Document Ozb8EY2V7eKO9QQk3gV02XRNQ

l j;NOMENCLATURE > El- ' ?* activatibnener^y, gram-cah /gram-mole ' ! ' e\ *;baseof:natu^^uog^UipiS:v. n' ; : f) *= combustible concentration exponent K == water gas equilibrium composition ratio at T k * rate constant, (liters)n"V(--K.)1/*(gram-moles)n_1 (sec.) njt moles inert gas/mole of oxygen in air N -- air rate, gram-moles/sec. 7i? - over-all reaction order P * pressure, atm. abs, . R -- gas constant, 1.987 gram-cal./(gram-mole) ( K.) or 0.08^06 (Hter) (atm. )/(gram-mole) (p K.) _r, riftj.,-ro- * radius at_probe tip, injector sphere, and reactor wall, respectively, inches T -- reaction temperature, K; To ' actual inlet mixture temperature, K. To e -- effective inlet mixture temperature (TQ corrected for heat loss), K. V = reaction volume, liters x = mole fraction' in combustion gases y = 4> for $ < 1; y ~ 1 for </> > 1 < =* fractional oxygen consumption efficiency <f> = equivalence ratio LITERATURE CITED (1) Avery, W. H,, and Hart, R. W., In d . En g . Ch e m., 45, 1034 (1953). (2) Brinkley, S. R.f JrM Smith, R. W.f Jr., and Sapsara, O., U. S. Bur. Mines, Rept. PX3-107/3, 1952. Semiannual , June 1954. 824 (1954). Mid Mattuck, A., (6).* Giasstqne;`yranik 'H., "Theory of Rate 1941. (7) Huff, V. Nv; amd Gordon, S., frail. Advisory Comm. Aeronaut., Tech. Note 2161,1950, (8) Longwell, J. P., in ``Selected Combustion Problems, Funda. mentals ^and. Aeronautical Applications," AGARD NATO _ Combustion- .CoUoquiiini, Dec. 7-*Tl, 1953, pp. 508-10. But' terworth; Lohdon^l954. (9) Longwell, J, P., Frost; E. E.,, and Weiss, M. A., In d . En g . Ch e m., 45, 1629 (1953). (10) Longwell, J. P., and Weiss, M. A., Ibid., p. 667. (11) Mullen, J. W., II, Feim, J. B,, Garmon, R. C., Ibid., 43, 195 (1951). (12) Mulready, R. C., United Aircraft Corp., East Hartford, Conn., Metqor Rept. UAC-9, 1947. (13) Prescott, R., Hudson, R. L., Foner, S. N., and Avery, W. H., -J. Chem. Pbys,, 22, 145 (1954). (14) Rosen, P., and Haft, R. W., J. Aeronaut. Bci., 20, 549 (1953). (15) Rossini, F. D., etal., "Selected Values of Physical and Thermo dynamic Properties of Hydrocarbons and Related Com pounds," Carnegie Press, Pittsburgh, 1953. (16) Spalding, D. B., Aircraft Eng., 25, 264 (1953). (17) Wohlenberg; W, J., Paper 53-A-103 presented at Annual Meet ing, Am. Soo. Mech. Engrs., New York, N. Y., December 1953. Re c eiv ed for review Ootoker 14, 1954. Ac c e pt e d February 9. 1955. Drying Oil Oxidation Mechanism, Film Formation, and Degradation S. B. CRECELIUS, R. E. KAGARISE, AND A. L. ALEXANDER Naval Research Laboratory, Washington 25, D. C. AMONG the more basic phenomena on which present radical chain mechanism which results in the formation of a protective coating technology is based is the oxidation of hydroperoxide group (--OOH) at the a-methylenic group ad drying oils. One would expect a process so basic to the coatingjsacent to the double bond carbon atom. He suggests that the industry to be the subject of intensive research as in fact it has oxygen adds first at the double bond to form a biradical been for the past several decades. Nevertheless, as Powers (18) has stated in a recent review, the oxidation mechanism of un saturated oils is still obscure due to the complexity of the oxida tion products and the uncertainty of their structure. HHH --C--0=C-- -f 02 H O* *O CH2 C--C HH Progress toward an understanding of the oxidation mech anism has been hampered in the past by the inadequacy of the This biradical then combines with another unsaturated group methods of direct chemical analysis. With recent develop to yield two radicals. ments of physical methods, however, particularly the techniques of ultraviolet and infrared spectroscopy, the scope of the analyti cal attack has been broadened considerably; it is by these means, coupled with the information already obtained on chemical structure, that the hope of obtaining more complete information of the autoxidation mechanism seems closer at hand. PREVIOUS WORK O *O 4- CH2C--C-- -- HH H O .* H H *0 --C--C=C~~ -j- --CH2--c--c-- H HH In recent years the older theory (7, 14, 17) that- the peroxida tion mechanism is due to the formation of a peroxide ring d;H HO--0 HR i,, + - ~ mU ot H^0 has been challenged by the hydroperoxide theory of Farmer (8, 9) who postulates that the autoxidation of drying oils proceeds by a These radicals continue the chain reaction and lead to the forma tion of a-methylenic hydroperoxides. As Powers (18) has commented, this explanation does not account for the uniform drop in iodine value during oxidation as found by previous workers (2, 6) exeept for the possible for mation of conjugated systems which would give a reduced iodine value. However, it is difficult to explain by this means alone the extensive and uniform drop in iodine value which has been observed. N41540  flggiw#,...* ss that*1"" mechamsmis du.f toth<e-acnrati^^ roethjdennioc. gcrrronuiiDpas Iwwhhiicchh aarreo^:flaainikkeedd~ro6nn :eeiittibhreerr ssiiddees-bhy^-^abuble bond. Tiiese radicals 'presumably 'undergd tba .Chapin rtbat result; hi increased conjugation ; T. ? /*' " ...;. . * -.; anal- \v--t. ^*M` fj 3$fUm.cri*?iv"';.*.'uii.'irik'jt *_*x H, H H. H II H rteoKi1o5S7^rm^pp^u^ha^. 14 X5bff ^exp8ecraimn ental HHHHH i (a) (b) HH -c--c==o -0=0--cEr,c=o-- (a) HHHHH HH ~0--C--0=0--C--CH,0=C-- procedureIws&A q 1 r-: ^ 1. liquid films oflinseed and dehydrated castor; oil, ^p^rp^iinaibely j0.p2 rnm. thick, were 8pJ'$aci, ;^'.$ie 8peetra were observed immediately. The oils'contained 0.03% (b) cobalt and 0.3% lead driers based on the weight of . the oil. Since the filni? weie.held in a vertical position during dig ^Gapping period, there were There have been many postulations as to the nature of decomr ireciable changes in thjcJ^iessV dqe to? flow- *nis was espe position products that are formed after tie initial stages of autoxidation Buch as the dehydration of peroxides to form ketone groups (18), formation of monoxide groups by loss of oxygen (5), cially serious during the first 4 hours of tne oxidation process. However, by reversing the orientation of the plates on alternate runs and by using relative absorbance values, the effect of this thickness change was minimized, : rearrangement of cyclic peroxides to form ketones and hydroxyls (7,11), and the breakdown of the carbon chain to form aldehyde 2. Between~r~u~ns th_e_f~ilmr *s --w--e--r-e---s--tox--red~ i-n- a j horizontal posit'io* n groups (18, SO). Since all these mechanisms apparently were based on the belief that the cyclic peroxide |^__J js the HH dominant form of oxidation, they are more or less inconsistent with the later theory of Farmer. These facts seem to be fairly well established 1. Oxygen is taken up by unsaturated drying oil films in the form o>ff peroxides, most likely hydroperoxides. 2. This uptake of oxygen is accompanied by a drop in iodine value indicating a decrease in the degree or kind of unsaturation. .3. Conjugation of C--C groups occurs in the oxidized films. 6.5, 7.5, 35, 90, 110, 155, and 200 hours. The infrared spectra of. a film of linseed oil obtained at various stages of the oxidation process are shown in Figure 1. Curve A was obtained immediately after spreading the oil over the surface of the plate and approximates the spectrum of nonoxidized linseed oil. Curve B shows the spectrum of the same film after drying 7.5 hours and Curve C after drying 90 hours. Figure 2 shows the same treatment of the infrared spectra of dehydrated castor oil. DISCUSSION With these facts in mind a study was undertaken in an attempt to clarify the autoxidation mechanism of drying oils by means of infrared spectroscopy. , Although a number of such studies have been described in the literature, discrepancies in the existing data and their interpretation emphasize the need for additional * work. In an initial effort, a preliminary spectroscopic study of the drying process in films of linseed and dehydrated castor oil was conducted. In addition, infrared spectra were obtained on similar films of these same oils that had been allowed to go through The observed data are summarized and compared with the results reported by previous investigators. As a matter of con venience the data are presented according to spectral regions. Unless otherwise stated, the observed results apply both to linseed and dehydrated castor oil films. 2 to 3 fi. Perhaps the most profound spectral change observed occurred in this region; namely, the appearance and rapid de velopment of an absorption band at approximately 2.93 The absorption peak reached its maximum value after drying t' * o .; '- >9 -Jiagi- ay *T . o ;Q V<7)> 5 ?y Vi x < * ?" oc h- !? J: L. LU1L _: u t- Ji t \> & f _c _ia-- ? _a Hz -9 <. ,,o 5J V> - Z* <^ OC ( I !_ J ? . _2_ .i 1 . - .-lQg- _ .12. \ -U !L li-J L_ r iHi 1-- rf- 22_ k KJ _! 9 A V. z / Jt "V 14 / L Y1 1 Lf J / t i / \A VH A If* u. 1z A4\ d/ 112 , / sS O l ( u _! SO / // * 19 f kn / 9 o 9 -JB2_ ?1 - 9 I i ... ( 7P r. fJ - X" }J --ji-- 4 r:: I ?X 4 t 4 & T * W 11 1* n1 Figure 1. Linseed oil infrared spectra at various autoxidation stages Sample: ------ -- --#-- -- Prism: Linseed oil 0 Hours (A) 7.5 Hours (B) 90 Hours (C) NaCI Resolution: 927 Response: 2 Gain: 5.5 Speed: 2.2 Suppression: 2 3? i;: > f &-A irL; n r~: c f- - 5 'r M -i : August 1955 INDUSTRIAL AND ENGINEERING CHEMISTp- RY 1645 10 hours and remained essentially constant during the remaining period (190 hours). Previous investigators have attributed this band to various compounds containing --OH groups, the favorite one being hydroperoxide. Nieholls and Hoffman (15), in studying the pentaerythritol esters of .linseed oil fatty acids, found a single band at 2.90 y which they attribute to the hydroxyl groups of alcohols and hydroperoxides. Dugan and coworkers (4) observed two distinct bands at 2.91 y and 2.88 y in the spec trum of methyl linoleate which they assign to hydroperoxide (R--OOH) and alcohol (--OH), respectively. Adams, Auxier, and Wilson (1) have observed the appearance of a single band at 2.78 to 2.83 y which they feel is due to the formation of hydro peroxides. The most intensive investigation of thi3 region toa date is probably that of Honn, Bezman, and Daubert (12). They examined the autoxidation of linseed oil over a 10-hour period and have interpreted the 2.9 y band as being a superposition of absorptions due to alcohol (R--OH), hydroperoxide (R--OOH), and carboxyl (R--COOH) groups. In addition, they have proposed a method whereby a quantitative estimate of the rela tive abundance of these three components is possible. This work is in disagreement, however, with the work of Overholt and Elm (16) who found little alcoholic content (about 2%) in oxidized oils. Because of the complexity of the oxidation process and the uncertainty of the reaction products, some of the assumptions of Honn and coworkers appear to be questionable. Thus it is generally agreed that the appearance of the band at 2.9 y in oxidized drying oils is due to the formation of some type of OH-group. However, the fundamental question as to the nature of these groups--e.g., alcohol, acid, hydroperoxide, or combinations thereof--does not appear to have a satisfactory answer in the existing spectral and chemical data. 3 to 4 fu The only spectral change observed in this region was the rapid disappearance of an absorption band at 3.3 y. Because of the limited resolution of the sodium chloride prism in this region, the band appeared as a shoulder on the strong absorp tion peak at 3.41 y. Adams, Auxier, and Wilson (1) have reported similar changes in the spectra of some dipeiitaerythritol esters by means 'of a double monochromator of high resolution. They interpreted this band (3.27 y) as being due to the C--H stretching mode of the carbon atom adjacent to the double bond-- i.e.} the <*-methylenic group. This interpretation was apparently based on a paper by Sutherland (23) in which he states "in unsaturated hydrocarbons (including aromatics) some of the C--H frequencies, arising from C--H bonds adjacent to the unsaturated bond, occur at a shorter wave length (3.25 y).n Sheppard (21), on the other hand, has interpreted the work of Sutherland and others on olefinic compounds by assigning the band at 3.27 y to C--H linkages attached directly to the double bonds--e.g., the group R--CH--CH--R2 has an absorption band at 3.30 yt whereas, the group Rj R2 C=CRaR4 has no absorp tion peak in this region. Thus the disappearance of the band under discussion does not substantiate the theory that oxidation occurs at the a-methylene group as concluded by Adams, Auxier, and Wilson (1) but rather indicates a reaction involving the double bond. One view (IP) that has been advanced is that vinyl polymeriza tion is one of the principal means by which the drying or solidify ing of an oil film takes place. If this mechanism occurs, the indicated action would be RC=CRi HH 4* HH RC--CRj HH R--O--CRj H CR, H Such a reaction mechanism is consistent with the observed decrease in the absorption at 3.27 y, since it results in fewer CH linkages adjacent to unsaturated bonds. Furthermore, the de crease in C===C linkages also accounts for the^drop in iodine value reported in the literature. In addition, it is well known that the presence of peroxide and hydroperoxide groups catalyze vinyl-type polymerizations. Although the spectral changes discussed do not substantiate the theory of Farmer (8) that peroxidation occurs at the amethylene carbon atom, they do not necessarily conflict with it. Thus the mechanism of Farmer and vinyl polymerization are compatible as this reaction occurs H O O RC--C=CR - HH 4* HH RC=CR H O OHH RC--C--CR H R--C, ^O-R H 4 to 6 y. The outstanding spectral change with oxidation in this region is the broadening of the ester carbonyl band at 5.72 u 1646 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 47, No. 8 toward longer wave lengths. Such a change ould result from the pro duction of carboxylic acids r ketones since both have absorption bands in the region from 5.75 to 5.85 y. An alternative explanation is that an unresolved C=C band is increasing in in tensity. Shreve, Heether, Knight, and Swern (22) have observed that per oxidation of unsaturated compounds (cyclohexenes) enhances the intensity of the C=C stretching mode. Nicholls and Hoffman (15) and Adams (I) have re ported increases in carbonyl absorption with drying time. 6 to 7 y. In previous in frared studies of drying oils no spectral changes were reported in this region for several reasons, the primary one being the presence of atmospheric water vapor absorption bands. More over, Shreve (22) has pointed out that bands due Figure 3. Increased con jugation effect on dehy drated castor oil infrared spectra to C=C stretching vibra tions are usually relatively weak. By using the double beam instrument ------ Untreated castor oil - -- - Conjugated castor oil dehydrated dehydrated the first of these difficul ties was eliminated. The changes observed in this region were 1. The weak band at 6.03 y rapidly decreases in intensity and disappears after about 6 hours. 2. A band appears at 6.14 y and increases in intensity; there is indication of a very weak band at this wave length in unoxidized dehydrated castor oil. 3. A weak band appears at 6.45 y after a period of 2 hours and reaches a maximum value after about 35 hours of exposure. After a period of 200 hours this band was not observed for the linseed oil film and was barely discernible in the case of the dehydrated castor oil film. Although these observations are preliminaiy in nature, they tend to show that a rearrangement of the double bonds occurs throughout the course of oxidation and such phenomena as cis-trans isomerization and conjugation are probably involved. In an attempt to ascertain whether the origin of these bands is due in any way to increased conjugation, a sample of dehydrated castor oil from the same source as that employed was isomerized by a modification of the method of Kass and Burr (IS) to increase the number of conjugated double bond systems present. This method consists of heating the oil in a medium of glycerol with an excess of potassium hydroxide necessary for saponification, for 24 hours at a temperature of 140 C. The acids are then freed, washed, dried, and re-esterified with glycerol. Diene values using the method of Ellis and Jones (5) are then determined for the isomerized product and the original untreated oil. The diene value of the treated product was 30.8 while that of the untreated dehydrated castor oil was 17.8. This indicates a substantial increase in conjugation in the isomerized product. This increase in conjugation resulted in the appearance of a weak band at 6.40 y as shown in Figure 3 which tends to substantiate -the view that the band at 6.45 y in the oxidized film is also due to conjugation resulting from the autoxidation process. 7 to 10 p. The spectral changes in this region were rather pootly defined and resulted primarily in a general increase in absorption with a subsequent loss of distinct absorption maxima. Changes in this region have been attributed to numerous func tional groups such as esters, epoxides, hydroperoxides, ethers, and secondary alcohols. Because of a vast number of groups ex hibiting characteristic absorption frequencies in this range, little if any reliable information appears to be available. ' 10 to. 12 y. Various investigators have previously reported changes in this region in the absorption spectrum of numerous drying oils. Nicholls and Hoffman (15) and Honn (12) have attributed the appearance of a band at 10.1 to 10.2 y to the pro duction of trans-ethylenic linkages. It is generally concluded that the naturally occurring fatty acids contain cis linkages only. Thus, *the conclusion is that the geometric structure of the double bond changes on oxidation. The band at 10.33 y increases in intensity. Moreover, this increase is accompanied by a decrease in the intensity. of the maxima at 10.15 and 10.98 y. If the correlations that exist in the literature are taken at face value, an increase of 10.3 y would indeed be indicative of an increase in the number of transethylenic linkages. By the same token the existence of bands at 10.15 and 10.98 y suggests the presence of an olefinic structure of the type RCH--CH2; a conclusion that is inconsistent with the accepted structure of the fatty acid components of linseed oil. Sutherland (23) has pointed out that many exceptions to the rule for classifying compounds with ethylenic linkages have been ob served. Thus, it is dangerous to form any conclusions concerning Figure 4. Percentage weight loss on ex posure to ultraviolet light cis-trans structures, particularly in those cases where two or three double bonds are. present in a single chain. Additional spectral studies with pure compounds of this type will be necessary before reliable correlations exist between the C--H bending mode fre quencies and the isomeric structures involved. INFRARED STUDY OF OIL FILMS AT VARIOUS PERIODS OF EXPOSURE TO ULTRAVIOLET Films of linseed and dehydrated castor were applied to potas sium bromide plates and allowed to stand for 24 hours in a con stant temperature room (24 C. and 50% relative humidity). At the end of this time the films were quite dry, having passed through the usual preliminary period of autoxidation. Prior to application of the films the potassium bromide plates had been weighed. After the 24-hour period the plates were weighed again and by difference the total weight of the dried films was calculated. The infrared spectra were then observed. The films were exposed to ultraviolet light emitted by a 110-volt Mineralight ultraviolet lamp Model No. 343 for intervals of 2, 6, 14, and 38 hours. The distance from the ultraviolet light source to the films was approximately 8 inches. The plates were weighed and the infrared spectra observed after each exposure period. After each exposure to ultraviolet light a substantial weight loss was noted which increased progressively with exposure time. The results are listed in Table I. August 1955 INDUSTRIAL AND ENGINEERING CHEMISTRY . V 1647 Table I. Effects of Ultraviolet Light Exposure Oil Linseed Original Wt. after 24-Hr. Drying, Gram 0.0116 Wt. Loss, Gram 0.0016 0.0044 0.0064 0.0112 Ultraviolet Exposure Time, Hr. 2 6 14 38 Original Wt. Loss, % 13.8 38.0 55.0 97.0 Dehydrated castor 0.0087 0.0006 0.0039 0.0051 0.0078 2 6 14 38 6.9 44.5 58.5 89.5 Table II. Elimination Rates for OH2, CII2, and 0=0 Groups from Films Oil Linseed Ultraviolet Exposure, Hr. 0 2 6 14 38 Absorbance Values 2.9 y (OH) 3.42 ft (CHi) 5.73 M (0=0) 0.265 0.239 0.178 0.171 0.095 0.900 0.760 0.630 0.451 0.169 1.040 0.990 0.940 0.645 0.207 Dehydrated castor 0 2 6 14 38 0.167 0.164 0.134 0.134 0.047 0.818 0.715 0.528 0.385 0.084 0.910 1.040 0.780 0.615 0.188 Figure 4 shows the percentage weight loss plotted against the time of exposure. These curves indicate that rates of weight loss for linseed oil and dehydrated castor when exposed to ultra violet light are quite similar. The infrared spectra observed after each period of exposure (Figures 5 and 6 for linseed and dehydrated castor oil, respec tively) show a general decrease in absorption indicating a loss in film thickness. These decreases are particularly evident for the strong bands at 2.93, 3.91, and 5.73 y previously assigned as being due to OH, CH2, and 0=0 groups, respectively. In order to determine the rate at which these various groups were being eliminated from the films, absorbance values were calculated for these bands, and the values are shown in Table II. Assuming that Beer's law holds for the materials being investi gated, the absorbance is directly proportional to the weight or concentration of the group causing the absorption. Therefore, by plotting absorbance percentage loss against ultraviolet exposure time, a curve should be obtained which would show the rate loss of the particular group comparable to weight loss. By this means it is possible to determine the extent to which each particular group contributed to the total weight loss of the films for each period of ultraviolet light exposure. Figure 7 shows absorbance percentage loss plotted against ultraviolet exposure time in the bands at 2.9, 3.4, and 5.73 m- The 2.9 band is indicative of OH groups either as hydroper oxide, alcoholic OH, or the OH of a carboxyl group. The de crease in absorbance indicates the loss of these groups. For the first 6-hour exposure linseed oil shows a rapid decrease in absorb ance in this region. For the next 10 hours a plateau is main tained in which no decrease in absorbance takes place. For the next 24 hours a steady diminution of absorbance is noted. The same general trend is observed for the dehydrated castor oil with a slightly lower rate for the first 6 hours. The same treatment is given to the absorbance data of the band occurring at 3.4 y--i.e., indicating CH2 groups. Here a uniform curve is obtained for both oils which begins to level off at longer periods of ultraviolet exposure. This indicates that CH2 groups are being lost at a gradually diminishing rate throughout the en* tire exposure. The slopes of the curves are practically parallel. Absorbance is plotted versus time for the carbonyl band at 5.73 y. The absorbance of the linseed oil decreases with exposure time in an almost linear manner. For the dehydrated castor absorbance increased slightly at 2 hours, continued to decrease at 6 hours, and then followed a linear decrease parallel to that of the linseed oil. It is hard to account for the increase shown at 2 hours. The instrument may have drifted due to the extremely low intensity at this point. On comparing these curves (Figure 7) with that of the total weight loss versus exposure time (Figure 4), it is fairly obvious that loss of the groups--e.g., OH, CH2, and carbonyl--indicated by the diminished absorbance in the bands at 2.9 y, 3.4 y, and 5.73 respectively, all contribute to the total weight loss of the films. The more rapid weight loss sustained by the films during the first 6 hours of exposure is probably due for the most part to the more rapid loss of OH groups (probably hydroperoxide) indicated by the steeper slope of the curve at 2.9 y during this period. The curves at 3.4 y indicating the loss of CH2 group are also steeper during this period although to a lesser degree than for 2.9 y. These weight losses are quite significant and illustrate the out standing tendency of ultraviolet light to accelerate the decom position of oil films in the presence of air. Overholt and Elm 1648 INDtf STBjLAL ANl^ 7jEN<3INEERIJIG CMEJAIST^Y Vq]L 47, No. 8 Figure 6. Dehydrated castor oil film infrared spectrum at various stages of ultraviolet light exposure (16) also reported weight losses of oil films on prolonged exposure to ultraviolet light. This decomposition is apparently accom plished at a number of points. The final breakdown products of attack have not as yet been identified, but they are obviously low molecular weight products, possibly C02 and water or vola tile organic compounds. An attempt to identify these products is planned by submitting similar films to ultraviolet exposure in a closed container in the presence of oxygen and subsequently de termining the infrared spectra of the gaseous products evolved- CONCLUSIONS The results of this experimental work may be considered in two parts 1. The study of the autoxidation of linseed and dehydrated castor oil films under artificial lighting conditions at normal temperature and humidity by means of infrared spectroscopy. 2. The effect of ultraviolet light on air-dried (24 hours) lin seed and dehydrated castor oil films as indicated by weight loss and change in infrared spectra. With reference to the first part of the study the spectral changes observed during the oxidation of linseed and dehydrated castor oil may be summarized: 1. The appearance and rapid increase in the 2.93 p region indicates formation of OH either as hydroperoxide or alcoholic hydroxyl together with a slight amount of influence due to carboxyl groups. These observations agree with the findings of earlier investigations. However, earlier studies have been able to determine chemically substantial amounts of peroxide in airdried films, but there is little evidence from a chemical standpoint for the formation of alcoholic hydroxyls. 2. The disappearance of the band at 3.27 p would indicate the replacement of hydrogens on the double bond carbon with some other radical, possibly by vinyl polymerization. 3. The broadening of the ester carbonyl 5.72 p indicates the formation of additional carbonyl groups either carboxyl or ketone. ~ 4. Changes in relative intensities of band 6.0 to 6.5 p regions indicate changes in the number and arrangement of 0=C link ages. The appearance of a band at 6.4 p is apparently the result of a shift toward conjugation. 5. Increase in absorption at 10.33 p with corresponding diminution of band at 10,15 and 10.98 p is also consistent with a rearrangement of 0=C linkages, particularly changes involving cis-trans isomerization. The second part of the study dealing with the effects of ex posure to ultraviolet light may be summarized 1. Both linseed and dehydrated castor films lost weight con tinuously with exposure to ultraviolet light. The rate of weight `loss was highest for the first 6 hours of exposure. After 6 hours the loss rate decreased slightly and continued uniformly throughout the remainder of the experiment until ultimately after 34-hour exposure the linseed oil film had lost 97% and the dehydrated castor 89.5% of its total weight. 2. Infrared spectra of the two oil films indicate a continuing decrease in absorbance at 2.91 p, 3.4 p, and 5.73 p for each period of exposure to ultraviolet. This decrease in absorbance was quite comparable for the two oils. 3. The decreasing absorbance at 2.9 p indicated the loss of OH-groups either as hydroperoxide or alcoholic hydroxyls. The rate of this decrease was quite high for the first 6 hours. After a 6to 14-hour exposure no decrease was noted. After 14 to 38 hours there was a uniform continuing decrease in absorbance at a some what lower rate than occurred during the first 6 hours. 4. The decreasing absorbance at 3.4 p indicated the loss of CH2 groups. The decrease in this region was continuous through out with a gradually lowering in rate for prolonged exposure. A. 2.9/* Figure 7. B. 3.4/', C. 5.7/i Absorbance percentage loss versus ultraviolet light exposure time August 1955 i N D U S T R l'A L A N D ENGIn EE 61 N'G t M i S T R' Y 1649 5. The decreasing absorbance at 5.73 y indicated the loss of carbonyl groups. The rate of decrease in this region was more or less uniform throughout the entire exposure for the linseed oil film; A slight increase in absorbance after a 2-hour exposure was noted for the dehydrated castor oil with a uniformly de creasing rate for the remainder of the exposure. 6, The loss of the individual groups as indicated by the infra red spectra agreed with the total weight loss of both oil films. This loss is apparently sustained by a breakdown of the film at several points--e.g., OH-group, CH2 groups in the chain, and carbonyl groups--and is greatly accelerated by the ultraviolet light in the presence of oxygen. The breakdown products are apparently highly volatile materials which escape and have not yet beeif identified. LITERATURE CITED (1) Adams, K., Auxier, R. W., and Wilson, C. E., Offic. Dig. Federa tion Paint & Varnish Production Clubs, No. 322, 669 (1951). (2) Coffey, S., J. Chem. Soc., 119, 1152-60 (1921). (3) Deatherage, F. E., and Mathill, H.A., In d . En g . Ch e m., 31, 1425 (1939). (4) Dugan, L. R., Jr., Beadle, B. W., and Henick, A. S., J. Am. Oil Chemists' Soc., 26, 681-5 (1949). (5) Ellis, B.A., and Jones, R.A., Analyst, 61,' 812 (1936). (6) Elm, A. C., In d . En g . Ch e m., 23, 881-7 (1931). (7) Fahrion, W., Z. Angew. Chem., 23, 722-6 (1910). (8) Farmer, E. H., Trans. Faraday Soc., 42, 228 (1946) (9) Farmer, E. H., and Sutton, D. A., J. Chem. Soc., 1943, p. 119(10) Farmer, E. H., Koch, H. P., and Sutton, D. A., Ibid., 1943, p. 541. (11) Frank, W., and Jerchel, D., Liebigs Ann. Chem.., 533, 46 (1938). (12) Honn, F. J., Bezman, I. I., and Daubert, B. F., J. Am. Chem. Soc., 71,812 (1949). (13) Kass, J. P.t and Burr, G. O., Ibid., 63, 1060-3 (1941). (14) Marcusson, J., Z. angew. Chem., 38. 148 (1925). (15) Nicholls, R. V. V., and Hoffman, W. H., Offic. Dig. Federation Paint & Varnish Production Clubs, No. 327, 245 (1952). (16) Overholt, J. L.t and Elm, A. C., In d . En g . Ch em., 32, 378 (1940); 32, 1348 (1940); 33, 658 (1941). (17) Paquot, thesis, "Autoxydation et Oxydation Par rOxygen." University of Paris, 1943. (18) Powers, P. O., In d . En g . Ch e m., 41, 305 (1949). (19) Powers, P. O., Overholt, J. L., and Elm, A. C., Ibid., 33, 1257 (1941). (20) Prill. E. A., Oil & Soap, 19,107 (1942). (21) Sheppard, N., J. Insti Petroleum, 37, 95 (1951). (22) Shreve, O. D., Heether, N. R., Knight, H. B., and Swern, D., Anal. Chem.,t 23, 282 (1951). (23) Sutherland, G. B. B., Trans. Faraday Soc., 41, 286 (1945). Re c e iv e d for review October 22, 1954. Ac c e p t e d February 23, 1955. Presented before Division of Paint, Plastics, and Printing Ink Chemistry. 126th Meeting, ACS. New York, September 1954. Thermodynamic Functions of Carbon Dioxide ENTHALPY, ENTROPY, AND ISOBARIC HEAT CAPACITY AT 100 TO 1000 C. AND 50 TO 1400 BARS DONNA PRICE Naval Ordnance Laboratory, White Oak, Silver Spring, Md. T THE present time, tables of thermodynamic functions Table I contains the smoothed values of the pressure-volume A based on pressure-volume-temperature (P-V-T) measure product, PVt as a function of pressure and temperature. The ments of carbon dioxide gas are available in the ranges 25 toterminal points and their sources are indicated. Most of the 150 C., 1 to 3000 atm. (<?, 7), and 0 to 600 C., 1 to 50 atm. (4). changes introduced by the smoothing were well within the pre Recently P-V-T data have been determined by Kennedy at cision limit of 0.2%. A few were larger but none exceeded the higher simultaneous temperatures and pressures (2). These maximum shift of 0.4% (2). data have been used in the present work to compute thermody namic functions in the range 100 to 1000 C:, and 50 to 1400 COMPUTATION OF THERMODYNAMIC PROPERTIES bars. The enthalpy was computed by SMOOTHING OF DATA H(T, P) = H(T, 50.66) -f The Kennedy P-V-T data (2) have a precision of 2/1000; these data were obtained by the use of reference values from the 150 C. isotherm determined by Michels and others (8, 10) (precision of 1/10,000), and agree to 0.2% or 0.0001 gram per cc. with the data of MacCormack and Schneider (3) (precision of 1 /10,000). The two sets of data of higher precision agree with each other to about 0.15% in the region of overlap. Difference tables of Kennedy's data (2) indicated that a func tional smoothing in two variables should be made. This was done by plotting the compressibility, - PV/RT (1) as a function of density, p, to form a mesh of isobars and iso therms. The data of higher precision (3, 8, 10) and a tempera ture extrapolation of the MacCormack and Schneider data (11) were used to obtain fixed terminal points. The remaining points, at corners of the mesh, were adjusted so that the smoothed values gave a satisfactory graphical net and also satisfied Equation 1. and the entropy by SIT, P) = S(T, 50.66) - pMp], where P is in bars, T in K., and V in Amagat units. The zero points for both. entropy and enthalpy have been chosen at 0 C. and 1 atm. The constants used for this adjustment are those of MacCormack and Schneider (4) and are included in the 50.66-bar (50-atm.) starting values. To obtain the starting values at 50.66 bars, the correction to the value at 1 atm. was plotted against temperature. Values for 75 to 150 C. (7) were assumed correct; those for 200 to 500 C. (4) and for 600 to 1000 C. (ll)t approximately correct, with greater weight given to the values at higher temperature. The final curve was visually chosen to pass through the first m