO 
 
 5 
 
 1 
 
 3 
 
 «< 
 
 0.45 
 
 5-o 
 
 0.1 108 
 
 IO 
 
 5 
 
 2 
 
 2 
 
 << 
 
 I.O 
 
 4-5 
 
 0.1 154 
 
 IO 
 
 5 
 
 2# 
 
 *X 
 
 «( 
 
 1.0 
 
 3-9 
 
 0.1 180 
 
 IO 
 
 5 
 
 *y* 
 
 *% 
 
 (C 
 
 1.0 
 
 3-9 
 
 O.II72 
 
 It will be noted that the amount of manganese found did not 
 agree at all with the amount found in the experiments 
 tabulated before. The amount of sodium formate present, as 
 well as the time necessary for the electrolysis were both shown 
 to have an influence on the result, the weight of the deposit in- 
 creasing with the longer time and the larger amount of sodium 
 formate. The hygroscopic nature of the deposits obtained was 
 very evident, it being hardly possible to weigh those gotten when 
 2,y 2 grams of sodium formate were present, even after they had 
 been ignited over the blast lamp. The most reasonable explana- 
 tion of this fact seemed to be that the sodium formate was also 
 partly decomposed under the influence of the current, with the 
 formation of caustic soda at the kathode. Some of this caustic 
 soda, being specifically heavier ttfan the electrolyte, falling down 
 to the bottom of the dish anode, might easily have been enclosed 
 
9 
 
 and held mechanically by the depositing peroxide of manganese. 
 When the precipitation was slower, the same amount of sodium 
 formate being present, the deposit had more time to seize upon 
 and enclose alkali, hence the greater weight of the precipitate, the 
 longer electrolysis was continued. On the other hand, when 
 much formate was present, a greater amount of it would be de- 
 composed and considerably more alkali could get into the deposit. 
 
 It should be noted in this connection that careless washing of 
 the precipitated peroxide was entirely excluded, as everyone of 
 the deposits was treated exactly alike. Whether the above ex- 
 planation is the correct one has not been ascertained, as it was 
 shown very soon by the following experiments that good results 
 could be obtained without the use of sodium formate. That it is 
 a perfectly rational one, in view of the results shown above and of 
 the hygroscopic character of the deposit, even after ignition, can 
 hardly be doubted. The tendency of the caustic alkali solution, 
 which is produced during electrolysis, to fall down to the bottom 
 of the vessel, is quite well known, as is also the fact that even in 
 ordinary quantitative analysis it is a most difficult thing to get a 
 precipitate free from alkali. 
 
 It was thus established that, though perfectly concordant results 
 could be obtained by using formic acid with the addition of sodium 
 formate, by adhering exactly to the same conditions, yet the 
 results contained a source of error, which would make them prac- 
 tically worthless. However, the good results obtained so far 
 with formic acid had raised quite a doubt as to the accuracy of 
 Kaeppel's statement with regard to formic acid preventing the 
 deposition of peroxide of manganese. 
 
 The next step was, consequently, to do away with the source of 
 error traceable to the addition of sodium formate and to try formic 
 acid alone. Accordingly, 10 cc. of the manganese sulphate solu- 
 tion were taken, 5 cc. of formic acid of 1.06 specific gravity added, 
 and the solution subjected to electrolysis in the usual manner. 
 The voltage required was pretty high, on account of the low 
 conductivity of the formic acid, it being 17.5 volts between the 
 electrodes at the start, with a current strength of 1 ampere. As 
 the solution got very warm in about an hour, by the passage of 
 the current, the voltage was decreased to 7.9 volts, giving an 
 amperage of 0.8. The deposit did not come down as quickly as 
 when sodium formate was present, but the manganese was all 
 
IO 
 
 out after four hours. The heating of the electrolyte, as had been 
 ascertained before, did not have a favorable influence, tending to 
 make the deposit blistery. The deposited peroxide of manganese 
 was washed carefully and ignited in the usual way. The weight 
 of the deposit, calculated as Mn, was 0.1063 gram, a value which 
 closely agreed with that calculated to be present. It was further 
 found that the precipitate was absolutely non-hygroscopic after 
 ignition, and could be weighed with perfect ease. It was left 
 purposely exposed to the action of the atmosphere over night and 
 reweighed on the following morning, when the weight was found 
 to be the same. The appearance of the deposit was the same as 
 that obtained before. Kaeppel 1 contends that it is possible to 
 weigh the deposit after drying for about two hours at 150 to 
 180 , thus doing away with the necessity of changing it to the 
 protosesquioxide by ignition. This was tried, but Kaeppel's 
 observation could not be confirmed, as the deposit attracted 
 moisture too quickly to make it possible to obtain reliable results. 
 
 Another solution of manganese sulphate was then made up, and 
 the manganese determined gravimetrically as pyrophosphate. 
 There was 0.1034 gram of Mn found present in 10 cc. As the 
 determinations with employment of the spiral of platinum wire 
 as kathode had necessitated such a high potential between the 
 electrodes in order to obtain the necessary strength of current 
 and had thereby caused objectionable heating-up of the electrolyte, 
 it was decided to adopt some means to get along with as low a 
 voltage as possible, which would give the necessary current. 
 
 This could, of course, only be accomplished by reducing the 
 resistance of the electrolyte, either by increasing its conductivity 
 by the addition of a better conducting material, or by increasing 
 the electrode surface. As the former course was intended to be 
 avoided, the latter presented itself as the only solution. Accord- 
 ingly, a so-called basket electrode was used, being a platinum 
 dish, which conformed in shape to the interior surface of the 
 roughened dish on which the precipitation of peroxide of man- 
 ganese took place. It had a surface of 60 sq. cm. and was per- 
 forated with holes to allow of a better egress of the gases set free 
 during electrolysis and a better mixing of the solution. The 
 following table gives the results of some experiments carried out 
 with varying quantities of formic acid and varying quantities of 
 manganese. 
 
 1 I,oc. cit., p. 274. 
 
II 
 
 p p 
 
 0> ON 
 
 00 00 
 
 oooooooooooooo 
 
 O Q C/i Gn C/« Cn Q C 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 OJ 
 
 Ui 
 
 C^i 
 
 OJ 
 
 (j» 
 
 £ 
 
 s 
 
 £ 
 
 Jfc 
 
 £ 
 
 o p p p p p p p p p p p p p p p p 
 
 KJ Jo W JoJoMMMMMt-tMMMMM M 
 
 00000-**C/iC/itnOOOOOOO O 
 
 mmOvOvOvOOKJi-<K)4^0iO>>Oi-P».-P>> O" 
 
 2 a 
 
 if 
 
 Cn O OtnCn O **J C/i Cn tn o OCnOttnCn 
 
 vO 3 
 0) 2. 
 
 " p 
 
 OtOl (A Ul A 
 
 „W Ol 
 
 OJ 
 
 * 
 
 s * 
 
 *& 
 
 5 I 
 
 5 n.3, ^ 
 
 On 
 
 ON ON "^ 
 
 b\ bo b 
 
 d> b 
 
 e .2 
 
 i 1 
 
 I 
 
 5 
 
12 
 
 As will be seen from the above table, the results obtained by 
 electrolysis agree very closely with the gravimetric. For some 
 reason, which has not been further investigated, the addition of 
 sodium formate appears to have a tendency to accelerate the 
 deposition, as the metal was out of the solution in one and one-half 
 hours when sodium formate was present, while it took two and 
 one-half hours for the formic acid alone. The deposits obtained 
 were all very fine and beautifully adherent, only with 0.2880 gram 
 of Mn 3 4 a slight disposition towards loosening was noticeable. 
 It was very slight however, and, owing to the scaly nature of the 
 material it was quite easy to filter it out quickly on a small filter, 
 wash, dry and remove it from the filter and add it to the bulk of 
 the material in the dish, together with the ashes of the filter 
 paper. 
 
 It is also apparent from the above table that the addition of 
 more than 5 cc. of formic acid of 1.06 specific gravity has no 
 beneficial effect. On the contrary, larger amounts of formic acid 
 tend to retard the precipitation. Much less than 5 cc. of formic 
 acid cannot be used, as there is a tendency to form flakes of a 
 brown precipitate around the kathode, which has to be brought 
 into solution again by the addition of more formic acid. 
 
 Other experiments were run with a view of establishing whether 
 the addition of ammonium formate would have any accelerating 
 or otherwise beneficial effect on the deposit. Such was not found 
 to be the case. When ammonium formate was present in small 
 amounts, no better quality of the deposit could be obtained than 
 could be gotten when it was not present. There was then also 
 no decided effect upon the time of precipitation, the time in all 
 cases being about the same as in the experiments run without 
 sodium formate, or being slightly in favor of the latter. If, 
 however, more ammonium formate was added, the quantity of 
 manganese sulphate and formic acid remaining the same, it was 
 quite apparent that a great retardation in the time of deposition 
 took place and a very decided change for the worse in the char- 
 acter of the deposit. It will be noticed in the following table, for 
 instance, that the complete deposition of the manganese was 
 obtained in one case where 0.1034 of Mn were present as sulphate, 
 and 5 cc. of formic acid, in two and one-half hours. The deposit 
 obtained in this case was very satisfactory. The addition of 2 cc. 
 of ammonium formate to this electrolyte retarded the time of 
 
UNIVERSITY 
 
 OF 
 
 complete deposition to four and one-half hours, and the deposit 
 obtained came off the dish to a considerable extent. The same 
 same thing happened when 4 cc. of ammonium formate were 
 added to an electrolyte, which had given satisfactory results with 
 2 cc. of the salt. The ammonium formate was produced by neu- 
 tralizing the 1.06 specific gravity acid with ammonia. 
 
 Mn 
 present. 
 
 Mn 
 found. 
 
 Formic 
 
 acid 1.06 
 
 sp. gr. 
 
 cc. 
 
 Ammon- 
 ium for- 
 mate, 
 cc. 
 
 Time 
 hours. Temp. 
 
 N.D.joo. 
 
 Volt- 
 age. 
 
 Character 
 of deposit. 
 
 0.1034 
 
 O. IO34 
 
 5 
 
 
 2 1 /, ordinary 
 
 0.8 to 1.0 
 
 7-6 
 
 good 
 
 O.IO34 
 
 O.IO36 
 
 5 
 
 2 
 
 4Vi 
 
 0.8 
 
 5.0 
 
 bad 
 
 O.I5SI 
 
 O.I552 
 
 5 
 
 2 
 
 37* 
 
 0.8 
 
 5-2 
 
 good 
 
 O.I55I 
 
 O.I55I 
 
 5 
 
 2 
 
 12 " 
 
 0.25 
 
 3-0 
 
 good 
 
 O.I55I 
 
 O.I547 
 
 5 
 
 4 
 
 7 
 
 0.8 
 
 4-2 
 
 very bad 
 
 O.I55I 
 
 O.I552 
 
 8 
 
 2 
 
 4 
 
 0.8 
 
 5-o 
 
 good 
 
 O.I55I 
 
 O.I550 
 
 10 
 
 2 
 
 5 
 
 0.8 
 
 4-8 
 
 good 
 
 It was further tried whether the addition of chrome alum, as 
 recommended by Engels, would have a beneficial effect when 
 added to a formic acid electrolyte. The electrolyte in this case 
 contained 10 cc. of manganese sulphate, containing 0.1034 gram 
 of manganese, 5 cc. formic acid, and 1 gram of chrome alum. 
 The current strength was 1.0 ampere with a pressure between 
 the electrodes of 7.0 volts in the beginning to 5.5 volts at the end. 
 The results obtained were, however, very unsatisfactory. The 
 manganese was not all out in seven and three-fourths hours, 
 instead of two and one-half hours without chrome alum. The 
 current was then reduced to 0.08 ampere and the experiment run 
 over night. The weight of the deposit of protosesquioxide of 
 manganese gave 0.1500 of manganese instead of 0.1551 gram. 
 Metallic specks were visible on the kathode and also in the 
 deposit immediately below the kathode, where they had evidently 
 fallen down from the kathode. A test of the deposit on the 
 kathode showed the presence of chromium. Chrome alum as an 
 addition to this electrolyte was therefore shown to be out of the 
 question. 
 
 Recapitulating the results so far obtained, the following con- 
 clusions may be arrived at : 
 
 1. The statement of Kaeppel, that it is not possible to precipitate 
 manganese completely in the presence of formic acid and that 
 more than 5 drops of this acid prevent the deposition of the 
 peroxide of manganese entirely, is not born out by the experi- 
 
14 
 
 mental facts, obtained under the conditions described above. It 
 is evident, therefore, that the conditions under which he per- 
 formed his experiments must have been different, and it is a 
 cause for regret that he did not publish them, as some inter- 
 esting comparisons might then be made. 
 
 2. It is possible to obtain the deposit of peroxide from a 
 formic acid solution in a beautifully adherent and satisfactory 
 form, which can be washed and ignited without any trouble. 
 Even the ignition has no tendency to loosen the deposit. The 
 deposits up to 0.2851 of Mn 3 4 do not show any tendency to scale 
 off. This tendency, however, manifests itself when the man- 
 ganese content of the solution is higher, but even then it is very 
 small. 
 
 3. Formic acid is a better electrolyte than acetic acid, it being 
 only possible, according to Classen's statement cited above, to 
 deposit 0.08 gram of manganese in the form of peroxide. 
 
 As it was thus ascertained that the deposition of manganese 
 could be performed easily and satisfactorily in an electrolyte which 
 contained formic acid, the question arose as to whether it would 
 also be possible to effect the separation of manganese from other 
 metals and especially from iron. The latter had not been satis- 
 factorily accomplished at all up to the present time, as will appear 
 from the following quotations : Neumann 1 says in regard to the 
 separation of iron and manganese: "A great number of experi- 
 ments have been carried out with all forms of salts in order to 
 discover a reliable method for obtaining a complete separation of 
 these two metals, but without success. In most of these experi- 
 ments the aim has been to obtain the manganese as peroxide at 
 the anode or to keep it in solution, while the iron is deposited at 
 the kathode. The results obtained showed that the deposition of 
 the iron was incomplete (at least for the first deposition), and 
 that when the manganese was separated as peroxide, this latter 
 contained iron. 
 
 This difficulty arises in connection with the method proposed by 
 Classen. 2 The solution of the two metals is prepared by treating 
 it with 6 or 8 grams of ammonium oxalate, and after heating to 
 50 or 6o° C. the electrolysis is conducted with a current of 1 
 
 1 " The Theory and Practice of Electrolytic Methods of Analysis," translated by Ker 
 shaw, London, 1898. 
 
 2 Electrolysis. 
 
15 
 
 ampere in density and of 3.1 to 3.8 volts as regards electromotive 
 force. Only a small portion of the manganese is obtained at the 
 anode as peroxide under these conditions. If less ammonium 
 oxalate be used, permanganic acid and its salts will be formed at 
 first at the anode and later a peroxide deposit will be obtained 
 containing iron. As a rule, the liquid is rendered completely 
 turbid by a brown flocculent precipitate, which partly settles in 
 adherent form upon the kathode. The method gives inexact results 
 in spite of all assertions to the contrary. 
 
 The method proposed by Brand, 1 in which a solution containing 
 sodium pyrophosphate and ammonium oxalate and ammonium 
 oxalate is used, also yields inaccurate results. 
 
 If one attempt to effect the separation of iron from manganese 
 in a solution containing 20 to 30 grams of ammonium acetate 
 an incomplete deposition of the manganese as peroxide occurs, 
 owing to the formation of a ferrous salt, which dissolves the per- 
 oxide again at the anode. Engels 2 has proposed to use oxidizing 
 agents in order to overcome this difficulty. If chromic acid be 
 used to oxidize the ferrous salt a complete deposition of the 
 manganese as peroxide can be obtained, but the deposit will be 
 found to contain up to 0.02 gram iron, probably in the form of 
 oxide." 
 
 Classen 3 says: "The hope that manganese in the presence of 
 iron might be separated and determined in an acetic acid solution 
 has not been fulfilled. Innumerable experiments, conducted under 
 the most varied conditions and with the most diverse substances, 
 have given no satisfactory results. In view of the great im- 
 portance which a method for the direct determination of man- 
 ganese in the presence of iron, etc., would possess, this investi- 
 gation will be continued." Nothing further has been published 
 by Classen up to now on this subject, consequently it must not 
 have been possible to make the method work. 
 
 Kaeppel, in his paper repeatedly referred to, publishes also a 
 method for the separation of iron and manganese. His acetone 
 method had failed completely to give satisfactory results. He 
 prepares his electrolyte by bringing a solution of 12 grams of 
 sodium pyrophosphate to boiling, and adding, while stirring, a 
 solution of ferrous ammonium sulphate and manganese ammo- 
 
 1 Ztschr. anal. Chem., a8, 581. 
 J Ztschr. Electrochemie, a, 414. 
 » Electrolysis, p. 149. 
 
i6 
 
 nium sulphate. After the solution has become clear he adds 5 
 drops of phosphoric acid. If a turbidity is caused by the addition 
 of the phosphoric acid the solution is clarified again by the addi- 
 tion of a few drops of sodium pyrophosphate solution. The metal- 
 lic salt solutions contained 0.1 to 0.15 gram of iron and 0.035 to 
 0.11 gram of manganese. The deposition of the iron takes eight 
 and one-half to nine and one-half hours, and it is not possible 
 to determine the manganese electrolytically in the solution after 
 the iron is out. Electrolysis takes place at a temperature of 30 
 to 40 C. 
 
 It was therefore an interesting question whether it was possible 
 to find a solution in which the iron and manganese could be 
 separated. As the manganese peroxide deposits on the anode 
 and the iron on the kathode it should be theoretically possible to 
 obtain both of them at the same time. The beautiful character 
 of the deposits obtained by the use of formic acid gave rise to the 
 hope that it might be easily possible to separate the two metals in 
 such a solution. Before going to this separation, however, the 
 method was first tried on the separation of zinc and manganese. 
 A solution of zinc sulphate, containing a calculated amount of 
 0.1 gram of metal in 10 cc, was used. Electrolysis was carried 
 out in the usual manner, a roughened platinum dish serving as 
 anode, while the platinum basket electrode was connected as 
 kathode. It was found that the deposit of zinc showed an extreme 
 tendency to come down in spots upon the kathode, and the basket 
 electrode had to be carefully adjusted so that it was equidistant 
 from the sides of the dish. It was also found that in this case 
 the addition of ammonium formate seemed to exert beneficial in- 
 fluence on the character of the zinc deposit, the best results being 
 obtained with a solution containing 10 cc. of a manganese sulphate 
 solution with 0.1034 gram of manganese, 10 cc. zinc sulphate 
 solution with 0.1000 gram of zinc, 10 cc. of formic acid of 1.06 
 specific gravity and 5 cc. of a solution of ammonium formate 
 obtained by neutralizing 1.06 specific gravity formic acid by 
 ammonia. The following table shows some of the results obtained : 
 
 Mn 
 
 Mn 
 
 Zn 
 
 Zn Time. Tempera- N.D.ioo 
 
 N.D.jqo 
 kathode. 
 
 Volt- 
 
 present. 
 
 found. 
 
 present. 
 
 found. Hours. ture. anode. 
 
 age. 
 
 O.IO34 
 
 O.IO37 
 
 O.IOOO 
 
 0.0998 11 ordinary 1.0 
 
 1.66 
 
 5-4 
 
 O.IO34 
 
 O.I033 
 
 O.IOOO 
 
 O.OIOOI 11^ " 1.0 
 
 1.66 
 
 5.4 
 
 O.I034 
 
 O.I036 
 
 O.IOOO 
 
 0.01003 IJ " I -° 
 
 1.66 
 
 5-4 
 
*7 
 
 After this question had been satisfactorily solved, the chances 
 seemed to be very promising that iron might be easily separated 
 from manganese in a formic acid solution. But these hopes 
 seemed to be doomed to disappointment, for when a separation 
 was tried under the conditions which had proved most successful 
 for manganese alone, not much of a result was obtained, even if 
 electrolysis was continued for twelve hours. The electrolyte in 
 that case consisted of 10 cc. of a ferric ammonium sulphate solu- 
 tion, containing o.iooo gram of metallic iron, 10 cc. manganese 
 sulphate, containing 0.1034 gram manganese and 5 cc. formic 
 acid. The basket electrode served as kathode and the dish as 
 anode. A long series of experiments were then started in order 
 to ascertain the best conditions for a separation. It was first 
 tried to obtain a better result by varying the amount of formic 
 acid in the solution. It was found that 10 cc. of the acid, with the 
 same amount of iron and manganese present as above, gave the 
 best results, but they were still far from quantitative. In order 
 to ascertain whether the substitution of a ferrous sulphate solution 
 for the ferric ammonium sulphate solution would prove beneficial, 
 electrolytes were tried, containing FeS0 4 , but the results obtained 
 were still less encouraging than those gotten before. It was 
 thought that various current densities at the anode, as well as at 
 the kathode, might have an influence on the result and perhaps 
 bring more success. Accordingly, the current density was varied 
 at the anode from N.D. 100 = 0.15 to N. D. 100 = 2.8. The current 
 density at the kathode was also varied from N.D. 100 = o.i5to47 by 
 using spiral, basket and gauze electrodes, but the result was unsatisr 
 factory. It was also tried to see whether a uniform current 
 density at the electrodes would prove beneficial, and therefore a 
 number of experiments were made by introducing the solution 
 into a beaker glass and suspending into it two platinum gauze 
 electrodes, arranged exactly opposite each other. This was again 
 unsuccessful, for not only was it impossible to get the two metals 
 out completely, but at the edge of the kathode there formed a 
 small deposit, consisting of a brown colored manganese compound, 
 as was ascertained by scraping some of it off and testing it. It 
 was noticed in all of this work that, in order to obtain the iron 
 free from manganese, it was absolutely essential to have the cur- 
 rent density at the kathode entirely uniform, that is to say, all 
 parts of the kathode had to be equidistant from the anode surface. 
 
i8 
 
 If there was one point of the kathode nearer to the anode there 
 was danger of some oxide of manganese depositing there. Per- 
 haps this fact is an explanation why it has not been possible to 
 obtain concordant results with methods proposed up to the present 
 time, when they were worked by different investigators. 
 
 Electrodes of sheet platinum were also tried. The results from 
 all these trials were disappointing, for, though the separation of 
 the two metals could be effected, one metal not containing any of 
 the other, yet it was impossible to get all of the metals out of the 
 solution even in periods varying from eleven to twelve hours. 
 
 As the addition of ammonium formate had proved beneficial in 
 the case of zinc, experiments were also carried out with the addition 
 of varying amounts, from I to 15 cc, obtained by neutralizing 
 formic acid of 1.06 specific gravity with ammonia, to varying 
 amounts of formic acid, from 5 to 20 cc. 1.06 specific gravity. 
 Its presence in small quantity was found to have a beneficial 
 effect, as a greater amount of iron was obtained, but it was not 
 possible to get all the iron out even when the electrolysis was 
 continued for more than twelve hours. Additions of formal- 
 dehyde, in small quantity, were found to retard the precipitation 
 of manganese, though they had no effect upon accelerating the 
 deposition of iron. 
 
 It was then decided to see whether it was not possible to separate 
 the two metals by retarding or preventing the deposition of 
 peroxide of manganese altogether. Accordingly, the use of formic 
 acid was abandoned for the time being, and as experiments had 
 proved that in an electrolyte of tartaric acid the manganese was 
 almost completely held up, it was decided to try this acid in 
 conjunction with formaldehyde, which had also been found to 
 have a retardent effect on the manganese. As ammonium formate 
 had been found to have a beneficial effect on the deposition of 
 the iron, it was used in conjunction with the other two. In order 
 to obtain as high a conductivity of the solution as possible and to 
 get along with as small a voltage as possible and yet obtain the 
 high current strength necessary for precipitating the iron, an 
 addition of 3 grams of ammonium sulphate was made to the 
 electrolyte. This finally led to the desired result. In an electrolyte 
 of this kind a complete and thorough separation of the iron and 
 manganese could be effected. Electrolysis was conducted in the 
 following manner: The electrolyte contained 10 cc. of man- 
 
19 
 
 ganese sulphate solution, containing 0.1034 gram of manganese, 
 and 10 cc. of a ferric ammonium sulphate solution containing 
 0.1003 gram of iron. To this were added 10 cc. of a ammonium 
 formate solution, made as described above, by neutralizing 1.06 
 specific gravity acid by ammonia, and 10 cc. of an ammonium 
 sulphate solution containing 3 grams of this salt. 1 cc. of a solu- 
 tion of 45 per cent, formaldehyde was also added. The platinum 
 dish was used as kathode, and the basket electrode, which had 
 been used before, was introduced as anode. The basket electrode 
 was chosen in order to provide as much anode surface as possible 
 for the deposition of any manganese peroxide which might be de- 
 posited there, and thus assure its adhering to the anode. The 
 solution was then made up to 130 cc, in order to cover as nearly 
 as possible 100 sq. cm. of the dish. Electrolysis was conducted 
 in the cold, with a voltage between electrodes of 3.5 volts and 
 an amperage of 1.4 amperes, the N.D. 100 at the anode being con- 
 sequently 2.33 and at the kathode 1.4. The iron deposited in 
 from three to five hours in a beautiful form, and not a trace of it 
 could be detected in the solution. The deposit, when tested, also 
 proved to be entirely free from manganese. A little bit of peroxide 
 of manganese precipitated on the anode in a beautifully adherent 
 form, but only from 0.0055 to 0.006 gram of peroxide had de- 
 posited at the end of the electrolysis when the iron was all out. 
 No peroxide was swimming in the electrolyte and the solution 
 was colored red from the permanganic acid formed. It was soon 
 found, however, that the deposit of iron could not be weighed, 
 as it contained varying and sometimes considerable amounts of 
 carbon. The odor of hydrocarbons was quite perceptible on dis- 
 solving the deposited metal in dilute sulphuric acid, and a residue 
 was left on the dish after this operation, which could be volatil- 
 ised by igniting the dish. This carbon content proved, however, 
 no obstacle to a successful determination of the iron, as it was 
 easily possible to dissolve the deposit of the dish very quickly by 
 introducing some distilled water and a few cc. of concentrated 
 sulphuric acid. The solution was then transferred to a beaker 
 and boiled for a few minutes in order to get rid of the hydro- 
 carbons. A little granulated zinc was then added in order to re- 
 duce any iron which might have become oxidized, and the solu- 
 tion was then titrated with a very dilute standardized solution of 
 potassium permanganate. The whole manipulation in this way 
 
20 
 
 consumed perhaps less time than the drying and weighing of the 
 dish would have done. The following table gives the results of a 
 number of experiments, run in the way above described. 
 
 The solution, when taken off, had a fishy smell, due to decom- 
 position products of the formaldehyde. The addition of the 
 latter was discontinued after it had been shown by experiments 
 that its presence had no influence on the result. The fishy smell 
 of the solution did not appear any more after formaldehyde was no 
 longer used. Results with and without formaldehyde are given 
 in the above table. 
 
 Though it was shown by these experiments that in an electro- 
 lyte composed as above, a successful separation of the iron and 
 manganese could be accomplished in much less time than by the 
 method of Kaeppel, as cited above, and a method, which would 
 also work in the presence of a large amount of manganese, yet it 
 was felt that only half the problem had been solved in this way. 
 The fact that the manganese could not be determined electro- 
 lytically after the iron was all out, and anyway, the fact that it 
 had to be held up and thus prevented from deposition, were suffi- 
 cient stimulants to renew the effort of finding a method by which 
 the two metals could be determined simultaneously. It was felt 
 that the experiments with formic acid had given considerable en- 
 couragement, and all that was needed was to find some addition 
 to the electrolyte for the purpose of accelerating the deposition. 
 Accordingly, the experiments with formic acid were resumed. 
 
 The electrolyte, which had been found most favorable formerly, 
 contained 10 cc. of manganese sulphate, 10 cc. of ferric ammonium 
 sulphate and 10 cc. formic acid of 1.06 specific gravity. The 
 desired result was first thought might be obtained by the use of 
 an oxidizing agent. As such nitric acid was chosen, and various 
 additions, from 10 drops to 10 cc. of concentrated acid were 
 made. It was found, however, that this had the effect of pre- 
 venting the deposition of iron altogether and that of manganese 
 to a large extent. It was therefore abandoned and the addition 
 of a reducing substance tried. Sodium sulphite was chosen as 
 such, and 34 gram of it added to the solution. This had the effect 
 of bringing out the deposit of peroxide of manganese in shorter 
 time than before, and a very fine deposit of iron, but in six and 
 one-half hours neither all the iron nor all the manganese was out 
 
21 
 
 ppoooooppppJ N 
 
 WWW W HWWWW^OJr 
 
 ppooooopppp » 
 
 (TOOHU K> m h ifi ^ W m • 
 
 o o p p p p p p p O PjJ 
 JsonooooOOOOrtes 
 
 o o o 
 
 o o o o o o o 
 
 2 2 M M M 
 
 B B 
 
 <?p2 
 
 2 2.3- 
 to • 2. 
 
 n 
 
 > 
 
 ?B 
 
 pSS 
 
 *l 
 
 a 
 
 
 5 m 
 
 B «!I 
 
 b n r 
 
 
 ***. .**££*! 
 
 - & 2 
 
 : s : : B 'B 
 
 
22 
 
 of the solution. A larger amount of sodium sulphite was evi- 
 dently needed, but when y^. gram of it was introduced into the 
 solution the salt was itself decomposed under the action of the 
 current, with a liberation of sulphur all through the electrolyte. 
 As it was considered undesirable to have this precipitate of sul- 
 phur in the solution, as it might contaminate the deposits and 
 interfere with the results, it was then tried to add the sodium 
 sulphite at intervals during electrolysis, a / 10 gram at a time. This 
 was carried out and was soon found to be effective in removing 
 the iron. Electrolysis was carried out as usual in the roughened 
 platinum dish with the basket platinum electrode, it having been 
 found by repeated trials that a spiral kathode could not be used, 
 as there was invariably a precipitate of oxide of manganese on it, 
 especially on he outer rim, which was nearest to the anode. 
 This was evidently due to a high current density at the kathode 
 and was in line with observations mentioned above. The elec- 
 trolyte was made up in such a manner that 10 cc. of a solution, 
 containing 1 / 10 gram of sodium sulphite, were introduced into the 
 dish and the ferric ammonium sulphate, 10 cc. containing 0.0997 
 gram of iron, was added. The solution immediately acquired a 
 very dark red color. The usual 10 cc. of MnS0 4 were then added, 
 and after that the formic acid solution. 5 cc. of formic acid of 
 1.06 specific gravity were found to answer well. It was found 
 that the addition of 1 cc. of a sodium sulphite solution, containing 
 V10 gram in 1 cc. for every half hour at the beginning for 2 
 times, and an addition of an equal amount every hour until the 
 iron was out, answered well. A very good indication of the 
 gradual deposition of the iron is furnished by the fact that on 
 each addition of sodium sulphite the color of the solution deepens 
 quite perceptibly, the coloring getting weaker, the less iron is 
 present. When only traces of iron are left in the solution it does 
 not color any more on the addition of sodium sulphite. It has then 
 to be watched and tested for iron from time to time, until no 
 reaction is obtained any more. Care must be taken in this connec- 
 tion that the electrode with the iron deposit on it be taken out as 
 soon as no reaction for iron can be obtained. If that is not done 
 there is danger of some of the iron going into solution again 
 if left too long. The basket electrode is then removed, washed off 
 and an auxiliary spiral electrode inserted as quickly as possible 
 
23 
 
 into the solution in order to prevent the peroxide of manganese, 
 which has precipitated, from dissolving in the electrolyte. Care 
 must be taken during the time that this spiral electrode is in the 
 solution, to insert resistance into the circuit, so as to cut down 
 the current to o.i ampere or less, in order to prevent anything 
 from precipitating on the spiral electrode. If this precaution be 
 not taken there will invariably be a brown-looking deposit on the 
 electrode. The basket electrode with the deposit of iron on it is 
 introduced into a small beaker and covered with distilled water. 
 A few cc. of concentrated sulphuric acid are then added and the 
 solution of the iron will take place rapidly. The basket is then 
 washed off and reintroduced into the electrolyte, after the spiral 
 electrode has been taken out. 5 cc. of formic acid of 1.06 specific 
 gravity is then added, and the current raised to its original value 
 again for the purpose of precipitating the rest of the manganese. 
 A little granulated zinc is then introduced into the iron solution 
 for the purpose of reducing any of the metal which has oxidized 
 during the time that the auxiliary kathode was introduced, and 
 the iron is then titrated in the usual manner. It is not possible 
 to obtain the iron by weighing, for, though a fine looking deposit 
 is obtained, there is invariably some slight oxidation during the 
 time necessary to change electrodes. The results obtained are 
 shown in the following table and are very satisfactory and 
 concordant. 
 
 Fe 
 present. 
 
 Fe 
 
 found. 
 
 Mn 
 present. 
 
 Formic 
 
 acid 1.06. 
 
 cc 
 
 Sodium 
 sulphite. 
 
 Time. 
 
 Temp. 
 
 kathode. 
 
 Volt- 
 age. 
 
 O.IOIO 
 
 0.1008 
 
 0.0988 
 
 5 
 
 6 additions 
 Viogm. each 
 
 4 1 /, 
 
 ordinary 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 0.1008 
 
 0.0988 
 
 5 
 
 <« 
 
 5 
 
 " 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 O.IOIO 
 
 0.0988 
 
 5 
 
 11 
 
 5 
 
 <« 
 
 1.2 
 
 44 
 
 O.IOIO 
 
 0.1008 
 
 0.0988 
 
 5 
 
 «« 
 
 5 
 
 <« 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 O.IOII 
 
 0.0494 
 
 5 
 
 11 
 
 4 1 /. 
 
 «« 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 0.1007 
 
 0.0494 
 
 5 
 
 11 
 
 4Vi 
 
 <( 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 O.IOIO 
 
 0.0494 
 
 5 
 
 << 
 
 5 
 
 (i 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 0.1009 
 
 0.0988 
 
 5 
 
 11 
 
 5 
 
 < t 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 O.IOII 
 
 0.0988 
 
 5 
 
 1 c 
 
 5 
 
 11 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 0.1007 
 
 0.0988 
 
 5 
 
 (< 
 
 5 
 
 11 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 O.IOII 
 
 0.0988 
 
 5 
 
 << 
 
 5 
 
 << 
 
 1.2 
 
 4.4 
 
 It was thus establshed that the separation of iron and man- 
 ganese could also be effected in an electrolyte, consisting mainly 
 of formic acid. It was found, however, that though the sodium 
 
24 
 
 sulphite was very effective as far as the iron was concerned, it 
 acted as a retarding agent upon the manganese, inasmuch as it 
 was found impossible to remove the last traces of manganese in 
 any reasonable time. When only 0.0494 gram of manganese were 
 present, the complete deposition could be effected in about ten to 
 eleven hours, but thirteen or fourteen hours were needed for 
 0.0988 gram of manganese. The precipitate obtained at the end 
 of electrolysis, when all the manganese was deposited, was not 
 adherent, though at the time when the electrode with the iron 
 deposit on it was removed the deposit of manganese peroxide 
 appeared adherent. This was probably due to the action of the 
 large amount of gas set free at the anode, as the electrolysis had 
 to be continued such a long time in order to bring out the last 
 traces of the manganese. However, this fact of the precipitate 
 being loose, was not considered very much of a drawback, as the 
 experiments mentioned in the beginning had shown that the 
 presence of a sodium compound in the electrolyte was not desirable 
 when the precipitate of manganese peroxide was adherent. The 
 precipitate, moreover, was scaly and very heavy, and settled rap- 
 idly down to the bottom of the beaker, to which the siphoned-off 
 solution was transferred. The solution in these, as in all other 
 experiments was replaced by distilled water, while the current 
 was still running, and the current then interrupted. As the 
 precipitate settled down to the bottom, the supernatant liquid 
 could be almost entirely decanted off and the precipitate filtered 
 quickly and washed thoroughly. The washing was performed 
 with hot water and continued for some time, in order to remove 
 any alkali which might be present. The filter was then dried, 
 the precipitate brushed into the dish, the filter burned separately 
 and the ashes added to the material in the dish, after which the 
 dish was ignited in the usual manner. The results obtained were 
 very satisfactory as far as separation is concerned, as is shown by 
 the following table : 
 
25 
 
 ft 
 
 ■o" 
 
 c 
 
 1 
 
 g 
 
 1 
 
 i 
 
 c 
 
 a 
 
 e 
 
 1 
 
 c 
 
 ■0 
 
 I 
 
 
 c 
 
 £ 
 
 s 
 
 1 
 
 do 
 
 8 
 
 s 
 is 
 
 e 
 1 
 
 s 
 
 1 
 
 a 
 8 
 
 (J 
 55 
 
 I 
 
 8 
 
 Q 
 
 
 
 fee 
 a 
 
 > 
 
 
 
 
 
 
 6inst 
 
 ments 
 
 
 ordi- 
 
 
 
 
 O.IOIO 
 
 O.IOII 
 
 0.0494 
 
 0.0499 
 
 5 
 
 V10 
 
 gm 
 
 each. 
 
 10 
 
 nary. 
 
 2.0 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 0.1007 
 
 0.0494 
 
 0.0497 
 
 5 
 
 
 
 1 
 
 n# 
 
 " 
 
 2.0 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 O.IOIO 
 
 0.0494 
 
 0.0497 
 
 5 
 
 
 
 « 
 
 ti 
 
 <« 
 
 2.0 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 0.1008 
 
 0.0988 
 
 0.0990 
 
 5 
 
 
 
 < 
 
 i2# 
 
 " 
 
 2.0 
 
 1.2 
 
 4-4 
 
 O. IOIO 
 
 0.1008 
 
 0.0988 
 
 0.0994 
 
 5 
 
 
 
 1 
 
 U 
 
 " 
 
 2.0 
 
 1.2 
 
 4-4 
 
 O.IOIO 
 
 O.IOIO 
 
 0.0988 
 
 0.0987 
 
 5 
 
 
 
 < 
 
 14 
 
 M 
 
 2.0 
 
 1.2 
 
 4.4 
 
 O.IOIO 
 
 O.IOII 
 
 0.0988 
 
 0.0992 
 
 5 
 
 
 
 ' 
 
 H 
 
 " 
 
 2.0 
 
 1.2 
 
 4-4 
 
 The problem of the separation of the iron and manganese and 
 the electrolytic determination of manganese in the same solution 
 had thus been solved, but unfortunately the time required was very 
 long and probably too long to make the method of practical value. 
 The next experiments were therefore undertaken with a view to 
 reduce the time necessary for the complete precipitation of the 
 manganese, after the iron was out. Various means were adopted 
 to accomplish this, the first being, as would naturally suggest itself 
 from the usual practice, to raise the current in order to get the last 
 traces out. It was soon found, however, that this method did not 
 work in this case, and that there was danger, if the current was 
 raised above 1.4 amperes, that some manganese compound would 
 precipitate on the kathode. This would take place to a consider- 
 able degree if the current was raised to 2 amperes, corresponding 
 to a current density at the kathode of N.D. 100 = 3.3, an at the 
 anode of N.D. 100 = 1.6 to 2 
 
 Varying amounts of ammonium formate were then introduced, 
 but did not prove beneficial. Additions of hydrogen peroxide up 
 to 10 cc. of a 3 per cent, solution also did not improve matters, for 
 while they evidently had some action on the deposit, in making it 
 more adherent, yet the last traces of the manganese did not come 
 out any faster, but rather the reverse. Additions of alcohol in 
 quantities up to 10 cc. and of acetone of 5 and 10 cc. did not have 
 a beneficial effect. Considerable time was thus spent in trying to 
 reduce the time for the precipitation of manganese, but no headway 
 could be made. 
 
 It was then decided to see whether ammonium acetate would 
 prove of value in this connection and a solution was made by 
 
26 
 
 neutralizing ordinary acetic acid of 1.06 specific gravity with 
 ammonia, to a barely acid reaction. An experiment was run in 
 the usual manner with addition of sodium sulphite, until the iron 
 was out and removed. Ten cc. of the ammonium acetate solution 
 were then added and the electrolysis continued. The result was 
 quite surprising, for after two hours the level of the solution was 
 raised and it was observed that no manganese peroxide deposited 
 any more. When the solution was tested, no manganese was 
 found to be present. The time of deposition of the peroxide had 
 thus already been cut down from thirteen or fourteen hours to 
 seven hours. The deposit weighed 0.0992 gram while 0.0988 was 
 present. 
 
 As it was felt that the presence of sodium sulphite in the elec- 
 trolyte was not altogether desirable and that it would be better, in 
 order to remove all possible sources of error, if the separation 
 could be effected without introducing it, the next endeavor was 
 made in the direction of reducing its quantity. Incidentally it was 
 also desirable to cut down the time required for the deposition of 
 the iron and to do away with the necessity of removing the basket 
 electrode, after the iron had all deposited on it, in order to prevent 
 any re-solution. Accordingly ammonium acetate was introduced 
 from the start and the quantity of sodium sulphite was cut down to 
 3 cc. = 0.3 gram and 2 ec. = 0.2 gram. For precaution's sake the 
 basket electrode, with the major part of the iron on it, was also 
 removed, but it was found that 93 per cent, of the iron had already 
 deposited in two hours, and the rest was found when the elec- 
 trode was removed at the end of the electrolysis. The surprising 
 result was found that the iron and manganese were all out of the 
 solution at the end of four hours. The iron was titrated as usual 
 and the manganese peroxide filtered off and the results agreed very 
 closely with the amounts present in the solution. It was then 
 tried to reduce the quantity of sodium sulphite still further and to 
 add only 10 drops, corresponding to only 0.05 gram of the salt. 
 It was found, that the separation could be effected just as satisfac- 
 torily as before, and that the electrode with the deposit of iron on 
 it, did not have to be removed from the solution. The complete 
 deposition of the two metals was again accomplished in four hours. 
 The next step was then to try and get along without the sodium 
 sulphite altogether and it was finally ascertained that if was not 
 
27 
 
 necessary to add any of it. The separation took place just as well 
 in formic acid solution, to which only ammonium acetate had been 
 added. However, as had already been noted in the beginning of 
 these experiments and mentioned there in connection with the use 
 of sodium formate, the sodium salts seem to have an accelerating 
 effect upon the precipitation of the manganese. Just in the same 
 manner as the deposition of peroxide of manganese in formic acid 
 solution took place more quickly when sodium formate was pres- 
 ent, than without it, so in this case the complete precipitation is 
 from one-half to one hour slower without than with, even as little 
 as 10 drops or 0.05 gram of sodium sulphite. Still without its 
 use there is no danger of any deleterious effect due to its presence, 
 and the character of the deposit of peroxide shows a decided im- 
 provement, it being almost entirely adherent. Only a trifle came 
 off during the washing and this small amount could easily be 
 collected on a small filter and added to the deposit in the dish. 
 The final experiments were run in the following manner: 10 cc. 
 of manganese sulphate solution containing 0.0988 gram of manga- 
 nese were introduced into the roughened platinum dish, and 10 cc. 
 of ferric ammonium sulphate solution contaning 0.0996 
 of iron were added. Then 5 cc. of formic acid of 
 1.06 specific gravity and 10 cc. of an ammonium ace- 
 tate solution, obtained by neutralizing 1.06 specific grav- 
 ity acetic acid by ordinary chemically-pure ammonia to barely 
 acid reaction, were added, the basket electrode introduced and a 
 current of 1.1 amperes, with a pressure of 3.9 volts between the 
 electrodes, was allowed to act for five hours. At the end of this 
 time the level of the solution was raised, and no further precipita- 
 tion took place. The solution was then siphoned off as before, 
 both of the electrodes washed carefully and the iron titrated in the 
 usual manner, while the manganese peroxide, after the small 
 amount of loose material had been filtered off and added to the 
 material in the dish, was ignited and weighed in the usual manner. 
 The following table shows the close agreement of the results 
 obtained : 
 
28 
 
 
 
 «j 
 
 
 I 
 
 
 
 1 
 
 
 s 
 
 f 
 
 V 
 
 1 
 I 
 
 ■ 
 
 1 
 
 I 
 
 a 
 | 
 
 i 
 
 a 
 2 
 
 T3 
 
 C 
 
 5 
 I 
 
 c 
 5 
 
 s 
 < 
 
 Formic t 
 
 sp.gr. 
 Time. 
 Hours. 
 
 2 
 
 B 
 
 & 
 
 s 
 
 
 
 | 
 
 * 
 
 
 si 
 
 O.O996 
 
 O.0996 
 
 0.0988 
 
 O.0990 
 
 10 cc. -f 10 
 drops Na^SOs 
 
 5 4 
 
 ordi- 
 nary. 
 
 I.I 
 
 1.83 
 
 3.9 
 
 O.O996 
 
 0.0994 
 
 0.0988 
 
 O.0991 
 
 t< 
 
 5 4 
 
 << 
 
 I.I 
 
 I.83 
 
 3.9 
 
 O.O996 
 
 0.0994 
 
 0.0988 
 
 O.0986 
 
 10 cc, no 
 NajS0 3 . 
 
 5 5 
 
 " 
 
 I.I 
 
 1.83 
 
 3.9 
 
 O.O996 
 
 0.0995 
 
 0.0988 
 
 0.0988 
 
 " 
 
 5 5 
 
 <« 
 
 I.I 
 
 1.83 
 
 3-9 
 
 O.O996 
 
 0.0995 
 
 0.0988 
 
 0.0990 
 
 » 
 
 5 5 
 
 << 
 
 I.I 
 
 1.83 
 
 3-9 
 
 O.O996 
 
 0.0994 
 
 0.0988 
 
 O.0990 
 
 << 
 
 5 5 
 
 ♦ r 
 
 I.I 
 
 I.83 
 
 39 
 
 The deposit of iron has a shining, metallic appearance and the 
 peroxide of manganese is obtained in a black, lustrous form, 
 almost altogether adherent to the dish. 
 
 It is apparent from the preceding description, that the results of 
 this investigation have been very satisfactory. It has been proven, 
 that a complete separation of iron and manganese can be effected 
 in a very simple manner and the fact, that both metals are obtained 
 at the same time, the one as metal on the kathode and the other as 
 peroxide on the anode, constitutes a considerable ad vantage over the 
 other methods proposed up to now and described above. The 
 time of deposition, compared with that necessary in Kaeppel's 
 method, has been reduced from eight and one-half to nine and one- 
 half hours, which are required to precipitate the iron alone by that 
 method, to four or five hours, when both metals are completely 
 precipitated. There is no floating matter of any kind swimming 
 in the dish in which the electrolysis is carried out, and conse- 
 quently no chance for contamination of one or both of the deposits. 
 Furthermore there is no necessity of watching the operation, as 
 owing to the low current used, the electrolyte does not heat up and 
 the voltage and amperage remain constant. Compared with the 
 usual gravimetric methods which demand considerable analytical 
 skill and, especially in the case of the basic acetate separation, con- 
 sume a great deal of time for the filtering and washing of slimy 
 precipitates, the electrolytic method described above has the ad- 
 vantage that it is easy of execution and requires practically no 
 manual labor and attention, after the electrolysis is once started. 
 
 While the work thus far described was in progress, investiga- 
 
29 
 
 tions were also conducted with other organic acids, to find out how 
 their presence in the solution would influence the deposition of 
 peroxide of manganese. The acids tried were propionic, butyric, 
 malonic, succinic, tartaric, citric and fumaric acids. They were 
 used in varying quantities either alone or in conjunction 
 with varying amounts of their ammonium salts. The results, 
 however, were not satisfactory for it was found that nearly all of 
 them retailed or entirely prevented the deposition of peroxide in 
 reasonable time. Citric acid was quite extensively experimented 
 with, in quantities from i gram to 18 grams, as it was hoped that 
 it might lend itself to a separation of manganese from iron. Some 
 good results were obtained by its use, but for some reason, which 
 has not been further investigated, it was impossible to obtain con- 
 cordant results. Sometimes the manganese would be out of the 
 solution in three to four hours, and in experiments run to all 
 intents and purposes in the same manner in order to duplicate the 
 results, it would not be out in seven or eight hours. It is alto- 
 gether probable that the peroxide goes into solution again, when 
 electrolysis is continued too long. Citric acid was therefore finally 
 abandoned. 
 
 The only one of these acids which worked satisfactorily was 
 fumaric acid. The best conditions for electrolysis were found to 
 be the following : 10 cc. of a solution of manganese sulphate were 
 introduced into the roughened platinum dish, the solution was 
 diluted and heated to 50°-6o° C. One gram of fumaric acid was 
 then added, the electrolyte made up to 130 cc, the basket electrode 
 was introduced and a current of 0.6 to 1.1 amperes was allowed to 
 act for three to three and one-half hours. The pressure between 
 the electrodes was 4.2 volts. The above temperature was main- 
 tained during the experiment. The deposit obtained was not en- 
 tirely adherent, about one-sixth of it being loose. The weights of 
 manganese protosesquioxide obtained were 0.1445, 0.1143, 0.1149 
 grams in three experiments, which agrees very closely with the 
 amount of manganese present. 
 
 The above work was undertaken at the suggestion of Professor 
 Edgar F. Smith, to whom the author is under great obligation for 
 many kindnesses. His constant interest in the work, his kind ad- 
 vice and many helpful suggestions were a never-failing source of 
 encouragement and inspiration. It is therefore with great pleasure 
 that the author begs to extend to him his heartiest thanks. 
 
 lABKAj^