THE ADSORPTION OX CATALYTICALLY POISOXOUS JIETdLS KTC. S PI.-The Adsorptio7a of Catulyticully PO~SOPLOZ~.~ Netals by Platinum. Part I . The Adsorptiaq? of Lead and Mercuyy. By EDWARD BRADFORD MAXTED. I~~EASIJEEWEKTS have been described (J. 1921 119 225 ; 1922, 222 1760) relating to the inhibitive influence of salts of certain metals on the hydrogenation of unsaturated substances and on the dccoinposition of hydrogen peroxide. In each case it appeared D 74 MAXTED THE ADSORPTION OF probable on first principles that adsorption of the metallic salt, or rather of the catalytically poisonous ion by the catalyst took place; but since the degree of adsorption was unknown the inhibi-tive effect of the poison measured by the depression in the activity of the catalyst could only be plotted against the initial bulk con-centration of the inhibitant in the reacting system in place of-as would have been more logical-its actual concentration on the surface of the catalyst itself.Accordingly the present work was carried out with the object of obtaining data relative to the variation of this adsorbed concentration with the bulk concentration of the inhibitant. The present paper deals with the adsorption of lead and of mercury by finely divided platinum. E X P E R I M E N T A L . An aqueous suspension of platinic oxide was reduced with hydro-gen the precipitated metal washed several times with hot distilled water and ground thoroughly in an agate mortar in presence of a little water. This method of preparing the platinum was found preferable to the reduction of a platinum salt with formaldehyde or other reducing agent on account of the difficulty in such cases, of freeing the preparation from adsorbed impurities.In carrying out each adsorption experiment 1 C.C. of a stock suspension of platinum containing 4.48 mg. of platinum per c.c., was added to a system containing a known quantity of a lead or mercury salt dissolved in 9 C.C. of water the mixture was thoroughly agitated kept over-night a t 20° and an appropriate portion of the clear liquid taken for analysis. The concentration of lead or mercury salt remaining unadsorbed was determined by observing the toxic effect of a known fraction of the solution on the activity of a standard catalyst for the decom-position of hydrogen peroxide for which reaction the inhibitive effect of various known concentrations of the lead and mercury salts in question had previously been mapped for the catalysts employed.It was thus possible to estimate with considerable accuracy small fractions of a milligram of lead or of mercury. Adsorption of Lead.-An aqueous solution of lead acetate usually containing 0.1 mg. of lead per c.c. was employed; for certain experiments solutions of one-tenth and ten times this strength, respectively were used. Preliminary experiments with potassium acetate showed that the acetate ion is a t any rate in the concen-trations employed without appreciable toxic action on platinum for the decomposition of hydrogen peroxide. The result of a series of adsorption experiments in each case with 4.48 mg.of platinum is shown graphically in Fig. 1. It wil CATfiYTICALLY POISONOUS METALS BY PLATINUM. PART I. 75 be seen on plotting the weight of lead adsorbed against its initial concentration that a linear relationship exists between these two quantities up to a stage a t which the platinum surface apparently approaches saturation ; and further adsorption from this stage onwards takes place only to a slight degree on increasing the initial bulk concentration of the lead salt. The two graphs in the figure were obtained by plotting the weight of lead adsorbed per gram of platinum first against the initial bulk concentration and secondly, against the final bulk concentration of the unadsorbed lead in the system. This result is analogous t o those obtained by Euler and Hedelius (Arlziv Kern.1920 7 31) for the adsorption of a silver FIG. 1. Bulk concentration of lead. X g . in 10 C.C. salt by silver and gold for which also an initially linear adsorption graph was obtained. Adsorption of Mercury .-A dilute solution of mercuric chloride was employed the other conditions being similar to those adopted for lead. Fig. 2 gives the results of a series of adsorption experiments, carried out as before with 4-48 mg. of finely divided platinum. For initial concentrations of mercury below that corresponding with the saturation of the platinum the adsorption of mercury is practically complete ; and the very small concentration of mercury remaining after adsorption could be estimated with the required degree of accuracy only by virtue of t,he method of analysis employed (see above).D* 76 THE ADSORPTION OF CATALYTICALLY POISONOUS METALS ETC. The linear nature of the adsorption graph for mercury and the abrupt break in this graph immediately before the saturation point, are very striking. The form of the adsorption graph both for lead and for mercury, on platinum may be compared with that of the poisoning curve. It has previously been shown (Zoc. cit.) that the decrease in the activity of a platinum catalyst caused by a poison such as lead or mercury is for by far the greater portion of the poisoning curve, a linear function of the initial bulk concentration of the poison in the system. It has now been shown that the mass of such a poison, e.g. lead actually adsorbed by platinum-i.e.the concentration of the lead on the surface of the platinum-is also a linear function of this initial bulk concentration of the lead in the system. From FIG. 2. 10 4 PI 3 8 d 0 0.04 0.08 0.12 0.16 Bulk concentration of mercury. Mg. in 10 C.C. this it follows that the activity of the catalyst in the presence of such a poison is at any rate for this first stage a linear function of the actual concentration of the lead on the surface of the catalyst. This result was to be expected on theoretical grounds since the free valency forces on the surface lattice of the platinum will if they are saturated by a difficultly evaporable poison such as lead, no longer be free for the adsorption of or association with a poten-tially reactive system it being presumed that catalysis normally takes place by reason of such association.Therefore since each atom of lead thus obstructively adsorbed will cause a given number (one or more) of such valency bonds to become no longer active for normal catalysis the poisoning of this catalytic surface should be a linear function of the concentration of the poison on the surface, a result which has just been obtained experimentally. From Fig. 1 it is seen that the linear relationship between the concentration of the lead on the catalyst and its bulk concentratio THE ZXPLOSIOX OF ACETYLENE AND NITROGEN. PART IV. 77 in the solution ceases as is necessarily the case in the region border-ing on the saturation of the surface of the platinum by the lead, in which region the initially linear adsorption graph becomes more or less abruptly converted into a line parallel with the bulk coiiceii-tration axis and denoting saturation.I n the poisoning graphs previously obtained for such cases the initial linear portion-during which the activity of the catalyst is a linear function of the initial bulk concentration of the poison in the system-is also followed in a very similar way by a break, followed in turn by n far less steep portion which like the adsorption curve becomes approximatc~ly parallel with the bulk concentration axis. It appears of great interest to ascertain whether the break in the poisoning graph obtained by plotting activity against bulk concentration of the poison occurs a t the region of incipient satur-ation.If this is the case then the activity of the catalyst would seem to be linearly proportional to the actual concentration of poison on the surface of the catalyst not only during the major portion of the poisoning graph but also up to the complete extinction of catalytic activity. It is hoped to discuss this point more fully iil a later paper. ~IASOR ROAD, PENS WOLVERHANPTOS. [Received October 14th 1921. THE ADSORPTION OX CATALYTICALLY POISOXOUS JIETdLS KTC. S PI.-The Adsorptio7a of Catulyticully PO~SOPLOZ~.~ Netals by Platinum. Part I . The Adsorptiaq? of Lead and Mercuyy. By EDWARD BRADFORD MAXTED. I~~EASIJEEWEKTS have been described (J. 1921 119 225 ; 1922, 222 1760) relating to the inhibitive influence of salts of certain metals on the hydrogenation of unsaturated substances and on the dccoinposition of hydrogen peroxide.In each case it appeared D 74 MAXTED THE ADSORPTION OF probable on first principles that adsorption of the metallic salt, or rather of the catalytically poisonous ion by the catalyst took place; but since the degree of adsorption was unknown the inhibi-tive effect of the poison measured by the depression in the activity of the catalyst could only be plotted against the initial bulk con-centration of the inhibitant in the reacting system in place of-as would have been more logical-its actual concentration on the surface of the catalyst itself. Accordingly the present work was carried out with the object of obtaining data relative to the variation of this adsorbed concentration with the bulk concentration of the inhibitant.The present paper deals with the adsorption of lead and of mercury by finely divided platinum. E X P E R I M E N T A L . An aqueous suspension of platinic oxide was reduced with hydro-gen the precipitated metal washed several times with hot distilled water and ground thoroughly in an agate mortar in presence of a little water. This method of preparing the platinum was found preferable to the reduction of a platinum salt with formaldehyde or other reducing agent on account of the difficulty in such cases, of freeing the preparation from adsorbed impurities. In carrying out each adsorption experiment 1 C.C. of a stock suspension of platinum containing 4.48 mg. of platinum per c.c., was added to a system containing a known quantity of a lead or mercury salt dissolved in 9 C.C.of water the mixture was thoroughly agitated kept over-night a t 20° and an appropriate portion of the clear liquid taken for analysis. The concentration of lead or mercury salt remaining unadsorbed was determined by observing the toxic effect of a known fraction of the solution on the activity of a standard catalyst for the decom-position of hydrogen peroxide for which reaction the inhibitive effect of various known concentrations of the lead and mercury salts in question had previously been mapped for the catalysts employed. It was thus possible to estimate with considerable accuracy small fractions of a milligram of lead or of mercury. Adsorption of Lead.-An aqueous solution of lead acetate usually containing 0.1 mg.of lead per c.c. was employed; for certain experiments solutions of one-tenth and ten times this strength, respectively were used. Preliminary experiments with potassium acetate showed that the acetate ion is a t any rate in the concen-trations employed without appreciable toxic action on platinum for the decomposition of hydrogen peroxide. The result of a series of adsorption experiments in each case with 4.48 mg. of platinum is shown graphically in Fig. 1. It wil CATfiYTICALLY POISONOUS METALS BY PLATINUM. PART I. 75 be seen on plotting the weight of lead adsorbed against its initial concentration that a linear relationship exists between these two quantities up to a stage a t which the platinum surface apparently approaches saturation ; and further adsorption from this stage onwards takes place only to a slight degree on increasing the initial bulk concentration of the lead salt.The two graphs in the figure were obtained by plotting the weight of lead adsorbed per gram of platinum first against the initial bulk concentration and secondly, against the final bulk concentration of the unadsorbed lead in the system. This result is analogous t o those obtained by Euler and Hedelius (Arlziv Kern. 1920 7 31) for the adsorption of a silver FIG. 1. Bulk concentration of lead. X g . in 10 C.C. salt by silver and gold for which also an initially linear adsorption graph was obtained. Adsorption of Mercury .-A dilute solution of mercuric chloride was employed the other conditions being similar to those adopted for lead.Fig. 2 gives the results of a series of adsorption experiments, carried out as before with 4-48 mg. of finely divided platinum. For initial concentrations of mercury below that corresponding with the saturation of the platinum the adsorption of mercury is practically complete ; and the very small concentration of mercury remaining after adsorption could be estimated with the required degree of accuracy only by virtue of t,he method of analysis employed (see above). D* 76 THE ADSORPTION OF CATALYTICALLY POISONOUS METALS ETC. The linear nature of the adsorption graph for mercury and the abrupt break in this graph immediately before the saturation point, are very striking. The form of the adsorption graph both for lead and for mercury, on platinum may be compared with that of the poisoning curve.It has previously been shown (Zoc. cit.) that the decrease in the activity of a platinum catalyst caused by a poison such as lead or mercury is for by far the greater portion of the poisoning curve, a linear function of the initial bulk concentration of the poison in the system. It has now been shown that the mass of such a poison, e.g. lead actually adsorbed by platinum-i.e. the concentration of the lead on the surface of the platinum-is also a linear function of this initial bulk concentration of the lead in the system. From FIG. 2. 10 4 PI 3 8 d 0 0.04 0.08 0.12 0.16 Bulk concentration of mercury. Mg. in 10 C.C. this it follows that the activity of the catalyst in the presence of such a poison is at any rate for this first stage a linear function of the actual concentration of the lead on the surface of the catalyst.This result was to be expected on theoretical grounds since the free valency forces on the surface lattice of the platinum will if they are saturated by a difficultly evaporable poison such as lead, no longer be free for the adsorption of or association with a poten-tially reactive system it being presumed that catalysis normally takes place by reason of such association. Therefore since each atom of lead thus obstructively adsorbed will cause a given number (one or more) of such valency bonds to become no longer active for normal catalysis the poisoning of this catalytic surface should be a linear function of the concentration of the poison on the surface, a result which has just been obtained experimentally.From Fig. 1 it is seen that the linear relationship between the concentration of the lead on the catalyst and its bulk concentratio THE ZXPLOSIOX OF ACETYLENE AND NITROGEN. PART IV. 77 in the solution ceases as is necessarily the case in the region border-ing on the saturation of the surface of the platinum by the lead, in which region the initially linear adsorption graph becomes more or less abruptly converted into a line parallel with the bulk coiiceii-tration axis and denoting saturation. I n the poisoning graphs previously obtained for such cases the initial linear portion-during which the activity of the catalyst is a linear function of the initial bulk concentration of the poison in the system-is also followed in a very similar way by a break, followed in turn by n far less steep portion which like the adsorption curve becomes approximatc~ly parallel with the bulk concentration axis. It appears of great interest to ascertain whether the break in the poisoning graph obtained by plotting activity against bulk concentration of the poison occurs a t the region of incipient satur-ation. If this is the case then the activity of the catalyst would seem to be linearly proportional to the actual concentration of poison on the surface of the catalyst not only during the major portion of the poisoning graph but also up to the complete extinction of catalytic activity. It is hoped to discuss this point more fully iil a later paper. ~IASOR ROAD, PENS WOLVERHANPTOS. [Received October 14th 1921.