134 DUERANT THE INTERACTION OF XV.-The Interaction of Stannous and Arsenious Chlorides. By REGINALD GRAHAM DURRANT. THE action of stannous chloride on arsenious oxide dissolved in hydrochloric7 acid was first noticed by A. Bettendorf (Sitzungsber. ATiederrheiw Ges. Bonn 1869 128 *) two years after his discovery of yellow arsenic [ibid. 1867 67 and (full paper) Annalen 1867 144, H e records the formation of a voluminous brown precipitate which proved to be arsenic (96-99 per cent.) with traces of non-removable tin. He showed that the rate of precipitation increase8 with ascending specific gravity of the arsenious solution. By dissolv-ing magnesium ammonium arsenate in acid he made a standard solution and treated this with stannous chloride in varying dilutions of hydrochloric acid.(His figures will be quohed later on.) From these results ho showed that the reaction is extremely delicate. It may be utilised for determining arsenic in sulpliuric or crude hydro-chloric acid. During a distillation with the latter he observed a faint yellow coloration which disappeared after a few hours. Arsenic was found to be present in this sample of hydrochloric acid, but he was unable to prove that the fading coloration was due to arsenic. The observation of this yellow coloration has decided me to record a v0ry remarkable yellow precipitate which in 1914 I exhibited as (( yellow arsenic ” to the Science Masters’ Association in London. The precipitate was quite bright yellow a t first and was always kept in the dark except when shown for short periods.After a year 1101. -* I a m indebted to Dr. Hatchett Jackson who recently procured me a rescript of this paper from the Bodleian Library STANNOUS AND ARSENIOUS CHLORIDES. 135 it had become a dull mustard colour being still in the original well-corked flask and surrounded by the original solution (a mixture of arsenious and stannous chlorides in nearly normal hydrochloric acid). Every effort was made t o repeat this but in vain. When filtered off the precipitate appeared very dull and shrunken on the paper. After washing it was specially tested for sulphur (sinoe arsenious sulphide is yellow) but no trace of sulphur was found. The presenoe of arsenic was proved. On many p i n t s I find that my observations have been anticipated by Rettendorf in particular the possibility of making the reaction a means of differentiating arsenic from antimony.E x P E R I ni E N T A L. The Nature of the Arsenic Prec,i@tnted. Arsenic is probably in a colloidal state before it is prscipitated, for (i) the precipitate invariably contains a trace of tin salts (chloride as well as tin) and this cannot be removed completely by prolonged washing ; (ii) two similar solutions (reactants 0.44N- and 3N-hydrochloric acid) were left corked for two days and remained quite clear. One was then diluted with an equal volume of water. After four days both had deposited arsenic. A third solution a t the start was made up to the lower of the above concentrations and remained perfectly clear for twenty-five days. The appearance of solid arsenic is always preceded by a pale buff tint; from this a buff-brown precipitate falls and is best observed from such admixtures as yield a very slow deposit.I f this deposit, after washing is immediately shaken with carbon disulphide arsenic is found to be dissolved. Tho yield is rather greater if carbon rlisulphide is shaken violently with the two chloride solutions while they are interacting. On five occasions small pale particles were observed to rise from the clear disulphide solution during spontane-ous evaporation. They moved about rapidly congregating in the centre of the surface then darkened and finally settled on tho bottom of the dish in the form of grey arsenic. Erdmann (Zeitsch. anorg. Chem. 1902 32 453) obtained arsenic soluble in carbon disulphide by reducing arsenious oxide with zinc dust in the presence of the solvent.Very small quantities were obtained by the author in this way. These results and those referred t o in the Introduction indicate that the very earliest deposit of arsenic is of the yellow type but that unless certain unascertained conditions obtain the yellow variety spontaneously becomes brown or grey 136 DURRANT TRE INTERAOTIO’N OF Nature and Conditions of the Reaction. The obvious equation is 2AsC1 + 3SnC1 = 3SnC1 + ZAs and when weights 04 the reactants correspnding with this equation are placed in hydrochloric acid of sufficiently high concentration the action rezches completion in a few hours. With other weights excess of either reactant corresponds with calculation.The action has been proved to be irreversible for if finely divided arsenic is boiled with solutions of stannic chloride in the presence of hydrochloric acid of varying concentration in no case d6es the resulting solution give any precipitate with mercuric chloride. A very careful experiment was made in order to ascertain if the anhydrcus chlorides react. Fresh arsenious chloride was so arranged that on movement of the handle of an air-pump some would drop on to dry powdered stan-nous chlo,ride-also under the receiver. After four days’ final drying with phosphoric oxide the experiment was made. Beyond the faintwt darkening no discoloration occurred. On exposure t o air a distinct brown colour overspread the powder and when a drop of water was added a heavy crusting of arsenic appeared immedi-ately.General Method of Estimu.tin,q the Rate of Progress of Action. Separate solutions containing known weights of the two chlo’rides were mzde u p in known concentrations of hydrochloric acid. Por-tions of these solutions were evaluated separately by means of standard permanganate. The results were found t o agree with the known concentrations. All stock flasks were re-tested from time t o timel. Small dry flasks were placed in a large thermostat and into these definite volumes of both chloride solutions were introduced by separate pipettes. After definite intervals water was added. The dilution effectively stops the action. The contents of each flask were then filtered and uniformly washed. Standard permanganate was used t o determino the amounts of stannous and arsenious chlorides remaining in the filtrates and wash-water.The action of the permanganate may be expressed : 2AsC13 + 2 0 = 2AsOC1 \ As the filtrates required less permanganate than did the sum of { 3SnCle + 3 0 = 3SnOC1, STANNOUS AND ARSENfOUS (3HLORTDES. 137 the separate solutions the deficit became a measure of the change which 'had occurred. Two-fifths of this deficit were due to the pre cipitation of arsenic and the rest to the formation of stannic chloride in the reaction : 3SnClB+ 2AsC1 + 3SnC1 + 2As. Errors.-The sources of error in this process are (1) imperfect washing (2) loss by adsorption (3) oxidation of stannous chloride due to access of air. The two first considered together were found t o give rise to an error probably less than 2 per cent.The third source of error was almost eliminated by keeping the stock solutions of stannous chloride in a well-coked flask and by introducing carbon dioxide immediately after use on every occasion. I n the same way the reaction took place in small corked flasks in which the air was displaced by carbon dioxide. Air had access only during the process of filtration. Calculations.-The recognised integration equations for first and second order reactions were applied to a large number of 'determina-tions. I n no case did the velocity constants conform to the second order. The results quoted are from the firsborder equation, 1 a - 1% t where a = 100 x = percenta.ge of change and t = time in minutes.Hence the mean value of k for each set of experiments represents a special figure by which the relative speeds may be com-pared. TABLE I. t = 12.5O. N/4-Permanganate used. Normalities S'nC1 = W507 AsC13= 0.584 HC1= 6.06. Complete oxidation should correspond with a deficit of 45.6 C.C. Interval, minutes. 2 5 8 10 12 15 20 30 40 50 65 100 120 180 Deficit, 3-66; 9.3 16.65 17.05 26-25 29-4 33.85 38.45 40.1 40.8 41.9 42-85 43-6 43-9 C.C. Percentage chmqe. 8.0 20.8 36.5 L37.41 57.5 64.5 74.2 84.3 87.9 89.5 91.9 93.9 95.6 96.3 k x 102. 1-81 2.03 2.46 r2.031 3-09 3.00 2-94 2.68 2-29 2-96 2.21 2.97 2-40 Mean 2-57 P.681 138 DURRANT THE INTERACTION OF Various further unimolecular values of k were obtained.These were found tci depend more on the concentration of hydrochloric acid than on anything else. The results made it pmsible to choose suitable concentrations for systematic study. Influence of IT ydroc hloric A cid. In the following experiments 0.2500 gram of stannous chloride ( 8 7 6 6 4-85 R W v e velocitiu of the reaction 3SnC1 i- 2AsC1 = 3SnC1 + 2As, See Table XI. Tgmperature = 1 2 9 due to altmt&ns of hydrochlol.ic concentration. acted on the equivalent weight of arsenious chloride in each case. The concentration of hydrochloric acid alone was varied. TABLE 11. t = 1 2 O . Reactant Normality = 0.298. Normality Range of change, of HCl. per cent. 10.09 40-6 1 8-10 36-64 7.25 32-63 6-77 13-63 6.60 2 1-69 6.11 41-63 7.09 34-78 6.34 26-64 4.85 19-63 Mean value, k x lo2.21.2 13.6 4.7 6 3.43 1-40 0.944 0.436 0.293 0.029 STANNOUS AND ARSENIOUS UHLOZUDES. 130 The curve obtainsd by plotting these relative velocities against concentration of hydrochloric acid between 4.85N and 8*10N is exceedingly regular. Its sharpest curvature is in the neighbourhd of 6.5N. I f the regularity persisted u p to the limit of possible hydrochloric acid concentration (about 10*3N) then the velocity a t 10*09~Y would be well over a thousand times what it is a t 4-85N-as measured it is only 723 times as great. Influence of Simultaneous Change ic Concentration of Reactants. I n all these experiments the concentration of hydrochloric acid remained constant a t 6 N .Four 250 C.C. flasks-A B A B,-contlained respectively, SnC1 = 1-74N AsCl3= 1-76N SnC1 = Om87N As(&= 0.88fl. The concentration of hydrochloric acid became 6N as soon as the mark in each flask was reached. Equal volumes from A and B were mixed in six small flasks, and after 4 6 8 10 12 and 14 minutes respectively their filtrates were titrated with AT/ 4-permanganate. In the same way equal volumes from A and 13 were treated from seven flasks after 24 32 48 64 80 96 and 112 minutes, respectively the filtrates being titrated with iV/ 8-permanganate. I n each set the range of progress was from 30 to 70 per cent. For A B set mean value k x 102=4'33. For A,B set mean value k x lO2=O0.557. Hence k/k =7*77 for the range between 30 and 70 per cent.t =18.4. The range between 30 and 40 per cent. however gave k l k = 5 * 5 . E'fJect of altering the Concentration of Each Reactant Sepmately. Preliminary work had appeared to show that arsenious chloride reacts as a second and stannous chloride as a first power. The following solutions were prepared tested and preserved with all possible care. Five C.C. of stannous chloride solution reacted with 5 C.C. of arsenious chloride solution in each case. The washing was strictsly uniform so that errors hence arising were similar. The mean results (2) and (4) in the following table are fairly concordant. The results from comparison of (l) (4) and (5) confirm the preliminary work with respect to arsenious chloride which is seen t o react! as a second power.The period preceding the first appearance of arsenic from a solution of its chloride a t one-fifth the original concentration was noticed to be just about twenty-five times as great as it had been. Those of (3) and (5) are more YO -1 40 TABLE 111. HCl= 6 N . t = 16.7'. (1) N-AsCl acting on N-SnC1 in 12 minutes required c.c.N/4-KMn04 Required by theory after 33.3 per cent. change. Mean of 2 readinns 27.85 26.66 (2) N-AsCl, (3) N-AsCl, (4) N-SnCl, (5) N-SnC1, This suggested Nesslerisation ; o ; N / ~ - s ~ c ~ in 12 minutes on N/3-SnC12 in 12 minutes on NIZ-AsCl in 48 minutes on N/3-AsC13 in 108 minutes ?) 22.3 23.3 $ 9 20.8 22.2 $9 22.55 23.3 ¶# 20.9 22-2 a method of working t o a standard tint as in moreover the method compares the earlier &ages of action on which calculations are more appropriately based.sulphide on a very dilute solution of lead acetate. The tlint used in table IV was obtained by t'he action of hydrogen TABLE IV. HCl= 61Y. t = 10'. Times to reach Stannous Arsenious chloride. chloride } N NP N/3 1 /c2 Total Ratio Staindard Tint are yiven in Seconds. + N . + N / 2 . + N/3. Total. Ratio. ,& 30 110 260 400 1.0 1.0 50 220 460 730 1-82 1-73 40 160 360 560 1.40 1.41 120 490 1080 1 4-08 9 1 4 9 A similar set of nine readings referred to another artificial standard tint gave ratios powers of AsCl, 1 4.3 10.0; powers of SnCl, 1 1-37 1-80. The N/3-stannous chloride solution on testing was found t o have deteriorated slightly; the others had not.A solution of N/4-stannous chloride was made. Using N / 2-arsenious chloride against N,' 2- and N /4-stannous chloride, the times were 230 and 320 seconds respectively giving a ratio 1 /1*39 again closely approaching 1/ d2. The results here given lead to the conclusion that in this reaction arsenious chloride reacts as a second power and stannous chloride reacts to the power of the s p a r e root of its concentration STANXOUS AND ABSENLOVS CHLORIDES. 141 The figure 5.5 nded in the last paragraph for the chasge between 30 and 40 per cent. is quite consistent with the results here given since 22 x 43 = 5.64. The action of stannous chloride to the square root of its con-centration is also in agreement with Bettevdorf's figures (Zoc.cit., 1869). H e took 0.001 gram of arsenic dissolved in 1 C.C. in each of five experimente adding this to a definite amount of stannous chloride solution in the presence of hydrochloric acid. I n the four last experiments he also added 50 100 200 and 400 C.C. of hydro-chloric acid (presumably of similar concentration). An immediate precipitate occurred in the first experiment and the arsenic appeared in 5 8 12 and 20 minutes respectively in the others. Neglecting 1 C.C. of arsenious chloride + an unknown volume of stannous chloride solution originally taken his concentrations were 1 2 4 8 his times were 1 1.6 2.4 4 figures which approach 1:1.41:2:2*83 but exceed them in each case because of the influence of the second power action of amenioua chloride present in very small relative amount.(His experimentcl were made to show the delicacy as regards arsenic.) Effect of Dihtirm with Water. The stock solutions when mixed were a t concentrations HC1=6N and reactants each a t 0*88N. When undiluted this mixture produced 70 per cent'. change in twelve minutes. The dilutions (in ten steps) finally brought ali the concentrations to one-third of the above. The hydrochloric acid normalities and the state of change after five days are noted in each case. t=16O. 6N 5*45N and 5N had reached complete change. 4.61N 89 per cent. 4.29N 59 per cent'. 4N 24-4 per cent. 3-75N 14.2 per cent. 3.53N 5.6 per cent., 3-3N 3 per cent. 3rV 1 per cent. and 2N no change and no subsequent sign of action after 29 days.This retarding action was made use of in all previous experiments when titrations with permanganate were made the dilution with water being sufficient to reduce the concentration to one-third or IAY. Summary. (1) There is evidence that arsenic in process of precipitation is I n certain circumstances partly soluble in carbon disulphide. arsenic may appear as a yellow deposit. VOL. axv 142 INTERACTION OF STANNOUS AND ARSENIOUS CELORIDES. (2) The anhydrous chlorides (arsenious and stannous) do not interact. (3) Acceleration of t'he action is caused chiefly by increase in the concentration of hydrochloric aoid next by that of arsenious chloride and least of +11 by that of stannous chloride. Arsenious chloride acts as a second power and stannous chloride to the power of the square root of its concentration.Conclusions. The various phenomena and the figures given can be account.ed for on the hypothesis that this action is between chloride ions, arsenious ions and the stannous complex H,SnCr,. Stoppage by dilution must be due to the destruction of arsenious ions by hydrolytic action. (1) Chloride ions proceed partly from arsenious chloride and partly from hydrochloric acid and they act as a first power. The velocity constants found in table I1 are thus explained. HCl normality. k x 10'. k/k. Cl'/cl'. 10.09 8.10 7-28 7-09 6.77 6-60 6-34 6.1 1 4-85 1-57 2-83 1.37 2.35 1.64 2-16 3.22 10.00 7.09 3.55 1.54 2.08 1-58 1.88 2.66 0.28 Aggregate 26.04 26.66 In the last column the numerator gives the sum of chloride ions due to arsenious chloride and those due to increased hydrochloric acid concentration ; the denominator is constant and represents the chloride ions due to the 0.298fl-arsenious chloride which is constant throughout the table.In the lower portion of the table, the arsenious chloride is not wholly ionised; in the upper portdon, hydrochloric acid becomes less ionised a t its higher concentrations. As is seen the aggregafe acceleration as directly proportional to the increase of chloride cmcentratiom. (2) Positively charged aqenious ions also act as a first power. Arsenious chloride as a whole appears t.herefore to act as a second power. (3) That a compound of hydrochloric acid and stannous chloride exists in solution was indicated by Young ( J . Amer. Chem. SOC., 1901 23 21 450) and several stannochlorides corresponding wit ELIMINATION OF THE CARBETHOXYL QROUP E"0. 143 the formula MzSnC14 have recently been described (compare Druce, Chem. News 1918 117 193). In the reaction this complex must be decompased in order t o produce stannic chloride and this decam-position may account for the complex acting to the power of the square root of its concentration. According t o accepted theory the order of a reaction is governed by the slowest reactant. The order here is unholecular and the slowest reactant is t-his complex. Essentially the action consists in the disintegration of the complex by circumambient ions. THE COLLEUE, MARLBOROUGH. [Rereiued July 30th 1918.