The Adsorption of Complex Ammonia Ions on Silica Gel

Department of Chemistry, University of Kansas City, Kansas City, Missouri. Received ... The strong adsorption of complex metal ammines by silica gel h...
0 downloads 0 Views 451KB Size
T H E ADSORPTION OF COMPLEX AMMOKIA IONS ON SILICA GEL GRANT W. SMITH Department of Chemistry, University of Kansas City, Kansas City, Missour4 Received July 80, 1958

The strong adsorption of complex metal ammines by silica gel has been reported previously by Smith and Reyerson (5). By reduction of the adsorption products a t elevated temperatures by hydrogen, metallic deposits are produced on the surface of the silica gel. These substances have proved to be very active catalytic agents in hydrogenation and other types of reactions, and the metal films of silver and nickel have been used by Smith and Reyerson (6) in streaming potential cells for measurement of electrokinetic potentials. Studies on the adsorption of the copper ammine on silica gel were made by Kolthoff and Stenger (2), who found that a final equilibrium condition is reached only after very prolonged shaking, and that the composition of the copper ammine ion undergoes a continuous change during the process. Reyerson and Clark (3) give results on the adsorption of the complex ammonio ions of silver and copper on silica gel, and report the influence of the pH of the equilibrium solution on the adsorption of the latter ion. The present study was undertaken in order to determine the relative adsorption values of several of the common metal ammines, namely, those of silver, nickel, copper, and zinc, since this knowledge would be of great value in preparing mixed metal surfaces on silica gel for future studies in catalysis. I t was felt, also, that this adsorption phenomenon offered an excellent opportunity to obtain information regarding the nature of adsorption from complex mixtures and its relation to the adsorption of the individual components in simple solutions. The adsorption isotherms for these four ammines were first obtained. Xext, the adsorption of each component in mixtures of all combinations of two, three, and four of these compounds was determined. In these mixtures the initial concentrations of all components in each mixture were equal, and the total concentration in every case was originally 1.0 N . EXPERIMENTAL

The silica gel used in these experiments was obtained from the Silica Gel Corporation. It was treated with nitric and hydrochloric acids to 637

638

GRANT W. SMITH

remove iron and other impurities, then washed with frequent changes of distilled water over a period of several days, and finally dried a t 300°C. The resulting product was quite transparent and glassy in appearance. The solutions of complex ammonio compounds were prepared from the nitrates of silver, nickel, copper, and zinc, which were of the best quality available commercially (c.P. grade or better) and were used without further purification. Approximately 1.0 N solutions of each were prepared in the following manner. The desired quantity of the salt was dissolved in distilled water. To this solution concentrated ammonium hydroxide was added carefully until the precipitate a t first formed just dissolved, then an excess of 5 nil. per liter of final solution was added. The solution was then diluted with-distilled water to the correct volume. It was noted that practically identical amounts of ammonium hydroxide were added for equal volumes of solutions of the four metal salts. The resulting solutions were assumed to contain Ag(NH&+, CU(NH~)~++, Ni(NH&++, and Zn(h”&++, and were, respectively, colorless, dark blue, medium blue, and colorless. It is quite possible that the nickel solution may have contained a small amount of the hexammine compound, but the color of the latter (pale violet) was not evident. Bjerrum (1) has shown that these solutions are quite complex, so it is probably impossible to obtain a pure solution containing but a single compound. Solutions of lower concentration were prepared by diluting the above stock solutions. Each sample for study was prepared as follows: 15.00 g. of silica gel was weighed into a 125-ml. Erlenmeyer flask (coated internally with paraffin wax to prevent reaction between ammonia and the glass), and 75.00 ml. of solution was added. The flask was stoppered with a paraffined cork stopper and placed in a shaking machine (4). The experiments were carried on in a room in which the temperature varied little from 25°C. The methods of analysis of the original solutions and of the final equilibrium solutions were the following standard procedures: ( 1 ) Silver: precipitation as silver chloride and weighing in filtering crucibles of Jena glass. (2) Nickel: precipitation as nickel dimethylglyoxime and weighing as above. (3) Copper: iodometric titration in acetic acid solution with sodium thiosulfate which had been standardized with pure copper. (4) Zinc: titration in sulfuric acid solution with potassium ferrocyanide, standardized with pure zinc, using diphenylamine as internal indicator. (6) Separations: in the case of mixtures, silver was always removed first by precipitation as silver chloride; copper was removed (when present) by precipitation in hydrochloric acid solution with hydrogen sulfide; nickel and zinc, when simultaneously present, were then separated by precipitating nickel dimethylglyoxime in the presence of some excess ammonium

639

ADSORPTION OF IONS ON SILICA GEL

chloride, and subsequently precipitating the zinc in the filtrate with hydrogen sulfide, followed by filtration, solution, and titration. The amounts of metal adsorbed per gram of silica gel were determined by calculation from the decrease in concentration as determined in the analyses. Analyses of the silica gels after hdsorption were not made as some experimenters have done, for it was believed that it would be impossible to wash the gel surface (both external and internal) free of the solution without disturbing the adsorbed materid and introducing unpredictable complications as a result. RESULTS AND DISCUSSION

It was found that these systems, especially in the cases of the more concentrated solutions, were slow to reach an equilibrium condition. This is in agreement with the results of Kolthoff and Stenger on the adsorp tion of the copper ammine. The time required to reach equilibrium apparTABLE 1 Influence of the time factor on the amount of silver complez ammine adsorbed on silica gel I N I T I G CONCENmATION (NORMALITY)

FINAL CONCENTRATION (EQUILIBRIUM)

-______

MILLIEQUIVALENT8 ADBORBED PER QRAM OFBILICAQEL

millicpuitalen(i pm

milliliter

milliliter

millicquiodcnu

0.9822

0.7777 0.7948 0.7932 0.7931

1.021 1.019 1.027 1.035

millicquiuolcnh per

0.9988

0.9988 1 .ooo1

,

TIME OF AD8ORPTION

I

i1 1 I

~

hour#

4 5 9 77

ently depends upon the concentration, and also largely upon the complexity of the system in question, since it is comparatively small for the silver ammine, which is obviously the simplest case. Table 1 shows the amounts adsorbed after various periods of time from solutions which were originally about 1.0 N . Note that the amount adsorbed after 77 hr. is practically identical with that after 4 hr., especially considering the difference in original concentrations. These results also indicate that the method of preparing the solutions of ammines gives quite uniform products, since three separately prepared solutions were used here. The results of adsorption studies on four complex ammines (silver, nickel, copper, and zinc) are presented in table 2. In each case the values for the more concentrated solutions were obtained after 77 hr. of shaking with the silica gel. These results are presented graphically in figures 1 and 2, in which the isotherms are drawn. The nearly linear relationship shown by the graphs

640

GRANT W. SMITH

of the logarithms of the amounts adsorbed (figure 2) is in close agreement with the empirical isotherm of Freundlich. The change in slope of these curves a t the higher concentrations seems to be characteristic of all these systems. It might be thought to indicate that the adsorption had not yet reached a constant value, but this seems improbable in the light of the results of table 1. A wide variation in the specific adsorption values of TABLE 2 Adsorption of complex ammines o n silica gel Temperature, 25°C. (approx.) INITIAL CONCENTRATION (NORMALITY)

~

FINAL CONCENTRATION (EQunlBRInm)

j

XILLIEQUIVALENTS ADBORRBD PER o n m OF BILICA OED

I. Silver ammino ion, Ag(NH&+ milliaquirdmta per millilihr

1 .OOol 0.4911 0.3274 0.2455

milliequivdenb per milliliter

0.7931 0.3269 0.1898 0.1274

millicguia&b

1.035 0.820

0.687 0.590

11. Nickel ammino ion, Ni(NHs)c++

1 ,0070 0.4888 0.3259 0.2444

0.6545 0.1851 0.0697 0.0251

1.763 1.517 1.279 1.095

0.9886 0.4428 0.2972 0.2234

0.3488 0.0185 0.0035 0.0012

3.199 2.121 1.469 1.111

IV. Zinc ammino ion, Zn(NH&++ 1.0229 0.5098 0.3410 0.2548

0.1717 0.0142 0.0038 0.0012

4.256 2.475 1.686 1.267

these four solutes is readily noted. The adsorption of the zinc ammine is particularly outstanding, and this fact is borne out in the ensuing studies on the mixtures of these compounds. It is worthy of mention, also, that if we were to plot the millimoles adsorbed instead of milliequivalents, the values for nickel and silver would be very nearly the same. In general, it seems that we may classify the zinc and copper ammines together as being very highly adsorbed, the

ADSORPTION OF IONS ON SILICA GEL

641

silver and nickel ammines being more moderate in degree of adsorbability. Results on mixtures will be seen to lead to the same rough classification. LO

40

a0

20

I O

no

0.1

a2

01

M

CONCEWIIUTION (NORWVJ

FIR 1. Adsorption isotherms for metal ammines on silica gel

FIQ.2. Adsorption isotherms for metal ammines on silica gel

Solutions consisting of two of the above ammines, each of 0.5 N concentration, i.e., with a total concentration of 1.0 N , were prepared and the

642

GRANT W. SMITH

adsorptions on silica gel measured. The appropriate solutions of mixed solutes were in all cases prepared first and then added to the silica gel. The results are given in table 3. I n figure 3 these values are shown graphically in such a manner that the influence on each metal ammine of the presence of the other ammines, arranged in order of increasing adsorbability, is demonstrated. In particular it will be noted that the relative order of adsorption is the same as in the case of the simple solutions, and that the adsorption of the zinc ammine is practically uninfluenced by the presence of other solutes. The zinc ammine is almost TABLE 3 Adsorption from mixtures of two complex ammines* Total concentration, 1.0 N ;temperatui 25°C. (approx.) .. ~~

IILLIEQUIVALIKn AD00BBED PEE OBAM OT aTLICA

XNITIAL CONCENTBATION

TINAL CONCENTBATION

milliequirdsnlr per milliliter

millispuioden& per millililer

mi(li.puimlmls

I

0.5000 0.5000

0.4017 0.3133

0.491 0.933

I1

0.4972 0.4972

0.4162 0.0678

0.405 2.147

111

0.5000 0.5000

0.4273 0.0281

0.363 2.359

IV

0.4989 0.4989

0.3450 0.1810

0.769 1,589

V

0.5000 0.5000

0.4280 0.0415

0.360 2.333

0.5000 0.5oOo

0.2850 0.0313

1.175 2.343

EXPIBIMENT

VI

AMMI-E

I N 8OODVTION

Qm

Time of adeorption in each caee, 9 hr.

completely adsorbed in all three cases in which it occurs. On the other hand, the amount of silver, nickel, and copper adsorbed is distinctly decreased by the presence of other ammines. The repressing influence in each case increases with increasing adsorbability of the other ions, Le., in the order silver, nickel, copper, and zinc. Table 4 gives the results of experiments dealing with mixtures of three and four of the ammines, in which the total concentration was 1.0 N and the initial concentration of all ions in any given solution was the same. I n figures 4 and 5 these results are presented graphically as in the

ADBORPTlON OF ION8 ON SILICA GEL

643

above case. Again the almost complete adsorption of zinc ammine is observed, and also the slight influence of the other solutes on this particular one. I n each case a more-or less marked decrease in amount of any particular ion adsorbed in the presence of other ions of increasing adsorbabilities is noted. It can also be shown by simple arithmetic that the amount of a given ion adsorbed in a complex mixture of ions can be predicted with an accuracy of about 5 to 10 per cent by averaging the values found in simpler mixtures of the same ions. Further study of this relationship is to be undertaken in the future in other experiments of this nature. It is also apparent that the values obtained in the mixture of all

FIG.3. Adsorption from mixtures of two solutes. Total concentration, 1.0 narmal four ammines are of the same relative magnitude as those indicated in the simple solutions of table 2 and figure 1. That the ammines are adsorbed in polymolecular layers of approximately colloidal thickness is known as a result of x-ray studies on the metallized gels obtained by reduction of the adsorbed ammines by hydrogen. Such studies have yielded the typical metal crystal patterns in the cases of silver and nickel, the only ones studied by this means by the author (6). The adsorption in the case of all of the mixtures was determined for the same time period,-namely, 9 hr. While it is likely that a h a 1 equilibrium value was not attained in this time, still it seems that this is not

644

GRANT W. SMITH

of much importance in the predent study. The time of adsorption was, of course, the same for each ion in the mixture and was sufficient to show the relative adsorption values clearly. I n line with the results of Kolthoff and Stenger, it seems that the slowness in reaching a final constant state is due to the occurrence of chemical changes in the adsorbed layer following the initial adsorption of the complex ion, rather than to a low adsorption rate. The near-equilibrium value which is attained in a few hours is thus TABLE 4 Adsorption from miztures of three and four complez ammines. Total concentration, 1.0N ; temperature, 25°C. (approx.) SXPmRIMENT

AMMINES I N LIOLUTION

INITIAL CONCENTRATION

FINAL CONCENTRATION GEL

milliaquimlenta per millililcr

millicquimlents per milliliter

millicquimlentr

VI1

0.3329 0.3329 0.3329

0.2716 0.2054 0.1007

0.307 0.637 1.161

VI11

0.3333 0.3333 0.3333

0.2747 0.2538 0.0010

0.293 0.397 1.661

IX

0.3333 0.3333 0.3333

0.2826 0.1250 0.0052

0.253 1.041 1.641

X

0.3333 0.3333 0.3333

0.2584 0.1620 0.0102

0.375 0.857 1.616

XI

0.2500 0.2500 0.2500 0.2500

0,2037 0.1786 0.1053 0,0064

0.231 0.357 0.723 1.218

* Time of adsorption in each case, 9 hr. likely to offer a truer representation of the primary adsorption of complex ions than the final equilibrium value, which is reached only after many weeks or months. Preparations are in progress for a detailed study of the effect of the time factor and also the pH factor on these adsorption phenomena, particularly in the mixtures. I n the latter case it is apparent that, although the pH values of the various solutions used in the present study probably differed somewhat, nevertheless all ions in any given mixture were adsorbed a t the same pH value.

ADSORPTION OF IONS ON SILICA GEL

645

It was found that no appreciable amount of silica was dissolved under the conditions of these experiments. This differs from the results of Kolthoff and Stenger but is logical, since the relative quantities of solution and silica gel used were so vastly different from those used by these authors.

FIG.4 FIG.5 FIQ.4. Adsorption from mixtures of three solutes. Total concentration, 1.0 normal FIQ.5 . Adsorption from mixture of four solutes. Total concentration, 1.0 normal SUMMARY

1. The adsorption isotherms of the complex ammonio ions of silver, nickel, copper, and zinc on silica gel have been determined. 2. The adsorption of the same ions from mixtures of two, three, and four components, originally a t a total concentration of 1.0 N , has also been measured. 3. I n all cases the adsorption in milliequivalents per gram of silica gel was greatest for the zinc ammine, and decreased in the following order: zinc, topper, nickel, and silver. 4. The time required to reach a constant state in the adsorption of silver ammine is considerably shorter than that for the divalent complex ions, and was about 4 hr or less. 5. In the mixtures the adsorption of each component is dependent upon the specific adsorption characteristics of the other components present,

646

GRANT W. SMITE

Le., it is decreased in the presence of highly adsorbable ions, and the amount of decrease is greater the greater the adsorbability of the other components. 6. The zinc ammine, which exhibits very strong adsorption, is only slightly less adsorbed in the presence of the other ammines studied. 7. Knowledge of the adsorption characteristics of the individual solutes, alone and in simple mixtures, enables one to predict, a t least qualitatively, the behavior in more complex mixtures of the same solutes. CONCLUSION

It is hoped that much insight into the nature of the adsorption behavior of these and other complex ammonio ions, as well as information concerning the nature and composition of the ammines themselves, may be gained in further studies now contemplated. The author also plans to carry out experiments to determine the catalytic behavior of the metal catalysts obtained upon reduction by hydrogen of the adsorption products resulting from the present work. The determination of the character of these metallic deposits in the case of the mixed ammines is also to be undertaken by x-ray methods, and should be of considerable value in explaining the phenomena. REFERENCES J.: Kgl. Danske Videnskab. Selsksb, Math.-fys. Medd. 11, No. 5 (1) BJERRUM, (1931); 11, No. 10 (1932); 12, No. 15 (1934). (2) KOLTHOFF, I. M., AND STENQER, V.: J. Phys. Chem. 88, 475 (1934). (3) REYERSON, L. H.,AND CLARK,R. E.: J . Phys. Chem. 40,1055 (1936). (4) SMITH,G.W.: Ind. Eng. Chem. 10, 282 (1938). (5) SMITE,G,W., AND REYERSON, L. H.: J. Am. Chem. SOC. 62, 2584 (1930). (6) SMITH, G.W., AND REYERSON, L. H.: J. Phys. Chem. 88,133 (1934).