ELECTRODE POTENTIAL OF MAGNESIUM BY RALPH KENNY HAMILTON THESIS lor the DEGREE OF BACHELOR OF SCIENCE IN CHEMISTRY COLLEGE OF LIBERAL ARTS AND SCIENCES UNIVERSITY OF ILLINOIS 1921 ■ / 9 2 ./ H I 82 UNIVERSITY OF ILLINOIS Ju rue. _ JL i92_l__ THIS IS TO CERTIFY THAT THE THESIS PREPARED UNDER MY SUPERVISION BY . j&lpJk _ _Lenr.:ey_ _ _ xt cn entitled Q tr od e_ _ _ P o ten_t i &1 _ sl£ _ _M&_gOfis iim. IS APPROVED BY ME AS FULFILLING THIS PART OF THE REQUIREMENTS FOR THE DEGREE OF HEAD OF DEPARTMENT OF Wo . ^ TABLE OF CONTENTS esarzosBEanwE Introduction Part I Electrolytic Potential of Apparatus Historic cil"* . — - - — — * Experimental ---• Discussion of Results Part II Electrolytic Potential of Historical — Experimental--*--- — Discussion of Results ------ Acknowledgement * page 5 Silver " 16 ” 19 Magnesium ------- • " 2 1 ” 24 -------- " 34 36 Digitized by the Internet Archive in 2016 https://archive.org/details/electrodepotentiOOhami ILLUSTRATIONS Figure Figure Figure Figure Figure Figure Figure I Silver Electrode System II Calomel Electrode Vessel III Silver Electrode Vessel IV Magnesium Electrode System V Hydrogen Electrode VI Magnesium Electrode and Ves VII Electrode Potential Curve- - „ INTRODUCTION The experimental determination of the electrolytic pot- ential of the elements is a field in which a relatively small amount of work has been done and one which presents a large opening for future experimental research. On account of the many difficulties in the way of making significant measure- ments, only a very few metals have been covered with any de- gree of certainty. The phenomena of passivity, overvoltage, hydrolysis of the sa,lts, and the extreme chemical activity of the alkalie and alkaline earth metals all enter in until it is little wonder that so much remains undone. Sven in the case of a metal as silver, in which the above mentioned difficulties do not present themselves, there still remains a considerable lack of agreement between the results obtained by different experimentors . This work was carried out with the purpose in view of determining the potential difference between metallic mag- nesium and a solution of magnesium salt normal with respect to magnesium ion. This potential difference had not, as yet, been determined with any degree of accuracy. Preliminary to this the same experiment was run with silver using some variations in experimental procedure which had been successful! u s e d wi t h copper. PART I ELECTROLYTIC POTENTIAL OF SILVER A APPARATUS 7 APPARATUS Two potentiometer systems uning the same working battery, standard cell, and galvanometer were used. The working battery was a two cell lead storage battery of large capacity. It was found that by allowing the battery to remain on dis- charge, a current of one mi Hi amp ere being continuously with- drawn from the lead cell, a remarkably constant working volt- age could be obtained, decreasing only one or two millivolts in twenty-four hours. Further, it was necessary to charge the battery only once in one to two months. The standard cell was a Weston cell of voltage 1 .0136, and the galvanometer a Leeds and Northrup Type K galvanometer. The figure of merit of the galvanometer was .343 X 10"' amperes per milli- meter deflection. The first potentiometer system consisted of a, Leeds and Northrup student’s potentiom ' : .3 ' y >f ±.'J volts and the second a system of resistance boxes giving an accuracy of ±. 1 millivolts. This combination proved to be very satisfactory. By means of the student's potentiometer a reading to ±1 millivolt could be obtained rapidly while with the system of resistance boxes, readings could be obtained to the next decimal place which was in keeping with the con- stancy and reproducibility obtained for the electrode potential of silver. Dial resistance ooxes were found very convenient for the two boxes on which the readings were made. The ac- companying blue print, Figure I, gives a diagramatic sketch of the set up as used. The potentiometer system is the same as is ordinarily 1 ' 9 used. To regulate tine working current to one mi Hi ampere, throw switches one and two, set the potentiometer at 1 .0136 volts and use tapping key (1). Regula.te resistance R ? to zero deflection of the galvanometer. For the unknown electro- motive force, throw switches one, three and nine and use tapping key ( 1 ) . The resistance box system is somewhat more complicated. The working current in this case is one-tenth milliampere. Box R ; has a resistance of from 1 to 9999 ohms and is used to regulate the working current . Box R^ has 11,000 ohms. To regulate the working current, switch five is thrown. Box R^ has a resistance of 9313 ohms, and box R 7 has 10,186 ohms making a total of 19 ,999 ohms. From box R 7 the circuit re- turns to the negative pole of the lead cell. In parallel with box R ? is the standard cell and galvanometer. Contact is made by key (2) . m determining an unknown electromotive force, switch four is thrown. When the unknown electromotive force is between zero and one volt it is placed in parallel with box R s } that is switches seven and eight are thrown and contact made by tapping key (3) . The total resistance in the dial boxes Rj and R^- is kept at 9999 ohms and the electromotive force read from box R^-, one ohm corresponding to one-tenth millivolt. When the unknown electromotive force is between one and two volts, switch six is thrown instead of seven. Figure II is an illustration of the type of calomel half- cell used. In the reservoir (A) is kept a saturated solution of mercurous chloride in tenth normal potassium chloride and 1 1 is used to flush out the arm occasionally to do away with the diffusion of the stronger potassium chloride solution in the salt bridge into the cell. Figure III shows the type of electrode vessel used for silver. The stop-cock is not greased and is kept closed to prevent syphoning over of the solution. 3 HISTORICAL 13 HISTORICAL Although a review of the literature on the electrode potential of silver in various solutions shows that a great deal of work has been done on this subject, still the in- formation at hand from which the electrolytic potential can be calculated is very small indeed. The value found in the older literature of +.771 volts as determined by Willsraore, (Zeitschrift ftir physikalische Chemie, vol. 35 , pp. 2 OI- 332 ) must certainly be in error by as much as .03 volt. Of the three or four papers bearing directly upon the subject, that by G. N. Lewis, (Journal of the American Chemical Society, vol. 23, pp. 166-3) is probably the least in error. His electrode was prepared in the following manner. A platinum wire was sealed into a glass tube and the projecting part wound into a spiral. This spiral was inserted into a tube containing silver oxide and the whole heated at 445 e C. until all the oxide was decomposed. The platinum spiral was now completely enclosed in a loosely cohering mass of finely divided silver. The cell, Pt: Ag: .IN AgN0 3 : .IN KN0 3 : .IN KCl: Hg^Cl, : Hg was found to give more constant results than had ever previously been obtained for silver. This type of electrode is undoubtedly superior to the plated electrode. The electrolytic potential was calculated to be +.302 volts. Lewis and Lacey, (Journal of the American Chemical Soc- iety, vol. 36 , pp. 804-10) working on the copper electrode have found that copper which has been precipitated electro- lytically from tenth normal copper sulfate solution at suf- 14 ficient current density to bring down a spongy, non-adherent mass of metallic copper gave a potential which was far more constant than could be obtained by copper plating. A copper plated platinum wire was used to make contact with the spongy copper . In making this review of the literature on electrode pot- entials, No. 5 (1911) and No. 3 (1915) of Abhandlung der deutschen Bunsen Gesellschaf t , on Measurement of Electromotive Forces of Galvanic Cells, compiled, by Abegg, Auerback and Leuther were found to be of inestimable service. c EXPERIMENTAL 16 EXPERIMENTAL Since the potential of even the most promising looking silver-plated electrodes show considerable variation, it was decided to prepare a silver electrode analogous to Lewis' copper electrode. Tenth normal silver nitrate was electro- lysed at sufficient current density to bring down a dark, very finely divided, spongy precipitate of metallic silver. The silver was washed with tenth normal silver nitrate sol- ution and allowed to stand for some time in tenth normal silver nitrate before reading the potential. Contact was made by means of a very bright silver-plated electrode. The cell, Ag: .IN AgNCL, : KN0 3 sat. sol: .1 N. E. gave the following readings: Volts Time (min.) •3973 0 .3977 5 .3979 10 .3973 65 .3973 75 This cell remained constant for several days. However, readings at other dilutions did not give the desired results. Further, it was found that a silver-plated platinum wire gave the same reading in a tenth normal silver nitrate solution whether in contact with the finely divided silver or not. The reading is this case was higher (.4002). Also, a smooth platinum wire placed in a tenth normal silver nitrate solution with finely divided silver gave no reading at all, showing that no contact was made between the platinum and the finely • I) 17 divided silver. In view of this fact it was decided that black amorphous silver obtained at high current density was of no value in determining the electrode potential of silver and that a smooth platinum wire instead of one plated as used by Lewis should be used in subsequent experiments since the plated wire gives a more positive potential. When silver is deposited at a very low current density on a platinum dish small white crystals of silver form. These small crystals in contact with a platinum w ire did not give satisfactory results. At a slightly higher current density, long sharp needles of crystalline silver form in clusters. It was this silver that gave all that could be desired in the way of constancy and reproduciabili ty . Working with freshly prepared solutions of silver nitrate, the following cells were set up and their electromotive force measured. The long needle-like crystals of bright silver mentioned above, in contact with a smooth platinum wire was used in making these measurements. Two tenth normal calomel electrodes were prepared and their potential found t,o be exactly the same. The potentials of the silver electrodes were abso- lutely constant, varying at no time by more than one tenth milli- volt from the value given. All readings were taken at 22° G . Ft: Cry st. Ag: AgN0 3 .001N: KtJ0 3 sat : . 1 N. E.--- .2350 If it rt ii .01N ft ft " -—.3415 If n it ti . IN ft II " --.3960 ■ ■ . 4 D DISCUSSION OF RESULTS • •• • 19 DISCUSSION OF RESULTS The potential of the tenth normal calomel electrode was taken as +-. 336 O volts on the hydrogen scale. With the aid of conductivity data taken from Landolt and Bornstein, the following table was constructed. 1 2 3 4 5 6 N Eq .Cond V.obs. E . P . Exp , . E.P.Calc. 0.000 1 15.3 100.0 0.001 113.15 97.7 .2350 .6210 .6203 0.01 106.31 93.1 .3415 .6775 .6775 0. 1 94.33 81.5 .3960 .7320 .7322 The e lectrode potentials calculated in the sixth column made assuming the value for one- hundreth normal to be correct, and with the aid of the Nernst formula: E = .0530 log c c' where E is the difference in electrode potential and c and c' are the concentrations of the silver ion. The potential in one-hundreth normal was taken as probably most nearly cor- rect because at greater dilution there is so much resistance in the cell that the galvanometer does not have the desired sensitivity and at greater concentration the thermodynamic environment digresses appreciably from that of pure water, so that the laws of ideal solutions would not be obeyed. The electrolytic potential calculated from the value obtained at one-hundreth normal is found to be +.7953 volts. This figure shows agreement with that obtained by Lewis. A comparison of values observed in column five and those . 20 calculated in column six shows an agreement more close than any obtained in any previously published work, and are, in fact, as close as could be expected, taking into consideration the possible error in alpha as obtained from the conductivity data. The value •+'.7953 volts is, therefore, probably the best value yet obtained for the electrolytic potential of silver. part ii ELECTROLYTIC POTENTIAL OF MAGNESIUM A HISTORICAL 21 HISTORICAL The electrode potential of magnesium has never been det- ermined. with any decree of accuracy on account of the form- ation of basic ss.lt s when metallic magnesium is placed in contact with a magnesium salt solution, (Kahlenberg, Action of Metallic Magnesium, Journal of the American Chemical Soc- iety, vol. 25, P* 3^0.) Magnesium attacks water slowly, hydrogen being evolved, but in the above case the action is much more rapid. The decomposition is probably according to the following equation: Mg + MgCl, + HOH—} 2 Mg ( OH) Cl -R H^ It is worthy of note at this point that magnesium amalgams attack ma,gnesium salt solutions much more rapidly than does the pure magnesium. M Le Blanc, (Zeitschrift fiir physikal- ische Chernie, vol. 5 , p. 467) measured the voltage of the cell: Mg )t (Hg): MgCl^: Zn x (Hg) obtaining a fairly constant electromotive force of 1.08 volts, but it was our experience, as will be shown later, that such a cell would give a rapidly decreasing electromotive force. Neuman, (Zeitscinrif t fttr physikali sc he Chemie, vol. 14, p. 215 ) states that solid magnesium amalgam decomposes a mag- nesium salt solution very rapidly and is quickly oxidized by the air. The potential drop between magnesium and magnesium sulfate was found to decrease with the time. As a mean value he gives the figure -1.239 volts. Comparisons with data ob- tained in this laboratory will be drawn later. Kahlenberg, in his article previously referred to, gives - I .366 volts as the potential of magnesium in magnesium sul- 22 fate. This value was found, to increase with, the time. Bab- orovsky, (Zeitschrift fttr Electrochemie , vol. 11, p. 465) found -1.55 volts to be the potential of magnesium in magnesium sul- fate and that the potential of magnesium amalgam changed very rapidly with the time. The work of Kistiakowsky, (Zeitschrift fiir Electrochemie, vol. 14, p. 113) is generally accepted as being the most sig- nificant. He measured the potential difference between a rotating magnesium electrode, which had the protruding part covered with paraffin, and a normal magnesium sulfate solution in an atmosphere of hydrogen. He took the view that it was necessary to rotate the electrode so that a homogeneous sol- ution will be maintained and that the presence of oxygen ef- fected the potential of magnesium. The voltage of tne cell: Mg: MgSO^: 3N KCl: .1 N. E. was found to decrease with the time, the highest value found being 1 .924 volts, which gradually decreased to 1 .790 volts at the end of an hour and a half. He gives the highest value found as the most nearly correct. A calculation of the single potential of magnesium by Kelvin's rule gives 2.54 volts. Although this calculation is based on an erroneous assumption it does give some basis for comparison. Thus, the individual experimentors disagree between themselves by as much as three tenths of a volt and on comparing experimental values with the above calculated value, one would say that probably all of them are far from the correct one. In view of this state of disagreement it was decided to investigate the subject. B EXPERIMENTAL 24 EXPERIMENTAL When experimental difficulties in determining the elec- trode potential of an element are involved, as is the case with magnesium, indirect methods have been employed making use of organic solvents, (Marquis, Electrode Potential of Arsenic, Journal of the American Chemical Society, vol. 42, p. 1 5 69 and Lewis, Electrode Potential of Sodium, Journal of the American Chemical Society, vol. 32 , p. 1459.) It is quite evident that such a solvent, in addition to dissolving a mag- nesium salt, must cause it to ionize appreciably into the pos- itively charged magnesium cation, and must not form addition compounds; nor can any phenomena similar to hydrolysis be pre- sent. On recalling the G-rignard reaction one 'would say offhand that a suitable solvent would be hard to find and such was found to be the case. Attempts were made to deposite magnesium electrolyticaliy from solutions of magnesium chloride, magnesium sulfate, or magnesium bromide in acetone, methyl alcohol, or pyridine. In any case, magnesium was found to go into sol- ution at the anode but only a gas could be obtained at the cathode, as is the case in aqueous solutions. The possibility of using an organic solvent was then discarded. Kahlenberg, in his investigations of the action of met- allic magnesium, states that alkalie solution are not attacked by metallic magnesium, the metall remaining bright for several hours. At Dr. Reedy's suggestion the following cell was set up: Mg: Mg(OH) ' .IN KOH: 3N KC1: .1 N . E. Theoretically this seemed quite promising for, even if decomp- osition did take place, all that could occur would be the for- . . 25 ation of more of the hydroxide. It was found that although metallic magnesium does not decompose potassium hydroxide sol- ution, hydrogen is evolved when magnesium hydroxide is present in the solution. Further, a constant electromotive force could not be obtained. The average was found to be -I.O3 volts for the potential of the magnesium half-cell, which is about the same as has already been obtained for magnesium in potassium hydroxide solution and to which no significance can be attached. Lewis, (Journal of the American Chemical Society, vol . 32, p. 1459) has obtained excellent results for the alkalie metals by making use of dilute amalgams. As was previously mentioned, M Le Blanc and Neuman did not agree on the matter of the electromotive force of magnesium amalgams, so it was decided to make an investigation. An amalgam was made up containing about . 5 $ of magnesium, but it was rapidly oxidized by the air forming a solid. A more dilute amalgam was made up then ( .01$) and it too was rapidly oxidized. When placed in a magnesium chloride solution a large black precipitate was found to be instantly formed. An amalgam containing .001$ magnesium gave this same black precipitate when in contact with magnesium chloride but not so rapidly. Its potential was measured and found to change rapidly with the time as is shown by the following taible . 26 Time Potential 1 min . -1 .298 volt; 4 " -1.332 7 " -1 .329 16 " -1.297 33 ” - 1 .220 " 56 " -0.364 With, this the possibility of making use of a magnesium amalgam was given up. A rotating magnesium electrode similar to that used by Kistiakowsky was next assembled with the idea in mind of as- certaining whether Kistiakowsky ’ s results were reproducible or not. The metal used was in a circular stick about one- half inch in diameter and. three inches long. Several meas- urements were made in an atmosphere of hydrogen but the result v/ere the same in either case. The following tables give a comparison between the best set of values obtained and those given by Kistiakowsky. The cell measured in either case is, Mg: MgSO : 3N KC1: .1 N. E. the only difference being that we used tenth normal magnesium sulfate where Kistiakowsky used normal. Our voltage should therefore be about .025 volts higher than his. An attempt to extrapolate our readings back to zero time was a failure because it was impossible to connect the points by any kind of a curve. However, the two sets of readings do agree that the potential of the magnesium electrode de- creases with the time. This is caused by the formation of II 27 Kistiakowsky This laboratory Voltage Time Voltage Time 1 .924 0 1 .379 0:34 1 .922 1 1 . 354 1:13 1 .916 5 1 .312 1 : 56 1 .900 9 1 .793 2:18 1 .900 13 1 .735 2:34 1 .376 32 1 .777 2:56 1 .346 77 1 .753 3:38 1 .790 37 1.631 123 : 00 the basic salt which appears on the electrode immediately on immersing it in the solution. This salt will dissolve in hydrochloric acid. The electrode potential seemed to reach a minimum at about the lowest value given, this minimum varying somewhat with the salt used. It is about the same as that obtained by Kahlenberg but could hardly be accepted as the true potential difference between magnesium and mag- nesium salt solution. Since this basic salt is soluble in an acid, the possi- bility of measuring the electrode potential in acid solution suggested itself. Although the acid would dissolve the mag- nesium, it was thought that by mixing half and half solution of magnesium sulfate and sulfuric acid of the same strength and obtaining the initial reading, a value to which some sig- nificance could be attaches might be obtained. However, the electrode potential was found to change very rapidly with the time and it was impossible to take a reading quickly enough. ... ,« 28 The final method used, which after taking a good many readings with different alterations gave some degree of success, was to measure the electrode potential in acid solution and simultaneously follow the hydrogen ion concentration by means of a hydrogen electrode. The first theory was to start out with a tenth normal solution of sulfuric acid and let it react with the magnesium until completion, when the solution will be tenth normal mag- nesium sulfate. The end point was to be determined by the hydrogen electrode which would show a potential of .406 volts. Theoretically, the basic salt should not commence to form until all the acid was gone but actually it did not wait that long so that this method had to be modified. When the solution is tenth normal acid, the action of the acid is strong enough to prevent the formation of the basic salt, so it was decided to start out with two tenths normal sulfuric acid and let the reaction run past the half way mark, again making use of the hydrogen electrode. Figure IV shows diagrammatic ally the simple arrangement make use of. To bal- ance the system, throw switch one; to measure the hydrogen electrode, switches two and three; the magnesium electrode, switches two and four. Figure V shows the kind of hydrogen electrode used, and Figure VI the electrode vessel. The voltage of the hydrogen-calomel cell at that point where the two tenths normal sulfuric acid has been changed to a mixture of tenth normal magnesium sulfate and tenth normal sulfuric acid was previously found to be .4075. On account of 31 it being impossible to take both readings at tbe same time, it was necessary to take readings alternately and plot them against time on tbe same paper. Obviously, tbe potential of tbe mag- nesium electrode at that time when tbe potential of tbe hydro- gen electrode is .4075 against tbe calomel is tbe one sought. A good many readings were taken, four sets of wnicb are given in tbe following table. Table I Table II Mg Time H Mg Time H 2.036 0:00 2.090 0:00 0:42 .399 0:40 .400 2.081 1:37 2.036 1 : 10 2:20 .4077 1 : 40 .405 2.069 3:15 2.030 2 : 15 4: 10 .4137 2:53 .4125 2.053 4:40 2.067 3:40 Table III laole xV Mg Time H Mg Time H 2.039 0:00 2.0345 0:00 0 : 43 .397 0:55 .397 2 . 035 1 :25 2 . 034 1:18 1:53 .4032 1 :53 .401 2.079 2:34 2.079 2:33 3:09 .41 1 04 O O .4065 2.0665 3:48 2.073 303 4:00 .41 15 c DISCUSSION OP RESULTS 4 34 DISCUSSION OF RESULTS Two graphs for each set of data were drawn on a single sheet, one of the hydrogen electrode against time and the other of the magnesium electrode against time. From this the pot- ential of the magnesium electrode at the same time that the potential of the hydrogen electrode showed that one half of the acid had been used up could be determined. Figure VII was drawn from the data given on Table III, which is about an average of all four. The values found in each case were: Table I 2.0765 " II 2.0307 " III 2.0733 " iv 2.0730 Average 2 . 077 This gives -1.741 volts as the potential of magnesium against the solution used. The solution is a tenth normal solution of magnesium sul- fate and tenth normal sulfuric acid. Tenth normal magnesium sulfate i 3 44.1^ dissociated and tenth normal sulfuric acid dissociated so that it contains .0625 moles hydrogen ion, . O 315 moles bisulfate ion, and *0155 moles sulfate ion per liter. Applying the isohydric principle, the concentration of magnesium ion is found to be .0258 normal. By means of the Nernst equation, the electrolytic potential is calculated to be - 1.700 volts. On account of the difficulty in determining exactly the concentration of magnesium ion, this figure may be in error by as much as ±*015 volts. . . * ACKNOWLEDGEMENT 36 ACKNOWLEDGEMENT The author wishes to take advantage of this opportunity to express his appreciation to Dr. Dietrichson, under whose direction this work has been carried out. To his many help- ful suggestions and frequent visits to the laboratory, which have caused this work to be a sourse of much personal pleasure and satisfaction, may be attributed the success attained in this investigation.