5 E 2 THE EFFECTS or ALUMINUM AND mow i ON ELECTRODEPOSITED NICKEL Thai: for tharDegroe'of Ph. D. MICHIGAN STATE vumvmsm Walter 0. Dow, Jr, 1955 “31‘! Hf"?! “ ‘ THE-"- 5!: This is to certify that the thesis entitled "The Effect of Aluminum.and Iron on Electrodeposited Nickel" presented by 'Walter 0. Dow, Jr. has been accepted towards fulfillment of the requirements for .DQnLnL's. degree in m1 cal Eng 0 . / I’ll/[riff I} V/é, ///\ [/V hfajor professor Date November 18, 1955 0-169 THE EFFECTS OF ADUMINUM AND IRON ON ELECTRODEPOSITED NICKEL By . \\e. \. ‘lhlter of'm, Jr. AN ABSTRACT Submitted to the School of Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Metallurgical Engineering Year 1955 Approved: ABSTRACT The effects of alunimn on the appearance, adhesion, ductility, hardness, ttn‘ouing power, and salt spray corrosion resistance of various types of nickel deposits were studied. The effects of iron on the same physical properties of nickel deposits were reviewed and correlated with those of aluminum. Believing these effects due to colloidal hydroxides of iron and alminmn , the author has included a review of the formation , coagulation, structure, and particle size of these sols in various solu- tions. This discussion also contains intonation concerning double twdroxide formation, effects of various ions on sol fornation, and the adsorption by. colloidal particles . The removal of almsinun was studied both electrolytically and by high pH precipitation methods. The removal of iron was more extensively studied because of the two vslences which exist in nickel solutions and because actual deposition of iron occurs in such solutions. THE EFFECTS OF ALUMINUM AND IRON ON ELECTRODEPOSITED NICKEL By . \"I a \o‘JI‘ l; waiter 0; Dow, Jr. A THESIS Submitted to the School of Graduate Studies of Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Metallurgical Engineering 1955 THESIS TABLE OF CONTDITS mmmCTIwOO0.00000000000000...0.0.0.000...0.000....0.0.0.0000... EFFETS 0F AIUMIWH ON NICKEL DEPOSUS............................. PmodureeseoeeeoseesseseesOoesseesoesessseeeeeeeeeeeeeeeeesee APpoar‘nceooeesoeeeeseeeoeseeeeeeeoseeeeeeesessseeoeeeeeessee. Ad11831°n0000soes0cease-oes-seeeoeseseeseeeoeeoeeeseeoeeseeesso mctflityosssooeoeeeseoeeaesooeeeeoeeseesssosoeosesOOeease-see amn°88000000000.0000000......00.0.0.0...00.00.000.000000000. vaing PMPOOOOOOOOO.OIOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO mt wray corm1°DOOOOOOOOICO ...... OOOOOOOOOOO‘0.00.00.00.00 norms armors on NICKEL DEPOSITS ..... couonm. paopmms or mos m AIDMINUM mam Fomtion...OOOCCCOOOOOOOOOOOOO0.00.00.00.0000.0.00...0.0.0... fimct‘mOOOOOOOOO00000000.00000.0.0000.OOOOOOOOOOOOOOOOOOsees. Adsorption by 3018............................................ REJOVAL 0? W FROM NICKEL WTIWS.......................... WALOPIRONFRQIINICKKL SODJTIONS.... ...... .................... CONCIUSIONS........................................................ WES......................................................... '1! a9; 0 H tosowocoow \n H “HEW I would like to acknowledge the help and astute guidance of the late Dr. D. T. Ewing under whose supervision a great deal of this work was completed. I an very grateful to Dr. A. J. Saith who conpleted my guidance with such interest and understanding. To my wife, Sally, who has borne the brunt of late hours, examination nerves and too little of the good things in life for these years, my heartfelt thanks. The large number of analyses done by both Chrysler and Sunbeam Corporation is greatly appreciated. DITROIXJCTION In the past two decades the natal finishing industry has made tremendous advancement in the use of science and technology. The I'art" of fomulating and raising paints has given way to fundamental and developmental research with the result that for the first tine in centuries wholly new paints have been compounded. Rubber base enulsion paints and the new silicone and open resin paints are samples. The ”art“ of polishing and buffing is developing into a science with de- tailed studies being node of polishing and buffing ccnpo'omds, surface speeds, types of wheels and belts, contact pressure and abrasive coupo- sition. The 'art" of electrcdeposition is an infant as compared to other types of finishing, but its scientific growth has outstripped any other finishing field. Much of this growth has occurred naturally with the general increase in technology, but also the tremendous complexity of plating solutions and electrode reactions forces an interest in research and developent out of proportion with the rest of the industry. In the beginning nore experienced electroplaters developed pro- prietary solutions for cleaning and plating. In order to be competitive with such solutions they also had to service then when sonething went aniss. This extra service called for a nuim of knowledge about solution characteristics . Thus an econsnic urge added incentive to research am developent. So it is that in the space of twenty years electrodeposition has progressed from testing a solution for pH detox-nin- ation to electronic pH meters; from observation of chrome plated articles for 'nisses' and "rainbow" to determine whether or not to add sulphuric acid to the analytical determination of the ratio of clu‘cnic acid , to acid radical by chenical neans. The scope of research and developnent in electroplating is wide. It includes netal cleaning, netal polishing and buffing, netallurgy of base netal an! electrodeposit, constituents in plating solutions , effect of inpuritiee in plating solutions , ad- dition agents, electrode behavior, current distribution, and easy rnifications of each . Recent researches in cleaning techniques now allow the scientific conpounding of electrolytic cleaners and provide means for testing then as to their ability to prepare a surface for the deposition of metal. Not only the cleanliness but the kind of naterial to be plated affects the character of the deposit. Studies have been and are being nade on the effects of the base netal on the physical properties of the electrodepoeit. i great deal of dovelopsent and fundanental research is being con- ducted to deternine the effects of variables in the constituents of plating solutions. The use of addition agents to enhance brightness, to reneve pitting, to cause either tensile or compressive stresses and to allow higher current densities have been assiduously studied by the industry. Research foundations and educational institutions have contributed a great easy of the basic concepts of electrodeposition. Such fields as polarisation, evervoltage , and nechsnisn of deposition have been done for the nest part by groups not directly connected with industry. Boiler” , industry has contributed singly and collectively to fundamental research through grants and fellowships, and it was in this nanner that the present work on iron and aluminum was made possible . Probably nore iron and steel have been electroplated than any other netal, and certainly iron is one of the nest cos-on inpurities in an electroplating solution. Therefore, it is of prime inportance to know the effects of even small counts of iron to deternins how nuch can be tolerated and even more imortant how to renove it. filing g; 5.1.. (l) have detemined the effects of ferrous ion in pure nichel plating solutions, but the novel studies were linited and the snall anounts of ferric ion were igmred. To round out this study of iron as an inpurity is one of the objects of this work. iluninn had as yet not been studied in the systenatio nanner of iron, and since its colloidal properties and high pH precipitation were sinilar to iron in most respects, it was decided to study the effects on electrodeposits and sons effects of its renoval fr- solution to see if sons correlation could be effected. The plating of notsls on alt-inn has developed into an important phase of industrial electroplating and with the possibility of increased contanination it is necessary to know what effect various concentrations night have and how they can nest easily be renoved. The experinents on the effects of aluninun will be presented first and the statutord tests are described for appearance, adhesion, throining power, hardness, ductility, and salt spray corrosion resistance. The effects of iron, which were studied in the sane nanner as alunimn, will then be discussed briefly. It is the thought of the author that the physical state of the iron and aluninun in plating solutions is the principal cause of changes in plvsical properties of deposits fren solutions centaninated with these ions. Therefore, a resume of the fornation, stntctutre , and special effects of aloninun and iron hydrox- ides not only in nickel solutions but in other aqueous solutions is included. By focusing attention on these colloidal properties it is hoped that other investigations will deternine how various structm'os , sizes of particles, age of sol, and so forth, will. affect electro- depesits. The novel of alt-inun was chilled both electrolytically all! by high pH precipitation . The equilibrium concentration at a given p11 as well as the conplete independence of rates of electrolysis are discussed. Finally the renoval of iron is discussed with respect to its two valent- state in solution and the electrolytic and high pH renoval of both ferrous and ferric ions. In recent years a great deal of interest has been shown in deposition of nichel and chroniun on aluninma. Te inprove adhesion, a sincate i-euion treatnent is co-only used to renove the tenacious alt-icon oxide filn. Thus aluninun can be carried into the nickel plating solution as the aluninato or can be dissolved fron the aluninun parts if electrical contact has been broken. Although several investi- gations of ease of the effects of slain on the deposition of nickel were published prior to 1916, it was decided to establish these effects in extrenely pure solutions (2). Watts reported that the hexalvdrate of aluninun chloride increased the throwing power of the solution trenendously while Barr reported that altmimn added as the sulptnte had a deleterious effect of the netal distribution in both high and low pa solutions (3,10 . However, Watts had added citric acid to keep a snall anount of aluninun in solution as a couple: ion while Barr had used as nuch as 120 grans per liter of alanimin sulphate, nest of which becane alt-inn hydroxide. Rarr and O'Sullivan obtained dark deposits which they attributed to occlusion or co-deposition of aluinun hydroxide, and O'Sullivan found traces of alunimn in these dark deposits which were not present in nernal greydeposits (LS). o. the other hand, several Mimexperinsntnrs reported increased. lustre and bright- nessd'ue totheadditianofanalunimhydrondesoloraluinu sulphate in .all anounts_(6,7) . All agreed that alunimn increased the stress of tin deponit_(6,8l. Host investigators were of the opinion that aluninm hydronde in the fem of colloidal particles rather than the aluninun ion in solution was the cause of the changes in the deposits. Io original references to alnninun as an inpuritw or addition agent appear to have been published since 19156. Effect of unlim- on Electredepositdon of Nickel Procedure -- he leddficatiens in the preparation and testing of the deposits fre- those set forth in an earlier publication were advis- able . (1) For better utilisation of the supply of low carbon rolled strip steel, the vertical section of the bent cathode was reduced free 2' x 3 l/Z' to 2" x 3". In order to obtain the greatest accuracy in the throwing power nemnts, both the top and botto- of the lip of the bent cathode sore considered. Briefly, the lips of the 0.002. panels with the pure deposits vere observed on a research netallegraph ecuippod with a Filer calibrated eyepiece. The distance fro- the loading edge of the lip to a point where the thickness of the pure deposit equaled 0.00? was accepted as a standard for that particular solution. The sane distance see then neasured on the deposits fro- the ilpnre solu- tions, ani the thickness at that point recorded. If it were 0.002' there had been no change in throwing poser. If it were less than 0.002' there had been a decrease in throdng poser, and if it were greater than 0.002“ there had been an increase in throwing power. It is apparent thatifthepenelwarpedinsuchasayastoraise theleadingedge of the lip even a snall mount, the top of the lip sould receive a dispro- portionallx .allaneunt of nickel due to a shielding effect, an! if only the top deposit were measured, a false decrease in throwing power would be reported. is a basis for detenining changes in physical properties, deposits were prepared fren pure solutions of four basic for-ulna i grey deposits were obtained free Watts' 2.2 and 5.2 pH solutions; bright deposits were obtained free a nickel-cobalt alloy type and fro- an organic type solution using nickel bensene disulphonate and tri- aninetolyldiphenylnethans (l) . Heasnrenents of these panels from pure solutions cenpared to these fre- solutiens containing varying anmnts of alueinun were used to calculate the percent change in plysical properties. The appearance of the grey deposits was reported as nusbers or half embers on the Eastman Grey Scale. The nickel-cobalt and organic bright deposits were arbitrarily Judged as 'Hirror Bright', 'Bright', “Sod-Bright', or 'Dull". The results of the ductility and adhesion tests were reported as trends rather tun unerical percentages. The results of the salt spray corrosion experiments were graphed as percent change from pure deposits. Considering the questionable status of this accelerated test it is advisable to consider the results as tendencies rather than as absolute values. Although it was dotsrnined fro- electrelytic and high pH precipita- tion renewal data that definite linits- of solubility of aluimsl existed in each of the four solutions, additions of 10, 20, 30, SO, and 1(1) ng./l. were nade in order to study the effects of dispersed precipitates of aluninun Indroxide as well as alminun in solution. Sodiun lanryl sulphate was used in amounts which Just prevented pitting of the face of the panel. Appearance - The effects on the appearance of deposits obtained free nickel solutions containing various nounts of aluninun as an inpurity were evaluated by observing the face of the panels under a diffuse light. The position of the illulinated panel under the light source and the line of sight of the observer were kept identical for all panels when it was found that differences of two units on the East-an Grey Scale could be detected simly by reversing a panel of the Vatts' type deposit. lie change occurred over the range of concentrations studied in the appearance of deposits fr. the Watts' 2.2 pH, the nickel- cobalt and the organic. type solutions . However , the hatts ' S .2 pa deposits were lightened in color upon the addition of 10 lg . [1. of denizens. No further change was observed untillOO ng./l. was reached. it this concentration a considerable darkening occurred which say have been caused by the occlusion of anall- counts of colloidal aluninus hydroxide. here wetting agent no required with ale-in- present in all four types of solution to prevent pitting, especially in the organic type solution. Adhesion - The adhesion of the deposits to the base netal was tested by bending the lip of the 0.001” panels 180 degrees around a 3/16” Iandrel in the niddle and across the short dimension of the piece (1). There was no indication of nan-adhesionin any of the deposits. Where cracb appeared, the separation was invariably in the base netal, not at the interface between the base metal and the deposit. motility .- Ductility neasurenents were node as described in a previous paper (1). The 0.001” deposits were stripped fro. the nickel plated panels , bent double and creased between the fingers , straightened , pressed flat and carefully creased again on the some fold. This was repeated until the foil parted. In order to nininise the hush elenent, several individuals tested strips fro. the seas foil, and an average was obtained. Results showed nonarked effect on any of the deposits regardless of alanine cementration. Hardness - Deposits of 0.032" thiclmess were tested for hardness ‘ at the National Bureau of Standards. A Tukon tester with a Knocp type of dinond inienter was enployed. Differences in hardness values are recorded in Table I as percent change from the hardness of deposits free pure solutions. TABLEI MWWWWGNEMDEPOSHS —' W ‘— v V— Percsnt Change Alminms Concentration Wat-r Tttev m ng,{1. pl 2,2 2.35.2 DH mg V 13 2,2 0 0 O O 0 10 13.8 -11 .2 6.9 10 .0 20 -O.6 1 .5 -l .7 10 ,7 30 -303 2 so -9 o3 5 oh so -5 co .2 so .9o3 11‘ e6 1w 5 so 2 6.2 .10 o7 1‘ .7 Two general observations are readily apparent . lluninun in solu- tion decreased the hardness of nickel-cobalt deposits and increased the 10 hardness of the organic type of deposit. Secondly, the effect of the alwinmn was greatest in the very low and the very high concentrations with only nominal changes in between. This effect was also observed in the appearance of the Watts' 5.2 pH deposits and in the salt spray corrosion resistance of the thin deposits. Throwing Power - As described earlier the throwing power measure- nents were performed on both the top and bottom of the lip of the panel in an effort to elininate errors due to the tilting of the lip. Such misalignment might be caused by strains in the deposit developed during electrolysis or by a error in the construction of the bent cathode. Repeated polishing, etching and neasuring tldclmessee gave a fair estimate of the precision possible with the nethod. It was concluded that both top and better: of the lip should he considered and that the average of all such measurenents should be used to evaluate the percent change in throwing power. The precision was never better than plus or mime 2 percent change in throwing power efficiency. Average readings indicated a 12 percent decrease in throwing power for the Watts' 2.2 pH solution upon addition of 10 ng./l. of alunimn. Changes at other concentrations of aluninun in this solution were insig- nificant. Doposits from Watts' 5.2 pH solutions showed a decrease in the throwing power of those solutions of 17 percent at 20 sujl. and about ll percent at '30 ng./l. of alulinun. Nickel-cobalt solutions exhibited a decrease in throwing power at nediun concentratiom of aluminun ani an increase when higher concentrations were present . The organic type of solution showed little change at any concentration of altminuu. TLBLEII NET 0‘ W Oil THE TERMS POWER OF NICKEL MING WIONS a w um _ _ __ Percent Change - Concentration Vatts' Watts' Niciiil-cobalt Organic ng./l. va2.2 pH 5_._2 __ __ pH 3.75 PH 3.2 0 O 0 0 0 10 -12 -2 0 3 20 h -17 -7 -3 30 1 -11 -3 O 50 -3 0 +3 2 100 h 5 9 5 Salt Spray Corrosion -- Triplicate panels were prepared frost each solution to be tested according .to 1.th Tentative Method of Salt Spray Testing, Bll7-h9'1‘. nickel deposits of 0.0003", 0.0015 and 0.0015' from the four types ofsolutions each containing 0, 10, 20, 30, 50, and 100 ng./l. of ale-inns .were tested. Magnesiu- oxide was used as a sur- face cleaning agent Just prior to the corrosion test. Arbitrary breakdown conditions were established for each thickness series. For example ,.-the 0.0003' panels corroded rather unfunny in seen rust spots over the face er the panel. In both the 0.001- and the 0.0015” panels, however, one, two, or three rust spots would appear, sonstines grouped, sonetinss scattered .over the face of the panel, and deterioration would proceed.by corrosim in depth. It is conceivable that one or two such highlynnedic spots night protect the rest of the panel by naintainingit relatively cathodic. Proof of the nature an! strength of these anodic. spots was the reaction on the reverse side of 12 the panel. Edges and backs of the panels were protected by'a continuous file of 'stcp-off"lacquer, but at a spot directly opposite the large rust spot, a bobble forned.under the lacquer. ‘The surface of the panel beneath the bubble was dry'and free of visible corrosion.products, indicating that the gas had famed on the back side -of the panel before a pinhole appeared. However, only a thin fill of moisture need have been electrolyzed by the gelvanic action.to evolve an appreciable quanp tity of gas. During the corrosion testing of the organic 0.0015” deposits a new phenomenon was noted. Nor-a1 corrosion took:place for the first one hundred hours, but after this point even previously'active rust spots ceased to corrode for hundreds of hours. A water insoluble grey film an the panels was found‘by qualitative analysis to be basic nickel carbon, ate, 23100..331(0B),oh808. Juicrosccpic exasination as shown in.Plates l and II disclosed that about 50 percent of the surface was covered by this film. Considerable difference in crystallinity was noted depending upon the type of nickel deposit. Regular, well-defined hexagonal crystals could be seen on the bright nickel deposits, while amorphous Iasscs occurred on grey nickel surfaces. The new centers of nucleation presented by the irregular grey’nickel surface evidently'prevented the growth of any’one crystallite. Repetition.of this experiment in which the organic type deposits were corroded isnndiately-after'plating and in an atmosphere from which a known source of carbon dioxide had been removed showed none of this basic nickel carbonate formation. These panels corroded much more quickly and in much the same manner as other 0.0015" panels . PLATE I.(TOP) PHOTOMICROGRAPH OF BASIC NICKEL CARBONATE ON WATTS TYPE NICKEL DEPOSIT. POLARIZED LIGHT. 250 X. PLATE II.(BoTToM) PHOTOIICROGRAPH 0? BASIC NICKEL CARBOVATE 0N ORGANIC BRIGHT NICKEL DEPOSIT. 250 x. IL is mentioned in a previous investigation, the thin deposits corroded so quickly that the results are greatly exaggerated (9) . However, the sane trends seemed to be present in both the 0.0003’I and the 0.001'l deposits in the lower pH solutions, i.e., the organic and the nickel-cobalt types (See Figures 1, 2, 3 and h) . With the exception of the nickel-cobalt series, the thin panels indicated an increase in corrosion resistance in the low concentrationf of almninum where the probability of a stable colloidal dispersion of the hydroxide was greatest. This indicated that the effects were due either to aluminum ions in solution or to small amounts of colloidal aluminum hydroxide. Alunimm is too elsctropositive to remain a free ion so the latter seems more probable. Since the cathode film is of a higher pH than the rest of the solution coagulation of larger concentrations of the hydroxide particles could alter the size factor and modify the effect on the deposit. O'Sullivan has shown that very large concentra- tions of aluminum and iron dispersions occlude or co-deposit (due to adsorbed positive ions) to form a gelatinous covering of the cathode. The few nickel ions which penetrate the gelatinous run deposit in thin plates. It is presumed that the nickel ions are discharged imediatsly when they penetrate the rifts in the gelatin rather than following the structure of the previously deposited nickel (S) . In general, the aluminum in nickel solutions seems to have its greatest effect in small concentrations and probably affects the deposit only near the base netal. Relatively large changes in corrosion resistance occur at low concentrations of aluminun in the thin deposits, somewhat slaller changes occur in the 0.001" deposits and the heavy 15 w 80 0 2 E a: 40 5 g 3 O O l ; 2 8 g \' -4o~ .\° -eoo 20 40 so so IOO ALumum CONCENTRATION , MG/L. FIOJRB 1. m <3 mm coscmna on me am span (roe) conaosxou assumes a DEPOSITS cs HIRING mmmssss FRO! ms wens 2.2 pH sownos. 1. 0.0003 INCH DEPOSIT 0 0010 DEE DEPOSIT 0 2. . 3. .0015 INCH DECS]! 9'. CHANGE IN CORROSION RESISTANCE 16 IZO 80 4O -4o 0 20 4O 60 80 . IOO ALUMINUM CONCENTRATION MG./L. FIGURE 2. EFFECT OP ALUMINUM CONCENTRATION 0N TRE SALT SPRAY (FOO) CORROSION RESISTANCE OF DEPOSITS OP VARYINO THICKNESS FROM THE WATTS 5.2pH SOLUTION. 1. 0.0003 INCH DEPOSIT 2. 0.0010 INCH DEPOSIT 3. 0.0015 INCH DEPOSIT 4O 7. CHANGE IN CORROSION RESISTANCE 6° 3 C) 2""; 17 O 20 4O 60 80 I00 ALUMINUN CONCENTRATION, MG./ L. FIGURE 3. EFFECT OF ALUNINUI CONCENTRATION ON THE SALT SPRAY (FOG) CORROSION RESISTANCE 0F DEPOSITS OF VARYING THICKNESS FROI THE NICKEL-COBALT ALLOY TYPE SOLUTION. 1. 0.0003 INCH DEPOSIT 2. 0.0010 INCH DEPOSIT 3. 0.0015 INCH DEPOSIT °le CHANGE IN CORROSION RESISTANCE 18 ALUWUM oouoENTRAmN, MGJL. new“. man ornmmmcoscmnososmw smut (roe) coaaosms assxsnncs or smarts or man": mmmssss PRO! m ORGANIC ms summon . 1. 0.0003 IEH DEPOSIT 2. 0.0010 mil DEPOSIT 3. 0.0015 INCH DEPOSIT 7.. In. Hand NINA, a a . , A r _ Enthuv‘ 19 0.0015'. deposits show almost no effect. This panels fron both high and low pH fitts ' solutions and the organic type of solution showed large increases in corrosion resistance at 10 and 20 ng./l. of alumina. The thin nickel-cobalt deposits showed a decrease in the same concentration range. For the 0.001' deposits from higher pH. solu- tions, namely the Watts' 5.2 and the Motel-cobalt 3.75 pH, the corrosion resistance was decreased. In the lower pH solutions, the Hatts' 2.2 pH deposits were unaffected and the organic type 3.2 pH solution deposits showedj, substamdal increase in corrosion resistance. Bfects of Iron on the Electrodeposition of Nickel -- in investi- gation of the effects of iron on the physical properties of various types of nichel deposits was conducted by Ewing 9; _._J_.. in 1952 (1). Results indicated that very little change occurred in deposits from solutions containing 0 to 200 Ig./1. of iron. A slight whitening occurred in the Watte' deposits which was very similar to the effect of aluminum in deposits nade using the same techniques. No change in appearance was evident in deposits from organic or nickel-cobalt alloy solutions . Again this paralled the effect of alminun. Insignificant changes in adhesion and throwing power were noted for all types of deposits, but some increase inhardness occurred if any iron were present in solution. Salt Spray -(I'og) corrosion resistance was affected only in the nickel-cobalt- series_of panels in uhiohan increase in resistance of about twenty percent was reported. It should be remembered that this investigation considered iron in otherudas pure solutions. Iron in nickel solutions of only noderate 20 purity causes dullness and brittleness of the electrodeposits. Colloidal Properties of Iron and lluninm Hydroxides - It appears surely in the case of elm-inn and at Just to some extent in the case of iron, that the colloidal hydroxide .deternines-the effect on the electrodeposit. Since no aluaim is removed by electrolysis, films of hydroxide interfering with deposition in sale nanner lust change the character of the deposit. an. iron does deposit electrolytically, it may be from the broaldown of a muroxide particle. Many of the investigations about the formation, coagulation, pre- cipitation and structure of aluinun and iron hydroxides were carried out in solutions which did not contain nickel ions. However, the onerous effects nay help to explain certain phenonena encountered in the purification of solutions for the electrodeposition of nickel. Femation of Colloidal Mrsddos - In solutions of ferric or alum. chloride tin hydrolysis.“ precipitation of iron and alt-inn. hydroxide occur in the follodng stops: aqueous solution, formation of the colloidal state, concentration of the colloidal particles and finally precipitation . According to Shikorr, there are three classes of salts which will coagulate an iron eunuch (10) . the first, including 01', as, 1", N0; anions, foul easily soluble salts of iron and coagulate at high concentrations. The second, e.g. SO: and P0: anions, fen difficultly soluble salts with iron and coagulate easily. Alkaline reagents, pre- cipitating Po.0.°XHOH, fem the third group, and they coagulate in much the sue-Inner as group two. 21 h. ouintin, working with mixtures of 0.91 Fool. and 0.94 Feel, precipitated under nitrogen by 1H [03, showed deviations from the theoretical curve (figure 5) which indicate a couple: precipitation process. The first precipitation occurs at pH 2.0 and only ferric hydroxide should be rucved at this point. Beaver, the slope of the line is double the predicted slope indicating either an abncrnal amount of ferric ion being renoved or that some ferrous ion is being co- precipitatod. it pa 5.0 only ferrous hydroxide should be precipitated, but it is thought that a snail mount of ferrous ion is oxidized to yield a ferrous-ferric product . Touperature changes the curves in the direction of the arrows, but the general shape renains the sane (11). Scale gases will also coagulate solo or n.0, (12). Stark found that oxygen, hydrogen, m and nitrogen is that era... effectiveness will cause coagulation when bubbled through such a sol . Carbon dioxide has no effect . Coagulation velocities vary with the concentration of electrolyte in accordance with Seoluchowski 's squat ion f The indications are that this type of coagulation involves the adsorption of colloidal particles at tin gas-liquid interface of the bubbles . In buffered solutions of nickel sulphate the pH of beginning oboe rv- able iron hydroxide formation was deterained by Rotinyan and Zel'des by n .1 +E'm:;;ot n - total particles at tine, t no:- total particles originally present D a diffusion constant r - radius. of sphere of attraction of particles o 5 9" I0 FIGURE 5. 0.5M FeC12 AID FoCls ---- THEORETICAL CURVE EXPERIMENTAL CURVE M. QUINTIN 22 —'-—l—_-- 23 potentionstric titration at SOP’C. on a glass electrode (13). The end, points corresponding to the appearance or disappearance of the hydroxide were detected visually and.by'ths observation of the Tyndall cone and were found to lie at tin sans pH. In the following tables are found the effects of sulphate, chloride, boric acid, and.tenperature as well as combinations_of several of these factors, on the precipitation of iron hydroxides. TIBLB III (Ni.a; (Hi a; (Ni.a; (n1 a; NiSO 11180 11130 8180 NaSO j $1.325: _ 4.4.521. 21 MA. N1 80/10 pH N‘sod 80/)” pH __§w1 80/10 par "261 80/10 pH 10 6.3 20 o 6.0 o ‘ 5.9 25 5 .9 no effect 5 5.7 S 5.7 39.6 5.7 , 20 5.6 20 5.5 61 5.5 80 50 5.5 59 5.h IABLE Iv (Ni as 011 as (Ni as men.) moo.) mesa, NaCl use.) mesa. NaCI ML :6 5,5“ ho ha. 5 5,1}, 2 6 g.&. to 5.2, 50 hp. 1301 80/10 pa __ H3803 80/1- pa A H3808 80/10 pa _._ _ o O 5.7 0 5.6 no effect 10 5.0 10 h.6 20 h.6 20 h.2 so to 3.9 ho 3.6 A 2h TABLE V NnSO‘ ho, Nec1 5, 11,30, 20 g./l. 20° 0 50° 0 1 _‘_7_0° 0 1 # (Ni as (Ni as (Ni as N180.) g./1. on 11130,) g./1. pH N131.) g./1.# pH 21 5.2 21 1.3 21 11.6 to I 35.0 to h.8 ho h.S 63 m 63 h.5 63 14.); Structure of Colloidal Hydrondes -- Ferric and alumina hydroxides are extremely structure sensitive. Variations occur due to method of preparation, the concentration of other materials in solution, the physical and chemical properties of these materials, the temperature and the amount of aging which the colloid has undergone . Changes in x-ray diffraction patterns are ascribed by Esoard to an increase in size of micelles in aging (lb) . X-ray diffraction patterns of Fe(OH) 3 formed by hydrolysing Fec1, at 100-180" c. were similar to thereof hematite (1e30,), but when mixtures of ferrous and ferric chloride were treated with alkaline solutions, a nagmtic black precipitate whose x-ray patterns correspond to magnetite -(Ee0.re,0,) was observed by Daturai. Supersonic urea changed it into a colloid with a characteristic magnetic field (15). According to rug-much amorphous forms of 11(03), and Roma), are quite stable and appear to be about 30 A in diameter with no regular 25 structure. Spheres 50 A in diameter have been prepared fran decompos- ing penta carbonyls . Aged iron oxide hydrates .fora needles 600 to 5000 l in length. The alpha fom (goethite) occurs in parallel bundles 80 to 165 A wide, and the gamma form has a random arrangement only 35 to 115 A wide (16). V ' Particle .1... has been determined with the ultrmnicroscope or by filtering through carborundun disks of various grain size by last and Duclaux (17) . Electron microscope studies allow a more direct measure- nent of .1... Because of the tremendous surface area of a sol, adsorption of other ions or of Pe(OK), on other crystals is constantly occurring. Kohlschutter and Nitsohmatm say that pseudomorphs formed by the hydrolysis of alkali ferrites exhibit properties which are due to their tepochemical formation on a crystal . Variations may occur due to purity, age, saturation effects, and oxidation by hydrogen peroxide (18) . In mltilminar films of ferric mdroxide the thickness of the unimicellar film is about 1.0 A , whereas ultrafiltration determinations indicate a size of 20 an (200 A). This indicates according to Hokrushin a plate-like particle shape which lies stack-lies in forming the film. nunimnn hydronde behaves in a similar answer, but the limiting thick- ness of the film is about 12 1 (19). Thiele and Kienast found hydrated ferric hydroxide sols are aniso- metric, that is, the physical properties in one direction through a crystal may differ from those measured- through another direction in the me crystal. Birefringence, or the property of double refraction, 26 is due to this asymetric nicelle. Birefringence is affected by the concentration of electrolyte present. Lithium, sadism, and potassium in 0.)! solutions have an increasing effect in that order, as do the bivalent ions berillium, magnesium, calcium, strontium , and barium. Hultivalent ions cause hydrolysis and aggregation of the sole (20) . In studying solutions containing both bivalent and trivalent metals Feitlmecht found that double hydroxides of the general formula n.1, maybe tor-ed (21). Minoan-mos)” hco(0u),-11(0s),, and hoe (OH).-Fe(OH), are examples of the type. The crystal structure consists of alternate lattice layers of the two hydroxides . With increasing radii of the trivalent ions the tendency to form 14.1,, compounds decreases ‘ as shown in Table VI . r1313 v1 new. A; A; +8 —— : ‘7 s‘ fies -; Ion __ Ni" 14;” co z "_ in ' ca‘ Ca onicw ' “— _ *— ‘— radius 0.78 0.18“ 0.82 0.83 0.91 1.03 1.06 +3 V EM H.“ t A]. 0.57 H‘A‘ H“1 “ML H“: 1.8t. D.H.2 CI‘I es “8m _ 93* 0.6!; " H 1 “eh 211° " " " _ a v I, 'w “n O ’ — :0 0.67 ‘, - H‘AI 60(03). 2110 H‘AI D.H;2 03“ Mn 0.69__ - n.1, - __ ZnO LD.H.1 - .- a i v _ gr 0.71%; 33.3». - 230 - 0.11.1. 6.11 a v 8 s T _ A La‘ 1.22 - La(_og).A - - - cams), Lama), 14,1, Type like Hgfl. double hydroxide. D.V.Str. Double combination unknown structure. 1) .H .1-h Double hydroxide with double layer lattice. 0.1,, Type like Dell double hydroxide. (v. Feitknecht) 2? If the basic character of the bivalent ion is increased, double hydroxide for-ation is enhanced, but the true H‘it crystal structure is obtained only when the radius of the bivalent ion is less than manganese. Larger ionic radii cause a modification of the crystal structure and finallyrvith oalciun it is completely altered. Adsorption‘by Sole -- Colloidal particles are unstable unless peptisod by'the adsorption of ions. Both positively and negatively charged sols of Few“), can be adsorbed on Pe(OH), with a resultant ring fbrnation (22). Both sulphate and chloride ions can be adsorbed on these particles (23,2h). Fran two to ten and one-half these sore nickel than copper is adsorbed on rows), unless the mm and 1m. ”mmnmmm.nWMmfimwummmmgg, the adsorption of the copper decreases to zero (25) . lajnik .e_t. 5;. found that adsorbability'dbpends not only on the valence of the adsorbed ion but on specific propertieejike couple: ion formation or the forn- ation of insoluble salts (26). Adsorption of'netal ions is very'pr0b- ably a double hydroxide fomation vith H“; type most effective. In efforts to purify Mon). and man). by three precipitation Okac 93 31. observed the following data (27): ’ TABLE VII PERCUT (1" ORIGINAL ADSORBED IONS W 00 Hi In (b 2n Gd ._ ##M rows). 20.? 8.5 17.7 1.1 1.2 0 11(05). 95.15 93 .1 119.6 0.8 23 .7 0 sun—H -—-—-.___—-—-—- w.— 28 It can readily be seen from this discussion of colloidal properties of sols that nany changes other than concentration of iron or alunimm any cause changes in electmdeposits.- For example, is it the effect of increasing iron concentration which causes a change in appearance or is it the aging of the sol and its subsequent change in physical state from amorphous to needlelikn stmctum? Do differences in nickel-cobalt solutions from regular- nic ksl solutions occur because of enhanced double hydroxide fomation? -(See Table VII) . We know that additional impuri- ties nalce iron. behave differently. Is it due to adsorption of foreign ions on the ferric hydroxide sol or nerely an additive effect? Fran this single aspect of the probleln, 1.e. uhst is the state of the iron and alumina- in nickel solutions, could originate years of research. Renoval of Alminun Fm Nickel Solutions Just as important as their effect on physical properties, are the effects of iron and aluminum sols on the rate of their removal from solution. This is true of electrolysis and of high pH precipitation nethods. The two methods of renoval of alunimn studied were high or low current density electrolysis and the high pH precipitation of aluminum hydroxide fr:- the solutions ,by the addition of nickel carbonate . Normal operating temperatures.“ 55° C. were used for both the electroly- sis and the precipitation nethods. Agitation of the electrolytic removal series was four feet per ninute of solution past the cathode. .-.._—v- .-—II———-._w ———- -flfi“ 29 Current densities of 2, S, and 1.0 amperes per square foot were used on a flat, 2" x 3 1/2' cathode. Samples were pipetted into 25 ml. erlenmeyers and tightly stoppered until analyzed Spectrographically. The 3082 A wavelength of aluminum was compared to the 3116 A wavelength of nickel. In repeating these experiments the colorimetric method of analysis aluminum proposed by Serfass was used. To determine the precipitation effects from solutions which original- 1y contained 100 mg./1., the pH was raised to about 5.3. Intermittent shaking of the high pH precipitation vessels insured uniformity of solu- tions and an average pH of 5.35. Samples from these high pH solutions were withdrawn) at intervals and centrifuged; then a ten milliliter aliquot was taken from the clear liquid. To prevent further precipitation of this ten milliliter sample it was imediately acidified with a few drops of concentrated sulphuric acid. .- .‘tpparently no aluminum is removed by actual deposition although it is possible a very small amount is occluded as the hydroxide, expecially from solutions of higher pH. One hundred milligrams per liter of aluminum was added as the anhydrous aluminum chloride. The hydrolysis of this compound, even in small amounts, liberated enough acid to lower the pH and therefore increase the amount of aluminum initially held in solution. Subsequent raising of the pH to its normal value by the addition of nickel carbonate lowered the aluminum content to its quasi-equilibrium value . Time , temperature, and pH were carefully controlled through the experiment and analysis was carried out as quickly as possible after sampling. It is difficult to interpret results from older samples since aging of the sol , coagulation and precipitation may cause erratic 30 sampling from the sample bottle . Precipitates and gels have been observed in week old samples. Figure 6‘ shows the results of 2 amperes per square foot electrolysis on the 2.2 pH watts, the 3.2 pH organic, and the 3.75 pH nickel cobalt solutions. All of the aluminum was pre- cipitated from 5.2 pm watts solution during the time it was being brought up to temperature in a water bath, therefcranc curves are aims: for that solution. Within the limits. of accuracy of the analytical procedure the curves are straight lines showing no removal of aluminum due to electrolysis. Figures 7 and 8 show results of 5 and to ampere: per square foot electrolysis respectively. Again no removal of aluminum due to electrolysis is noted in any solution. however, there is a little more distinct spread in quasi-equilibrium concentration according to pH. The concentration of alumina seems not to be entirely a function of pi! and the type of solution as is shown by a composite set of curves in Figure 9, with concentration plotted as a function of pH at various rates of electrolysis . These curves were prepared from previous experi- ments with electmlysis during which the pH was not as carefully con- trolled. Also the analysis and pH of the sample were taloen several months after removal from the plating solution. Below the buffering range at pH h.0 to LS there is very nearly a straight line function between pH and the concentration of alumimm during 2 amperes per square foot electrolysis . is higher cm'rent densities are employed, the curve changes to indicate a greater solubility in the medium pH range, i.e., the organic and nickel-cobalt types of solution. This might be due to 31 . IZO “d ‘5 1 E (2, so 2 ,: < a: p. 5 o 60 z o o 2 3 E 30 2 a .1 < O O I 2 AMPERE HOURS PER GALLON FIGURE 6- EFFECT OF 2 AIPERES PER SQUARE FOOT OI THE ELECTROLYTIC REMOVAL OF ALUMINUN FROI VARIOUS NICKEL SOLUTIONS. 1. 2.2 pH HATTS SOLUTION 2. 3.2 pH ORGANIC SOLUTION 3. 3.75 pH NICKEL-COBALT SOLUTION IZO 60 30 ALUMINUM CONCENTRATION, MG./L. 32 O 2 4 6 AMPERE HOURS PER GALLON FIGURE '7. EFFECT OF 5 AMPERES PER SQUARE FOOT ON THE ELECTROLYTIC RENOVAL OF ALUNINUN FRO! VARIOUS NIGEL SOLUTIONS. 1. 2.2 pH WATTS SOLUTION 2. 5.2 pH ORGANIC SOLUTION 3. 3.75 pH NICKEL-COBALT SOLUTION 33 I20 90 60 30 ALUMINUM CONCENTRATION, men. O I6 32 48 AMPERE HOURS PER GALLON FIGURE 8. EFFECT OF 40 AMPERES PER SQUARE FOOT ON THE ELECTROLYTIC REIOVAL OF ALUMINUM FRO! VARIOUS NICKEL SOLUTIONS. 1. 2.2 pH WATTS SOLUTIM 2. 3.2 pH ORGANIC SOLUTION 3. 3.75 pH NICKEL-COBALT SOLUTION 3 I00 — / .i \. o I ' 2 80l— I O z. 9. 2 E g 60 C O 3 g 40 2 Z) d 20 $5 3.0 4.0 5.0 6.0 PH name 9. Immune CQICMRATIGIS (1" Alumni as A menus or pH. 1. 2 AHPEES Pm flUARE FM 2. 5 AHPEES PE QUAKE FOOT 3. ho ANPEES Pm QUAKE POM . . V N, ,t ‘ .flfl blip! x ,. . .. . can}? .t 2?... V -5 35 the formation of organic complexes of aluminum. However , considering that the almimm hydroxide is known to be in a colloidal state at least for part of its existence in the plating solution, it is more likely that under the influence of increased potential more ions are adsorbed on the hydroxide particles, thus holding them in suspension. Figure 9 indicates that most of the almimn can be removed from any nickel solution by raising the pH above h.5 electrometric. Further experiments showed tint in one hour it 55° c. a concentration of 100 mg./l, of claims can bsnreduced to leeathanlO mg./1. in am of the four types of solutions by such high pH precipitation. Removal of Iran from Nichol Solutions To remove iron frm nickel solutions in the shortest possible time the pH should be raised to about 5.0 electrometric with nickel carbonate, the iron oxidized with hydrogen peroxide , the solution then heated to 55 to 65° 0. for an hour with agitation. By this time the iron will have been precipitated as the ferric hydroxide and most of the excess peroxide decomposed and the solution can be filtered substantially free of iron. Nickel carbonate and ludrogen peronds are ideal reagents because no fweign ions are added to the solution. While the nickel carbonate is expensive and not very soluble compared to sodium hydroxide, the gain in nickel metal content salewhat compensates the cost. In spite of the difficulty of introducing it into the solutions, nickel carbonate additionereven in excess, will cause the pH to rise only to S or 5.5 electrometric . Considering that above a pH of 6 appreciable 36 quantities of nickel precipitate as the hydroxide, this buffer action of the nic bl carbonate is invaluable . In nam- comercial installations it is not economically feasible to stop production in order to carry out the precipitation purification often, if at all. If nickel plating solutions are operated at a pH of 3.5 or above, with continuous filtration most of the ferric iron will be removed as ferric hydroxide . This is in accord with Veisner's data which show that ferric iron is completely precipitated from nickel sulphate solutions buffered with boric acid when the pH is 3.5 to 3.75 electronetric (28). The ferrous iron concentration will increase, however, due to galvanic reduction by nickel and in this reduced fom will not be precipitated. ’ Figure 10 sumarizes the results of a previous investigation by the author using a 3.75 pH nickel cobalt solution (29). The most ilportant information derived from these curves is that precipitation of ferric hydroxide at this pH exceeds some rates of electrolysis. Figure 11 clearly shows the advantage of electrolysis in low pH solu- tions as compared to high pH solutions where precipitation automatically helps to purify the solution. In solutions of greater acidity in which iron remains as the ferrous ion, contimms low current density electrolysis in separate 'chnm" tanks becomes advantageous. The variables affecting the optimum electrolytic methods of removal of iron from 2 .1 pH nickel-cobalt plating solutions will be considered in- detail . |20Ir IOO ~ _i \. to 80 ' 2 z. 9 E J z m o z o o z o 9:. t' l fl l L. 2 I 5 8 I0 TUME , HOURS FIGIRE 10. mm W MTROLISIS AND PRESIPHATION 0N Rama. 0? IRON mm 3.75 pH NICKEL-COBALT SOLUTION. SOLUTIQJ CURRm TmmATUHE CATHODB minnow Dmsm °c . 1mm” 102; com: 1 20 Fin/uni. 10 AS!" 65 mega curve 2 10 n./um. 10 ASP 8;} 0 15:22; CURVE 3 20 FLAIR , 5 ASP 8’) 0 3 EEC-H CURVE h 20 Fifi/MIN. 375 SF so «3.5 INCH CURVE S 20 Flu/MIN. 1:“... A33 3) 0.75 11:10:»: CURVE 6 20 PLAIN. 10 .95? 83 C .5 1m}; CURVE 7 PRmI ITATIQN arm MGJ L. IRON CONCENTRATION, .38 |20 IOO 80 60 - E: 8 l2 I6 20 me , HOURS FIGURE 11 . WECT or HIDROGH ION CWCMRATION ON THE ELETROLITIC WOVAL (1' IRON F304 NICKEL-COBALT SOLUTIONS . CURVE 1 - mscnrmnou mam IN A 2.1 pH soeron cum: 2 - orrmm counnxons roe muousxs or non mm A 2.1 pH sonmos CURVE 3 - PRECIPITATION mm'r n A 3.75 pH ennrrlou cum I; - OPTIMUM coan'Ious FOR mamxas or IRQI mm A 3.75 pH SOLUTION 39 Relatively large volumes of solutions were used for the electrolytic removal experiments in order to minimize variations in temperature, pH and concentrations. Right to eleven liters of solution were heated in round Pyrex Jars immersed in a water bath. Nickel anodes containing one percent cobalt were submerged to give a projected area of about one-fourth of a square foot. Immerable small cavities produced a much greater effective anode area. Solution agitation of about four feet per minute past the cathode was maintained by two stirring notors driving two-bladed propellers which lifted the solution from the bottom toward the top of the Pyrex jar. The 1 on. x 2 an. blades were tilted forty- five degrees from the vertical and placed at one-quarter and three- quarters of the tank diameter . Corrugated cathodes with indentations one-half inch deep were constructed of upper or steel to give a pro- jected area of 1.125 square feet. After standard purification of the solutions by high pH precipitation, activated carbon trea‘hnent , and low current density electrolysis, the. pH was adjusted with sulphuric acid and 100 ng./l. of ferric ion was added as FeCl,°6HOH. The solutions were analyzed for total iron using a aethod proposed by Serfass and Levine (30) and modified by the author to give more accurate and repro- ducible results (29) . The ferrous ion was oxidized quantitatively with standard potassium dichronate solution_and the endpoint detected with a Beclman Medal B pH meter using platinum an‘l calonel electrodes. Figure 12 demonstrates- the effect of current density on the electrolytic removal of iron tron a 2 .1 pH. nickel-cobalt plating solu- tion. As the current density was raised from five anperes per square I20 .i \ o' 2 z“ 2 .- g g... § 0 40 2 g 20 O 4O ‘0 2 A 3 O 20 40 '60 80 '00 AMPERE HOURS PER GALLON FICURE 12. mm: (s CURRM DEBIT! (I THE WING REMOVAL OF m PRC)! A 2.1 pH IEKEL COBALT SJLUTION CURVE l S “PF-RES Pm QUARE PM CURVE 2 1.5 WEEKS Pm QUARE FOOT CURVE 3 10 .HPERES PER SUI-RE FOOT CURVE b 15 WEISS PHI fiUARE FOOT foot, the removal of iron was enhanced until 10 amperes per square foot was reached. Higher current densities had a faster initial rate of renoval as, for example, curve 1; for 15 amperesper square foot. Honour, depletion curve 1: levels out at a much higher value than does curve 3 for 10 alperes per square foot. 01' those current densities tested 10 amperee per square foot appears to be the Optimum current demity for the electrolytic removal of iron fron low pH nickel-cobalt plating solutions with mixed valences of iron . Snell additions of hydrogen peroxide will facilitate this Optimm rate of removal as will be shown later. The dip in curve 3 for 10 anperes per square foot was caused by inserting a capper cathode of the same size and shape as the previous steel base cathode into the sac test solution. Variation of the solution agitation gave results which at first glance seemed inconpatible with the physical picture (Figure 13). The removal rate increased as expected when the agitation was increased from 10 to 20 feet per minute of solution past the cathode, since more iron ions were placed in contact with the cathode per unit time. However, the removal rate decreased when the agtation was increased from 20 to 50 feet per minute. A second experinent was performed at 50 feet per minute solution agitation, and the points. fell exactly upon the previous curve . Probably this phenomena -is due to the relatively faster buildup of ferrous ions in a solution which is also more often in contact with the nickel anodes. 42 l20 :1 I00 if 80 ' \(’¥ '4 a so 5 E 95 so 2 I g 0 . 3 \A z 40 E. V 20 O 20 4O 60 80 IOO MERE HOURS PER GALLON FIGIRE 13 e HINT (P SOLUTION AGI'I‘ATION ON THE WROLITIC RHOVAL or IBM FRO! A 2.1 pl! NICKEL-COBALT SENIOR. CURV'El-IOFBTPEHINUTE CURVBZ-ZOPEETPEHUUTE CURVBB-SOFEETPEHIIUTB 143 The construction of the cathode has some effect on the mount of iron removed free solution, but, as shown in Figure 11:, the material from which the cathode is constructed is even here important. The copper base cathode moved iron lore quickly than either its corrugated counterpart in steel, or a flat copper cathode. There may be more fundamental reasons for this difference than corrosion of the steel base through cracks in the heavy nickel coating, but from these data they are not apparent. Tenperature has very little effect on the removal of iron from nickel solutions when it is varied from 51.3" 0 (125° F) to 65° C (150° F). The difference which appears in Figure 15 can be explained by the amount of ferrous ion in solution at the beginning of electrolysis. Figures 16 ani 17 show the buildup of ferrous iron as the total iron is depleted by electrolysis at 65° and 51.3° c respectively. Since ferrous ion is more slowly electrolyzed free solution due to the swaller change in free energ of that reduction it is obvious that the initially higher ferrous ion content in the 65° 0. experiment would at least slow the effective removal of iron. Figure 18 illustrates an attespt to determine the effect of anode area upon the removal of, iron. Reduction of both anode and cathode areas in order to maintain the some current densities as in previous experiments caused such a fast buildup of ferrous ion compared to the removal of total iron that. it was abandoned. Leaving the cathode area 1.125 square feet arr! reducing the anode area by half had the effect of doubling the anode current density when the standard conditions of IRON CONCENTRATION , MG/L. 44 80' 4O 20 O 20 4O 60 80 IN AMPERE HOURS PER GALLON FICURE l4 . W O? CATKODE CWSI‘RUCTION ON THE WROLI‘I‘IC RWY“. 01" m FRO! A 2.1 pH NICKEL COBALT $111110“. CURVE 1 - PLAT CATHODE CURVE 2 - 0.5" CORRUGATED sum. CATHODB CURVE 3 - 0.5" C(RRUGATSD COPPfli CATHODE MGJ L. IRON CONCENTRATION , 45 IZO IOO 60 20 O 20 4O 60 80 IOO AMPERE HOURS PER GALLON FINES 15. mm G 1mm 08 THE WROLITIC REJOVAL 0! mos PRO! 2.1 pH NICKEL-OCEAN PLAYING sownous. cum: 1 - 65°C. cum: 2 - 51°C. IRON CONCENTRATION , MG./L. 46 l2 4O so; ol Ji O 20 4O 60 80 AMPERE HOURS PER GALLON FICURE 16. Home: IN reasons 10' coucmaa'rlou wants mrmxas AT 65°C . CURVE 1. TOTAL mos COMMRATION CURVE 2. FMS ION CQCMRATIOI IOO IRON CONCENTRATION, MGJL. IZC lOO—-— 0 40 - 20 0L O 20 40 60 80 AMPERE HOURS PER GALLON 31mm. news: IN renews ION communion wmc mmousxs AT 52°C. CURVE 1 - TONI. IRON CMCDITRATION CURVE 2 - FERRWS mm CONCMRATIG CONCENTRATION , MC./L. IRON EC 48 80F 20 O 20 4O 60 80 I00 AWFERE FWURS PER (flLLON FINES 18.. mm or DEBRASED ANDE AREA on 3130180111710 RHOVAL (I IBM. CURVE 1 - TOIAL IRON COUCBHRATIOI WRVE 2 - PMS ION COEMRATIOU ‘3‘“ .I...‘ lag. 4‘s“ ._ 149 10 anperes per square foot cathode current density were employed. Apparently the anode current density has no effect on the galvanic reduction of ferric to ferrous ion. Over the same interval of lo ampere hours per gallon, the rate of ferrous ion production was just half as fast as with the larger area. The reduction of ferric ion to ferrous ion appeared to be directly preportional to the niche]. anode area in solution. This correlates quuist's work with comartnented electrodes where he found very little buildup of ferrous ions in the ccthclyto (31). seen anode area, therefore, enhances the removal of total iron free solution . The fact that ferrous ion was deleterious to electrolytic removal nethods suggested two experieunts to eliminate it. One was the use of wall amounts of twdrogen peroxide {and the other was the generation of nascent chlorine with a graphite anode. In Figure 19 is illustrated the first experiment and the effect of the peroxide was to increase the rate of renoval of iron. Figure 20 denonstrates the specific effect of hydrogen peroxide on ferrous ion. It is apparent that larger initial additions of peroxide should have been made and later ones reduced in size. Ten milliliter additions of 30 percent hydrogen peroxide were added every 10 ampere hours .per gallon up to 60 ampere hours per gallon, S milliliter additions- at 60 and -70 amperehours per gallon, none at 80 and 90 alpere hours per gallon. -1 halite precipitate as noted at 30 ampere hours per gallon which was probably ferric mdroxide which can precipitate to sons extent at this pH under oxidising conditions. MG./L. IRON CONCENTRATION , 50 l2C’ IOO 80? 60f — 4. j; 20* O 20 4O 60 80 IOO AMPERE HOURS PER GALLON FIGURE 19. ”EC? or HURON P31011133 OR THE WROLITIC RMOVAL OF 1301 FRO! A 2.1 pH NICKEL-COBALT SLUTION. CURVE 1 - WITHGJT HIDROGH PEROXIDE CURVE 2 - WITH moms PEOXIDE F; k2. -n-s : ‘3‘; IRON CONCENTRATION , MGJL. 51 IZC O IOOF———- 80 i n O T__ ID l l O 20 40 60 80 IOO AMPERE HOURS PER GALLON Plums 20 mm or mam Pmoms ON THE summon or renews ION DURING mamas AT 51°C . CURVE l - TOTAL IRON COIEMRATIOR CURVE 2 - FERGJS ION CONCENTRATION 52 The upmrd swing of total and ferrous iron at the end of the experiment was probably a redissolving of this precipitate. The graphite anode experiment met be considered only as a guide to future experiments. Figure 21 shows that no ferrous ion was allowed to fore in the presence of an active graphite anode. However, the pH was lowered to 1.5 electronetric very quickly. Also the tremendous ' excess of chlorine in the solution caused dark powdery deposits with only a 25.7 percent cathode efficiency. The nickel content of the solution was at the sac tile reduced from about 90 g./l. to about 60 g./1. A graphite anode in commotion with a nickel anode but with different voltage sources night be balanced to losep the ferrous ion content down tdthout intedering as, such tdth the ordinary sequence of deposition. Conclusion -- The effects on the physical preperties of nickel deposits tron solutions conteining aluminum and iron seen to be caused by hydroxides in solution or in the cathode filn no not by the inclusion of almimn or iron atone in the deposit. , The following tab1s_indicates the changes in the physical properties of the four types of deposits as iron and alminun are added -to the solution. So new different physical properties being affected the same by both iron and aluaimn is lore thenooincidenoe. The studies of the colloidal structure, formationrnndphysical properties of iron and alanine: hydroxide have also been shown to be similar in most respects. Thus if aluminu- affects the deposit without being occlllied or co- dOpacited then iron may well do the ease. Codeposition of the iron nay CONCENTRATION , MGJL. IRON I20 F IOO 80F— GO 20 (LI—r O 20 4O 60 80 I AMPERE HWRS PER GALLON FIGJRE 21. WET OP WIPE MODE (ll WHIC RHOVAL 0? IR“ FRO! A 2.1 pH NICKEL-Cam MIDI. CURVE 1 - MAL m CWCDITRATION CURVE 2 - ruinous ION CWCDITRA‘I‘IOR 514 A . t —: em——#Tm Vatts 5.2 Watts 2.2 , N100 Org. __ fi _ W i L .11 _ _:‘_ i -—; t ' Iron whitened mottled no effect no effect Appearance _ w w M ECW— Aluninun " no effect " . " : _._ ._ E .1": . . + —— 5 Iron __ -no change no change no change no change F Adhesion *f ‘ +— _‘ v; ‘ f j um I I I C L}? W321; w _— ‘3:— - * '13:...- Iron slight- slight - slight slight Ductility decrease __ decrease decrease decre;a_s;_e__ Ali-im- no change no change no change no change 1 Iron increase increase increase varies up Hardness “ ~— and “9“ Aluminum decrease varies up initially; decrease increase and down then increase at 100 gm . Throulng Power A no change Iron slight. -- no appreoi- slight . - decrease able change increase Aluminum- decrease in no change slight no change medium. con- except at decrease trations. 10 ppm M —— In only two- instancesxe the results. diamtzicflly oppOsed. In all other cases the results are similar or exactly the same. 55 tales place £33; a cathode fill reaction which alters the physical properties of the deposit. “bother this is an adsorption phenomenon or a shielding mechanism camot be ascertained from this data. The one place shore alumina and iron differ completely in all four solutions is in the rates of moral by electrolysis ._ iluninm simply does notdeposit from the solution. In fact, it mold seen that inclusion does not occur to a measurable extent. The removal of both ferric ironand aluminum can be easily effected by high pH precipitation_and subsequent filtration. Aluminum cannot be removed electrolytically at high or low current densities from these four solution. Iron nay be removed electrolytically but the process is cloned by the reduction .of the ferric ion at the nickel anodes. The optima conditions for the electrolytic removal of iron from low pH solutions are 10 amps per sq. ft. current density, 15 feet per minute of solution agitation post the cathode, a minimnn anode area, and a 1/2" indented corrugated copper cathode. Above 3.75 pH electrouetric, iron will be continuously precipitated from solution as the hydroxide. -Peroxide is an aid to both precipita- tion and to the electrolytic removal of iron in that it keeps the iron in the state of higher oxidation. 56 RUMORS (De-131.93 AL, American Electrophter's Soc. Report Serial ~ No. 23 (1953). (2) Ewing, D. T. and H. D. Gordon, American mectroplater's Soc. Report Serial No. 15 (1990). (3) Wattshm O. P. “HonthlyRev An. Beetroplater's Soc. 111,110. 14, 5-11 (h) Ehrr, R., Transactions Am. Electrochen. Soc. §_§, 1125-1111 (1935). (S) O'Scllivan, J. 3., Trans. Faraday Soc., _2_6_, 533-9, 5110-3 (1930). (6) Puri, v. s. and 0. .Juneja, J. Indian Chen. Soc. 11 (19140) c. A. 133, 5710 (19112). (7)3a1anki, 0. u. 6and51nfh, D.J.,Indian0haa. Soc. 18, h23-hs (19141) C.A.i6_,hh21 19142). (8) Martin, 3., Proc. Am. Electmplater's Soc.-, June 33, 206-17 (19th). (9) Mrs. D. 1. :1 3., Plating 112 I12. 1391-11400 (1953). (10) Shim-r, 0., Kolloid 2. 2, 25-31 (1930) c. A. gt, A686 (1930). (11) Quintin, Harguerite, Compt. Bend. 2 2, 1303-5 (1951). (12) Stark, H. 11., .1. Am. Chem. Soc. 2, 2730—142 (1930). (13) Rot an, A; L v 1a. Zel'des. Zhur Priklad Khim. 2 , 717-23, (1950 .811, {am (1950). (114) Jaqueline-Longuet-Escard, Odette Bagno, Men. Services chim. etat (Paris) 3331, 77-9 (1951’), c. A. A_7_, 380 (1953). (15) Tominosuke Daturai Chem. Soc. Japan _6__1, 255-6, (19110); c. A. 16, 6930 (1952). (16) mrkevich, J., J. Hillier, Anal. Chem. 21, 1:75-85 (191:9). (17) Amt, M andJ. Ducllaux J. chim. mm 6, 252-62 (1939); c A 3);, 2678 (19110). (18) Kohlschutter, £1.14. and a. Nitschnann, z. anorg. angels. Chem. _2___29, 115-8 (1936), c. A. 3;, 25111 (1937). r .._ g ‘8' e . -.—-¢ 57 (19) Mokrushin, s. 0., J. Gen. Chem. USSR _1__6, 11-16 (19116), C. A. A0, 6935 (19116). (20) Thiele, a. and G. Kienst, Kolleid 2.127.13h-hh (1952), C. A L7 25 (1953). (21) Feitknecht, w., Helv. Chin. Acte 25, 555-69 (19h2). (22) Mittra R. 11., Pros. Nat'l Aced. Sci. India6, 321-32 (1936), ..C A. 3;, 5655 (1937). (23) Gr. Balanescu and Vintila T. Ionescn, Bul. Soc. Chim. Romania _2___OA, 139-83 (1938); c. A. 31,, 2229 (19110), (21;) Ibid., 19A, 93-131 (1937), c. A. 3;, 21 (1939). (25) Stromerg,1. G., R. V. Dityatkovskaya Lab. _1A_A_, 919-25 (19118), C. A. _3, 25191 (19119). (26) Yajnik, N. A. ,.P 1.. Kafur ..L Malhotra, Kolloid z. 80,152-5 (1937); c. A. 3;, 77201937). (27) 01am, 1.,M. Bezdek, Chen. Lietyhh, 300-5 (1950), C. A. 95, 3682 (1951). (28) Heiener, H. J., Colloidal State and Precipitation of Certain Hetallic Hydroxide: in Concentrated Solutions of Nickel Sulphate. Thesis for Ph. D. Degree, Michigan State College , 19h}. Milovanova, Zarodskaya (29) Dow, w. 0., meetroixtic Separation of Iron tron Concentrated . Solutions of Nickel and Cobalt tor the Electrodeposition or Nickel. Thesis for the H. S. Degree, Michigan State College (19119). (30) Serfass, n. J. am w. s. Levine, Monthly Review 2;, 1189-97 (19116). (31) Nyquist, R. 1., Eiectro1ytie Behavior or Iron in a Buffered Nickel fnlpflhgte Solution. Thesis for M. S. Degree, Michigan State College, 19 i. C) USE 0‘94)“- hr A w..—