THE P R E C I P I T A T I O N OF AN'fIMONY, ETC.
87
some results which were not so good, but, on the whole, the method is quite promising and will be the subject of further study by me, which I trust may result in a somewhat more accurate and more convenient method than those now in general use. AGRICULTURAL EXPERIMENT STATION, MORGANTOWN, W E S T VIRGINIA, December IO. 1894.
ON CERTAIN PHENOMENA OBSERVED IN THE PRECIPITATION OF ANTIMONY FROM SOLUTIONS OF POTASSIUIl ANTIIlONYL TARTRATE. BY J. H .
LONG.
Received December 19,1~34.
I
H A V E elsewhere called attention to the behavior of solutions of tartar emetic when treated with solutions of other salts, (see Am. Jour. Sci. and Arts, October, 1889 and October, 1890) and with Mr. H. E. Sauer have determined the conditions of A.n a l . precipitation by carbonates, acetates, and phosphates. (1 APPC. Chem., March, 189r.) When to solutions of the antimony salt sulphates, chlorides, nitrates, and oxalates of the alkali metals are added no precipitation occurs, even with elevation of temperature. With carbonates, acetates, phosphates, borates, thiosulphates, sulphites, tungstates, and some other compounds, clear solutions can be made at a low temperature, but precipitation follows a t a higher point. T h e precipitate, in nearly all cases, consists of hydrated antimony oxide, and its amount is a function of time, temperature, and amount of added salt. With mixtures of the tartrate and sodium carbonate, for instance, it was found that in the cold, a t the end of twenty-four hours, amounts were precipitated as shown in the following table. I n each test five grams of the tartrate were dissolved in Then different sixty cc. of warm water and cooled to 20'. weights of pure sodiuiii carbonate were dissolved in thirty-five cc. of water; these solutions were added to the others and the mixtures were brought up to 100 cc. They were allowed to stand until precipitation was complete, usually over night or longer. An aliquot part of the clear supernatant liquid was
88
J . H. LONG.
THE P R E C I P I T A T I O N O F
taken and the aniount of antimony in solution determined. This was calculated to tartrate in the whole : NO. of experiment.
KSl)OC,H 0 iHIO
Na Cu
left in s&&on.
ad&L3
............... 0.1 grain 2 . . ................ 0.3
99.93 per cent.
I..
3 ..................
0.j
''
8j.22
..
70.36 '' 56.76 40.87 29.1; " 13.94 " 3.88
..
4 .................. 0 . ; 5 .................. 0.9 ' . 6 . . ................ I . 2 granls 7 .................. 1.5 ' I 8 .................. 2 . 0
q t
I '
1'
' (
I'
'' "
'' " I '
I n another series of experiments the solutions of carbonate and tartrate were mixed as before a t 20' and then brought to I O O ~ , and maintained a t this heat one hour. T h e precipitates formed immediately, and a t the end of the hour were separated by filtration. T h e filtrates were tested for antimony remaining. T h e results are shown in the third column below. No. of
experimeut.
KSbOC,€I 0 i H I O
S a CO
ad4ed.3
left in sJu?ion.
................ 0 . 2 g r a m I O . . ................ 0 . j 1 1 .................. 0.3 1 2 . . ................ I , I granis 13.. ................ 1 . j ' ' 14 .................. 2 . 0 I S . . ................ 3.5 1 6 . . ................ 5 . 0 9..
79.23 per cent. 46.70
' I
' I
"
21.74
"
8.69 6.53 4.42 4.66 4.74
" I'
I'
,' 8'
('
(' "
'
O.
O.,;IZ
2 graiiis of
LNTIMONY F R O N SOLUTIONS O F T A R T A R EMETIC.
93
T h i s solution was polarized in the- 2 0 0 mm. tube and gave UD
= 3 . jgo",
which agrees very well with the result of the first experiment. Another solution, containing in 100cc., a t zoo, 0.1.50 gram of KSbOC,H,O,.$H,O, 4. I 19 grams of KXaC,H,0,.4H2O, 1.811grams of H,BO,, gave a, = 3.661". While boric acid increases the rotation of tartrates and tartaric acid I have elsewhere shown that borax decreases the rotation of Rochelle salt slightly. T h e equatiori probably represents the facts properly. I n the above nothing has been said about certain peculiarities observed in the formation of the precipitates. When cold dilute solutions of borax and the tartrate are mixed no reaction takes place immediately, but with warm, strong solutions, a precipitate seems to form as soon as the two liquids are poured together. A x a l . A@l. Chem., loc. c i f . ) I pointed I n a former paper (1. out the important and exceedingly curious fact that in the reaction between carbonates or acetates on the one hand, with the tartrate on the other, while no precipitate may appear ininiediately, perhaps not in hours, indicating a decomposition, the polarimeter shows that such has taken place. Here, also, we have evidence that a reaction has taken place even without precipitation, and this the polarimeter furnishes. T h e matter can be best explained by giving the details of several experiments. I made five solutions by dissolving five grams of the tartrate a s before in fifty-five to sixty cc. of warm water, cooled to zoo, and added certain weights of borax in small volumes of water, making the solutions finally to 100cc. a t 20". These solutions were polarized immediately in the 2 0 0 mni. tube, with the following results : NO.of
experiment,
KSbOC,H,08.fH,0 taken.
.......... 5 grams ........ 5 3.... ........ 5 4...,........ 5 I..
2,...
"
"
5....,.
...... 5
Na,B,O .roH,O added.
a,
0.5 gram
12.08"
1.0
I'
1,o grams
9.40"
5.5.3*
"
3.0
I'
4,IO"
"
4.0
'I
3.7P
94
J. H. LOSG.
THI3 PRECIPIT~\TIOS OF
T h e noriiial rotation of the tartrate at 20' in the 2 0 0 nin. tube, with a concentration of five granis iii 100 c c . , I have show1 to be (YI,
= 14. r c q '
T h e effect of the borax is therefore iiiarked, but the tsterit of the decrease in rotation tiepeiitls on the iiuniber of minules intervening between the iiiixiiig of the solutions and tlie coriipletion of tlie observation iri the polariitieter. -% gradual de8:rease in the readings vias iii all cases obserrmi, until the solutiors becaiiie finall!. too turbid for observation froiii the begirinirig precipitation. Tlie first solution, for illstarice, iii the above table was read as follows : I O hours, 30 minutes, aI> = 12.08; IO " 4j =I I ,sj3 "
IO
"
gj
"
% ,
"
= I1.SO'
After standing sonie hours, the solutions deposited a precipitate and cleared up. On again polarizing I fourid : SO.
(Y!>
.................... 11.52 2 ........... 9.01. 3 . ................... 5.29' 4 .................... 3.02 5 ............ 3.65 I
These obsenations were niade iii a zoo nini. tube, but similar solutions were polarized in a 400 111111. tube with perfect sharpness, the readings agreeing within o.ozo,as is possible with the large and excelleiit instrument used. I niention this to prove the perfect transparency of the liquids, and to show that the decreased rotations observed at the start ryere not due to any loss through precipitatiori, but were in consequence of changes preceding precipitatiori, these changes taking place very gradually. e& '\ have here a phenonienon reniindirig one of the birotation of solutions of certain sugars, but depending 011 a different cause undoubtedly. 111 the reaction between the same tartrate arid sodium carbonate the sanie change was observed, brit through a longer period. A solutioii containing iii 1 0 0 cc. one-tenth gram of the carbonate and five granis of the tartrate gave, at the end of five minutes i n a 400 nini. tube, a rotation of 25. j8Z0,after
ANTIJIOSY FROM SOLUTIOXS O F T A R T I R
EMETIC.
95
thirty minutes, 25. j80", that is, practically the same, but after twelve hours, 24.480". A perfectly clear mixture can be made containing five grams of the tartrate and nine-tenths gram of sodiuin carbonate in 100 cc. IYlieii polarized immediately, I found with this in the 400 nini. tube (xD = I I . j7', after ten minutes, I I . jo", after twenty-five minutes, I I . 132", and after sixtyfi1.e minutes, 10.5j". In the norinal reaction between carbonates or borates and the potassium antimonyl tartrate a precipitate should he formed, but we find that at a low temperature this is niuch delayed. If precipitation alone were taken as the indication of a reaction it would necessariiy appear that at the outset no reaction takes place, but the behavior with poiarized light shows the error in this view. I t is evident that a reaction begins immediately and progresses far toward completion in some of the cases considered before even tlie first polarization can be made, that is within two or three minutes. This first part of the reaction is the beginning stage of precipitation and niay corisist in the formation of some intermediate product, which filially decomposes. I have elsewhere shown ( A m .Joiiy. S r i . mid .1rZs, lor. c i f . ) that the rotation of potassium sodium tartrate is decreased by the addition of sodium, thallium, and lithium salts, but is increased by the addition of potassium and aniiiioniuni salts and that this reaction is fully accounted for if we assume the formation of sodiuiii tartrate, sodium tlialliuni tartrate or sodium lithium tartrate in the one case or of potassium or potassium amnioniuni tartrate in the other. I n the present instance we evidently must admit the formation of sotliuni potassiuin tartrate from tlie iiistaiit the solutions are mixed, but that the reaction is a progressive one. T h e potassium antiiuonyl conipouiid with a high rotation gives place to the potassium sodium compound with a niuch lower rotation. There is nothing to shorn, however, in what form the antimony is held. Possibly tlie reaction inay be explained by assuming the formation of an intermediate product according to this equation : ~ K S h O C , K , O , ~ N a , B , O= , 2KNaC,H,O,+ (SbO),B,O,. If the last coiiipound is formed it must break up in thismanner : (Sb0),B,0,+6H10+XH10=~H,BO,fSb,0,.XH,0,
96
J . H. LOSG.
THE PRECIPITATIOX O F
leaving a hydrated oxide of antimony with more water than the final precipitate contaiiis. By loss oi water. possibly, this hydrated compound must, in time, settle out as R precipitate. I t has been explained that by boiling, the precipitate forms arid subsides soon. At 20'~ even after what I have called the preliminary stage of precipitatioii, may have occupied hours. the actual formation of the precipitate may coiisuiiie nil equall)~Ioiig time. T h e precipitate is a growth through an invisible and a visible stage, and what is true here is true of the next case to be given. R E A C T I O N W I T H SODILTM T U S G S T - I T E .
Cold solutions of the tartrate give no immediate precipitate when mixed with cold solutions of ordinary sodium tungstate, but on standing the mixtures gradually become turbid aiid finally deposit a sediment. T h e coniposition of this depends largely on the temperature and concentration, T h e precipitate formed in the cold, collected and dried at 10s'-1 roo C . consists, essentially of antimony oxide. A precipitate formed by mixing hot solutions has practically the sanie coniposition. but if obtained after long heating it contains a relativeiJ- larger amount of tungstic acid. In a series of tests made by mixing hot solutions of the two salts, and allowing the niixtures to stand several hours to cool, the following results were obtained : KSbOC,H,08.)Hp0 1u I o 0
I 2
4 5
cc.
NalWO,.zH,O I U 50
cc.
gram grams
2
grams
2
"
"
2
I'
"
2
I'
W t . of
mt.
0.3087 0.6442 0.5728 0.7222
Sb,S, from same.
Per c e n t . of Sb.i n ppt.
0.3467
80.22
0.7291
80.84 80.88 80.79
0.6486 0.8170
T h e mixtures were made in platinum dishes holding about 2 0 0 cc., and as the precipitates formed as a coherent coating on the dishes they were easily washed, dried at IIO', and weighed. They were then dissolved in diluted hydrochloric acid, which left a small amount of tungstic acid in each case in flocculent form. T h e solutions were then filtered, and, after the addition of some tartaric acid, were precipitated by hydrogen sulphide in the usual manner. T h e sulphide precipitates were collected 011 a Gooch, dried at IIO', atid weighed. It will be seen that the
A N T I M O N Y FROM SOLUTIONS OF T A R T A R EMETIC.
97
results are a little low to correspond to pure antimony oxide a s the composition of the white precipitate. This compound contains 83.3 per cent. of antimony. T h e compound Sb,O,.H,O contains 78.4 per cent. T h e lower results are doubtless due to the small amounts of tungstic acid left in each case on treatment with hydrochloric acid, and referred to above. I n a second series of experiments constant amounts of the tartrate in hot solution were mixed with varying amounts of the tungstate likewise in hot solution. T h e white precipitates which formed were collected and weighed as before, with the following results, which show the effect of excess of tungstate on the amount of precipitate. I n each case two grams of the tartrate was taken in 100 cc., and the tungstate in fifty cc. : Tungstate taken.
Ppt. obtained.
0.5 gram
0.1070 0.4598 0.5762 0.5861 0.6143 0.6185
1.0 " 2.0 grams
3.0 4.0 5.0
" 'I '(
From this, it is plain that the amount of precipitate is not much increased by great excess of the tungstate beyond a certain point. I n the cold, precipitation is much less perfect, while, by boiling, fully three-fourths of the theoretical yield of oxide from the tartrate can be obtained. By working with cold solutions a mixture may be made which remains clear long enough to permit polarimeter observations to be taken. I dissolved five grams of the tartrate in sixty cc. of water, cooled to 20°, and added five-tenths gram of the tungstate in twenty cc. of water. T h e mixture was made u p quickly to 100 cc. at 20' and polarized in the z o o nini. tube immediately and after intervals of five minutes. I fouiid without refilling the tube : 1st observation a, = I I .66" 2nd " - 12.03' " " = 12.13' 3rd 4th " - 12.53' 6 6
$ 6
T h e solution became now too turbid, from separation of a precipitate, for further tests. On standing some hours the remainder
98
J . H . LONG.
THE: P R E C I P I T A T I O S O F
of the solution in the flask cleared aiter subsidence of its precipiI portion of this exanlined gc1x-e tate. . 12.j-F'.
(?:>=
Another portion of the same solutiou, heated and then cooled to 20°, gave the same. A new solution prepared in the same inaiiner gave, after standing sonie time, cy>)=
ILj.II9,
I made next a solution containing in tartrate and one gram of the tungstate. diately, a t zoo, gave (ID=
IOO cc.
five grams of the This examined, iinrne-
g.4zJ,
b u t it soon became turbid and deposited a precipitate. clearing, I found
After
= I I .&io,
which increased to 11.87' by heating the liquid. These reactions are especially interesting when compared with those between the tartrates arid other salts. \Ve have here, as before, a marked decrease in the specific rotation on mixing the solutions of the active and inactive substances. But in the case of the tungstate on standing, there is a n increase instead of a decrease in. the rotation observed in the other cases. This behavior finds its explanation probably i n the action of the liberated tungstic acid. T h e reaction between the two salts undoubtedly follows this equation : 2KSbOC,H,O,+ lia,WO,=zKNaC,H,O,f Sb,O,+ (U'0,)x. T h e precipitation of the tungstic acid is very slow and incomplete. \Vhile in solution, it ilia)- conibine with the soluble tartrate to form a body with iiicreased rotation, the possibility of which is shown by the researches of Gernez and others. T h e delay in the appearance of the precipitate niay be due to the formation and slow breaking up of intermediate products containing the antimony and tungstic oxides in teniporarily soluble condition. With liberation of the tungstic acid we have a gradual increase in the already decreased rotation. This change in the rotation, before precipitation, is well s1ion.n in the following observations. I dissolved five grams of the tartrate and two and five-tenths grams of the tungstate, mixed at a low tempera-
A N T I 3 I O N Y FROM SOLUTIOXS OF. T.\RTAK
EMETIC.
99
ture and made up to 1 0 0 cc. as before, and at exactly 2o'C. A reading with the 2 0 0 niin. tube was made as soon as possible and, without changing the solution, this was repeated a t irequent intervals, T h e results were as follows : 3 hours, 28 min. aD= 5.66" " = 5.96" = 6.45' = 6.83' = 7.15" < '= 7.32' 3 hours, 44 min. = 7.43@ 3 hours, 58 min. = 7.90' At this point the liquid began to grow slightly turbid, so that the observations could not be continued. T h e remainder of the liquid was then heated in a closed flask to complete the precipitation, cooled to 20', and tested. I found now (YD = 9.13'. T h e liquid still remaining was filtered, fifty cc. of the filtrate taken and precipitated with hydrogen sulphide, yielding finally 0.4045 gram of antimony sulphide. From this it appears that of the tartrate originally taken, I . jgg grams remained in solution in the 100 cc. (no allowance being made for the volume of the precipitate), From this we have KSbOC,H,O,.+H,O decomposed = 3.401 grams. 'I remaining I ,599 '' KNaC,H,0,.4€1,0 formed 2.889 Now, the rotation corresponding to the tartar emetic reniaining is 4.50' and that to the Rochelle salt formed is 1.25' from which we should expect a total rotation of 5.75'. This, in fact, corresponds very nearly to what was observed a t the begiiining of the test, and seems to bear out the suggestion made above : s i z . , that the principal reaction here occurs before actual precipitation appears. A splitting of the tartar emetic is indicated b y the ininiediate decrease in the rotation aiid then the complex effect of addition of the liberated tungstic acid to the alkali tartrate in solution appears from the gradually increased rotatic11 < (
i'
L L
' (
' 8
I '
J . H . LONG.
IO0
THE PRECIPIT.LTION O F
Precipitatioii finally follows as the elid of the reaction ; the separated part assumes the insoluble form. I t will be recognized that the phenomenon in this case is much more complex than i l l tlie other. There is nothing to show that while the rotation is 1)eiiig increased the action of the tuxgstic acid there is not also R tentleiicy toward decrease because of progressive decoiiiposition of the potassium antinion!-l tartrate. 111a11 probability tlie observed rotation is a resultaiit effect. KEACTION LVITH S O I > I r X THIOSI-LPII.ITE.
Cold dilute solutions of the thiosulphate arid potassium antimonyl tartrate can be mixed without immediate precipitation. Application of heat, however, produces a light yellow precipitate which grows deeper and finally heconies bright red. T h i s precipitate is the substance commonly known as antimoiiy cinnabar, used as a pigment, arid on the large scale is made by decoinpositig the thiosulphate b y antiinoriy chloride. In the reaction between the thiosulphate and tartar enietic the precipitate appears very heavy, but the decomposition is far from coiiiplete as shown by the figures given below. There seeins to be some doubt as to the coniposition of this precipitate. Roscoe and Schorleiiiiner (2, part 2 , 324) give it as, probably, SbS,O, referring, however, to other formulas. Dammer’s Handbuch gives Sb,O,.Sb,S, as the probable formula. Others are also given. Recently Baubigny (Comnflf.rend., No. 1 7 , 18~4)has given reasons why the formula Sb,S, should be considered the correct one and the proof he presents appears to be satisfactory. There remains a possibility, however, that tlie composition may, under certain circumstances, vary with t h e method of preparation. I n fact, some of my owti results seeill to show this aiid I ani now engaged in studying the question further. But as made in the reaction i n hand the composition seems to be zSb,S,.H,O. This was determined by the following considerations : T h e precipitate dissolves in hydrochloric acid without liberation of sulphur, yielding a perfectly clear solution. A solution iiiade in this way was heated, mixed with a little tartaric acid solution, and precipitated by hydrogen sulphide in the usual manner. On filtering off the orange-yellow precipi~
xs'mIow
FROM
sorxrIoNs ov
TAKTXK EMETIC.
IOI
tate so obtained in a Croocli crucible, drying at 1 2 o o , and neighing:.,the weight was always found less than that of the aiitiiiiony ciiixabar taken. Sulphur deteriiiinations were made by dissolving a gram or less of the substance i n strong S-free solution of potassiuiii hj.clroxide and tlieii oxidizing the sulplio-salt foriiied by waslied chlorine gas, (iiiethotl of liivot). T h e results of these tests were as follow : Calciilated for zSb,S,.H,O.
SI,. ............... 69.56
s .................. 27.83
Found.
69.80
27.72
No deterinination of the water was made, but its presence was s h o ~ v niii the substance dried at 120' by heating to a liiglier temperature in a narrow glass tube. 111 a series of experiiiieuts on tlie precipitation a number of portioiis of the tartrate, of five granis each, were weighed out am1 dissolved in 150 cc. of water. Varying aiiiounts of the thiosulpliate i n fifty cc. of water were added and then water enough to iiiake exactly 2 5 0 cc. T h e flasks holding tlie mixtures were closed with perforated stoppers coiitaining long glass tubes and then heated iii the n-ater-bath one hour. I n this way evaporation was practically avoided. At the end of tlie hour the precipitates were collected 011 a Goocli funnel, dried at 1 2 0 ' ~ and weighed with tlie following results : K O . o! expcriI l l C Ilt,
................. 2 ................. 3 ................ 4 ................. 5 . ................ 6 ................. I
7 8
Weight of precipitate.
..
0.1 grain. 0.2
0.00~39gram. 0.0068 ''
0.4 I ' 0.8 I .6 grams.
0.0363
3.2
' I
0.0803
(' ''
'I
0.2112
((
................ 6.4 ................. 12.8
''
0.0111
"
0.0178
'(
0.4809
In a second set of tests I dissolved, i n each case, ten grains of the tartar eiiietic i n roo cc. of hot water aiid added the tliiosulphatr i i i fifty cc. of hot water, T h e iiiixtures were kept a t 100' two hours aiid tlieii filtered. Results as follows :
THE PRECIPIT.ITION OF
J . 1%. LONG.
I02
Weight of A7alS 0 , j H , O
N o of espcri-
abed
mriil
9 ..................
......
I2
........................
20
W c i p h t of precipitnlr. 0.1y
15
..
0 . ~ 0 i I
..
o. j I 6 2 0.6818
It is evident from these figures that in both sets of experiments the reactiori is far from complete a n d not easily determined. I t is, perhaps, quite coniplex. I noticed in no case the escape of hydrogen sulphide or sulphurous oxide and the gradual change in color during precipitation from very light yellow to bright red ti. suggests tlint it takes place i n two stages. Vortrnaiiii (11~~. them. G f s . , 22, 2307) has studied the general problem of deconiposition of thiosulphuric acid arid states that i t l m a k s up into H,iS, 0 , and SO,. In presence of certain metals, tetra- and pentathionates seein to be formed. I t is certain that no sulphate is foriiied in the reaction in hand, but tlie proof of formation of the several thionic acids is dificult because of the incoinpleteness of the reaction and t h e presence of the great excess of thiosulphate. In cases of complete reaction, however, Vortiiiann and Vaubel, also, ( B e y . d. chenz. Gcs., 22, 2703) have shown that these acids are formed. I suggest, therefore, this explanatiori of the present reaction. * i t the outset there m a y be, as with borax, a deconipositioii according to this equation, 2I
