EXCHANGE 'AT ' New Results in Electro-Analysis THESIS Presented to the Faculty of the Department of Philosophy of the University of Pennsylvania, in Partial Fulfil- ment of the Requirements for the Degree of Doctor of Philosophy BY THOMAS POTTER McCUTCHEON, JR. PHILADELPHIA, PA. 1907 PHILADELPHIA THE JOHN C. WINSTON CO. 1907 New Results in Electro-Analysis THESIS Presented to the Faculty of the Department of Philosophy of the University of Pennsylvania, in Partial Fulfil- ment of the Requirements for the Degree of Doctor of Philosophy BY THOMAS POTTER McCUTCHEON, JR. PHILADELPHIA, PA. 1907 PHILADELPHIA THE JOHN C. WINSTON CO. 1907 The writer takes this opportunity to express his entire indebtedness to the suggestions, advice and unfailing kind- ness of DR. EDGAR F. SMITH for any merit which this investigation may possess. INTRODUCTION Previous workers in electro-chemical analysis have directed their efforts almost entirely to the determination and separation of metals. A year ago, J. H. Hildebrand,* working in this Laboratory, showed that such anions as Br, I, PO 4 , Fe(CN) 6 , CNS could be determined in the elec- trolytic way. He employed the mercury cathode and an attackable silver anode in a cell especially devised for the purpose. Sodium and potassium can be determined simul- taneously, so that a complete analysis of such salts as sodium chloride and potassium ferrocyanide can now be made with ease and accuracy. The purpose of the first part of this investigation was to extend this work to the estimation of other anions and to learn whether other metals than silver could be advan- tageously used as anodes. In the second part of the work, the same form of cell was used in the separation of certain metals. APPARATUS The decomposition cell, devised by Edgar F. Smith and Hildebrand, consisted of a crystallizing dish, the bottom of which was covered with a layer of mercury. Inside of this was placed a beaker of smaller diameter, without a bottom and supported so that its lower edge extended just below the surface of the mercury. The inner cup was held in position by means of three corks placed radially. The mer- cury was connected with the cathode by means of a stout platinum wire enclosed in a glass tube. The inner compartment contained the solution to be analyzed. Water was placed in the outer compartment, to *Jr. Amer. Chem. Soc., XXIX, 4, 447. (3) 444325 which a few drops of sodium chloride solution WQFC added to increase the conductivity and a nickel wire to aid in the decomposition of the amalgam. The anode consisted of two disks of platinum gauze, placed one above the other and supported by a platinum rod. It was plated with silver by making it the cathode in a bath of silver potassium cyanide. It was then washed, dried and weighed. It was rotated 250-300 times per minute by means of a small motor. The anode was lowered into the inner compartment until the upper gauze was covered by the solution. The procedure for the analysis of such a salt as sodium chloride was as follows: the salt was placed in the inner compartment and diluted to about 50 cc. and the anode placed in position, while in the outer compartment, as stated above, pure water was placed and a little sodium chloride solution so that the current would be conducted more rapidly at first. On electrolyzing the sodium chloride in the inner compart- ment, the sodium passed into the mercury and formed an amalgam. In a short time this amalgam found its way to the outer compartment, where it was decomposed with formation of sodium hydroxide. At the conclusion of the experiment the sodium hydroxide was estimated by titration with standard acid, using methyl orange as indicator. The chlorine appeared as silver chloride at the anode and was perfectly adherent. As nothing remained in the inner com- partment but pure water, the needle of the ammeter gradu- ally fell almost to zero. This indicated the end of the experi- ment and the anode was removed, dried and weighed. The increase in weight represented the chlorine in the sodium chloride. EXPERIMENTAL PART On experimenting with other anions than those men- tioned above, it appeared that the silver anode was unsatis- factory. In some cases, as with a borate and sulphide, the silver salt formed was not perfectly adherent ; silver fluoride, on the other hand, was somewhat soluble in water. The following experiments were instituted, with these facts in mind, in the hope that other anodes than silver might prove helpful. LEAD ANODE The platinum gauzes were plated with lead in a -bath of lead sulpho-cyanide dissolved in potassium hydroxide. In the experiments to be described 4-5 volts were used. This ordinarily gave a current of about 0.5 ampere, falling to 0.02-0.01 ampere when the salt was completely removed from the inner compartment. The anode was rotated 250 times per minute. Potassium cyanide White lead cyanide formed readily, but did not adhere to the anode. B or ax Amperes 0.7-0.03. The lead borate was non-adherent. Potassium fluoride Amperes 0.7-0.01. After many trials it was again found impossible to obtain a satisfactory deposit. A considerable amount of lead peroxide formed on the anode. . (7) Sodium sulphate Some lead sulphate formed on the anode, but the forma- tion of lead peroxide prevented the estimation of SO 4 . By titrating the caustic formed in the outer compartment, it was possible to determine the sodium in the salt with fair results. Sodium sulphate present 0.8520 grams amperes volts time " 0.0276 " o.i-o.oi 5 30 min. found 0.0274 " 0.0274 " 0.0276 " Sodium sulphide The lead anode became black as soon as placed in the solution of sodium sulphide. However, as the electrolysis proceeded, pieces of the lead sulphide became detached, making an exact estimation of the sulphur impossible. A curious phenomenon was noticed during this experiment. The solution of sodium sulphide used was quite colorless. In a few minutes the solution in the inner compartment took on a very pronounced yellow color. The same color was noticed when cadmium and bismuth anodes were used with soluble sulphides. The cause of this color wil shortly receive further investigation. CADMIUM ANODE The anode was plated in a bath of potassium cadmium cyanide. Sodium sulphide The formation of yellow cadmium sulphide became apparent before the current was passed. The solution in the inner compartment became yellow, but after a few minutes became colorless again. The deposit of cadmium sulphide on the anode was adherent if carefullv handled. It was dried in an oven at 115 and weighed. A low voltage should be used and care be taken not to rotate the anode too rapidly ; otherwise small pieces of cadmium sulphide may become detached. Sodium sulphide present 0.0429 grams volts amperes time " 0.0252 " 3.5 0.1-0.03 15 min. found 0.0252 " 0.0256 " 0.0251 " Sodium chromate The cadmium chromate was gelatinous and non-ad- herent. Sodium arsenate White cadmium arsenate formed abundantly, but did not adhere to the anode. BISMUTH ANODE The anode was plated from a solution of bismuth nitrate containing sulphuric acid. Sodium chromate The color of the liquid in the inner compartment altered from yellow to green and finally a green gelatinous precipi- tate separated, probably chromium hydroxide. No bismuth chromate formed on the anode. Sodium sulphide The liquid in the inner compartment again became yellow. When a high pressure was used (14 volts) black bismuth sulphide separated from the liquid. Little bismuth sulphide formed on the anode. 10 Sodium ar senate. No bismuth arsenate appeared, indicating that the bis- muth was not attacked by the arsenate anion. The liquid in the inner compartment became acid to litmus, probably due to the formation of arsenic acid. Sodium iodide The solution in the inner compartment assumed a deep orange color at first. On continuing the electrolysis the color became much paler. The bismuth anode did not change in appearance. ZINC ANODE The anode was plated from a solution of sodium zincate. Sodium phosphate White zinc phosphate formed, but did not adhere. Sodium tungstate When the anode was rotated slowly, the surface of the mercury in the inner compartment took on a blue iridescent tarnish. No hydrogen was evolved in the outer compart- ment and the current did not fall. Rapid rotation of the anode caused a slight evolution of hydrogen in the outer compartment and a non-adherent, white, flocculent precipi- tate, probably zinc tungstate, formed in the inner compart- ment. In the latter case the current fell from o.i-o.oi ampere. Five volts were used in each case. Potassium cyanate A white, non-adherent precipitate formed. II Further, a copper anode was used with a solution of sodium arsenite. The green arsenite of copper was formed readily, but little of it adhered to the anode. When an iron anode was used with potassium ferrocyanide, the anode was not attacked and hydroferrocyanic acid seemed to form in the inner compartment. It is clear, then, that the anodes, described above, behave in several ways toward the different anions. In some cases an insoluble precipitate is formed which may or may not adhere to the anode. In other cases, for example, the iron anode with a soluble ferrocyanide, the anode is not attacked, but the free acid is generated. These experiments, while not at all exhaustive, seem to indicate that silver is the most suitable anode for receiving most anions. It is entirely probable that further study will reveal conditions under which the other anodes will give good results. ELECTROLYSIS OF SOLUTIONS OF METALLIC CHLORIDES WITH A SILVER ANODE. Up to this time only salts of sodium and potassium had been electrolyzed in this apparatus. Attention was next turned to other metals, using their chlorides with a silver anode, to learn whether they formed amalgams and whether the amalgams, if formed, decomposed in the outer compart- ment with formation of hydroxides. A solution of calcium chloride was placed in the inner compartment and electro- lyzed in the usual manner. From the appearance of the surface of the mercury it was evident that an amalgam was formed at first, but in a short time it decomposed in the inner compartment, giving rise to a large quantity of calcium hydroxide. H. S. Lukens, working in this Laboratory at the same time, found that the amalgams of barium and. strontium deported themselves like those of sodium and potassium and decomposed in the outer compartment with 12 formation of hydroxides. He succeeded in making a com- plete analysis of barium chloride, weighing the chlorine collected at the anode and titrating the barium hydroxide formed in the outer compartment in the manner to be de- scribed later. These facts seemed so suggestive that solu- tions of a number of the metallic chlorides were made up and electrolyzed with the following results: Amalgams of lithium, sodium, potassium, calcium (see below), strontium and barium decompose in the outer com- partment with formation of the corresponding hydroxides. Amalgams of cadmium, tin, antimony, iron, aluminium, chromium, manganese, zinc, nickel, cobalt, titanium, vana- dium, zirconium, thorium, lanthanum, cerium, neodymium, praseodymium, magnesium and uranium decompose in the inner compartment with formation of hydroxides. ELECTROLYSIS OF CERIUM CHLORIDE A peculiar result was observed on electrolyzing a solu- tion of cerous chloride. At the beginning- the appearance of the surface of the mercury indicated the formation of an amalgam. A subsequent examination proved conclusively that considerable cerium had passed into the mercury. Later the solution in the inner compartment took on a .pink color, which finally resembled a somewhat dilute solution of potas- sium permanganate. This color was observed by trans- mitted light. It had a greenish, fluorescent appearance by reflected light. The solution was filtered without losing any of its properties. Addition of common salt produced a brownish red precipitate, which differed in appearance both from the hydrated dioxide of cerium, produced by conduct- ing chlorine into an alkaline cerous salt and from the hydrated trioxide, produced by the action of ammonia and 13 hydrogen peroxide on a cerous salt. The precipitate did not dissolve readily in hydrochloric acid, but dissolved in concentrated sulphuric acid with a yellow color. In this latter respect it resembled the dioxide. The same precipitate formed on allowing the purple solution to stand several days or on continuing the electrolysis for about an hour. A pressure of 8 volts seemed to be most favorable for the formation of this compound. It appeared to be a derivative of cerium in a colloidal condition. A further study of this compound will be made immediately. ELECTROLYSIS OF CALCIUM CHLORIDE A more careful study was next made of the behavior of calcium chloride in the cell. Lukens found that the salt with which we had been working was contaminated with considerable amounts of magnesium. The calcium was puri- fied by a number of precipitations as oxalate. On elec- trolyzing the pure chloride, he found that a part of the calcium appeared in the outer compartment like strontium and barium. He was never able to obtain all of it. On mixing magnesium chloride with it again, none of the cal- cium appeared outside. With these facts in mind, Lukens made a mixture of barium, calcium and magnesium chlorides and was able to completely separate the barium from the calcium and magnesium. A determination by the writer gave the following result : Barium chloride 0.1049 grams volts amperes time " present 0.0691 " 3.5 0.5-0.02 2 hours found 0.0692 ' (in presence of calcium and magnesium chlorides) All the previous work had been done with a pressure of 5 volts or under. It seemed probable to the writer that by using a higher voltage; all the calcium could be removed 14 to the outer compartment. The following examples show that such is the case: Calcium chloride 0.0771 grams volts amperes time " present 0.0278 8 0.1-0.02 2 hours " found 0.0272 " 0.0280 0.0278 " 0.0276 " 0.0280 The estimation of the calcium was troublesome at first, as the calcium hydroxide which separated in the outer com- partment was not readily soluble in standard acid of con- venient strength. The difficulty was removed by adding to the solution in the outer part of the cell a slight excess of standard hydrochloric acid at the beginning of the experi- ment. At the conclusion the excess 'of acid was estimated with standard sodium carbonate solution, using methyl orange as indicator. SEPARATION OF CALCIUM FROM MAGNESIUM By using a still higher pressure it was found possible to separate calcium from magnesium, although considerable time was necessary. To remove the last traces of calcium it was found advantageous to add a drop of hydrochloric acid to the solution in the inner compartment from time to time. Magnesium chloride o.iooo grams volts amperes time Calcium chloride 0.0771 " 9 0.3-0.02 3 hours " present 0.0278 " found 0.0282 " 0.0276 " 0.0281 " The calcium hydroxide could be titrated directly in the 6uter compartment as described above, but to obtain accu- rate results it was necessary to stir the mercury a long time 15 with a glass rod tipped with a piece of rubber to completely decompose the amalgam. It was found more convenient to remove the anode, siphon out the liquid in the inner com- partment with .the magnesium hydroxide formed there, and wash the inner compartment thoroughly with pure water. The remaining contents of the cell were then poured into a large beaker. After stirring the mercury well, the titration could be made without difficulty. This procedure was fol- lowed in all subsequent analyses where an insoluble hy- droxide was formed in the inner compartment. SEPARATION OF BARIUM AND CALCIUM FROM MAGNESIUM The thought occurred that by mixing barium, calcium and magnesium chlorides, it would be possible to remove the barium to the outer compartment with a low pressure, estimate it and then remove the calcium by increasing the pressure. Time did not permit of the completion of this work, but a single determination will show that the separa- tion may be realized: Barium chloride 0.1049 grams Calcium " 0.0771 " Magnesium" o.iooo " Barium present 0.0691 found 0.0691 " volts 3.5 Calcium present 0.0278 " found 0.0274 " volts = 9 SEPARATIONS In view of the fact that the amalgams of the metals divide themselves into two classes, some decomposing in the outer and some in the inner compartment, attention was directed to separations. Obviously there is the possibility of separating any metal in the first class from any metal in the second class and in most of the cases tried the separation proved a success. The only aim in this investigation was i6 to discover whether the separations could be made. No effort was made to reduce the time factor. The experiment was usually begun at two o'clock and stopped between five and six. Nor was an effort made to use larger quantities of the metals. This would materially decrease the percentage error. In order to work with larger quantities it may be found advantageous to use anodes with a larger surface. These points will be worked out in detail and the most favorable conditions found for each separation. SEPARATION OF THE ALKALIES AND ALKALINE EARTHS FROM URANIUM These separations are usually troublesome, and it was thought that an electrolytic separation might be useful. SODIUM FROM URANIUM No difficulty was experienced in this or the following separation, although the uranium exercised a retarding influ- ence on the sodium. Silver chloride formed on the anode as usual and the inner compartment became full of yellow uranium hydroxide, which later became black. The sodium hydroxide in the outer compartment was titrated with standard hydrochloric or sulphuric acid. Uranium chloride o.iooo grams Sodium chloride 0.1172 volts amperes time present 0.0461 3-5 0.3-0.02 3 hours found 0.0463 " 0.0459 " 0-0457 POTASSIUM FROM URANIUM Uranium chloride o.iooo grams volts amperes time Potassium chloride 0.1467 " 3-5 0.5-0.01 2 hours present 0.0768 " found 0.0771 " 0.0771 " 0.0766 I? LITHIUM FROM URANIUM As lithium chloride had not been previously analyzed in this cell, a solution was made up and electrolyzed with the following results : Lithium chloride 0.0846 grams volts amperes time " present 0.0140 " 5 0.03-0.01 I hour " found 0.0143 0.0143 0.0144 The separation was made as for sodium and potassium from uranium. Uranium chloride o.iooo grams volts amperes time Lithium chloride 0.0846 " 5 0.03-0.02 2 hours " present 0.0140 found 0.0143 0.0142 " 0.0141 " 0.0141 0.0142 " 0.0143 " BARIUM FROM URANIUM In this separation the addition of a few drops of hydro- chloric acid during the electrolysis was necessary to separate the last traces of barium from the uranium. As under calcium, a slight excess of standard acid was added to the solution in the outer compartment at the beginning of the experiment to prevent tfie formation of insoluble barium hydroxide. The excess of acid was estimated with standard alkali. Uranium chloride o.iooo grams volts amperes time Barium chloride 0.1040 " 5 0-15-0.01 I hour " present 0.0685 " " found 0.0685 " 0.0688 " 0.0682 " 0.0682 i8 STRONTIUM FROM URANIUM. In this separation strontium bromide was used. Uranium chloride o.iooo grams volts amperes time Strontium bromide 0.1456 " 5 0.4-0.02 2 hours present 0.0513 grams found 0.0513 " 0.0513 " 0.0516 " 0.0510 " SEPARATION OF BARIUM FROM THORIUM, CERIUM, LAN- THANUM AND NEODYMIUM The following 1 separations were made to further test the applicability of the method. As before, traces of barium were apt to remain with the hydroxide in the inner compart- ment unless a few drops of hydrochloric acid were added during the electrolysis. BARIUM FROM THORIUM Thorium chloride 0.1300 grams volts amperes time Barium chloride 0.1049 " 5 0.4-0.02 2 hours " present 0.0691 " found 0.0689 " 0.0691 " 0.0689 " BARIUM FROM CERIUM Cerium chloride o.iooo grams volts amperes time Barium chloride 0.1040 " 5 0.4-0.02 2 hours " present 0.0685 " " found 0.0685 " 0.0684 " 0.0686 " 19 BARIUM FROM LANTHANUM Lanthanum chloride 0.0500 grams volts amperes time Barium chloride 0.1049 " 5 0.3-0.01 2 hours " present 0.0691 " " found 0.0693 " 0.0689 " BARIUM FROM NEODYMIUM Neodymium chloride 0.1500 grams volts amperes time Barium chloride 0.1049 5 0.5-0.01 2 hours " present 0.0691 " found 0.0693 " 0.0690 " 0.0693 " THALLIUM The similarity of thallium to the alkalies in many of its behaviors raised the hope that its amalgam would decom- pose in the outer compartment of the cell. Owing to the insolubility of thallous chloride, the sulphate was used in connection with a lead anode. The platinum gauze was plated with lead in a very satisfactory manner by making it the cathode in a bath of hydrofluosilicic acid. The anode was a strip of pure lead. When thallous sulphate was electrolyzed in the cell, the solution in the outer compartment soon gave a very distinct test for thallium with potassium iodide. While no quan- titative results can be given as yet, it is hoped that thallium can be estimated in this way and its separation from other metals accomplished. AMMONIUM CHLORIDE A solution of ammonium chloride was electrolyzed with a silver anode. An amalgam was formed in the inner com- partment, which swelled up enormously. It was interesting 20 to note that the surface of the mercury in the outer compart- ment became covered with bubbles and the solution showed a strong alkaline reaction, indicating that the ammonium amalgam had passed to the outer compartment like the amalgams of sodium and potassium. ELECTROLYSIS OF A MIXTURE OF SODIUM AND POTASSIUM CHLORIDES The following experiments were made to ascertain whether larger amounts of chlorine than formerly used could be estimated with the present form of anode. Hildebrand used a solution of sodium chloride containing 0.0708 grams of chlorine. In these determinations double the amount was estimated with an error of less than 0.0005 gram. The method might be used in the estimation of a mix- ture of sodium and potassium as chlorides. The mixed chlorides could be weighed, dissolved in water and elec- trolyzed. The chlorine could be weighed and the combined sodium and potassium titrated in the outer compartment. The sodium and potassium could be readily calculated from this data. Sodium chloride 0.1166 grams volts amperes time Potassium chloride 0.1487 " 3-S~5 0.5-0.02 45 min. Chlorine present 0.1416 found 0.1412 " o. 1420 " * 0.1418 " 0.1420 " 0.1414 " YD 05100 UNIVERSITY OF CALIFORNIA LIBRARY