LIQUID FILM FORMATION A Criticism of the Balanced-Layer Theory THOMAS H. HAZLEHURST AND HARVEY A. NEVILLE Lehigh University, Bethlehem, Penna.
T
HE only general theory extant relative to the stability of liquid films and foams is the balanced-layer theory introduced by Edser (1) and extended by Foulk ( 2 ) . Foulk and Miller (3) later obtained data which were published as experimental proof, and the theory and data have been quoted as authoritative in widely used textbooks on surface chemistry and in papers on films and foams. Acceptance of the basic assumptions and experimental data in these papers has been without criticism up to the present, although the technical importance of the subject warranted careful scrutiny. It has now been found that the paper by Foulk and Miller (5) to test the theory experimentally contains significant errors, and its conclusions are unreliable. These facts as well as theoretical objections t o the premises of the balanced-layer theory are discussed here. Experimental Evidence
Three methods of experimentation were used to produce data for the paper under discussion (3)-measurements of static, us, and dynamic ug, surface tension to calculate values of Au = us - uD as a measure of the difference in concentration between the surface and bulk liquids, of the foam-zone height on different solutions, and of percentage film formation by the two-bubble experiment correlated with the other two. These measurements will be considered separately. MEASUREMENT OF Au = us - uD. No description of the technique used for measuring dynamic surface tension is given, but reference is made t o the work of Stocker (8) who used elaborate equipment and great care in operation. However he claimed an accuracy of only 0.15 dyne. Consequently we assume the error in the paper discussed (3) to be a t least ten times as large as Stocker’s in view of statements that “the quantitative results are not very precise”, and “no attempt a t . . special accuracy was made”. The curve of surface tension of sugar solutions us. concentration was drawn through every experimental point and resulted in an irregular line. These breaks have been found by no one else and are misleading. The points on both the us and uD curves can be represented by a single straight line within the limits of accuracy of Stocker’s work so that Au is zero over the measured range. The majority of the reported values of A u are positive. This is not possible thermodynamically. Solutes are positively or negatively adsorbed because by this adsorption they lower the surface tension from the value it would have if surface adsorption were eliminated. Dynamic surface tension measurements are designed to eliminate the effect of surface adsorption by measuring the tension of a freshly formed surface. Therefore, any real difference between static and dynamic surface tensions must show the dynamic to be greater than the static. Thus A u must always be negative, regardless of the sign of adsorption. However, Figure 6 or Table VI (5) shows positive values of A u as high as 3.4 dynes. We ascribe this result to experimental error. The actual limits of error, a t least in this case, must be as high as 3.4 dynes, a conclusion not unlikely when the extreme difficulty of measuring dynamic surface tensions is taken into account. Yet no steps were made to smooth out the discontinuities in the two sets
.
of measurements. We conclude that the complete set of measurements of Au must be discredited. FOAM-ZONE HEIGHTS. These measurements are made with a dynamic foam meter, an upright glass tube containing the solution above a porous disk. Through this tube air is blown t o rise continuously in bubbles to the top of the column of liquid, the thickness of the layer of bubbles or foam above the liquid being considered a measure of relative stability of the foam. Measurements of foam-zone height are recorded to 1 mm., with differences of a few millimeters apparently considered significant. Attempts to repeat these measurements with various types of porous septa have failed to yield consistent data of such accuracy as reported (3) by Foulk and Miller. T o produce a foam zone of any considerable thickness from solutions of inorganic salts requires a rapid stream of air ascending the column of liquid. This prohibits a steady or sharp interface between the foam zone and the underlying liquid. Also, the upper surface of the foam zone is irregular and fugitive. The individual evanescent bubbles have diameters of several millimeters, and those in contact with the tube wall are usually larger. Practically any desired foam-zone height can be obtained by varying air pressure. Uncontrollable fluctuations of air pressure are an inherent feature of the apparatus due to discontinuity of the escape of air bubbles. Air pressure must build up to force the liquid from a capillary in the porous disk; this is released by the escape of one or more bubbles, the liquid again seals the opening of the capillary, and the cycle then repeats. The opposing hydrostatic pressure against which the bubble is formed a t the capillary opening varies with this cycle. Bubbles from different capillaries in the porous disk will vary in size, since the capillaries are not uniform and will operate a t different air pressures. We performed a series of experiments with different salt solutions, maintaining the air pressure as constant as feasible. The solutions were arranged in order of foam-zone heights. When the
OF PERCENTAGE FILMFORMATION BY TABLE I. MEASUREMEKT THE
Salt SasSOA
TWO-BUBBLE EXPERIMENT Bell Sepn., yo Film Formation
Molar Concn. 0.035
Mm. D
11
Distd. H10
H. & N. 0.0
6
100
0.01 0.1
5 6
1.01
11
1084
J
1
F. & M. ( 9 ) 1.8
0.2
NaSSOd
1.0
1%
NaCiViS
1.0
1%
1
28
KCSS
1.0
NaCl
1.0
1%
}
44
KCI
1.0
4 6
82
August, 1941
INDUSTRIAL AND ENGINEERING CHEMISTRY
series was repeated a t several other controlled air pressures, the solutions fell in different orders on the basis of this measurement. I n general, then, it appears that differences in foam-zone heights in this experiment are attributable less to differences in concentration or composition of the inorganic salt solute than t o other variables and to inability to obtain precise values. TWO-BUBBLE EXPERIMENT.The results obtained from this experiment do not seem reliable. The technique is intriguing in simplicity of operation and apparent directness, but there are too many factors to be controlled to make it even semiquantitative. Using this technique in an effort to reproduce some of the reported results, it was found that the distance between the ends of the small “bells” by means of which the bubbles are formed and manipulated is an all-important factor. All glassware used was first carefully cleaned with chromic acid and thoroughly steamed. Solutions were made from c . P. reagents] carefully filtered and preserved from contamination. Table I shows a comparison of data with those of Foulk and Miller, who did not state the bell-separation distance. The importance of bell separation is clear. The effect probably is connected with the fact that rate of bubble formation has a strong influence on results. I t is likely that these factors which affect the time of contact of the bubbles should both be controlled to make the times of contact comparable. It is not possible to do this simply by using a fixed bell separation since the size of the bubble is determined in part by surface tension which varies with the solution. Hence the results are not sufficiently reproducible to justify use in a critical experiment. Applying this technique to the nonfoaming (and therefore nonfilm-forming) mixtures of Foulk and Miller, results shown in Table I1 were obtained.
TABLE 11. FILMFORMATION IN NONFOAMINQ MIXTURES Mixture, Co. 100 NaCNS 71 NaCl 100 NaCNS 3 2 . 5 NarSOi 100 KCNS 7 2 . 5 KC1
+ ++
Bell Sepn., AMm. 5 5
5
% Film Formation H. & N. F. & M. (3) 20 0.8 0.6 58 0.3
11
With each of these mixtures a small increase in the bell separation resulted in 100 per cent film formation. Very little film formation is obtained with distilled water. This fact was used by Foulk and Miller (S) to substantiate the statement that pure liquids do not form films. Of all pure liquids, water is the one least able to form stable films (4). It is the only pure liquid reported by Foulk and Miller. The two-bubble technique was accordingly tried with other pure liquids, with results listed in Table 111. If the two-bubble experiment is reliable, pure liquids do form films.
TABLD 111. FILM FORMATION IN PURE LIQUIDS Liquid Ethyl aloohol Benzene Carbon tetrachloride Xylene n-Butyl aloohol Diethylcarbinol
Bell Sepn., Mm.
% Film Formation
3 5
0 100 0 0 0 0
8
4
4 4
Bell Sepn., Mm. 5
% Film Formation 100
7
100 100
6 6 6
100 100
5
100
It was possible to increase the percentage film formation greatly by clarifying the solutions used. This was done by producing gelatinous precipitates in them and allowing these to settle. For example, addition of colloidal iron oxide to a solution of sodium chloride, followed by settling of the coagulated colloid, made it possible not only to increase the measured percentage film formation a t a given bell separation, but also to make films practically 100 per cent stable a t almost any bell separation.
1085
Consequently, even if the rather dubious results of two-bubble experiments are admitted, evidence shows nothing in favor of the balanced-layer theory of film formation since absence of fine suspended matter has a far greater effect on film stability than concentration of the solution, and pure liquids do form films. Therefore it is concluded that the experimental justifications of Foulk and Miller (3) for the balanced-layer theory of film formation are invalid, and the theory must be judged by its inherent plausibility. Balanced-Layer Theory of Film Formation I n Foulk’s presentation of the balanced-layer theory of film formation (g, S), the following facts are t o be reconciled with the theory: “(1) Pure liquids do not foam, (2) some impurity must be present in a liquid to make it foam, (3) it does not matter whether the added substance is positively or negatively adsorbed, (4)certain substances have the power of destroying the foams produced by positively or negatively adsorbed substances.” Foulk interprets foam formation in terms of film formation, with the obvious corollary that.liquids which do not foam will not form films. In particular, pure liquids will not form films. In essence the theory holds that a film is formed by the close approach of two already present surfaces. If there is no opposing force, the surfaces approach each other so closely as to reach the vanishing point. However, if there is an opposing force (presumably one which increases in intensity as the film thins), then it is possible for equilibrium to be reached and a film to remain in existence. The opposing force in question is osmotic according to Foulk (following the earlier suggestion of Edser) and arises from the difference in concentration between the surface layer and the bulk liquid. Pushing the two surfaces together tends to mix the bulk and surface layers and so cause an equalization of concentrationhence to do work “against osmotic forces”. Foulk and Miller advance two corollaries (3): (A) “In the case of solutions containing both positively and negatively adsorbed substances, there should be certain mixtures which do not foam because positive and negative adsorption cancel each other and thus produce equality in concentration between surface layer (B) In the case of those solutions which contain and mass. only one solute, but which either do not foam or foam at certain concentrations only, it will be found that the condition of nonfoaming is also a condition of equality of concentration between surface and mass of the solution.” In addition ( 2 ) , ‘
