chromat ography in t he pai nt & al l i ed i ndustri es. Raw mat eri al s: Par t 2 Ann J. Wal t on Cashew nutshel l l i qui d The liquid extract obtained from the natural cashew nutshell is rich in phenolic substances which are derived from anacardic acid C 6 H 3 (OH) (CO 2 H) (C 15 H 31 -n), where n may have values of 0, 2, 4 or 6 and represents various degrees of unsaturation in the aliphatic C 15 side-chain. Industrial decarboxylation of this material affords cardanol C 6 H 4 (OH) (C 15 H 31 -n) plus other substituted phenols and polymeric residues. Tyman et al. (197, 198) have studied the analysis of all these products using GC, molecular distillation, TLC and mass spectrometry. After hydrogenation and the formation of the corresponding methyl esters, the products were analysed by GC using glass columns (5ft x 3/16in) packed with acid washed and silanized Diatomite as support material and which was coated with non-polar stationary phases such as SE30, SE25 or APL, or semi-polar phases such as 0V17, Dexil 300 or PEGA. Alternatively, the samples were sub- jected to an acetylation procedure prior to GC exami- nation on columns containing Dexil 300, SE30 or SE52. The GC equipment consisted of a Pye-Unicam model 104 instrument operated with nitrogen carrier gas (flow rate 45cm 3 min -1 ) and equipped with FID. The saturated (15:0), monoene 15: 1) , diene (15:2) and triene (15:3) constituent phenols were de- termined by mass spectrometry following a TLC sepa- ration, whilst both TLC and GC separations required that correction factors were applied in order to account for different response characteristics of the detector system. The more highly-polymeric samples were handled by molecular distillation techniques, which gave fractions that were subsequently examined by GC. It was concluded that the bulk of cashew nut- shell liquid was sufficiently volatile to allow GC analysis along the lines indicated above and, since the derivatives employed could be quantitatively pre- pared, the GC peaks could be handled by standard integration procedures that permitted quantitative de- termination of the component phenols. Resi n aci ds The methyl esters of abietic and dehydro- abietic acids were separated by Medyantsev et al. (134) using a column (1.8m x 2mm dia) containing 4.2% Varnisch 3R-728 as stationary phase that opera- ted at 125C and with helium as carrier gas at a flow rate of 100cm 3 min -1 . The methyl esters of laevopi- maric, palustric and neoabietic acids, which were obtained by isomerising abietic acid at 180 under carbon dioxide or 200C under argon, were analysed by Fomin et al. (155) using GC, infrared and ultra- violet spectroscopy. Pine rosin acids including pimaric, isopimaric, laevopimaric, abietic and neoabietic, alongside C12 to C22 fatty acids were resolved simultaneously in a GC procedure described by Lapshina and Kosynkova (117). The acids were converted into their corres- ponding methyl esters using diazomethane and sub- sequently separated on a column packed with 5% polyethanediolsuccinate on Chromatron N-AW-DMCS which was operated under a temperature program- ming schedule between 150 and 230C. In an alternative procedure, suggested by Mahood and Rogers (129), the fatty acids were sepa- rated initially from the resin acids by preliminary methylation followed by the addition of 2,2,4-tri- methylpentane prior to extraction with 5% potassium hydroxide solution in water, whereby the resin acids were carried into the aqueous phase. After applying suitable recovery techniques each of the acid samples were subjected to treatment with diazomethane, to ensure total methylation, followed by GC analysis on a column containing EGSS-X on Gas Chrom P at 175C. A review dealing with the analysis of the lower fatty acids, with up to 18 carbon atoms, by capillary GC has been published by Krupcik et al (114). Amongst the topics discussed were the separa- tion of free fatty acids as opposed to the correspond- ing methyl esters, qualitative and quantitative aspects including the calculation of Kovats indices, plus fea- tures of the equipment such as ovens, detectors, stationary phases and the use of support-coated as opposed to wall-coated open tubular columns. Mi scel l aneous compounds The lower primary aliphatic amines were separated, as their corresponding fluorine-containing Schiff bases, on a temperature-programmed GC column containing Tenax SE-30 as stationary phase and with an electron-capture detector, according to Hoshika (88, 89). Residual secondary and tertiary amines were analysed as the free bases. Amongst the samples examined were methylamine, ethylamine, n- and iso-butylamine, n- and iso-amylamine, dimethyl- amine, diethylamine, di(n-propyl)amine, trimethyl- amine and triethylamine. Zatkovetskii and Safonova J ANUARY 1981 PIGMENT AND RESIN TECHNOLOGY 5 (210) reported the GC separation of toluene diisocy- anate and hexamethylene diisocyanate using sintered PTFE powder (Polychrom 1) as column packing material. The same substances were determined in their respective prepolymers or adducts by J anmot (96) who used a sample solution in ethylacetate which was injected on to a GC column containing silicone gum rubber and operated at 150C. In the latter procedure 1-chloronaphthalene served as inter- nal standard. Dimethylacetamide is present as an impurity in concentrations below 2% in regenerated synthetic glycerol. The former substance was determined using a GC column (1m x 3mm i.d.) packed with 20% of a 3:1 mixture of poly(oxpyhenylene) and poly(ethylene- glycol) supported on Chromatron N-AW-DMCS by Sergeeva et al. (169). The column was eluted with nitrogen as carrier gas (30cm 3 min -1 ) and operated at 135C. After the analysis the column was cleaned up by heating it to approx. 300C. A comprehensive list of GC retention data on 50% methyl and phenyl silicone columns, operated under temperature programming, was published by Mattson and Peterson (132) for the trimethylsilyl de- rivatives of aromatic compounds such as phenols, naphthols, alcohols, carboxylic acids, chloro-com- pounds, aldehydes and ketones, the latter two being analysed both as oximes and O-methyl oximes. An interpretation of the correlation between structure and retention behaviour was given. Alkyphenols produced from the olefin fractions of petroleum were analysed according to a GC method developed by Shchegol et al. (170). The technique was applied in particular to the analysis of p-tert-octyIphenol produced from 2,4,4- trimethylpent-l-ene. A total of 14 impurities, which included acetone, methylacrylate, methyliso- butyrate, ethyl methacrylate and methyl 2-hydroxy-2- methylproprionate, were resolved in samples of methyl methacrylate by Sokolowska et al. (181) using a GC column packed with Ucon-LB-550X on Chromo- sorb W that was operated at 85C. Acetaldehyde, as a source of interference in the quantitative GC analysis of lower levels of vinyl chloride, was eliminated using a mixture of glass wool and sodium bisulphate prior to the GC column proper according to the method of Krishen and Tucker (113). White (206) proposed the use of a column of OV-1 on acid-washed Diatomite, that was temperature programmed from 180 to 380C, in the identification of waxes that were taken from museum objects. The presence of free parent acids in cyclic anhydrides was determined by Haas and DuBois (68). Each anhydride sample was initially reacted with morpholine so that the residue of free carboxylic acid could then be con- verted into the corresponding silyl derivative using bis(trimethylsilyl)acetamide reagent. The sample so obtained was analysed by GC on a column of 5% SE52 on Gas Chrom Q under temperature programme control from 125 to 250C. TLC, on plates comprising a mixture of silica gel and boric acid, was used to separate , , and tocopherols from the unsaponi- fiable matter of vegetable oils, prior to their determi- nation by a GC technique employing a silicone SE-30 column according to Riera (160). Squalane served as an internal standard and tocopherols were measured in samples of soya bean, safflower and other oils. A very similar technique for the GC analysis of toco- pherols was also recorded by Dompert and Beringer (46). ADDITIVES Pl asti ci sers A GC procedure for the qualitative and quanti- tative analysis of dicarboxylic aliphatic acid esters, which are frequently used as plasticiser components in surface coatings, has been reported by Bloom (26). A 0.5-microlitre quantity sample dissolved in carbon disulphide, and equivalent to between 2 and 5 micro- grammes of ester, was injected at 260C on to a spirally-wound glass column (8m x 1.8mm i.d.) packed with 1% QF-1 (a trifluoropropylmethylsilicone liquid phase available commercially from Applied Science Labs.) on 100 to 120 mesh Chromosorb Q. The instrument used was a Hewlett Packard Series 7400, having a FID which operated at 260C, and which provided an oven temperature programme from 160 to 250C at a heating rate of 3 deg. min - 1 . The carrier gas was nitrogen admitted at a pressure of 29 psi which enabled a flow rate of 16cm 3 min - 1 to be maintained. Amongst the esters examined were 14 alkyl adipates, 8 alkyl azelates and 9 alkyl sebacates and their retention times were compared to that of methyl arachidate which was employed as the inter- nal standard. Even when these materials were used as mixtures to test the GC technique, at least 25 of them could be resolved and identified unequivocally. The pre-determined elution properties of any of the above-mentioned compounds allowed them to be readily identified as components in two commercial plasticisers namely Duroplaz DIOA and Adimoll BB. Imia (91) described the determination of the alcohol components of ester-type plasticisers by GC in a pro- cedure that required a preliminary ester interchange reaction to be carried out on the sample using metha- nol. Dioctylphthalate (DOP), 2-ethylhexanol (2EH) and other substances which arose during the reaction of 2EH with phthalic anhydride were separated by Mosinska (144) using a GC column packed with 5% DC 200 deposited on Chromosorb G-AW-DCMS. The column was temperature programmed from 110 to 250C at 20 deg. min - 1 and the 2EH was determined at 110C with a carrier gas (hydrogen) flow rate of 60cm 3 min - 1 whilst DOP was determined at 240C with a carrier gas flow rate of 120cm 3 min - 1 . Hydro- chloric acid is a possible degradation product of PVC that may react with the epoxy ester added as stabili- ser and plasticiser to form chlorohydrins. Although these would decompose during GC to form oxo com- 6 PIGMENT AND RESIN TECHNOLOGY J ANUARY 1981 pounds, it was reported by Startin et al. (187) that the chlorohydrins could be converted into the corres- ponding trimethylsilyl ether derivatives which were sufficiently stable to allow their examination by GC- mass spectrometry. Results were quoted for a series of fatty acid methyl ester chlorohydrins. Kochel and J aroszynska (105) employed a GC column packed with Chromosorb W and coated with OW25, SE30 or Apiezon L as stationary phase in the identification of 13 common ester plasticisers, inclu- ding tricresyl phosphate, and 17 frequently-used anti- oxidants present in nitrile and neoprene rubbers. It was found that the plasticisers could be examined by evaporating them directly from the samples at 480C, prior to GC analysis at 250C, and that no interfer- ence was evident from the degradation products of the rubbers. The antioxidants needed to be extracted from the rubbers over a 3-hr. period using acetone at the ambient temperature followed by GC examination of the extracts. Surf act ant s Surface-active agents, derived from the re- action of a long-chain aliphatic alcohol with ethylene oxide, have been extensively studied by GC. Stancher and Favretto (186) reported the fractionation of samples having the general formula RO(CH 2 CH 2 O) 13 H, where R represents an n-alkyl radical, and found that only the components having n = 1 to 4 were capable of resolution on a glass column packed with 3% SE30 on Gas Chrom Q. The volatility of the oligomers was improved by conversion to the corresponding tri- methylsilyl derivatives, but the resolution was no better than before even when temperatures above 330C were employed. In order to quantify the alkyl radical distribution of ethoxylated alcohols, Sones et al. (183) first converted the samples into the corres- ponding alkyl iodides prior to GC analysis. Farkas (53) studied the products obtained by catalytic oxyethylation of lauryl alcohol on a GC column packed with 60-80 mesh Chromosorb and coated with 7% Dexsil. In order to eliminate tailing effects, the hydroxyl groups present in the sample were subjected to ketene acetylation. The same author also examined the product distribution of do- decylalcohol oligomers at 190-290C by GC on 30- 60 mesh Chromosorb W-AW-DMCS which was coated with 10% SE301 as liquid phase. Acetylation or methylation of the samples prior to chromato- graphy led to the elimination of peak tailing, but it was found that the methylated compounds gave rise to superior separation performance. Goto et al. (65) used a GC column (75cm x 3mm i.d.) that was packed with 2% OV-I on Chromosorb W (80-100 mesh) and which operated at 232C using nitrogen as carrier gas (40cm 3 min -1 ) to study the state of solu- tion of the non-ionic surfactant hexaoxyethylene n- lauryl ether. The technique was used in conjunction with the initial GPC (gel permeation chromato- graphic) separation of the surfactant solution on Sephadex G-200. The results suggested that a con- tinuous self-association of the surfactant existed at low concentrations, whereas a more constant micellar size was assumed at higher concentrations. Sones et al. (183) have also reported the alkyl homologue dis- tribution in alkyl suphates and alkylether sulphates, whereby the samples were initially subjected to acid hydrolysis prior to their conversion into the corres- ponding alkyl iodides and analysis by GC. Shilov and Molova (171) on the other hand proposed to calculate the polarity index of sodium alkylmonosulphate samples from their GC retention volumes and to plot this parameter against hydro- carbon chain length in order to determine the length of the hydrophobic part of the surfactant. Miscellaneous Ermilova and Maiorova (49) described a GC technique for the determination of formaldehyde in formalin solutions. Polyethyleneglycol adipate. (10% w/ w) on Polychrom 1 formed the column packing material and detection of the samples was by thermal conductivity. Impurities such as oligomeric glycols, which may be present in formalin solutions, were quantified by Dankelman and Daeman (41) who sepa- rated the corresponding silylation products by GC. The results of these experiments were corroborated using a 220MHz NMR spectrometer. Benzophenone and benzotriazole, which may be used in plastics to absorb UV light, were deter- mined by FID on a GC equipped with dual stainless steel columns containing 10% SE-30 on Chromosorb W, and which were temperature programmed from 180 to 270C, in a method developed by Horacek (87). The absorber was initially extracted from the sample of plastic using chloroform or ethyl ether, and subsequently applied to the GC column as a solution in chloroform, ethanol or heptane. Using calibration graphs or internal standards, it was claimed that 0.1 microgram quantities of absorber could be deter- mined and that the error did not exceed 20%. According to a paper published by Uden (199), both high performance liquid chromatography and GC analytical techniques have been applied to the analysis of metals. GC required neutral molecules combined with low molar mass and typical ligands studied contained -alkyl, -aryl, dienyl, carbonyl or nitrosyl functionality. For example, -diketonates of beryllium, chromium, aluminium, iron ( I I I ) and cobalt (III) have been used in realistic analytical deter- minations, whilst -thioketonates were found to be valuable in the determination of nickel, cobalt, zinc, palladium, platinum, cadmium and lead. The most suitable detector for this approach was a direct cur- rent argon plasma emission spectrometer. The sample was transferred from the chromatograph via a heated line with annularly introduced argon sheath gas to op- Continued on Page 10 10 PIGMENT AND RESIN TECHNOLOGY J ANUARY 1981 APPLICATIONS OF GAS CHROMATOGRAPHY Continued from Page 6 timise the form of the plasma discharge, to ensure a minimum of peak broadening, and the sensitivity to sub-nanogram levels of the metals. Gas number, which is a parameter for characterising the foaming properties of blowing agents, was determined from the nitrogen content of the thermal degradation pro- ducts in a pyrolysis-GC approach reported by Makagon et al (130) Each determination took from 20 to 30 minutes and the results showed a standard deviation of 0.012 to 0.038.