D E T E R M I N A T I O N O F POTASH, ETC.
453
the precipitate in a warm room and its consequent partial drying. There is no evidence here nor in the previous cases of the formation of any definite thiosulphate and we would question its existence under ordinary conditions. There is no .evidence of the formation here of an hydroxide as one of the authors quoted states. Basic salts seem to be the only products. UNIVERSITY O F N O R T H CAROLINA.
March,
1%~.
ON SOIlE CONDITIONS AFFECTING THE ACCURACY OF THE DETERIllNATlON OF POTASH A S POTASSIUM PLATINICH LORI DE. BY A. I,. WINTON,Connecticut Agricultural Experiment Station. Received April I, ~ 8 9 5 .
1
N the course of some determinations of potash in potassium
chloride, made by direct evaporation with platinum solution preliminary to a further study of methods, it was observed that the Dotassium platinichloride was more finely divided and the results were higher when the solution of the potash salt y a s concentrated on adding the reagent, than when it was dilute. I n order to ascertain the real connections of the facts observed, I then made a series of experiments which were all conducted in exactly the same way, except for the kind of dish used, the dilution of the solution of potassium chloride on adding the platinum solution, the presence or not of added hydrochloric acid, the temperature of the water-bath, and the strength of alcohol used for washing. T h e potassium chloride used was from a lot sent by Dr. H. J. Wheeler, Reporter on Potash, for the Association of Official Agricultural Chemists. After being finely ground, it contained 0.45 per cent. of moisture, which was determined by heating below a red heat over a lamp until constant weight was secured. I n each case, approximately one-half gram was weighed out from a weighing bottle. After solution in water and adding platinum solution, the mixture was evaporated to a pasty condition, and about twenty-five cc. of alcohol were added and allowed to stand on the potassium platinichloride for one hour with repeated stirring. T h e precipitates were then collected in
454
A . I,. X V I N T O N .
I)ETEK311ShTIO?; 01:
Gooch crucibles, waslied with aboui ioo cc. of alcohol, and dried to constant weight, first at I O O ' C . and finally at 130' C. T h e results calculated to potassiuni chloride b!. the use of the factor 0.30j6 and expressed i n per ceiits. oi the quaiitit!. taken, are given in Table I . I t will lie seen from the table that neither the kind of dish used, nor tlie teiiiperature of the solution of potassiuni chloride on adding the platiiiuiii solutioii, nor the pres-lice of a large excess of hydrochloric acid, during evaporation, nor the teniperature of evaporation, had any perceptible influence on the results. These poiiits, therefore, ne& not be furtlier considereti. In most of the expcriiiients ninety-five per cent. :ilcohol \vas used i n order to diminish the solubility of the p0ta:isiulTi platinichloride. 111 experiments Nos. 9, 19,2 9 , a n d 30, Iiowever, the alcohol was eighty per cent., and the results are, for this reason. soiiie what low e r . The e f e c t of' Dilutioii of the Solution zuheTi Potassium Chloi-ide with no Adwii.t.tuw i s ziscd.-Tlie principal point brought out by the experinients is tlie effect of tlie dilution of the potassium chloride solutioii \vlieii chloroplatinic acid is added, on the character of the potassiuiii platitiichloride, ant1 the results. If the potassiuiii chloride were dissolved in fifty cc. of water, the potassium platinichloride was either iiot precipitated at all on addition of the platiiiuni solutiori, or else went ixto solution when the liquid became heated. and was then deposited slowly on evaporation. The potassium platinichloride thus obtained x a s invariably in brilliant, orauge-colored granuiar crystals, which were often united to forin plates froni one to five nini. across. Constant weight was secured after drying tlie double sait for three hours a t 100' C., the results of several deterii!iiiations being concordant and slightly below theory. T h e loss 011 iurtiier drying at 130' was very small. But if, on the other hand, ten cc. or less of water were used to dissolve the potassium chloride, a coiisiderable portion of the potassium platinichloride precipitated as a fine powder immediately on adding the reagent, the remainder being deposited on evaporation. After treating with alcohol and filtering,
POTASH AS POTASSIUM PLATINICHLORIDE.
so.of
rsperimeut. water used to dissolve
:,
KCI.
cc. HCI :oncentrated added. HlPtCl added to t o t or cold solutiou.
Volume per cent. of alcohol used for washing.
Kind of dish used lor evaporation.
>rams of dry KCI taken.
I
4-F-95
456
- 2 . L . WIS'l'OS.
D E T G K h I I S A T I O N OF
the salt presented the appeararice of a fine, yellow powder. with generally an occasional granular crystal. Constant weight a t roo7 was only secured after heating froin nine to fifteen hours, the results, althougli not always concorda~it, being in every case considerably above theory. Heating for'maiiy hours at 130' C.further reduced the results, although even after this treatment they were still slightly above theory. T h e difference i n the results obtained in dilute and coiicentrated solutions, is brought out strikingly by the averages given in the following table : TAHI,I;11. .4VERAGES
01;
KlCSULTS OIlT.AIKE1) IN DII,LT€$ .%XI1 COXCEXTKATED SOLI'I'cr cent. KC1 fouiid.
Dilutioii of t h e solution 011 addiiig
p r r cent. c,f
H2l'tCI6.
alcc1hol.
j0 C C .
95
I O cc. or
less 50 cc. j ''
Yol.
Xvera:.c of r e s u l b of ;ire practicall!. equivalent to determinations on potassium sulphate with the additioii of hytlrochloric arid sulphuric acids, the aiiiount of the latter acid being about double that necessary to combine with the potassium to form the sulphate. Since in these experiments the results were practically the same as when no sulphuric acid was added. it is reasonable to assume that the effect of the dilution of the solution is the same whether the potash exists a s chloride or sulpha t e. I n both dilute and concentrated soluticns the results obtained i n the presence of sodium chloride or magnesium sulphate, were lower than when determiiiatioiis were niade on potassium chloride with no admixture, or with only sulphuric acid. In the presence of sodium chloride this vias due partly, if not entirely, to the smaller quantity of potassium chloride taken. Since about the same amount of alcohol was used in all cases, the percentage error due to solubility of potassium platinichloride iii alcohol would be greater, the snialler tlie quantity of the salt take;], although the actual weight lost might be the same. In the :.WO experiments where magnesium sulphate was present, the washing with Gladding's reagent and the secoiid washing with a!cohol undoubtedly increased. the minus errors. I'ossibli. the n x r e presence of sodium and magnesium salts depresses the results. T h e fact that the results are lower does not impair their value for the present purpose. as in each case the experiments with the saiiie atiriiixture were both niade oil nearly the same weight of potassium chloride and in exactly the same manner, except as regards the dilution of the solution. Rcszdfs o j Fw.re?zixs.--R. Fresenius' in an article published in I 87 j describing his method for determining potash a s potassium platinichloride, directs to add platinum solution to a co?zccntrafed solution of the potash salt, wash with cight3, Per ccnf. (by volume) alcohol, arid dry the potassium platinichloride to constant weight a t 130' C. I n a later article' lie gives tlie iollowing results obtained by this 1 Zl.cfi?. una!. C h m i . . 1 6 , 6 3 . 2 Zlschv.
a n a l . Ciiem.,
2 1 , 218
461
POTASH AS POTASSIUM P L A T I N I C H L O R I D E .
method on pure potassium chloride, with and without the addition of sodium chloride, using the factor 0.3056. H e states that the potassium platinichloride was obtained in the forin of a j n e powder.
TABLEV. RESULTSOF R. FRESENIUS ON POTASSIUM CHLORIDE WITH AND WITHOUT ADDITION O F SODIUM CHLORIDE. ( Thejigures show the quantity of Potassium Chloride obtained, expressed in p e r cents. of the g u a d i f y taken. Wt.K2F'tC18X 0.3056 = wetght. KCI.) KCland gram NaCl taken.
o.6S1zgram KCI and 0.4524 gram NaCl taken.
5 8
100.1g
100.13
100.11
100.06
11
100.04
100.00
14
I00.OZ
99.93
20
100.00
99.98 99.97 99.94
Hours
Hours dried at
1.0378gram
KCl taken.
dried at
130'. 2
0.9563gram KC1 taken.
100.67
100.42
7
100.40
I2 21
100.23 100.19
100.16 100.04
30
100.10
130'.
0 . ~ 9 gram 0 0.11%
T h e results of Fresenius without addition of sodium chloride are corroborated by my results'in experiments Nos. 29 and 30in Table I, and No. 32 in Table I V . In experiments Nos. 19 to 29, Table I, the results are a little higher, due to the lesser solubility of potassium platinichloride in ninety-five per cent. alcohol. T h e results given in Table I V , where sodium chloride was present, are somewhat lower than those of Fresenius, owing probably to the smaller weights of potassium chloride taken. If the factor, based on the latter and now generally acccpted atoiiiic weight of platinum a s determined by Seubert, had been used, the results would be 0.42 per cent. higher. There must be then some considerable plus error which more than compensates for the errors due to the solubility of the salt in alcohol and the use of the wrong atomic weight of platinum. This error is largely attributable to the presence of water in the potassium platinichloride. Seubert' finds that it is difficult to so prepare this salt that it will be free from '' verknisfcru?zgs wasser," which is not coiiipletely removed even at 160' C. T h e slow drying of'the pulverulent potassium platinichloride a t 100' C., and the further loss in weight after subsequent drying at 1 3 0 " C. is in hariiioiiy with Seubert's experience. T h e results after drying for many hours 1 A ~ I I JCirem., . 2 0 7 ~I .
462
A . I.. W I N T O N .
DETERMIN.ITION
OF
at 130' C. are very close to theory, where eighty per cent. alcohol was used, notwithstanding the solubility of the salt, and are higher if the alcohol were ninety-five per cent. ~Wicrosco~ic Examiiia fion oJi f lt c Pofassiii i n Platiizichloride.--I t has been already stated that the potassium platinic salt, when deposited on evaporation of dilute solutions, was coarsely crystalline, while obtained from concentrated solutions, it was finely pulverulent. Examined uiider the iiiicroscope, the difference was even more striking. T h e coarse form was found to consist of octahedrons, generally attached to one another, foriiiing thin plates, aiid corresponded to the descriptions usually giveii in the books, but when the salt was pulverulent, it consisted largely of curious radiating crystals, a characteristic form coiisisting of three bars intersecting a t right angles to each other. Resting 011 three of the six arms, they presented at first sight, the appearance of stars with six arms iii the same place, but by careful focussing, their true form was revealed. Prof. S . L. Peiifield, of Pale University, has very kindly examined my precipitates, and reports as follows : Both fornis of crystals belong to the isonietric system. T h e coarse crystals being deposited slowly. are normal octahedrons ; the pulverulent crystals, however, owing to their rapid formation, are very iiiuch distorted beiiig developed into rods, crossing at 90", parallel to the direction of tlie isonietric axes. In both the octahedrons and the interpenetrating rods, there were globular cavities which, before drying, were undoubtedly filled with mother-liquor. Only ari occasional cavity could be seen i n the octaliedrons, but i n tlie rods they were very abundant. Even after the long continued heating at 130' C. ocular proof was secured that the octahedrons were not entirely dry. Patient search revealed the preseiiqe of minute cavities from 0.003 to 0.01 mni. across, which were still partially filled with liquid. On incliiiing the stage about 45' and revolviiig it about its center, a bubble could be seen to move from one side to the other exactly as the bubble moves in a spirit level. T h e facts disclosed by Prof. Penfield's examination, explain the slow drying in my experiments. T h e octahedrons, being comparatively free from inclosed liquid, did not lose greatly in
POTASH AS POTASSIUM PLATINICHLORIDE.
463
weight on heating ; the interpenetrating rods, however, contained a very considerable amount of inclosed liquid, part of which was slowly given off on heating at IOO', more at 130°,and still more at 160' C. T h e quantity of liquid enclosed in the octahedrons could not have been considerable, but it shows the persistence with which water, once enclosed, is retained. Conclusions.-The method of precipitating in concentrated solut i o ~ and ~ ~ drying , the potassium platinichloride at 130' depends, for its accuracy, on the compensation of three errors, due ( I ) to the solubility of the potassium platinic salt in eighty per cent. alcohol, ( 2 ) to the presence of water in the crystals which is not driven off at 130', and ( 3 ) to the use of a factor based on the wrong atomic weight of platinum. Such a method is certainly open to criticism. T h e solubility in alcohol occasions an error that can hardly be avoided. I t could be diminished by using ninety-five per cent. alcohol, but further experiments would be necessary in order to ascertain whether for other reasons the stronger alcohol could be advantageously substituted for the weaker. T h e results given herewith show that fhe error occasioned by the presence of wafcr can be p e a f l y reduced and fhe process of drying sinzplzjZed, by adding fhe pCa finum solution fo a dilufe solufion of the pofash salt (om p a r t of pofassiunz chloride or six-tenth part potassium oxide f o 100 cc. of wafer) and dryinf the poiassium plafinichloride a t 100' C. I t is not clainied that the double salt thus obtained is perfectly dry, but it is necessary to heat the pulverulerit forni for many hours at 160' C. in order to secure it as free from water as the granular crystals obtained by the above siniple process. T h e liquid enclosed in the crystals, although largely water, must contain also solid matters, which remain behind, however complete may be the drying. I t must then follow that the less liquid inclosed at the outset, the smaller the error due to this cause after drying. With such a simple method at our disposal for obtaining the potassium platinic salt comparatively free from moisture, the necessity for using an erroneous factor at once disappears. T h e factors based on the atoinic weights as revised by F. W.
464
A . L. W I N T O X .
U E T E K M I N A T I O Z ; OF
Clarke', tip to January I , I 894, woiild be 0.30688 for potassium chloride, and 0.1939 for potassiurii oxide. These factors would make all tlie foregoing results 0.42 per cent. higher iii potassium chloride, which is equivalent to ( 3 . 2 6 per cent. higher in potassium oxide. In Table VI are given comparative results by the oldand new factors, in tlie cases of all the preceding experiments i n which one part of potassium chloride was dissolved in approximately 100 parts of water, and the potassium platinichloride was washed with eighty per cent. alcohol and dried two or three hours at
c.
IOoC
T a u x , ~V I . COMPARISON O F RESV1,TS O e T A I N E D BY THE F A C T O R S
0.3056 AND 0.30688
FOR CALCUL.4TING POT.kSSICM CIILOKIDE FROM POT.kSSIUX
PLATISICHLOXIDE. (Chloroplaii?ric Acid added to l l i l t i f c Solufio?rs of Pofassiiinz C h h i d e , P o t t i ~ ~ i uZ'Lnfi?iiiChloride ~~t zcashca' 7edth eighty per celzf. Alcohol and dried two t o ihrce hours u t zoo' C.! P e r c e n t . KCI fouiid using factor.
Expt
r
Material taken.
No
9 0 . 4 j grain ~ 10 o.jm5 gram 31 0.5331 grain 33 0.2416 grain 35 0.3768 grain 37 0.4765 gram 39 0.4734 giani
KC1.. ......................... K C ] . . ......................... K C l . ......................... KC1 a i d 0.4 gram Sac1 ........ KCl and 0 . 2 grain SaC1.. . . . . . KCl am1 0.6 gram XlgSO,. ..... KCl aud 1.0 gram H,SO,. ......
O . p j 6
0.p5'rj
9?.ji
ICXj.19
99.76 99.71 99.30 99.57 99.26 99.73
100.18
1cx1.13 99.72 99.99 99.68 10 ) . I j
I t will be seen that when potassiutii chloride, either alone, or mixed with sulphuric acid was taken, the results using the factor 0.30688 are 0.13 to 0.19 per cent. above theory. This would indicate either that the water reniaining in the double salt after drying for two or three hours at 100' C. a little more than compensated for the solubility in alcoliol, or else that there are other slight plus errors. When, however, other salts are present, and particularly if 1 r.S . Dept. A g r . , Cheiii. I>iv., Bu!!. 43, p. 340. Potassium = 39 I I . Platiiiuni = 195.00. = r6.m
Oxygen Cbloriiie
=
35.45.
POTASH A S POTASSIUM P L A T I N I C H L O R I D E .
465
the Gladding method were followed, the tendency is toward lower results. Taken as a whole the new factor gave, in these instances, more satisfactory results than the old factor. A method which so reduces the error due to inclosed water in the double salt, that it about compensates for the solubility in alcohol, and which employs the factor based on the revised atomic weights, is, in my opinion, to be preferred to one which depends, for the accuracy of results, on the compensation of three errors, one of which is due to the use of a factor known to be incorrect, especially when the former method gives more concordant results and is more readily carried out. Summary.-When chloroplatinic acid was added to a concentrated solution of potassium chloride, a la'rge part of the potassium platinichloride formed was precipitated at once in a pulverulent form, the remainder being deposited on evaporation, After treating with alcohol, filtering, and drying, the double salt was in a fine powder which, examined under the microscope, was found to consist largely of radiating crystals, a characteristic form having six arms, formed by the intersection at right angles of three bars. Numerous globular cavities in the crystals gave evidence that during their rapid formation mother-liquor was enclosed. Owing to this enclosed liquid the double salt dried slowly at IOO", lost further in weight after heating subsequently and further still a t 160"C. T h e loss at for many hours at 130"~ these temperatures compared with the weight after drying three hours at IOO', was equivalent to about six-tenth's per cent. potassium chloride. When, however, the solution of the potash salt was so dilute that no precipitate was formed on adding the reagent, the potassium platinic salt being deposited on evaporation, the latter was obtained in octahedrons, generally attached to one another forming crystalline plates. These octahedrons were comparatively free froin cavities, and practically the same results were secured after drying three hours at im0,as after drying for many hours at 130' and 160"C. Somewhat lower results were obtained in the presence of sodium chloride and magnesium sulphate, whatever the dilution of the
solution, but iii other respects the above staternetits will hold true. T h e kind of dish used, the temperature of the evaporation, the presence of free hydrochloric w i d a:id free sulphuric acid, did not appear to influence tlie results. X method based on the evaporation of dilute solutions, drying tlic potassiuni platinichloride at I O O ” , arid the use of the factors 0.~;oh8S for potassium chloride and 0.1939 for potassium oxide, gave good results and is beliel.ed to be more satisfactory for scieiitific and practical reasons, tlian the method of Fresenius, i n which the platinum solution is added to concentrated so!utions. tlie double salt is dried niaiiy hours at 130’ C. and the factors 0.30j 6 arid 0.19308 are used.
THE INDIRECT ANALYSIS O F MIXTURES CONTAINING A c o n n o N CONSTITUENT. I>\’ E I ) W . 4 X I l IC. K c < c i , r i i hl3.r.b
I..\SI>IS 1;.
r89j.
F I R S T METHOD.
C
ALCULATION of mixtures containing a coniinon constit-
uent. h t a Kive)i : U’eiglit of mixture. V‘eight of coniriioii constituent in mixture. Let x = one salt. Let y = tlie other. Let n = ainouiit of corntiion constituent in one part of x . Let 6= airiount of coiiinioii constituent i n one part of y. Let z = weight of mixture. Let Q = weight of coninion constituent in mixture. ‘1‘0 find values of x a n d y ; .r =z . a x + by= Q. -+?I
a(z--y)fby= p. az--ny+by= 9. by-an_l,= Q - a z . Q - a z and x = z-y. J = b -a
