- Email: [email protected]
Gas chromatography of esters
XIII’;. INTERRELATIONSI4IP OF EQUIVALENT CHAIN AND RETENTION INDEX VALUES OF FATTY ESTERS
LENGTH
(ECL)
J_RASHES CSIRO,
Div&ion of Am-ma! Production, P-O_ Box 239, Btocktowtr, N.S. W.
J. EC.HAKEN” Depmnnart of Polytr~er (AKrzr&j
Science,
Universfry of New
South
Wales, P-0.
2148(AK&~G&z~
Box
1, RezMgm~
2033
and s. c. MrJxs
CSIRO,LkXdm
of Anhal
Proaktion,
P-0. &xc
239,Blacktown,
N.S. W. 2148 (Aurtrahb)
(Received August23rd,1979)
SUMMARY
The interrelationship of equivalent chain Ie~gth and retention index values of straight- and branched-chain fatty esters is demonstrated on a typical polyester phase (DEGS), a phenyl-substituted poIy.siloxane (GV-17) and on two Iclw-polarity, thermally stable hydrocarbons. Hydrogenated Apiezon M and the synthetic hydrocarbon G&&,6 were used, and have lower polarities than the simple Apiezon greases that have been used with fatty esters. These stationary phases have polarities comparable to that of squakne and retention increments of fatty esters may be considered comparatively with those of simple esters which have been compared with squahme.
INTRODUCXON A variety of relationships have been used to specify retention behaviour and the interrelation of nine of the commoner structure retention parameters was shown ma+&ematicaIly by Guerin and Banks*. The growing use of automated gas chromatographic equipment with accompanying computer processing of retention data will continue to eJ.iminatetedious measurementsfrom charts and provide data virtually simultaneously with the elution of compounds. The determination of mathematical dead-time has recently been reviewed2 and this, together with the interpolation of
l
* FartXII: see3.K. E&ken,A. NguyenandM. S_Wkuvright,A C%ro~ogr., 178(1979)471. - To whautcerrespandeuce shouldbe &dressed.
the results for a compound on a graph prepared with homologous compounds, e.g., n-afkanes, to provide retention indices are simple procedures that can be rapidly handled by‘ on-line processors_ Retention indices3 are established by the relationship - log fm) + 100.x log
fR
--%tRN
(1)
where = tE?s ZRN = ta
retentionof unknown n-substrate; retentionof n-alkane with N carbon atoms; retentionof n-alkane with N + 1 carbon atoms; integerequal to carbon number of tnr.
The slope (b) of the semi-logarithmic plot of carbon number V~CYZCS retention of n-alkanesis common to many of the schemesconsideredby Guerin and Bank9 and is determinedfrom two points on a graph prepared with n-alkanes, as shown in eqn. 2, although in a computer plot the slope is determined statistically by considering all of the n-alkanes used in the calibration:
where
ami
t R(N+xl reteniionof n-alkaneswith N and N + x carbon atoms, tRCn, = res~tively The retentionindex is thereforegven by
The methyieneunit (MU) nomenclaturepropoc-I by VandenHeuvel et aL4 is essentiallyidenticalwith the retentionindex system.A eAibrationusingn-docosane and n-tetracosanewas used and the meth$ne unit multigr:-d by 100 gives the retention index. The proposalbased on a two-point caiibrationline is less satisfactorythan the cormalJydeterminedmulti-pointn-alkane calibrationline. The difhcuhyin selectinga referenceseriessuitablefor all polaritiesof solute and solvent is well known. Lorenx and Roger9 consideredthe effect of sorption phenomena on retentionindices and found significanterrors when the solute and solvent polaritieswere widely variant. These Endings were similarto those of VandenHeuvei and Horn&$, who recommendedthat the polaritiesof the two species should be similar. The use of secondary standards with n-alkanes has heen suggested, and Hawkes’ discussed the dif&ulty of using primaryn-alkanolsas proposedby Groblefl, who had critic&d the retentionindex system using polar stationaryphases owing to the rapid ehrtion of hydrocarbonsand indicatedthat compounds with very dS ferent boiling points cannot be chromatographedunder strictlyidenticalconditions. n-Propyl ethers were recommended as standards by Hawkes’, with ketones and
GCOFESTEECLX5
299
aldehydes as alternatives. Z-Alkanones had earlier been suggested as standards by Dymond and Kilburn9 and the merits of using ketones as secondary reference standards were later detailed by Ackmanzo. The di&ulty or impossibility of providing a universally acceptable reference series has been demonstrated by a study of the behaviour of compounds ofmcderately polar character, e.g., ;r-a&y1esters. With dimethyl polysiloxane, the solvent of lowest polarity included, it is apparent that methyl acetate (b-p. 57.1”) has a retention index of 569, Le., identical with that of n-pentane (b-p. 36.2”), while :r-hexyl acetate (b-p. 169.2”) has a retention index of 988, Le., almost identical with that of n-decane (b-p. 174.0”), so that for a given retention index the esters have higher boiling points than the comparable n-alkanes. With the introduction of two phenyl groups, i.e., using DC-710, the situation is reversed, with methyl acetate (I = 609) being comparable to n-hexane (b-p. 69.0°) and n-hexyl acetate (I = 1092) being comparable to n-undecane(b.p. 196”).With weaker acceptor phases,Lc. Z-60, these retention indices are increased to 737 and 1221,respectively, the former being equivalent to an n-alkane with a boiling point of approximately 220”. With a more polar stationary phase (Silar SCP), retention indices of 894 and 1313 are obtained, which are equivalent to n-nonane (b-p. 151°) and n-tridecane (b-p. 235”), respectively. The great differences in boiling points at each end of the scale reflect the differences in the slopes for the n-alkanes and simple esters on the various stationary phases. As alternative secondary reference standards, acetate esters and symmetrical esters with equal acid and alcohol chain lengths have been suggested”. The use of esters as standards has found considerable acceptance with fatty esters In the form of the equivalent chain length (ECL) relationship develop.ti by Woodford and Van GenF and Miwa et QZ_‘~.The relationship Is equivalent to tZ:e retention index system, results for an unknown ester being interpolated on graph prepared using homologous long-chain methyl esters. The relationship is shown in eqns. 4 and 5 and is identical with eqns. 1 and 3 except for the reference series used. EcL=
logt,--ogt,-i-N log
fR
-
1%
fRN
(4)
The quantities are as shown above but refer to methyl fatty esters, while the calibration line is usually constructed using even-bon-number esters. With computer processing of chromatographic data it is possible to provide retention data relative to multiple reference series and with the retention jndex and ECL representations a simple consideration of the slopes of plots for standard homologous series is necessary. In this paper we consider the retention of straight- and branched-chain fatty esters with the data expressed in. terms of both retention indices and ECL values. The data were determined on a typical polyester phase (DEGS) and on two lowpolarity, thermally stable hydrocarbons. Hydrogenated Apiezon M and the synthetic hydrocarbon C&I, were used, and have lower polarities than the simple Apiezon greases that have been used with fatty esters. These stationary phases are comparable
11.39 13,37 9.40 15.37 19,34 17.36
Clb Cl0 GO Cl2 C,,
gl Cll d: Cl2 Go
nAlknncs
16.00 18.00 20.00 22.00 7.43
Cl1
Cl,
Cl4
10.00 12,ao 14.00
C!,o
Methyl
~
13,37 15.36 19,35 17,36
1400 tm 1800 2200 2000
0,176 0,371 0.083 0,789 3,520 1,670
16.00 18.00 20.00 22.00
1859 2059 2257 2454 :tz
Z 4i79 9,499 0.040
lO.cIO 12.otl 14.00
1256 1457 1659
ECL
c87&b
Oiii0,221 0,471
JYyilrogenatcfl Aplezon M (230’) --. J??CL RR I
OSkCS
-_.
nmlrcr
Conlpounrls Ccrhon
-
1800 2200 2000
0.762 4.184 1.791
1860 2061 2262 2460
1.000 2.350 5,530 13.109
1600
1266 1468 1659
0,075 0,177 0,424
0,342
I
-
RR
(,zao’)
VI
7
__-
9,89 7.96 11,87 13,86 15,85
2015 2215 2414
1.000 2.170 4.550
---
1400 1200 1800 1600 2ov0
1410 1613 1814
0,091 0,211 0.465
0.093 0,044 0.436 0.201 0,943
I
RR
(2OOO)
16,OO 18,OO 20.00
10,oO 12.00 14.00
ECL
0
0,205 0.585 0.343
1,756 3,021 0,995
18.03 20.15 15,98
1,000 1.703 2.907 4.963
0.202 0.347 0.585
RR
IO,05 13,99 11,98
16,OO l&O0 20,oo 22,OQ
1om 12,00 14.00
ECL
DEGS (ZOO’)
Iri
2600 2800 2400
E
gn 1”1
I & p
pl . P
1800 2200 2000
z: 2796 2993
1794 2003 2201
I
EQUIVALENT CHAIN LENOTHS @CL), ROLATIVERETENTIONS(RR) AND RETENTIONINDICES (I) OF METHYL ALWNOATES AND /I-ALKANES
TABLE I
GC OF ESTEES
XIII.
301
in behaviourto squalaneand retentionincrementsof fatty estersmay be considered comparativelywitk tkose of simple esterswhick have been comparedwith squalane.
Gas chronmogrraphy
A Hewlett-Fackard57OOAgas ckromatograpk,fitted with dual flame-ionization detectors, was used. Nitrogen (30 ml/nun) was used as the carrier gas, and hydrogen (30 ml/mm) and air (200 ml/min) were suppliedto tke detector.All results were determinedisothermally at 200”. Four stationary phases were used, viz., in order of increasing polarity, kydrogenatedApiezon M (refs. 14 and 15), C&i,,, (ref. lo), OV-17 and DEGS, all coated to tke extent of 10% on Chromosorb W AW DMCS (SO-100 mesh) and packed into 2 m x 3 mm 0-D. stainless-steelcolumns. Tke gas chromatogmphwas connectedto a Hewlett-Packard3370B integrator with facilitiesfor paper-tape output on a teletype (STC, ASR-33). The ckromatograph was also fitted witk a Hewlett-Packard7671A auto-sampler. Preparation of fatty acid methyl esters from wool WQX Wool from Merino sheep was extractedwith light petroleum (b-p. 40-&Y), using a Eoxklet apparatus. Small portions of wax were saponitied with akzoholic sodium hydroxide solution. Unsaponifiable components were removed with light petroleum, the solution was then acid&d and fatty acids together with hydroxy fatty acids were extracted with light petroleum. Fatty acids were separated from hydroxy fatty acids by thin-layer chromatographyon silica gel G using n-hexanediethyl ether-acetic acid (70:30:1). The bands were detected under UV ligkt after sprayingwitk 0.1% etkanolic 2,7-dickloro~uoresceinand identi&d using standards (Applied Science Labs., State College, Pa., U.S.A.) run on the same plate. Fatty acid bands were scraped from the plate and extracted with dietkyl etker. Methyl esterswere preparedby adding diazomethanesolution. The data were processed by means of a PDP 1l/l5 computer fitted witk a high-speedpaper-tapereaderand disc pack (RKOS), dead-timesbeing calculatedby the procedureof Guardino et al.“. ECL values were computed from tke modi&i regressionequation first obtamed by tke procedure of Guardino eb al. from a homologous series of straigktchain faw acid methyl esters. In order to allow for tke normal variationstkat occur with net retention times with time (as up to 36 samples could be automatically processed), all experimentaldata were converted into retentionsrelativeto methyl paknitate (Cl6). Tke net retention time of methyl paknitate was constantly remeasuredin eack sample, as it eitheroccurrednaturally or, if not, was added as an internal standard. Retention indices were similarly determinedusing homologous IHlhane&
RESULTS
AND
DISCUSSION
Retention data for methyl alkanoates and n-alkanes on tke four stationary phases are shown in Table I as retentionsrelativeto metkyl palmitate, EC. dues
Conrpolmd’
9.99 10.72 11.35 11.62 11,99 12,33 12.71 13.35 13.62 14.00 14.71 15.00 15.33 15.62 16,OO 16,31 16.71 16,99
9.63
0,290 0.369 0.408 0.471 0.615 0,618 0.777 0.867 1.00 1,119 1.304 1.452
1217 1252 1325 1390 1417 1454 1487 1527 1590 1618 lG56 1727 1756 1789 1818 1856 1886 1927 1955
0,091 0.104 0,137 0.174 0.192 0.221 0.243 0.360 0.424 0,572 0,849 1,000 1,345
12,69 13,61 13,99 1469 1562 16,OO 16,69
1928 lG.69 17.00
1820 15.58 1858 16.00
1528 12,69 12,99 1619 13,58 1658 13.99 1727 14.70 1752 15.00
1425 11.57 145s 11.98
9.56 9.99 10.69
11,Gl 11.96
I
1336 0,153 0.177
RR
10.71
ECL
1,310 1,476
1,000
0,849
0.276 0,310 0,390 0.458 0.603 0,677
0,179 0,210
0.082 0,096 0,127
RR
ECL
I
ECL
RR
0 Y-17(200”)
Hydrogenated Apiczon M (230”) C&,,, (ZOO’)
1,206
0.709
14,71
2085 16.70 2115
0.516 0.587
13.52 14.00
0.871 1.000
0.421
1973 15.48 2015 16,OO
1683 1713 1772 1814 1885
12.75
0,306 1571 lGl2
11,55
RR
0,249
ECL
--
1365 1407 1481 10.78
I
DEGS (200”)
9
l
2478 E
2357 1”1 2409
2280 I Fn
2161 2210
2085
1963 ii*
1883 iR
I
EQUIVALENTCHAIN LENGTHS@CL), ROLATIVERETENTIONS(RR) AND RETENTIONINDICES(I) OF WOOL WAX METHYL ESTERS
TADLE11
’ U = Unknown,
17.26 17.60 18.00 18,70 19.01 19.19 19.60 19.98 20.26 20.68 21.18 21.31 21.61
3.880 4,479 4,962 5,821 7,036
1.605 1,823 2,120 2.763 3,101 1982 2016 2056 2126 2157 2175 2216 2254 2281 2321 2374 2386 2412 1.957 2.340 3.153 3.1133 4.636
17.24 17.57 18.00 18.68 19.14 19.59 2175 2220
1978 2016 2059 2128 19,55
17,57 17.98 18,68 3.969
1.840 2.158 2.835
2873 2948 3001 3069 3144 3197 3265 3342 3462
3.511 4.310 4.976 5.985 7.345 8.476 10,198 12.574 17,441
20.71 21.48 22,02 22.72 23,49 24.02 24.72 25.50 26.93
2555 2605 2676
2481
2,533 2.902
1.484 1.701 2.056
19,49 20,Oo
17.48 17,99 18,71
2370
2172 2213 2283
B
8
8
and retentionindices, and Table II gives similardata for a seriesof wool wax fatty esters on the same pti. From Table II, it is evidentthat on t!heDEGS ahmn the iso-esters(terminal isopropyl group) have shorter retentiontimes than the anteiso-esters(terminalisobutyl group); the same elution pattern is, as expected, observed with the non-polar statiouaryphases as the threehomologous seriesare all of the same functionalclass. Semi-logarithmicplots for the differentseriesare esseutiallypemllel, and the separationfactors are essentiallyidentical,as indicatedpreviouslyby GersorP and Haken=. The separationfactors are given in Table III; as expected,the separations are enhancedon the stationaryphases of lower polarity TAELE
111
SEPARATION SERIES
FACTORS
OF
BRANCE3ED-
AND
STRMG!3T-CHAIN
Structure
Hy&ogemfed Apiezon iU
C&&
oy-17
DEGS
E..
211
235 2.35
2.14
1.76
Ad?Zk3TJ
211 211
23.5
2.14
1.76
FATTY
ESTER
The polarity has a greateriutluenceon the elution of iso-estersthan on that of anteiso-csters; +& is shown in Table IV, where variations in ECL values are given. TABLE IV REDUcXION IN RETENTION VARYING FGU-
rS&
AtieiW-
;r
-
(ECL
UNJXS)
OF ESTERS ON STATKONARY
Hydrogenated Apiezon M
CSl~IlS
ov-I7
DECS
-0.37 -0.28
-0.38 -0.31
-0.43 -0.32
-0.53 -0.30
1.00 kO.65
1.00
1.00
PHASES OF
L.00
The improved separationson *he non-polar colllmns are furtherevident by the appearanceof a fourth seriesof minor componentson hydrogena’kdApiezon M cobm shown as U in Tables II-IV. The saue separation factors are exhibited as for the other serieson this stationaryphase, z&d OQ a non-polar column separation of an unsaturatedspecies would’ be very unlikely. The composition of wool wax fatty acids has recently been reviewed= and, in addition to the saturated esters described,seriesof normal, iso- and auteiso-hydroxyestersare presentand the peaks observedare due to small amounts of the materialsthat remainafter the separation reportetj. Tables I and II show that the retentionis essentiallyidenticalon both of the non-polar phases, with retentionindicesbeing approximately250 units or 2.5 carbon numbersgreaterthan tie ECL values. With increasedpolarityof the stationaryphase,
GC OF ESTERS. Xm.
305
retention rek&ive to the n-alkanes is increased by 410 units or 4.1 carbon ~urnbers on the OV-17 cofumn and 800 units or 8 carbon numbers on the DEGS coiumn. These values show the same trends as in a previous study in which the retention of simple esters on polysiloxane stationary phases of increasing polar character was considered*‘. The results indicate that the inter&at&ship of retention indices and ECL values or other indices relative to some other homologous calibration line can be readily achieved by susingsimple data reduction units or by manual operation. REFEREZCES 1 2 3 4 5
M. R. Guerin and C_ V. Bznks. f_ Gas Chromato~r., 4 (1966) 428_ M. S. Wainwright and J. K. -H&en. J. Chromatogr., fS4 (1980) in press. L. S. E$tre, Anal. Chem.. 36 (1964) 31A. W. J. k VandenHeuvel, W_ L. Gardner and E. C. Homing, Anal. Chem., 36 (1964) 1550. L. J. Laxenz and L. B. Rogers, Ar!zzl.Chem.. 43 (1971) 1593. 6 W. J. A. VandenHeuvel and E. C. Homing, Biochim. Biophys. Acta, 64 (1962) 41. 7 S. J_ Hawks, J. Chronzarogr_SC& 10 (1972) 535. 8 A_ Grobler, I_ Chronzrugr_ SC& 10 (1972) 128. 9 H. F. Dymond and K, D. Kilburn, in A. B. Littlewood (Editor), Gcs Chromotogruphy 1966. Institutesof Petroleum, London, 1967,p. 353. 10 R. G. A&man, L Chromatogr. Sci., 10 (1972) 536. 11 J. K. H&en, J. Chronzatogr.,99 (1974) 329. 12 F. P. Woodford and C. M, van Gent, J. Lipid Res-, 1 (1960) 188. 13 T. K. Miwa, K. L. Miko&z&, F. P. Earle znd I. A. Wolf?, Annl. Chem., 32 (1960) 173Y. 14 E Vernon and C. 0. E. Ogundipe, L Chromarogr..,132 (1977) 181. 15 J. K. Haken and F. Vernon, J. Chronra~ogr., 186 (1979) 89. 26 J. K. Haken and D. K. M. Ho, J. Chrorn&ogr_, 142 (1977) 203. 17 X Guardho, J. A&zig&, G. Fiio, R_ Rodriguez-Vii and M. Gassiot, J. Cl;romatogr_, 11s (1976) 13. 18 T. Gerson. J. Chronrafogr.. 6 (1961) 178. 19 J. K. Haken, J. Chromrogr., 26 (1967) 17. 20 K. Mot%&, J. Amer. Oil Chem. Sot., 56 (1979) 55. 21 J. R Ashes and I. K. H&ken, J. Chromdogr., 101 (1974) 103.
LENGTH
(ECL)
J_RASHES CSIRO,
Div&ion of Am-ma! Production, P-O_ Box 239, Btocktowtr, N.S. W.
J. EC.HAKEN” Depmnnart of Polytr~er (AKrzr&j
Science,
Universfry of New
South
Wales, P-0.
2148(AK&~G&z~
Box
1, RezMgm~
2033
and s. c. MrJxs
CSIRO,LkXdm
of Anhal
Proaktion,
P-0. &xc
239,Blacktown,
N.S. W. 2148 (Aurtrahb)
(Received August23rd,1979)
SUMMARY
The interrelationship of equivalent chain Ie~gth and retention index values of straight- and branched-chain fatty esters is demonstrated on a typical polyester phase (DEGS), a phenyl-substituted poIy.siloxane (GV-17) and on two Iclw-polarity, thermally stable hydrocarbons. Hydrogenated Apiezon M and the synthetic hydrocarbon G&&,6 were used, and have lower polarities than the simple Apiezon greases that have been used with fatty esters. These stationary phases have polarities comparable to that of squakne and retention increments of fatty esters may be considered comparatively with those of simple esters which have been compared with squahme.
INTRODUCXON A variety of relationships have been used to specify retention behaviour and the interrelation of nine of the commoner structure retention parameters was shown ma+&ematicaIly by Guerin and Banks*. The growing use of automated gas chromatographic equipment with accompanying computer processing of retention data will continue to eJ.iminatetedious measurementsfrom charts and provide data virtually simultaneously with the elution of compounds. The determination of mathematical dead-time has recently been reviewed2 and this, together with the interpolation of
l
* FartXII: see3.K. E&ken,A. NguyenandM. S_Wkuvright,A C%ro~ogr., 178(1979)471. - To whautcerrespandeuce shouldbe &dressed.
the results for a compound on a graph prepared with homologous compounds, e.g., n-afkanes, to provide retention indices are simple procedures that can be rapidly handled by‘ on-line processors_ Retention indices3 are established by the relationship - log fm) + 100.x log
fR
--%tRN
(1)
where = tE?s ZRN = ta
retentionof unknown n-substrate; retentionof n-alkane with N carbon atoms; retentionof n-alkane with N + 1 carbon atoms; integerequal to carbon number of tnr.
The slope (b) of the semi-logarithmic plot of carbon number V~CYZCS retention of n-alkanesis common to many of the schemesconsideredby Guerin and Bank9 and is determinedfrom two points on a graph prepared with n-alkanes, as shown in eqn. 2, although in a computer plot the slope is determined statistically by considering all of the n-alkanes used in the calibration:
where
ami
t R(N+xl reteniionof n-alkaneswith N and N + x carbon atoms, tRCn, = res~tively The retentionindex is thereforegven by
The methyieneunit (MU) nomenclaturepropoc-I by VandenHeuvel et aL4 is essentiallyidenticalwith the retentionindex system.A eAibrationusingn-docosane and n-tetracosanewas used and the meth$ne unit multigr:-d by 100 gives the retention index. The proposalbased on a two-point caiibrationline is less satisfactorythan the cormalJydeterminedmulti-pointn-alkane calibrationline. The difhcuhyin selectinga referenceseriessuitablefor all polaritiesof solute and solvent is well known. Lorenx and Roger9 consideredthe effect of sorption phenomena on retentionindices and found significanterrors when the solute and solvent polaritieswere widely variant. These Endings were similarto those of VandenHeuvei and Horn&$, who recommendedthat the polaritiesof the two species should be similar. The use of secondary standards with n-alkanes has heen suggested, and Hawkes’ discussed the dif&ulty of using primaryn-alkanolsas proposedby Groblefl, who had critic&d the retentionindex system using polar stationaryphases owing to the rapid ehrtion of hydrocarbonsand indicatedthat compounds with very dS ferent boiling points cannot be chromatographedunder strictlyidenticalconditions. n-Propyl ethers were recommended as standards by Hawkes’, with ketones and
GCOFESTEECLX5
299
aldehydes as alternatives. Z-Alkanones had earlier been suggested as standards by Dymond and Kilburn9 and the merits of using ketones as secondary reference standards were later detailed by Ackmanzo. The di&ulty or impossibility of providing a universally acceptable reference series has been demonstrated by a study of the behaviour of compounds ofmcderately polar character, e.g., ;r-a&y1esters. With dimethyl polysiloxane, the solvent of lowest polarity included, it is apparent that methyl acetate (b-p. 57.1”) has a retention index of 569, Le., identical with that of n-pentane (b-p. 36.2”), while :r-hexyl acetate (b-p. 169.2”) has a retention index of 988, Le., almost identical with that of n-decane (b-p. 174.0”), so that for a given retention index the esters have higher boiling points than the comparable n-alkanes. With the introduction of two phenyl groups, i.e., using DC-710, the situation is reversed, with methyl acetate (I = 609) being comparable to n-hexane (b-p. 69.0°) and n-hexyl acetate (I = 1092) being comparable to n-undecane(b.p. 196”).With weaker acceptor phases,Lc. Z-60, these retention indices are increased to 737 and 1221,respectively, the former being equivalent to an n-alkane with a boiling point of approximately 220”. With a more polar stationary phase (Silar SCP), retention indices of 894 and 1313 are obtained, which are equivalent to n-nonane (b-p. 151°) and n-tridecane (b-p. 235”), respectively. The great differences in boiling points at each end of the scale reflect the differences in the slopes for the n-alkanes and simple esters on the various stationary phases. As alternative secondary reference standards, acetate esters and symmetrical esters with equal acid and alcohol chain lengths have been suggested”. The use of esters as standards has found considerable acceptance with fatty esters In the form of the equivalent chain length (ECL) relationship develop.ti by Woodford and Van GenF and Miwa et QZ_‘~.The relationship Is equivalent to tZ:e retention index system, results for an unknown ester being interpolated on graph prepared using homologous long-chain methyl esters. The relationship is shown in eqns. 4 and 5 and is identical with eqns. 1 and 3 except for the reference series used. EcL=
logt,--ogt,-i-N log
fR
-
1%
fRN
(4)
The quantities are as shown above but refer to methyl fatty esters, while the calibration line is usually constructed using even-bon-number esters. With computer processing of chromatographic data it is possible to provide retention data relative to multiple reference series and with the retention jndex and ECL representations a simple consideration of the slopes of plots for standard homologous series is necessary. In this paper we consider the retention of straight- and branched-chain fatty esters with the data expressed in. terms of both retention indices and ECL values. The data were determined on a typical polyester phase (DEGS) and on two lowpolarity, thermally stable hydrocarbons. Hydrogenated Apiezon M and the synthetic hydrocarbon C&I, were used, and have lower polarities than the simple Apiezon greases that have been used with fatty esters. These stationary phases are comparable
11.39 13,37 9.40 15.37 19,34 17.36
Clb Cl0 GO Cl2 C,,
gl Cll d: Cl2 Go
nAlknncs
16.00 18.00 20.00 22.00 7.43
Cl1
Cl,
Cl4
10.00 12,ao 14.00
C!,o
Methyl
~
13,37 15.36 19,35 17,36
1400 tm 1800 2200 2000
0,176 0,371 0.083 0,789 3,520 1,670
16.00 18.00 20.00 22.00
1859 2059 2257 2454 :tz
Z 4i79 9,499 0.040
lO.cIO 12.otl 14.00
1256 1457 1659
ECL
c87&b
Oiii0,221 0,471
JYyilrogenatcfl Aplezon M (230’) --. J??CL RR I
OSkCS
-_.
nmlrcr
Conlpounrls Ccrhon
-
1800 2200 2000
0.762 4.184 1.791
1860 2061 2262 2460
1.000 2.350 5,530 13.109
1600
1266 1468 1659
0,075 0,177 0,424
0,342
I
-
RR
(,zao’)
VI
7
__-
9,89 7.96 11,87 13,86 15,85
2015 2215 2414
1.000 2.170 4.550
---
1400 1200 1800 1600 2ov0
1410 1613 1814
0,091 0,211 0.465
0.093 0,044 0.436 0.201 0,943
I
RR
(2OOO)
16,OO 18,OO 20.00
10,oO 12.00 14.00
ECL
0
0,205 0.585 0.343
1,756 3,021 0,995
18.03 20.15 15,98
1,000 1.703 2.907 4.963
0.202 0.347 0.585
RR
IO,05 13,99 11,98
16,OO l&O0 20,oo 22,OQ
1om 12,00 14.00
ECL
DEGS (ZOO’)
Iri
2600 2800 2400
E
gn 1”1
I & p
pl . P
1800 2200 2000
z: 2796 2993
1794 2003 2201
I
EQUIVALENT CHAIN LENOTHS @CL), ROLATIVERETENTIONS(RR) AND RETENTIONINDICES (I) OF METHYL ALWNOATES AND /I-ALKANES
TABLE I
GC OF ESTEES
XIII.
301
in behaviourto squalaneand retentionincrementsof fatty estersmay be considered comparativelywitk tkose of simple esterswhick have been comparedwith squalane.
Gas chronmogrraphy
A Hewlett-Fackard57OOAgas ckromatograpk,fitted with dual flame-ionization detectors, was used. Nitrogen (30 ml/nun) was used as the carrier gas, and hydrogen (30 ml/mm) and air (200 ml/min) were suppliedto tke detector.All results were determinedisothermally at 200”. Four stationary phases were used, viz., in order of increasing polarity, kydrogenatedApiezon M (refs. 14 and 15), C&i,,, (ref. lo), OV-17 and DEGS, all coated to tke extent of 10% on Chromosorb W AW DMCS (SO-100 mesh) and packed into 2 m x 3 mm 0-D. stainless-steelcolumns. Tke gas chromatogmphwas connectedto a Hewlett-Packard3370B integrator with facilitiesfor paper-tape output on a teletype (STC, ASR-33). The ckromatograph was also fitted witk a Hewlett-Packard7671A auto-sampler. Preparation of fatty acid methyl esters from wool WQX Wool from Merino sheep was extractedwith light petroleum (b-p. 40-&Y), using a Eoxklet apparatus. Small portions of wax were saponitied with akzoholic sodium hydroxide solution. Unsaponifiable components were removed with light petroleum, the solution was then acid&d and fatty acids together with hydroxy fatty acids were extracted with light petroleum. Fatty acids were separated from hydroxy fatty acids by thin-layer chromatographyon silica gel G using n-hexanediethyl ether-acetic acid (70:30:1). The bands were detected under UV ligkt after sprayingwitk 0.1% etkanolic 2,7-dickloro~uoresceinand identi&d using standards (Applied Science Labs., State College, Pa., U.S.A.) run on the same plate. Fatty acid bands were scraped from the plate and extracted with dietkyl etker. Methyl esterswere preparedby adding diazomethanesolution. The data were processed by means of a PDP 1l/l5 computer fitted witk a high-speedpaper-tapereaderand disc pack (RKOS), dead-timesbeing calculatedby the procedureof Guardino et al.“. ECL values were computed from tke modi&i regressionequation first obtamed by tke procedure of Guardino eb al. from a homologous series of straigktchain faw acid methyl esters. In order to allow for tke normal variationstkat occur with net retention times with time (as up to 36 samples could be automatically processed), all experimentaldata were converted into retentionsrelativeto methyl paknitate (Cl6). Tke net retention time of methyl paknitate was constantly remeasuredin eack sample, as it eitheroccurrednaturally or, if not, was added as an internal standard. Retention indices were similarly determinedusing homologous IHlhane&
RESULTS
AND
DISCUSSION
Retention data for methyl alkanoates and n-alkanes on tke four stationary phases are shown in Table I as retentionsrelativeto metkyl palmitate, EC. dues
Conrpolmd’
9.99 10.72 11.35 11.62 11,99 12,33 12.71 13.35 13.62 14.00 14.71 15.00 15.33 15.62 16,OO 16,31 16.71 16,99
9.63
0,290 0.369 0.408 0.471 0.615 0,618 0.777 0.867 1.00 1,119 1.304 1.452
1217 1252 1325 1390 1417 1454 1487 1527 1590 1618 lG56 1727 1756 1789 1818 1856 1886 1927 1955
0,091 0.104 0,137 0.174 0.192 0.221 0.243 0.360 0.424 0,572 0,849 1,000 1,345
12,69 13,61 13,99 1469 1562 16,OO 16,69
1928 lG.69 17.00
1820 15.58 1858 16.00
1528 12,69 12,99 1619 13,58 1658 13.99 1727 14.70 1752 15.00
1425 11.57 145s 11.98
9.56 9.99 10.69
11,Gl 11.96
I
1336 0,153 0.177
RR
10.71
ECL
1,310 1,476
1,000
0,849
0.276 0,310 0,390 0.458 0.603 0,677
0,179 0,210
0.082 0,096 0,127
RR
ECL
I
ECL
RR
0 Y-17(200”)
Hydrogenated Apiczon M (230”) C&,,, (ZOO’)
1,206
0.709
14,71
2085 16.70 2115
0.516 0.587
13.52 14.00
0.871 1.000
0.421
1973 15.48 2015 16,OO
1683 1713 1772 1814 1885
12.75
0,306 1571 lGl2
11,55
RR
0,249
ECL
--
1365 1407 1481 10.78
I
DEGS (200”)
9
l
2478 E
2357 1”1 2409
2280 I Fn
2161 2210
2085
1963 ii*
1883 iR
I
EQUIVALENTCHAIN LENGTHS@CL), ROLATIVERETENTIONS(RR) AND RETENTIONINDICES(I) OF WOOL WAX METHYL ESTERS
TADLE11
’ U = Unknown,
17.26 17.60 18.00 18,70 19.01 19.19 19.60 19.98 20.26 20.68 21.18 21.31 21.61
3.880 4,479 4,962 5,821 7,036
1.605 1,823 2,120 2.763 3,101 1982 2016 2056 2126 2157 2175 2216 2254 2281 2321 2374 2386 2412 1.957 2.340 3.153 3.1133 4.636
17.24 17.57 18.00 18.68 19.14 19.59 2175 2220
1978 2016 2059 2128 19,55
17,57 17.98 18,68 3.969
1.840 2.158 2.835
2873 2948 3001 3069 3144 3197 3265 3342 3462
3.511 4.310 4.976 5.985 7.345 8.476 10,198 12.574 17,441
20.71 21.48 22,02 22.72 23,49 24.02 24.72 25.50 26.93
2555 2605 2676
2481
2,533 2.902
1.484 1.701 2.056
19,49 20,Oo
17.48 17,99 18,71
2370
2172 2213 2283
B
8
8
and retentionindices, and Table II gives similardata for a seriesof wool wax fatty esters on the same pti. From Table II, it is evidentthat on t!heDEGS ahmn the iso-esters(terminal isopropyl group) have shorter retentiontimes than the anteiso-esters(terminalisobutyl group); the same elution pattern is, as expected, observed with the non-polar statiouaryphases as the threehomologous seriesare all of the same functionalclass. Semi-logarithmicplots for the differentseriesare esseutiallypemllel, and the separationfactors are essentiallyidentical,as indicatedpreviouslyby GersorP and Haken=. The separationfactors are given in Table III; as expected,the separations are enhancedon the stationaryphases of lower polarity TAELE
111
SEPARATION SERIES
FACTORS
OF
BRANCE3ED-
AND
STRMG!3T-CHAIN
Structure
Hy&ogemfed Apiezon iU
C&&
oy-17
DEGS
E..
211
235 2.35
2.14
1.76
Ad?Zk3TJ
211 211
23.5
2.14
1.76
FATTY
ESTER
The polarity has a greateriutluenceon the elution of iso-estersthan on that of anteiso-csters; +& is shown in Table IV, where variations in ECL values are given. TABLE IV REDUcXION IN RETENTION VARYING FGU-
rS&
AtieiW-
;r
-
(ECL
UNJXS)
OF ESTERS ON STATKONARY
Hydrogenated Apiezon M
CSl~IlS
ov-I7
DECS
-0.37 -0.28
-0.38 -0.31
-0.43 -0.32
-0.53 -0.30
1.00 kO.65
1.00
1.00
PHASES OF
L.00
The improved separationson *he non-polar colllmns are furtherevident by the appearanceof a fourth seriesof minor componentson hydrogena’kdApiezon M cobm shown as U in Tables II-IV. The saue separation factors are exhibited as for the other serieson this stationaryphase, z&d OQ a non-polar column separation of an unsaturatedspecies would’ be very unlikely. The composition of wool wax fatty acids has recently been reviewed= and, in addition to the saturated esters described,seriesof normal, iso- and auteiso-hydroxyestersare presentand the peaks observedare due to small amounts of the materialsthat remainafter the separation reportetj. Tables I and II show that the retentionis essentiallyidenticalon both of the non-polar phases, with retentionindicesbeing approximately250 units or 2.5 carbon numbersgreaterthan tie ECL values. With increasedpolarityof the stationaryphase,
GC OF ESTERS. Xm.
305
retention rek&ive to the n-alkanes is increased by 410 units or 4.1 carbon ~urnbers on the OV-17 cofumn and 800 units or 8 carbon numbers on the DEGS coiumn. These values show the same trends as in a previous study in which the retention of simple esters on polysiloxane stationary phases of increasing polar character was considered*‘. The results indicate that the inter&at&ship of retention indices and ECL values or other indices relative to some other homologous calibration line can be readily achieved by susingsimple data reduction units or by manual operation. REFEREZCES 1 2 3 4 5
M. R. Guerin and C_ V. Bznks. f_ Gas Chromato~r., 4 (1966) 428_ M. S. Wainwright and J. K. -H&en. J. Chromatogr., fS4 (1980) in press. L. S. E$tre, Anal. Chem.. 36 (1964) 31A. W. J. k VandenHeuvel, W_ L. Gardner and E. C. Homing, Anal. Chem., 36 (1964) 1550. L. J. Laxenz and L. B. Rogers, Ar!zzl.Chem.. 43 (1971) 1593. 6 W. J. A. VandenHeuvel and E. C. Homing, Biochim. Biophys. Acta, 64 (1962) 41. 7 S. J_ Hawks, J. Chronzarogr_SC& 10 (1972) 535. 8 A_ Grobler, I_ Chronzrugr_ SC& 10 (1972) 128. 9 H. F. Dymond and K, D. Kilburn, in A. B. Littlewood (Editor), Gcs Chromotogruphy 1966. Institutesof Petroleum, London, 1967,p. 353. 10 R. G. A&man, L Chromatogr. Sci., 10 (1972) 536. 11 J. K. H&en, J. Chronzatogr.,99 (1974) 329. 12 F. P. Woodford and C. M, van Gent, J. Lipid Res-, 1 (1960) 188. 13 T. K. Miwa, K. L. Miko&z&, F. P. Earle znd I. A. Wolf?, Annl. Chem., 32 (1960) 173Y. 14 E Vernon and C. 0. E. Ogundipe, L Chromarogr..,132 (1977) 181. 15 J. K. Haken and F. Vernon, J. Chronra~ogr., 186 (1979) 89. 26 J. K. Haken and D. K. M. Ho, J. Chrorn&ogr_, 142 (1977) 203. 17 X Guardho, J. A&zig&, G. Fiio, R_ Rodriguez-Vii and M. Gassiot, J. Cl;romatogr_, 11s (1976) 13. 18 T. Gerson. J. Chronrafogr.. 6 (1961) 178. 19 J. K. Haken, J. Chromrogr., 26 (1967) 17. 20 K. Mot%&, J. Amer. Oil Chem. Sot., 56 (1979) 55. 21 J. R Ashes and I. K. H&ken, J. Chromdogr., 101 (1974) 103.