THE .ELECTRC'LY'UC DEPOSITION CF COLL-CIDAL BERYLLIA C‘N METAL SURFACES Thesis for die Dagree cf M. S. WCHIGAN STATE COLLEGE - Cecr‘ge W. jernstedt 1940 .‘luuu \‘u’fi. . DI at.» . .l. ~00- .‘li- .0 ~ . a -‘< r-~ 5A “ . {“35 _:‘ ‘1.u/“ X- \p. , I I 1 'I- ..y f .‘ _..,. , . tv‘.'_ "A! , A - . ‘11 ‘L I I I' ~..‘. ..- ‘ 'iv-r; UL 'f‘ st, 34. is ”..‘v‘ '- I ‘ < a. . . ~ ‘ -:¢’€ L ;_ 1. h‘ I f’_ ' I ' ." 'I‘. - a 5‘ J r" .».~ '. 1:62'.‘ 4 NwIQ . “.r-ri 'gf.""". v ‘. ¢ ‘_ a] 4. K 4. ‘00 ' ‘ V -.0'.’ fi‘ 1 . .l ‘ .- ;r-:“’\ «,2 - cg; '_ - ., A ’ L _. .gf 4...",V.,,;~_-._ .. . _ J‘ r’: .' ‘ph'. 1.x 0 ‘n;.:d“' ‘133'3'2; «- ‘ .f H": .- .4 ‘. "‘ 2.4-: ‘1: a vfiwssvvzz .fi'fwg -. I. . 4:11: we «Wax 2+ .. 5 .‘fsw’z 3.55.1?“ =3; V t V' ‘ . . I “L" ’ " “7'1: “‘ "‘ l ““5? ‘y’I‘Wys’fi-Qfl’i ‘ ( .‘1:%fi. ‘ ‘43“; 6. ( '! g '6." "4f 7" L“: A . ~ . - 3"; . ..' . taxuh‘tt). ' ”.4496 V;.’ t"h¢. ,. ’3 37%.». ’ T - . 2f ‘, “3.7.1" .' w. . a *4 I, 0".,' -'-\‘ ' ‘V‘ 1-, .€ “ 0’ ‘ ‘J it)": LE“ if": _' _ . _ ' n . ‘ . . .‘k —| 2") "I _ .. “I :‘At W. _~' 1 . : ' w. . SKI"; ‘3‘..'7 Ik' 1‘.’ _?I: j ‘. ,Ith; 3"}ifi‘.‘ J .4 ,r‘ 'v.0_ <. - . .’ A ‘ ‘f' . . A. "I” ’5 2;. I‘Y’fi ‘ i ‘ .- ' . ‘ "'.I‘-' ‘l" ‘ . ‘~ . .V." ‘ ., - < . -‘ ~ '.‘x‘ h- -' ' " : W .‘. g, r I . ~£f~wu , _. ;. .- - _. ~ . It.” {,- ; H ’ A "J. ' '9‘ . ' " .. “I. .. - . :’ I‘l“'|‘. "xii. 1 ‘ ‘1; | ‘v I “' o. r "Il-"’ , < ‘0¢_,.&~ . - . .r'. l- " |‘. . ‘ 6/ I ~‘ ‘ ‘\ . , ‘1 tun-F 5: ‘ ‘ ,\’.‘,I‘:~~-:"'I:\4\‘¢.I.AV. ’ . 2‘ "fl. '1. l '. fi'gi.‘ V (O tang-eff .‘ f, 45%: 1°» dL:“ a. ‘ . .'_ ‘ e”; ‘t-L ‘ $I - . .» . .-‘. . H. "' .‘ :. "II. n ’ tF-s fiv-‘t‘ “2* ~ VI»? Ii?" ‘ ‘ ‘u. ’3 ' ' c 2‘ ""1 ~ h ”3 ,- _._ .: .4 ‘ ' .‘. {I .4 . ' W‘s ' ' .' m ’2‘ ‘ . #1:”. in. .‘.~’;‘ \ fie "3"" W53" “'5‘ 25 ft - “at" . ‘ 437"" ‘ lat-4|" ‘ ‘ ."-‘| . ' ' 1‘3 -'. A .‘Nt'i‘ )0 . x .' . ‘ 1 I‘.’ '1')". . ’1‘ _ . ~ -. ..' . é‘L 4: :.\\V ‘ . . n“ ~; -' -‘,-~». I ~ I ’ I - "v {I .. -. ‘; quflrrgx, ‘ J . . , - - - P . - ‘ ‘x - J. u _ V. IL; . . c .U' J‘Ma' ' “_ ‘ v? ...~ ‘~.' . ‘ 7:“.-44‘1‘4. map «gs/9’ )43q1‘5. . «a ‘V 1n‘v;-,'.w 1r 5...; - ‘ 1; ‘ ,7. {3.1.43.3 3’1 L31)”. if.4{£_lg; *. __,, r ‘ a.» . . . '.' ,‘. . 5 ..1 ‘1’“. ‘ .1 :.‘E . , . a —~—' I - ‘ I _ _ - '1'— .1 ,. I ‘\'}-‘ 1 _ :, _ . .- _‘_. -.'.._ Q I ,' “"'," ".TM _ L ,1- “? _. ; ”/— _..——- ,— - ——- =.i it .——--—:_— __.————— ,_..——-—-— ——' ___. ,_.- _ (D fa:— ‘ r- ’— _.__~_ ’— -—-—fi’-— 1 l A A U _H\l|ll|l||t + a o I ' Fig. 2. Wiring Diagram (A) Cathode bar (E) Milliammeter (B) Lead-tin alloy anodes (F) Voltmcter (C) Colloidal plating solutions (G) Slide-wire resistor (D) Temperature control tank (H) Edison cell available 7 per cent tin - 93 per cent lead.alloy as generally employed in the chrome-plating industry. The anodes were allowed to stand in water for 24 hours and were then electro-cleaned before use. Solutions in.which these anodes were used for over a month, did not show the presence of either element even when analyzed by spectrographic methods. Duplication of Past Work: Next, it was desired to reproduce the results which were already reported on this type of coating. A concentrated solution of beryllium.sulphate (10 per cent by wt. B6504.4H20) was prepared by dissolving the crystalline salt in dis- tilled water'and.then filtering several times. The solution was allowed to stand.over night and then de— canted, this being done in order to remove as much as possible, any suspended matter which might absorb a charge or be absorbed. It was noticed that there was an appreciable amount of the anhydrous beryllium sul- phate which settled out. This would definitely inter- fere in the electrodeposition of a colloid. Spectro- graphic analysis of the concentrated solution showed that there was a small amount of aluminum present. This was not detrimental since it was shown that salts of aluminum (1) react mmch.in the same way as salts of beryllium. The stock solution, thus prepared, was then diluted to a concentration of 3.4 g./l. as suggested by Thomas -5- and Price (1) and ammonium hydroxide added until the pH was between 5.75 and 5.85. The precipitate which formed was given 1 hour to settle. Approximately 100 copper plates were electroplated in solutions as indicated above, the results proving that excellent protection could be secured, but that the appearance was not entirely satis- factory. The entire range of pH from.3.4 to 7.0, as shown in figure 1, was investigated with results showing that only between a pH of 5.5 and 5.9 could there be a protective coating secured. This coating in all cases had some degree of coloration which was attributed to the thinness and un- evenness of the film as described in detail later in this work. As the ammonium hydroxide is added to the beryllium sulphate solution, in making up the bath, a gelatinous precipitate results which.settles out in about 30 minutes. As shown later, colloidal particles of beryllium.hydroxide remain in suspension. With this in mind, it was believed advantageous to investigate the effect of higher tempera— tures, hoping to increase the speed of travel of the colloid which is ordinarily relatively slow. Temperatures up to 55°C. were tried with but a slight improvement in appearance and a decrease in the protective powers of the film. After 2 one-hour periods at 55°C., the solution no longer contained a colloid. This was evidenced by a change -7- in pH from.5.75 to below 5.00 and also the fact that no protective film could be secured. It is probable that the colloid lost its absorptive powers and precipitated. Weiser (3) states that "Like most gelatinous oxides, the absorbability and solubility of hydrous beryllia decrease slowly on standing at room.temperature and rapidly at higher temperatures, particularly if heated in a current of steam or in the presence of a solution of ammonia or of alkali hydroxide or carbonate." Agitation of the solution at room temperature did not improve the appearance of the film, but slight movement of the plate itself did help remove any bubbles of gas or the larger particles which were carried over. Original Research: It was decided that sufficient work had been done with the solution as established by Thomas and Price (1), and that if a suitable coating were to be obtained, it would be necessary to make some funda- mental improvement in the process itself. One of the difficulties encountered was the mainten— ance of the pH with.any degree of accuracy over an oper- ative period of several days. It was always necessary to add ammonium hydroxide and this addition was critical since it would require smaller and smaller quantities to arrive at the correct pH value. In the event this value was over- stepped, the solution had to be discarded, since the colloid was destroyed. This impediment was the subject of the -8— first part of the investigation to effect a change in the process. The most natural correction for the instability of the acidity of a solution would be the addition of a buffer. Several were tried, including ammonium acetate, ammonium formate, ammonium sulphate, and boric acid. The effect of the first two was not noticeable in concentrations below 5.0 g./l., and above that, it became increasingly more diff- icult to get any film deposited at all. In the case of the ammonium sulphate, results were the same except more pro- nounced, probably because of the common ion effect. Boric acid proved to be quite different. In concentrations as low as 1 g./l., an improvement in the appearance of the coating was apparent, but the buffer action was not very effective (see figure 1). Addition of Boric Acid: In figure 3 are shown the re- sults of the addition of varying amounts of boric acid. It will be noted that as the quantity is increased, there is a corresponding decrease in the total rise in potential for a given period of time. The plates from which these data were taken all showed better than 95 per cent protec- tion against ammonium.polysulphide. As the content of boric acid increased in the bath, there was a notable de- crease in film coloration. At 5 g./l., plates could be obtained which showed no indication of interference bands. At 10 g./l., the potential rise was almost the same as at Volts Rise per Plate 206 2.4, 2.2 2.0 1.8 1.6 1.4 fl {3* (A) V 4? /’/' 2 (E)‘ _g / i / ////?”// 0) 59¢ h l// 1///::::A O l 2 . 3 4 5 6 Time in Minutes Fig. 5. Effect of boric acid on potential rise. Solutions of (A) 0.0 g./l. H3E03. (5) 1.0 g./l. 55505. (c) 5.0 g./l. 55505 5 g./1., but the deposit was cloudy in appearance. This work was duplicated several times and 5.0 g./l. of boric acid was taken as optimum concentration. Later, it was 1 shown that 5.0 g./l. could be employed with Just as much effect if the solution was allowed to stand 48 hours before use. This allowed time for the larger particles to pre- cipitate and collect on the bottom. From figure 1, it is apparent that the buffer action of the acid consumes only a small quantity of the ammonium hydroxide added to obtain a pH of 5.70. The shape of the curve has not changed appreciably, so that quantity could have been taken up in just neutralizing the acid. The action of the boric acid was not that of a buffer. An attempt was made to deposit boric acid itself,‘but without any success. It was not precipitated at all by the ammonium hydroxide. In figure 4, the change in pH with concentration of boric acid is shown. Figure 5 is the same for beryllium sulphate. Another phenomenon which was observed as an action of the boric acid was the change in pH on standing. This is best shown in figure 6. The slight drop in the case of solutions not containing any boric acid is due to the de- crease in the absorbing power of the colloidal beryllium hydroxide according to Weiser (3) as stated previously in this work. Solutions containing 5 g./l. of boric acid ex- perience a steep drop in pH during the first 48 hours from -10- 6.0 5.0 \‘k\\\ pH '\ \ 4.0 1 ~\“‘l(-I=:_—__ 3.0 O l 2 3 4 WGIght % H5805 Fig. 4. Concentration of boric acid vs. pH. 4.0 3.0 \ Nl\“n~.q,‘~ pH 2.0 1.0 O) a: 0 2 4 Weight % BeSO4-H20 Fig. 5. Concentration of beryllium sulphate vs. pH .comnm .H\.m o.n .Ho.o mo on .Ne.n an Aha .mm.n mo Adv .ommvvowom .a\.m w.m msficflwonoo wgowpdaow .oaflp spas mm CH omcomu .m .mam when a“ mafia m m w m N H o m.v o.m mm 1/ v.m m.m :1 an entirely different cause. It is believed that the boric acid causes the slow coagulation of the larger colloidal particles of beryllium hydroxide. That the boric acid is absorbed on the colloid was indicated by the presence of relatively large amounts of the compound in the precipitate which.settled out. After solution (C) from figure 6 had dropped to a pH of 4.9, a plate was coated at a current density of 80 ma./sq. ft. The appearance of the plate was excellent and its corrosion resistance 100 per cent according to pre- vious standards mentioned. It was noticed at the time of plating that the potential rise was unusually high for this current density. Several solutions of similar concentra— tion were prepared and plates coated as soon as they were prepared. The potential rise was recorded. Two days later the same procedure was followed on these solutions in order to observe the difference in potential rise. These data were correlated and appear in figure 7. In all cases, as was expected, the plates which gave the greater rise in potential for a given current density also gave higher corrosion resistance. This is in accordance with the work by wagner (4) which infers that the higher the electrical resistance of an oxide film, the greater will be its tarnish resistance. Conductivity Measurements: Next, the conductivity of the bath was investigated since it was desired to determine -11- Volts Rise per Plate 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 O 1 2 3 4 5 Time in Minutes Fig. 7. Effect of ageing bath on potential rise. Solution containing 3.4 g./l. EeEO oéHkC, 5.0 g./l. H7203, pH 5.61, C.D. 100 ms. / sq.ft. Curve (A) at time bath was made up and curve (E) 48 hours later. the effect of the boric acid. The amount of boric acid was varied in a solution containing 2.5 g./l. of beryllium sulphate. The results of the measurements are illustrated in figure 8. All conductivity determinations were made at 25°C. It readily can be seen that very little change takes place. The conductance of various strength beryllium sul- phate solutions was also observed and a typical curve for salts of that type resulted as shown in figure 9. The data reported in the previous paragraph served their purpose, but really were not very enlightening since actual operating conditions are at acidities'between a pH of 5.0 and 6.0. In order to determine the conductivity over this range, it was measured as the ammonium hydroxide was added. The pH was determined simmltaneously and the results plotted in figure 10. The increase in conductivity which appeared in figure 8 can be seen to diminish as the pH rises, or in other words, the amount of ammonium hydroxide becomes greater. One other important point to be observed is that the con- ductivity does not change appreciably beyond aij of 6.0 which shows it is not a function of pH. This is no doubt due to the fact that very little base was added to incur this change and that the base is actually carrying the greater part of the current. This explains why the current does not go to zero as the potential rises during plating. The colloid obviously could not be carrying 100 ma./sq. ft. -12- Grams per Liter Grams per Liter 8 7 / e / 4 [T l// 2 I 0 , .0017 .0018 .0019 .0020 Specific Conductance in MHOE/cm‘ Fig. 8. Change in conductance with concentration for boric acid in O.25%-beryllium sulphate solution. 10 A///fy / 8 Z///r /’ 6 ////A 4 // . f 2 i .I/ / 0 .001 .002 .005 .004 .005 .006 Specific Conductance in MHOS/cm2 Fig. 9. Change in conductance with concentration for beryllium sulphate solution. 7.0 6.5 6.0 ///‘ 44// 5.5 ////’ pH 5.0 l/ m / 4.5 /’ j /(5) 4.0 / 5.5 / // I 500 ' .0022 .0026 .0050 .0054 .0058 .0042 .0046 Specific Conductance in MHOS/cm2 . 10. Change in conductance with pH. (A) 3.4 g./l. Fl 1:. O' C) C u he 04-4H20. (B) 5.4 g./l. EeEO4o4HL0, 5.0 g./l. H5506. at peak voltage. Other data which were recorded for pH's beyond 7.0 showed that the conductivity again increased as the ammonium hydroxide was added. Action of Citric and Tartaric Acids: Since the boric acid was rather unusual in its action, it was thought that perhaps compounds such as citric acid and tartaric acid might also be worth investigating. Quantities of 5 g./l. of either compound completely complexed the beryllium sul- phate and no precipitation resulted even when great amounts of the base were added. The effect of adding 1.0 g./l. of citric acid was determined by potentiometric titration of a solution containing 3.4 g./l. of beryllium.sulphate. A slight break in the curve is detectable (figure 11) and very slight precipitation occurred. Upon electnflyzing the solution, no rise in potential was obtained nor was there any evidence of deposit. It is highly possible that there was some free beryllium.sulphate in the solution which went to the hydroxide as the base was added, but action of the complex salt was not conducive to the formation of a colloid. Investigation ofng Rang_: The pH range of a bath containing only beryllium sulphate and ammonium hydroxide was reported (1) to be approximately 5.6 to 6.1. In the process of investigating this solution, no protective coat- ing could be secured above a pH of 5.90. Plates coated at a pH of 5.95 failed 100 per cent in the sulphide tests. -15.. 8.2 7.4 6.6 5.8 pH 5.0 4.2 3.4 2.6 4 8 12 16 ML/L of 3N NH40H Fig. 11. Potentiometric titration curve for a solution containing 3.4 g./l. of EeEC4-4Hk0 and 1.0 g./l. of ‘citric acid. At the other extremity of the range there was not such a definite break. At a pH of 5.50, excellent plates were secured, but at 5.40, the protective value of the coating fell off to about 50 per cent. This was true only of the baths not containing any boric acid. However, when this agent was added to the bath in quantities of from 3 to 5 g./l., the initial adjustment of pH had to be between 5.5 and 5.9, but as the solution was operated, the pH could drop as low as 4.9 without any decrease in the pro- tective powers of the film deposited. The upper limit re- mained at a pH of 5.90. Effect of Concentration: In determining the effect of various concentrations of beryllium sulphate solutions, the potential rise per plate was again recorded (figure 12). The rate of rise per plate varied with the change in concen- tration, but the peak voltage was practically the same in almost all cases. Protective coatings, however, could be secured only over a limited range, as illustrated in table I. The plates from which these data were taken were plated at a current density of 100 ma./sq. ft. for 6 minutes. Table I Conc. BeSO4.4H20 Amt. of Tarnish 0.5 g./1. 95% 2.0 " 5% 5.0 I 0% 4.0 " 0% 5.0 " 2% -14- Volts Rise per Plate 2.2 2.0 1.8 1.6 1.4 1.0 0.8 0.6 <2. ._._ ////%W°:/;5 0— X / i/ ' / / fl/ / { Time in Minutes Fig. 12. Rise in potential vs. concentration of beryllium sulphate solutions containing 5.0 g./l. H5505 at pH 5.62 and a current density of 100 m&./sq.ft. (A) 5.0 g./l. (B) 3.4 g./l. (C) 2.0 g./l. (D) 0.5 g./l. Several observations may be made from these results. First, the peak voltage reached is not necessarily indic- ative of the protective value of the coating. Second, the concentration of beryllium salt is not critical and a sat- isfactory range would be 3 to 4 g./l. It should be noted that this is in agreement with the value of 5.4 g./l. as given by the published work of Thomas and Price (1). Effect of Current Density: Current density also proved to be a factor which influenced the nature of de- posit. From figure 13 it can be seen that the peak voltage becomes greater as the current density rises. Again it was made apparent from tarnish tests (tables II and III) that the peak voltage does not necessarily indicate the nature of the film.which has been deposited. Table II Bath conditions -- 2.0 g./l. B330 .4320 and 5.0 g./l. H5803 at a pH of 5.37 Current Density Amt. of Tarnish 20 ma./sq. ft. 0% 55 I 0% 100 I 5% 150 I 50% 190 " 95% 240 " 100% Table III Bath conditions -- 3.4 g./l. BeSO .4H20 and 5.0 g./l. H5803 at a pH of 4.30 Current Density Amt. of Tarnish 20 ma./sq. ft. 0% 50 " 0% 100 I 15% 195 I 25% -15- Volts Rise per Plate 2.6 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 (A) (B) EFL . (C) (D), Fig. 13. H ( 5 at pH 5.07. (A) 240 ma./5q-ft-. 5 4 Time in Minutes Effect of current density on rise in potential for solution containing 2.0 g./l. LeSOA-éHPO, 5.0 g./l. - (B)V190 ma./sq.ft., :20 t) 150 ma./sq.ft., (D) 100 ma./sq.ft., (E) 55 ma./sq.ft. The current density range is greater at the optimum concentration of 3.4 g./l. beryllium sulphate than at 2.0 g./l. as shown in tables II and III. Actually, plates were coated much lower than 20 ma./sq. ft., but the time of plating was longer and the relative value of the read- ings was lost. The optimum range would seem to be 50 to 100 ma./sq. ft., considering both the protective value and time of deposition. Time of Deposition: The time of deposition is another factor in the consideration of protection and appearance. Obviously a heavier deposit will give greater protection, but unless the particles have been restricted somewhat to the smaller, secondary sizes, the interference colors will then be present. In any event, the thickness is not a linear function of the time of electrolytic treatment. As would be supposed, there is little change after the max- imum potential has been reached and this usually takes from 5 to 5 minutes, depending upon bath composition and current density. From a beryllium sulphate solution containing boric acid, plates can be secured having maximum thickness but no coloration of film. If the bath contains only beryllium sulphate, the interference colors start to appear within 2 minutes of plating and less than the maximum thickness. Heat Treatment: While determining the proper method for drying coated plates, it was discovered that boiling -15- water would remove the film. This effect gradually dis- appeared as the film aged. Plates tested after standing at room temperature for a period of two weeks were not affected by the boiling water. The test consisted of sub- merging the lower half of the plate in.boiling water for one minute and then subjecting the whole plate to the sul- phide fumes to determine if the film.had been removed. It was found later that films which had been given the oxida- tion test at 500°C. for half an hour also were not affected by the water. Correlating all the data presented in the preceding paragraph, the necessity for a heat treatment was readily seen. One of the limiting factors which immediately be- came apparent was the effect of the heat treatment upon the base metal. Most metals would be annealed after one half hour at 30000. An investigation was made to determine the proper heat treatment to give the film.stability against the boiling water. As little as 1 minute at 500°C. was found sufficient to render the coating inert to this test. At 25000., approximately 2 minutes were required. In attempting to strip the film from some cepper samples which had been coated and tested for resistance against oxidation at 50000., it was found that the film was apparently inert to the acid bath. The material had been finished in the regular manner and heated -17- for one half hour. The acid strip bath consisted of half-and-half concentrated nitric and sulphuric acids. A detailed investigation disclosed that this unusual resistance was not developed until the film was heat-treated for 8 to 10 minutes at 500°C. Heat treatment for less than that time would not give copper any resistance whatsoever against the acid dip. Heat treatment for longer periods of time did.not increase the resistance of the film which was approximately 2 minutes in the strip bath. Plates coated in solutions not containing any boric acid were re- sistant for about 1 minute and it was noticed that the heat treatment removed most of the interference patterns from.these plates. This indicated a change in composition of the film, and an explanation of this change is advanced under the theory of the process. Thickness Measurements: Thomas and Price (1) cal- culated a theoretical value for the thickness of this type of coating. Basing their calculations upon the deposition of either the oxide or hydroxide and a 100 per cent effic- ient process, they reported 50 A. If this were true, the film would.be in the order of only one colloid particle deep. Desiring at least to determine the order of the thickness accurately, a method of physical measurement was sought. The only instrument suitable to measure such small films is the interferometer. Several extra mirrors which -18.. had been employed on a Michelson interferometer were re- silvered by the rochelle salts reduction process and washed in distilled water. A small phosphor-bronze clamp was made to hold a single mirror, Just making contact with the silver film along the edges. While the mirrors were still wet, they were placed in the fixture and the lower half of the surface coated with a film.of beryllia, the first for 4 minutes and the second for 6 minutes. They were dried in alcohol and ether and then maintained at 22°C. until the measurements were taken. .An attempt was.made to adjust the current density to approximately 100 ma./sq. ft. When the mirrors were located in the interferometer and properly adjusted, a displacement of one quarter of a wave length of sodium light was detected on.both mirrors. These measurements indicated two facts; first, that the thickness was in the neighborhood of 700-800 A., second, that the thickness did not increase with time. The latter factor was quite obvious from the very nature of the process. Other Metals and Solutions: At first, only copper was employed as a base metal. Later, when the process was sufficiently advanced, silver plates were coated also. The same protection was secured, but in some cases, a slight interference coloration appeared if the plates were coated longer than the 5-minute period. This was always entirely removed by the heat treatment. -19- Nickel-plated, brass, and steel articles were also coated with films of beryllia. There was evidence of the protective powers of the coatings but no quantitative in- vestigation was made. A bath employing aluminum sulphate instead of beryllium sulphate was prepared and several copper plates coated. The same protection was secured as in the case of the beryllium salt. No detailed investigation was made. The impurities which will affect this type of bath most are copper, silver, and iron. This is because they are attacked.by the solution and go into solution. Their oxides are not protective and make the resulting film porus. Spectrographic analysis disclosed that simply hmmersing a copper plate or wire in the bath was suff- icient to contaminate it. CHARACTERISTICS CE FILM A thickness of 700 to 800 A. is approximately 0.000005 inch. That is rather thin for a protective coating, but tests have proved that the film.can be very effective even in such thicknesses. If one of the copper plates used for the experimental work were coated with beryllia and then heat-treated for one minute at 30000., it would not be affected by H28 fumed for 5 to 4 hours. Copper may be protected from oxidation at elevated or room temperatures in this manner also. Test plates heated at 500°C. to 550°C. for 5 hours showed no oxidation to the copper or any discoloration to the film. There are 5 types of coatings obtainable by varying the heat treatment; first, no heat treatment at all, second, 1 minute at 500°C. or its equivalent of 2 minutes at 25000., and third, 10 minutes at 500°C. These films may'be characterized'by the following tests: The first will withstand the action of H28 fumes but will not be stable to the boiling water test; the second will give better resistance to H28 fumes and will withstand the ac— tion of boiling water but will not give any protection against the concentrated acid mixture; and the third.will be the best film, not affected by the boiling water test and not attacked for about 2 minutes in the concentrated acid mixture. An oxide film.three millionths of an inch thick would have one inherent weakness when applied on the surface of a different metal. That weakness is its low resistance to mechanical abrasion or scratching. This film.will with- stand a great deal of handling and rubbing with cloths, but the continuity of the film is destroyed.when an abra- sive scratches the surface. In general, it may be stated that the adhesion of a heat-treated sample is excellent in spite of this condition. -21- Since the deposited particles of colloidal beryllia have a high electrical resistance they are not deposited to any great depth, but this has one advantage and that is the increase in throwing power. The film.builds up uni- formly even in recessed areas. This fact was evidenced when the plates were put in the concentrated acid mixture. The copper was attacked evenly over the whole surface. Part of this was no doubt due to the uniformity of the colloid particles which had been deposited. THEORY It is important to consider what happens to a solu- tion of beryllium.sulphate as ammonium hydroxide is added. In figure 1 will be seen the change in pH of such a solu- tion as the base is added. This same effect was reported by Britton (2) and Thomas and Price (1). Britton's work was a little different in that the base was sodium hydrox- ide, but the resulting curve was much the same as that found in figure 1. The effective range for the initial pH adjustment is 5.5 to 5.9 (figure 1). This may be called the flat por- tion of the curve. Britton maintains that the composition within this range would be BeSO4.Be(0H)2. This is no doubt true, but is not a clear enough picture because no mention is made of the colloid particles. A more complete analysis is suggested below. -22- As ammonium hydroxide is added to a solution contain- ing between 5 and 4 grams of beryllium sulphate, precipi- tation does not start until a pH of 4.8 is reached. As more base is added the pH rises more quickly and.precipi- tation continues. This increase in rise is due to the ab- sorption of BeSO4 on the precipitate to form.what is equivalent to the compound BeSO4.Be(OH)2 as advanced by Britton. That hydrous beryllium oxide is highly absorp- tive, is indicated by Weiser (5). As still more ammonium hydroxide is added to go beyond, say a pH of 5.5, the BeSO4 is removed from.the precipitate by neutralization and the flat portion of the curve results. The equivalence point is reached at a pH of 6.0. Now the reason that it is impossible to deposit a protective coating as soon as a precipitate is formed may be because there has not been any peptizing action to form a colloid. The precipitate itself will migrate towards the cathode but will not form an adherent deposit. This was shown in cataphoresis experiments. The charge is pos- 1tive from.a pH of 4.8 up to at least 9.5, but measurements were not made any higher. Once the colloid has been fonmed (pH 5.5 to 5.9) it will remain stable even if the pH drops back to 4.8. The protective power of the film.deposited begins to fall off after a pH of 4.9 has been reached. The previous paragraphs alone still do not prove that there is a colloid in the solution. The one pertinent fact -23- which does, however, is that the Tyndall phenomenon may be observed only when the solution can be used to secure a protective coating. By continuous plating, a bath may be deplenished of its colloid and then no Tyndall phenomr enon can be observed. The effects of boric acid may be several in number, but it does serve at least one important purpose. That purpose is to cause the coagulation of the larger primary particles and leave the secondary particles for plating. much evidence is advanced by Thomas (5) for this grouping of colloidal particles into primary and secondary classes. Weiser (6) goes so far as to present the actual sizes of the particles in each group. The fact that these larger particles are removed is evidenced in at least three ways. First, because during a 48-hour period of standing, these particles gather in clusters large enough to be seen; second, because the de- posit is thinner as shown by the disappearance of the interference colors; and third, because of the higher potential rise per plate for a given solution (figure 7). The potential rise per plate does not necessarily in- dicate the protective powers of a coating. This was shown in the investigation of both the concentration of beryllium salt (figure 12) and the current density (figure 15). How- ever, the work of wagner (4) still holds and it may be said that for a given set of conditions, i.e., bath composition -24.. and current density, the higher the potential rise per unit area the more will be the protective value of the film, If the particle sizes as advanced by Weiser (6) are correct, that would explain why Thomas and Price (1) were unable to eliminate the interference colors. Particles equivalent to the primary sizes were given dimensions of 5 by 10"5 to l by 10"7 cm. This being true, it would be possible for films just 2 or 5 particles deep to become visible, interference colors starting at a depth of the order of 10'4 cm. The secondary or molecular dispersed particles were given a range of from.lO'7 em, down to the size of the molecules themselves. It is quite obvious that the secondary particles could build up a much more dense and protective film. It might be thought, at first, that the larger par- ticles could be removed by plating out. This is not the case. Thomas (7) in the following equation shows that the mobility of a colloid micelle is inversely proportional to the mass of dispersed phase per cc. of original hydrosol, but this does not concern the size of the individual colloid particles: of the original hydrosol - mass of dispersed phase trans— orted from.one compartment to he other specific conductivity of the original hydrosol U - mobility of particle t - time in seconds I - current in amperes U =‘%%t M - mass of dispersed phase per cc. m. w l -25- This fact was also shown in the experimental work on the solution as established by Thomas and Price (1). The bath gave interference colors even after many plates had been coated. The exact composition of the film which is deposited is difficult to determine. Spectroscopically, the film was shown to contain mainly beryllium'but also a little boron. There can be little doubt that the colloid particle is Be(OH)2. The charge is probably due to the absorbed BeSO4 on the surface of the particle. Some boric acid is also carried over with the colloid, but it would be diffi- cult to say whether it was absorbed in any quantities or not. In any event, the film.deposited must consist mainly of Be(OH)2 or BeO.H20 as Weiser (5) puts it. When heated, hydrous beryllium oxide loses its absorp- tive powers and consequently loses part or all of the water molecule depending upon the temperature. Weiser gives data on temperatures up to 280°C. where all but 0.15 H20 has been driven off. He also makes the statement that moisture will again be absorbed unless the oxide is heated higher. These data would indicate that the film consists mainly of Be0.0.15 H20 after heat treatment at 500°C. For convenience it is well to call the substance beryllia, regardless of its composition. As mentioned before, films deposited in the manner suggested by Thomas and Price (1) contained interference -25- colors which were removed almost entirely by heat treat- ment. This can now be explained as the reduction in thickness due to the removal of 0.87 molecules of water from the film, No decrease in interference colors results from the change in index of refraction because the oxide has a higher index than the hydrous oxide. SUMMARY 1. Many metals, including copper, silver, and their alloys, may be treated cathodically in a solution of beryllium sulphate and.ammonium hydroxide to give them excellent protection against oxidation in air at elevated temperatures or tarnishing from.su1phur fumes. 2. Deposition is made from a bath containing 5 to 4 g./l. of beryllium.sulphate nearly neutralized with ammonium hy- droxide. The film is probably composed of colloidal beryllium hydroxide which.has a positive charge in the solution and migrates towards the cathode when a potential is applied. 5. The addition of 5 to 5 g./l. of boric acid to the bath causes coagulation of the larger colloidal particles. This leaves the smaller, secondary particles to be plated out. A transparent and colorless film is thus formed. 4. General plating conditions include: anodes of 7 per cent tin and 95 per cent lead, a current density range of -27- 50 to 100 ma./sq. ft., a period of plating of 4 to 6 minutes, and a pH which is initially adjusted to between 5.5 and 5.9 and may be operated as low as a value of 4.9. 5. The film is limited to a thickness of approximately 5 millionths of an inch, this being attained after the maximum potential has been reached during electrolytic treatment. The film, although limited to this small value, will withstand rubbing with. non-abrasives. The main weakness is its low resistance to mechanical abrasion. 6. Heat treatment of the film will materially increase its protective powers. Two types of treatment are suggested depending upon whether it is desirable to anneal the base metal or not. The first consists of 1 minute at 500°C. or its equivalent of 2 minutes at 250°C.; the second consists of 10 minutes at 500°C. 7. The final film exhibits remarkable resistance to oxida- tion in air, to tarnish from sulphur fumes, and even will stand 2 minutes in.a concentrated mixture of half sulphuric and half nitric acids. 8. The film deposited consists of hydrous beryllium oxide (Be0.H20) and after heat treatment, probably has the composition BeO.0.l5H20. -28.. (1) (2) (5) (4) (5) (6) ('7) REFERENCES G. J. Thomas and L. E. Price, J. Inst. Metals, . Q5, 21, (1958). " " J. Inst. Metals, ‘gg, 29, (1958). " " J. Inst. Metals, .Qg, advance cOpy, (1959). H. T. S. Britton, J. Chem. Soc., 2121, (1925), 422, (1927). H. B. Weiser, Hydrous Oxides McGraw Hill Book Co., 1926, p. 159. C. wagner et al., Z. physikal chem.,.§1, 25, (1955) B. n I .gg, 447,(l956) B. I " 29, 455,(1958) B. A. W. Thomas, Colloid Chemistry McGraw Hill Book Co., 1954, p. 175. H. B. Weiser, Colloid Chemistry John Wiley and Sons, Inc., 1959, p. 2. A. W; Thomas, Colloid Chemistry McGraw Hill Book Co., 1954, p. 210. ’ -. -1116?!) . 127475 J55 T546 Jernstedt ”-1 L; Irmhr .. Pilll‘rr‘ nu? («th .- It I l .rqilr ivdrlul‘ n‘h .2- - .I~ :(h 1....1: . 31293 0244