Relationship between Unsaturation and Ultraviolet Absorption Spectra

Relationship between Unsaturation and Ultraviolet Absorption Spectra of Various Fats and Fatty Acids. R H. Barnes, I I. Rusoff, E S. Miller, and G O. ...
1 downloads 14 Views 252KB Size
Relationship between Unsaturation and the Ultraviolet Absorption Spectra of Various Fats and Fatty Acids R. H. BARNES, 1. I. RUSOFF, E. S. MILLER’,

AND

G. 0. BURR, University of Minnesota, Minneapolis, Minn.

The spectral absorption of several unsaturated fatty acids and natural fats have been measured from 2500 to PI00 Angstrom units. Data are presented that show a definite relationship between the degree of unsaturation and extinction coefficients at PI 00 Angstrom units. From the composition of natural fats it is possible to predict the extent of absorption at this wave length.

I

T IS well known that absorption by the carbon to carbon double bond, one of the most important chromophores (3, e ) ,

is modified by factors such as the cis- and trans-configuration,

weighting by substituent groups, the number of double bonds in the carbon chain, and their positions relative to each other ( 3 ) . The natural fatty acids, their esters, and isomers constitute a most important group of aliphatic compounds differing chiefly in the number and position of double bonds. In a recent review (2) it was pointed out that except for the saturated fatty acids (15) the absorption curves of the well-known members of t’his series of compounds and the naJura1 oils have usually not been extended below 2200 or 2300 A. (1, 4, 6, 7 , 9, 14, 19). The measurements that have been made in the longer wave lengths show such great irregularitmiej that it must be assumed that impurities with strong absorption band3 are affecting the r e d t s . The effect of increa~ingnumbers of unconjusated double bonds in simple hydrocarbons is so marked at 2100 A. (3,6) that it was decided t,o study the absorption spectra of the highly purified fatty acids. Although it was not possible t o extend the curves of tllc unsaturated acids to their maxim? (below 2000 A.) because uf the limit of the spectrograph (2100 A), there nevertheless was found a large and consistent effect of increming unsaturation which seemed of practical importance, since this wave length is within the rang? nf many .spectrographs now in use.

The curves in Figure 1 show the markei effect of unsaturation on spectral absorption below 2250 A. At 2100 8. the long-chain fatty acids have the following molecular extinction coefficients: stearic 60; oleic 180; methyl linolate 2500; methyl linolenate 10,000; and methyl arachidonate 14,500. I n other words, arachidonic acid with an iodine number approximately 4 times that of oleic acid has a spectral absorption a t 2100 d. which is roughly 80 times as great. The suggestion of an absorption band a t 2350 8. may be due to trace impurities. This is the region of maximum absorption by conjugated dienes and it is known that in the saponification, bromination, debromination, and distillation of highly unsaturated fatty acids some conjugation may take place (16, 17, 18). However, since the conjugated dienes have molecular extinction coefficients of 20,000 to 30,000 in this region ( 2 ) , there could not be more than a fraction of 1% present in any of these preparations. This would not measurably affect the absorption values of the highly unsaturated acids a t 2lOtL. although it is sufficientto throw the curves out of line a t 2300 A. The large a?d regular effect of unsaturation on light absorption at 2100 A. is contracted n-ith the smaller and irregular effect- at the longer Tzave lengths in Figure 2. It is clear that if nbwrption a t 2100 A. can be measured with sufficient accuracy the values can be used as constants in simultaneous equations for calculating the fatty acid composition of oils. This direct measurement may well be used instead of the one described by Kass ef al. (12) and later extended by Mitchell et al. (15), which depends upon the measurement of the conjugated linoleic and linolenic acids after saponification a t a high temperature. The chief disadvantage of the present method comes from the requirement that the measurements be made a t a wave length shorter than that reached by many instruments.

EXPERIMENTAL

he atxorptioii meawrements from 2100 to 2250

45

1

.3.

\\ere w’rc ma$e with a Gaertner Littrow spectrograph. From 2300 to 2500 A. absorption was measured with a photoelectric spectrophotometer similar to that described by Hogness et al. (8). The solvent employed was purified ethyl alcohol (commercial 95’;b alcohol freshly distilled from potassium hydroxide) for all qamples except stearic acid which was dissolved in ethyl ether {freshly opened anesthesia grade). The absorption values of the pure compounds are plotted as the logarithm of the molecular extinction coefficients, e, while the values for the oils are expressed as E:?&., where 1% means 1 gram in 100 cc. of solution. The values were calculated from Lambert’s and Beer’s law,

I VETHYL ARACHIDONATE

I

I

I

I

I

log 2 = ecl

Z

Measurements were made on samples of the highest purity obtainable. (The authors are indebted to J. P. Kass and J. Nichols for the preparation of these materials.) The stearic acid melted a t 69.6” in a capillary tube and had no measurable iodine number. Oleic acid was prepared by repeated recrystallization a t low temperature of the fatty acids of olive oil until a sample with iodine number (Wijs) of 88 was obtained. The chief impurity probably wa5 palmitic acid. The methyl esters of linoleic, linolenic, and arachidonic acids were made from the recrystallized polybromides by debromination in methyl alcohol. The iodine number of each preparation was within 2 units of the theoretical value.

-2500 -

2400

a00 WAVE

Fi y e 1.

2200

2100

LENGTH

Absorption Spectra of Five Fatty Acids with bifferent Numbers of Unconjugated Double Bonds

Deceased.

385

1

INDUSTRIAL AND ENGINEERING CHEMISTRY

386 Table

I.

Vol. 16, No. 6

Comparison of Calculated and Experimentally Determined Extinction Coefficients for Natural Fats Extinrtion Coeffirients. Iodine ThioryValue anogen (Wija) Value

Fat Coconut Olive Corn (Alasola)

Saturated Acids

Oleic Acid

Linoleio Acid

1%

'lorn.

Calculated

a t 2100 K. Measured

8.6 83.2

7.3 74.8

87.4 11.8

6.6 74.0

9.7

1.5

3.8 15.2

5.9 18.2

134.2

82.8

6.5

29.6

50.4

61.6

60.0

I n Figure 3 the curves of four plant fats of widely different composition are compared with those of t'he fatty acids. The extinction coefficients, E:lpm,r 1% a t 2100

Ai,are in the range that

would be expected from the composition of the fats. Iodine numbers (Wijs) and thiocyanogen numbers were determined for coconut oil, olive oil, a,nd corn oil. (The authors arc indebted to H. G. Loeb for these determinations,) From these analytical constants the composition of the three fats was r:alculated (Table I). The corrected value for the thiocyanogen number of linoleic acid as given by Kas5 e1 al. (1 1 ) was 3ubstituted in Equation 3 of the following.simultnneous equations described by .Jamieson (10):

z+y+s=l 86.01 1: 86.01 N

4- 173.20y

+

90.59 y

+o

+

I. N. o = 2'. N . =

N

+ 100 y + 2.1 s

=

2100

A.

Figwe 3. Absorption Spectra of Four Vegetable O i l s Compared with Their Constituent Fatty Acids Ir

(1)

and for stearic acid, 2.1. The E;?:.

(2) (3)

then compared with the value determined experimentally. The resulk (Table I) are Peen t o be of the right order of magnitude. Both c o c p u t oil and olive oil have very low spectral absorption a t 2100 .4. and are thus subject t o considerable error introduced by trace? of highly absorbing materials. .4 .imall error in the calculation of the linoleic acid content would also have a large effect. For example, if t'he coconut oil really contained 2.5% linoleic acid instead of the calculated 1.50j0,the E:?&. would be raised t o 5.6. However, corn oil absorption is of such magnitude that the effect of contamination is minimized and, consequently, it is possible t o show a close agreement bct'ween the calculated and experimental values for. this fat.

where z is the amount' of oleic acid glyceride; y, the linoleic acid glyceride; and s, the *saturatedacid glyceride present in the fat. After determining the composition of the fat, and converting to the free acids (95.570 of the glycerides) the value. were substituted in the equation:

7.1

2500 2450 2400 2350 2300 22% 2200 2150 WAVE LENGTH

E:?&.

The E;:&. et 2100 A. for oleic acid is 7.1 : for linoleic acid, 100;

calculated inthismannerwas

L I T E R A T U R E CITED

0

I

2

3

4

NUMBER OF COUELE BONDS Figure 2. Effect of Number of Unconjugated Double Bonds in Fatty Acids on Extinction Coefficient at Different Wave Lengths Experimental points are connected bv lines to aid in following values for same wave

length. Mixtures of two fatty acids differing by one double bond would give intermediate valuer on rbaighl line between them but an oil averaging one double bond b v having an equal amount of saturated acid and linoleic acid would not have absorption of oleic acid glyceride (1 double bond)

Bradley, T. F., and Richardson, D., IXD. ENG.CHEM.,32, 968 (1940). Burr, G . O., and Miller, E. S., Chem. Reu., 29, 119 (1941). Carr, E. P., and Stiicklen, H., "Proceedings of Seventh Summer Conference on Spectroscopy and I t s -%pplications", p. 128, New York, John Riley & Sons, 1940. Chevallier. A , , Guillat, J., and Chahre, P., Bull. m c . chim. h i d . , 15,358 (1933). Devlaux, E.. .I.pharm. Belg., 18, 131, 153 (1936). Dimroth, K., Angeto. Chem., 52, 54.5 (1939). Gillam, A . E., Heilbron, I. M.,Hilditch, T. P.. a n d Morton R . A , Biocheni. .I., 25, 30 (1931). Hogness, T. It., Zscheile. F. P., and Sidwell, A . E . , J . Phljs. Chem., 41, 379 (1937). Hnlst, L. d . N. van der, RBC. t r a v . chim., 54, 639, 644 (1935). Jamieson. G. S., "T'eget,ahle Fats and Oils", A.C.S. Monograph, p. 397, New York, Reinhold Publishing Corp.. 1943. Kass, J. P., Lunrlherq, W. O., and Burr, G . 0.. Oil ond S o a p , 17,50 (1940). Kass, J. P., Miller, E. S..Hendrickson. >I.. and Burr. G. 0.. Abstracts of papers of 99th Meeting, -kxEmc.m- CHEMIC.AI' SOCIETY, Cincinnati, Ohio, April, 1940. Ley, H . , and Arends. B., Z . physik. C'hena., B17, 177 (1932). Manecke. R..and Volbert, F., Ii'urbanseitvng, 32, 2287 (1927). Mitchell, J. H., Jr.. Kraybill. H. R.. and Zscheile, F. P., I s n . ENG.CHEM.,.\N.AL. ED., 15, 1 (1943). Moore, T., Biochem. J . , 31, 138 (1937). Norris, F. A., Rusoff, I. I., Miller, E. S., and Burl,. G . O., ,I. Bioi. Chem., 139, 199 (1941). Zbid., 147, 273 (1943). Ramart-Lucas, Biquard, and Gounfeldt, Compt. rend., 190, 1196 (1930). AIDEDh y grants from the Graduate School of the University of Ilinne8ota and from the Rockefeller Foundation. Assistance in the preparation of these materials was furnished by the personnel of Work Projects .4dministration, Offirial Project No. 165-1-71-124, Subproject N o . 331.