Analyst, February, 1967, Vol. 92, pp. 91-97 9L The Determination of Dicumyl Peroxide in Polystyrene Materials BY J. A. BRAMMER, S. FROST AND V. W. REID (“Shell” Research Limited, Central Laboratories, Egham, Surrey) A method has been developed for the determination of dicumyl peroxide in polystyrene plastic materials that may contain other organic peroxides. The dicumyl peroxide is extracted from the plastic with acetone and separated from the other additives present by thin-layer chromatography on silica gel. The silica gel in the area that contains the dicumyl peroxide is transferred to a small reaction flask and the peroxide determined by a micro-titration procedure. Formulations containing 0-25 to 0-5 per cent. w/w of dicumyl peroxide have been analysed by this method with a precision of f 12 per cent.of the determined value. Other organic peroxides commonly used in polystyrene formulations do not interfere. DICUMYL peroxide (DCP), which imparts a considerable degree of fire resistance to self- extinguishing grades of polystyrene and is often included in fire-resistant formulations, has the structural formula- CH, CH3 I I CH3-C-0-0-C-CH3 I I To maintain good self-extinguishing properties it is important that the DCP content of the product should not be materially reduced below the original level, either by the manu- facturing process or by subsequent conditions of usage. To investigate the stability of DCP in commercial products during processing and storage, a method for its determination in polystyrene was required. The method should be capable of application in the presence of other organic peroxides used in the formulation.Dicumyl peroxide is one of the most unreactive of the organic peroxides, and is not readily reduced by chemical or electrochemical methods. The only analytical method we have noted that has a specific application to DCP is the assay method used by the Hercules Powder C0mpany.l This is a titrimetric procedure, similar to that published by Mair and Graupner,2 and depends upon the liberation of iodine when refluxed with a sodium iodide - glacial acetic acid mixture. This procedure would be open to interference from other peroxides, however, which would also liberate iodine under the conditions used. It was expected that low molecular weight polystyrene, which is always present to some extent in polystyrene products, would also interfere.Separation of the DCP from the other additives would thus be a necessary pre-requisite to the determination. Thin-layer chromatography was used as a separation procedure, and a micro-titration procedure was then developed to determine the small amounts of DCP separated by this technique. EXPERIMENTAL DEVELOPMENT OF A SEPARATION PROCEDURE- Knappe and Peteri3 have published a method for the separation and detection of a number of organic peroxides by thin-layer chromatography. They used siIica gel as substrate and developed the chromatogram with a toluene - carbon tetrachloride solvent. Peroxide spots were detected by spraying with a solution of NN‘-dimethyl #-phenylenediamine dih ydrochloride.92 BRAMMER, FROST AND REID: DETERMINATION OF [A?th?bSt, Vol.92 This method was examined with respect to the separation of the peroxides that may be present in self-extinguishing grades of polystyrene, for example, dicumyl peroxide, t-butyl perbenzoate, benzoyl peroxide and cumene hydroperoxide (associated with the dicumyl peroxide). Solutions of these peroxides in carbon tetrachloride were applied to a thin-layer plate coated with silica gel, and developed according to the Knappe and Peteri method. Application of the spray reagent to the dried plate failed to reveal the position of the DCP, however, although the other peroxides were rendered plainly visible. The same separation procedure was carried out on plates coated with silica gel which embodied a fluorescent indicator, Merck Kieselgel GF254, and the dried plates were viewed by ultraviolet light at a wavelength of 254 mp.The DCP, t-butyl perbenzoate and benzoyl peroxide appeared as dark spots on a green fluorescent background. Only the cumene hydroperoxide spots remained invisible. By applying both methods of rendering the spots visible, a good separation was achieved for each of the compounds, as shown in Table I. TABLE I DETECTION OF PEROXIDES AFTER SEPARATION BY THIN-LAYER CHROMATOGRAPHY Visual indication r 1 NN’-Dimethyl GF254 plates Additive p-phenylenediamine and ultraviolet light RF values Dicumyl peroxide .. .. negative positive 0.55 t-Butyl perbenzoate . . .. positive positive 0.15 Benzoyl peroxide .. .. positive positive 0.30 Cumene hydroperoxide .. . . positive negative 0.05 Having established the conditions for separating the DCP from the other peroxide additives, the procedure was applied to the analysis of a sample of expanded polystyrene board containing 0.2 per cent. w/w of DCP. A solution of this sample in chloroform was applied to a prepared fluorescent plate, which was developed as before. Examination of the plate by ultraviolet light revealed that the polystyrene present in the solution had streaked badly and had obliterated most of the peroxide spots. A procedure was required, therefore, to enable the additives to be extracted from the polystyrene before separation of the DCP from the other peroxides. EXTRACTION OF THE ADDITIVES FROM THE POLYSTYRENE- A 25-g sample of the same polystyrene containing 0.2 per cent.w/w of DCP was extracted several times with 50-ml portions of acetone, and the separate extracts were reduced in volume to 10 ml. Aliquots of each extract were spotted on to a plate coated with the fluorescent silica- gel adsorbent, and the plate was developed and inspected under ultraviolet light. By com- paring the spots produced with those resulting from known weights of DCP, it was possible to estimate the amount of peroxide in each extract. The chromatogram obtained showed no visible DCP spot on the eighth extract. It was shown that if 1 per cent. of the original DCP had remained in this extract, then a visible spot would have appeared. A minimum of eight extractions was thus used in the developed procedure. Some polystyrene was visible near the starting line on this plate, but the streaking effect had been sufficiently reduced by the pre-extraction procedure to ensure that the small amount of polystyrene remaining in the acetone extract would not interfere with the DCP determination.DETERMINATION OF THE SEPARATED DICUMYL PEROXIDE- In view of the relatively small amounts of material that can be effectively separated by thin-layer chromatography, an attempt was made to determine the isolated DCP by a colorimetric procedure. Trials were made with two methods for the colorimetric determination of organic peroxides. In the first method, Vioque and Vioque4 used NN’-dimethyl p-phenylene- diamine dihydrochloride as the colour producing agent. In the second method (Eiss and Giesecke5) zirconium naphthenate and benzoyl leuco methylene blue were used. ’These reagents failed to produce a colour with DCP.Titrimetric procedures based on the liberation of iodine were then examined, particularly the procedures used by Mair and Graupner and the Hercules Powder Co. In these proceduresFebruary, 19671 DICUMYL PEROXIDE IN POLYSTYRENE MATERIALS 93 the sample is refluxed with a glacial acetic acid solution of sodium iodide, and the liberated iodine titrated with sodium thiosulphate solution. These titrimetric methods were designed for the determination of about 0.3 g of DCP, so that it was necessary to use a modified pro- cedure in order to determine the smaller amounts of peroxide separated by thin-layer chromatography. The volume of the apparatus was scaled down by a factor of ten, and an Agla micrometer syringe was used as the titration burette.This micro-scale procedure proved to be satisfactory after eliminating certain causes of high blank values. For example, it was found that the xylene, which was used as a solvent in the preparation of standard DCP solutions, contained trace amounts of peroxide impurities. These impurities could be removed by passing the solvent through an activated alumina column, as proposed by Dasler and Bauer.6 Also, the nitrogen used to displace air from the flask while refluxing and titrating contained oil and other impurities that liberated iodine. These were removed by passing the nitrogen through a tube packed with molecular sieves and cotton-wool. The micro-titration procedure was then applied to synthetic solutions of DCP in xylene that contained between 0.4 and 1.2-mg amounts of the peroxide.The recoveries obtained are shown in Table 11, where it may be seen that good recovery of peroxide, together with a reasonable degree of reproducibility, is obtained when the amount of DCP titrated is about 1 mg. A t this level, the blank values normally obtained were in the region of 3 to 5 per cent. of the sample titre. The method published by the Hercules Powder Co. introduces a correction factor based on the assumption that there is 93 per cent. reaction of dicumyl peroxide with sodium iodide. This factor was not applied in calculating the results shown in Table 11, and has not been included in the final procedure. TABLE I1 MICRO-TITRATION OF DCP WITH SODIUM THIOSULPHATE DCP present, mg 1.205 1.205 1-205 1-174 1.174 1.174 0.782 0.782 0.782 0.391 0.391 0.391 0.02 N Thiosulphate titre (less blank), 410 425 420 430 420 425 270 280 275 145 130 140 P*.1 DCP found, DCP recovery, mg per cent.1.11 92 1.15 95 1.14 95 1.16 99 1.14 97 1.15 98 0.73 93 0-76 97 0.74 95 0.39 100 0.35 90 0.38 97 We then examined the possibility of transferring the adsorbent containing the DCP spot to the reaction flask and subsequently titrating it by the micro-procedure. The silica-gel adsorbent gave negligible blank values when the procedure was applied in the absence of sample. We then applied known solutions of DCP in xylene to thin-layer chromatographic plates coated with Merck Kieselgel GF,,, adsorbent, and the chromatograms were developed with toluene - carbon tetrachloride solvent as before.When the plates were dried, the DCP zones were marked out under ultraviolet light and the silica gel from each zone was quanti- tatively transferred to separate reaction flasks. The peroxide content of each zone was then determined by the micro-titration procedure. The results obtained are shown in Table 111, which includes values for the amount of sample and the lengths of the sample stripes applied to the plate. At the 0.6-mg level, 12 pl of DCP solution applied evenly along the starting line for a distance of 5 cm gave good recovery of the peroxide; 1.2 mg distributed along a 7.5-cm length gave recoveries of about 80 per cent. When the amount of DCP separated was about 1 mg, and the sample was applied to a 15-cm length of the starting line, good recoveries were obtained.This last procedure was therefore adopted.94 BRAMMER, FROST AND REID: DETERMINATION OF [Analyst, Vol. 92 DCP solution (strength, normal), mg per ml 50 50 50 50 50 5 5 5 5 5 TABLE I11 RECOVERY OF DCP FROM THIN-LAYER CHROMATOGRAPHIC PLATES Volume of solution on TLC plate, 12 12 12 25 25 300 300 300 200 200 P1 Length of sample stripe on TLC plate, cm 5 5 5 7.5 7-5 15 15 15 15 15 Sodium Weight of thiosulphate DCP present, titre (less blank), mg E.1 0.578 220 0.578 215 0.578 205 1.205 335 1.205 345 1.174 420 1.174 425 1.174 415 0.782 285 0.782 300 DCP recovery, per cent. 103 101 96 75 78 97 98 96 99 104 The extraction, separation and titration steps were then combined into one procedure, which was applied to samples of polystyrene products in the form of beads and expanded boards. METHOD APPARATUS- Thin-layer chromatography (TLC) equipment-This consists of TLC plates, 20 x 20 cm, plate coating equipment (any commercially available equipment is suitable), drying oven, desiccator, chromatographic tank capable of holding the TLC plates and a TLC spotting template.Ultraviolet lamp-The lamp used should give maximum emission at 254 mp. Reaction and titration equipment-This is assembled, together with the nitrogen purifica- Micro-reaction flask-This consists of a 25-ml round-bottomed flask with B 14/23 neck, Agla micrometer syringe-This is fitted with a glass capillary delivery tube and mounted Magnetic stirrer-This is fitted with a plastic-coated stirrer bar, 15mm in length. tion train and micro-reaction flask, as shown in Fig.1. fitted with a capillary side-arm, as shown in Fig. 1. on a retort stand. REAGENTS- All reagents should be of analytical-reagent grade, unless otherwise stated. Kieselgel GF,,, (Merck)-TLC silica gel with binder and fluorescent indicator. A cetone. Toluene. Carbon tetrachloride. Dicumyl peroxide-Technical grade (obtainable from Hercules Powder Co., Hercules Acetic acid, glacial. Sodium iodide. Sodium thiosulphate solution, 0.02 N-Prepare a 0.1 N solution of sodium thiosulphate. Tower, Wilmington 99, U.S.A.). Dilute 20 ml of this solution to 100 ml. PROCEDURE- Preparation of TLC equipment-Coat a batch of 20 x 20-cm TLC plates with Kieselgel GF,,, to give a layer thickness of 0-25 mm.Allow the layer to set, then dry the plates at 110" C for 30 minutes. Store the prepared plates in a desiccator. Prepare the developing solvent by mixing two volumes of toluene with one volume of carbon tetrachloride. Pour sufficient of the mixed solvent into a TLC tank to give a depth of liquid of about a & inch. Line the inside of the tank with filter-paper, replace the lid and allow the tank to stand for at least 4 hours to enable the atmosphere to become saturated with solvent vapour.February, 19671 DICUMYL PEROXIDE IN POLYSTYRENE MATERIALS 95 EXTRACTION OF THE ADDITIVES- sample as indicated in Table IV. (;) Treatment of expanded boards-Weigh to the nearest 0.1 g a suitable amount of TABLE IV OPTIMUM SAMPLE WEIGHTS FOR LEVELS OF DCP UP TO 1 PER CENT.DCP content, Sample weight, per cent. g 0 to 0.25 25 0.25 to 0.5 10 0.5 to 1.0 5 Before weighing cut the sample into pieces 3 x 1 x 1 inches. Introduce the weighed pieces, one at a time, into a 250-ml beaker containing about 100 ml of acetone. The pieces will immediately collapse to form a rubber-like mass on making contact with the acetone. Pummel the mass for 5 minutes with a glass rod flattened at the end, and then decant as much of the acetone as possible into a clean 100-ml beaker. Evaporate the acetone solution to about 10 ml. Add a further 30 to 40 ml of acetone to the polystyrene sample, repeat the extraction procedure and decant the excess of acetone into the same 100-ml beaker. Carry out a total of eight extractions in this way, reducing the acetone volume in the beaker to 10 ml each time.Finally, reduce the volume of acetone solution to less than 10 ml, quantitatively transfer this solution to a 10-ml calibrated flask and make up to the mark with acetone. After mixing the solution, allow the flask to stand until any emulsion formed has subsided. The solution is then ready to be applied to the TLC plate. (ii) Tyeatment of polystyrene beads-Weigh a suitable amount of sample (Table IV) into a 250-ml beaker and add 100 ml of acetone. Pummel the beads until the sample has com- pletely softened to a rubbery mass. Carry out eight extractions with acetone to produce a 10-ml volume of acetone solution as described above. CHROMATOGRAPHIC SEPARATION OF DCP- Take a prepared TLC plate and apply 0-2 ml of the acetone extract along a starting line, 2.5 cm from the bottom edge of the plate, either by applying a series of small (5 p1) drops, or by using sample striping equipment.Whichever method of application is used, it is essential that the correct volume of solution be applied as uniformly as possible along a 15-cm length of the starting line. Parallel with the starting line, and 15 cm away from it, draw a line through the adsorbent to mark the limit of travel of the developing solvent. Allow the acetone solvent to evaporate from the plate, and then develop the chromatogram in the pre-saturated tank. This process should take about 45 minutes. Remove the plate from the tank, transfer it to a fume cupboard and completely evaporate the solvent in a stream of dry nitrogen. Place the plate under an ultraviolet lamp that is emitting light at 254 mp, The DCP will then be visible as a dark zone against the green fluorescent background, and will normally be located nearest the solvent front.Should any difficulty be experienced in locating the correct zone, a solution of DCP in acetone may be used as a marker. Draw round the DCP zone, leaving sufficient margin to include adsorbent that may contain traces of DCP not visible under the ultraviolet light. Transfer quantitatively the adsorbent within the marked area to a small test-tube, in readiness for the determination of the DCP. DETERMINATION OF THE SEPARATED DCP- Transfer by pipette 5 ml of glacial acetic acid into the micro-reaction flask and add a few anti-bumping granules. Connect the nitrogen line to the side-arm and adjust the flow of nitrogen so that a slight dimple is produced on the surface of the liquid.Assemble the reaction apparatus as shown in Fig. 1 and pass a steady flow of tap water through the condenser jacket. Increase the nitrogen flow to a steady stream as indicated by the water bubbler, and then heat the flask contents to boiling. Reflux for 5 minutes to remove dissolved air.BRAMMER, FROST AND REID: DETERMINATION OF [AutabSt, VOl. 92 A = Molecular sieves B = Cotton-wool C = PVC tubing D = Capillary side-arm, 6mm o.d., Imm i.d. E = Condenser (I50mm) F = Reaction flask, 25:ml (Quickfit and Quartz Ltd., FR 25/15) = Water Fig. 1. Apparatus for the reaction of DCP with sodium iodide After refluxing, increase the nitrogen flow further, and cool the flask by immersing it in water.Ensure that the gas flow-rate is sufficient to prevent air being sucked back into the flask. Remove the flask from the condenser, add 0.6g of sodium iodide and swirl the flask to dissolve it. Transfer by pipette 0.3 ml of distilled water into the flask and mix the liquid phases. Momentarily turn off the nitrogen supply and quickly pour the silica gel powder, on which the DCP sample is adsorbed, into the flask. Again attach the flask to the condenser and reflux the contents for 15 minutes, with a steady nitrogen flow. Cool the flask contents as before, then remove the flask from the condenser, add 10 ml of water and introduce a small magnetic-stirrer bar. Clamp the flask over a magnetic stirrer and adjust the stirring-rate so that a shallow vortex is produced.By using an Agla syringe burette titrate the contents of the flask with 0.02 N sodium thio- sulphate to a colourless end-point. Carry out a blank titration as described above with 5 ml of glacial acetic acid, 0.6 g of sodium iodide and 0.3 ml of water. Do not include silica gel in the blank titration. CALCULATION- The percentage by weight of DCP in the sample is given by- (Ts - TB) x N x 270.4 x 100 - (Ts - T B ) x N x 676 - 2 x 1000 x 0.02 x w W where Ts = millilitres of sodium thiosulphate solution used for the sample, TB = millilitres of sodium thiosulphate solution used for the blank, W = the weight in grams of polystyrene taken, and N = the normality of the sodium thiosulphate solution. RESULTS The procedure has been applied to samples of polystyrene in expanded board and in bead form.The results obtained on ten determinations carried out on the board, which had been prepared by the addition of 0-5 per cent. w/w of DCP* during manufacture, are shown in Table V (sample No. 1). * The technical DCP used contained 98 per cent w/w of dicumyl peroxide.February, 19671 DICUMYL PEROXIDE IN POLYSTYRENE MATERIALS TABLE V DETERMINATION OF DCP IN EXPANDED POLYSTYRENE BOARD { 3 0.5 0.25 0.25 12 Sample DCP added, Sample weight, DCP found, per cent. w/w 0.43 0.51 0.46 0.47 0.47 0.44 0.51 0.45 0-45 0.47 Mean . . 0.47 0.20 0.20 25 0-17 0.20 Mean . . 0.19 0.19 25 0.18 No. Der cent. w/w g Mean . . 0.185 97 These results are reasonabl 7 reprodi cible and close to the known addition, the average result being 0.47 per cent.with a spread of k0-04 per cent. The standard deviation is 0.03 per cent., which represents 95 per cent. confidence limits of some k12 per cent. of the JXP content in single determinations. The reproducibility at the 0-25 per cent. w/w level (sarr ples Nos. 2 and 3) is of a similar order. The results obtained at similar DCP levels, when the sample was in the form of :mall polystyrene beads, are shown in Table VI. Here again, satisfactory agreement was obta ined, the DCP content determined being of the same order as the amount of DCP added i i the bead formulation. TABLE VI DETERMINATION OF DCP IN POLYSTYRENE BEADS Sample DCP added, Sample weight, DCP found, per cent. w/w g per cent. w/w 0.43 0.46 4 10 0.44 0.48 Mean . . 0.45 No’ { 0.5 0.20 0.19 0.20 I Mean . . 0.20 The results obtained on both board and bead samples are slightly lower than the amount of DCP added during manufacture; this is to be expected as some DCP would be lost during processing. CONCLUSION 5 [ 0.25 25 0.20 A method has been described for the determination of dicumyl peroxide in polystyrene, either in the form of expanded boards or beads. The reproducibility of the method is of the order of k12 per cent. of the determined value at the 0.25 and 0-5 per cent. w/w level. Good recovery of dicumyl peroxide has been obtained in the analysis of formulated products. REFERENCES 1. 2. 3. 4. 5. 6. “Assay Analyses for Dicumyl Peroxide,” Technical Data Bulletin PR-107, Hercules Powder Co., Mair, R. D., and Graupner, A. J., Analyt. Chein., 1964, 36, 194. Knappe, E., and Peteri, D., 2. analyt. Chem., 1962, 190, 4, 386. Vioque, A., and Vioque, E., Grasas Aceit., 1962, 13, 5, 203. Eiss, M. l., and Giesecke, P., Analyt. Chem., 1959, 31, 1558. Dasler, W., and Bauer, C. D., I n d . Engng Chem. Analyt. Edn, 1946, 18, 52. Hercules Tower, Wilmington 99, Delaware, U.S.A. Received June 30th, 1966