Infrared Absorption Spectra of Some Hydroperoxides, Peroxides, and

67, 412 (1945). (6) Richards, R. E., and Thompson, ... Holland and Hughes (5), however, have reported that in the autoxidation of the polyolefin, squa...
0 downloads 0 Views 513KB Size
ANALYTICAL CHEMISTRY

282 epoxy compounds that contain 5- and 6-membered rings. I n addition to strong ether absorption in the 9-micron region, these compounds show strong bands in the 11- t o 12-micron region. Although additional reference spectra will be required before generalizing, the common band near 11.4 microns in the Fpectra of tetrahydropyran and dioxane is probably characteristic of the 6-membered heterocyclic owgen ring in these compounds. Similarly, the 11-micron h i d in the tetrahydrofuran spectrum is probably characteristic of the 5-membered heterocyclic oxygen ring. The spectra of other compounds that contain similar ring. confirm this assumption, but the puritv of these additional compounds was not sufficientlv certain to justify presentation of thc,ir spwt ra here. LITERATURE CITED

(3) Barnes, R. B., Gore, 11. C y . , Stafford. 11. IT., and Williams, V. %.,

ANAL.CHEM.,20,402 (19481. Field, Cole, and Woodford, J . Chem. Phys.. 18, 1298 (1950). ( 5 ) Findley, T. W., Swern, D., and Scarilan. J.T.. J . A m . Chcrn. .Cor., 67,412 (1945). (6) Rirhards, R. E., and Thompson, H. IT,,Proa. Roll. Soc., A 195, l(1948). (7) Shreve, 0. D., Heether, 11. K., Kuight, H. R., and Swrrn, D., ANAL.CHEM.,22, 1498 (1950). (8) Swern, D., Billen, G. S., and Eltldy. C . R., J . .lm. Chern. S t i c . , 70, 1228 (1948). (9) Swern, D., Billen, G. S.. :md Knight, H. B., Ibid.. 7 1 , 1162 (1949). (10) Swern, D.. Billen, G. S . , a n d Scankin, ,J, T., Zbid., 68, 1504 (1946). (11) Sm-ern, D., Findley, 1’. JY., Hillrn, G.S., and Scanlan, J. T., A N A L . CHEM., 19. 414 (1947). (4)

(1) Anderson, J. -4., and Seyfried, JT, D., ANAL. CHEW.,20, 998

(1948). (2) Barnes, R. B., Gore, R. C.. Liddel, U., and Williams, 1.. Z.. “In-

frared ,:Spectroscopy, Industrial Applications and Bibliography, Yew York, Reinhold Publishing Corp., 1944.

RECEIVED M a y 31, 1950. Prrsentcd i n part before the Division of Organic Chemistry a t the 116th Meeting of the .IMERICLS C H E h i r c a L SOCIETY, Atlantic City, 1.J. Paper V I in the series “Reactions of F a t t y XIaterials with Oxygen.” Paper Vis ( 7 ) .

Infrared Absorption Spectra of Some Hydroperoxides, Peroxides, and Related Compounds 0. D. SHREVE AND R.1. R. HEETHER, E. I . du Pont de Nemours & Co., Znc., AND H. B. KNIGHT AND DANIEL SWERN, Eastern Regional Research Laboratory, Philadelphia, Pa. The formation of peroxidic substances during autoxidation of unsaturated materials derived from fats and other substances has long been known, but their structure, quantity, and mechanisms of formation have not been adequately established. Infrared absorption spectra of a series of pure hydroperoxides, peroxides, and related compounds from 2 to 15 microns have heen obtained and interpreted. On the basis of empirical analysis of the spectra of the hydroperoxides and their parent compounds it has been tentatively concluded that the hydroperoxide group gives rise to a characteristic absorption band near 12 microns. Study of the spectra of peroxides indicates that the peroxide linkage probably gives rise to a strong absorption band in the 10- to 12-micron region, but the frequency corresponding to this hand is sensitive to changes in the structure of the groups attached to the peroxide linkage. The spectra are primarily intended to serve as reference data in the application of infrared spectrophotometric methods to the analyses of autoxidation mixtures.

T

HE formation of peroxidic substances during the reaction of unsaturated materials with oxygen has been known for a long time, but the structure of these reaction products, as well as the mechanism of their formation, has not been completely established. Originally, i t was proposed t h a t all the oxygen combined directly with the double bond, yielding a product which was saturated and contained some kind of cyclic peroxide structure (9-13). The recent isolation from oxidized nonconjugat’ed olefins of pure a-methylene hydroperoxides, in which the double bond wa6 still intact ( 7 , 17, 18, 21), suggested to some investigators that hydroperoxides must be the initial products of oxidation. Consideration of the high energy requirements for hydroperoxide formation, however, coupled with the fact that conjugated compounds containing a-methylene groups autoxidize by addition of oxygen a t the double bond, prompted Farmer (14-16) to suggest that autoxidation of olefins is universally initiated by addition of oxygen a t the double bond of a few molecules only, forming free radicals. Subsequent reaction occurs by chain reactions in which the free radicals at,tack the a-methylene position. I n contrast to t h k , Hilditch and co-workers ( 2 , 20) have reported that at 20” peroxidation of methyl oleate occurs , at significantly t,o a large estent at a-mrthylrnr ~ I ’ O I I ~ Pwhrwas

higher temperatures double bond attack predominates. Bolland and Hughes (5), hoiTever, have reported that in the autoxidation of the polyolefin, squalene, two of the oxygen at,oms form a hydroperoxide group and the remaining two form a n intramolecular peroxide ring. These differences in opinion have prompted t,he authors t o consider the use of a physical method in studying the naturcl of the peroxidic substances formed during the initial stage of oxidation. The main advantage of a physical method in oxidation studies would be the possibility of eliminating the need for isolating labile substances present only in small amounts. I n a relatively short time, infrared spectroscopy has achieved marked success in the qualitative and quantitative determination of oxygen-containing functional groups. in a large numbcr of both short-chain and long-chain compounds (1, 3, 4, 24, 2.5). The literature on orgaiiic peroxides, however, is extremely sparse, and in only a few isolated cases has infrared spectroscopy heen employed in oxidation studies (6, 8, 19, 22). I n these investigations, however, no rc>pc)rtw:ts rnatle of the use of infrared techniques in tietermitling the nature of the peroxidic substances i’ormd.

V O L U M E 2 3 , NO. 2, F E B R U A R Y 1 9 5 1 &fore it is possible t o use infrared spectroscopy as a tool in determining the constitution and quantity of the various types of peroxidic substances iornied during the autoxidation of compounds derived froni fats and other materials, it is necessary t o have available reference spectra on pure model compounds. I n t h i ~paper are reported the infrarrd ahsorption spectra of a serirs of pure hydrop,rroxides, peroxides, and related conipounds

I

283 from 2 to 15 micwxir. t,he discussion.

Interpretntion of the spectra is given in

EXPERIMESTAL

Spectrophctometer. The spectrophotometer and the technique for obtaining the spectra have beendescribed (24). All compounds which were liquid a t room temperature were run directly in a standard Beckman liquid cell. Tetralin hydroperoxide, a solid, was run as a 10% solution in .03M M . C E L L c a r b o n d i s u l f i d e . Benzoyl peroxide, whose solubility in carbon disulfide is comparat'ively low, was run as an approximately 4vo solution. Materials Used. Tetralin: Tet,ralin melting point hvdroDeroxide. 54.0-54.5"; cumene hydroperoxide, boiling p o i n t 65'at 0.1 mm. and n v 1.5221; cyclohexene; cyclohexene hyd r o p e r o x i d e , boiling point 40-41" a t 0.2 mm. and n1,0 1.4892: t e r t - b u t v l h v d r o peroxide, boiling point 34:5-35 O a t 13 mm. and TZY1.3987; and di-tert-butyl peroxide, boiling point 38.3" a t 51 mm. and n y 1.3865, were prepared as drscribed in a Drevious mihlioation ( 2 5 ) . Cimene (isipropilbenzene), boiling point 152" anti nZ,U 1.4910, and tert-butyl alcohol, boiling Doint 82" at

--c

-

n

a zz"t

/AEf-kA u \-/ \ I

.03MM. CELL

distillation of the purest commercial grades. Benzoyl perF purest commeroside W ~ the cial grade and was used as recrivcd. liethy1 oleate hydro-

60

4o

m

20

m -

7

:.03 MM. CELL

.OSMM. C E L L

The spectra of the four pure hydroperoxides studied (terthu tyl hydroperoxide, cyclohexene hydroperoxide, Tetralin hydroperoxide, and cumene hydroperoxide) and those of the parent compounds from which they are derived are shown in Figures 1 and 2. Figure 3 shows the spectra of two pure peroxides. Sample form is indicated in the lower right-hand corner and exact wave-length positions of absorption maxima arr indicated on each curve. Common Absorption Bands Attributable to Vibrations of Hydroperoxide Group B e cau'e of the imuortance of hydroperoxides in studies of oxidation mechanisms, it is of primary interest to examine the sprctra of Figures 1 and 2 for ahsorption bands which might he related to vibrations of the 0-0-H (hydroperoyitle, linkage, .4ll an infrared spectrum arise from vilirations of the molecule as a

.

V;;I

21

I I

131 I

Figure 2.

1

I41 I

AAnm,.nnMn

I I 191 I I llbl I l 1 1 ' 1 1 I Illkl I I l l 3 I I1141 I 1 W A V E LENGTH. M I C R O N S

1151 I 161 I 111 ' 1I I I 8 l ~

Infrared Spectra of P u r e Hydroperoxides and Parent Compounds

I15

ANALYTICAL CHEMISTRY

284

n

r

whole, and it is not strictly 0 correct t o think in terms of DI-TERTIARY vibrations of a particular linkBUTYL age w i t h i n t h e m o l e c u l e . PE ROXl DE Nevertheless, it is a well-known 2 20 e m p i r i c a 11y established fact fn io 2: that many functional groups and specific structural units do give rise to bands whose fre560 BENZOYL N -IN quencies are substantially inACTIVE ID K PEROXIDE ID dependent of the structure of * 4 % S'LN. IN cs2 .I MM.CEL I- 4 0 the remainder of the molecule. 20 IOne band whose wave-length 1 position can almost always be relied on t o remain nearly conh stant with change in molecular structure is that due to 0-H . 0 3 M M . CELL stretching vibration. As ex20 401 pected, all four hydroperoxides exhibit this band near 2.8 21 I i3!.i i I 4 1 i i 151 I I - l s I I I l ? l I I I S I I i i41 I i ilbl i 1 I l ' l I I l l l k l I Ill31 I 11141 Ill5 m i c r o n s . Comparison with WAVELENGTH, M I C R O N S the tert-butyl alcohol spectrum indicates no appreciable Figure 3. Infrared Spectra of Two Pure Peroxides difference between the hydrcperoxidic and the alcoholic I n general, the introduction of a substituent into a ring compound 0-H stretching frequency. Dilution with carbon disulfide brings about marked changes in the spectrum. Because the shifts the 0-H absorption maximum to shorter wave length preparation of cyclohexene and Tetralin hydroperoxides involves in both cases, indicating considerable hydrogen bonding in the group into a ring, the spectra of these the introduction of 0-0-H condensed state. compounds do not retain many of the spectral features characIn attempting to select an absorption band that might be teristic of the parent compounds. In the case of cumene hydroattributed t o vibrations within the hydroperoxide group, or peroxide, however, no new substituent has entered the ring and vibrations of that group moving as a unit, it seems best to conmost of the spectral features of the parent cumene persist in aider first the spectrum of tert-butyl hydroperoxide in relation the hydroperoxide spectrum, although some wave-length shifts to that of its parent compound, tert-butyl alcohol. These comand changes in intensity are apparent, The following bands for 0-H in pounds differ only in the substitution of 0-0-H which appear in the cumene spectrum are usually observed at the molecule, and their spectra are less complex than those of the approximately comparable wave lengths in a variety of monoother compounds of Figures 1 and 2. T h e following bands which substituted aromatic compounds: the strong bands a t 13.2 and appear in both spectra can be assigned with reasonable certainty 14.3 microns associated with the benzenoid structure; sharp as indicated: a band near 2.8 microns due to 0-H stretching; bands near 9.3 and 9.8 nlicrons; two bands near 6.2 and 6.7 bands near 3.3 and 6.8 microns due t o C-H stretching and bendmicrons; the 3.3- and 6.9-micron bands; and the doublet near ing respectively; a band near 7.3 microns related to symmetrical 7.2 and 7.4 microns. The last bands mentioned are related deformation vibrations of the methyl groups ( 3 , 2 6 ) ; and two to C-H stretching, C-H bending, and methyl group deformabands a t 8 and 8.4 microns related to vibrations of the terltion, respectively. The additional sharp band seen in the 9- to 10butyl structure ( 4 , 26). The two bands at 9.8 and 13.4 microns micron region (near 9.6) may possibly be related to vibrations are of uncertain origin, but both are present in both spectra. Thus, of the isopropyl group in this compound, although a shorter the only marked consequence of replacing 0-H by 0-0-H wave-length range has been assigned to vibrations of this group in this case is the disappearance of the strong band a t 11 microns \ (4,26). All these bands persist, in the cumene hydroperoxide (possibly related to OH bending motions in the -COH strucspectrum, if it can be assumed that the 9- and 9.3-micron bands / in the hydroperoxide represent the 9.2- and 9.6-niicrori bands ture) and the appearance of two new bands, a weak one near shifted to somewhat shorter wave length and that the 6.2-nlicron 11.3 and a strong one near 11.8 microns. From this it is conband has merely undergone a decrease in intensity. Thus it cluded that if a band exists between 2 and 15 microns (other than would seem that all but t>hreeof the strong bands in the cumene the 0-H stretching band) which can be attributed to a vibration hydroperoxide spectrum ran probably be ussigned to various viwithin the hydroperoxide group (or of that group moving as a brations of the cumene wsidue in the molecule. This leaves three unit), the 11.3- and 11.8-micron bands represent the only posstrong bands t o account for. Of these three, the two a t 7.9 a n d 8 . i sibilities. Because bands arising from vibrations of oxygenated microns (unlike the 12-micron band) are not common to all four groups are usually strong, the strong 11.8-micron band is the hydroperoxides. This leaves the strong 12-micron band as the most likely choice. most probable assignment to a vibration involving the peroxide Having thus established a limited spectral region as a definite linkage in the 0-0-H group or possibly a bending motion of the possibility, it is now of interest t o examine the spectra of the hydroperoxide group as a unit. remaining three hydroperoxides in relation to those of their parent Because of the complications incident t o ring substitution, the compounds in the vicinity of 11.8 microns. On so doing it is assignment of the 12-micron band to a vibration of the hydrofound that each of the three does indeed exhibit an absorption peroxide group in cyclohexene hydroperoxide and Tetralin band near 12 microns which is absent in the corresponding hydroperoxide is less certain. However, in view of the above parent compound. arguments for the probable reliability of the assignment in the The argument for assigning the band near 12 microns to a cases of tert-butyl and cumene hydroperoxides, together with the group is strengthened on further convibration of the 0-0-H fact that the band does appear in all four cases, the authors have tentatively concluded that the presence of the hydroperoxide sideration of the cumene and cumene hydroperoxide spectra. h

"y

-

(v

V O L U M E 2 3 , NO. 2, F E B R U A R Y 1 9 5 1 group in a n organic molecule gives rise t o a n absorption band in the vicinity of 12 microns. Aside from the assignments already mentioned, no further interpret'ation of the complex cyclohexene and Tetralin hydroperoxide spectra will be attempted. One interesting feature of the cyclohexene hydroperoxide spectrum, however, should be noted. While the 6-micron C = C stretching band in the cyclohexene spectrum is only moderately strong, the intensity of this absorption in cyclohexene hydroperoxide has been great,ly enhanced. The change in molecular symmetry resulting from group on a carbon adjacent to the substitution of the 0-0-H double bond has evidently brought about a marked increase in the infrared activity of this vibration. Included in Figure 3 is a n absorption curve run on methyl oleate hydroperoxide (purity 69%). T h e broad absorption near 11.7 microns represents a considerable increase over that observed in the spectrum of a sample of pure methyl oleate. Some of this increased absorption probably arises from the hydroperoxide groups present, but conclusions on this point must await the availability of a pure sample of methyl oleate hydroperoxide, which the authors are now attempting to prepare. T h e failure to observe a stronger increase in absorption in the 12-micron region than was actually obtained might seem surprising. I t has been observed, however, in studies (24, 26) of other oxygenated long-chain compounds containing oxirane and hydroxy groups, t h a t in the liquid or solution state the absorption bands attributable to these oxygen-containing groups aril broad and weak, whereas in the solid state they arp sharp and intense. This suggests that infrared spectra of mixtures of oxidation products be detrrmined a t low teniperatures in the solid state to bring out bands which are broad and weak in the liquid state. T h e strong 0-H stretching band near 2.8 microns arises from hydroperoxidic hydroxyl and other types of hydroxyl groups present in this comples mixture. T h e st,rongband a t 10.36 niicrom (not present in methyl oleate) is characteristic of trans-octadecenoic acids and esters, thus indicating that. geometric isomers may be formed under the oxidizing conditions emploj-cd in preparing this mixture from methyl oleate. PEROXIDES

Only two pure pwoxides ivailahle for study a t thc time of this report. Their spectra are show.ri in Figure 3. Di-tert-Butyl Peroxide. The spectrum of this compound bears a strong ovw-all resemblaiicc t o those of krt-butyl alcohol and tert-butyl hydroperoxide. Fi.oni the discussion of the latter two spectra i t will be clear t h a t all bands in the di-tert-butyl peroxide spectrum can probably be attributed t o vibrations of the alkyl residues, with the exception of the strong band near 11.4 microns and the very weak band a t 11 microns. T h u s by reasoning similar t o t h a t employed in the analysis of the tertbutyl hydroperoxide spect'rum it may be concluded t h a t the 11.4micron band is the only band t h a t could possibly be related to vibrations of the peroxide linkage in this molecule. Benzoyl Peroxide. Although, in vie\v of the above, this compound might have been espected to show an absorption band near 11.4 microns, no surh band is present. I n the henzoyl peroxide spectrum only three strong band3 occur in the longer wave-length region. T h a t at 14.3 microns is related to vibrations of the phenyl ring ( 4 , 8 6 )and t h a t a t 8 microns is very likely 24-26). If, therefore, related t,o a C-0 stretching vibration (4, a n y band is assignable a t the present time to a vibration of the peroxide linkage, the IO-micron band represents the most probable choice. Thus if the highly tentative assignment. of the 11.4micron band in di-teit-butyl peroxide and the 10-micron band in benzoyl peroxide t o a vibration of the 0-0 linkage is correct in both cases, the vibration frequency involved must be sensitive to changes in the structure of the two groups attached to the 0-0 linkage. Although the authors have not yet studied additional pure peroxides, infrared spectra of several additional peroxides

285 (some claimed to be pure and others of uncertain purity) are included in a commercially available catalog of spectra. Examination of these spectra reveals the interesting fact t h a t those containing aryl groups attached to the peroxide linkage (phthaloyl peroxide, p-chlorobenzoyl perouide, and benzaldehyde peroxide) all show a strong band near 10 microns in common with t h a t observed in benzoyl peroxide. Two acyl-type peroxides in this file show common absorption near 9.4 microns. Several others containing widely variable substituent groups show no coninion band t h a t could be attributed to a vibration of the peroxide linkage. From these observations, i t seems likely t h a t absorption due t o the peroxide linkage will vary widely with the nature of the attached groups. I n view of the existence of a common band (10 microns) in the four aryl peroxides mentioned above, however, i t may be possible to assign a fairly narrow wave-length range to this type of vibration in each of several classes of peroxides whose individual members are closely related. COMMENTS AND CONCLUSIONS

The potentialities of the infrared method as applied to a n y type of chemical problem cannot be assessed until spectra of appropriate pure compounds are available for reference. Thus the spectra presented here should be useful in connection with t h e possible application of this method to a wide variety of problems involving peroxides and hydroperoxides. When sufficient reference spectra of this type are available, i t may be possible to distinguish various types of peroside compounds involved in oxidation and other studies and to determine one type in the presence of others bp this physical method of analysis. ACKNOWLEDGMENT

The authors are grateful to R. E. Koos of the Eastern Regional Research Laboratory for the preparation and analysis of some of the reference compounds, and to C. E. Swift of the Southern Regional Research Laboratory for the methyl oleate hydroperoxide (purity 69%). LJTERATURE CITED

(1) Anderson, J. A., and Seyfried,

IT.D . ,

. ~ N A L . CHEM.,20,

998

(1948). (2) Atherton, D., and Hilditoh, T. P., J . ( ' h e m . SOC., 1944, 105. (3) Barnes, R. B., Gore, R. C.. Liddel, C., and Williams, V. 2., "In-

frared Spectroscopy," Sew York, Reinhold Publishing Corp., 1944. (4)

(5) (6) (7) (8)

Barnes, R. B., Gore, R. C., Stafford, R. W., and Williams, V. Z.. ANAL.CHEM.,20, 402 (1948). Bolland, J. L., and Hughes, H., J . Chem. Soc., 1949, 492. Cole, J. O., and Field, J. E., I n d . Eng. Chem., 39, 174 (1947). Criegee, R., Pilz, H., and Flygare, H., Ber., 72, 1799 (1939). Dugan, L. R . , Beadle, B. IT.,and Henick, A. S., J . Am. Oil Chemists' Soc., 26, 681 (1949).

(9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19)

Engler, C., Her., 33, 1090 (1900). I b i d . , p. 1097. I b i d . , p . 1109.

Engler, C., and Weissberg, J., I b i d . , 31, 3046 (1898). Engler, C., and Wild, IT., I b i d . , 30, 1669 (1897). Farmer, E. H., J . SOC.Chem. I n d . , 66, 86 (1947). Farmer, E. H., Trans. Faradat,' Soc.. 42, 228 (1946). Farmer, E. H., Trans. I n s t . Rubber I n d . , 21, 122 (1945). Farmer, E. H., and Sundralingam, A , J . Chem. Soc., 1942, 121. Farmer, E. H., and Sutton, D. A , , I b i d . , 1943, 119. Gamble, D. L., and Barnett, C. E . , I d . Eng. Chem., 32, 376

(1940). (20) Gunstone, F. D., and Hilditch, T. l'., J . Chem. SOC.,1945, 836. (21) Hock, H., and Schrader, O., ,~aturu,issenschuften,24, 159 (1936). (22) Honn, F. J.,Bezman, I. I., and Daubert, B. F., J . Am. Chem. Soc.. 71, 812 (1949). (23) Knight, H. B., and Swern, D., J . Am. Oil Chemists' Soc., 26, 366 (1949). (24) Shreve, 0. D., Heether, h l . R . , Knight, H. B., and Swern, D., ~ N A L CHEM., . 22, 1498(1950). (25) I b i d . , 23, 277 (1951). (26) Thompson, H. IT,, J . Chern. Soc., 1948, 328. RECEIVEDM a y 19, 1550. Presented before the Division of Organic ChemSOCIETY.Philadelistry a t the 177th Meeting of t h e AXERICANCHEMICAL phis, Pa. Report of a s t u d y i n which certain phases were carried on under the Research a n d Marketing Act of 1546. Paper VI1 in the series "Reactions of F a t t y Materials with Oxygen." Paper V I is ( 8 5 ) .