Chemistry and Physics o f Lipids 15 (1975) 53-85 © North-ttolland Publishing Company

THE PMR ANALYSIS OF NON-CONJUGATED AND ALKYNOIC

ALKENOIC

ACIDS AND ESTERS

D.J. FROST Unilever Research, Vlaardingen, The Netherlands and KD. GUNSTONE Department of Chemistry, The University, St. Andrews, Scotland Received December 12, 1974,

accepted May 2, 1975

The 220 Mltz PMR spectra of 143 non-conjugated ",dkenoic and alkynoic acids and esters are correlated so as to provide a method for the structural analysis of such compounds in general. The spectral data are explained in terms of long-range deshielding of the double bonds, triple bonds, acid and ester groups in the molecules, and parameters are derived to quantify the influence of these groups on the chemical shifts of methyl and methylene protons up to six carbon atoms distant along an alkyl chain. It is shown that, by the application of these parameters, 220 MHz PMR spectroscopy can be used to determine both the stereochemistry and position of double bonds, and the position of triple bonds, in the majority of fatty acids and esters. The 2to 9- and 13- to 17-cis- and trans-isomers of octadecenoic acid may be readily identified in this way, whilst for the octadecynoic acids all positional isomers may be characterized. Examples are also given of the structural analysis of several polyenoic compounds, including methyl cis-5, cis-8, eis-11, eis-14, cis-5, cis-8, cis-1 l, trans-14, and trans-5, cis-8, cis-I 1, cis-14-eicosatetraenoates, and methyl trans-5, cis-9, cis-12-octadecatrienoate.

I. Introduction Recent advances in PMR spectroscopy, namely the development of superconducting magnet spectrometers and shift reagents, have so greatly enhanced the potential of this technique that it may now lay claim to being the most powerful tool for the structural analysis of fatty acids. While it is often possible to provide a complete analysis of an unknown fatty acid on the basis of PMR measurement alone, such an approach has little practical merit since other techniques very readily provide information which may well obviate the need for some of'the more sophisticated PMR procedures. Information such as the chain length, number of double bonds and the presence of a trans double bond, may usually be obtained from combinations o f IR, GLC, UV or MS measurements, while Raman spectroscopy [1]

54

D.J. Frost, b]D. Gunstone, PMR of alkenoic and alkynoic acids

promises to provide a convenient method for the identification of triple bonds. The problem then resolves itself into establishing the position and configuration of the unsaturated bonds along an alkyl chain of given length. The extent of this problem is demonstrated by the fact that a C18 fatty acid, possessing one cis and one trans double bond, not conjugated with each other, may exist in 156 isomeric forms, together with another 26 isomers where one of the double bonds is at the terminal C17 position. Previous work [2] has shown how PMR spectroscopy at 220 MHz provides a method for determining the position ofcis double bonds in non-conjugated fatty acids, while other publications [3-8] have been concerned with the 60 and 100 MHz spectra of various series of alkenoic and alkynoic acids. In the present work we have extended our measurements at 220 MHz and attempted to establish a method for the structural analysis of all non-conjugated alkenoic and alkynoic acids and esters.

II. Experimental All measurements were carried out on solutions of carbon tetrachloride and the concentrations were usually 10 to 20% by volume. Chemical shifts were measured relative to internal TMS and are reported as a-values. All spectra were obtained on Varian HR 220 spectrometers operating at 13-20°C and 220 MHz. Since the measurements were carried out over a period of 3 years on two different spectrometers (sited at TNO, Delft and SRC, Runcorn) by many operators, the accuracy attained was less than optimal. In general the a-values are accurate to -+0.02 ppm, but in most cases the accuracy is better than -+0.005 ppm (i.e. about 1 Hz). Coupling constants are given to an accuracy of-+0.3 Hz. Spectral simulations were carried out on an 8K Varian 620i computer using the Simeq programme written by C.W.F. Kort of the University of Amsterdam. The normal minimal sample size for obtaining PMR spectra from fatty acids (of about C18 chain length) is about 2 - 5 mg, but the use of spectral accumulation and microtechniques reduces this to 25 50 btg. When necessary such accumulation was carried out using a dedicated Varian 620i computer. The acids and esters measured were all synthesized in our laboratories [6 16] with the exception of the five (n-4), ( n - l ) alkadiynes which were generously supplied by Kunau [I 7]. The compounds measured are given below, with Cn: indicating the length of the carbon chain, and the numbers after the colon the positions of the unsaturated functional groups. Cis and trans double bonds are indicated by "c" and "t", and a triple bond by "a".

Monoenes [6, 8 - 1 1 ] C6: 2c/t; 3c C8: 2c; 3t C9: 5c CIO: 3t; 4c; 5c; 6c Cll: 2t C12: 6c; 9c C13: 2t C14: 8c C16: 7c C18: 2c/t through to 16c/t; 5c, 18-carboxy; 17 C2O: l l c C22: llc; 13t.

D.J. Frost, I'ID. Gunstone, PMR of alkenoic and alkynoic acids

55

Dienes [7, 8, 12, 13] C9: 2t5c C12: 2t5c C14: 5c8c C18: 6c9c; 6c10c; 6 c l l c ; 6c12c; 7c12c; 8c12c; 8c15c; 9c12c; 9c12t; 9t12c; 9c15c; 9t12t; 10c15c; 11c15c. Polyenes [11, 14-16] C12: 2t5c9c C13: 2t5c10c C18: 5t9c12c; 6c9e12c; 8cl lcl4e; 9c12c15c; 3c6c9cl 2c; 5c8cl lc14c C19: 4c7c10c13c C20: 4e8cl lc14c; 5c8cl lc14c; 5c8el lcl4t; 5t8cl lc14c; 8cl lc14c18c C21: 5C8C1 lC14C. Monoynes [7] C6: 3a C9: 8a ClO: 6a C15: 8a C22:1 la; 19a.

C16: dioic acid

5a C18: 2a-17a;

dioic acid 5a

Poly-ynes [10, 11,14, 17] C8: 4a7a C9: 5a8a C10: 6a9a Cll : 7al0a C]2: 8al la C13: 8a12a; la4a7a alkyne C14: 5a8a C16: la4a7alOa alkyne C18: 9a12a C20: 8al lal 5a. Polyunsaturated containing both double and triple bonds [7, 11 ] C9: 2t5a C12: 2t5a; 2t5a9a C13: 2tSal Oa; 2t5a8a12a C14: 2t5a8a C18: 9a12t; 9a12c; 9t12a; 9cl 2a; 9cl 2a C20: 2t5a8al lal4a; 5a8al lal4t.

III. PMR spectral features of double and triple bonds Tile presence of alkenyl protons in a compound is readily signified by their characteristic PMR absorptions which are usually in the region 6 5 - 6 . The most significant parameter in PMR for the determination of the configuration of a double bond is the coupling constant between the alkenyl protons. Although this coupling is always larger for trans than for cis bonds, the actual values are dependent upon the electronegativity of the substituents.For alkyl substituents, Jtrans ~ 15 Hz and Jc/s ~ 10 Hz. In those cases where this coupling constant can be measured the configuration of tile double bond can be unambiguously determined, but for most isolated bonds this is not possible, the very small chemical shift difference between the alkenyl protons producing a complex absorption not susceptible to first-order analysis. However, even when there is no chemical shift difference between the alkenyl protons, it has been shown [3] that the different coupling constants involved cause isolated cis and trans bonds to give different and distinctive absorptions, independent of the magnitude field strength. This is illustrated in fig. 1 by spectra simulated from the parameters obtained by Schaumburg and Bernstein [18] for this system. The chemical shifts of these absorptions are also different, being 6 5.28 and 8 5.32 in CC14 solution for the cis and trans respectively. The absorptions of the methylene protons adjacent to a double bond are also dependent upon the configuration of the bond - for a cis bond this absorption (at 6 1.985) is noticeably sharper than for the trans isomer (at ~ 1.940). The simulated spectra in fig. 1 show that this is also a result of the different coupling constants involved. The n-electrons of a triple bond produce a well-known ring current effect [19] in a magnetic field which results in an alkynyl proton coining at a surprisingly high

56

D.J. Frost, F.D. Gunstone, PMR o f alkenoic and alkynoic acids

A5.32

I! IL

61935

CH2-:

"-H

65.28

J

61985 -10 6-, 70,':~ H\C ----c/H ~\\

Fig. 1. Computer simulated PMR signals for alkenyl and allylic protons in isolated cis and trans double bonds (using the parameters obtained by Schaumberg and Bernstein I18]). field (about 6 1.8) in the PMR spectrum in comparison with alkenyl protons. Such an alkynyl absorption is readily recognisable since it usually appears as a narrow triplet, coupled with the methylene adjacent to the triple bond by about 2.5 Hz, but the acidic nature cf the proton causes the chemical shift to vary appreciably (about 0.3 ppm) with concentration. Non-terminal triple bonds are more difficult to identify, especially at 60 and 100 MHz, since the absorptions from the adjacent methylene protons are often overlapped by those from similar groups next to double bonds or carboxyl functions. At 220 MHz, however, these absorptions are usually resolved. For compounds where the methylene protons on each side of the triple bond have different chemical shifts, the long-range coupling (about 2.3 Hz) between them becomes apparent, and often serves to identify the presence of such a group. For an isolated bond the adjacent methylene protons give a broad triplet (J about 6 Hz) absorption at 6 2.070.

IV. PMR substituent effects

The long-range deshielding effects of the functional groups in non-conjugated cis alkenoic acids and esters have been determined earlier [2]. The present work, encompassing 220 MHz measurements on 62 cis- and 26 t r a n s - u n s a t u r a t e d fatty acids and

Table 1 Deshielding effects of functional groups along an alkyl chain (in p p m in CCI 4 solution). Basic chemical shift for methyl group, 63 = 0.885. Basic chemical shift for m e t h y l e n e group, 6 z = 1.255. For methylene groups between two alkynyl or b e t w e e n an alkynyl and a carboxyl group 0.100 p p m m u s t be added. ~3

CH2

CH 3

CH2

3'

CH 3

CH2

6

E

CH 3

-COOH

1.035

0.360

0.095

0.055

0.030

0.005

-COOCH3

0.955

0.320

0.060

0.035

0.020

0.005

- C O O C 2 H5

0.940

0.320

0.060

-Ctt3

0.030

0.000

0.000

0.000

0.000

0.000

0.065

0.020

0.015

0.005

0.000

0.075

0.025

0.025

0.010

0.000

0.050

0.005

0.000

0.000

0.000

0.040

0.025

0.015

0.045

0.025

0.015

tt H -C=C

0.730

tt H H tt -C=CCH2 C=C

0.770

0.685

0.025

H

-C=C

0.685

0.735

H

H -C=CCH2 CH=CH

0.700

0.065 a

H

C~2R -C---CH

0.820 0.875

-C-=CCH2 C-=CR - C = C C H 2 C-=CH

0.865 0.875

a One measurement only.

0.835 -

0.160 0.220 0.210

0.210

0.130 0.145

0.070

58

D.J. Frost, F.D. Gunstone, PMR o f alkenoic and alkynoic acids

esters, together with 34 alkynyl compounds and 21 compounds containing two different unsaturated functions, has enabled these parameters to be estimated for further flmctional groups. The deshielding effects obtained from the total data are given in table 1. The parameters have been rounded off" to the nearest 0.005 ppm (i.e. about 1 Hz). Nevertheless, apart from the few exceptions discussed below, use of these parameters (i.e. basic chemical shift 82 or 83 plus appropriate deshielding effects of neighbouring functional groups) provided a calculated 6-value for all o f the many hundreds o f measured absorptions which agreed by better than 0.02 ppm with tile experimental value. Measurements on one spectrometer under more controlled conditions, and with frequent spectrometer calibration, provide an even better correlation, making differences of