INDUSTRIAL AND ENGINEERING CHEMISTRY
834
NOMENCLATURE
Vol. 47, No. 4
(6) Markstein, G. H., Fourth Symposium on Combustion, pp. 44-
D = average base diameter of flame front distortions H = thickness of flame zone h = measured amplitude of flame front distortions L = length of flame front 1 = scale of turbulence as defined in ( 3 ) 11 = Lagrangian scale of turbulence k Eulerian scale of turbulence 1’ = relative scale of turbulence r =distance from axis SO= normal burning velocity St = turbulent burning velocity = average turbulent burning velocity t = time U = average approach stream velocity u‘ = intensity of approach stream turbulence 6 = mean angle between flame front and approach velocity
-
(7)
59, Williams & Wilkins, Baltimore, 1953. Markstein, G. H., “Selected Combustion Problems,” Combustion Colloquium, Cambridge, England, AGARD, NATO, pp. 263-5, Butterworths Scientific Publications, London, 1954.
(8) Scurlock, A. C., and Grover, J. H., Fourth Symposium on Combustion, pp. 645-58, Williams & Wilkins, Baltimore, 1953.
Scurlock, A. C., and Grover, J. H., “Selected Combustion Problems, “Combustion Colloquium, Cambridge, England, AGARD, NATO, pp. 215-47, Butterworths Scientific Publications, London, 1954. (10) Simmons, L. F. G., and Salter, C., Proc. Roy. Soc. (London), A145, 212 (1934). (11) von Karman, T., Proc. Fifth Internatl. Congress Appl. Mechanics, p. 347, Cambridge, Mass., 1938. (12) Williams, D. T., and Bollinger, L. M., Third Symposium on Combustion, pp. 176-85, Williams & Wilkins, Baltimore, (9)
1949.
LITERATURE CITED
(1) Dryden, H. L., Quart. A p p l . Math., 1 , 7-42 (1943). Harris, M. E., Grumer, J., von Elbe, G., and Lewis, B., Third
(2)
Symposium on Combustion, pp. 80-9, Williams & Wilkins, Baltimore, 1949. (3) Hottel, H. C., Williams, G. C., and Levine, R. S., Fourth Symwosium on Combustion, _ww. _ 636-44, Williams & Wilkins, Baltimore, 1953. (4) Karlovitz, B., “Selected Combustion Problems,” Combustion Colloquium, Cambridge, England, AGARD, NATO, pp. 248-62, Butterworths Scientific Publications, London, 1954. ( 5 ) Karlovitz, B., Denniston, D. W,, Jr., and Wells, F. E., J. Chem. Phys., 19, 541-7 (1951).
Wohl, K., and Schilly, R., Project Squid Semiannual Progress ReDort. D. 69 (ADril 1. 1953). (14) Wohf, K.; *and Shore, L., Project Squid Semiannual Progress Report, p. 162 (Oct. 1, 1952). (15) Wohl, K., Shore, L., von Rosenberg, H., and Weil, C. W., Fourth Symwosium on Combustion. -ww. - 620-35, Williams & Wilkins, Baltimore, 1953. (13)
RECEIVED for review March 5, 1954. ACCEPTED November 24, 1954. Research conducted under auspikes of Project Squid, jointly sponsored by the Office of Naval Research, Department of the Navy. Office of Scientific Research, Department of the Air Force, and Office of Ordnance Research, Department of the Army, under Contract No. N6-ori-105.
Effects of Shear on Lithium Greases
and Their Soap Phase THEODORE A. RENSHAW Naval Air Experimental Station, Naval Air Material Center, Philadelphia 12, Pa.
ELECTRON MICROSCOPE STUDIES
...show
some unusual results contrary to accepted ideas of grease structure and rheology
...emphasize
the attractive forces between adjacent fibers a s a primary factor responsible for special properties of greases
A
CONSIDERABLE number of studies of greases have already been made with the electron microscope and much information has been revealed concerning the varied physical forms that the soap component may take depending on different compositions and different treatments. T h e basic objective of such studies is always to achieve understanding and, hence, control of all the variables that contribute to the properties of the bulk product. The soap phase, being responsible for the special nature of greases, is studied to determine what factors influence its capacity for gelation. One study (8)describes the mechanical breakdown of greases in terms of the disintegration of the soap fibers constituting the thickening agent.
Additional information on sheared greases is reported here, and some of the generally accepted concepts concerning the rheology and structure of greases are re-evaluated. Further study of the attractive forces acting between soap particles may clarify grease phenomena. EXPERIMENTAL
The materials studied in this investigation were all greases currently on the military qualified products list of Military Specification MIL-G-3278, Grease; Aircraft and Instruments (For Low and High Temperatures). Attention was concentrated on this class because of the widespread interest in its properties a t temperature extremes. It was also considered a favorable group to study because of t h e uniformity of the constituent materials. All greases thus far submitted were composed of a lithium base soap and an oil which is essentially a synthetic dibasic ester oil. The fact that small amounts of petroleum oils and other diverse additives were sometimes present did not appreciably affect the similarity found among the greases. A means for inducing extremely high rates of shear was required to investigate the effect of maximum fiber degradation on grease properties. Such a n apparatus was developed by modifying a high temperature grease performance unit of the type described in Method 33.1 of Federal Specification VV-L-791e. This modification comprises a means of forcing the test grease through the rotating ball bearing a t a controlled rate. T h e grease enters the rear of the bearing housing through copper tubing inserted in a hole opposite the ball cage and, after passing through the balls, is exuded onto the front surface of the housing.
INDUSTRIAL AND ENGINEERING CHEMISTRY
April 1955
Figure 1.
835
Electron micrographs (14,000 X ) of fibers of AML-8 grease showing effects of different shearing devices A . As received, unworhed B . After 100,000 double strokes of N a v y Worker C.
After high speed ball-bearing test
Grease flow was developed by screw feeding a plug into a cylinder loaded with sample. The bearing, which rotating a t 10,000 r.p.m., was flushed with the test grease a t a flow rate of 0.38 cc. per minute, whereas in the actual test the rate was lowered to 0.038 cc. per minute, which was sufficient to effect a uniform degradation of the fibers. A 20-cc. sample took about 9 hours to run through. The second method of shearing greases employed the Navy Grease Worker in the test outlined in Method 313.1 of Federal Specification VV-L-79le. The equipment is the same as the ASTM Worker (D-217-52T) except t h a t 1/16-inch holes are employed in the churn. The test requires that 100,000 double strokes of the perforated plate be made through the sample. The technique used for preparing the specimens for the electron microscope is relatively simple: Dilute dispersions of the grease in a light naphtha or petroleum ether were dropped on the Formvar-covered specimen screens and allowed to dry. These screens were then washed in additional solvent to remove any remaining oil. Finally, the specimens were shadow cast with chromium to enhance the otherwise poor contrast between the soap particles and the Formvar substrate film as well as to enhance the configurations and surfaces of the various particles. RESULTS
The first material evaluated in this series of tests was the AML-5 grease which contained fibers characteristic of the majority of the approved greases. The results of its performance
Table I.
Grease AML-5 2 1 . 6 % Lithium stearate
AML-:I I S . % Lithium hydroxy stearate AML-8 Lithium stearate
25%
Working Test Results
Condition As received After 100,000 strokes of Navy Worker After highspeed ball bearing As rereived After 100,000 strokes of Navy Worker After hiah meed ball bearing As received After 100,000 strokes of Navy Worker After hjgh speed ball bearing
I
Average Fibers Length, Width, Micro- Length/ X X 10-8 pene- Width om. om. tration Ratio 13 5.5 75 24/1 9
10.2
210
9/1
2.8 2.9 6 5 No appreciable change from grease as received
101 84
lO/l 12/1
143
12/1
80
77
5/1 13/1
2.9 19
5.8 15
20
20
185
10/1
7
10
73
7/1
on the high speed ball bearing test, recorded with other working tests in Table I, were quite unexpected by comparison with the test run on the Navy Grease Worker. It was anticipated that the higher shear conditions in the ball bearing would result in the exuded sample having a softer consistency-Le., a higher penetration-but this was not the case. Figure 1 shows electron micrographs of the soap fibers of this grease: ( A ) when received; ( B )after the Navy Grease Worker test; and (C) after the high speed ball bearing test. The fibers from the ball bearing test suffered substantial degradation in size. The fibers of the Navy Grease Worker sample were much larger, but the grease itself was softer. In order to evaluate the changes that took place on a quantitative basis, an average fiber size was determined in each case by a semistatistical method. A longer, more tedious, fully statistical method was avoided because it was not necessary in this study. Although the method may not be exact, it does serve the purpose of supplying comparative information which is more accurate than the limited observations that can be made from the accompanying micrographs. Five or more micrographs were taken of the most representative areas of the specimens. Each frame was projected t o a magnification of five times, and dimensional measurements were made in all of some 30 to 40 fibers which were considered to be near average size. These measurements were made only where the dimensions were clearly seen. The averages of these measurements were then used t o calculate the length/width ratio. The values given in Table I are averages of three separate determinations of the fiber dimensions. One was made in each case by another operator, and agreement between operators was as good, in most cases, as that found in separate determinations by the same operator. Differences between operators developed not in the evaluation of sizes but primarily in the evaluation of the shape characteristics of the fibers-i.e., what one considered a cylindrical fiber another interpreted as being rectangular and where one found only dimensions of length and width, the other found dimensions of length, width, and height. I n order t o average these determinations, they were modified where necessary t o fit the convention of considering all the fibers to be square-ended and rectangular. The results of the working tests performed on the AML-3 and AML-8 greases are also given in Table I and the respective micrographs of these results in Figures 2 and 3. Each grease experienced a large increase in its micropenetration value after 100,000 double strokes in the Navy Grease Worker, and each grease maintained a low micropenetration after the high speed
INDUSTRIAL AND ENGINEERING CHEMISTRY
836
Figure 2.
Vol. 41, No. 4
A Electron micrographs (14,000 X) of lithium hydroxy stearate fibers of AML-3 grease showing effects of different shearing devices A.
A s received, unworked
B . After 100,000 double strokes of Navy Worker C. After high speed ball-bearing test
test; in two cases the value was lower than that of an unworked sample. The changes in the length/width ratios after the Navy Grease Worker test for the AML-5 and AML-8 greases are caused not so much by changes in lengths but by an apparent increase in the widths. Figure 1B illustrates well how the adhesion of several parallel-oriented fibers accounts for this increase in width. These laminated fiber aggregates were considered to be effectively single fibers in this study. Although the length/width ratios of the fibers, after the ball bearing tests, were considerably reduced, the even greater changes in the sizes of the particles cannot be disregarded. Throughout the data of Table I no correlation is possible between the length/ width ratios and the consistency of the samples in any of the tests. GREASE WORKING
A standard idea in the study of the rheology of greaselike materials has been that the greater the shear applied, the greater the resultant lowering of consistency. The results of the high speed ball bearing test apparently belie this notion, since what
Figure 3.
can be called extremely high shear rates have not effected any softening of the test greases. A possible explanation is that previous work has been primarily concerned with a secondary effect of shear stress due t o the conditions under which it is applied. T h a t is, if the primary effect was orientation of the fibers, then lowering of consistency might be a result of that orientation rather than a direct result of the shear stress. Thus, if shear stress were applied without orientation of the fibers, no change in consistency need necessarily result. The idea is that it is not necessarily the level or duration of shear stress which causes a change in consistency in a nonNewtonian material such as a grease but rather the mode in which the shear stress is applied. Shear stress in the Navy Grease Worker is applied by causing the grease to flow through holes, and each stroke supplements the previous stroke in slowly accumulating the effect of the repeated stresses; later paragraphs discuss this more fully. However, in the ball bearing test, the sample is subjected to very rapid turbulent actions which have no describable pattern. The effect of the stresses here is to degrade the solid particles on the micro level, but because of the mixing
Electron micrographs (8400 X) of lithium stearate fibers of AML-8 grease showing effects of different shearing devices A. B. C.
As received, unworked After 100.000 double strokes of Navy Worker After high speed ball-bearing test
April 1955
INDUSTRIAL AND ENGINEERING CHEMISTRY
and dispersing actions, they have no effect on the macro level. T h e results have confirmed that the macro level of the sample-Le., consistency-does not change when the flow rate through the bearing is changed from 0.38 to 0.038 cc. per minute even though the degree of particle degradation does n change markedly. I n the Navy Grease Worker, the sample changes markedly on the macro level (consistency) but little on the micro level (particle size). It is already well established that soap fibers in a grease tend to orient their long axes in the direction of flow and that they tend to B maintain this orientation ( 2 , 3, 6, 9). T h e soap in a grease is normally a threedimensional n e t w o r k of randomly oriented fibers, but the fluid dynamics of the S a v y Grease Worker cause a uniform orientation t o develop. I t is assumed that the degree of uniform orientation increases slowly C as more and more fibers yield to the high velocities in the orifices of the movable plate of the worker. After thousands of strokes, the supporting soap network begins t o break down as the unidirectional orientation proceeds with diminishment of the cross linkages necessary for plastic rigidity. Thus, while there is a minimum of change in the size of the fibers, there is a large disturbance to the soap mesh structure and, consequently, to the measured consistency of a grease. Many other devices that produce orienting effects in greases will also reduce consistency by the same mechanism. It follows then, that soap fibers or particles having high length/ diameter ratios will be the most readily affected by the orienting dynamics of the S a v y Grease Worker, whereas those particles having length/diameter ratios approaching unity will be the least affected. T h e small-fibered lithium hydroxy stearate grease, AML-3, has the smallest' increase in penetration after 100,000 strokes of the Navy Worker (Table I). Another factor whose influence on consistency has not been clearly established is what may be called secondary aggregation. It appears reasonable to assume that clumps of tightly bound fibers are produced in a grease by milling the hard greases that are initially formed. These clumps are associated with each other by the formation of contact bonds, but they are considered to have weak interfaces so that, on shear, they are readily separated and moved about as units. Work in this laboratory and others (8) indicates that greases often undergo a n initial thickening or increase in consistency during the course of ordinary working tests. Moore and Cravath ( 8 ) believe this is due to fibers splitting and shredding down their long axis, causing an increase in the length/diameter ratio. However, as observed in this laboratory, it is difficult to differentiate under the static conditions in an electron microscope between fibers that are splitting and fibers that are coming together to form laminated aggregates. The evidence presented here (Table I ) indicates that in some cases, fiber size increases in the Navy Grease Worker, and this
837
appears to be due to the formation of coherent laminated fiber bundles, created by the bonding of adjacent f i b e r s t h a t h a v e been oriented in a parallel manner. If the groups of fibers that are present as secondary aggregates were broken up and dispersed more u n i f o r m l y b y t h e initial working, then the initial increase in consistency might be accounted for. To illustrate how the factor of secondary aggregation is visualized, five conditions are represented in Figure 4. Condition A is that of a newly formed grease showing a perfect meshwork of fibers so well cross linked E that the grease is actually brittle. Condition B is that of most greases offered for Figure 4. Soap structures use; they have had a miniof a grease subjected to different conditions of mum of milling but possess treatment a great deal of secondary a g g r e g a t i o n . After the A . Newly formed-hard B. After minimum millinggrease has been sheared in stiff C. After short period in Navy the Xavy Grease Worker Worker-stiffer than B for a period, its condition is D . After long period i n Navy Worker-soft considered to be that of C in E. After high speed ball-bearing test-stiff which greater dispersion has been achieved and hence a stiffened consistency. After a long period of working in the same test, the orienting effects cause a small cross section t o appear, as D. Condition E illustrates the degradation of the fibers produced by the high speed ball bearing test, but it also shows how its turbulence has induced a very uniform dispersion of the fibers so that they may develop a highly cross-linked network. Indications of secondary aggregation ,have been observed in this laboratory; however, the method of specimen preparation, since it alters the soap fiber mesh, limits the observations that can be made. Some work on the elucidation of this mechanism has been done b y others (10) employing an alternate specimen preparation scheme. GREASE GEL STRUCTURE
The mechanism generally accepted in explanation of the peculiar plastic behavior of grease gels has been clearly outlined and advocated by Moore, Cravath, and coworkers ( 1 , 8). They speak of the form of the soap fibers as not unlike a mass of wood shavings or excelsior, and thereby they are prevented from settling out. As such, the soap fibers encompass a large volume of oil which is retained by capillary attraction. The irregular shapes are thus related to the structure of greases, and the length/diameter ratio of the fiber dimensions is considered the key influence in the resistance to flow of a grease. Thus, when a grease was subjected to shear stress so that the fibers were degraded in this length/diameter ratio, a correlation was found between the new ratio and the resultant lower consistency. I n direct contrast with these findings, the results of this study imply that the length/dianieter ratios of the fibers and the consistencies of the greases studied are essentially independent factors. The electron micrographs show that the fibers of the greases sheared in the high speed ball bearing have substantially diminished length/diameter ratios with only slight changes in
838
INDUSTRIAL AND ENGINEERING CHEMISTRY
penetration. This demonstrates that plasticity in these highly sheared greases cannot be attributed to the mechanical interference of the fibers with each other since such a pronounced change in length/diameter ratio would have required a comparable change in penetration. It is considerably more reasonable to relate the plasticity of these greases, containing such fibers as those shown in Figure 2C, to the ability of the fibers or particles to attract each other. It is possible that special conditions developed during this work to obscure the direct effects of a changing length/diameter ratio on the consistency of the samples, but it is definitely indicated that particle interaction plays a major role in the phenomenon of plasticity in greases, Whatever the exact nature of the interactive forces may be, there are other indications of their existence and importance, As Dean has suggested (6) syneresis or the contraction of a gel may be due to the forces operating between fibers. Likewise (‘settingup” or rebodying in a grease that has been sheared may be due to the re-establishment of many disrupted bonds. The different properties that have been produced in greases by varying the composition or chain length of the fatty acid portion of soaps (6, 7), may well be resolvable as quantitative differences in the interactive forces. Analogously, Chawalow (4)has shown that in sodium soaps the Van Der Waals forces acting between adjacent hydrocarbon chains and the forces prevailing in the ionic layers directly affect the crystal forms that the soaps assume. He has shown how variations in chain length produce variations in crystal form when formed under the same conditions. It is likely that more complete researches along these lines will result in sufficient understanding of grease gels so that it will be possible to make quantitative predictions of their physical properties from a fuller knowledge of the average soap fiber or particle. This knowledge, of course, would include the nature and strength of interparticle attraction and the size and geometry of the particle; it would, of necessity, presuppose a standard state of particle distribution such that variations in secondary structure and particle orientation would not distort the prediction. I n a strict sense the observations made here should be restricted to the MIL-G-3278 greases since the test work was performed only on this group, but is likely that they will apply as well to
Vol. 47,No. 4
many other classes. It is also recognized that there may be deviations in other classes of greases whose plastic nature differs sharply from that of the greases studied. CQNCLUSIONS
As a result of the test work performed, the following conclusions are made for the MIL-G-3278 greases and extended to other types: 1. Progressively larger amounts of shear or increasing shear rates applied to a grease do not necessarily result in a progressively lower consistency. The manner in which shear stress is applied must be considered as well as its magnitude and duration. 2. Mechanical breakdown of grease consistency could not be directly related to a lowering of the length/diameter ratio of soap fibers. Other factors, apparently, influence the breakdown of consistency. 3. A greater recognition must be given t o particle interaction as a basic factor contributing t o the plastic nature of greases. The role of mechanical interference of the particles to movement should be re-evaluated. ACKNOWLEDGMENT
The author wishes to express his appreciation for the assistance given him by E. R. Lamson who developed the ball bearing shear apparatus and who cooperated in evaluating the soap fiber dimensions. LITERATURE CITED
Bondi, A. A., Cravath, A. VI.,Moore, R. J., and Peterson, W. H.,Inst. Spokesman, 13,No. 12 (1950). Brown, J. A., Hudson, C. N., Loring, L. D., Ibid., 15, No. 11 (1952). Brbwning, G. V., Ibid.,14, No. 1, 10 (1950). Chawalow, M.L. E., J . Phys. Chem., 57,354-8 (1953). Dean, W. X., Inst. Spokesman, 16, No. 12 (1953). Dean, W. K., presented at annual meeting of National Lubricating Grease Institute, Chicago, Ill., Oct. 30, 1950. Aileyer, H. C., Inst. Spokesman, 16, No. 12 (1953). Moore, R. J., and Cravath, A. M., IND. EN^. CHEM.,43,No. 12 (1951). Vold, 14.J., Inst. Spokesman, 16, No. 8 (1952). Vold, R. D., Coffer, H. F., and Baker, R. F., Ibid., 15, No. 10 (1952). RECEIVED for review December 21, 1953. ACCEPTED December 3, 1964. The opinions expressed herein are those of the author and do not necessarily represent the views of the Naval Air Experimental Station or the Department of the Navy.
Adjustment of Vapor-Liquid Equilibrium Data D. B. BROUGHTON AND C. S. BREARLEY Universal Oil Products Co., 30 Algonquin Road, Des Plaines, I l l .
A
-
LTHOUGH methods are available for testing the thermodynamic consistency of equilibrium data, no methods have been proposed for refining data when they do not meet such tests. Since many of the systems reported in the literature fall in this class, such a method is highly desirable to permit maximum utilization of available data. A rational method of making such adjustments is based on an analysis of the systematic errors likely to be encountered in operation of the usual type of equilibrium still. T h e method has been applied successfully on a number of binary systems to bring into agreement the discrepant sets of data from different sources on the same system. One set of thermodynamically inconsistent ternary data, adjusted by the proposed technique, gave results in good agreement with consistent data on the component binaries as reported from other sources.
EQUILIBRIUM STILL CHARACTERISTICS
Most of the reported data have been obtained in recirculating stills of the types developed b y Othmer and Scatchard, with various modifications, Systematic errors may be introduced by: 1. Rectification, resulting from partial condensation of vapor on the still walls 2. Partial ‘Lflashing”of the condensate returped to the still, with failure of this flashed vaDor to come to equilibrium wlth the bulk of the still contents 3. Entrainment
Items 1 and 2 would tend to give an overhead vapor too rich in the more volatile component. It appears reasonable t o express this effect analytically by assuming that the still comprises the equivalent of s theoretical contacts a t total reflux. The cor-