MONOMOLECULAR FILMS. * THE SOLID-LIQUID INTERFACE AND T H E SEDIMENTATION AND FLOCCULATION OF POWDERS I N LIQUIDS' BY WILLIAM D. HARKINS AND DAVID M. G h N S
1. Introduction
While it is known that the volume occupied by a powder after long settling in a liquid is dependent upon the composition of the liquid, the fact that monomolecular films at the solid-liquid interface have an extremely great effect upon the settling, has been unknown. The phenomena of settling in both aqueous and organic liquids present many remarkable characteristics. While experiments in both classes of liquids have been carried out, the present paper will be restricted almost entirely to those of the organic type, since they have revealed more new unknown and striking relations. The energy of immersion of a solid oxide in the form of a powder, is very low in a non-polar liquid such as hexane, benzene, or carbon tetrachloride. However, if a small amount of an organic acid, alcohol, amine, or any other similar liquid is added to the non-polar liquid, the heat of immersion is greatly increased. Molecules of this type may be designated as hetero-homopolar, and are commonly represented by the symbol, in which the circle represents a polar group. The data2 on the heat of immersion indicate that the polar grouDs in such molecules are oriented toward the oxide at the solid-liquid interface. Obviously the orientation of the molecules in this interface is the same with a pure liquid if its molecules are of the heterohomopolar type. The purpose of the work reported in this paper was to determine the effect of such an oriented interfacial film upon the final sedimentation, and upon the rate of sedimentation of a powder in the liquid, since these indicate the extent of the flocculation of the powder. The presence of a monomolecular oriented fJm of molecules of the heterohomopolar type is found to prevent flocculation. If, however, molecules of water are also present, the mixed film does not prevent flocculation to the same extent, and if et considerable amount of water is present in the mixed film, the flocculation may be as great as in a pure non-polar liquid.
9
I
-
* Contribution from the George Herbert Jones Laboratory of the University of Chicago. 1 1
This research waa made possible by the support of the Titanium Pigment Co., Inc. Harkins and Dahlstrom: Ind. Eng. Chem., 22, 897 (1930).
MONOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
87
2. Experimental Procedure
In a part of the experiments on settling, about 48 cu. cm. of the liquid was drawn by a syphon into a graduated glass-stoppered cylinder and the stopper immediately inserted. The powder, dried at a high temperature in a high vacuum (8.0 grams), was then weighed quickly and poured into the cylinder, exposing the powder to the atmosphere as little as possible. The tightly stoppered cylinders were then vigorously shaken and allowed to stand, and the volume of the suspension read off at intervals. Other more carefully conducted experiments were carried out in sealed pyrex tubes, with special care to exclude moisture, and were used to check the values obtained in the stoppered cylinders. 3. Preparation of Liquids and Drying of Powder
The titanic oxide (Ti02) used in this work was dried in a high vacuum a t 450'to 550'C. for 24 hours. Thiophene-free benzene was dried over oil-free sodium wire for several weeks. Fresh sodium wire was added several times during this period. The benzene was then distilled, and the distillate stored with a large quantity of sodium wire distributed through its volume. The oleic acid was purified by the method used by Harkins and Beeman, who applied the well known procedure in which the lead and barium salts are used. 4. An Adsorption Method for the Determination of the Area of
the Surface of a Powder In problems which concern settling or those which involve surface energy it is of fundamental importance to have a method for the accurate determination of the area of the surface of a powder. The method described below is believed to be the most accurate of all known methods. The powder, titanic oxide or silicon dioxide for example, is dried in a high vacuum a t as high a temperature as can be used without affecting the area of the crystals. The cool, dry powder is immersed in a solution of oleic acid, butyric acid, or some other suitable acid, in very dry benzene, and the suspension is shaken until equilibrium is attained. After the powder has settled, a sample of the supernatant liquid is drawn off. The benzene of this sample and of a sample of the initial solution is evaporated off, if oleic acid is the solute chosen, and the oleic acid left from each solution is dissolved in gsyo ethyl alcohol. The alcohol solutions are then titrated with carbonate free sodium hydroxide dissolved in water. The difference in concentration is considered to give the amount of acid adsorbed by the surface of the powder. Only air that is thoroughly dried by phosphorus pentoxide is allowed to enter the vessel in which the solution is prepared and the flasks in which the adsorption experiments are carried out. Fig. I shows how the amount of oleic acid adsorbed by the surface of one gram of powder varies with the concentration of the final or equilibrium solution of oleic acid in benzene. At concentrations above 0.01to 0.02 moles per
88
WILLIAM D. HARKINS AND DAVID M. GANS
kilogram of benzene the adsorption becomes practically constant, and the oleic acid on the surface of the grains of powder may be said to form a condensed film. There is some evidence which seems to indicate that this film is monomolecular. It may be assumed that the area occupied per molecule of oleic acid a t the interface is 2 0 . 0 sq. A,, which is about the mean value for condensed films of the acid on water at a zero film pressure. On this basis, the area for Ti02-I was calculated as 22.9 sq. m. per cu. cm. of titanic oxide (3.89 grams), and for TiOpII as 14.4 sq. m. per cu. cm. The ratio of these values is 1.59, while
FIQ.I Adsorption Curves for Dried Titanic Oxides
the ratio for the energy of immersion1in ethyl acetate for the same two powders is I .53, and the ratio of the weights of propyl alcohol adsorbedl from the vapor per unit weight of these powders at a constant vapor pressure equal to one-half the saturation pressure a t room temperature is 1.53. The area of the powder TiO2-I1 as determined for us by a microscopic method: applicable to crystalline powders, is 13.8 sq. m. per cu. cm. of powder, on the assumption of a spherical shape for each particle. This is in good agreement with the value 14.4 obtained by the adsorption method. However, the closeness of the agreement is to some extent accidental, since the adsorption method includes the area of the colloidal part of the powder, which is not included by the microscopic method. However the agreement between the two methods indicates that the film is monomolecular and not polymolecular.
5. Effect of a Monomolecular Film of Oleic Acid on the Settling of Powders The experiments described below show that if a powdered crystalline oxide is allowed to settle in benzene the most complete settling is obtained when the grains of powder are coated with a film of oleic acid one molecule thick. 2
Harkins and Dahlstrom: to be published. Gans and Brooks: to be published. Dunn: Ind. Eng. Chem., Anal. Ed., 2, 59 (1930).
MOKOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
89
A fine powder consisting of titanic oxide was dried in a high vacuum a t It was then suspended in extremely dry benzene in tubes about 2 cm. in diameter. After several weeks the top of the powder attained a constant level. The density of the titanic oxide used is 3.89, and so 3.89 g. of the material may be considered to occupy I cu. cm. This amount of material was found to settle to a final volume of 20 cu. cm., that is, until 19of the 20 cu. cm. was occupied by the benzene between the grains of powder. It was found that if oleic acid was added to the benzene any initial concentration up to 0.005 moles per kg. of benzene produced no great effect 400' C.-for 24 hours.
3
FIQ.2 Variation of Extent of Settling with Amount of Oleic Acid present
(Table I) but a concentration of 0.008 moles of oleic acid gave complete final settling to a volume of 6.5 cu. cm. per cu. cm. of oxide, while if the suspension were kept very dry in sealed tubes as little as 0.0016moles per kg. of benzene produced settling to 16 cu. cm. per cu. cm. of material with sample I and to 8 CL. cm. with sample 11. It!has been shown in Section 5 that oleic acid is highly adsorbed from its solutions in dry benzene by titanic oxide. Harkins and Dahlstroml have shown that this is an effect which is obtained whenever an active or polar group is present in the organic compound used. The values for the heat of immersion show that the active group is oriented toward the oxide and the non-polar group toward the benzene (Fig. 6). They show that powders which consist of the oxides of silicon, tin, zinc, or of barium sulfate, also exhibit a similar great adsorptive effect. The effect of the adsorbed film of oleic acid in producing settfing is shown by Fig. 2 . Harkins and Dahlstrom: Ind. Eng. Chem., 22, 897 (1930).
WILLIAM D. HARKINS AND DAVID M. G W S
90
Thus the presence of 0.005 moles of oleic acid per kg. of benzene in the final equilibrium solution is sufficient to cause complete settling, in which the initial volume of 21.5 cu. cm. per cu. cm. of solid in dry benzene alone is reduced by four times to 5 cu. cm. The addition of larger amounts of oleic acid is seen to produce no further effect.
TABLE I Apparent Volume of Powder after Settling in Benzene in Cc. per Cc. of Solid C = Initial Concentration of Oleic Acid Moles per Kg. Benzene A. T i O r I in Cylinders
7;Water in the Powder
C = o.oooo
0.00
20.3
21.5
19.8
0.05
20.6
21.9
21.1
20.8
20. I
20.0
0.25
20.7
22.0
22.I
21.6
21.7
21.j
0.95
22.2
22.1
22.1
21.6
21.8
22.0
'/b Water in the Powder 0.00
0.05
0.33
0
00145
0.00498
0.00827
0.0208
6.5
6.1
0.04gO
6.3
B. T i O r I in Sealed Tubes C
= o.0000
21.4 -
o.0016
0.0111
5.'
15.8
-
-
0,0354
0.1111
-
5.2
-
14.4 -
18.9
C. Ti0,II in Sealed Tubes % Water in the Powder
C = o 0000
0.00
19.4
0.05
0
33
-
o
0016
7.9
o
0111 5.0
-
-
-
-
o 0354
-
14.I
o
1111
4.7 9.4 -
The curves of Section 4 indicate that at 0.005 moles per kg. of benzene the adsorbed film of oleic acid is monomolecular, but isnot quite tightly packed. It may be noted that the above refers to titanic oxide I or I1 when extreme precautions were taken to keep both the benzene and the titanic oxide very dry, and when the suspension was kept in sealed pyrex tubes. When the suspension was kept in glass stoppered cylinders an even smaller final concentration of oleic acid (about 0.003 moles per kg. of benzene) gave complete settling, but to only 6 instead of 5 cu. cm. That is, the minute amount of water which evidently was able to get into the film, decreased the amount of oleic acid necessary in the film and decreased the settling only slightly. It will be shown in Section 7 that slightly larger amounts of water in the oleic acid film decrease the extent of settling much more.
MONOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
91
6. Effect of Vibration on Settling The minimum volume for one cu. cm. of the solid after settling is found to be about j cu. cm. The density of titanic oxide is 3.89 so that the mass of one cu. cm. of the solid oxide is 3.89grams. The volume of this amount of powder in the absence of the dispersing liquid was found to be 5 . 5 cu. cm. for TiOpI, and 3.j cu. cm. for TiO,II, as obtained by a gentle pressure which was used to remove air pockets. The former of these values is approximately that obtained for the final volume of the suspension, while the latter value is considerably less. I n order to determine the effects of a gentle mechanical agitation upon the suspensions, the sealed tubes were put vertically in a large test tube rack which was caused to vibrate horizontally 400 times per minute with an amplitude of I cm. I p order to increase the effect the tubes were left loose in the racks. After 30 hours of vibration almost no further settling occurred, but the vibration was continued until the total period was 64 hours. TABLE I1 Effect of Tapping on Apparent Volume of Powder after Settling in Benzene C = Initial Concentration of Oleic Acid Moles per Kg. Benzene
A . Apparent Volume in Cc. per Cc. of Solid
yo Water in the
Powder
C
= o 0000
0.0016 I.
0.00
0.Oj
0.33
8.9 -
6.5 -
2.
0.00 0.0s
0,33
0.0354
.O.IIII
3.9 -
6.j
3.8 5.4 -
-
2.8
Ti02-I1 in Sealed Tubes 3.5
8.0 -
0 0111
Ti02-Iin Sealed Tubes
-
3.1
-
-
j.1
3.1
-
B. Ratio of Apparent Volume with Tapping to Apparent Volume without Tapping I. Ti02-I in Sealed Tubes 0.00
0.42
0.oj
0.33
-
0.00
0.41
0.41 2.
0.oj
0.33
-
0.76
-
-
-
-
0.27
0.73 0.38 -
0.36
0.60 0.33 -
T i O r I I in Sealed Tubes 0.44
-
-
0.62 -
92
WILLIAM D. HARKING AND DAVID M. GANS
The final volume (Table 11) of the suspension was thus reduced to 8.9 and 8.0 cc. for TiOTI and TiOrII respectively in the absence of oleic acid, and to 3.9 and 3.1 cc. for an initial oleic acid concentration of 0.0111moles per kg. of benzene. Thus, when the particles are covered with a monomolecular film of oleic acid, the final volume of the suspension becomes in this way as small as or smaller than the volume of the dry powder. After the suspenvions which contained oleic acid had settled to such a compact mass, it was very difficult to cause a resuspension of the powder by shaking, and to this extent the behavior is analogous to that known as the hard-settling of paints. For example, 45 minutes of vigorous shaking by hand was necessary to resuspend the powder in one of these tubes.
FIQ.3 Variation of Extent of Settling with Water Content of Powder
7. The Effect of Water The results of certain experiments on the adsorption of oleic acid by titanic oxide have been given, but an important additional result obtained was that powder with a content of 0.05 per cent water, in the form of an adsorbed layer, adsorbs only about one-half as much oleic acid as the dry powder. Since the oleic acid film is very effective in producing settling of the suspension in benzene, it seemed important to determine the relation between the water content and the settling. The results of the experiments are shown in Fig. 3. Powders which have settled completely to about 5 cu. cm. per cu. cm. of solid, are found to settle less and less as the water content is increased. The data are summarized in Table I. I n the first series (A) of experiments the presence of 0.33 per cent of water with the powder was sufficient to entirely
MONOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
93
prevent the settling due to oleic acid, while in the second and third series the presence of this amount of water was found to overcome largely but not entirely the effect of the acid. 8.
Texture of the Suspension
As has been shown, titanic oxide in dry benzene gives a suspension with a large specific volume, about 2 1 cu. cm. per cu. cm. of solid. The suspension, after settling, is found to exhibit a coarse structure, sometimes with cavities one mm. or more in diameter. Such a suspension is often described as “flocculated.” If the suspension is shaken in the sealed tube, the walls of the glass become coated with a flocculated coating, which does not completely cover the glass, and which may be easily detached. , Very little of the powder adheres to that part of the glass which is under the liquid. Bn interesting phenomenon was observed with such tubes when the powder and dry benzene only were present. If such a sealed tube is shaken vigorously, it gives a spark over a gap of about 3 mm. if the glass is brought near the hand. Tubes in which oleic acid is also present were not found to exhibit this phenomenon, though it is possible that the glass becomes charged to a lesser extent. When the suspension contains enough oleic acid to produce the maximum settling, the texture is very uniform and fine, and the suspension may be said to be deflocculated. With such suspensions the deposit on the glass is much more fine and more adherent. If water is added, the suspension, after it is shaken, is found to be flocculated, and the structure of the suspension, with 0 . 3 3 7 ~of water to powder, is apparently the same as if neither oleic acid nor water were present. 9. Rate of Settling a s influenced by the Composition of the Liquid
The rate of settling of the powder in benzene solutions which contain different amounts of oleic acid and of water is exhibited in Fig. 4. A striking feature of these curves is that in pure benzene the powder settles the most rapidly a t first but settles least in 24 hours. If only oleic acid is added the rate of settling is often practically constant for an hour, and increases with increasing acid concentration up to a concentration of acid (about 0.01 m.) which gives a complete monomolecular film. A peculiar relation is shown by curves b and c. The system is the same for both of these (0.0016 moles of oleic acid per kg. of benzene). Both settle a t the same rate for three minutes and to the same extent in 24 hours. However, at the end of 3 minutes suspension c began to settle much more rapidly than b and departed very widely from the linear course of the latter. Fig. 5 illustrates the behavior of a suspension in a tube which was called the “clock-tube,” since it behaved with such regularity as to give a moderately accurate measure of an interval of time, four minutes in length, and also a shorter interval of about three minutes. The titanic oxide contained 0.33 per cent of water, and was suspended in a solution which contained 0.035 moles of oleic acid per kg. of benzene. For four minutes the powder settled
94
WILLIAM D. HARKINS AND DAVID M. GANS
21 Hours
FIG.4 Typical Settling Curves for Titanic Oxide
with extreme slowness (from 24.4 to 24.2 cu. cm.) a t the top, and at the end of this period it fell with considerable rapidity from 24.2 cu. cm. a t 4 minutes to 17 cu. cm. a t 5 minutes and to 15 cu. cm. a t 6 minutes, while in 60 minutes it fell only to 14 cu. cm. This procedure was repeated many times, after the tube was shaken.
:ime in MUuteS FIG.5 Abnormal Settling Curve. Titanic Oxide plus 0.33% of Water, Solution of 0.035 Moles Oleic Acid per Kg
MONOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
95
At the beginning of the period the suspension appeared to be uniform (not flocculated), and this condition persisted for 4 minutes near the top of the cylinder, but only 3 minutes at the bottom. A t the end of 3 minutes the suspension seemed to begin to settle suddenly and rapidly near the bottom of the tube. This settling drew the powder mostly away from a region about 2 cm. above the bottom of the tube. Then powder fell rapidly into this vacant space from just above, and by the continuation of this process the region almost free from powder moved rapidly up the tube. The effect of n-butyric, n-heptylic, stearic and oleic acid on the rate of settling of graphite, barium sulphate, and silica gel in benzene, has been studied by Rehbinder and his collaborators.’ Their conclusions agree t o a certain extent with those of the present paper, but a comparison is difficult since they studied the settling for about a one minute period only. 10.
Orientation of the Molecules at the Interface Solid-Liquid
The values of the energy of immersion of oxides, such as titanic oxide, stannic oxide, silicon dioxide, and zinc oxide, indicate that in suspensions in benzene the carboxyl group of the acid is in all cases oriented toward the surface of the solid oxide (Fig. 6). Liqdid
Solid
FIQ.6 Orientation of an Acid at the Surface of an Oxide
11. Discussion
It is often considered that the settling of powders in liquids and other related characteristics of the suspensions, are a simple function of the “wettability” of the powder by the liquid, but this is not true. Thus, it has been seen that dry titanic oxide in dry benzene settles only slightly. If oleic acid is present the heat of immersion is much higher, and the powder settles much more. However, the presence of small amounts of water in addition to the oleic acid may altogether prevent any additionalsettling due to the acid. The energy of immersion of the solid oxide in water, or in benzene which contains water, is even higher than the similar values with oleic acid. Thus the correlation of extent of settling with energy of wetting is not a simple one. It has been seen, however, that the presence of a monomolecular film of oleic acid on the grains of powder greatly increases the amount of settling, and causes deflocculation. The effect might be ascribed to a lubricating action of Rehbinder, Lagutkina, and Wenstrom: Z. phyeik. Chem., 146 A, 63 (1930).
96
WILLIAM D. HARKINS AND DAVID M. GANS
the monomolecular film which does not allow one particle of powder to adhere to another. This behavior may be assumed to be due to the electrical relations a t the interface solid-liquid. Thus the particles may be considered as more highly charged when oleic acid is present than when it is absent, but no evidence in favor of this conclusion has been obtained. The energy of adhesion and of cohesion may be involved as well. More work is being done on the various factors which affect the flocculation. The writers wish to thank L. W. Ryan, Director of Research, Titanium Pigment Co., Inc., for many useful suggestions.
summary I. A monomolecular film of oleic acid on the surface of a fine powder suspended in a very dry non-polar liquid, is found to deflocculate the powder, and to cause it to settle to one-fourth the volume it would occupy if no oleic acid were present. This effect is produced by a monomolecular fiIm of oleic acid even if it is not quite a condensed film. Thus the final volume occupied by I cu. cm. of titanic oxide in its flocculated suspension in pure and very dry benzene is 20 cu. cm., but this is reduced to 5 cu. cm, by a monomolecular film of oleic acid on each grain of powder. 2. Vibration of the tubes which contained the suspensions greatly reduced the above volumes to about 8 and 3 cu. cm. respectively. Thus the reduction in volume did not annul the great effect of the film of oleic acid, 3. The addition of very small amounts of water to the solution of oleic acid in benzene was found either to destroy entirely or else to remove partly the effect of the oleic acid in deflocculating the solution and causing settling to a small volume. Thus water caused the flocculation to reappear. The final volume of T cu. um. of titanic oxide powder in dry benzene was found to be 20.3 cu. cm., which was reduced to 6.5 cu. cm. by an initial concentration of 0.00827 moles of oleic acid per kg. of benzene. The presence of 0.05% of water in the titanic oxide increased this volume of the suspension to 20.8 cu. cm. or slightly more than the volume when no oleic was present. Thus water partly replaces the oleic acid in the adsorbed film at the solid-liquid interface. 4. An adsorption method for the determination of the area of the surface of a fine powder is described. The powder is immersed in extremely dry benzene to which small varying amounts of oleic acid are added. The maximum adsorption is found to correspond t o a monomolecular film of the acid. It is assumed that the area occupied per molecule of oleic acid in this saturated film is the same as in a condensed film of oleic acid on water, that is 2 0 sq. A., or 2 0 X 10-l~ sq. cm. The area thus determined has been found to agree as well as could be expected with the results of a microscopic determination. 5 . It is shown that the extent of the flocculation and of the settling is not, as is sometimes supposed, a simple function of the “wettability” of the powder by the liquid. Thus titanium oxide exhibits a high energy of immersion with 12.
’ 1
MONOMOLECULAR FILMS : SOLID-LIQUID INTERFACE
97
both oleic acid and water and a low energy with benzene, yet oleic acid in benzene deflocculates the suspension, while if water is added, the suspension becomes re-flocculated. 6. The composition of the organic liquid in which the powder is suspended has a remarkable influence upon the rate of settling of the suspension. In general the titanic oxide settles most in the end when it settles least rapidly initially.
7 . In general the volume of the suspension, when it settles without vibration, varies inversely with the extent of the flocculation. Cniversity of Chicago, July 87, 19~1.
