T H E MECHANICAL CONDITION O F COAGULA AND I T S BEARING ON THE THEORY OF COMPLETE COAGULATION. I * BY SHANTI SWARUP BHATNAGAR, KRISIINA KUMAR MATHUR AND
DASIIARATH LAL SHRIVASTAVA
Much valuable illformation regarding the mechanisni of the coagulation of colloids has been obtained by a rigorous chemical esaniination of the precipitates formed as a result of the action of electrolytes on colloidal solutions. In fact the experiments of Linder and Picton’ and Whitney and Ober2 on the complete chemical analysis of the precipitates form the pillars on which thc theory of coagulation by adsorption has been raised by H. Freundlich. A thorough and accurate examination of the physical and mechanical condition of coagula ought to be of considerable importance in elucidating the properties arid beliaviour of colloidal solutions from which the coagula havc becn formed. It can be seen a priori, for example, that the coagulum of “primary particles” of a colloidal solution would have smaller aggregates if the original sol consisted of smaller particles and of larger aggregates if it was composed of larger particles. Thus it would be possible to get roughly at thc average original Pizes of various colloidal solutions from the sizes of aggregates in the coagula which preserve their identity and which are similarly obtained by the action of an electrolyte. For even a rough comparison of the sizes of colloidal solutions very coniplicated apparatus and mathematics have to be used. Tt is probablj. on account of these difficulties that the effect of the sizes of particles lias not been taken into consideration in any of the well-known theories of coagulation, although such an effect can be easily shown to be important in the following manner:Let E be the electrical charge on the ions required to produce “the sign of turbidity” which indicates coagulation of a sol containing particles of equal sixes of radius T r2 ThenQ=V --k d where V is the potential difference, k the dielectric constant of water and d the thickness of the electrical “double layer” and Q the quantity of charge on each particle. If we assume that the theory of electrical neutralisation of particles represents the process of coagulation faithfully then we have E=nQ where n is the number of particles coagulated by the ions. Keeping the coni
d . Chem. SOC.61, 114 (1892).
Z.physik. C h m . 39, 630
(1902).
388
S. S. BHrlTNAGAR, K. li. ,MATHUR AND
I).
I,. SHRIVASTAVA
centration of the dispersed phase as before let us assume that the size of intlividual particles is reduced to a smaller radius q3 r? then also Q1 = V --k d The number of particles is now evidently increased in the ratio 1.3: rla The amount of electrical charge required to neutralise all the particles in the second sol is ra E l = n -QI r?
1.9
=n -V-k rla
r9
v
rl
d
= n -.-k
E
rl? d
r -E; rl
.E1
.-
E
- -r
rl
i.e. the amount of electrolyte required t o coagulate a given concentration of a sol would increase with the decreasing size of the particles. Experimental evidence of Sven Oden on sulphur sol is in agreement with the theoretical conclusions deduced above. Sven Odenl investigated the effect of the degree of dispersity of the sol on the coagulation of sulphur suspensions and came to the conclusion that the electrolyte concentration necessary for coagulation increased with decreasing size of particles. If the theory of coagulation based on the mere preferential adsorption of ions without any reference whatsoever t o the electrical nature of the adsorbent ions, be adopted as our guide to the process of coagulation, one would not expect any relationship between the valency of the precipitating ions and the structure of the precipitate. On the other hand if the adsorption of ions by the particles is an electrical process, the structure of the coagulum would indicate to some extent the nature of steps in which combination took place t o yield the particular precipitate. The structure of the coagula would probably show some relationship with the valency of the precipitating ion and the degree of dispersity of the sol. edition.
Cf. Hatschek: “Introduction to the Physics and Chemistry of Colloids”,p. 58, 4th
MECHANICAL CONDITION OF COAGULA
3 89
Further, the fact that the coagulation’ “at any stage” is quant,itatively reproducible, postulates that a particular combination of particles (which represents the particular stage) will always be formed provided the conditions of the experimentation remain unaltered. Thus the prospects of obtaining valuable information regarding the mechanism of coagulation by studying the coagulum at any particular stage of coagulation seem to be very promising. It is in order to get evidence from this new direction, namely, from the structure of the coagula, for or against the present theories of coagulation and to find any relationship if it exists between the valency of the precipitating ions, the original sizes of the colloidal solutions and the particles present in the coagula that the present investigation was primarily undertaken. Experimental Stokes’ Law for the fall of particles in a viscous medium can be applied to the sizing of particles of coagula. According to this law
where V = cpnstant velocity of the particles p and pl the specific gravities of the parti-
cles and the liquid respectively. 9 =viscosity of the latter, and g = gravity constant.
The above law has been applied by the geologists for the separation of sand, mud and silt in the mechanical analysis of the soil and various forms of elutriators for work have been devised by Schoene, Stadler, Cook and others. Sven Oden2 has developed a method of calculating the distribution of sizes from “the accumulation curve” obtained by plotting time against the weight of accumulated material on a pan suspended at the bottom of a cylinder containing a suspension of clay. This fascinating method was first employed t o obtain information regarding the mechanical condition of coagula of arsenious sulphide, antimony sulphide, manganese dioxide and copper ferrocyanide. The sol of arsenious sulphide was prepared in the manner advocated by Linder and Picton by dissolving Kahlbaum’s pure arsenious oxide in about 1 2 litres of twice distilled water and passing pure sulphuretted hydrogen in the cold solution. The excess of sulphuretted hydrogen was removed by bubbling hydrogen in large quantities for several days. The absence of the uncombined gas was tested for by the lead acetate reaction. The solution was filtered before use and was stocked in two large hard glass bottles which were kept in a dark room. Similar precautions were taken in the preparation of large quantities of colloidal solutions of antimony sulphide and Merck’s pure potassium antimony tartarate was used in the preparation.
* Cf. Smoluchowski: Z.physik. Chem. 92, 192 (1917);Willows: “General Discussion on Colloids,” Trans. Faraday, SOC.and Phys. SOC.Oct. (1920). * Trans. Faraday SOC.17, Part I1 (1922).
’
S. S. BHATNAGAR, K. K. MATHUR AND 1). L. SHRIVASTAVA
390
The colloidal solution of copper ferrocyanide was prepared according to the method described in Zsigmondy’s book on colloidal chemistry and Merclr’s pure reagents were used throughout. The preparations of colloidal solution of manganese dioxide presented some difficulty. The precipitation method of getting the sol proved quite useless. The sol was never stable for more than an hour and fractional coagulation set in automatically during the process of dialysis as well as in the stock bottles. The method of producing colloidal solution by the reduction of potassium permanganate by manganous sulphate according to the equation z
K Mn 04+3Mn S04+7H20=zKHS0,+5 MnO2+H2SO4+gHz0
is also useless for our purpose; the sol so obtained being unstable unless some protective colloid like gelatine is added to it. The addition of a substance like gelatine complicates the process of coagulation and interferes with the purity of the sol. The following method was finally adopted for preparing a colloidal solution of manganese dioxide. N/zo solution of Merck’s pure potassium permanganate was prepared and boiled vigorously. Fairly concentrated solution of ammonia was gradually added’to the boiling solution and a small portion of the solution was taken out and coagulated with sodium chloride from time to time until it showed complete absence of potassium permanganate. This could be determined very accurately by noting the absence of colour in the supernatant liquid after the coagulum had settled down. The colloidal solution thus obtained could be preserved for months in hard glass bottles well protected against the entry of dust by suitable stoppers. 500 C.C. of colloidal solution were as a rule coagulated by a known amount of an electrolyte. The two were thoroughly mixed by a clean glass stirrer. The coagulum was next transferred to the cylinder for noting the sedimentation equilibrium according to the method of Sven Oden. Although all the precautions necessary for the constancy of conditions were observed, the results obtained for a particular coagulum were never concordant even when the number of trials made was very large. As a great deal of time was spent in repeating the experiments it would be of interest t o show the best results obtained by this method and they are given in Tables 1 and 11.
TABLE I Coagulum of Manganese Dioxide obtained by coagulating 500 C.C. of Manganese Dioxide sol with 50 C.C. of M/IO solution of Barium Chloride. Weight.
Time for Trial I.
Min. gm. . IO gm. .IS gm. .18 gm. .os
See.
Time for Trial 11.
Time for Trial 111.
Min.
Sec. 5 50 5
4
50
55
3 4 8
10
45
0
24
0
17
IO
3
20
5
30
11
47
Min. 2
See. 45
391
MECHANICAL CONDITION OF COAGULA
TABLE I1 Coagulum of Antimony Sulphide obtained by coagulating Ant,imony Sulphide sol with 50 C.C. M/IO Barium Chloride. Weight.
Time for Trial I
gm.
5 Min.
o Sec.
.IO
))
IO
))
20
.I2
))
16
”
IO
.I5
’)
39 53
)’ ))
45 30
of
Time for Trial 11.
.os
.I 4 )’
500 C.C.
)’ )’ ))
1,
8 Min. o Sec. 14 )’ 40 ’’ 24
;,
33 45
’) )’
35 45 40
>I
” ))
It is evident from the results that the time required to accumulate a given quantity of the coagulum on the pan is not the same in all experiments which should be the case if Sven Oden’s method was applicable to these coagula. Therc is probably a slight tendency in the coagula to form a net-like structure if allowed to stay a t rest for a long time and this is likely to interfere with the reprotlucibilihy of results by Sven Oden’s me.thod. Reproducible results coiild, however, be c.btaincd by adopting an elutriator which one of us (K.K.M.) had used at the Royal School of Mines, London. A complete drawing of the elutriator used is given in Figure I . The elutriating vessel E was set in position according to the epecial requirement of the experiment as shown in the figure. It was attached t’o the piezometer tube P, by a long rubber tube to thc sliding cistern C. The wate was kept at a particular level-head in this vessel by adjusting the entry of distilled wa8terfrom a large reservoir R so that the water just oozed out into theoverflow veeel 0 through a glass tube. The arrangemcnt for pumping distilled water to the reEervoir It are self-explanatory in the figure. The pumps served a double purFIG.I pose vie., that of pumping wat.er and of filtering the particles as they came out through the nozzle. Precautions were taken to see that the distance between the nozzle and the piezometer scaIe remained constant during experimentation. The nozzles were standardized
6 . S . BHATNAGAR, IC. K. MATHUR AND D. L. SHRIVASTAVA
392
by placing the elutriator in position as shown in the diagram and weighing the quantity of water that flowed out in a minute at various heads through the nozzle. Three nozzles were tried. The area of the cross-section of the elutriating vemel was next determined a t various positions and the mean was found to be 9.6 sq. cm. The velocity cms./min. was calculated by dividing the amounts of water obtained at various heads by the area of the cross section. The velocity head curves are shown I20 in Figure 2. From the curve it is possible to read out accurately the velocity a t 100 any position for any nozzle. 80
Hydrosols of Antimony and Arsenic Sulphides
60
500 C.C. of the colloidal solution of arsenious sulphide were taken and coagulated with 50 C.C. of M/IOsolution of 20 barium chloride, this amount being the 0 minimum quantity required to produce immediately the “sign of turbidity” ccrresponding to complete coagulation. The supernatant liquid was removed by dccantation and the whole of the prccipitate was transferred to the elutriating vessel. The apparatus was then put into operation and different fractions of the particles were collected a t various velocities. These fractions were separately analysed. The method of analysis used for arsenic was the well-known iodometric method. This method did not give satisVEL.WAT€R /N €iUTR/AT/NG !-’&S€L factory results. At this time there apFIG.I1 peared a paperf in which the defects of this method were pointed out and also some precautions given to counteract these. But not being satisfied with the results obtained even with the observation of these precautions another method by Kessler2 which gave us quite concordant results was taken up. This method is also applicable in estimating antimony. The various fractions obtained in the case of antimony sulphide sol vith different electrolytes are as given in the following Tables. 40
1
Analyst. August (1922). 118, 1 7 (1863)
* Pogg. Ann.,
MECHANICAL CONDITION O F COAGULA
3 93
TABLE 111 500
C.C.
of Antimony Sulphide sol coagulated with M/I Sodium Chloride
Solution. Nozzle Head.
Nozzle I. 3.8 c.m. 81.7 ’) 126.0
)’
Equivalent quantity of Sb203 present in the fraction.
0.127 gm. 0.885 )’ o , 1 3I ”
Velocity of wat.er in the elutriator.
4.5 cms./min. 11.0 ” 13.5
’?
TABLE IV 500 cc. of Antimony Sulphide Coagulated with 50 Chloride.
Nozzle Head.
Nozzle I. 4.5 em.
of M/IO Barium
C.C.
Equivalent quantity of Sb2O3 present in the fraction.
Velocity of water in the elutriator.
0.087 gm.
4.6 cms./min. 6.6 ’’
22.0
”
0.011
’)
80.0 90 . o
”
0.079
”
10.9
!’
”
0.022
”
11.5
:‘
0.397 0.513
)’ ,)
25.7
))
28.0
)’
Nozzle 111. 40.3 cm. 50.0
Jl
TABLE V 500 C.C. of Antimony Sulphide coagulated with 1.3 gm. of Aluminium Chloride in Solution.
Nozzle Head.
Nozzle I. 35.5 cm. 8.9
”
Nozzle 111. 90.0 cm. 130.o cm.
Equivalent quantity of SblOl present in the fraction.
0.513 gm. 0.126 ” 0.065
0.307
” 99
Velocity of water in the elutriator.
7 . g cms./min. 11.4
37.1 41.8
>)
” ”
Similar fractions were obtainable in the case of arsenious sulphide hydrosol. Colloidal Solution of Manganese Dioxide 500 C.C. of the colloidal solution of Manganese Dioxide prepared in the manner described above were coagulated with sodium chloride, barium chloride, and aluminium chloride solutions respectively. I n all the three cases it was found that the coagula consisted of only one size of particles and the results are shown in Table VI.
3 94
S. 8. BHATNAGAR, K . K. MATHUR AND D .
L.
SHHlVASTAVA
TABLE VI Precipitating Electrolyte. Quantity of the electrolyte used.
Piezometer Head
Sodium Chloride. solut J?
The above results were surprising and it was thought t,hat these extraordinary results might be due to the fact that the original colloidal solution consisted of only one size of particles. It was, therefore, considered necessary t o verify the conclusion on colloidal solutions of one size of particles only. Recourse was taken to ultra-filtration after the manner of Bechhold and the rrsults obtained are shown in Tables VU-VII1:-
TABLE VI1 Copper Ferrocyanide Sol after Ultra-filtration. Precipitating Electrolyte
Piezometer Head
Sodium Chloride Barium Chloride Aluminium Chloride
cm. 57 . o cm. 9 1 . 4cm. (?)
Nozzle Number
11. 11. 11.
3.2
Elutriation Velocity
4.8 cm./min. 9.95cm./min. 12.0 cm./min.
TABLE VI11 Antimony Sulpliidc so1 after Ultra-filtration Prrcipitating Electrolyte
Sodium Chloridc Barium Chloride Aluminum Chloride
Piezometer Head
6.1 cm. 65.9 l 1 0.7
Nozzle Number
I1 I1 111
Elutriation Velocity
5.4 cms./min. 10.5
14.5
11
(1
Conclusion A close examination of the results of elutriation shows that an interesting relationship exists between the precipitating electrolytes and the sizes of the particles in the coagula. In the case of sols of one size of particles only the ratio between the elutriating velocities for the particles of the coagula obtained by mono-, bi-, and trivalent electrolytes are as I :2 :3. In case of coagula obtained from colloidal solutions having more than one size of particles and giving a series of graded fractions on elutriation, th(e above ratio still holds good when the largest, elutriation velocities alone are compared. The results are shown in Table IX for the sake of closer comparison. It is evident, from these results that the valency of the precipitating electrolyte is an important factor in determining the size of particles formed in their precipitation. In the case of a spherical particle falling freely in a viscous
MECHANICAL CONDITION OF COAGULA
395
S. S. BHATNAOAR, K . K . MA'IHUR AND D. L. SHRIVASTAVA
396
liquid the velocity of fall which corresponds to the elutriation velocity is proportional to the square of the radius, that is, the surface of the particle. If we can assume that the elutriation velocities in the case of coagula are proportional to their surfaces these results will have more than an empirical importance. It would appear that the magnitude of the electrical charge present in the precipitating ion would determine the surface of the ultimate particles formed tis a result of coagulation.
Summary It has been shown that the elutriator can be used for grading the sizes of the particles present in a coagulum. Concordant results have been obrained by using a special adaptation of this instrument. 2 . It has been shown that the largest elutriation velocities required to I.
separate the final size of particles in coagula formed by precipitating the colloidal solution by mono-, di-, and trivalent electrolytes are approximately in the ratio of 1:2:3. 3 . This provides a new evidence in favour of the electrical adsorption theory of coagulation of colloids. 4. It has been shown that colloidal solutions containing one size of particles form a single size only in the coagulum also. 5 . Further experiments with the object of throwing light on the phenomenon of coagulation are in progress on the lines indicated above and it is hoped that a physical explanation of the interesting relationships so far observed will be forthcoming. Geological and Chemical Laboratories, Renares Hindu U n i c e r s i t ~ , Benares.
