Document k9yLLya229wyOMxe9D1JD19wn

Studies in the Drying Oils XVII. Influence of Several Factors on the Mechanism of Drying of Oil Films1 J. S. Lo n g , A. E. Rh ein ec k , a n d G. L. Ba l l , Jr ., Lehigh University, Bethlehem, Pa. PkREVIOUS work has in dicated that the service able life of a paint film is determined in no small measure by the mechanism of the drying The first stage in the drying of linseed oil films is a matter of oxidation or, more generally, of making polar molecules. The mechanism of reaction in this stage is greatly influenced by of bodied oils, varnishes, and resins can be considered with that of paints and other prod ucts containing raw oil. We simply recognize that the first process. This in turn is greatly factors such as temperature and humidity. stage of the process is somewhat dependent on the composition High humidity greatly inhibits and modifies the different in the different cases. of the oil used and on the drying conditions, especially humidity and temperature. Abnormal environment, particularly in reactions occurring. The second stage in the drying involves association of the polar molecules produced in In the case of raw linseed oil, the first stage is largely a process of oxygen addition. Direct addi tion of oxygen at the double early stages, stunts the life of the first stage to form the solid gel-like coherent bonds has used up the free the film. This paper describes structure of the film. Some further oxidation energy of the oil. In the case of certain experiments on (1) the mechanism of the so-called dry ing process and (2) the influence of humidity and temperature. occurs after the oil has "set," but the film is soon oxidized to a stage corresponding to the addition of two oxygen atoms at each double bodied oils the free energy of the double bonds has been partially used up by the association or polymerization process, the It is a part of a long investiga bond, and the percentage of oxygen remains union of oil molecule to oil mole tion being carried on, one ob jective of which is the lengthen ing of the serviceable life of paint and other protective coat ing films in exterior exposure. practically constant thereafter. The aging of films of linseed oil or of trilino- lenic glyceride consists in a gradual transition of polar liquid phase to solid phase of sub cule. In either case a larger, more complicated, and polar molecule has been formed which has a strong tendency to associate with neighboring mole It is well known that in the stantially the same ultimate analysis. Em cules to form a solid gel. The first stage of transition from liquid drying oil to solid film as carried out by exposure to air, oxidation playB a major brittlement and failure of drying oil films is pri marily a matter of reduction of the percentage of liquid phase to low values. process consists of -two major phases irrespective of whether we are considering raw oil, bodied oils, or varnishes. role. The interesting ability of This view of the situation linseed oil to change to a solid film was known as far back seems helpful because it means one picture instead of several. as the tenth century. Understanding of the part played by However, there are definite differences in the primary mole the air waited until after the discovery of oxygen in 1774. cules built in the first part of the process. This means that Since that time we have had a fair the constitution and structure of picture of the nature of the process the resultant film will be different although, until recent years, inves in the various cases and suggests tigators failed to see that the later the possibility of producing desired stage of the process is not oxidation differences in a product by pur but is a matter of building up or posed variation of one of the factors association of the primary particles affecting the result; also it seems to form the solid film or gel. This that the possibilities inherent in part of the drying comes under the combination processes have not realm of colloid chemistry and been exhausted. This applies should be considered in the light especially to combinations of heat-, of colloid behavior. The later bodying and oxidation, of varnishes stage of drying and the aging of from oils processed to various ex films involves some oxidation but tents, etc. this is not the most important feature, and this oxidation is Fir s t o r Ox id a t io n St a g e masked by the other phenomena. If we recognize that the later stages in the setting and aging are colloidal association processes, then h the process of the setting and aging The simple idea of catching the volatile oxidation products given off and of determining the weight of each was extended to include comparison of the effects of high and 1 The first seven articles of this series appeared as unnumbered papers in In d . low temperatures and high and low humidity as follows (Figure 1): En o . Ch b m. as follows: 17,138,005 (1925); 18, 1245, 1252 (1926); 19, 82, 901, 903 (1927). Parts VIII to XVI appeared in Holes were bored in the bottom and lids of screw-top mason jars to admit October, 1933 INDUSTRIAL AND ENGINEERING CHEMISTRY 1087 Fig u r e 2. Wa t e r Ev o l v e d d u r in g Ru n I (Dr y Air ) Fig u r e 3. Ca r b o n Dio x id e Ev o l v e d d u r in g Ru n I (Dr y Air ) Fig u r e 4. Ac id s Ev o l v e d d u r in g Ru n I (Dr y Air ) were sealed on to the jars with de Khotinsky cement. Each jar was filled loosely with glass wool which had been carefully sprayed with the oil, oil + drier, or oil + drier + pigment. The weight of oil or paint sprayed on to the glass wool was adjusted so that each jar contained the same weight of oil. Precautions were taken that oil did not run down off the glass wool and collect on the bottom. In the beginning difficulty was experienced at this point. The loose packing of glass wool permitted fight from the tamp to reach all parts of the jar with fair uniformity. The results of duplicate experiments agreed within 6 per cent. Air let in the bottom of each jar struck a-shield of wire gauze to distribute it uniformly through the glass wool. Volatile products of oxidation of the oil passed with the excess air to a train consisting of (1) a calcium chloride tube to absorb moisture, (2) gas-washing bottles containing ether, ether +- water, ana potassium hydroxide to absorb fatty acids and carbon dixoide. At regular intervals these were replaced by fresh ones and the weights of water, carbon dioxide, and fatty acids determined. Aldehydes were found in the gaseous products in every case, but the weight of these were so much smaller that it was finally decided not to involve quantitative determination of aldehydes. For run I, samples were made up as follows: 1 Raw linseed oil .... Benzene (equivalent to that in the drier in 2-4) 2 Raw linseed oil Paste_--_--_-.>. (f4554.5% XraXwblilnascekezdinociloxide 3 2 Drier 143 Paste 4 73.4% basic carbonate white lead 26.6% raw linseed oil 1.5 Drier Samples 2 to 4 contained drier equivalent to 0.2 per cent metal (lead, cobalt, manganese) based on the weight of the oil. The drier contained: Lead resinate Manganese resinate Cobalt resinate Gram* 25 5.62 1.75 Linseed oil Benzene Grams 35 92 surrounded by an aluminum cylinder which protected the samples against direct light and thus tended to equalize the effect of light in the different parts of the jars. At the end of 30 days, or when the weight of volatile products coming off had decreased to a small figure, the jars were broken and the films examined. In all cases the oil or paint on the glass wool was thoroughly dry. The results are plotted graphically in Figures 2 to 4. In run II, the jars held the same materials as in run I, but the air led' through them was saturated with moisture by bubbling it through water in a series of gas-washing bottles. Results of this run are given in Figures 5 and 6. In this case water evolved could not be determined, but only carbon dioxide and fatty acids. However, these serve sufficiently for comparison. 1. Oils or paints exposed in diffuse light to very dry air, dry thoroughly at either moderately high or moderately low temperatures. The temperatures used are representative of the extremes likely to be met in painting practice. It is doubtful whether it would be logical to paint when the tem perature is much below 7 C. Yet the oils and paints dried perfectly. True, as expected, there is a difference in the rates of drying at 7 and 52 C., but the oil or paint dried thor oughly in each case. The weights of volatile products are, however, quite different, thus indicating very real differences in the mechanism of drying and in the. composition and nature of the resultant films. An interesting side light on this is afforded by the following experiment: Using sample 1, raw linseed oil alone was to all indications of touch perfectly dry after 30 days in the re frigerator. It had given off 1.654 grams of volatile oxidation products. This jar and sample were then put in the oven at 52 C., and the air stream was continued. After 5 days in the oven the sample had given off: In run I weights of these to give 5 grams of oil in each case were sprayed on the glass wool in the jars, and air was bubbled through sulfuric acid and calcium chloride to dry it, and over soda lime to remove carbon dioxide, and was passed through the four samples at the same rate of 2.5 liters per hour for 30 days. Two sets of jars were included in this run. The first set was kept in an electric refrigerator ' at 7 C., the secpnd set in an oven at 52 C. Each set was exposed to indirect light reflected from the white walls of the oven or the refrigerator from a 100-watt Mazda lamp, Fatty acids CO, Moisture Total Gram/10 gram* oil 0*240 0-495 0.881 Adding this to 1.654 grams gives 2.535 grams. The cor responding sample in the oven gave off 2.475 grams in 30 days. The agreement is close. 2. Excessive moisture or humidity in the air inhibits dry ing of the oil or paint. When the jars were opened after 30 1088 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 25, No. 10 days, the oils and paints which had been exposed to air prac tically saturated with moisture at the two temperatures were still undried or wet. They had been "drowned" by the high humidity. Drying had been inhibited to a large degree. This series brought out in striking fashion the inadvisability of painting under conditions of continued high humidity. This is not new but the work furnished quantitative data to sub stantiate the work of Schmutz and Palmer (4). the liquid changes to solid more rapidly under the stimulus of the sun's rays. Complete detailed study of the film after the oil has set will naturally involve separation of the solid from the liquid and a study of each phase separately. For tunately, this can be done simply, though it is a time-con suming job. If the film is removed from the surface and put in an extractor, such as a Soxhlet extractor, the liquid oil portion can be extracted from the film by means of organic solvents. Whitby has used acetone. The present authors' experience with it has been rather satisfactory though they have also used other solvents. The process called "solution" savors of chemical reaction in many instances, and in this extraction of the film with solvents it is felt that the solvent exerts a disassociating effect on the solid phase to some extent. The results are therefore relative or comparative rather than absolute. However, as the film ages from its initial set-up to its final failure, there is a steady and inter esting decrease in the percentage of (liquid) material extracted by a solvent such as acetone. This is true for both inside and outside exposure. It has been followed for many series of films. In other Words, as time goes on, one thing of major importance in the film is the steady increase in the percentage of the insoluble solid part. The film is becoming harder and more rigid. It is losing elasticity. These changes which are of interest in determining the life of the film may be conveniently observed by following the change in the per centages of liquid and solid in the film. In several cases it has been found that failure on outside exposure started soon after the percentage of liquid had decreased to low values. 3. Pigments and driers exert definite influence on the drying process. In some experiments not reported, the percentage of pigment or pigment-vehicle ratio was varied and was found to produce appreciable differences in the volatile products evolved. Se c o n d St a g e on Or ie n t a t io n a n d As s o c ia t io n o p t h e Mo r e Co mp l ic a t e d Po l a r Mo l e c u l e s t o Fo r m t h e So l id Fil m After the oil has absorbed oxygen corresponding to the formation of one peroxide group per molecule of acid, HH HH <U or three peroxide groups per molecule of oil, it sets or gels. A more general statement would be that, when the mole cules have grown to a certain degree of complexity or polarity or both, they associate and form solid gel. After this stage some further oxidation occurs, but this is outweighed by other considerations and soon becomes very small. The oxidation has practically.run its course before the film begins to undergo appreciable changes leading to failure. As shown by the data below, ultimate analysis seem? to indicate that oxidation has stopped or decreased to a very low rate before the films begin to undergo appreciable changes in physical character. Rather the gradual embrittlement or final failure of the films seems to be closely associated with steady and progressive change of thfe fully oxidized liquid phase to solid association products with only small changes in composition. Me t h o d o p St u d y o p Se c o n d St a g e . After the oil has set to what is called a "film," a new feature enters the situa tion. The so-called solid film really contains both solid "and liquid. Oil films exposed indoors contain appreciable per centages of liquid oil even after a year. Exposed outside, Ru n II (We t Air ) Duplicate films inside were still perfect and contained much larger percentages of extractable liquid oil. These findings are at least suggestions. They led further to the question, what factors determine the gradual change from liquid to solid? Is it a matter of slow oxidation catalyzed by the driers? Is it a change prompted by absorption of light energy, particularly of certain wave lengths or, more exactly, certain quanta of energy? Finding the cause may supply the means of hindering the transition of liquid to solid and thus increase the life of paint and other protective coatings. Nelson (#, S), from criticisms of stress-strain curves, premised and discussed the disappearance of liquid phase in films as they age. Linseed oil is a mixture of glycerides of at least four acids, stearic, oleic, linoleie, and linolenic. In starting this more detailed study, it was decided to avoid all possible compli cations due to mixtures and deal with pure substances. No.'10 imulus e; } . X syX- ie-conad put aid oil irg&nic ithors' ti they utiou" in thiB olvent some native im its interracted ie and ries of major mtage ar and which ay be e per uses it i soon ralues. { 1 1 much idings stion, ud to Iy the light actly, apply 1 and i igs. i urves, ; ise in I mplivnces. October, 1933 INDUSTRIAL AND ENGINEERING CHEMISTRY 1089 Trilinolenic glyceride was chosen as the material: (1) be cause it is the most active and therefore the most interesting component of linseed oil and (2) because the authors had learned how to make it in a relatively pure condition. There fore, this material was synthesized through the bromine derivative and was used to make up a number of series of panels, a few of which were on wood, but most of which were on glass since this permitted samples to be scraped off with a razor blade for extraction and study at regular inter vals. In some series driers were present, in others no driers were used. Supplementary experiments to confirm various leads were run at intervals. Results are given in Tables I to VII. Certain features of the process seem to be established by the data already in hand. These refer to trilinolenic Ta b l e I. Vo l a t il e Pr o d u c t s Fo r me d d u r in g Dr y in g Pe r io d o f Tr il in o l e n ic Gl y c e r id e Dr ie r % None 0.2 None 0.2 1.0 (Determined as carbon and hydrogen) Ti me of Ro n In c r e a s e O IN in C H. Fil m We ig h t Ev o l v e d Ev o l v e d Howe 100 100 ISO 150 160 % 27.16 28.26 31.10 29.41 28.00 % 18.10 17.76 22.80 19.20 16.60 % 3.19 3.08 8.10 7.50 6.02 (1.65). 0.61 1.42 1.00 0.99 To t a l Ra t io Ev o l v e d C:H 3.69 5.05:1 9.52 6.7:1 8.50 7.8:1 7.01 6.1:1 Ta b l e II, So l id a n d Liq u id Ph a s e s o f Tr il in o l e n ic Gl y c e r id e Fil ms So l v e n t Ethyl acetate Methyl acetate Isopropyl ether Methyl alcohol Ethyl alcohol N-propyl bromide Ethylene dichloride Carbon tetrachloride Chloroform -* Benzene Acetone Petroleum ether Ethyl alcohol Carbon tetrachloride Chloroform Benzene Acetone B. P. OF Liq u id So l v e )n t PHA8H Ox y g e n Co n t e n t Solid Liquid phase phase c. % % % 88 DAYS WITH 0.2 PER CENT DRIER 77 57 67 66 78.4 71.6 84 76 61.6 80 56.6 44.4 36.6 11.4 50.8 50.1 15.3 32.9 1.5 30.6 13.4 29.5 31.56 30.80 29 .*04 29.10 30.47 28.84 30l75 29.40 *29.80 o 33a.73 35.20 30a.43 35f.t92 32.40 25.72 38.34 42 DATS WITH NO DRIER 35-76 78.4 76.0 61.6 80.0 66.6 7.70 100 36*80 26.00 36.70 31.10 29.95 30.10 28.15 a 36f.t85 41.50 26a.92 a Sample too small. & Decomposition occurred. Fig u r e 7. w it h Time , er id e w it h In c r ea s e in We ig h t , o f Tr il in o l e n ic Gl y c 0.2 Pe r Ce n t Me t a l Dr ie r glyceride, not oil, unless otherwise stated. Although this material is in no sense of the word commercial, a study of it throws light on the drying of commercial products. Ch a n g e s in We ig h t The gain in weight (Figure 7) for trilinolenic glyceride with drier equivalent to 0.2 per cent metal shows that there is practically no induction period. Oxidation begins at once and proceeds steadily until the film is dry to touch. The net gain in weight amounts to 15 per cent at this point, which corresponds to a take-up of 8.1 atoms of oxygen per molecule of glyceride. When no drier is present, the net gain in weight up to the dry point is 19.25 per cent. It took 30 hours to reach the dry-to-touch point in contrast to 5 when drier was present. Drier therefore accelerates the rate of absorption of oxygen, but the results seem to indicate that the drier has other functions than that of an oxidation catalyst. Elm (1) has previously determined the gain in weight vs. time curve for trilinolenic glyceride. Figure 7 and Table VII give further data on this, extending Elm's work to show the influence of metallic driers. A smaller percentage increase Ta b l e III. Pe r c e n t a g e o f So l id Ph a s e in Tr il in o l e n ic a n d Tr io l e ic Gl y c e r id e Fil ms a t Va r io u s Fil m Ag e s Oil Co mp o s it io n Linoleum Oleic %% 100 94.65 89.20 77.70 66.20 49.60 0 5.35 10.80 20.30 33.80 6e0s..4io0 11.20 88.80 No determination, films not dry. 4 38.5 20.8 2e3..8s 0aa.0 (0.2 per cent cobalfc-manganeae-lead drier) -------Ps r c e j STAGS OF A.c e t o n e -In SOLUBLE PeCASE AFTER Da t s :------- 7 18 29 47 88 92 114 175 64.6 34.9 38.0 31.7 23.5 20.2 10.2 12.4 65.6 68.0 80.0 60.8 39.9 24.5 16.9 12.4 57.6 69.4 69.8 68.5 43.7 33.1 29.2 14.7 89.4 67.9 70.0 58.6 62.8 36.1 29.3 20.0 70!4 68.7 54.5 61.3 40.1 `34.1 18.6 74.8 74.8 74.2 66.0 64.1 38.5 34.7 ** 83.7 79.6 75.0 64.6 66.8 41.3 23.2 44 78.2 81.9 73.8 64.7 48.3 35.8 32.4 22.4 385 82.0 81.0 70.0 68.0 63.0 . 44 i-. Ta b l e IV. Ox y g e n in Ac e t o n e -In s o l u b l e Ph a s e o f Fil ms o f Tr il in o l e n ic a n d Tr io l e ic Gl y c e r id e s (0.2 per cent drier, exposed inside) Co mp o s it io n 4 Linoienic Oleic -------- JrBRCSlSTAGE 07 UXTGBN AFTER d a y s :-------4 7 18 29 ' 47 68 92 114 176 385 ' % 100 94.65 SO. 20 311 .90 11.20 % 5.36 10.80 20.30 33.80 50.40 69.10 88.10 29.50 29.48 30.40 30.28 28.71 30.00 29.78 29.65 31.18 32.16 34.52 35.22 28.48 29.48 32.15 32.94 32.84 32.44 32.64 31.25 31.69 30.92 31.35 29.80 34.12 32.78 31.45 33.30 32.47 31.91 33.54 32.72 34.08 31.72 34.90 36.12 31.13 31.66 29.48 31.50 32.90 31.35 32.58 37.48 31.12 30.55 30.63 31.61 32.49 34.13 32.45 30.88 30.44 31.80 31.74 32.67 34.08 32.42 36.36 31.50 30.04 31.90 30.90 33.40 34.27 32.29 38.05 31.20 29.76 37.10 30.97 34.75 . The method of study used consisting in (1) extraction to ; determine percentages of solid and liquid and (2) ultimate analysis of the separated phases to see whether the change from liquid to solid took place only after the liquid phase had [ been oxidised to a certain composition. in weight of the films up to the dry point or after 12 days is obtained as the percentage of drier increases. The differences are striking; thus, the percentage increase in weight when 3 per cent of metal is present is less than half of that when no metal is present. As shown by Table VIII, variation in 1090 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 25, No. 10 percentage of drier does not produce any definite drift in If the dry films are heated, there is definite loss of weight oxygen content. The figures vary slightly as is to be ex with evolution of corresponding amounts of water and carbon pected from the technic used, but are substantially constant dioxide. Thus, 16-day films heated for 3 hours at 130 C. at any given time. In one series (Table I, column 3) a down lost 8.17 per cent, and 40-day films heated for 20 hours at ward drift in oxygen content with increase in drier was 130 C. lost 7.63 per cent. observed. These two observations taken together indicate The mechanism of drying characteristic of the glyceride that increase in the percentage of metallic driers increases was not appreciably changed by the presence of a red iron the volatile oxidation products evolved during the early oxide pigment; thus, a mixture of 32.3 parts by weight stages of drying. iron oxide (99.4 per cent Fe203) and 67.8 parts by weight trilinolenic glyceride showed a gain of 8.15 per cent up to Ta b l e V. So l id Ph a s e in Fil ms the dry point. This is equivalent to (100/67.8) X 8.15 = (0.2 per oent drier exposed outside) 12.0 per cent on the basis of the glyceride. Fx l u Ag e Daya S 7 16 35 52 Ac BTONB-In BOIAJBLB Ph a s e Trilinolenic Linseed glyceride % 6.0. 5.8 13.0 14.3 -41.2 % 75.5 86.5 82.8 81.2 79.6 Fil m Ag e Days 17030 123 165 Ac ETONE-In SOLUBLE Ph a s e Trilinolenic Linseed glyceride % 3668..20 60.2 53.8 % 74.5 78.2 79.6 When baked for 20 hours at 130 C., this film lost 9.0 per cent calculated on the basis of the oil. These results are quite comparable to those for the glyceride above. Ch a n g e s in Co mp o s it io n Although the net change in weight is small after the glyceride is dry to touch, oxidation is proceeding at first Ta b l e VI. Ox y g e n in So l id a n d Liq u id Ph a s es a f t e r Ac e t o n e Ex t r a c t io n o f Fil ms o f Lin s e e d Oil a n d Tr il in o l e n ic Gl y c e r id e Fil m Ag s Days 3 7 16 35 52 73 100 123 165 Lin s e e d Oil OUTBIDS. i% DRIES Solid Liquid phase phase %% 26.60 30.08 34.29 33.30 30.88 32! 87 38.40 34.10 34.56 40...7. 2 Ox y g e n in Tr il in o l e n ic Gl y c e r id e INSIDE NO DRIER OUTSIDE, 0.2% DRIER Solid Liquid Solid Liquid phase phase phase phase % 29.90 29.70 30.48 32.45 29.90 % 30.14 30.79 31.05 34.13 37.04 35.80 35.00 31.16 % 28.83 26.30 29.65 30.76 32.06 34.10 34.78 34.45 % 32.43 37\88 33.40 ... rapidly, but the rate soon falls off to a very low value, and after some months the values for ultimate analysis remain sensibly constant. After this point further changes in the film are a matter of change of liquid phase into solid with little or no change in percentage of oxygen, rather than an oxidation. Further, the physical changes in the film which are considered as film failure occur long after gain in per centage of oxygen has ceased, as far as the writers are able to detect. This leads to the thesis that failure of films is not a matter of progressive oxidation catalyzed by' metallic driers but of progressive change of liquid to solid by reactions of an "association" nature. Substances which decrease the velocity of this change will be of service in prolonging the life of films. From this viewpoint, increasing the life of Ta b l e VII. In c r e as e in We ig h t , w it h Time , o f Tr il in o l e n ic Gl y c e r id e Fil ms Dr ie r % None 0.1 0.2 0.5 1.0 3.0 Dr y in g Tt h h Hours 40:00 6:14 6:21 2:56 2:33 2:13 --------In c r e a s e in We ig h t ------ At dry After 12 After point days heating % 19.30 14.90 14.80 13.00 11.89 9.43 % .. 16...0. 0 14.50 13.90 12.25 % 4.83 3*65 1.75 0.95 Loss ON He a t in g % ii*..,.i7 . 10.95 12.15 11.30 paints is not a hopeless task but is a matter of patient search for materials which will not retard the initial oxidation phase of the drying process, but will retard the latter stages of transition of fully oxidized liquid oil into solid gel. The first work had been done on films exposed in the laboratory. A second series which has been outside on the roof exposed at an angle of 45 facing south has so far given identical results. The percentage of liquid phase extractable from the film Ta b l e VIII. Ox y g e n Co n t e n t o f Tr il in o l e n ic Gl y c e r id e Fil ms Dr ie r % 0.0 0.1 0.6 1.0 3.0 At dry point 23.67 24.67 23.07 23.24 --------- Pe r c e 1NTAGB OF O XYGEN-------- After 3 days 8 days 12 days heating9 27.15 27.40 27.90 26.44 26.70 27.85 27.50 27.25 27.70 28.05 27.15 25.30 28.85 27.90 28.02 29.10 20.25 by acetone or other solvents decreased much more rapidly than in the indoor samples. The films also reached the maximum oxygen content much more rapidly but have not yet failed physically although they have reached what seems to be the maximum degree of oxidation. The work on the gross films has been supplemented by studies on the separate phases after extraction with acetone. The figures for the composition of the solid, acetone-insoluble phase are constant to an unexpected degree, showing that the 12-day films were heated at 130 C. lor 20 hours. solid fundamental structure of the film is fairly definite and constant in composition but that the percentage of it simply The effect of light in promoting gelation by. polymerization increases as time goes on. or association rather than oxidation is shown by the data On the basis of the formation of a peroxide group at each from trilinolenic glyceride with drier equal, to 0.2 per cent double bond, one molecule of trilinolenic glyceride would metal. This showed a gain in weight of 15.3 per cent before take up 9 X 2 = 18 atoms of oxygen. If no volatile oxidation it set in the dark, whereas the duplicate samples dried at the products were evolved, the solid oxidized product would same time in diffuse daylight in the room showed a net gain contain 33.1 per cent oxygen. It is known that some carbon of only 12.1 per cent when dry to touch. dioxide, water, and other volatile products are evolved, but The gain in weight measured is the net gain. Volatile such measurements as have been already completed (Table I) oxidation products are being given off at the same time indicate that the quantity is small and that this loss should but at a very slow rate. However, after the film is dry to decrease (Table IV) the percentage of oxygen in the solid touch, the two rates become nearly equal, the net loss in film to approximately 31-32 per cent. one week being only one per cent. With the smaller percentages of trioleic glyceride, the As shown by Figure 7, the change in slope of the curve figures agree closely with this. In those cases where the for trilinolenic glyceride is sharp enough so that the time percentages of admixed trioleic glyceride are high, there is a when the film is dry to touch may be determined by following definite drift of the figures towards 34 per cent. Apparently the gain in weight. the presence of fluid trioleic glyceride facilitates complete October, 1933 INDUSTRIAL AND ENGINEERING CHEMISTRY 1091 oxidation and minimizes the formation of volatile oxidation products. Whole unextracted films of trilinolenic glyceride with drier equivalent to 0.2 per cent metal (lead, cobalt, manga nese) contained 31.55 per cent oxygen (exposed inside 288 days) and 32.27 per cent oxygen (exposed outside 164 days). These figures indicate that the composition of the entire unextracted film is nearly the same as that of the acetone insoluble (solid) phase and has apparently not only reached a maximum, but the maximum is nearly the theoretical maximum for this material. It was logical also to follow the changes in the liquid phase by changes in ultimate analysis. However, the acetone apparently reacts with the liquid or else dissolves preferen tially from the "embryonic structure" those portions richer in oxygen. At any rate, the acetone-soluble phase contains a higher percentage of oxygen than the solid insoluble phase (Tables II and VI). This helps to justify the premise that the liquid is not simply waiting for progressive oxidation to reach the percentage corresponding to the solid phase and then change over to solid. It has been previously established by other workers that the peroxides first formed during the early stages of oxidation, as mentioned in the foregoing, undergo rearrangements involving formation of C=0 and 0--H groups as typified by the equation: hand, the stearic and oleic chains increase the percentage of the film disassociated and dissolved by hydrocarbon solvents. Tables HI and IV should be considered together. In running down any vertical column, it can be seen that the percentage of acetone-insoluble solid phase decreases as the percentage of oleic acid increases. Trioleic glyceride, like other fatty oils, is soluble in acetone. The results are there fore exactly as expected. They are, however, suggestive of the conditions in oil films, particularly those containing high percentages of oleic and stearic chains. In running along any horizontal line, the percentage of acetone-insoluble solid phase increases steadily as time goes on. This is not paralleled by steady increase in percentage of oxygen. On the contrary, the percentage of oxygen in the films, after reaching a maximum early in the life of the films, remains constant or decreases slightly except in the case of the films ridiculously high in oleic. Ac k n o w l e d g me n t The writers wish to thank the Archer-Daniels-Midland Company under whose research program this work is being carried out. The work on effect of humidity and temperature was suggested by the New Jersey Zinc Company, to whom deep appreciation is also due for helpful suggestions. Lit e r a t u r e Cit e d (1) Elm, In . En g . Ch e m., 23, 881 (1931). (2) Nelson, H. A., Proc. Am. Soc. Testing Materials, 21, 1111 (1921). (3) Nelson and Rundle, Ibid., 23, pt. 2, 356 (1923). (4) Schmutz and Palmer, In d . En g . Ch e m., 22, 84 (1930). AA HH Gelation and film formation (drying) are ascribed to the attraction and association of these polar groups. If it is a matter of association (secondary attractions) rather than polymerization (primary valence change) or condensation, it should be possible to disassociate the films by use of sol vents. Accordingly, the action of different types of solvents on films was studied. The results not only seem to strengthen the general premise of association but, on the basis that "like dissolves like," offer specific evidence of the existence of the polar C=0 and 0--H groups. This is brought out in Table II in which a series of different types of solvents was used which had boiling points as close together as practicable. The larger percentages of liquid phase were obtained with those solvents that contained C==0 or 0--H groups. Hydrocarbons and ethers extracted a smaller percentage of the film. In the course of this work many gels and films made from linseed oil were also extracted. The action of solvents on these was somewhat different from that in the case of tri linolenic glyceride. This is to be expected if we keep in mind that linseed oil is a mixture of mixed glycerides of stearic, oleic, linoleic, and linolenic acids. Acetone is a good solvent for raw linseed oil and for its polar oxidation products as discussed. Acetone therefore attacks the linseed oil films strongly. Alcohol is not a good solvent for linseed oil. The linseed oil films at various stages in their oxidation contain some 0--H groups which would be attacked by alcohol. The film, however, contains up to 10 per cent of stearic acid chains which do not oxidize and 10 to 15 per cent more of oleic chains which can have only one polar spot as against three in the linolenic chains of the trilinolenic glyceride. In general, therefore, alcohol has less action on the linseed oil films than on trilinolenic glyceride films, but, on the other Rec eiv ed April 6,1933. Presented before the Division of Paint and Varnish Chemistry at the 85th Meeting of the American Chemical Society, Wash ington, D. C., March 26 to 31, 1933. A. E. Bheineck and G. L. Ball, Jr., were Archer-Daniels-Midland Fellows at Lehigh University. Courtesy U* 8, Department of Agriculture Pe a n u t s Dr o p in t o t h e Tin Ca n a t t h e To p o f t h e Pe a n u t Bu t t e r Ma c h in e a n d Ar e Fo r c ed Al o n g b y a Re v o l v in g Sc r e w Wh ic h Ca r r ie s t h e Pe a n u t Ma s s t h r o u g h a Pa ir o f Gr in d in g Dis k s t o t h e Dis c h a r g e Sp o u t