Phenolic Antioxidants for Paraffinic aterials HERBERT iMORAWETZ Bakelite Corporation, ,Yew Yolk 17.
Thirty-four phenolic compounds and two xyleneformaldehyde condensation products were evaluated as antioxidants for paraffin wax at 163" C. and several relations betw-een chemical structure and antioxidant efficiency were established. Alkyl substitution in the reactive ortho and para positions and halogen substitution in the para position were shown to improve the stabilizing efficiency of phenolic antioxidants. High activities were observed with bisphenols linked by methylene, sulfur, or thionyl. Increased antioxidant efficiency was also obtained by introducing a second hydroxyl group into a substituted phenol. Substituted resorcinol and pyrocatechol were equally effective. The principles outlined above are applicable to the stabilization of paraffin wax and chemically similar materials.
A
NTIOXIDAIYTS perform a n import,ant role in niany chcmical industries in protecting foodstuffs, rubbers, resins, and a variet,p of chemicals against' atmospheric oxidation on storage or during processing at elevated temperatures. A wide variety of antioxidants is available, tm-o of the important groups being phenolic compounds and aromatic amines. The choice of antioxidant for a given application is determined by a number of considerations, such as compatibility, toxicity, color, effect on color stability in sunlight, odor, stabilizing efficiency, and price. The present investigation was stimulated by a request for a n antioxidant t,o be used in protecting paraffin wax against discoloration and the development,uf a rancid odor during processing a t temperatures above 100" C. Amines were excluded because of their discoloration characteristics. The investigation w a j restricted t o phenolic compounds and a systematic study of the effect of various substituting groups on the antioxidant efficiency of these compounds was undertaken. The results obtained were expected t o be applicable t o the stabilization of a wide range of materials cheniically similar t o paraffin mix. It mas also hoped t h a t the results of this study n-ould contribute toward clarification of the mechanism by whic3h phenolic antioxidant,s perform their function and would allon- the synthesis of conipounds of increased effectiveness. A number of methods have bccn used in the past for the evaluation of antioxidants. They all measure some characterist,ic change, physical or chemical, brought about by oxidation- and determine the modification of this change when antioxidants are added. Changes in electrical pori-er factor are characteristic of the oxidative degradation of polyethylene, mhile changes in mechanical properties of rubbers are usually taken as indications of the extent of their oxidation. Chemical analysis is commonly used to follow the oxidation of fats. Studies of the oxidation of liquids which can be saturated with oxygen or air are t o be preferred over studies on solids where surface attack and oxygen diffusion are limiting factors n-hich make interpretation of the results difficult'. When paraffin wax is subjectcd to the action of oxygen a t elevated temperaturcs no chemical change is detected for some time. At the end of the induction period peroxides are formed and their concentration rises rapidly t o a maximum. At t,his
.\-.Y .
point, "atty acids appear-probably by decompodtion of the peroxides-and the peroxide concentration diminishes gradually. h typical result illustrating the action of oxygen on Essowax a t 163" C. is shown in Figure 1. Tests in this laboratory have shown t h a t the suppression of peioside formation in paraffin wax by antioxidants correlates a it11 the preservation of electrical and mechanical characteristics oi chemically similar resin-. Therefore, a study of peroxide formation in paraffin wax containing substituted phenols was chosen as the method for evaluating their antioxidant efficiency. EXPERIMESTAL
The paraffin Ti-ax used in this work was supplied by the Standard Oil Company of Kerv Jersey, and is knorrn in the trade as Essowax. I t melted in the range of 50.5" t o 51.5" C. and had a n average molecular \?-eight of 370. A 50-gram sample of the n'ar was melted and the antioxidant was added as a n aliquot of a n alcoholic solution. The sample was heated t o 110" C., in order to boil off the alcohol, and poured into a U-tube fitted with a 1.25-inch diameter fritted-glass plug. Oxygen was passed through t,he cell a t a rate of 8 ml. per minute. the flow of the gas being started before the wax saniple was added i o as t o prevent the was from flowing through the fritted glass. The cell was immersed in a n oil bath controlled thermostatically a t a temperature of 163" C. The oxidation of the \\ax n a s follo\~eilby determining tlie peroxide number at hourly intervals. I 0.5Lgram sample was dissolved in a n Erlenmeyer flask in 40 ml. of carbon tetrachloride and a stream of carbon dioxide n-as passed through the solution. To the solution were added 45 mi. of 311 acetic acid solution of hydrochloric acid (made b>- mixing 4 i d . of concentrated hydrochloric a.cid v i t h 1 liter of glacial acet,ic acid), followed b>1 nil. of a 20% aqueous solution of potassium iodide. After 5 minutes, 50 ml. of water were added and the iodine liberated sodium thiosulfate. Thc induction was tit,rated with 0.008 period vas arbitrarily defined as the time. (in hours j required to obtain a peroxide number of 50 (50 milhcquivalents of sodium thiosulfate per 1000 grams of paraffin). This induction period v;as found t o be 1 hour for the unstabilized Essov-as. The "stabilization coefficient'' of an antioxidant, was dcfincd as the ratio of induction period incrctiv (hours j to antioxid:int, C O ~ C C I I -
i?
8 -
300
e 200 _I
-
I
100
0
0
1
2 TIME
3
4
OF OXIDATION, HOURS
Figure 1. Formation of Peroxides ant1 Fatty Acids in Paraffin W a x Oxidized a t 163" C .
1442
I N D U S T R I A L A N D E N G I N E E R'I N G C H E M I S T R Y
July 1949
tration (per cent by weight). This ratio was found t o be constant for the concentration range of 0.01 t o 0.05% of 4,4'-methylenebisphenol, but it is known t o decrease with most antioxidants at higher concentrations. The break a t the end of the induction period wss very sharp, the peroxide number rising within 30 minutes from vafues lem than 10 t o 200 and more. S o acids were formed before the peroxide concentration reached a high value. The results obtained with the 36 compounds evaluated arc listed in Table I.
1443
DISCUSSION,
The mechanism of antioxidant action was first studied by Moureu and Dufraisse ( 7 ) . I n 1917, they discovered t h a t the autoxidation of acrolein could be inhibited by pyrogallol and this discovery became the first industrial application of an antioxidant. Moureu and Dufraisse believed t h a t antioxidants react with organic peroxides in a series of reactions in which the peroxides are decomposed and the antioxidant is regenerated. h
TABLE I. EFFICIEWY O B ANTIOXIDANTSIN PARAFFIN WAX AT 163" C. CIsw O F Compound Substituted phenols
Comooiind 1
Concn., Wt. yo 0.1
2
0.1
10
90
Dow Chemical Co.
0.1
4
30
Eastman Kodak C.0
0 01
1 5
50
Bakrlite Corp.
5
0.01
3.5
250
Bakelite Corp.
G
0.01
1
0
Bakelite Corp.
0.01
5
400
0.01
1
0
0.01
2.5
150
Givaudan-Delawanna, Jnc.
0.01
4.5
350
Givaudan-Delawanna, Inc
11
0.01
5
400
GivaudaIi-Delawanna, Inc
12
0.01
8
700
Givaudan-Delawanna, Inc.
0.01
1
0
Givaudan-Delawanna, Inc.
0.01
*.
90
Structural Formnla
3
OH
4
'
CH1
H CIIi
CHI
A CHs
HO-O-B?
Induction Period4 5
Stabilieation Coefficientb Source Dow Chemical Co. 40
Br 7
C(CH3)1
\
CX-C-,-OH
Gulf Reiearch and Development Co.'s multiple fellowship, IClellon Institute
I
C(CH?!i S
Biphenyl derivatives Methvlene hisnhenol derivatives (A) Unsubstituted methylene
9
NO?-C>-OTT HO-C>-C>OII
10
OH
OH
Eimer & Amend
C1
13
Br
OH
Br 14
a b
Bi
0 1%
0 €I
Time required to obtain a peroxide number o i 50, in hours. (Hours, induction period stabilized wax) (hours, induction period unstabilized wax) % antioxidant eoncentration
Insoluble
Givaudan-Delawanna, Ino.
-
(Continued on pave 1444)
INDUSTRIAL AND ENGINEERING CHEMISTRY
1444
Val. 41, No. 1
~~
O F ASTIOXID.\STs TABLE I. EFFICIESCY
Class of Compound
Compound No.
Methvlene bimhenol deriratives (A) Unsubstituted methylene (Contd.)
15
(B) Substituted methylene
1G
I S P B R 4 F F I N \T4X AT
Striiotural Formula
OH
c. (CO?Lti7ZUf?d)
Induction Periodn
0.01
3
Stabilization CoefEcientb
SoLrce
OH
700
Giraudan-DF.lavc-anna. Inc.
100
Bakelite Corp.
HO-
18
9-Cresol-formaldehyde condensation products Purified fractions
163"
Concn.. Wt. %
19
0 I1
011 I
OH
0
0.01
5
400
Givaudax-Deiawanna, Inc
0.01
Ll
800
Givaudan-Delawanna, I n c
0.01
20
1900
Givaudan-Delxmanna, Inc,
0.01
13
1200
Givaudan-Delawanna, Ino.
0.01
7
600
Bakeiite Corp
0.01
7
600
Bakelite Corp.
I
dH3 20
Bakelit? Cory>.
1
CHa
OH
OH
CH3 21
Crude resin
OH
OH
OH
BH
22
CHr OH
I
Dihydroxybenzene derivatives
24
OH
OH
.
0.01
0
Bakelite Corps
(Concluded ur; page 1446)
INDUSTRIAL AND ENGINEERING CHEMISTRY
M y 1949
TABLE I. EFFICIENCY OF ANTIOXIDANTS Class of Compound
Compound NO.
Dihydroxybenzene derivatives (Contd.)
Structural Formula
25
YH 0
-
0
I N PARAFFIN W A X AT
163'
1445
c. (Concluded)
wt. %'
Induction Period"
Stabilization Coefficient b
0.01
2.5
150
Merck 8: Co., Inc.
0.01
7
600
Winthrop Chemical Co.
0.01
6
500
Bakelite Corp.
Concn
Source
.
27
28
1200
Nordigard Corp.
OH Phenol ethers
29
Sulfur hisphenols
0.01
1
0
30
0.01
18
1700
Givaudan-Delawanna. Inc,
30a
0.01
17
1600
Givaudan-Delarvanna, Inc.
0.01
11
0.01
1
33
0.01
8
700
34
0.01
11
1000
Bakelite Corp.
0.5
1
0
Bakelite Corp.
0.5
1
O\-CHs
O\-CHs
CHS
CHS
31
32
?H
0
"-j-0 1
Gallic acid esters
Diary1 methylene hydrocarbons
b
?H
'
1000
Givaudan-Delanranna, Inc.
0
Givaudan-Delawanna, Inc.
cI1
Time required to obtain a peroxide number of 50 in hours. (Hours, induction period stabilized wax) (houis, induction period unstahilized wax). % antioxidant concentration
-
Givaudan-DelaKanna, Inc.
..
Heyden Chemical Go.
Bakelite Corp.
1446
INDUSTRIAL AND ENGINEERING CHEMISTRY
close rela tion bet,iveen oxidation catalysts and inhibitors m-as demonstrated so that some compounds were eirher posidve or negative catalysts, depending on the chemical to which they Tl-ere added or depending on the p H of the solution. After following the changes taking place in antioxidants during their activity, Xlyea and Backst.roni ( I ) concluded that antioxidants interrupt oxidative chain reactions and are themselves oxidized in the process. Compounds stable to atmospheric oxygen may be excellent antioxidants as long as they reduce the peroxide fornied in the chain reaction which they are t o inhibit.
Vol. 41, No. 7
halogen has an effect similar to that of para alkyl. Also, Reaction 1 shows that the ring of a phenol is broken before the alkyl side chain is attacked by an organic peroxide. K i t h a comparat.ively mild oxidizing agent such as an organic peroxide, the 2-, 4, and 6-substituted phenols cannot undergo any of the oxidative reactions outlined above except for Reaction 3, leading to t,he formation of dimer peroxides. The evidence appears conclusive that this is the react,ion by vr.hich alkyl phenol antioxidants perform their function. Blocliing the ortho and para po3itions probably prevents undesirable side reactions Jvith the direct oxidation of the phenolic body by atmospheric oxygen. The oxidation of substituted plieriols uwd as antioxidants i n Thus, all of the stabilizer is available for the reduction of organic this study can take various courses clcpending on the position 01' peroxides in the material being protected. the substituting groups. Boesclien, l\letz, and Pluim (3) have While halogen in the para position effects an improvement in shown that p-qninone and muconic acid arc the main rcac tion the ant,ioxidant efficiency of the compound, similar t,o that of products of the oxidation of phenol by organic peroxides. II a para alk5-l or halogen substituted phcnol is oxidized no quinones para methyl substitution (Table I, compounds 10, 12, and 19), are formed and the reaction products arc substit,uted muconic two compounds cont,aining several halogen atoms in one benzene acids. ring are ineffective (compounds 6 and 13). l l a r s h and Butler reported a similar effect in a study of the fungicidal activity of OH COO11 COOH I I halogen substituted phenolic compounds ( 6 ) . In comparing the antioxidant efficiency of various bisphenols, it is evident that the methylene group of methylenebisphenols and their derivat,ives is of great importance. This is demonstrated not only by the much higher antioxidant, efficiency of methyleneR hisphenol derivatives as compared t o analogous single ring compounds, but also by the great loss in activity vherever one or both In s o n i ~m 3 e q , n-here the para position is irloc.lii:d dimer methylene hl-drogcns are substituted (Table I, compounds 11, Eovmatioii is observed on oxiitatioii of phcnolii~ c~nipouiida. The niechnnisrn of this reaction is discussed L y Garnett and co16, 17: and 18). On the othcr hand, compounds containing wor!rers ( 2 ) . Cn oxidizing &ii~iphthol\=,-it11ferriL. cliloricle, t he diary1 methTlene, but no phenolic hydroxyl (Table I, compounds link is formed in t h e 1-position (5). 2!3> 35, and 36) exhibit no antioxidant activity. It appears, thevefore. that the high antioxidant activity of the methylenebisphenol derivatives is due t o some interaction of the methylene groups with the phenolic hydroxyl. I t is possible that quinonemethide, n.hich has often been assumed to be a t'hermsl decommay be the active position product of methylenebisphenol (4, stabilizer. The p-cresol-foviiial~ehydecondensation products constitute an inkresting series. The great increase in antioxidant efficiency in passing from the 2-ring t o the 3-ring compound may be acThe forination of ellagic acid 011 oxiclizin:: gallic acids foilons a counted for by the higher niethylene concentxation in the trimer similar mechanism (8). .In interesting case is the oxidation of 0-chlo~~n-lO-ph~naiitl~rol.and also by the 2-, 4-, and 6-subslitulion of the middle ring. and 6- positions arc In this compound, all the reactive 2 - , S o explanation is offered for the decrease of activity in passing ldoclied and on oxidation the dinier peroxide is obtained (5). from the trimer to the tct.ramcr. However, commercial p cresol-formaldehyde resin is less efficient as an antioxidant, than any of the pure Eract,ions t,ested (compare pure compounds 19, 20, and 21 wit,h the crude resin 2 2 ) . I t is suggested that this may be due t o t,he presence of free p-cresol in the resin. Cresol oxidizes comparatively easily and thus part of the actual antiPbO? (3) oxidant may be consumed in t,erminating oxidation chain reactions of p-cresol. This view is supported by evidence that antioxidant efficiencies of phenolic resins are greatly improvcd aft'er small amounts of monomeric phenols are removed by stea,ni distillation. INTERPRETATION OF RESULTS A sulfur link between two phenolic nuclei was shon-n to contribute t,o antiosidant efficiency even more t,han methylene. The The results of the experirnental v-ork listed in Table I justify sulfoxide link is comparable to methylene vc-hile a sulfone link the conclusion that, alkyl substitution of phenolic compounds in completely inactivates the phenol (Table I, compounds 30, 31, the react,ivr 2-, 4-,and 6- positions greatly increases their antiand 32). oxidant efficiency. This may be strikingly illustrated by comUnsubstit,uted resorcinol is ineffective and pyrocatechol is a paring the antioxidant efficiency of three monocyclic phenols in poor antioxidant but they become excellent antioxidants on paraffin wax. alkylation. Wherever a second hydroxyl is introduced into a S u m b e r and benzene ring of an antioxidant, the st,abilieing efficiency is greatly Position of Stabiliiation increased (compare Table I, compounds 1 and 4 with compounds Compound Alkyl Groups CoefficienL 26 and 2 7 ) . On the other hand, the similarity of the antioxidant 1,1,3,3-Tet~~arnetlis.Ibutglybenol 1 (ilara) 50 Di-tori-bntylphcnol 2 ( o r t h o , para) 250 characteristics of pyrocatechol derirat,ives and resorcinol derivaDi-tert-butS1-p-cresol 3 ( o r t h o , o r t h o , para) 400 tives is additional proof that no oxidation to quinones is involved The increased activity of the phenols on alkylation cannot be in the mechanism of antioxidant action. due t o the oxidizability of the alkyl side chain since tert-butyl side From this study it appears that the best antioxidants are cornchains are known to be unaffected by oxidizing agents and para pounds which are stable in the presence of atniospheric oxygen
July 1949
INDUSTRIAL AND ENGINEERING CHEMISTRY
but which react rapidly with organic peroxides. It may be expected that a more reactive compound may be required for best antioxidant efficiency a t low temperatures than for use at high t.emperatures. A number of isolated facts seem to support this assumption. A systematic study of the change in relative stabilizing efficiency with temperature, for various antioxidants, would be of fundamental importance. ACKNOWLEDGMENT
The author is indebted t o C. S. Myers of Bakelite Corporation, under whose direction this work was carried out, for his many helpful suggestions. The assistance given by many of the writer's associates of Bakelite Corporation is also gratefully acknowledged. The experimental procedure was based on information received from R. M. Koppenhoefer of the Socony Vacuum Laboratories. W. S. Gump, of Givaudan-Delawanna, Inc.,
1441
helped the project greatly by supplying many of the compounds evaluated. LITERATURE C I T E D
Alyea, H. N., and Backstrom, H. L. J., J . Am. C h a . SOC.,51, 90 (1929). Barnett, E. de B., Cook, J. W., and Matthews, M. A . , J . Chem, Soc., 123, 1994 (1923). BGeseken, J., M e t a , C. F., and Pluim, J., Rec. trav. chim., 54,345 (1935). Euler, H.v., ;Idler, E., aud Caspersson, A. O., Chem. Zentr., 1943 11, 20. Fieser, L.F.,J . Am. Chem. Soc., 52, 5204 (1930). Marsh, P. B., and Butler, M.L., IND. ENG.CREM.,38,T O 1 (1946). Moureu, C., and Dufraisse, C., Chem. Revs., 3, 113-62 (1926). Richter, G. H., "Textbook of Organic Chemistry," 2nd ed., p. SOB, New York, John Wiley & Sons, 1943. R E C E I V EApril D 30, 1848.
Sulfite Disap earance in Dehydrated Vegetables during Storage R. R. LEGAULT, CARL E. HENDEL, WILLIAM F. TALBURT, AND LOIS B. R4SMUSSEN Western Regional Research Laboratory, Albany, Calif.
Sulfite disappearance in dehydrated sulfited carrot, white potato, and cabbage stored at temperatures ranging from 24' to 49' C. proceeds approximately as a first-order reaction. The apparent activation energies, calculated according to the Arrhenius equation, are high, ranging from 33 to 43 kg.-cal. The rate increases markedly as the moisture content is raised; for carrot and white potato at 38' C., the increases are about 3- and s-fold, respectively, over the moisture ranges of 5.4 to 8.0, and 5.3 to 9.2%. For sulfited vegetables stored i n air as compared to similar samples stored in nitrogen, the respective rates of sulfite disappearance are in the ratio of about 1 to 1 at 49" C., and 2 to 1 at 24" C. Little change in the sulfite disappearance rate in dehydrated carrot is caused by varying the blanching time from 2 to 8 minutes, by sulfite application during or after blanching, or by application of the sulfite from solution or from gas. On the basis of the generalizations presented herewith, it is feasible to estimate the life expectancy of the sulfite in other samples of dehydrated sulfited \ egetables.
I
T IS generally recognized t h a t the presence of sulfite decreases
t h e rate of discoloration of dehydrated fruits and vegetables, both during dehydration and during storage of the dehydrated product (2-4, 7 , ll).' It is known also that the sulfite disappears fairly rapidly on storage of t h e dehydrated fruit or vegetable a t temperatures above 35" C. and that the rate of disappearance increases as the temperature is raised. Since the rate of discoloration increases after t h e disappearance of the sulfite, knowledge of the rate of sulfite disappearance is of importance t o the dehydration industry. Stadtman and co-workers (9) have reported on the rate of sulfite disappearance in dehydrated apricots during storage, but little information is available for dehydrated vegetables. The principal purpose of this paper is t o report studies of the rate of sulfite disappearance in dehydrated sulfited carrot, white potato, and cBbbage during storage, as a function of the
moisture content, the temperature, and the atmosphere during storage. The results of studies of this rate as a function of the blanching time and the method of sulfite application for one lot of carrot also are reported. These results were obtained in connection with studies of the rate of browning of dehydrated vegetables during storage ( 5 ) . MATERIAL AND M E T H O D S
The commodities studied were dehydrated carrot, white potato, and cabbage. Analytical data on the samples studied and description of the methods of preparation and storage were reported previously ( 5 ) or are included in Tables IV and V. The dehydrated vegetables were packed either in air or in nitrogen containing 1.5 * 0,5'% oxygen and were stored at temperatures ranging from 24" to 49" C. Samples were removed from storage periodically for evaluation. The sulfite content of these samples was determined by the method of Prater, Johnson, Pool, and Mackinney (8). I t is reported as parts of sulfur dioxide per million parts of vegetable on a moisture-free basis (M.F.B.); this is equal t o the parts of sulfur dioxide per million parts of moisture-free solids. Table I records the standard deviation of replicate sulfite determinations on one sample of each commodity after it had been stored at 49' C. until approximately half of the sulfite had been destroyed. I n obtaining these results eight determinations were made on each sample; each of these determinations was carried out on a different day so t h a t day-to-day variation in the personal factor might be taken into account. Replicate determinations on samples of t h e undeteriorated vegetables, reported by Prater et al., show approximately the same degree of reproducibilitj-. Somewhat greater deviations were observed among results for storage samples, presumably because of lack of uniformity among the samples. The moisture content of the dehydrated vegetables Kas determined by the vacuum-oven method of Makower, Chastain, and Nielsen (6). This method \\as used because it gives results
