I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
570 TABLE Iv.
COMPARISON O F PUBLISHED VISCOSITIE5 WITH DAT.4 FROM FIGURE 2 \. iscosity, Centipoise3 _ _ _ _ _
Temp.. O c . 30 50 60
Author Rawitsch (61
NO.
30 40
50
Bauer end Markley ( 1 )
98.9 98.9
. ,
98.9
44.6 30.6 22.1
From Fig. 2 40.0 21 .o 16.0 12.4 9.5 7.7 45.0 31.5 22.4
9.4 8.6 8,3
9.3 8.5 8.3
lished 42.0 21 .6 16.0 11.6 9.4 7.5
115 115 115 115 115 215 105 105 105
90 80 90
Boekenoogen ( 2 )
Pub-
Iodine
6.8 45.6 56.3
_
AND DENSITIES OF UNDEODORIZED AND TABLE V. VISCOSITIES
DEODORIZED HYDROGENATED COTTONSEED OILS FROM DIFFERENT
RATCHES
Janiple S o . Batch No. Deodoiized Iodine h o e
6 B
7 B
8 c 1
9 c 1
10
I1
c 2
No
No
Yes
65
Yes 65
No
c 2 Yes 61
1O:Ol
1O:OS 7.07 3.48 2.51 1.87
20.25 9.72 6.77 3.36 2.38 1.82
20.97 10.02 7.00 3.48 2.49 1.87
!.
No 66
66
66
61
Vol. 3Q No. 6
generality stated by Strevens ( 7 ) and confirmed 'by Kaufmanii and Funke (6)that the viscosity at a given temperature decreases with increasing iodine number. _ .The ~ effect of thermal change such as decomposition or polj rnerization of the oils upon viscosity values should be small brcause of the short time the oils were kept hot for the high teirrperature measurements. Bauer and Markley ( 1 ) showed thsr heating a cottonseed oil of iodine value 50.2" to 230" C. for 2 hours increased the viscosity by about 3%. Fatty acid contenth of the oils removed from the viscometers after the complete ten1 perature range had been covered were found to have increased in some cases as shown in Table I. Some unavoidable inaccuracies due to thermal change are therefore to be expected, ailti the viscosities reported in Tables I11 and V may he in error bi about 1yo,especially at the higher temperatures. Table V compares the viscosities and densities of samples 3 t I J I I, inclusive. These data indicate, in general, that cottonseed oil5 from different batches but with similar iodine numbers have essentially the same viscosities and densities. A refining treatment, such as deodorization in which little change in composition probably occurs, does not affert these properties to a n y appreciable extent I
59,s 88.5 l06.6 152.3 179.9 210.0
20.95 10.01 5.78 3.47 2.49 1.85
7.01 3.49 2.51 1.88
59 . 5 88 . 5 106 . 6 152 . 3 179 . 9 210 .o
0,8848 0,8664 0.8546 0.8248 0.8055 0.7853
. I . ;
0.8663 0.8546 0.8253 0.8071 0 7883 ~
___Density-----. , . . 0.8822
0.8664 0.8546 0.8258 0.8077 0.7882
0.8661 0.8545 0.8244 0.8069 0,7886
22.78 10.48 7,l5 3.56 2.54 1.88
22.72 10.18 6.91 3.47 2.49 1.85
---0.8834 0.8663 0,8546 0,8250 0.8071 0.7879
0.8832 0.8838 0.8649 0.8654 0.8531 0.8536 0.8237 0.8241 0.8054 0.8057 0.7858 n.7861
GCKNOWLEUGMEh'l
The authors art: indebted to S. To Bauer of this laborarory for supplying hydrogenated cottonseed oil samples; to \I Gtansbury and D. @. Heinzelman for determining the iodine numbers, free fatty acid contents, and refractive indices; and to J. J. Ganucheau of the Southern Cotton Oil Company for sirpplying many of the oil samples used. LITERATURE
by interpolation and extrapolation from Figure 2. The kinematic viscosities of Bauer and Markley were converted to absolute viscosities by multiplying by densities read from Figure I , The agreement shown in Table IT' seems to show that the data here reported may be used as an indication of the viscosity of any refined or hydrogenated cottonseed oil of which the iodine nurnber is kn0.IT-n. Furthermore, the curve%in Figure 2 t'ollo\~ the
Craxton, F. c.,I X D . E N G . CHEM., ANAL. E D . , 14, 593-5(1948) Hershberg, E. B., Ibid., 8, 312 (1936). Xrtufmann, H. P., and Funke, S., Fette u. SeiferL, 45, 255 (1W-4) Rawitsch, G . B., Kolloid-Z., 76, 341-5 (1936). Strevens, J. L., J. Soc. Chem. I d . ,33, 109-11 (1914). Zeitfuohs, E. H.,Natl. Petroleum S e w s , 31, 2621 (1939); I'roc. A m . Petrolewn Inst.. 111, 20, 104 (1939)
INHIBITION of F
ING In Solvents >)
Containing SYDNEY ROSS
T
AND
.J.
crrm
Bauer, S.T.,and MarkIey, K. S., Oil & S o a p , 20, 1 (1943). Boekenoogen, H. A., Chern. Weekblad, 34, 759 (1937).
))
Foamers
w. MCBAIN, Stanford [Jriitersity, Calif.
HE foaming of liquids is a, frequent, c a u w of trouble in industrial and laboratory processes. Methods of preventing or reducing the amount of foam may be divided into t\vo groups--mechanical and chemical. The former employ pulsating strea,ms of gas above the liquid, perforated spiral canals, wntrifuges, continuous pumping of liquid from bottom to top of rontainer, change of pressure, heating elements, ultraviolet r a p , x-rays, canal rays, supersonic waves, rotating fans or disks or adjustable gratings above the liquid surface, sharp rorners in the design of the apparatus, etc., with varying degrees of success. h o n g chemical methods the addition of small quantitic+ of caprylic alcohol, amyl alcohol, octyl alcohol, linseed oil, castor oil, rapeseed oil, trimethylcyclohexanol, phenyl ether, isonmyl'isovalerate, milk, etc., has been recommended for various
aqueous ioammg syatemb. .1r11ong the intlu~trial procesws where i t has been found necessary to use foam preventive measures are purification of beet juice&, manufarture of glue, sepnration of cream, production of steam in boilers, preparation of paper and coated papers, heat dehydratioii of c r u d e oils and tars, purification of sewage, and thr boiling. v : + ( ~ i i i r n i wapnratinn, distillation. or filtering of many solutiorr~. The present paper reports a h e r r ~ 5of experinwnts to determine the effect of incorporating certim agents in different ~ ~ 1 1 defined systems capable of forming foam. The object of t h e investigation 1s to uncover some operative factors in the inhibition of foaming by means of antifoaming additives. ,1 knowledge of those factors would k~ of out importance
INDUSTRIAL AND ENGINEERING CHEMISTRY
June, 1944
Pure liquids do not foam. Foaming is due to added foaming agent. Practically every industry has a twofold problem of suppressing undesired foams and of stabilizing foam where it is wanted. Such antifoams as ethyl ether and amyl alcohol operate only when directly applied to foam and,are usually inactive when dissolved in the liquid. All of the additives here studied were first dissolved or dispersed in the liquids, and some were highly effective as antifoamers. A number of additives have been tested as chemical defoaming agents for five systems, aqueous and nonaqueous, containing known added foamers. Systems which have a chemical similarity in either solute or solvent, when compared with respect to the antifoaming effect of the same additives, show a corresponding effect in three out of every four cases;
571
when systems have a more complete chemical resemblance, the correspondence of the effects is even greater. Complete correspondence is shown, for example, between Aerosol OT in triethanolamine and Aerosol OT in diethylene glycol. The most complete defoamers are usually but not always insoluble. I n some cases excess of an "insoluble" antifoam nullifies its effect; hence in these cases there is an optimum amount where solubility is slightly exceeded. Results of bulk foaming tests are shdwn to bear a large degree of correspondence with the action of an additive on a single film. The spreading of additives on single films is discussed from the point of view of surface tensions. More than one mechanism must be postulated to account fully for all types of foam inhibition by chemicals.
FOAM FORMING SYSTEMS
frequently observed but could be traced to a difference in solubility or viscosity. Differences between these nonaqueous foaming systems and aqueous systems are both physical and chemical. The chief physical differences are viscosity and rate of viscosity change with temperature. Everything else being equal, a decrease in viscosity is accompanied by a decrease in foam stability, ab shown by Bartsch ( 1 ) . The parallel variation of foam stability and viscosity with temperature is probably due to their causative phenomenon-the reduction of intermolecular forces which results in a simultaneous loosening of the cybotactic structure both in the bulk liquid and a t the liquid-air interface (3, 4, 7). While a decrease in viscosity leads t o a less pronounced foam stability by facilitating drainage, it is nevertheless conducive to greater ease of foam formation upon shaking or beating; the latter effect is responsible for troublesome foaming in industrial processes a t higher temperatures. Ease of foam formation is inversely proportional t o viscosity; if the viscosity is constant, i t may depend on some factor such as the degree of Carbonation (6) and bears no direct relation to foam stability. Solubilities of organic compounds are in general greatly enhanced in nonaqueous solvents. Foaming ability is, however, very different-for example, the failure of saponin or Aerosol 1B t o form foam in the absence of water. TABLE1. FOAM BEHAVIOR OF SOLUTES IN TRIETHANOLAMINE EFFECT O F CHEMICAL ADDITIVES ON WELL-DEFINED AND DIETHYLENE GLYCOL
The liquids used in the present investigation are not specially purified. They are, nevertheless, incapable of forming bubbles of stability greater than about half a second upon exposure to the atmosphere, until a known foaming agent is added. The nonaqueous solvents are triethanolamine and diethylene glycol (Eastman Kodak Company), chosen because of their low vapor pressure. Various solutes were then tried in those two liquids with the primary purpose of determining if a foaming system would be obtained. The results offer marked contrast with the effects of the same solutes in water. Table I gives data on the foaminess and solubility of the solutes in the two nonaqueous solvents. The capacity for forming a foam was tested by shaking the solution by hand; preliminary tests were conducted a t about 70" and a t 100" C. The higher temperatures were first chosen because a t room temperature the solutions were generally too viscous to form a foam by shaking. Room temperature tests were made by bubbling air through the solution and dispersing the gas with a sintered glass disk. The solutions exhibited similar behavior a t room temperature, 70" and 100" 0. A difference i i i the amount of foam formed a t different temperatures was
.FOAMlNG SYSTEMS
Solute
Solubilityu
Aerosol O T Triethanolamine oleate Aluminum oleate Soybean lecithin Sodium stearate Benzene chloroform Diethylene glycol Benzene Cerotip acid Saponin Cholesterol
+
Commentsa
TRIETHANOLAMINE S Foam S Foam SH Foam, H or C S Foam SH Foam 8 Clear soln.: foam S No foam I No foam: chromatic emulsion SH No foam, H or C SH No foam, H or C SH DIETHYLENB GLYCOL
Aerosol O T Benzene Aluminum oleate Sodium palmitate Sodium stearate Calcium palmitate Aerosol lB, Butyric acid Oleic acid Saponin Cholesterol
-
S S
I
SH
SH
dH S S S
I
bf water
S
SH
S = soluble. I insoluble. H = hot. C = cold; foam = solution markedly foamy:, slight foam = o h y a few bubbles capable of foamin Turbid solution foams t o 148' C.; clear solution above 148% foams elightly (148' C . is melting point of cholesterol).
A selection of five foaming systems, three nonaqueoub and two aqueous, and a selection of twenty-two chemical agents was made. The foaming systems were the following: 7.74% Aerosol OT in triethanolamine, 5.55% Aerosol OT in diethylene glycol, 34.0% triethanolamine oleate in triethanolamine, 1.55% Nacconol N R in water, and 0.8570 Aerosol OT in water. The chemical agents are listed in Table 11. The ability t o form foam was first tested a t 100' C. by a shaking method, similar t o that employed by Wilson and Ries (IO). Thereupon the additive was incorporated to the extent of 1% of the volume of the original solution. This is a relatively high concentration for an antifoam. The results in every case (columns 1, Table 11) could be classified in three easily distinguished groups: complete inability to form foams, marked E; marked loss of ability to form foam, &/I; and no observable effect on the ability to form foam, N. Columns 2, Table 11, record the solubility of the additive. The effect of the additive when added to a single film of the liquid system was tested by a method developed by J. V. Robinson in this laboratory. A film of the foaming liquid (columns 3, Table 11) is formed on a loop of platinum wire and another
572
INDUSTRIAL AND ENGINEERING CHEMISTRY
discrepancies in tho reverse direction, 6 are in aqueous solution, 2 in triethanolamine solution. Since there are 44 aqueous mixt . w s and 66 nonaqueous, this means that the reverse discrepancies occur 4.5 times as often in aqueous as in nonaqueous media. This is to be expected, since any discrepancies in this direction would surely happen only if there was already strong surfaw adsorption of the inhibitor so that adding more produced little or no effect. Another interesting regularity is that in four of the six aqueous reverse discrepanries, the inhibitor is a phosphate. Of all the additives tried, in the concentrations reported, only four are effective foam inhibitors for all five foaming systems. The remainder are variously effective, on from zero to four out of the five systems. The action of organic phosphates appews to be specially pronounced and more widely applicable than the other additives. I n some cases the effect of varying the concentration was observed. Although not a n invariable rule, in general an optimum concentration seems to exist for defoaming action. This optimum amount sometimes lies definitely above the solubility of the agent. When this concentration or amount is exceeded, the foaming ability of the mixture frequently reasserts itself. This is true, for example, in the cases of additives 8 and 9, and 12 and 18,when added to the nonaqueous solutions of Aerosol OT. The arbitrary choice of a concentration of 1% in Table I1 does not always permit the true character of a defoaming agent to be properly represented. I n higher concentrations some of the agents reverse the effect that they exercise a t a concelitration of 1%. Such examples are noted in Table 11. The marked effect caused by the presence of even minute amounts of water has been noted in some of our studies of the foaming of nonaqueous systems. I n the present investigation it was found that, after defoaming action had been achieved, addition of small amounts of water to the nonaqueous system8 did not generally affect the result. This was true even in cases where the solubility was greatly altered by the presence of water. When the best defoamers were used, the solution could be diluted with water to three or four times its original volumes before foaminess began to reappear.
platinum wire, previously dipped in the additive under test, is gently touched to the thick part of the liquid film. Immediate breaking of the film by the additive is marked E, a delayed rupture but still definitely due to the influence of the additive is marked M, and no effect on the stability of the film is marked N. DISCUSSION OF RESULTS
The behavior of different foaming systems with respect to additives can be compared according to the following plan: Aerosol OT in triethanolamine and in diethylene glycol compared with Aerosol OT in water; Aerosol OT in water compared with Nacconol NR in water; Aerosol O T in triethanolamine compared with triethanolamine oleate in triethanolamine. Comparisons are made of the first column in each system, between inhibiting effect (E or M) and no effect (N) on the foaming capacity of the system. Thus E and M are counted as an agreement but M and N count as disagreement. On this basis, examination of Table I1 reveals that more similarities or regularities are to be found than differences. The different additives exhibit completely parallel behavior (0% deviation) in the nonaqueous solutions of Aerosol OT. The deviation of behavior between the nonaqueous and the aqueous solution of Aerosol OT is 8 in 22, or 36y0; that between the two different solutes in triethanolamine is 7 in 22, or 32%. The behavior of additives on foaming systems thus bears a certain relation, in contradiction t o our preliminary impression, based on less widely distributed data, that the defoaming action of a n additive is highly specific. Systems which have a chemical similarity in either the solute or the solvent, when compared with respect to the antifoaming effect of the same additives, show a correspondence in effect in three of every four cases. When systems have a more complete chemical resemblance, then the correspondence of the effects is even greater (complete correspondence is sho-ivn, for example, between Aerosol OT in triethanolamine and Aerosol OT in diethylene glycol). The following further regularities have been pointed out (8): I n "pop" tests, inhibitor was necessarily placed directly at the liquid-air interface. An inhibitor could hardly have any effect unless it did reach that region, but it may not do so when present in the body of the solution. Of the 110 entries in Table 11, 43 are discreuancies between columns 1 and 3. Of these 43, no less than 35 (Slyo)indicate inhibitor more effective in film than in foam, which is to be expected on the above grounds. Of the 8
THEORY OF FOAM 1NHIBlTION
CORRELATION WITH SOLUBII.I~N OF ADDITIVE. Sonic? wvntei~ have stated that a dcbfoarning additive should be, or m i s t be,
O F ADDITIVES ON FOAM INHIBITION TABLE 11. EFFECT
1. 2. 3. 4. 6. 6.
7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22.
Aerosol OT in Triethanolamineh 1 2 3 Additive E I E Ethyloleylglycol o-phosphate S E Trioctyl tripolyglycol tetrapolyphosphate E E S G1 ceryl monoricinoleate S M 2 - 1 mino-2-methyl-1-propanol ' E I E Tetraoctyl pyrophosphate E E I Diethanol aminoethyl phosphatidic aci d N S N Carbitol malqate E I E Mono-oleyl dipolyglycol o-phosphate E I E Diethylene glycol mono-oleate E I E Diglycol dinaphthanate N S M 2-Amino-2-ethyl-l ,a-propanedi 01 E I E Sapamine MS M I E B. 0. 1 M I E B.0 . 2 M I E Penatrol 60 N S N Diethylene glycol M I E Ethyl phosphate E E S Tween 80 M S E n-Nongl alcohol N S N Hydroquinone N S N Phenol E I E n-Butyl phthalate
Triethanolamine Oleate in Triethanolamineb 1 2 3 E 8 M E E S
N
%e
E
N
N N
E N
N E
E N
N
M N M N
N
M
S
S s 1
S
S S
6 S
S I 1 S
S
S S S S
S S
N M E E h.1 N N E M N
E E
E N
E
N E N N E
OF
FOAMING SYGTEhfS'
Naccqnal NR in Water at 24' 1 2 3 I hI
cbl _
Ec
e:
sc
E
Sd
kfN
I
N
I I
N
I I I
M
N M &f
N
N
N N N N i?I N E N
S
S
N
N N N N
M N M
I I I I
M N
S
N N
S S
I S
S I
M
E
N M M
Aerosol OT in Diethylene Glycol b 1 2 3 E I E &l I E E S E S E I E E I E N I M E" I E E" I E E I E N 6 El I M I E hl I E M I E N S N M S E E I E M S E N S N N N S E I E
2
E = excellent inhibiton; M = moderate inhibition; N = no effect; S = soluble; I = insoluble; H hot: C b 1 = test for foaming by mere shaking; 2 = solubility of antifoaming additive; 3 = "pop" or film breaking test. r E in higher concentrations than the 1% here referred to. d E in greater amounts when the solubility is exceeded, although 1 % was soluble and was without effect (N. 6 N in greater am,ounts, even although 1 % was incompletely soluble and was a &sa E defoamer. I N in concentrations above and below 3%, the only example of any but 1 % cited (compare footnotes d and e ) . a
Vol. 36, No. 6
E"
-
oold.
Aer0:ol 01' in Waterb 1 2 3 E I M 8 N N M I M N B M
M
I
N
8
S I . hl
M
M N M M
9. I
P
S
H
I
1
hi M
S
M N
S B
$ N N
8 S I
M
s
x
M
N N
M
A4
M
I M
N M M M
M M M M N N hl
c
INDUSTRIAL A N D ENGINEERING CHEMISTRY
lune, 1944
insoluble in the foaming system to which it is added. of the present study follow: Effect of Additive
No. Ineol. Additives
E
23
M N Total
Results
No. Sol. Additivas 6 16 34
18 13 .-
-
54
56
This table shows that the most pronounced foam inhibitors are preponderantly insoluble, those that exhibit a partial defoaming action may or may not be insoluble, and a majority (600/,)of the soluble additives are without effect. However, an additive need not be insoluble to be an effective foaming inhibitor or an effective foam breaker. CORRELATION WITH FILMBREAKINGTEST. Some of the mechanisms of defoaming action are best observed when a drop of the additive is placed directly on a film of the foaming system. I n some cases the film snaps or pops immediately (E), in other cases the reaction is more delayed (M), or possibly does not occur a t all (N). This test can therefore be used to make a classification of the addikives into three groups. Correlation of this classification with the foaming tests i6 bulk follows: Effect of Additive in Bulk Test
E
M N
No. Additives in Film Breaking Test
E
M
N
26 18 2
3 12 15
0 4 30
it is evident that immediate rupture of the film on addition of the additive generally denotes an excellent defoaming agent; when no effect on the f3m is obtained, the additive, although probably without effect, may still exercise a partial defoaming action but never complete defoaming action, in any case yet met. The degree of correlation at the extremes (E to E and N to N) is the most pronounced. In those cases where the additive does not cause immediate rupture, if it is soluble i t diffuses into the body of the film without visual effect on the surface; if insoluble it is left as either a solid or as a liquid lens on the surface. When rupture of the film is due to the additive but is so delayed that the surface can be (a) watched, three different mechanisms can be observed: When a drop of additive is added to the center of the film, the solution slowly draws back from that spot, becoming thinner in the center and thicker at the edges, until the central portion becomes so thin that it snaps. ( b ) On a thick liquid film the additive is without effect, merely sitting as a separate drop on top of the liquid without spreading. On a thin film it spreads out rapidly and breaks the film. (c) In aqueous systems in particular, where evaporation plays a part in maintaining a thin film in the presence of a thick film (9),the spreading on the thin film was not visible but the film ruptured immediately when its thinnest part came in contact with the additive. Those differences in visible effect may be due, respectively, to slow spreading on a thick film, slow spreading on a thin film, and rapid spreading on a thin film. There are two hypotheses to account for the instability of a liquid film when an additive is spread on the surface. Hardy (3)regards differences in concentration at the two liquid-air interfaces as exercising an unsymmetrical restraint on the liquid between them and causing an ultimate collapse: “When the interfaces are not similar, it is like an unannealed plate.” Neville and Haelehurst (6, 9) state that a film can persist only as long as evaporation (and its resulting temperature gradient) provides a flow of liquid into the thin portion to counteract drainage, an effect opposite to those discussed by Gibbs as examples of the tendency of the interior of the film to flow out and downward. An additive, by lowering surface tension, might decrease the flow of liquid into the thin film and cause a rupture a t that point. A delayed film-rupture test, as described in mechanism b above, is shown by n-butyl phthalate added to Nacconol NR solution and
573
by Carbitol maleate added to the solution of triethanolamine oleate in triethanolamine. Bulk surface tensions and interfacial tensions were measured on those two systems in the du Notry tensiometer: Liquid a n-Butyl hthalate b.* NaooonnPNR a. Carbitol maleate b. Triethanolamine nlnatn triethmolamine
__
Temp.,
c.
Tension, Dynea/Cm. Surface Interfacial
27
1. d
-. 27
25
in
25
36.6
a
The spreading coefficient for a liquid, a, to spread upon another liquid, b, is given by Harkins ( 4 ): 8 = Yb
-
(Yo
+
yab)
From this equation it is apparent that under normal circumstances Nacconol NR dissolved in water (1.55%) will spread on n-butyl phthalate, and triethanolamine oleate dissolved in triethanolamine will spread on Carbitol maleate. Observations of those two effects, using powdered talc on the surface of the lower liquid, c o h e d the prediction. It hm already been stated that the lower liquid, in each case, does not spread on a thick film or on bulk volume of the other liquid, and this too is in conformity with the spreading coefficient. However, since liquid a does in each case spread on a thin film of liquid b, the value of yb for a thin fib appears to be greater than Y b for a thick film or bulk volume. From the equation, Yb for a thin film of Nacconol NR solution must be greater than 37.1 dynes, ,and yb for a thin film of triethanolamine oleate solution in triethanolamine must be greater than 42.1 dynes. I n both cases this represents a n apparent increase in surface energy of the thin film by more than 6-7 ergs per ema. The two liquids in each pair have nearly the same surface tension and a low interfacial tension. The increase in surface tension of the thin film is just sufficient to overcome the (small) difference between the two surface tensions added to the low interfacial tension. It is therefore to be expected that the phenomenon of delayed rupture as described for the two cases would only be apparent in those examples where the two liquids meet the specified conditions. The results of the present MECHANISMS OF FOAM INHIBITION. investigation indicate that the chemical inhibition of foaming may be accomplished by more than a single mechanism. In some cases, additives capable of breaking single films are still not effective in inhibiting foaming of a liquid when present in bulk, and there are also different ways in which an additive is capable of destroying a film, I n more complex cases it is to be expected that several factors are simultaneously operative. No single theory of mechanism can be applied to all possible examples. ACKNOWLEDGMENT
The information contained in this paper was obtained in large part in connection with an investigation sponsored and financed by the National Advisory Committee for Aeronautics. LlTEFtATURE ClTED
(1) Bartsch, O.,Kolloidchem. Beihefte, 20,1-49 (1924). (2) Hardy, W.B.,J . Chem. SOC.,127,1207-27 (1925). (3) Hardy, W.B.,J . Qen. Phvsiol., 8, 641 (1927). (4) Harkins, W. D., “Theory and Application of Colloidal Behavior”, Vol. l, pp. 142-211,New York, McGraw-Hill Book Co.,1924. ( 6 ) Hazlehurst, T. H., and Neville, H. A., J . Phy8. C h m . , 41, 1206-14 (1937); 44,692-600 (1940). ( 6 ) Helm, E., and Richardt, O., J. Zmt. Brewing, 42, 191 (1936). (7) McBain, J. W.,and Davies, G . P., J . Am. Chem. SOC.,49,2230 (1927). (8) Neville, H. A.,private communication. (9) Neville, H. A., and Hazlehurst, T. H., J . Phus. C h m . , 41,545-61 (1937). (10) Wilson, R. E., and Ries, E. D., Colldd Symp08iurn MMacrgtrph, 1, 145-73 (1923).
