724 Ayzalyst, November, 1974, Vol. 99, pp. 724-728 Antioxidant Analysis Incorporating a Thin-layer Chromatographic Separation Procedure BY DAVID T. MILES j.4 ustralian Post Ofice Reseaych Laboratories, 59 Little Collins Street, Melbourne 3000, Australia) The method described consists of three distinct and separate steps. Initially, a thin-layer chromatographic technique, involving the use of mixed silica gel - aluniinium oxide plates, is used to separate and identify the ex- tracted antioxidants. This step is followed by confirmation of the suspected identity and determination of the antioxidant by using infrared and ultra- violet spectroscopy. This method is considered applicable, over a range of concentrations, to the analysis of both antioxidants and ultraviolet screening compounds possessing ultraviolet absorbing properties, in polyolefinic polymers such as polyethylene, polypropylene, poIy(4-methylpentene), etc.I t can be used t o detect such additives down to a level of 200 p g g-I, and possibly even less, in a given polymer. In addition, the procedure has been successfully applied to the separation of isomeric antioxidants and to the identification of antioxidants present in lubricating oils. CURRENT interest in the analysis of various polymers at the Australian Post Office Research Laboratories led to a decision to investigate methods available for the determination of antioxidants. Those applicable to the six antioxidants most frequently used in Australia for the production of polyolefins were of prime importance. It was immediately realised that numerous rnethod~l-~ existed for the identification of specific antioxidants and mixtures thereof.While accounts of some of these methods reported a subsequent quantitative deter- minati0n,~$~,6 none appeared to report a separation of the extracted antioxidants by thin- layer chromatography using 1 + 1 silica gel - aluminium oxide plates. Therefore, our studies were directed towards this aspect and eventually we developed a procedure that possessed improved antioxidant resolution and facilitated their identification. When linked with a step for the confirmation and quantitative determination of the antioxidants it formed the basis of the procedure reported in this paper. Users of this procedure will realise that new antioxidant compounds are constantly being encountered and conflicting conclusions may therefore be drawn. EXPERIMENTAL APPARATUS- All polymer extractions were performed with a standard Soxhlet apparatus.Any reputable thin-layer chromatographic equipment was considered to be suitable for Infrared spectra were obtained by using a Jasco IR-G diffraction grating spectrophoto- A comparable instrument would A Perkin-Elmer 124, double-beam, ultraviolet - visible spectrophotometer or a similar use in the separation step. meter covering the frequency range 4000 to 400 cm-l. suffice. unit was used for measurements. MATERIALS- Type E, mixed 1 : 1. reagent grade chemicals. Thin-layer plates were prepared from Merck Kieselgel G and Merck aluminium oxide, The solvents required, methanol, chloroform, and methylene chloride, were all analytical- 0 SAC and the author.MILES 725 Antioxidant reference standards with a concentration of 1000 mg 1-1 were prepared from pure samples of : A, 1,1,3-tris(2-methyl-4-hydroxy-5-t-butylphenyl)butane; B, 4,4’-thiobis(6- t-butyl-m-cresol) ; C, NN’-di-P-naphthyl-j5-phenylenediamine ; D, 2,2’-methylenebis[6-( 1- methylcyclohexyl)]-fcresol; E, 2,6-di-t-butyl-@-cresol; and F, 2,2’-thiobis(4-methy1-6-t- butylphenol).PROCEDURE- Cut the polymer specimen into small pieces and use accurately weighed portions of between 2 and 5 g, depending on the amount and type of antioxidant in the polymer. If these are not known use a maximum mass of 5 g. Extract the sample with 50 ml of methylene chloride or chloroform, by use of a Soxhlet extractor, for a period of 8 hours to ensure complete removal of the antioxidant.Cool and filter the extract if necessary; in isolated instances pigment particles will be observed to pass through the extraction thimble. Evaporate the cool, filtered extract nearly to dryness, immediately re-dissolve the residue in 500 p1 of methylene chloride and transfer the solution to a stoppered phial ready for the thin-layer chromatographic separation. Coat the thin-layer chromatographic plates with a slurry of aluminium oxide and silica gel (1 + l), containing calcium sulphate as binder, by use of the following technique. Dis- perse approximately 40 g of the powder mixture in 60 ml of distilled water and stir gently to ensure complete wetting of the powder and to assist the removal of any entrapped air.Allow the slurry to stand for about 5 minutes to thicken. Adjust the spreader to give a 300-pm coating to the glass plates in one steady pass. After gelling, load the plates horizontally into a thin-layer chromatographic plate rack and place the rack in an oven at 110 “C for 2 hours in order to activate the plates. Subsequently, store the rack, with the prepared plates in a vertical position, in a desiccator until required. Plates not used within 3 days should be re-activated in the oven for 15 to 30 minutes at 110 “C. Spot the solutions on to the plates in the usual manner by using a graduated spotting pipette and a template. Apply 5-pl aliquots of the reference standards and concentrated sample extract, such that the resulting spots are kept to a minimum size, e.g., 3 mm or less in diameter.This can be achieved by spotting several small aliquots, e.g., 2, 2 and 1 p1 in succession, while directing a current of air to the point of application. Place the spotted plate in an unlined developing tank that has been saturated at room temperature with chloroform vapour, chloroform being the eluting medium, and allow the solvent front to travel to a distance of 100 mm from the origin. Afterwards, dry the plate in air and locate the resolved component spots by spraying the plate with a 50 g 1-1 solution of iodine in methanol. As an alternative, the plate can be placed in a developing tank containing iodine vapour. In both instances the resolved spots appear as light yellow to red - brown areas on a near-white background.Confirm the identity of the antioxidant that has been indicated by thin-layer chromato- graphy by obtaining an infrared spectrum of a second sample extract, concentrated to the required level, and comparing it with the spectrum of a pure reference compound or by reference to a recognised atlas of spectra. When the initial extract contains components other than antioxidants or mixtures thereof, it may be necessary t o repeat the thin-layer chromatographic step. A streak of the concentrated extract, i.e., 30 p1 to 40 pl, is applied to the thin-layer plate together with a control spot. After elution, location of the area of interest is achieved as before with a solution of iodine in methanol, while the streaked portion of the plate is kept covered with aluminium foil.The foil is subsequently removed and the area indicated to contain the antioxidant is scraped off. This portion is then extracted with chloroform, filtered, the chloro- form evaporated down and the resulting concentrate applied to a potassium bromide disc in sufficient volume to yield an acceptable infrared spectrum following complete removal of the chloroform. The identity of the resolved component is then confirmed in the manner described above. When quantitative results are required it is advisable to remember that the normal concentration of antioxidant in a manufactured polymer ranges from 0-2 to 6 mg g-l. Assum- ing that the sample contains 2 mg g1 of an antioxidant, and taking 5 g for extraction, the final extract, adjusted to 50 ml, will contain 20 mg 1-l.In most instances the amount taken is adjusted to obtain a concentration of about 20 mg 1-l in the final extract.726 MILES : ANTIOXIDANT ANALYSIS INCORPORATING A [AnhdySt, VOl. 99 The final quantitative determination of the indicated antioxidant is obtained by use of ultraviolet spectrophotometry. Record the wavelengths at which the suitably diluted standard solutions show maximum absorbance in a 10-mm path length cell. Then, plot calibration graphs of concentration ‘ueysus absorbance for the range of concentrations from 0 to 50 mg 1-1 for each antioxidant studied at the characteristic wavelengths recorded in Table I. Alternatively, the required information can be obtained more conveniently from a table of A,,,. and absorbance per 10 mg 1-1 for each antioxidant.TABLE I CALIBRATION RESULTS FOR THE ANTIOXIDANTS STUDIED 2,6-Di-t-butyl-P-cresol a t 283 nm A I > Concentration/ mg 1-l Absorbance 10 0.112 20 0.227 30 0.330 40 0.434 50 0.535 4,4’-Thiobis(G-t-butyl- m-cresol) at 247 nm Concentration/ mg 1-1 Absorbance 10 0.380 20 0.800 30 1.190 40 1.580 60 1.920 2,2’-Methylenebis [6- (1 -methyl cyclohexyl)]-p-cresol at 287 nm Concentration/ n I 3 mg 1-1 Absorbance 10 0.115 20 0.232 30 0-360 40 0.475 50 0.600 NN’-Di- j3-naphthyl-p- phenylenediamine at 321 nm mg 1-l Absorbance 5 0-442 10 0.890 15 1.265 20 1.690 r A \ Concentration/ - L 1,1,3-Tris(2-rnethyl-4-hydroxy- 5-t-butylpheno1)butane at 281 nm Concentration / mg 1-l Absorbance 10 0.120 20 0.235 30 0.340 40 0.460 50 0-570 2,2’-Thiobis(4-methy1-6- t-butylphenol) at 296 nm I A > Concentration/ mg 1-’ Absorbance 10 0.242 20 0-460 0.702 30 40 0.915 60 1.1 18 From these results and a known mass of extracted polymer, the concentration, and hence the mass, of antioxidant in the original sample is calculated.DISCUSSION AND RESULTS An initial literature survey revealed that the majority of reported procedures involved the use of mixed solvents for elution on thin-layer plates used in the resolution of antioxidants. This probably resulted from necessity rather than choice. Mixed eluting solvents suffer from anomalous effects arising from the effect of varying humidity conditions on the evaporation rates of the components. These anomalies, in turn, affect the partitioning of the extracted antioxidants between the liquid and solid phases.Our experiments were therefore aimed at finding a suitable single solvent to replace the various mixtures advocated and they revealed that chloroform performs satisfactorily when used in conjunction with 1 + 1 silica gel - aluminium oxide thin-layer plates. The decision to investigate a mixed silica gel - aluminium oxide plate for the thin-layer chromatographic separation of the five antioxidants of interest resulted from the observations made when either simple silica gel or aluminium oxide plates were employed. Neither coating gave a completely satisfactory result ; either the resolved components were bunched near the solvent front or they were held back in the region of the origin. This bunching was attributed to the adsorptive forces of the coating.With silica gel, the adsorptive forces were too weak as a result of the inability of the sterically hindered phenolic antioxidants to achieve a silica gel - solute interaction by means of hydrogen bonding; hence, the components moved with the eluting solvent. With the simple aluminium oxide plates the resolved components were confined to an area near to the origin, probably because of a property of aluminium oxide whereby the liydroxyl groups of the antioxidant are superficially built into the crystal lattice of the oxide. It was consequently concluded that a 1 + 1 mixture of aluminium oxide and silica gel should provide a state midway between these two extremes. The results we achieved tended to support this conclusion. Experiments to determine the optimum thickness of the silica gel - aluminium oxideNovember, 19743 THIN-LAYER CHROMATOGRAPHIC SEPARATION PROCEDURE 727 layer indicated that a 300-pm coating gave the best resolution of sample extracts.Extracts in methylene chloride yielded fewer unwanted components, i e . , processing additives and degradation products, than those in chloroform. If the amount of antioxidant C is to be determined the analyst should be aware that chloroform will not completely extract this antioxidant from a polymer specimen. As antioxidant C is also far less soluble in methylene chloride at room temperature than at elevated temperatures it was necessary to heat extracts of this component nearly to the boiling-point prior to spotting. In the author's experience, when chloroform was replaced by methylene chloride in the initial extraction procedure, quantitative extractions of the order of 980 mg g-1 were obtained.The usual precautions for thin-layer chromatographic studies were observed, e.g., tightly fitting developing tank lids, saturation of the developing tank with solvent vapour, etc. Experiments with lined and unlined tanks showed that there was little difference between them, except perhaps a reduced elution time with a lined tank. Failure to take the above precautions did not necessarily mean that the plate concerned was wasted. If sufficient standards were spotted across the plate (a minimum of three was desirable) the variation in retention values could be recorded and an identification of the unknown obtained.In this event confirmation of the apparent identity was imperative. The migration distances of the extracted sample components on the thin-layer chromato- graphic plate were measured by determination of their individual R, values. Owing to the variations in R, values between plates, it was considered more reliable to relate the R, values of the components to the R, value of a reference standard, i.e., antioxidant E, The new ratio was known as the R, value' (see Table 11). TABLE I1 RETENTION VALUES OBSERVED FOR THE ANTIOXIDANTS STUDIED Observed R p and RT values of antioxidant- Antioxidant as a single substance as one of a mixture n f 1 -- RJ? H T RF R T A 0.25 0.25 0.25 0.25 B 0.43 0.44 0.51 0.52 C 0-87 0.89 0.90 0.92 D 0.95 0.97 0.95 0.97 E 0.98 1.00 0.98 1.00 F* - - - - * Reasons for the non-inclusion of antioxidant F are explained in the following paragraph.From Table I1 it will be noticed that the retention values varied according to whether the component of interest was present as a single substance or as part of a mixture. Those spots which did not correspond to any known antioxidant or its breakdown products were attributed to the processing aids mentioned earlier. Antioxidant A did not exhibit the behaviour under illustration; it performed in a similar manner to antioxidant D in this respect. Antioxidant F is not included because it has not been encountered in mixtures in our experience to date, and, from information received, it appears that this is most unlikely to occur. In addition to the determination of antioxidants in polyolefins, the determination of ultraviolet screening compounds4 in this type of polymer can be carried out in a similar manner.A prior separation step is necessary if the latter compounds are to be determined on the same sample as the antioxidants, as they are very strongly absorbing over the ultraviolet wavelength range and could interfere in any attempt to determine the amount of an antioxidant. This method should also prove useful for the determination of antioxidants in other classes of polymer. The procedure as reported above has, in addition, provided information regarding the identities of antioxidants, e.g., styrenated phenol, in various lubricating oils.6 CONCLUSIONS Mixed silica gel - aluminium oxide thin-layer plates eluted with chloroform can be used to improve the resolution and ease of identification of antioxidants extracted from polyolefins. Coupled with the confirmatory and quantitative steps, the reported thin-layer technique728 MILES forms the basis of a general procedure that can be used for determining antioxidants in polyolefins both qualitatively and quantitatively. The permission of the Australian Post Office to publish this paper is hereby acknowledged. 1. 2. 3. 4. 5 . 6. 7. REFEREKCES Slonaker, D. F., and Severs, D. C., Analyt. Chem., 1964, 36, 1130. Woggon, H., Kom, O., and Jehle, D., Nuhrung, 1966, 9, 496. Crompton, T. R., Eur. Polym. J., 1968, 4, 473. Crompton, T. R., “Chemical Analysis of Additives in Plastics,” Pergamon Press, Oxford, New Zubkova, N. D., Turskii, Yo. I., Genkina, V. I., and Klyacho, G. V., Khim. Tekhnol. Topl. Masel, Simpson, D., and Currell, B. R., Analyst, 1971, 96, 616. British Standard 2782: 1970, Part 5, in preparation. York, Toronto, Sydney and Braunschweig, 1971. 1964, No. 8, 60. Received October 15th, 1973 Amended A p i l 23vd, 1974 Accepted May 21st, 1974