186 Analyst, March, 1968, Vol. 93, @@. 186-190 Use of Anthrone in the Determination of Trace Amounts of Glycerol BY F. A. LYNE, J. A. RADLEY AND M. B. TAYLOR ( J . A . RadEey Research Institute, 220-222 Elgar Road, Reading, Bevkshire) A fluorescent product is formed when a sample containing glycerol is heated with a solution of anthrone in sulphuric acid, and based on this a method for the quantitative determination of glycerol in amounts from 10 to 75 pg is described. The original concentration of glycerol is determined from the intensity of the fluorescence of this product in sulphuric acid solution, after first producing a calibration graph from known amounts of glycerol. The method is particularly useful for determining the concentration of glycerol in spots separated from interfering substances by chromatography.THERE are several methods available for determining glycerol in milligram amounts,l including periodate oxidation, acetylation and Zeisel and Fanto’s method.2 Raveux3 outlines methods for extracting small amounts of glycerol from a mixture; this may be done by a form of steam-distillation with ethanol instead of water, which might be useful with the anthrone method of analysis. His suggestion for determining glycerol, by boiling with potassium di- chromate in nitric acid for 20 minutes, followed by back-titration of the dichromate with ferrocyanide, with diphenylamine as indicator, is only suitable for amounts of glycerol in excess of 100 pg. As a qualitative test for glycerol, Schutz4 introduced the reaction between anthrone and glycerol in concentrated sulphuric acid solution, in which a solution showing a dull orange fluorescence in ultraviolet radiation was formed.Later this was adapted by Radley5 for microgram amounts of both glycerol and ethylene glycol. This latter has now been put on a quantitative basis. The reaction involves the condensation of an a-p unsaturated aldehyde with anthrone; the formation of acrolein by dehydration of glycerol with sulphuric acid and completion of the reaction by cyclisation.6 (1) CH20H.CHOH.CH,0H-2H,0 __+ CH,=CH.CHO. Acrolein 0 0 0 (2) & + CH,=CH.CHO Acrolein * (A@ from $-9 (@ + H,O Anthrone H,SO* C1-r Benzanthrone. kH=CH, Intermediate EXPERIMENTAL The excitation spectrum of benzanthrone shows that there is a wide excitation wave band between 350 and 500mp so that no filters were necessary in the excitation beam.With the Locarte single-sided fluorimeter Mk. 4, the maximum in the fluorescence spectrum was observed at 575mp and all subsequent readings were taken at this wavelength. 0 SAC and the authors.LYNE, RADLEY AND TAYLOR 187 SOLVENTS- The intensity of fluorescence of a solution of benzanthrone in sulphuric acid was found to increase uniformly when the concentration of acid was between 85 and 98 per cent.: fluorescence of the solution in less than 85 per cent. acid was erratic and in various organic extractants, such as benzene, carbon tetrachloride, dimethylformamide and toluene, was insufficiently intense for determination. Anthrone itself has a strong blue fluorescence in dimethylf ormamide because of the formation of anthranol.When it was necessary to separate the benzanthrone from irrelevant material, it was extracted with benzene, the solvent removed and the residue dissolved in 98 per cent. sulphuric acid. CONCENTRATION OF ACID IN REACTION MIXTURE- The concentration of sulphuric acid and the temperature of the reaction were adjusted to give minimum decomposition of anthrone with the maximum production of benzanthrone in a reasonably short time. Rapid decomposition was caused by 98 per cent. sulphuric acid at temperatures over 100" C, and the rate of formation and the yield of benzanthrone fell rapidly at acid concentrations below 85 per cent. This concentration of sulphuric acid (85 per cent.; sp.gr. 1.783 at 16" C) was, therefore, chosen as the reaction medium.TEMPERATURE OF REACTION- Of the range of temperatures investigated, 120" C was chosen as the most suitable, as the production of the fluorescent compound was rapid, and little decomposition of the anthrone by the sulphuric acid occurred during the time necessary for the maximum production of benzanthrone. The development of a slight red colour in the solution did not interfere with the measurement of fluorescence intensity. TIME OF REACTION- In examining the development of the fluorescence with time of reaction at 120" * 1" C, the maximum fluorescence was obtained between 15 and 20 minutes; a longer reaction time caused a decrease in the yield of fluorescent product, probably because of the destruction of the benzanthrone. The intensity of fluorescence of the blank solution containing no glycerol increased with time after 15 minutes' heating, so that this heating period was standardised at 15 minutes.The volume of the reacting solution was standardised at 1 ml of glycerol solution and 1 ml of 0.1 per cent. w/v anthrone solution, each in sulphuric acid; this amount of anthrone theoretically allows the determination of glycerol up to about 500 pg. For determination of amounts of glycerol less than 2Opg 0.01 per cent. w/v anthrone solutions were found to be suitable. The method can be used to determine glycerol in the concentration range of 10 to 75 pg with an accuracy of f5 pg. The accuracy will be less if charred material is formed by decomposition of readily carbonisable substances present with the glycerol, in which case it is necessary to resort to the extraction technique referred to under Solvents. In the presence of many of the interfering substances the glycerol may be separated from the sample by chr~matography,~~~ and the amount of glycerol determined in the spot.REAGENTS- Anthrone-As supplied by British Drug Houses Ltd. GZyceroLAnalytical-reagent grade. Sulphuric acid, 85 and 98 per cent.-Analytical-reagent grade. Benzene-Analyt ical-reagen t grade. PROCEDURE FOR SOLUTIONS CONTAINING ONLY GLYCEROL- Make up a 0.1 per cent. w/v solution of anthrone in 85 per cent. sulphuric acid. This solution is stable for about 14 days. Transfer, by pipette, 1-ml portions of the anthrone solution into a series of stoppered, 15-ml Pyrex test-tubes, each containing 1 ml of a solution of glycerol in 85 per cent.sulphuric acid in known amounts from 0 to 100 pg. Heat for exactly 15 minutes in an oil-bath at188 LYNE, RADLEY AND TAYLOR: USE OF ANTHRONE IN THE [A%a&St, VOl. 93 120” C . Allow to cool, dilute with 5 ml of 98 per cent. sulphuric acid and thoroughly mix the two layers. Measure the intensity of fluorescence of the solutions at 575 mp and prepare a calibration graph of intensity of fluorescence against amount of glycerol, with a solution containing 1 pg per ml of quinine sulphate to set the instrument sensitivity. The graph should be linear over the range of concentration 5 to 75pg, after which the concentration quenching effect is as shown in Fig. 1. Glycerol, pg per ml Fig. 1. Intensity of fluorescence against original concentration of glycerol Dissolve a suitable amount of the sample in 85 per cent.sulphuric acid, then to 1 ml of this solution add 1 ml of the anthrone solution, heat and dilute, as above, before measur- ing the intensity of fluorescence. By reference to the calibration graph the amount of glycerol can then be determined. It is preferable to prepare the standard solutions at the same time as those containing the sample. PROCEDURE FOR DETERMINATION OF GLYCEROL IN OTHER MATERIALS- Many organic materials, e.g., sugars, are easily carbonised by hot concentrated sulphuric acid and interfere with the anthrone reaction, so that it is usually necessary to effect a quantitative separation of the glycerol from the bulk of the interfering substances, e.g., by extraction with a suitable solvent such as dry acetone, before carrying out the determination.Liquid samples may be evaporated on sand or treated with anhydrous sodium sulphate, and the mass extracted with acetone, which is subsequently removed by evaporation. With oil or fat-containing materials the residue from the acetone extract can be taken up in light petroleum or chloroform and the glycerol recovered from this solution with water. After removal of the solvent used to extract the original material, the glycerol residue may be sufficiently pure to proceed with its determination directly, or it may require further purification and separation by chr~matography.~ When using the “crude” extract we find sometimes that considerable darkening occurs when the samples are heated in the reaction mixture.We have endeavoured to separate the benzanthrone from the darkened solutions that would otherwise interfere with the measure- ments by Tyndall scattering and the consequent variations in the adsorption of light. To overcome this difficulty, dilute the darkened reaction mixture with 6ml of water and add 6 ml of benzene. After thorough shaking, centrifuge the mixture, and then remove the benzene layer with a dropping pipette. Repeat the extraction twice, then evaporate off the benzene on a water-bath and re-dissolve the red - brown solid in 10 ml of concentrated sulphuric acid before measuring the intensity of fluorescence of this solution. A separate calibration graph must be drawn carrying the standards right through the extraction process.Providing gross amounts of easily charred material are not present in the extract we have found that this method gives good results. If gross charring is encountered we have found no solvent that gives a 100 per cent. recovery of benzanthrone, perhaps because it is adsorbed on to the charred particles. Further dilution of the reaction mixture before extraction gives a greater yield of benzanthrone, and neutralisation of the acid with 10 per cent. of sodium hydroxide to give solutions of pH 2,March, 19681 DETERMINATION OF TRACE AMOUNTS OF GLYCEROL 189 7 or 10 seems to have no effect, except that caused by dilution. When this occurs the original extract must be purified or the extraction repeated on the original material with a more suitable solvent.The formation of a sulphonated bemanthronelo may account for the loss in yield of fluorescent product with increased time described above. It may also account for the reddish brown compound produced by evaporation of the benzene. THE EFFECT OF INTERFERING SUBSTANCES IN THE GLYCEROL EXTRACT- The reaction between anthrone and sugars is well known, although the sensitivity of the reaction is not as high as for glycerol. Experiments were carried out to determine the extent of interference with the glycerol (50pg) when the determination was carried out in the presence of much larger amounts (5000pg) of sugars, sorbitol, gelatin and fat, and followed by the benzene extraction of the reaction mixture. The figures in the “blank” column were obtained by carrying out the determination in the absence of glycerol with 5000 pg of the interfering substance.TABLE I EFFECT OF READILY CARBONISABLE MATERIAL ON GLYCEROL DETERMINATION Apparent amount Substance added, of glycerol, 5000 pg I-% Dextrose . . .. 54 Fructose . . .. 140 Sucrose . . .. 97 Gelatin . . .. 0 Sorbitol . . .. 46 Blank (interfering substances only), CLg 16 100 45 0 0 Amount actually found, Percentage Icg recovery 38 ‘76 40 80 52 104 0 0 45 90 Although gelatin completely quenched the fluorescence, glycerol can be completely extracted from it with acetone, and the anthrone reaction carried out on the residue after removal of the solvent. Other substances mentioned by Radley5 as giving fluorescent products have been briefly studied under similar reaction conditions. Formaldehyde produces a compound with a single fluorescent peak at 585mp when heated at 120” C for 15 minutes.Ethylene glycol forms a green fluorescent product at a higher temperature with a peak at 540mp. Acetone, in about l-mg amounts, produces a yellow - green fluorescent compound that is fluorescent in daylight with peaks at 454 and 580mp. Tartaric acid reacts with the anthrone solution, when heated at 150” C for 5 minutes, to form a green fluorescent product. If this is diluted with 60 per cent. sulphuric acid (5 + 1) the product can be shown to re-arrange photochemically to give a dark brown solution, which has a blue fluorescence (485 mp) after a period of about 30 seconds, with a high pressure mercury lamp as the ultraviolet source, and the green fluorescence reported by Radley5 is apparent during the first few seconds of observation. Unlike the sugars, which, as shown in Table I, only interfere significantly when present in 100-fold amounts, tartaric acid, formaldehyde, ethylene glycol and acetone cause far more serious interference.Tartaric acid and formaldehyde, if present in 10-fold amounts, suppress the fluorescence of the benzanthrone to an insignificant level, ethylene glycol to a lesser but still significant extent, and acetone enhances the fluorescence. In the presence of these substances recourse would have to be made to c h r o m a t ~ g r a p h y . ~ ~ ~ ~ ~ The reaction mechanism may indicate other compounds that might react to form benzanthrones. Further work is in progress to investigate the reaction conditions necessary for production of fluorescent products with anthrone, and also to examine the relationship and the original concentration of the compound.190 LYNE, RADLEY AND TAYLOR APPLICATIONS The reaction can be used to determine milligram amounts of glycerol such as are found in spots obtained by paper chromatography.Further work is being carried out to apply the method to the determination of small amounts of glycerol in fermented beverages, rancid fats and in the metabolic products of moulds and yeasts, as well as in a variety of other substances when the amount of sample available is restricted or the amount of glycerol is small. It is also being examined for the determination of glycerol in drugs6 in forensic work. The analysis of some duplicating inks and writing papers that had been in contact with hectograph “beds” was used qualitatively many years ago by J. A. Radley. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. REFERENCES Miner, C. S., and Dalton, N. N., Editors, “Glycerol,” American Chemical Society Monograph Series No. 117, Reinhold Publishing Corporation, New York; Chapman & Hall, Ltd., London, 1953, p. 201. Zeisel, S., and Fanto, R., 2. analyt. Chem., 1903, 42, 549 and 579. Raveux, R., Annls. Chim. Analyt., 1943, 25, 70 and 95. Schutz, F., Papierfabrikant, 1938, 36, 55. Radley, J. A., J . Sci. Fd Agric., 1950, 1, 222. Allen, C. F. H., and Overbaugh, S. C., J . Amer. Chem. SOC., 1935, 57, 1322. Hough, L., Nature, 1950, 165, 400. Buchanan, S. G., Dekker, C. A., and Long, A. G., J . Chem. SOC., 1950, 3162. Sporek, K., and Williams, A. F., Analyst, 1954, 79, 63. Pritchard, R. R., and Simonsen, J. L., J . Chem. SOL, 1938, 2047. First received November llth, 1964 Amended November 23rd, 1967