Isolation of Fatty Acids and Identification by Spectroscopic and Chemical Degradative Techniques
Sections C and D. Identification by spectroscopy and chemical degradation
C. Spectroscopy of Fatty Acids
A detailed discussion of the principles of spectroscopy is beyond the scope of this text, but it is hoped that sufficient information is given to point the reader in the right direction. Most spectroscopic methods are based on empirical collations of vast amounts of data obtained with model compounds of known structure, and in the interpretation of the spectra of unknowns, some knowledge of these data is required. It will only be possible to reproduce a few selected examples here. Applications of spectroscopic procedures in lipid analysis have been reviewed in some detail elsewhere [153,483]. The mass spectra of fatty acid derivatives are described in Chapter 7.
l. Infrared absorption and Raman spectroscopy
Infrared (IR) spectra are obtained when energy of light in the infrared region at a given frequency is absorbed by a molecule, thereby increasing the amplitude of the vibrations of specific bonds between atoms in the molecule. The most useful and conveniently measured region of the IR spectrum (limited by sodium chloride optics) is over a range of wavelengths of 2.5 to 15 μm (equivalent to wave numbers of 4000 to 667 cm-1).
Fatty acids may be subjected to IR spectroscopy in the free (unesterified) state, bound to glycerol or as the methyl ester derivatives, although an esterified form is to be preferred as a band due to the free carboxyl group between 10 and 11 μm may obscure a number of other important features in the spectra. Most information on the chemical nature of fatty acid derivatives can be obtained when they are in solution and Figure 6.10 illustrates the IR spectrum of soybean oil in carbon tetrachloride solution. The sharp band at 5.75 μm is due to the esterified carbonyl function, which is also responsible for a band at 8.6 μm. With free fatty acids, the first of these bands is displaced to 5.9 μm and there are also broad bands at 3.5 μm and 10.7 μm. cis-Double bonds give rise to small bands at 3.3 μm and 6.1 μm that are useful as diagnostic aids and are considered sufficiently distinct for use in quantitative estimations in some circumstances [43,59]. Most of the remaining bands are absorption frequencies of the hydrocarbon chain.
Figure 6.10. Infrared spectrum of soybean oil in carbon tetrachloride solution. The insert (dotted line) at 10.3 μm illustrates an absorption band due to a trans-double bond.
Many other functional groups give rise to characteristic absorption bands, which can be used to identify or to estimate the amount of a given acid. The most important of these is a sharp peak manifested by trans-double bonds at 10.3 μm (967 cm-1), and IR spectroscopy has long been one of the principal methods for estimating compounds with this structural feature in the fatty acid chain. In the recommended procedure , a calibration curve for the absorbance at 10.3 μm is drawn up from standard solutions of methyl elaidate in carbon disulfide with equivalent methyl stearate solutions run simultaneously to improve the background correction; the IR spectrum of the unknown is scanned in the region 9.5 to 11 μm and the trans-double bond content is derived by reference to the standard curve. If a trans-double bond is part of a conjugated system, the band maximum may be shifted to about 10.1 μm.
Free hydroxyl groups give rise to bands at 2.76 and 10.9 μm, epoxy groups produce a double band with maxima at 11.8 and 12.1 μm, allenes give a small band at 5.1 μm, a cyclopropene ring produces a small band at 9.9 μm and a cyclopropane ring produces two small bands at 3.25 and 9.8 μm. Some of the absorption frequencies caused by the more unusual functional groups have been discussed by Wolff and Miwa , and more detailed general information can be found in a number of reviews [153,266,490]. Characteristic frequencies are not altered markedly by the position of a group in the aliphatic chain, unless it is at either extremity of the molecule or immediately adjacent to another functional group. IR spectroscopy is therefore a valuable method for detecting the presence and, on occasion, for estimating the amount of certain functional groups in fatty acids, especially when these are in esterified form in a natural lipid mixture, but other spectroscopic or chemical procedures must be used to fix the exact position of the group in the alkyl chain.
Fourier-transform IR (FTIR) spectroscopy appears to have been little used by lipid analysts to date, but this situation will no doubt change soon. Two applications of GC-FTIR to some standard fatty acids and to cyclic monomers produced in heated sunflower oils have been published, and may serve to illustrate the power of the technique [589,931].
Structural features in unsaturated fatty acid methyl esters (in carbon tetrachloride solution) can give rise to distinctive bands in Raman spectroscopy [208-210]. For example, characteristic Raman bands are found for cis-double bonds (1656 cm-1), trans-double bonds (1670 cm-1) and triple bonds (2232 and 2291 cm-1); a terminal triple bond gives a single band at 2120 cm-1 (this group does not give a distinctive band in IR spectroscopy). Again, the position of a double bond does not affect the spectrum significantly unless it is at either extremity of the molecule.
2. Ultraviolet spectroscopy
The ultraviolet (UV) spectrum of a compound is generally measured over the range 220 to 400 nm. It is nowadays used principally to detect or to confirm the presence of fatty acids containing conjugated double bond systems in natural oils, or to observe chemical or enzymatic isomerization of fatty acid double bonds in which conjugated systems are formed. With such acids, series of broad bands of increasing intensity the greater the number of double bonds in the conjugated system, are found in the UV region at successively higher wavelengths. For example, with conjugated dienes, λmax is 232 nm (ε = 33,000), with conjugated trienes (e.g. α-eleostearic acid), λmax is at 270 nm (ε = 49,000) and with conjugated tetraenes (e.g. α-parinaric acid), λmax is at 302 nm (ε = 77,000). Different geometrical isomers have slightly different spectra; the greater the number of trans-double bonds, the higher the extinction coefficient but the shorter the wavelengths of the band maxima. Isolated cis-double bonds exhibit a specific absorbance at 206 nm, but with a relatively low extinction coefficient so this feature is of little use as a diagnostic aid. Conjugated triple bonds also affect the spectra. A peak of absorbance at 234 nm is characteristic for the cis,trans-conjugated double bond system produced by the action of lipoxygenase, a property that is utilised in the estimation of polyunsaturated fatty acids with methylene-interrupted cis,cis-double bonds in lipids  and more widely to detect and quantify lipid peroxidation in general [476,726]. A review of applications of UV spectroscopy in lipid analysis has been published .
3. Nuclear magnetic resonance spectroscopy
The nuclei of certain isotopes are continuously spinning with an angular momentum, which can give rise to an associated magnetic field. If a very powerful external magnetic field is applied to the nucleus and made to oscillate in the radio frequency range, the nucleus will resonate between different quantised energy levels at specific frequencies, absorbing some of the applied energy. Such very small changes in energy can be detected, amplified and displayed on a chart. The trace obtained, of the variation in the intensity of the resonance signal with increasing applied magnetic field, is the nuclear magnetic resonance (NMR) spectrum. In organic compounds, the isotope of hydrogen, 1H, displays this phenomenon whereas the main isotopes of carbon, oxygen and nitrogen do not, so the resonance frequencies of hydrogen atoms in molecules are those most often measured and the technique in this instance is referred to as proton NMR spectroscopy. One of the less naturally abundant isotopes of carbon, 13C, also exhibits the phenomenon but until comparatively recently, the inherent 6000-fold loss of sensitivity relative to proton NMR limited the value of the technique. Developments in instrumentation and data processing have made 13C NMR much more accessible and since all carbon atoms in organic compounds give distinctive signals, whether or not they are linked to protons, a great deal of structural information can be obtained from the spectra. As the power of modern instruments has increased, the size of sample needed has decreased to about 1 mg for an analysis time of 4 hours at 300 MHz (much more information on NMR spectroscopy is available on this site here...).
Compounds must be in solution for analysis and for proton NMR analysis, the solvent should preferably not contain the isotope 1H. Carbon tetrachloride is suitable for nonpolar lipids, but deuterochloroform may also be used and deuterated methanol has been added to this to effect solution of phospholipids. Chemical shifts are not measured in absolute units but are recorded as parts per million of the resonance magnetic field. Tetramethylsilane is added to the solvent as an internal standard, and in the conventional system it is given the arbitrary value zero on the so-called δ (delta) scale. (On the usual charts, values increase from right to left with the increasing strength of the magnetic field). In an alternative system (the τ (tau) scale), tetramethylsilane is given the value 10. To convert:
δ = 10 - τ
Fortunately, the same solvents can usually be used for both 1H and 13C NMR, so that both spectra can be obtained from a single preparation. With 13C NMR, chemical shifts are reported as upfield or downfield from tetramethylsilane. Whereas the spectral width of a proton NMR spectrum is about 10 ppm, that of a 13C spectrum is approximately 200 ppm, and this gives a corresponding improvement in the resolution attainable.
In recent years, NMR spectroscopy (especially proton NMR, but also 13C NMR to some extent) has been increasingly applied to the identification of lipid structures, and in particular to the detection, and often the location, of double bond systems in fatty acid chains (again in the form of methyl ester derivatives). The topic has been reviewed most recently by Pollard  and by Chapman and Goni . The NMR spectrum of methyl linoleate obtained at 60 MHz was illustrated and described previously by the author , and it has been published elsewhere . When the same spectrum is obtained on a 300 MHz instrument, each of the main features is of course much better resolved . Here, the proton NMR spectrum of methyl linolenate (9,12,15-octadecatrienoate) obtained at 300 MHz, is illustrated in Figure 6.11. There are eight main features: a multiplet at 5.33δ for the olefinic protons, a sharp singlet at 3.64δ for the protons on the methoxyl group, a triplet at 2.78δ for the methylene groups between the double bonds, a triplet at 2.28δ for the protons on the carbon atom adjacent to the carboxyl group, a multiplet at 2.06δ for the methylene groups on either side of the double bond system, a triplet at 1.6δ for the protons on carbon atom 3, a broad peak at 1.3δ for the protons in the chain (carbons 4 to 7) and a triplet at 0.98δ for the terminal methyl protons (this signal is at 0.88δ in the spectrum of methyl linoleate). Integration of the signals assists in confirming the assignments to particular protons.
Figure 6.11. Proton NMR spectrum of methyl linolenate in deuterochloroform solution at 300 MHz (tetramethylsilane as internal standard). Features in the spectrum and corresponding protons in the fatty acid are labelled "a to h". The spectrum was kindly obtained on a Bruker instrument by Professor F.D. Gunstone.
A wide range of unsaturated fatty acids have been subjected to proton NMR spectroscopy on 60 to 100 MHz instruments. They include the complete series of methyl cis-octadecenoates (2- to 17-18:1) , all the methylene-interrupted cis,cis-octadecadienoates , several cis,cis-octadecadienoates with more than one methylene group between the double bonds  and very many natural polyunsaturated fatty acids [320,399]. From studies with such compounds, it is now known how variations in the positions of double bonds affect the NMR spectrum of a long-chain fatty acid. For example, the 2- to 5- and 14- to 17-18:1 isomers can be distinguished by this technique, largely because of small changes in the signal associated with the olefinic protons, but cis- cannot be distinguished from the corresponding trans-isomers. The terminal methyl group of polyunsaturated fatty acids of the (n-3) series produces a well-separated triplet at a slightly lower field than that for the (n-6) family, and this feature has been utilised in the estimation of such compounds in natural mixtures . NMR spectra of some natural conjugated trienoic acids (α- and β-eleostearic acids) have been determined; the olefinic protons give rise to a complex multiplet at approximately 6.0δ . (Much of these data are tabulated on this site here...).
More informative spectra are obtained with more powerful instruments (220 to 300 MHz), as is seen in Figure 6.11, and data are available for a large number of different unsaturated fatty acids, including cis- and trans-monoenes, polyenes and acetylenic compounds [268,269]. Only the 10- and 11-isomers of the methyl octadecenoates cannot be distinguished, for example, while all the corresponding acetylenic compounds have unique features in their spectra. Some evidence as to the configuration of the double bond can be adduced, especially in the presence of lanthanide shift reagents, which also permit the location of centrally placed double bonds . The use of chemical shift reagents for structure determination with lipids has been reviewed .
Free hydroxyl groups in fatty acid chains give rise to two separate signals; that due to the -OH proton is indistinct, and its intensity and position may vary because of hydrogen bonding effects, but that due to the -CHO- proton at 3.6δ is quite characteristic. All the isomeric hydroxy stearates have been examined by NMR spectroscopy, and all can be distinguished from each other by this technique when quinoline is used as the solvent . In addition, much valuable information on the structures of hydroxy acids can be obtained if lanthanide shift reagents are added [977,978]. When the hydroxy ester is acetylated, the acetoxy protons give rise to a sharp signal at 2.1δ. Keto groups influence the alpha-methylene protons, which produce a signal similar to that for protons adjacent to a carboxyl group. Many other functional groups give rise to distinctive signals: epoxide ring protons at 2.8δ , cyclopropene ring protons at 0.8δ  (suggested as a means of quantification ), cyclopropane ring protons at 0.6 and -0.3δ , and olefinic protons in a cyclopentene ring at 5.7δ .
Methyl branches on aliphatic chains do not give signals that are helpful in locating their positions, unless the branch is immediately adjacent to either end of the molecule; iso-compounds can therefore be recognised, but anteiso-compounds cannot be distinguished from fatty acids having the methyl group in a central position .
There has recently been great interest in natural-abundance 13C NMR spectroscopy of fatty acids, as features in the spectra can be assigned to virtually every carbon atom in fatty acids differing widely in structure. Model compounds are used to obtain chemical shift assignments, and these are essentially additive for particular functional groups in an aliphatic chain. Most effort has gone into an understanding of the spectra of conventional polyunsaturated fatty acids [141,142,332,448], but much work has also been done with saturated fatty acids [84,448,930], and those containing conjugated double bonds [933,937,939], acetylenic bonds [141,142,331], hydroxyl groups [748,931] and oxo moieties . (Again, much of these data are tabulated on this site here...).
The 13C spectrum of methyl linolenate is illustrated in Figure 6.12, and the detailed assignments of the signals to particular carbon atoms are listed in the legend. It can be seen that nearly every carbon atom has a distinct signal, in essence only those from carbons 4 to 6 overlapping.
Figure 6.12. 13C NMR spectrum of methyl linolenate in deuterochloroform solution at 75.5 MHz (tetramethylsilane as internal standard). Chemical shifts: C-1, 174.16; C-16, 131.92; C-9, 130.24; C-12/13, 128.29; C-10, 127.8; C-15, 127.18; -O-CH3, 51.36; C-2, 34.11; C-7, 29.63; C-4 to 6, 29.21 to 29.18; C-8, 27.25; C-11, 25.68; C-14, 25.58; C-3, 24.99; C-17, 20.60; C-18, 14.29. The spectrum was kindly obtained on a Bruker instrument by Professor F.D. Gunstone.
>cis- and trans-Isomers of unsaturated fatty acids are readily distinguished, an effect that is enhanced by lanthanide shift reagents , and 13C NMR spectroscopy has been suggested as a means of quantification of trans-unsaturation in lipid mixtures [708,776]. Unfortunately, 13C NMR spectroscopy is relatively insensitive for many purposes, requiring approximately 1 to 10 mg of fatty acid for a usable spectrum, so it is only likely to be used to complement mass spectrometric analysis with unknowns, for example.
D. Identification of Fatty Acids by Chemical Degradative Procedures
l. Chain length determination
One of the first steps in the determination of the structure of an unsaturated fatty acid is to establish its chain length, and this can be ascertained simply by catalytic hydrogenation to form the saturated compound, which is then identified positively by gas chromatography. A procedure suitable for the purpose is described in Chapter 4. Ideally, the reaction should be carried out on the single fatty acid of interest, but valid results can be obtained in some circumstances with natural mixtures or with fractions isolated by silver ion chromatography, for example. A procedure suitable for determining the chain length of oxygenated fatty acids is described in Section D.3 below.
2. Location of double bonds in fatty acid chains
The positions of double bonds in alkyl chains are generally determined by oxidative fission across the double bond, followed by gas chromatographic identification of the products. Lipid analysts have largely accepted two procedures as suitable for the purpose: oxidation with permanganate-periodate reagent (frequently termed "von Rudloff oxidation"), and ozonolysis followed by oxidative or reductive cleavage of the ozonide. The former method yields mono- and dibasic acids as the products, while the latter can give either these or aldehydes and aldehydo-esters. Both procedures have been reviewed in some detail elsewhere . Ozonolysis techniques have the advantages that over-oxidation and spurious by-product formation are negligible, recovery of short-chain fragments is less of a problem, and it can be used when other functional groups (e.g. hydroxy- or epoxy-) are present. It is of course necessary to have equipment to generate ozone (although this is probably a worthwhile investment if many samples must be analysed). The permanganate-periodate procedure, on the other hand, uses readily available inexpensive chemicals and over-oxidation is minimal. In the author's opinion, the drawbacks of this method are sometimes painted in too dark a light. It is especially valuable when only an occasional sample must be analysed.
(i) Permanganate-periodate oxidation: In this procedure, the methyl ester of the unsaturated fatty acid in tert-butanol solution is oxidized by a solution containing a small amount of potassium permanganate together with a larger amount of sodium metaperiodate, which continuously regenerates the permanganate as it is reduced, while the whole is buffered by a solution of potassium carbonate. When the reaction is complete, the solution is acidified, excess oxidant is destroyed by addition of sodium bisulfite or preferably by passing sulfur dioxide into the solution, and the products are extracted thoroughly with diethyl ether before being methylated for GC analysis. Very little over-oxidation occurs with this reagent if the reaction is carried out properly, but it is not easy to achieve quantitative isolation of short-chain mono- and dibasic acids or half-esters of these for GC analysis. The problem has been partially resolved by injecting the free short-chain fatty acids directly on to a GC column of Porapak™ Q , although Carbowax™ 20M-terephthalic acid might be a better choice of liquid phase, or by pyrolysing the tetramethylammonium salts of the acids, thus converting them to methyl esters, in the heated injection port of the gas chromatograph . Where the compound to be oxidised consists of more than one positional isomer, only the longer-chain fragments are obtained in reproducible yields, and wherever possible, these alone should be considered when determining the amounts of individual positional isomers . The recommended procedure is as follows (only the highest purity reagents should be used) [958,959].
A stock oxidant solution of sodium metaperiodate (2.09 g) and potassium permanganate (0.04 g) in water (100 mL) is prepared. This solution (1 mL) together with potassium carbonate solution (1 mL; 2.5 g/L) is added to the monoenoic ester (1 mg) in tert-butanol (1 mL) in a test tube, and the mixture is shaken thoroughly at room temperature for 1 hour. At the end of this time, the solution is acidified with one drop of concentrated sulfuric acid, and excess oxidant is destroyed by passing sulfur dioxide into the solution, which is then extracted thoroughly with diethyl ether (3 × 4 mL). The organic layer is dried over sodium sulfate, before the solvent is removed carefully on a rotary evaporator or in a stream of nitrogen at room temperature. The products are methylated for GC analysis, preferably by reaction with diazomethane, freshly prepared by the procedure of Schlenk and Gellerman  (see Chapter 4).
The procedure can be scaled up for polyunsaturated fatty acids, but the proportion of water to tert-butanol should be kept as close to 2:1 (v/v) as possible. Malonic acid, formed by oxidation of methylene-interrupted double bonds, is oxidised further and is not detected. Ambiguity may result when polyunsaturated fatty acids with more than one methylene group between the double bonds are oxidised, as it is then not possible to state which dibasic fragment contained the original carboxyl group. However, this difficulty can be resolved by reducing the carboxyl group to an alcohol prior to the analysis .
(ii) Ozonolysis and reductive or oxidative cleavage: Ozone attacks olefins rapidly and quantitatively to form ozonides with no over-oxidation and very few side reactions, if the reaction is carried out at low temperature in an inert solvent such as pentane. The ozonide can be cleaved reductively to aldehydes and aldehydo-esters by a number of reagents of which the most convenient are dimethyl sulfide in methanol  and tetracyanoethylene , although triphenyl phosphine has been more widely used . Reduction of the ozonides with sodium borohydride to yield products which are alcohols or alcohol esters is reported to be less subject to interference in GC analysis with packed columns . The following procedure is generally satisfactory .
A solution of ozone in pentane is prepared by bubbling oxygen containing ozone through purified pentane at −70°C until a blue colour is obtained. The unsaturated ester (1 mg) is dissolved in pentane (1 mL) and cooled to 0°C, when the ozone solution is added drop-wise until the blue colour persists. After 1 min, the reagents are removed in a stream of nitrogen. Methanol (0.3 mL) then dimethyl sulfide (0.5 mL), both pre-cooled to −70°C, are added and the mixture is maintained at this temperature for 20 min, before it is allowed to warm up to room temperature. The excess reagents can be removed in a gentle stream of nitrogen, provided that the shortest-chain aldehydes are absent, and the products are dissolved in pentane or hexane for injection into the gas chromatographic column for analysis.
With larger amounts of ester, it may be necessary to add more methanol to the reaction mixture to assist the reaction. There may be difficulties in recovering short-chain aldehydes or aldehyde esters quantitatively for analysis. Dimethyl sulfoxide is the other product of the reaction, but does not interfere with the GC analysis as it tends to elute ahead of dodecanal.
Oxidative ozonolysis became a more attractive technique when Ackman and colleagues [12,826] demonstrated that ozonolysis could be carried out in a solution of 7% boron trifluoride-methanol reagent in such a way that acid ozonolysis products were formed, with a minimum of secondary reaction, and they were then methylated in a one-pot reaction. The reaction is carried out as follows:
A stream of 2-4% ozone in oxygen (120 mL/min) is bubbled into a solution of the fatty acid or its ester (1 mg) in 7% boron trifluoride-methanol (2 mL) for 1 min. The tube is sealed tightly and heated at 100°C for 1 hour, then the tube is cooled, water (2 mL) is added, and the methyl esters of the acidic fragments are extracted with two portions of methylcyclohexane.
The products of the two types of reaction can be identified on almost any GC column or phase; di-functional compounds are more rapidly eluted on silicone liquid phases, but are less likely to be confused with mono-functional compounds on polyester phases. With both types of liquid phase, temperature-programmed analysis is generally necessary if all the fragments must be determined. Dicarboxylic acid standards are readily available for comparison, but aldehydo-esters are not and it may be necessary to make up a suitable standard by ozonolysis of monoenoic esters of known structure that are commercially available, e.g. methyl petroselinate, methyl oleate and methyl vaccenate. A GC trace of the ozonolysis products of such a mixture on an EGS column is illustrated in Figure 6.13.
Figure 6.13. GC recorder trace of the products of reductive ozonolysis from methyl esters of 18:1(n-12), 18:1(n-9) and 18:1(n-7) fatty acids on a packed column of EGS . (Reproduced by kind permission of the authors and of the Journal of Lipid Research, and redrawn from the original paper.)
With oxidative ozonolysis, standards are freely available for identification purposes. The procedure has been applied successfully to monoenoic fatty acids  and, with some modification, to polyunsaturated fatty acids with other than methylene-interrupted double bonds [26,756].
(iii) Special procedures for configurational isomers or for polyunsaturated fatty acids: The procedures described above give optimum results with mono- and dienoic fatty acids. When the structures of polyunsaturated fatty acids must be determined, especially when they contain double bonds of both the cis- and trans-configurations or are part of conjugated double bond systems, more positive identification can be obtained by partially reducing the compounds prior to oxidation . Hydrazine, a reagent that does not cause any double bond migration or stereomutation, is used for the purpose, under conditions such that a high proportion of monoenoic compounds are formed; isomers are then found with double bonds in each of the positions in which they were present in the original polyunsaturated fatty acid. cis- and trans-Monoenes are separated by silver ion chromatography (see above), and their structures are determined separately by one of the above methods so that the original compound is identified fully. Hydrazine reduction, which is better performed on the free acid than on the ester, is carried out as follows (c.f. Chapter 4):
The free fatty acid is heated in air at 35°C with 100 volumes of 10% hydrazine hydrate in methanol for a predetermined time (typically 1.5 to 2 hours), found by trial and error with a standard polyunsaturated acid, such that there is an approximately 50% yield of monoenes. Excess methanolic hydrogen chloride (6% w/w) is then added to stop the reaction, and the mixture is refluxed for 2 hours to convert the acids to the methyl esters, which are recovered for further study as described earlier.
A number of variations on the procedure exist and have been reviewed by Privett . For example, the positions of double bonds in monoenes containing both cis- and trans-isomers can be determined by conversion to epoxides, of which the configurational isomers are readily separable, and cleavage of these with periodic acid . Similarly, partial oxymercuration of the double bonds followed by mass spectrometric identification of the products has been applied to the analysis of polyunsaturated fatty acids .
3. Location of other functional groups in fatty acids
Spectroscopic aids to the recognition and location of functional groups in fatty acid chains are discussed above and in Chapter 7 (mass spectrometry). Chemical methods for characterization are also of value, sometimes as an aid to mass spectrometric identification.
Isolated triple bonds in fatty acids are not easy to recognise, as they do not exhibit particularly distinctive features when examined by any of the spectroscopic techniques other than Raman spectroscopy, but a specific TLC spray consisting of 4-(4'-nitrobenzyl)-pyridine (5%) in acetone, that gives a violet colour with acetylenes, has been described . The position of a triple bond can be located by mass spectrometry (Chapter 7), and triple bonds are readily cleaved by the permanganate-periodate reagent. Ozone reacts with triple bonds, although more slowly than with double bonds, and the products of reductive cleavage of the ozonide are mono- and dibasic acids rather than aldehydes and aldehydo-esters, so that double and triple bonds in a single fatty acid can be differentiated . In addition, double bonds are hydroxylated by peracids while triple bonds remain unchanged .
Allenic groups have distinctive IR and NMR spectra, and all the natural fatty acids containing this functional group exhibit a marked optical activity. The position of the group in the fatty acid chain can be determined by partial reduction and oxidation of the monoene fragments as described above .
A number of spectroscopic procedures are of value for the detection and location of oxygenated functional groups, such as keto, hydroxyl or epoxyl, in fatty acid chains. In addition, several chemical techniques are available, and for example, the presence of hydroxyl groups can be confirmed by GC analysis before and after the preparation of volatile derivatives, such as the acetates or trimethylsilyl ethers, as described in Chapters 4 and 5. The first step in identifying an acid of this type is to determine its chain length. To accomplish this, the acid is first hydrogenated to eliminate any multiple bonds before free hydroxyl or epoxyl groups are converted to iodides by the action of iodine and red phosphorus. The iodine atom in the aliphatic chain is then removed by hydrogenolysis with zinc and hydrochloric acid in methanol, when the resulting saturated straight-chain compound is identified by GC . Keto groups should be reduced to hydroxyl groups by the action of sodium borohydride, prior to analysis by the above procedure . A chemical procedure is available for determining the position of the hydroxyl group , but mass spectrometry would be favoured by most analysts.
Epoxyl groups are cleaved directly with periodic acid in halogenated solvents  or in diethyl ether , and the position of the ring is established by GC analysis of the products. Furanoid fatty acids, separated on TLC plates, can be detected by spraying with a 2% solution of tetracyanoethylene in acetone, which reacts to give blue spots on a yellow background .
The chemistry of cyclopropane and cyclopropene ring-containing fatty acids has been reviewed elsewhere . The presence of such functional groups can be detected by a number of spectroscopic procedures and also by various chemical techniques. For example, cyclopropene rings give a pink coloration with carbon disulfide (the Halphen test) and a brown coloration with silver nitrate. The ring is disrupted by ozonolysis or permanganate-periodate oxidation, and a β-diketo compound is formed, which can be identified by mass spectrometry . Similarly, the reaction with silver nitrate can be used with GC to quantify individual cyclopropene fatty acids in seed oils (see Chapter 5) and in conjunction with mass spectrometry for identification purposes (see Chapter 7).
Cyclopropane fatty acids behave as normal saturated fatty acids on silver ion chromatography, but react with bromine and so can be distinguished by this means . They form methoxy-derivatives with boron trifluoride-methanol reagent , and they are converted to methyl-branched fatty acids by vigorous catalytic hydrogenation . Both types of derivative can be characterized by mass spectrometry.
Methyl branches in fatty acids are not easily located by chemical means as such compounds are comparatively inert to most chemical reagents. However, their positions can be determined chemically by oxidising the fatty acids vigorously with acidic potassium permanganate; homologous series of normal and branched-chain acids are formed together with a neutral keto compound that identifies the point of branching [639,669]. Both the acidic and keto products can be isolated and identified by GC. These and related procedures have been reviewed by Polgar .
The following abbreviations are employed at various points in the text of these chapters:
amu, atomic mass units; BDMS, tert-butyldimethylsilyl; BHT, 2,6-di-tert-butyl-p-cresol; CI, chemical ionisation; DNP, dinitrophenyl; ECL, equivalent chain length; ECN, equivalent carbon number; EI, electron-impact ionisation; FCL, fractional chain length; GC, gas chromatography; GLC, gas-liquid chromatography; HPLC, high-performance liquid chromatography; IR, infrared; MS, mass spectrometry; NMR, nuclear magnetic resonance; PAF, platelet-activating factor; ODS, octadecylsilyl; TLC, thin-layer chromatography; TMS, trimethylsilyl; UV, ultraviolet.
This document is part of the book Gas Chromatography and Lipids by William W. Christie and published in 1989 by P.J. Barnes & Associates (The Oily Press Ltd), who retain the copyright.
Return to Part 1 of this Chapter.
Updated: July 5, 2011