Document DMKLb4zjZ2DrNqG0NbNkLXJ44

0.12 i have ed is from sm as ; of a t per . that r and orre- over Jy on pe of been )le to frac; move I ,rcap; iy be : cm a imul; jome: orber , and . 1930). . j) H ed., )25. ; York, , 1921). : 27-35, Engi .w-Hill ; ;roleum ; jhester, Oxidation of Unbodied Linseed Oil Antioxidant Influence of Phenol-Formaldehyde Resins on Unbodied Linseed Oil in the Presence of Driers HE phenomenon known as dry Ting of certain unsaturated polymerizable oils has been long C. C. VERNON AND W. W. RINNE University of Louisville, Louisville, Ky. double bonds particularly susceptible to polymerization. This study was limited to unbodied oil, and the results known. Many factors govern this drying, and many suba ttained should not be compared to those where bodied oils stances when mixed with oil accelerate or retard it. Many were used. theories havebeen proposed in an effort to explain this behavior. It is generally recognized that oxidation and polymeriza Method tion are the principal factors involved in the process. The After a study of the limitations of such methods as ultimate oxidation is believed to be confined to addition of oxygen to analysis, volumetric measurement of oxygen absorption, and unsaturated carbon atoms of those acids whose esters com blowing with conditioned air, the simple gain-in-weight prise the major part of a drying oil. The gain in weight fol method was chosen. lowing such addition of oxygen provides an approximate ThiB method gave only the apparent rate of oxygen absorp criterion of the rate of oxidation, despite the fact that vola tion, but in a study such as this one where conclusions were tile products are evolved during the drying process, especially to be drawn from a comparison of data collected under iden at the end of it. tical experimental conditions, it was considered adequate. Moureu and Dufraisse (5) proposed a theory relative to It had the further advantage of rapidity, making possible inhibition of autoxidation processes in general, which has many readings in the time available. been helpful in planning this study and which has indicated some of the procedures followed. This theory holds that The resin-oil solutions were prepared in concentrations cal culated as per cent by weight. Each constituent was weighed antioxidants are themselves oxidizable substances with the to 0.5 mg. on an analytical balance; for each gram of oil in the antioxygenic property localized in the oxidizable part of the molecule. solutions, 0.0427 gram of naphthenate drier was included, the metal content of which is given in the section headed Prominent among the antioxidants "Materials." The mixture was of Moureu and Dufraisse are phenols, which are used in the pro duction of many synthetic resins. Several investigators (4,9,10) have This investigation, using commercial materials, showed that the six phenolic heated to 150 C. for 5 minutes to ensure complete dispersion, sealed in vials, and stored in a dark cabinet 2 days before use. The mixture so pre pared was then spread on slightly : shown that phenolic compounds inhibit the oxidation of linseed oil during the induction period. Other investigators (1, 8) have inferred resins tested increased the induction period in the oxy gen absorption of unbodied etched, tared glass plates so that an area of approximately 95 sq. cm. was exposed. These plates were weighed immediately and at timed intervals thereafter; the intervals were the presence of free phenols in syn linseed oil in the presence of governed by the rapidity of drying. thetic resins. Careful tests of the samples used in this investigation, driers, in direct proportion to the amount of resin The initial weights of the films so pre pared varied from 0.1000 to 0.2580 gram, depending on the thickness of using diazotized p-nitroaniline as well as ferric chloride, showed the absence of any considerableamounts of unreacted simple phenol left present. The behavior of the oil after the induction period was apparently un the film which, in turn, was governed by the viscosity of the original solu tion. No attemptwas made to govern the thickness of the films, except to keep it within the limits prescribed by 1 there from the manufacturing proc affected, since it eventually Wise and Duncan (11) who showed ess. In one instance a somewhat ' complex phenol was detected and i removed. , Synthetic resins have been used dried completely. Mixtures of synthetic resins with ester gum had the same effect as that reasonable variation did not ap preciably affect the rate of oxidation. The present writers' experience in securing checks with films of the same resin mixture but varying in weight, in the manufacture of varnishes ' (or some time. Such varnishes have dried satisfactorily, although the pure resins, but to a less degree since the proportion of resin was less. Treatment and therefore thickness, confirmed the finding of these investigators. Since temperature (<?,10), humidity (7), and light (3) all influence the 1 the effect of the resins on such dry ing has varied. This behavior must not be confused with the results , reported for this study. The oil of two of the three resins that showed the most marked antioxidant effect drying of oils, it was necessary to carry out the experiments under conditions such that these factors were as nearly constant as possible. During the day the films were exposed to diffused used in the manufacture of these with an oxidizing agent re light, and the temperature variation varnishes was heat-bodied to begin with, became heat-bodied during the process, or else was tung oil sulted in a diminution of this effect. for the test periods was held within 2 C. limits. The humidity was much more difficult to control, but it was found that data collected when which has a conjugated system of the humidity varied from 50 to 60 1393 1394 INDUSTRIAL AND ENGINEERING CHEMISTRY agreed within the limits of experimental error. Most of the data were oollected at an average of 55. Weighings were made over a period of 180 hours, the intervals being longer after 60 hours. The most satisfactory agreements were secured during the first 15 hours, after which time uncon trollable factors caused some variation. VOL. 29, NO. 12 DECEM K 13 12 Materials The resins investigated met the usual industrial specifica tions and contained the amount of phenolic resin indicated in the following table. The diluent for those resins speci fied as less than 100 per cent phenolic was ester gum. The numbers given provide the key for reference in the discussion: Resin No. 1 2 3 Approx. Phe- Approx. Phe- nolle Resin Teat for Resin noHc Resin Con- Test for Content, % Free Phenol No. tent. % Free Phenol 100 - 4 45 + 100 -5 20 100 - 6 14 - The alkali-refined linseed oil used had the following values: Iodine No,, Wife method Acid No. Saponification value Specific gravity 135.3 0.65 193,20 0.931 The metal content of the naphthenate drier used was 0.30 per cent cobalt, 0.30 per cent manganese, and 4.60 per cent lead; the specific gravity of the drier was 0.855, Consideration of Curves The curves presented represent the percentage gain in weight of the oil in the resin-oil mixtures, not that of the entire mixture. It was shown in preliminary studies that thin films of the resin, spread from ether suspensions or solu tions, did not gain in weight when exposed to air under the experimental conditions which prevailed during this study. The amount of oil present in any given sample was easily determined from the original weight of the film and the per centage composition of the solution. The "raw" data were all recalculated by the use of an appropriate conversion fac tor so that the new data showed the percentage gain in weight of the oil. The conclusions were drawn by comparing the curves for the resin-oil mixtures with curves for linseed oil without resin; both curves were from data collected under identical experimental conditions. This method had the advantage that the inherent defects of the gain-in-weight method were eliminated. Each of these curves represents data collected in several runs under similar conditions and agreeing within the usual experimental limits. In addition to curves representing the resin-oil mixtures is the curve showing the behavior of linseed oil (Figure 2,1) and a curve (Figure 5,1) showing the behavior of ester gum, the diluent for resins 4, 5, and 6. The data for these curves were collected under the experimental conditions prevailing during this study and the curves have been included for comparison with the resin-oil curves. The curves for the 100 per cent phenolic resins (Figures 1, 2, and 3) show that each of these had antioxidant influence, in direct proportion to the amount of resin present. This effect was confined for the most part to the induction period. There was a loss in weight after the initial weighing in the higher concentrations of resin 1 and in all concentrations of resin 3. This portion of the curves was checked many times in an effort to determine the cause. It was found that heating the weighed sample of resin to incipient fusion under a vacuum before incorporation in the linseed oil re sulted in the elimination of this behavior, without affecting the rest of the curve. Boiling weighed samples of resin in water, drying, and incorporating with oil had precisely the same effect. Naturally, there was a loss in weight of resin by both of these treatments. Some volatile water-soluble substance was eliminated by these treatments, but it could not be identified with certainty, although occluded or ad 1120 hours sorbed formaldehyde was suspected. Elm (2) noted a simi did not, h lar initial loss of weight in the drying of trilinolenic glyceride, porated in which of course was not due to formaldehyde. 5, I). Th In order to test the applicability of Moureu and Dufraisse's treated in theory to the work in hand, 4.0 grams of resins 1 and 3 were The amt oxidized for one hour with 60 cc. of 30 per cent hydrogen solutions v peroxide. After thorough washing and drying, these sam in the figui ples were spread and weighed in the usual manner. Although phenolic re the antioxygenic properties were not completely eliminated per cent j (Figure 1,1, and Figure 3,1), they were sufficiently decreased scarcely be to substantiate the theory. nolic resins This same antioxidant effect was noted when commercial interface of resins made up of mixtures of phenolic resins and ester gum the latter were tested (Figures 4, 5, and 6). Ester gum itself was The temj found slightly oxidizable, gaining 1.7 per cent in weight after 260 rather ro. 12 DECEMBER, 1937 INDUSTRIAL AND ENGINEERING CHEMISTRY 1395 to 260 was identical with that heated to 150 within the limits of experimental error, and therefore was not included. An interesting point concerning resin 4 was that it yielded, on prolonged treatment with boiling water, 19.93 per cent of 4,4-dihydroxydiphenylmethane. When this substance had been removed, the dried residue, incorporated in oil and spread in the usual manner, showed markedly less antioxi dant effect (Figure 4, II). The other resins when so treated did not show any change in their antioxidant effect. With resin 3 an initial loss of weight was eliminated by this treat ment, as has already been mentioned. The increased viscosity of the solutions containing the larger amounts of resin was undoubtedly a factor in the rate at which the oil in the mixture absorbed oxygen, in that the diffusion of oxygen was rendered more difficult. No way of eliminating this was found. Then, too, there was less oil exposed at the surface of the,more concentrated solutions, which would tend to slow up the oxygen absorption. It was believed that the slope of the curves representing more con centrated solutions was due at least in part to these factors. Conclusion The six phenol-formaldehyde resins and mixtures of these resins with ester gum studied in this investigation showed marked antioxidant effect on pure alkali-refined linseed oil in the presence of driers. This effect was specific for each individual resin, and was confined to the induction period. This antioxidant effect was partially due, in two cases, to some oxidizable part or component of the resin, since it was mark edly decreased when the resin was treated with oxidizing agents. This last fact was in accord with the theory of Moureu and Dufraisse (5) relative to antioxygens, and indi cated a possible treatment of such resins should their antioxygenic properties be undesirable. The resins studied were apparently free from simple phenols used in their manufacture, although one of the resin-ester gum mixtures yielded a considerable amount of a complex phenol. The presence of free formaldehyde could not be demonstrated; yet in one case a behavior was noted which indicated the possibility of its presence. The results of this investigation must not be compared to those attained with heat-bodied oils, since the oil used here was not heat-bodied. Neither must it be concluded that these resins prevented the complete drying of the oil, for the oil did dry eventually. The results of interest are those shown during the induction period of oxygen absorption. TIME - HOWS 120 hours when spread as a film from an ether solution. It did not, however, show antioxidant properties when incor porated in linseed oil and treated in the usual manner (Figure 5, I)... The synthetic resins did not gain in weight when treated in the same manner as the ester gum. The amount of synthetic resin actually present in these I solutions was less than the percentage given for the curves in the figures. Thus a 30 per cent solution of a 20 per cent ^phenolic resin-ester gum mixture actually contained but 6 per cent phenolic resin. Curves for such mixtures can scarcely be compared to 6 per cent solutions of purely phe nolic resins, since the amount of oil exposed at the surface -interface of the former was much less than that exposed in rthe latter case. The temperature required to disperse resins 4 and 5 was 260 rather than 160 C. The curve for the linseed oil heated Acknowledgment This work was made possible by a fellowship granted to W. W. Rinne, the junior author, by the Louisville Oil, Paint, and Varnish Production Club. The authors wish to express their deep appreciation, not only for the fellowship, but also for the generous cooperation of the members of the club in furnishing materials and helpful suggestions. Literature Cited (1) Bradley, R., In d . En g . Ch e m., Anal. Ed., 3, 304 (1931). (2) Elm, A. C., In d . En g . Ch e m., 23, 881 (1931). (3) Long, J. 8., Rheinook, A. E., and Ball, G. L., Ibid., 25, 1086 (1933). (4) Morrell, R. S., J. Oil Colour Chem. Assoc., 10, 278 (1927). (5) Moureu, C., and Dufraisse, C., Chem. Res., 3, 113--62 (1927). (6) Rodgers, W., and Taylor, H. S..J.Phys. Chem., 30,1334 (1926). (7) Schmutz, F. C., and Palmer, F., In d . En g . Ch b m., 22, 84 (1930). (8) Seebach, F., U. S. Patent 1,891,455 (Deo. 20, 1932). (9) Tanaka, Y., and Nakamura, M., J. Soc. Chem. Ind. Japan, 33, 126 (1930). (10) Wagner, A. M., and Brier, J. C,, In d . En g . Ch bm., 23,40 (1931). (11) Wise, L. E., and Duncan, R. A., Ibid., 7, 203 (1915). Re c e iv e d December 14, 1936. Abstracted from a tbeaia presented by W. W. Rinne to the faculty of the Graduate Sohool, University of Louiaville, in partial fulfillment of the requirements for the M.S. degree, June, 1936.