摘要:

JULY, 1967 THE AhTALYST Vol. 92, No. 1096 An Indirect Sequential Determination of Phosphorus and Silicon by Atomic-absorption Spectrophotometry BY G. F. KIRKBRIGHT, A. M. SMITH AND T. S. WEST (Chemistry Defiavtment, ImGerial College, London, S. W.7) Phosphorus and silicon are determined sequentially as phosphate and silicate in a single aliquot of solution by an amplification procedure. Phospho- molybdic acid, H,PO,(MoO,),,, is formed and selectively extracted away from silicate and excess of molybdate reagent into isobutyl acetate. After adjust- ment of the acidity of the remaining silicate solution, silicomolybdic acid, H,SiO,(MoO,) ,,, is extracted into butanol. The twelve molybdate ions associated with each phosphate and silicate ion are determined by direct atomic-absorption spectroscopy in the isobubtyl acetate and butanol phases in a nitrous oxide - acetylene flame a t 3132 A.With this procedure 0.08 to 1 p.p.m. of phosphorus, and 0.08 to 1.2 p.p.m. of silicon can be determined in the same solution. The most suitable conditions for the determination have been established, and a study of the effect of other cations and anions is presented. MOST methods for determining traces of phosphorus and silicon are based on the absorption spectrophotometry in solution of orthophosphate and silicate as yellow phosphomolybdic and silicomolybdic acids, or as their blue reduction products in aqueous or organic media. These methods are neither very sensitive nor selective, although selectivity is frequently obtained by judicious choice of solvent for extraction of the heteropoly a~id19~J or use of reagents such as tartrate or oxalate to inhibit formation of the heteropolymolybdate of one of the elements present.4 Increasing use, however, is being made of the considerably more sensitive phosphovanadomolybdate method, especially in the metallurgical i n d ~ s t r y .~ More recently Djurkin, Kirkbright and West6 have published an even more sensitive and selective procedure for determining phosphorus by an amplification procedure. In this procedure phosphorus is converted to phosphomolybdate and selectively extracted into chloroform - butanol away from the excess of molybdate reagent. It is then broken down by equilibration with alkali, and the twelve molybdate ions associated with each original phosphate ion are determined by their spectrophotometric reaction with 2-amino-4-clilorobenzenethiol. This method is sensitive and selective, and it is possible to devise a similar amplification procedure for the determination of silicon, or both phosphorus and silicon sequentially in the same sample solution.Determination of the molybdenum in these procedures by solution spectro- photometry, however, extends the number of operations required so that the method becomes tedious and time consuming. This is particularly true when both phosphorus and silicon are determined sequentially and the sensitive 2-amino-4-chlorobenzenethiol method is used to determine the molybdenum. We have, therefore, devised a more rapid procedure for deter- mining phosphate and silicate in the same sample, in which the molybdenum is determined directly in the organic phases containing the phosphomolybdic or silicomolybdic acids by atomic-absorption spectrophotometry. In the procedure described in this paper samples containing both phosphate and silicate are treated with excess of molybdate reagent in acid medium, and the phosphomolybdic acid is extracted away from excess of reagent and silicate into isobutyl acetate.The molybdate associated with the phosphate is then determined by direct atomic-absorption spectrophotometry of the molybdenum in the organic phase in a nitrous oxide - acetylene flame at 3132 A. The residual aqueous phase from the phosphate 41 1412 [Aqzalj’st, Vol. 92 determination, which contains the silicate, is treated with ammonia solution to lower the acidity and the silicomolybdic acid is extracted into butanol.This organic phase is then washed free of excess of reagent and its molybdenum content determined in exactly the same way. This procedure is less sensitive than when the molecular spectrophotometric finish6 to the determination is used. This is due entirely to the less sensitive determination of the molybdenum by the atomic-absorption method. The ease of operation and rapidity of the procedure, however, compensate for the loss of sensitivity, and the sensitivity can be increased, if necessary, by suitable adjustment of the phase volumes in the extraction procedure. The method loses nothing in selectivity by the atomic-absorption method ; when foreign ions extractable into isobutyl acetate or butanol from dilute hydrochloric acid are present, a gain in selectivity is obtained as these may interfere with the molecular spectrophotometric determination in solution. KIRKBRIGHT et al.: INDIRECT SEQUENTIAL DETERMINATION OF EXPERIMENTAL APPARATUS- The equipment used in this work was the Unicam SP9OOA flame spectrophotometer with a Unicam molybdenum hollow-cathode lamp. The instrument conditions used through- out were: slit width, 0.06 mm; wavelength, 3132 A; lamp current, 15 mA. The burner for atomic-absorption measurements in air - acetylene was replaced by a brass burner of similar dimensions for the nitrous oxide - acetylene flame. The burner, which has a slit 2 inches long and 0.015 inch wide, has already been described in a previous communication.7 The sample was aspirated directly with nitrous oxide at 15 p.s.i., and sufficient acetylene was pre-mixed with the nitrous oxide in the burner base to produce a flame with a red-€eather of about 1 to 2 cm in height.The burner height was adjusted so that as much as possible of the light from the hollow-cathode lamp passed through this red zone. REAGENTS- hydrate, (NH,),Mo,O,~.~H,O, in distilled water and dilute to exactly 1 litre. KH,PO,, (AnalaR grade) in distilled water and dilute to 1 litre. 1 ml of solution contains 25 pg of phosphorus. Ammonium molybdate solution-Dissolve 10.69 g of AnalaR ammonium molybdate tetra- Standard phosphorus solution-Dissolve 0.1098 g of potassium dihydrogen orthopliosphate, Standard silicon solution-Fuse 0.3385 g of powdered sodium silicate with about 3 g of sodium hydroxide pellets (AnalaR grade) in a platinum crucible, and leach the melt with hot water.Allow the solution to cool, and dilute it to 1 litre in a calibrated flask. This solution was standardised gravimetrically by the hydrofluoric acid volatilisation procedure. 1 ml of solution contains 100 pg of silicon. Store the alkaline stock solution in an opaque polythene bottle. Solutions containing 40 pg per ml of silicon were prepared by neutralisation of this solution and dilution to volume as required, and these were also stored in polythene bottles. Ammonia solution, auzalytical-reagent grade-Supplied by May and Baker, Proanalys. Hydrochloric acid, analytical-Yeagent grade-Supplied by May and Baker, Proanalys.The hydrochloric acid and ammonia solution used were chosen for their low silicon content (not more than 0.0001 per cent. of 50,). All other reagents used were of analytical- reagent grade. Wash liquid for butanol phase-This is dilute hydrochloric acid, 0.15 M, saturated with butanol, and is stored in polythene bottles. CALIBRATION GRAPHS FOR PHOSPHORUS AND SILICON- To a series of six 100-ml separating funnels add 10 ml of molybdate reagent, 10 ml of distilled water and sufficient concentrated hydrochloric acid to make the solution 0.96 M with respect to hydrochloric acid. Add 0, 0.2, 0.4, 0.6, 0.8 and 1-0-ml aliquots of the standard phosphorus solution and 0, 0.1, 0-2, 0-3, 0.4 and 0.6-ml aliquots of standard silicon solution to the funnels. Mix the solutions and allow them to stand for 5 minutes.Add 10ml of isobutyl acetate, shake the mixture for 1 minute, and allow the phases to separate. Transfer the lower aqueous phases ( A ) to a second series of 100-ml separating funnels for the subsequentJuly, 19671 PHOSPHORUS AND SILICON BY ATOMIC-ABSORPTION SPECTROPHOTOXETRY 413 silicon determination. Wash the isobutyl acetate phases by shaking the solution for about 30 seconds with 10 ml of 2 M hydrochloric acid. Discard the lower aqueous phases and aspirate the isobutyl acetate solutions into the nitrous oxide - acetylene flame for the deter- mination of the molybdenum by atomic absorption at 3132 A7 against a solvent blank of isobutyl acetate. To the aqueous phases ( A ) containing the silicon, add 4ml of 4~ ammonia solution and allow the mixture to stand for 15 minutes to ensure complete formation of the silico- molybdic acid.Introduce 10ml of butanol into each funnel and shake them for 1 minute. Discard the lower aqueous phases, and wash the organic phase three times with 10 ml of the prepared butanol-wash solution. Spray the butanol phases directly into the nitrous oxide - acetylene flame and determine the molybdenum as before against a butanol solvent blank. When more than 25 pg of silicon are present with 5 to 25 pg of phosphorus, the above procedure should be modified by using a larger volume (15 to 20 ml or more) of butanol to enable the silicon concentration in the organic phase to be kept within the range of the calibration curve. When more than 25 pg of phosphorus are present with 2 to 20 pg of silicon, the procedure should be changed to enable this greater weight of phosphorus to be extracted quantitatively, and also to keep the concentration of phosphorus in the organic phase within the range of the calibration curve.This is easily accomplished by the use of a larger volume of isobutyl acetate. Thus for the determination of 20pg of silicon and 2OOpg of phosphorus, two extractions of the phosphomolybdic acid with 20 ml of isobutyl acetate should be made. RESULTS The optimum conditions for the determination of molybdenum by atomic-absorption spectrophotometry in the nitrous oxide - acetylene flame have already been established and described el~ewhere,~ and were therefore not studied here. The optimum conditions for the selective extraction of phosphomolybdic acid into isobutyl acetate were investigated, together with an examination of the requirements for the solvent extraction and determination of silicon by the amplification procedure.EXTRACTIOK OF PHOSPHOMOLYBDIC ACID- The extraction of phosphomolybdic acid from aqueous solution was conducted under similar conditions of hydrochloric acid concentration (0.96 M) and molybdate reagent concen- tration (0.0024 M) to those used by Wadelin and Mellon,l but isobutyl acetate was used in place of a chloroform - butanol mixture as the extractant. Repeated extractions were made to establish that quantitative recovery of at least 50 pg of phosphorus as the phosphomolybdic acid could be achieved with 10 ml of isobutyl acetate from 25 ml of aqueous solution. Blank experiments, with no phosphorus present, revealed that no appreciable amount of the molybdate reagent is extracted into isobutyl acetate.In the procedure finally adopted, however, the isobutyl acetate phase is washed with 10 ml of 2 M hydrochloric acid before aspiration into the flame for atomic-absorption measurements. This washing procedure serves to remove from the isobutyl acetate phase any molybdate reagent which is mechanically transferred from the aqueous phase because of incomplete separation of the solvent from the molybdate solution. Isobutyl acetate retains the selectivity of the chloroform - butanol mixture that was used previously in the spectrophotometric procedure,6 as is shown from the results of the study of interferences described below. It has the advantage that it may be sprayed directly into the nitrous oxide - acetylene flame whereas the chloroform - butanol solvent cannot, and this leads to an increase in sensitivity due to the greater efficiency of nebulisation of organic solvents relative to aqueous solutions.The extraction procedure based on isobutyl acetate also has the advantage of rapidity, as only one extraction and wash are required before the final determination, and as it forms the top phase, the sequential procedure does not require as many transfer operations as when chloroform - butanol is used. EXTRACTION OF SILICOMOLYBDIC ACID- In an aqueous silicate solution we have found that the optimum hydrochloric acid concentration for the formation and extraction of silicomolybdic acid, H,SiO,(MoO,) 12, into butanol is 0.15 M when the total molybdate reagent concentration is 0-024 M.In the sequential414 [APzaZyst, Vol. 92 determination of phosphorus and silicon, therefore, when the phosphomolybdic acid is formed and extracted at 0 . 9 6 ~ hydrochloric acid, it is necessary to lower the acidity after the phosphorus extraction. Slightly less than the calculated a.mount of 4 M ammonia solution is used to lower the acidity to 0.15 M, because a small amount of hydrochloric acid is lost during the separation of the phosphorus, probably by extraction into the isobutyl acetate phase. The optimum volume of ammonia solution required was established experimentally by adding 4 M ammonia solution to the aqueous phase containing silicon before its extraction, and after it had been taken through the phosphorus-extraction procedure.As shown in Fig. 1, 4 ml of 4 M ammonia solution are required for maximum absorbance in the butanol phase after extraction. KIRKBRIGHT et al. : INDIRECT SEQUENTIAL DETERMINATION OF D, 51 n 6 0.1 - 4M Ammonia solution, ml Fig. 1. Effect on net absorbance for 20pg of silicon of volume of ammonia solution used to adjust acidity before silicon extraction I I I I I j 0.10 0.20 031 Molarity of hydrochloric acid - butanol washings Fig. 2. Effect on net absorbance for 16 pg of silicon of molarity of hydrochloric acid used to wash the butanol phase We have established that a single extraction with one 10-ml aliquot of butanol quantita- tively extracts up to 50 pg of silicon as silicomolybdic acid.An appreciable amount of free molybdate reagent is also extracted with the silicomolybdic acid into the butanol phase, and this gives rise to high blanks. It was found that to remove this molybdate preferentially without removing the silicomolybdic acid, it is necessary to wash the butanol phase with an acid solution. Fig. 2 shows the effect of the acidity of the wash solution on the molybdenum absorbance obtained for 16 pg of silicon extracted from 0.15 M hydrochloric acid. It is apparent that 0-15 M hydrochloric acid is a suitable wash solution for removal of the molybdate. A single back-wash with 0.15 M hydrochloric acid is insufficient, and we have established experimentally (Fig. 3) that three rapid washes with 10-ml aliquots of 0 .1 5 ~ hydrochloric acid are sufficient. Because of the miscibility of butanol with water, the butanol phase would be depleted seriously during this washing procedure. We, therefore, adopted the use of a 0.15 M hydrochloric acid wash solution which had been pre-saturated with butanol. This ensures that the butanol phase is maintained at 10ml. Even when the established optimum conditions are adopted for the extraction of the silicomolybdic acid into butanol and the back-washing procedure, a significant blank is always obtained for silicon. This is caused by the silicon content of the reagents used. Thus it is difficult, even when the best available grades of hydrochloric acid and ammonia solution are used, to avoid the introduction of less than about 10 pg of silicon during a determination.This blank is reproducible, however, and the molybdenum absorbance arising from it may be subtracted from the absorbance produced by the standards taken through the procedure.July, 19671 PHOSPHORUS AND SILICON BY ATOMIC-ABSORPTION SPECTROPHOTOMETRY 415 CALIBRATION GRAPH AND OPTIMUM CONCENTRATION RANGES- When the recommended procedure is used, the calibration graph for the determination of phosphorus in the presence of up to 200 pg of silicon is linear over the range 2 to 25 pg of phosphorus. The absorbances corresponding to these concentrations of phosphorus in the original aqueous solution (0-083 to 1 p.p.m.) are 0.048 and 0-64. The calibration graph for the determination of silicon in the presence of up to 200 pg of phosphorus is linear over the range 2 to 25 pg of silicon.The absorbances corresponding to these concentrations of silicon in the original aqueous phase (0.08 to 1 p.p.m.) are 0.043 and 0.54. Y Y 0 5 06 c( 0.4 m ai u 0.3- n - m m L 2 0.2- 6 - - 1_11__ Number of washings with 0 . 1 5 ~ hydrochloric acid - butanol Fig. 3. Effect of number of washings with 0.15 M hydrochloric acid on absorbance produced by 12 pg of silicon 06 O d 0 4 2 s 2 0.2 6 - m m n D Weight, pg Fig. 4. Calibration graphs for A, silicon in the presence of 200 pg of phosphorus; and B, phosphorus in the presence of 200 pg of silicon PRECISION- A series of absorbance values was obtained from the replicate analysis of 25ml of a solution containing 20 pg (0.8 p.p.m.) of phosphorus and 20 pg (0.8 p.p.m.) of silicon.The average absorbance for phosphorus was 0.510 and the standard deviation was 0.005 absorbance unit or 1.03 per cent. The average absorbance for silicon was 0.404 and the standard deviation was 0.01 absorbance unit or 2.45 per cent. EFFECT OF DIVERSE IONS- The effect of a selected group of elements on the absorbances produced in the deter- mination of 10 pg of phosphorus or 12 pg of silicon by the recommended procedure has been investigated. An ion was considered not to interfere, for the purpose of this study, when less than 5 per cent. error in absorbance was produced for the test solutions compared with the standards. The presence of a 100-fold excess by weight of the following ions produced no interference: Al, Au, Bi, Ca, Cd, Co(II), Cr(III), Cu, Fe(III), Ni, Pb, Mg, Mn, Se(IV), Te(IV), Ti(IV), Zn, fluoride, EDTA, NO,- and Sod2-.The effect of arsenic and germanium, which are known to form the central atoms of heteropoly acids in which molybdate is the co-ordinated group, was also examined. Neither of these elements (as arsenic(V) and german- ium(1V) interferes in the determination of phosphorus, because of the selectivity of the isobutyl acetate for extraction of the phosphomolybdic acid. Both arsenic(V) and germanium(IV), however, interfere in the silicon determination and yield high results. This is caused by the extraction of arsenomolybdic and germanomolybdic acids into the butanol with the silicomolybdic acid. The presence of a 100-fold excess by weight of antimony(V) and vana- dium(V) causes no interference in the phosphorus or silicon determination.A 100-fold excess of tungsten(V1) causes no interference in the determination of silicon, but produces416 KIRKBRIGHT, SMITH AND WEST low results in the phosphorus determination. This is presumably caused by formation of some phosphotungstic acid that is not extracted into the isobutyl acetate. If was estab- lished, however, that a 10-fold excess of tungsten(V1) causes no interference in the deter- mination of phosphorus. It seems, as reported elsewhere,6 that it is possible to tolerate larger excesses of tungsten when Tiron (1,2-dihydroxybenzene-3,5-disulphonic acid) is used as masking agent. ACCURACY- A measure of the accuracy of the method was obtained by the determination of phosphorus and silicon sequentially in synthetic solutions treated as unknown samples. The results of these analyses are shown in Table I. We are grateful to the Ministry of Aviation for support of this work. TABLE I DETERMINATION OF PHOSPHORUS AND SILICON IN SYNTHETIC MIXTURES Phosphorus present, p.p.m. 1.0 0 0.10 0.34 0.64 0.88 0.10 0.20 1.00 1.00 Silicon present, p.p.m. 0 0.80 0.24 0.32 0.50 0.62 0.80 0.80 0.16 0.32 Phosphorus found, p.p.m. 1.02 0 0.10 0-32 0-57 0.88 0.10 0-20 1.01 0.98 Silicon found, p.p.m. 0 0.80 0.23 0.30 0.49 0.61 0-8 1 0.78 0.17 0.32 REFERENCES 1. 2. 3. - , Mikrochim. Ada, 1965, 830. 4. 5. 6. 7. Wadelin, C., and Mellon, M. G., Analyt. Chem., 1953, 25, 1668. Paul, J., Analytica Chim. Acta, 1960, 23, 178. Chalmers, R. A., and Sinclair, A. G., Analytica Chim. Acta, 1966, 34, 412. Elwell, W. T., and Wilson, H. N., Analyst, 1956, 81, 136. Djurkin, V., Kirkbright, G. F., and West, T. S., Ibid., 1966, 91, 89. Kirkbright, G. F., Smith, A. M., and West, T. S., Ibid., 1966, 91, 700. Received January 19th, 1967

ISSN:0003-2654    

DOI:10.1039/AN9679200411

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

A ~ d y ~ t , July, 1967, Vol. 92, $9. 417-422 41 7 The Determination of lStrontium-90 in E'nvironmental Materials by Ion Exchange and Preferential Chelation Techniques BY R. D. IBBETT (Ministry of Agriculture, Fisheries and Food, Fisheries Radiobiological Laboratory, Hamilton Dock, Lowestoft, Sajjfolk) An investigation has been made into the conditions necessary for the preferential chelation of calcium and magnesium over strontium by EDTA and citric acid, and subsequent isolation of the strontium by ion exchange. This has been used as a basis for a convenient method for the determination of strontium-90 in a wide range of environmental materials, down to levels as low as 0.001 picocuries per g of material in its natural state. ENVIRONMEKTAL monitoring of nuclear sites, and radiobiological and radio-ecological research commitments at this laboratory have made necessary the development of a routine method for the determination of low levels of strontium-90 in a variety of environmental materials, principally of marine origin, such as sea water, sediments, fish and seaweeds. This requires separation of strontium from naturally occurring and fission-product radioisotopes, including radium-226, ruthenium - rhodium-106, zirconium - niobium-95, cerium - praseodymium-144, yttrium-91 and barium - lanthanum-140, in addition to large amounts of calcium, magnesium and phosphate.The classical fuming nitric acid precipitation methodl was found to give satisfactory results, but because of the tedious and unpleasant manipulations involved it was decided to investigate the possibility of adapting the more convenient ion-exchange techniques available for other materials, with a prospect of eventual semi-automation. The potentialities of the various column operations involving complete alkaline earth absorptions, followed by elutions with citrate, lactate,2 etc., were investigated, but these methods were rejected because of the relatively large and variable amounts of calcium and magnesium present in the samples and, in consequence, the large columns and elution volumes that would have been necessary.Preferential chelation of calcium and magnesium over strontium seemed a more attractive proposition; the unchelated strontium would then be taken up on a cation-exchange column in preference to calcium and magnesium, most of which would pass through.It has been shown that in the presence of ethylenediaminetetra-acetic acid (EDTA), strontium may be preferentially retained from a calcium-containing solution by a cation-exchange column, provided that the correct pH is ~elected.~ Similarly, in the presence of citrate ions, strontium may be separated from magnesium.2 The combination of these two complexing agents offered possibilities of a simultaneous separation of strontium from both calcium and magnesium. EXPERIMENTAL PRELIMINARY SAMPLE TREATMENT All solid samples were dried and ashed in a muffle furnace by conventional techniques to eliminate water and most of the carbon. This was followed by wet ashing with perchloric acid which was chosen because, in addition to destroying the remaining carbon, it was found to volatilise almost all of the ruthenium-106, a major contaminant of many of the samples.418 IBBETT: DETERMINATION OF STRONTIUM-90 IN ENVIRONMENTAL [Analyst, Vol.92 As many samples contained a considerable amount of phosphate, the elimination of which proved to be difficult and time consuming, it was decided to use a phosphate-precipita- tion step as a method of initial alkaline earth concentration. This was performed in the presence of ammonium chloride, thereby considerably reducing the magnesium content of the fraction. Unlike most of the samples, sea water has a low phosphate content and it was found most convenient to eliminate much of the magnesium in sea-water samples before the stron- tium phosphate precipitation step.This was achieved satisfactorily by precipitation direct from the original sample with carbonate-free alkaline hydroxide. It was also found that a useful preliminary activity decontamination could be incorporated into this stage by the addition of iron(II1) ions. The strontium was then recovered as phosphate, as with the solid samples, except that the presence of ammonium chloride was no longer necessary. Strontium yields obtained after these stages were always in excess of 90 per cent. CONDITIONS FOR THE PREFERENTIAL COLUMN UPTAKE OF STRONTIUM SEPARATION FROM CALCIUM AND MAGNESIUM- Previous work3 ,* and experiments in the laboratory revealed the information summarised in Fig. 1, concerning the uptake of strontium, calcium and magnesium on to a cation-exchange column in the presence of EDTA and citric acid over a range of pH conditions.From these graphs it may be seen that the highest pH at which strontium will be quantitatively retained from a solution containing both complexing agents is 5.0. It may also be seen that optimum separation should be achieved at this pH, and this was confirmed by experimental runs with prepared mixtures of strontium, calcium and E DTA magnesium. Citric acid S r L v) .- 2 100 u .- SOD $ o Kl 3 4 5 6 7 Fig. 1. Relationship between the per- centage of (a) strontium, (b) calcium and (c) magnesium retained on the resin, and the pH of the solution CHOICE OF CITRATE CONCENTRATION- A study was made to determine the effect of citrate concentration on the uptake of strontium from a solution by a cation-exchange column. Accordingly, 10 mg of strontium were added to each of 3 separate litres of citric acid of 0 6 , 1.0 and 2.0 per cent.concentration,July, 19671 MATERIALS BY ION EXCHANGE AND PREFERENTIAL CHELATION 419 and the pH of each solution was adjusted to 5-0 with N sodium hydroxide solution. The solutions were then run through 10-ml Dowex 50W ion-exchange columns in the sodium form; the strontium leakage in the effluent is shown in Table I. TABLE I OPTIMUM CITRIC ACID CONDITIONS AND STRONTIUM LEAKAGE Citric acid Strontium leakage, Volume, Concentration, Column litres per cent w/w per cent. A 1 0.5 (0.1 B 1 1.0 1.0 C 1 2.0 16.7 D 2 0-5 <0*1 w Quantitative uptake of strontium was observed on column D despite the use of 2 litres On this evidence 0.5 per cent.was adopted as the maximum citric acid of feed solution. concentration. CONCENTRATION OF EDTA- A similar series of experiments was undertaken to study the effect of EDTA concentration on strontium uptake. Three solutions of citric acid (0.5 per cent.) each containing 10 mg of strontium per litre were made 0-5, 1-0 and 1.5 per cent., in EDTA tetrasodium salt. The pH of each solution was adjusted to 5.0 with sodium hydroxide or hydrochloric acid, and 1 litre of each was run through similar 10-ml Dowex 50W columns. No significant leakage of strontium was detected in any instance and it was, therefore, decided that any concen- tration of EDTA up to 1.5 per cent. could be tolerated. COLUMN FEED CONDITIONS- It was observed that the alkaline earth phosphate precipitate obtained from samples could be dissolved in either citric acid or EDTA, but whereas this was readily accomplished in citric acid, in EDTA the step was too time-consuming for this to be incorporated into a routine method.Attempts at using Eriochrome Black T as an indication of completed dissolution, i.e., full chelation of alkaline earths by EDTA, were discontinued because the dye came out of solution during the column absorption step and blocked the column. An attempt was therefore made to relate the EDTA necessary for chelation to the amount of citric acid required to dissolve the precipitate. First, the precipitate was dissolved in citric acid solution, the amount of this being kept to a minimum as the concentration of citric acid in the feed solution was to be no more than 0.5 per cent., and it was obviously preferable to keep the volume of the feed as small as possible.The pH of the solution was then adjusted to 11 with sodium hydroxide to re-precipitate the phosphates. EDTA was added very slowly to the suspension with vigorous stirring until the precipitate dissolved. In all cases from all types of initial sample it was found that less than twice as much EDTA as citric acid accom- plished dissolution. As it had been shown already that an excess of EDTA could be tolerated it was decided to adopt this as a basis for the routine method; thereafter, sample phosphates were treated simply by dissolving in a minimum amount of citric acid and then adding twice this amount of EDTA before dilution and pH adjustment.DECONTAMINATION BY EDTA - CITRIC ACID- I t can be seen from published literature4 that at pH 5 the citrate radical will form chelates with aluminium, copper, nickel, cobalt, zinc and manganese, in addition to mag- nesium. Also5 at pH 5, EDTA will complex rare earths, iron and zirconium, in addition to most of the calcium. As no radioactive contamination could be detected by half-life measurements of the yttrium-90 precipitate obtained at the end of this method it was considered unnecessary to check individual decontamination for this stage. This was verified on samples known, by gamma spectrometry, to contain substantial amounts of these con- taminants. It was considered that a 500-ml wash with EDTA - citric acid applied after the feed solution was adequate to free the column of all remaining contaminants, except barium and radium.420 IBBETT : DETERMINATION OF STRONTIUM-90 IN ENVIRONMENTAL [Analyst, Vol.92 ELUTION OF STRONTIUM FROM THE COLUMN As it was proposed to elute the strontium from the column with hydrochloric acid, and as the solubility of EDTA decreases considerably with increasing acidity, it was necessary to remove interstitially held EDTA - citric acid before the elution step by applying a water wash. For the same reason, the column was converted before use from the hydrogen form to the sodium form to prevent hydrogen ions, released in the strontium absorption step, from lowering the pH and precipitating EDTA within the column, thereby impairing the flow-rate. After some preliminary runs it was decided to use 2 N hydrochloric acid to elute the strontium.It was found that 100 ml of this reagent would completely elute all of the strontium, no barium being observed until 110 ml had passed through. This was confirmed by using barium-133 tracer and counting consecutive 10-ml fractions of the eluate for y-activity. Group IIIA elements are not readily eluted with 2 N hydrochloric acid and therefore the daughter activities, yttrium-90 and lanthanum-140, remained in the column, being subse- quently eluted along with the barium and radium with 6 N hydrochloric acid before the column was regenerated for re-use. YTTRIUM MILK A conventional iron(II1) hydroxide scavenge was applied to the strontium fraction and, after estimation of recovery and a waiting period of at least 14 days for equilibration, the strontium-90 activity was determined, the daughter yttrium-90 activity being separated by co-precipitation on iron(II1) hydroxide, which was also used as the counting source.6 Yttrium-90 milking was undertaken as a necessary step because of the recovery estimation method used and because of the probable presence of strontium-89 in many of the samples.ESTIMATION OF YIELD Several materials contain natural strontium, which presents a separate problem in estimating yield if undertaken gravimetrically. This was overcome by adding strontium-85 tracer (obtained from the Radiochemical Centre, Amersham) to all samples and pilot runs, and determining recoveries by direct y-counting. Yields from all samples were generally in excess of 80 per cent.REAGENTS A 10 per cent. solution of EDTA was prepared by dissolving 100 g of the ethylene- diaminetetra-acetic acid (tetrasodium salt) (laboratory-reagent grade) in 1 litre of de-ionised water, and passing the mixture through a Whatman No. 42 filter-paper to remove undissolved matter. A dilution of this same solution was used to produce the EDTA - citric acid wash solution. All other reagents used were of analytical-grade specifications and all solutions or dilutions were made with de-ionised water. PRELIMINARY SAMPLE TREATMENT- Sediments and biological materials-Dry the sample in an oven at 120" C, ash in a muffle at 500" C until most of the carbon has been removed, and grind it to pass through a 60-mesh sieve. Place 20 g of the resulting ash in a 1500-ml beaker and add to it 1 ml of strontium carrier solution (containing 10 mg of inactive strontium and about 500 disintegrations per minute of strontium-85 tracer), 100 ml of de-ionised water and 100 ml of analytical-reagent grade concentrated nitric acid.Heat the mixture on a hot-plate to boiling-point. Add 100 ml of analytical-reagent grade 60 per cent. perchloric acid to the mixture and heat to dryness. Leach the residue thoroughly with a mixture of 300 ml of de-ionised water and 100 ml of analytical-reagent grade concentrated hydrochloric acid. Filter, and reject residue. Add, with stirring, 10 ml of orthophosphoric acid and 100 ml of concentrated ammonia solution to the filtrate. Spin the mixture in a centrifuge and reject the supernatant liquid.Stir the precipitate with 200ml of de-ionised water, spin it in a centrifuge, and reject the supernatant liquid. Retain the alkaline earth phosphate precipitate for dissolution to produce the column feed solution. Sea water-To 1 litre of sea water add 10 ml of concentrated hydrochloric acid, 1 ml of strontium carrier solution (as above) and 10 ml of iron(II1) chloride solution (containing PROCEDUREJuly, 19671 MATERIALS BY ION EXCHANGE AND PREFERENTIAL CHELATION 421 1 mg of iron per ml). Stir thoroughly and add sufficient analytical-reagent grade sodium hydroxide pellets to raise the pH to 11. To the supernatant liquid add 10 ml of orthophosphoric acid and re-adjust the pH to 11 with sodium hydroxide. Spin the solution in a centrifuge and reject the supernatant liquid.Stir the precipitate with 200 ml of de-ionised water, and spin it in a centrifuge; reject the supernatant liquid. Retain the alkaline earth phosphate precipitate for the production of a column feed solution. Spin the mixture in a centrifuge and reject the precipitate. PREPARATION OF COLUMN FEED SOLUTION- Add 200 ml of de-ionised water to the alkaline earth phosphate precipitate and stir. Add slowly sufficient 5 per cent. citric acid solution to dissolve the precipitate. Note the volume of citric acid used. Add an equal volume of 10 per cent. EDTA tetrasodium salt solution and transfer the solution to a 2-litre beaker. Dilute the solution with de-ionised water to produce a 0.5 per cent. concentration of citric acid, then filter the solution through a Whatman No.42 filter-paper. Adjust the pH of the solution to 5.0 with N hydrochloric acid. PREPARATION OF COLUMN- Preparation of resin-The resin chosen for the determination was Dowex 50W x 8, 50 to 100-mesh cation exchanger, initially in the hydrogen form. Some of the resin is preliminarily treated by stirring it in a beaker with N hydrochloric acid and decanting to remove fines. CoZumn-The column consists of a Pyrex glass tube, 15 cm long and 1 cm i.d., with a coarse glass sinter set into the lower end to retain the resin. A vertical Pyrex tube, 60 cm long by 6 cm i.d., is connected to the upper end to serve as a reservoir for feed and wash solutions. A tube, 2 mm i.d., is connected to the column below the sinter and is constructed to deliver liquid at a level just above that of the top of the resin, thus preventing the bed from running dry.Preparation of the resin bed-The column is filled, to a height of 12 cm above the sinter, with pre-treated resin, which is then covered with a glass-wool plug to prevent the bed from being disturbed by the introduction of washes. Before use the resin bed is converted to the sodium form by passing through it a wash of 100 ml of 5 per cent. sodium chloride solution followed by 50 ml of de-ionised water, both at a rate not exceeding 2 ml per minute. COLUMN OPERATION Run the feed solution through the column at a rate not exceeding 4ml per minute, and discard the effluent. Pass 500 ml of 0-5 per cent. citric acid - 1 per cent. EDTA solution at pH 5 through the column at a rate not exceeding 4 ml per minute, discard the effluent, then pass 50 ml of de-ionised water through the column at about 2 ml per minute and discard the effluent.Pass 100 ml of 2 N hydrochloric acid through the column at about 1 ml per minute to elute the strontium, and collect the effluent in a 250-ml beaker. The column may now, if required, be regenerated for further use by passing through it 200 ml of 6 N hydrochloric acid, backwashing with 50 ml of de-ionised water, and then running through it 100 ml of 5 per cent. sodium chloride and a final 50 ml of de-ionised water, in that order. It is then ready for the next sample determination. RECOVERY AND ESTIMATION OF YTTRIUM MILK- Add 1 ml of iron(II1) chloride solution (1 mg of iron per ml) to the strontium fraction in a 250-ml beaker, then add sufficient carbonate-free ammonia solution, with stirring, to make the solution alkaline, and add a few drops of 100-volume hydrogen peroxide.Filter the mixture through a glass sinter (porosity 4) and reject the precipitate. Add sufficient 6 N hydrochloric acid to the filtrate in a 250-ml beaker so that the pH is below 4. Add 1 ml of iron(II1) chloride solution (1 mg of iron per ml) to the mixture and dilute to 200 ml with de-ionised water. Prepare a strontium-85 standard by diluting 1 ml of the original strontium carrier solution to 200 ml with de-ionised water in a similar 250-ml beaker.422 IBBETT Compare the strontium-85 activity levels of the contents of the two beakers by counting each in turn under the same conditions of counting geometry, with a shielded 3 x 3-inch sodium iodide crystal connected to a conventional pulse height analyser.Derive a strontium recovery factor. Store the strontium fraction from the sample for at least 14 days to allow the yttrium-90 daughter to grow into equilibrium, then add carbonate-free ammonia solution and hydrogen peroxide as before to precipitate the ingrown yttrium-90 together with the iron. Filter the precipitate on to a 24-mm, 0.45-pm Millipore filter membrane in a de-mountable all-glass filter holder, and wash it with 100 ml of 0.1 per cent. ammonium chloride solution. Dry the membrane and precipitate at 120" C for 10 minutes, mount on a stainless-steel counting planchet, and /3-count against a strontium-90 - yttrium-90 standard source with a low background Geiger - Miiller assembly.Calibration is performed by separating yttrium-90 from a standard solution of stron- tium-90 - yttrium-90 at equilibrium containing inert strontium carrier, on to iron(II1) hydroxide as before. The procedure adopted for filtering, drying, mounting and counting is as for the sample. The periods allowed between the separation and counting of the yttrium-90 and the beginning and end of the counting period are similar to those for the sample, to eliminate the need for yttrium-90 decay corrections. CONCLUSIONS A procedure is presented for the determination of strontium-90 in a wide variety of environmental materials without resorting to the customary use of fuming nitric acid. The method has been successfully applied for several years at this laboratory, giving chemical recoveries normally in excess of 80 per cent., and results in close agreement with those obtained by conventional means, as shown in Table 11.TABLE I1 THE DETERMINATION OF STRONTIUM-90 IN VARIOUS SAMPLES BY THE METHOD DESCRIBED AND BY A CONVENTIONAL FUMING NITRIC ACID PRECIPITATION TECHNIQUE Strontium-90, picocuries per g wet Ion-exchange Sample method Brown trout (flesh) . . .. . . . . . . 0.0016 Plaice (flesh), Windscale . . . . . . . . 0.006 Brown trout (flesh) . . . . . . .. . . 0.009 Porphyra, Seascale . . .. . . . . .. 0.020 Fucus, Drigg . . . . .. . . . . 0-074 Mussel (flesh), Braystones . . . . . . 0.169 Shrimp (whole), Braystones . . .. . . 0.432 Sea water, Windscale, picocuries per litre . . 19.9 Conventional method 0-0015 0.006 0.009 0.018 0.068 0.160 0.460 20.0 Although rapid determinations are not a feature of this method, the small handling times involved enable batches of determinations to be undertaken at any one time with comparative ease. Handling times could be reduced even further by using relatively simple automation of the wash, elution and regeneration stages. By the use of 20-g ash samples and overnight counting, total strontium-90 contents of 2 picocuries may be detected with a standard deviation on counting statistics of +9 per cent. For some samples this represents levels as low as 0.001 picocuries per g of the material in its natural state. I thank Mr. F. Morgan for the suggestion to exploit ion-exchange techniques, and Mr. J. W. R. Dutton for many useful suggestions and encouragement, and other colleagues for the collection and preparation of the sample material. REFERENCES 1, 2. 3. Davis, P. S., Nature, 1959, 183, 674. 4. 5. 6. Willard, H. H., and Goodspeed, E. W., I n d . Engng Chem. Analyt. Edn, 1936, 8, 414. Milton, G. M., and Grummitt, W. E., Can. J . Chem., 1957, 35, 541. Povondra, P., and Sulcek, Z., Colln Czech. Chem. Comvnun. Engl. Edn, 1959, 24, 2398. Fritz, J. S., and Umbreit, G. R., Analytica Chirn. A d a , 1958, 19, 509. Sugihara, T. T., James, H. J., Troianello, E. J., and Bowen, V. T., Analyt. Chew., 1959, 31, 44. Received December 14th, 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200417

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Analyst, July, 1967, Vol. 92, @. 423430 423 The Examination of Tocopherols by Two-dimensional Thin-layer Chromatography and Subsequent Colorimetric Determination BY K. J. WHITTLE* AND J. F. PENNOCK (Deeartment of Biochemistry, University of Liverpool, Liverpool 3) A two-dimensional thin-layer chromatographic system has been used to separate tocopherols and tocotrienols before determining them by the Emmerie - Engel reaction. Reproducible recoveries of about 92 per cent. were obtained with authentic tocopherols and the method was used to deter- mine tocopherols and tocotrienols in several vegetable oils. A compound, apparently plastochromanol, has been identified in some oils. THE existence in nature of four tocopherols and four related tocotrienols has been reported by Pennock, Hemming and Kerr.l The four naturally occurring tocopherols known so far (M-, p-, y- and 8-tocopherols) all have a methyl group at position 8, as do the four related tocotrienols (see Table I).The purpose of this paper is to describe a method by which the tocopherols and tocotrienols can be separated and determined quickly and simply. Although many methods have been used for the analysis of tocopherols, perhaps the best and most widely used at this time is that recommended by the Vitamin E Panel of the Analytical TABLE I THE NATURALLY OCCURRING TOCOPHEROL COMPOUNDS Tocopherol R= (CHz-CHz-CH-CHz)3H I Methyl substitution CH3 5,7,8-Trimethyl . . wTocophero1 5,s-Dimethyl . . .. 8-Tocopherol 7,s-Dimethyl . . . . y-Tocopherol 8-Monomethyl . . &Tocopherol Tocotrienol R = (CHZ-CH = C-CHJ,H I CH3 cr-Tocotrienol or C,-tocopherol 8-Tocotrienol or etocopherol y-Tocotrienol* 6-Tocotrienol * A tocopherol found in rice and palm oil, and denoted ?-tocopherol, was previously thought to be 7-methyl tocol but is now known t o be y-tocotrieno1.l Methods Committee of the Society for Analytical Chemistry (A.M.C.).2 The tocopherol fraction obtained by column chromatography of the unsaponifiable lipids is subjected to two-dimensional paper chromatography, the first system being adsorption and the second, reversed-phase chromatography.The tocopherols are eluted and determined by the Emmerie - Engel method. B0oth~9~ has had success without the preliminary purification by column chromatography. Disadvantages are that the paper has to be prepared for reversed-phase chromatography while the tocopherols are adsorbed on the paper and, further, it is not possible to separate P-tocopherol from y-tocopherol or P-tocotrienol from y-tocotrienol.The papers are susceptible to overloading and the tocopherol spots obtained are large. * In receipt of a Science Research Council Research Studentship.424 WHITTLE AND PENNOCK EXAMINATION OF TOCOPHEROLS [Analyst, VOl. 92 Seher5 showed that separations of cc-tocopherol from /I- plus y-tocopherol and from &tocopherol could be achieved by thin-layer chromatography. By measuring spot sizes of tocopherols in a concentrate, and comparing them with those of known amounts of authentic tocopherols, Seher was able to make an estimation of the tocopherol content.A major advance in tocopherol separations was made by Stowe,6 who used a five-component solvent system for chromatographing tocopherols on thin layers of silica gel and was able to separate p- from y-tocopherol. Thus with a mixture of cc-, p-, y- and &tocopherols, Stowe obtained four distinct spots by thin-layer chromatography. Recently, Rao, Rao and Achaya' have used this chromatographic system for separating tocopherols in the unsaponifiable lipids of several seed oils before determining them colorimetrically. It was also shown by Stowe that good separation of 6- from y-tocopherol could be obtained by using a 20 per cent. v/v solution of di-isopropyl ether in light petroleum, boiling range 60" to 80" C. Pennock, Hemming and Kerrl found that Stowe's system was also useful in separating tocopherols from tocotrienols, and they described a two-dimensional thin-layer chromatographic system for separating tocopherols.Chloroform was used in the first dimension to separate the trimethyl, dimethyl and monomethyl compounds, and then 20 per cent. di-isopropyl ether in light petroleum to separate a-tocopherol from cc-tocotrienol, 6-tocopherol from P-tocotrienol plus y-tocopherol and from y-tocotrienol, and finally, 6-toco- pherol from 6-tocotrienol. Thus seven spots were obtained for the eight compounds, with /I-tocotrienol and y-tocopherol overlapping (see Fig. 1). 01 'B2 7m5 u 6 - c 0 "Blanks" taken from this area - ? 9 0 0 8 vl t I x - Solvent B Origin Fig. 1 . Two-dimensional thin-layer chromatogram showing several lipids: spot 1, ubiquinone; spot 2, c(- tocopherol ; spot 3, a-tocotrienol ; spot 4, plastochromanol ; spot 5, P-tocopherol ; spot 6, P-tocotrienol or y-tocopherol ; spot 7, y-tocotrienol; spot 8, S-tocopherol; spot 9, S-tocotrienol.Spots 2 to 9 inclusive were stained with Emmerie - Engel reagent, and spot 1 was stained with Emmerie - Engel reagent after reduction with sodium borohydride in ethanol. Solvent A was chloroform and solvent B was 20 per cent. v/v solution of di-isopropyl ether in light petroleum, boiling range 40" to 60" C EXPERIMENTAL REAGENTS- Ethanol, absolute-Purify by refluxing over zinc dust and potassium hydroxide (8 and 16 g, respectively, per litre of ethanol) for 4 hours, and then distil and collect the fraction boiling at 78" C. I t contains 1 per cent.v/v of ethanol as stabiliser. Chloroform-Use as purchased from May and Baker Ltd., Dagenham.July, 19671 BY TWO-DIMENSIONAL THIN-LAYER CHROMATOGRAPHY 425 Cyclohexane-Purify by distillation, followed by passing it through silica gel (30 to 120 mesh). Dietlzyl ether-Dry over sodium wire and distil over reduced iron just before use. Di-isopropyl ether-Technical-reagent grade, obtainable from British Drug Houses Ltd. Light petroleum, boiling range 40" to 60" C-Dry over sodium wire and distil. Sodium sulphate, anhydvous-Contains not more than 0.002 per cent. w/w of iron. Iron(II1) chloride hexahydrate solution-Prepare a 0.2 per cent. w/v solution of analytical- 2,2'-Dipyridyl solution-Prepare a 0.5 per cent. w/v solution of analytical-reagent Fluorescein solution-Prepare a 0.01 per cent.w/v solution of general-purpose reagent Silica gel G. It contains 0.01 per cent. w/v of hydroquinone to inhibit peroxide formation. reagent grade iron( 111) chloride, FeC1,.6H20, in absolute ethanol. grade 2,2'-dipyridyl in absolute ethanol. grade fluorescein in absolute ethanol. MATE RIALS- D-CC- and D-y-Tocopherol-Obtainable from Distillation Products Industries, Rochester, New York, U.S.A. DL-p- and DL-6-TocopheroZ-These were given by Dr. J. Green, Vitamins Ltd., Tad- worth, Surrey. The tocopherols were treated to preparative thin-layer chromatography on silica gel by using as solvent either chloroform or a 20 per cent. v/v solution of di-isopropyl ether in light petroleum. Vegetable oils-Crude samples of the vegetable oils : palm, soyabean, cottonseed, rapeseed and maize were given by Dr.H. Jasperson of J. Bibby and Sons Ltd., Liverpool. QUANTITATIVE RECOVERY OF TOCOPHEROLS- Samples of a mixture of a-, p-, y- and 6-tocopherols were treated to two-dimensional thin-layer chromatography and, after recovery from the chromatogram, the tocopherols were determined by the Emmerie - Engel reaction.8 AP$licatio.tz to the plate-About 60 pg of the tocopherol mixture (containing 15 pg of each tocopherol) in 25p1 of cyclohexane solution was applied to a thin layer of silica gel (250 p thick and 20 x 20 cm). Chromatography-The chromatogram was developed first with chloroform (with a tank lined with filter-paper to assist saturation of the vapour phase), and secondly, with 20 per cent. v/v of di-isopropyl ether in light petroleum, boiling range 40" to 60" C (unlined tank).To obtain optimum separations the chromatotanks containing the solvent were shaken just before use. During the development in each system the tanks were covered to exclude light. Care was taken to remove all of the chloroform from the thin layer before running the second system. The chloroform tank could be used several times, but it was found that the best results were obtained if the second solvent system was freshly prepared each day. The chromatography required 30 minutes in the chloroform tank and 40 minutes in the di-isopropyl ether - light petroleum tank, the solvent front travelling 16 cm in each instance. Detection-The tocopherols were detected by spraying the chromatogram evenly with the ethanolic fluorescein solution.They were seen as purple spots on a fluorescent background in ultraviolet light (270 to 280 mp), and as pink spots in visible light (Bolligerg). Four distinct spots corresponding to a-, p-, y- and &tocopherols were seen (see Fig. l), the lowest limits of detection being 1 to 2 pg of tocopherol. Extraction of tocopherols-The area of silica gel corresponding to each tocopherol was scraped off and transferred into. conical-shaped 10-ml centrifuge tubes that were shielded from light in enclosed test-tube racks. "Blank" areas of silica gel (equal in size to the tocopherol spots) were also scraped off the chromatogram from positions within the solvent fronts where no lipids were to be found. To each tube 4 ml of ethanol were added, and the silica gel was stirred up and spun by centrifuge (1000 g) for 2 to 3 minutes.The silica gel was re-suspended in the ethanol and spun by centrifuge once again. The time elapsing between the removal of the chromatogram from the solvent and elution of the tocopherol was kept to a minimum to prevent destruction of the tocopherols.426 WHITTLE AND PENNOCK : EXAMINATION OF TOCOPHEROLS [Analyst, Vol. 92 Colorimetric determination-To 3 ml of the total of 4 ml of ethanolic extract 0.5 ml of the 2,2’-dipyridyl solution was added, followed by 0.5 ml of the iron(II1) chloride solution. The tube was well shielded from light and the extinction at 520 mp was measured on a Unicam SP600 spectrophotometer. The readings were taken 2 minutes after adding the iron(II1) chloride solution, except for &tocopherol, for which an interval of 3 minutes was allowed (to be discussed later).The glass spectrophotometer cells were rinsed with ethanolic iron(II1) chloride, water and ethanol before use to remove any reducing materials that might interfere with the tocopherol determinations. An average value for the blank readings was obtained, and the corrected reading was converted to micrograms of tocopherol per 4 ml of Emmerie - Engel reaction mixture by using the spectrophotometric factors recommended in the A.M.C. Report (p. 365),2 i.e., a-tocopherol, 98; P-tocopherol, 96; y-tocopherol, 90; and &tocopherol, 75. Table I1 indicates that between 91 and 93 per cent. of the tocopherols could be recovered from the chromatogram and that the recoveries were reproducible.TABLE I1 PERCENTAGE RECOVERY OF TOCOPHEROLS FROM TWO-DIMENSIONAL THIN-LAYER CHROMATOGRAPHY Tocopherol Percentage recovery of tocopherol Average percentage recovery a 93.7 93.7 93.7 90.3 92-85 92.8 92.8 92-8 94.6 93.25 96.1 90.3 88.7 92.8 9 1.975 Y 6 90.5 91.1 92.4 93.0 91-75 B DETERMINATION OF SOME TOCOPHEROLS IN CRUDE VEGETABLE OILS- The tocopherols contained in several seed oils (see under Materials) were determined. The seed oils’(1 g of each) were saponified by the procedure given in ihe A.M.C. Report (p. 361).2 Establishment of the tocopherol pattern-The ethanolic extract was dissolved in 0-2 to 0.4 ml of cyclohexane and the tocopherol pattern in each seed oil was established initially by qualitative two-dimensional thin-layer chromatography.Chromatograms were run in duplicate, one being sprayed with Emmerie - Engel reagent to locate the tocopherols and the other with diazotised o-dianisidine reagent,2 which forms a stain with y- and %tocopherols and y- and 8-tocotrienols. The colours obtained with these two spray reagents, and the positions of the tocopherol spots on the chromatograml were usually sufficient to identify the tocopherols or tocotrienols. Authentic tocopherols were added to the chromatogram before running it to confirm the identifications. As mentioned under Reagents, the di- isopropyl ether contained 0.01 per cent. of hydroquinone and this substance was strongly adsorbed on the silica gel where the plate was dipped in the solvent. The hydroquinone stained strongly with either the Emmerie - Engel reagent or the diazotised o-dianisidine.Quantitative determination of tocopherols-Having established the tocopherol pattern, aliquots of the unsaponifiable lipids (1 to 2 mg), usually in 10, 25 or 50 p1 of the cyclohexane solution, were chromatographed and the tocopherols were separated and determined as TABLE I11 TOCOPHEROL CONTENT OF CRUDE VEGETABLE OILS Tocopherols found in oil, lipid from oil, (---*-, Unsaponifiable tLg per g Oil mg Per g a B Y 6 Soyabean .. 23.5 34 trace 326 204 26 trace 308 195 Cottonseed . . 29.8 489 - 296 - 470 - 275 - Rapeseed . . 30-8 238 - 424 11 244 - 430 12 Maize . . .. 35.0 60 trace 437 20 63 trace 448 24 Total 564 529 785 745 673 686 517 535 Tocopherols found in oil, per cent. of total tocopherol 7*- 7 a B Y 6 6.0 - 57.8 36.2 4.9 - 58.2 36.9 62.3 - 37.7 - 63.1 - 36.9 - 35.3 - 63-1 1.6 35.6 - 62.7 1.7 11.6 - 84.5 3.9 11.8 - 83.7 4.5July, 19671 BY TWO-DIMENSIONAL THIN-LAYER CHROMATOGRAPHY 427 TABLE IV TOCOPHEROL AND TOCOTRIENOL CONTENT OF CRUDE PALM OIL Tocopherols* and tocotrienols* found in oil, pg per g Tocopherols and tocotrienols found in oil, per cent.of total tocopherol Sample mg per g aT aT-3 pT-3-f yT-3 8T-3 Total aT ctT-3 /3T-3 yT-3 6T-3 Unsaponifiable lipid from oil, r A > < A > 1 13.8 207 121 38 266 73 705 29.4 17.2 5.4 37.7 10.3 209 120 26 285 61 701 29.5 17.1 3.7 40.7 8.7 2 15.5 204 149 24 302 69 748 27.3 19.9 3.2 40.4 9.2 209 138 19 303 68 737 28.4 15.7 2.6 41.1 9.2 3 17.4 200 147 29 268 71 715 28.0 20.6 4.0 37-5 9.9 200 181 25 269 72 747 26.8 24.2 3.3 36.1 9.6 4 20.2 188 152 30 281 67 718 26.2 21.2 4-1 39.2 9-3 239 134 38 311 72 794 30.1 16.9 4.8 39-2 9-0 * The following abbreviations are used: a-tocopherol = aT; a-tocotrienol = aT-3; etc.-f fl-Tocopherol was also detected but the level was too low to be determined. described in the section on Quantitative recovery of tocopherols. The results are shown in Tables I11 and IV. With palm oil (see Fig. 2) several spots were seen that corresponded to CC-, p-, y- and 6-tocotrienols, as well as a- and P-tocophero1s.l To determine the tocotrienols the Emmerie - Engel conversion factors for the related tocopherol were used. It was found that the total unsaponifiable lipid could be chromatographed without previous separation of tocopherols or removal of sterol.Good agreement was obtained with duplicate runs on the same specimen of unsaponifiable lipid (Table 111). I I Solvent front A I ‘ I I I X - Solvent B I I I Origin Fig. 2. Two-dimensional thin-layer chromatography of unsaponifiable lipid from palm oil: spot 1, cr-tocopherol; spot 2, a-tocotrienol; spot 3, p-tocopherol; spot 4, 8-tocotrienol ; spot 5, y-tocotrienol ; spot 6, 8-tocotrienol. All the spots shown on the chromatogram were stained with Emmerie - Engel reagent The analysis for tocopherols and tocotrienols in palm oil was less straightforward than with the other seed oils. Several tocopherols were present, and if sufficient of the unsaponi- fiable lipid was applied to the thin-layer chromatogram to allow the determination of the minor constituents, e.g., P-tocopherol and 16-tocotrienol, then the spots of the major com- ponents overlapped each other.The position on the chromatogram occupied by cc-tocopherol and cc-tocotrienol was also taken by a carotenoid, which made detection of the tocopherol difficult.428 WHITTLE AND PENNOCK: EXAMINATION OF TOCOPHEROLS [APzahSt, VOl. 92 Non-tocopherol components encountered-In determinations on rapeseed and maize oil the analysis of a-tocopherol was found to be difficult, as a reducing substance (which also gave a dianisidine stain) was found to run next to, but a little behind, a-tocopherol in both systems (see Fig. 1). Several reducing spots other than tocopherol or tocotrienols were noticed. These were less polar than the toco- pherols and it may be that they correspond with the dimeric oxidation products of tocopherols identified in seed oils by McHale and Green,12 and also by Shone.13 It was found that all of the oils gave a spot that showed up well with fluorescein and with the phosphomolybdic acid spray reagent,14 and corresponded to ubiquinone (see Fig.1). The spot could also be sprayed with a solution of sodium borohydride in ethanol (a modi- fication of the method used by Lester and Ramasarma15), followed by a 1 per cent. solution of hydrochloric acid in ethanol to destroy excess of borohydride. The spot then reacted positively with the Emmerie - Engel reagent, indicating that the organic substance was a quinone. SPECTROPHOTOMETRIC FACTORS- When examining recoveries of authentic tocopherols some difficulty was found with the factor for 8-tocopherol.A survey of the literature revealed that there are some doubts about the conversion factor for 6-tocopherol (Green, Marcinkiewicz and Watt16). Thus the 2-minute factor for 8-tocopherol (75) suggests a much greater colour development than the other toco- pherols, whereas Stern and Baxter17 and Tsenls reported that longer periods than 2 minutes are required for 8-tocopherol to give colour development as great as that given by other tocopherols. Booth4 allowed 3 minutes, at least, for colour development with 8-tocopherol. This compound appears to be plastochromanol.lOJ1 - .. ...., c.. . . . . . . . ... .. . . . . . . . . . . . . .... .. 1. . .. . . .+ B.. :./: ---- * -----___ _c ______ i/ Fig. 3. The rate of colour development a t 520 mp for A, a-tocopherol; B, /3-tocopherol; C, y-tocopherol; and D, &tocopherol, with the Emmerie - Engel reagent (equal parts of 0.5 per cent.2,2'-dipyridyl in ethanol and 0.2 per cent. iron(II1) chloride in ethanol). The con- centration of each tocopherol solution was 4 mg per 100 ml of ethanol The colour produced by 6-tocopherol and the Emmerie - Engel reagent at 520 mp was measured spectrophotometrically against a reagent blank in a Unicam SP800 recording spectrophotometer and compared with that produced by the other tocopherols. As shown in Fig. 3, full colour development is not given after 2 minutes, the value being about 85 per cent. of that after 10 minutes; and at 3 minutes it is about 94 per cent. of the latter. However, for tocopherol solutions of equal concentration the colour development for 6-tocopherol is similar to that of the other tocopherols after only 80 or 90 seconds. In recent work19 at Liverpooh the spectrophotometric factors were determined for 6-tocopherol and 8-tocotrienol, and were found to be similar.It is our view that the A.M.C. factor for 8-tocopherol is reason- able only for a 3-minute reading.July, 19671 BY TWO-DIMENSIONAL THIN-LAYER CHROMATOGRAPHY 429 POSSIBLE CONTAMINATION BY RUBBER PRODUCTS- It has been shown (Dunphy, Whittle, Pennock and Morton20) that a-, y- and ?i-tocotrienols, both free and esterified, occur with cc-tocopherol in rubber latex. Samples of crepe rubber and smoked crepe rubber (kindly provided by Dr. E. G. Cockbain of the Natural Rubber Producers’ Research Association, Welwyn Garden City, Herts.) were found to have about 2 mg of tocotrienyl ester and 1.5 mg of free tocotrienol per g of smoked crepe, and 1.3 mg of ester and 0.8 mg of tocotrienol per g of crepe rubber. I t is advisable when assaying toco- pherol to avoid the use of various rubber products, e.g., tubing and bungs, DISCUSSION Quantitative experiments with authentic tocopherols indicated that between 91 and 93 per cent.of the material applied to the chromatogram was recovered. When chromatograms were stained quite a strong reaction was given at the origin spot, and we must assume that most of the breakdown occurred before chromatography or during the application. We have not done any quantitative experiments with tocotrienols, but our experience with them suggests that they are slightly more unstable than tocopherols on thin layers of silica gel.Experiments were carried out in a nitrogen atmosphere to see if higher recoveries could be obtained. Running the chromatogram in an atmosphere of nitrogen was technically more difficult, the development time was considerably longer (presumably the atmosphere was not saturated with solvent vapour) and recoveries were lowered. Other recovery experiments were carried out in a darkened room, but although the tocopherols suffered less breakdown in the absence of light, all of the procedures were slower in the dark and recoveries were again lowered. In the method for separating tocopherols and tocotrienols, y-tocopherol and P-tocotrienol overlap.The problem of determining these two substances is not difficult as they can be cut out and subjected to reversed-phase chromatography,2 which will effectively separate them, or they may be converted to their nitroso-derivatives,21 which are also easily separated. Rather the problem is knowing which compounds are present. The Emmerie - Engel reagent will stain both, but only y-tocopherol gives a blue colour with diazotised o-dianisidine. There- fore, if a pink colour is obtained with the Emmerie - Engel reagent and no blue colour with dianisidine, then p-tocotrienol is present. If a blue dianisidine stain is given y-tocopherol is present, but j3-tocotrienol may also be present. The compound appearing to be plastochromanol in seed oils is interesting as it is usually associated with plastoquinone and the photosynthetic apparatus.1° ~1 Plastoquinone has not been found in the oils but, as they are saponified before analysis, it is almost certain that any plastoquinone present would be degraded.In adsorption chromatograms plastochromanol runs very close to a-tocopherol and might have been determined with it, but reversed-phase chromatography would separate the two substances. Thus it would seem that the unidentified reducing material observed by Herting and Drury22 in castor bean and linseed oils, which moved with a-tocopherol on adsorption chromatography and failed to move on reversed-phase chromatography, was plastochromanol; the amounts here were quite high compared with the total tocopherol present. The determination of tocopherols and tocotrienols in palm oil was a difficult task, but this tissue was selected for a thorough investigation because it was believed that if the method could be successfully applied to palm oil it would be suitable for most other determinations.Palm oil contains a large amount of carotenoid, which made visual identification of the tocopherols difficult. I t contains several tocopherols and tocotrienols, some in high and others low concentration, and in particular, it contains all the known tocotrienols, which seem to be more unstable than the related tocopherols. The results show that quite good determinations were possible and, in particular, 6-tocotrienol gave reproducible values. There is much to recommend the two-dimensional thin-layer chromatographic system as a method for separating the tocopherols and tocotrienols before determining them.No preliminary concentration of the tocopherols is required (either by chromatography or removal of sterol); the separations are very rapid; only two of the eight members of the “tocopherol” family merlap; no paraffin is present to be extracted with the tocopherols; and the method can be applied directly to the determination of tocopherols in whole lipids from 1eaves.ll430 WHITTLE AND PENNOCK REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Pennock, J. F., Hemming, F. W., and Kerr, J. D., Biochem. Biophys. Res. Commun., 1964, 17, 542. Analytical Methods Committee, Analyst, 1959, 84, 356. Booth, V. H., Biochem. J . , 1962, 84, 444. -, Analyst, 1963, 88, 627. Seher, A., Mikrochim. Acta, 1961, 308. Stowe, H. D., Archs. Biochem. Biophys., 1963, 103, 42. Govind Rao, M. K., Venkob Rao, S., and Achaya, K. T., J . Sci. Fd Agric., 1965, 16, 1. Emmerie, A., and Engel, C., Red. Trav. Chim. Pays-Bas Belg., 1939, 58, 283. Bolliger, H. R., in Stahl, E., Editor, “Thin Layer Chromatography,” Springer-Verlag, Berlin, Whittle, K. J., Dunphy, P. J., and Pennock, J. F., Biochem. J., 1965, 96, 17c. Dunphy, I?. J., Whittle, K. J., and Pennock, J. F., “Biochemistry of Chloroplasts,” Volume 1, McHale, D., and Green, J., Chew. 6% Ind., 1963, 982. Shone, G., Ibid., 1963, 335. Waldi, D., in Stahl, E., 09. cit., p. 498. Lester, R. L., and Ramasarma, T., J . BioE. Chem., 1959, 234, 672. Green, J., Marcinkiewicz, S., and Watt, P. R., J . Sci. Fd Agric., 1955, 6, 274. Stern, M. H., and Baxter, J. G., Analyt. Chew., 1947, 19, 902. Tsen, C. C., Ibid., 1961, 33, 849. Whittle, K. J., Dunphy, P. J., and Pennock, J. F., Biochem. J., 1966, 100, 138. Dunphy, P. J., Whittle, K. J., Pennock, J. F., and Morton, R. A., Nature, 1965, 207, 521. Marcinkiewicz, S., and Green, J., Analyst, 1959, 84, 304. Herting, D. C., and Drury, E. J. E., J . Nutr., 1963, 81, 335. 1965, p. 210. Academic Press Inc., London, 1966, p. 165. Received May l l t h , 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200423

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Analyst, July, 1967, Vol. 92, $9. 431-435 431 The Qualitative Analysis of Complex Carbonyl Mixtures by Thin-layer Chromatography BY J. H. DHONT AND MISS G. J. C. DIJKMAN (Central Institute for Nutrition and Food Research, T N O , Zeist, The Netherlands) By the consecutive use of two thin-layer chromatographic systems, complex mixtures of 2,4-dinitrophenylhydrazones are resolved into two frac- tions, one containing the derivatives of the aromatic carbonyl compounds and the other containing the aliphatic and terpene-type carbonyl compounds. Systems are described for the separation of mixtures of aromatic derivatives. IN the past the systematic separation of 2,4-dinitrophenylhydrazones derived from carbonyl compounds by thin-layer chromatography has been restricted to the separation of the aliphatic 2,4-dinitrophenylhydrazones into methyl ketones, aldehydes and unsaturates.l v 2 93 94 In a previous paper5 the separation of a few aromatic 2,4-dinitrophenylhydrazones in mixtures containing aliphatic-type, but not including terpene-type, carbonyls has been investigated.In many instances, however, for example, when foods and beverages are examined, even more complex mixtures of carbonyls may be encountered, consisting of aromatic and terpene- type carbonyl compounds, as well as the aliphatic type. The application of any of the methods mentioned above may therefore lead to erroneous results, and a differentiation into classes and groups becomes necessary before separation into individual components. During investigations by thin-layer chromatography of 2,4-dinitrophenylhydrazones, the results of which are now reported, a two-step procedure was developed first to separate complex mixtures into two groups, one containing the 2,4-dinitrophenylhydrazones derived from aliphatic carbonyl compounds, including those of terpene-type, and the other containing the derivatives of aromatic-type carbonyl compounds.Further attention was then given to the separation of 2,4-dinitrophenylhydrazones of the aromatic group of carbonyl compounds. EXPERIMENTAL Although activated plates are often used in studies on carbonyl-2,4-dinitrophenylhydrazone separation, the chromatographic procedures used in the experiments described below were carried out on non-activated plates. Previous to their use the plates were left to equilibrate with the atmosphere in the laboratory for at least 3 hours.Test solutions of the compounds were prepared in suitable solvents. In most instances ethyl acetate, dioxane or methanol was used. The concentration of the 2,4-dinitrophenylhydrazones was adjusted to enable 1 to 5 pl to give a good visible spot when applied to the plate. All of the plates were prepared by using the standard Stahl equipment. DESCRIPTION OF CHROMATOGRAPHIC SYSTEMS- The chromatographic systems used in our experiments for the chromatography of the 2,4-dinitrophenylhydrazones are shown in Tables I and 11. Table I presents the types of plates used and their method of preparation. Table I1 gives the solvent systems applied in combination with these plates.432 DHONT AND DIKJMAN: QUALITATIVE ANALYSIS OF COMPLEX [Analyst, VOl.92 TABLE I TYPES OF CHROMATOPLATE COATINGS AND THEIR PREPARATION Types of plate Magnesium oxide 2-Phcnoxyethanol Silica gel . . Zinc carbonate Anasil B . . Acetylcellulose .. . . .. .. .. .. .. .. .. .. .. .. SOLVENT SYSTEMS System No. Coating Preparation of coatings Fifteen grams of kieselguhr G (Merck) and 15 g of magnesium oxide (Merck p.a., sieved to ( 5 0 mesh) were mixed and thoroughly shaken in an Erlenmeyer flask with 80ml of a mixture of equal parts of water and methanol Twenty-five grams of kieselguhr G (Merck) and 5 g of 2-phenoxy- ethanol were shaken in an Erlenmeyer flask with 60ml of the same water - methanol mixture as used for the magnesium oxide plates Twenty-five grams of silica gel G (Merck), 40 ml of ethanol and 20ml of water were shaken in an Erlenmeyer flask for about 1 minute Thirty grams of zinc carbonate, B.D.H.Ltd. (sieved to <40 mesh) and 2 g of plaster of Paris (<40 mesh) were mixed, and suspended in a mixture of 30 ml of methanol and 30 ml of water by shaking in an Erlenmeyer flask Fifteen grams of Anasil B (Analabs Inc.) were suspended in a mixture of 15 ml of methanol and 15 ml of water Fifteen grams of acetylcellulose powder (Machery and Nagcl 300 G/Ac) were suspended in a mixture of 35 ml of ethanol, 15ml of chloroform and 15 ml of water. The suspension was stirred in a small beaker with a thick glass rod TABLE I1 USED WITH THE COATINGS SHOWN IN TABLE I Number of Solvent combination irrigations I I1 111 IV V VI VII ? T I 1 I IX Magnesium oxide 2-Phenoxyethanol Xcetylcellulose Acetylcellulose Acetylcellulose Silica gel G Silica gel G Zinc carbonate Anasil B Hexane - chloroform (7 + 3) .. .. .. .. .. 1 Light petroleum .. .. .. .. .. .. 5 t o 6 Light petroleum - methanol - ethyl acetate (9.5 + 1.5 + 1.0) Light petroleum - dibutyl ether - ethyl acetate - methanol Light petroleum - dibutyl ether - ethyl acetate - methanol 2 (6.5 + 1.0 + 1-5 + 1.0) . . . . . . .. .. 1 (8-5 + 0.5 + 0.5 + 0.5) . . . . . . . . . . 1 Light petroleum - benzene (1 + 1) . . . . . . .. 5 Light petroleum - benzene - dibutyl ether (2 + 1 + 1) Light petroleum pyridine (8 $. 1) .. . . 1 Light petroleum - benzene - pyridine (8 + 1 + 0-25) . . 1 2 . . . . . . . SOLVENTS- Light petroleum-Boiling-point 60" to 80" C, free from aromatic hydrocarbons.Benzene-Distilled commercial sample. Dibutyl ether-Pure, as supplied by Schuchardt, Munich, Germany. Pyridine-As supplied by Merck for chromatography. Methanol, commercial. Ethyl acetate-As supplied by Merck for chromatography. PROCEDURE FOR THE SEPARATION OF COMPLEX MIXTURES OF 2,4-DINITROPHENYLHYDRAZONES INTO GROUPS- The solution of 2,4-dinitrophenylhydrazones to be analysed is first applied as a streak on to the magnesium oxide plates, e.g., by using the apparatus described by Ritter and Meyer .6 $-Toluylaldehyde 2,4-dinitrophenylhydrazone is placed as a marker on the same plate. After development by using system I, several different groups are usually seen as separate bands. The bands having an RF value equal to, or lower than, that of the marker contain most of the aromatic-type 2,4-dinitrophenylhydrazones admixed with, when present, the 2,4-dinitro- phenylhydrazones of keto-acids and the osazones.Keto-acids, however, if present in the mixture of 2,4-dinitrophenylhydrazones, are easily extracted with aqueous carbonate solutions before applying the mixture to the plate. Osazones are recognisable on the magnesium oxide plates as dark blue bands with very low R, values. They will separate out during further analysis of the aromatic compounds.July, 19671 CARBONYL MIXTURES BY THIN-LAYER CHROMATOGRAPHY 433 The mono-derivatives of dicarbonyls are seen as intensely red or pink bands between the marker and the aliphatic-type bands. Diacetyl 2,4-dinitrophenylhydrazone, having the lowest RE' value of the dicarbonyls, may, if present, be partially mixed with the aromatic compounds below the marker band.Some aromatic compounds (benzyl methyl ketone and those like hydratropic aldehyde) with their carbonyl group in a long side-chain, if present in the mixture, will have moved into the area above the p-toluylaldehyde 2,4-dinitrophenylhydrazone marker of system I. All of the bands therefore with RP values larger than that of the marker are now extracted and re-run in system I1 (which was first described by Urbach3) with the 2,4-dinitrophenyl- hydrazone of acetaldehyde as a marker. The aromatic compounds will now appear in bands below the acetaldehyde 2,4-dinitrophenylhydrazone. Formaldehyde may join them but is easily recognised during further analysis. A check for the presence of diacetyl by spraying part of the chromatogram with ethanolic sodium hydroxide will indicate its position by the development of a pink colour, which clearly differentiates it from all aromatic Compounds. All other aliphatic compounds and mono-derivatives of dicarbonyls, together with any terpene-type components of the original mixture, will remain above the marker in system 11.If no terpenes are present the aliphatic compounds may be further investigated by procedures described by other authors,l 9 2 7 3 7 4 but if terpene-type components are admixed with them a special method will have to be applied. This method will be published in our next paper. SEPARATION OF A MIXTURE OF AROMATIC-TYPE 2,4-DINITROPHENYLHYDRAZONES INTO INDI- VIDUAL COMPONENTS- For the separation of the mixture of aromatic carbonyl compounds that have accumulated by extraction from the bands of system I below P-toluylaldehyde and the bands in system I1 below acetaldehyde, some or all of the systems I11 to IX are used.Rs values of these com- pounds, measured against the 2,4-dinitrophenylhydrazone of P-toluylaldehyde as a standard, are tabulated in Table 111. The R, values for the standard in the different systems are given in Table IV. TABLE I11 R, VALUES FOR SOME AROMATIC-TYPE 2,4-DINITROPHENYLHYDRAZONES System 111 IV v Vl VII VIII IX 2,4-Dinitrophenylhydrazone of 2As RS as 5& Rs RS ' R s p-Toluylald6hyde . . Anisaldehyde . . . . Xcetophenone . . . . Benzophenone . . . . Benzaldehyde . . . . Cinnamaldehyde . . Cuminaldehyde .. . . Ethylvanillin . . . . Benzyl methyl ketone . . a-Pentylcinnamaldeh yde @-Isopropyl-a-methylhydro- cinnamaldehyde . . Vanillin . . . . . . Ethyl styryl ketone . . Furfural . . .. .. 5-Methylfurfural .. Piperonal . . .. Isojasmon . . . . Salicylsldehyde . . . . Veratraldehyde . . . . 5-H ydroxymethylfurfural Hydratropic aldehyde . . Methyl a-naphthyl ketone Methyl P-naphthyl ketone Phenylacetaldehyde . . Phenplpropionaldeh yde p-Hj-droxybenzaldehy de Syringa aldehyde . . . . 1.00 . . 0.51 . . Streaks . . Streaks . . 0.93 . . 1.52 . . 0.82 . . 1.12 . . Streaks . . 1.68 . . 0.03 . . 0.02 . . Streaks . . 0.60 . . 0-65 . . 0.72 . . 0.78 . . 1.42 . . 1.85 . . 0.72 . . 0.72 . . 0.45 . . 0.03 . . 1.23 . . 1.32 . . 0.03 . . 0.02 1.00 0.75 1.02 1.21 0-93 1.12 0.90 1-35 1.20 1.29 0.61 0.54 1.14 0.87 0.96 0.93 1.13 0.68 0.48 0.63 1.13 1-35 0.95 0-95 0.79 0.69 1.08 1.07 0.66 0-35 1.00 0.44 1.15 1.80 0.78 1.62 0.70 2-06 1.54 2.16 0.20 0.13 1.50 0.45 0.90 0.70 1.25 0-24 0.80 0.30 1.45 2.30 Streaks Streaks 0.33 0.25 1.17 1.25 0.14 0.08 1.00 0.44 1.06 1-22 0-96 0.87 0.86 1.28 1-17 0.98 0.05 0.04 1-12 0.46 0.72 0-52 0.82 0 0.50 0.90 1-30 0.98 1.02 0.45 0.04 0.82 0.7 1 0.10 0.02 1.00 1.00 0.51 0.58 1.02 Streaks 1.58 1.15 1.04 0.75 1.17 1.14 0.86 0.74 1.35 1-15 1.20 1.10 1.23 1.30 0.11 0.01 0.06 0.01 1.36 1.22 0.42 0.94 0.67 1.10 0-29 0.14 0 0.60 0.51 1-03 1.23 1.40 1.65 1.00 0.98 1.02 0.89 0.41 0.01 0.08 0.35 0.87 0.92 0.75 0.80 0.09 0.02 0.01 0-02 0'43 0.84 1.00 Streaks 1-12 1.35 Streaks 1.10 0.75 1.20 1.15 1-50 0.07 0 1.23 0.88 0.97 0.53 0 0.45 1.11 1-70 Streaks Streaks 0.21 0.01 0.88 0.80 0 0434 DHONT AND DIKJMAN: QUALITATIVE ANALYSIS OF COMPLEX TABLE IV [Analyst, Vol.92 RF VALUES FOR fi-TOLUYLALDEHYDE 2,4-DINITROPHENYLHYDRAZONE I N DIFFERENT SYSTEMS System No. I11 IV V VI VII VIII IX Number of experiments (n) 9 13 6 20 26 26 17 Mean RF from n experiments 0.49 1 0.713 0.410 0.400 0-548 0-390 0.310 Standard deviation for a single observation 0.033 0.050 0.015 0.069 0.067 0.076 0.059 Osazones, if present in the mixture of the aromatic-type compounds, have low R, values in all of these systems and can therefore be separated from them by the choice of a suitable system. They can be differentiated from the aromatic-type 2,4-dinitrophenylhydrazones by spraying the chromatogram with an ethanolic solution of potassium hydroxide.The osazones then show up as intense blue spots while the latter compounds give brown to red - brown spots. When fractions are obtained from the 2-phenoxyethanol plate of system 11, the immobile phase will be extracted from the plate, together with the 2,4-dinitrophenylhydrazones. The phase material, which is very difficult to remove by evaporation, can easily be eliminated from the solution by using an ion-exchange resin, as described by Schwartz, Johnson and Parks.' RESULTS AND DISCUSSION A large number of synthetic mixtures containing both aliphatic and aromatic-type 2,4-dinitrophenylhydrazones has been analysed by using the procedure described in this paper. Chromatograms obtained in two of these experiments are shown in Fig.1. From Fig. 1. Separation of two different mixtures on non- activated magnesium oxide plates. 2,4-Dinitrophenylhydra- zones of: 1, syringa aldehyde; 2 , furfural; 3, phenylacetaldehyde; 4, mono-derivative of diacetyl; 5, valeraldehyde ; 6, butyl methyl ketone ; 7, pyruvic acid ; 8, anisaldehyde ; 9, propionaldehyde ; 10, methyl propyl ketone. The 2,4-dinitrophenylhydrazone of p-toluylaldehyde has been used as a marker on both plates these chromatograms it is seen that under favourable conditions of composition, even aliphatic aldehydes and methyl ketones are obtained in separate bands. The only pair of 2,4-dinitro- phenylhydrazones tested that did not resolve into their components was that of the cc- and p-methyl naphthyl ketones. In a few instances it was found that some decomposition occurred on the magnesium oxide plate.Nona-2-enal and cinnamaldehyde 2,4-dinitrophenyl- hydrazones when extracted from these plates sometimes showed a faint second spot on being re-chromatographed on a silica-gel plate. Very good separations of aromatic-type 2,4-dinitro- phenylhydrazones were obtained on the acetylcellulose plates and some of the chromatograms obtained on these plates are shown in Fig. 2, the notations A, B and C indicating the use of the solvents of the systems 111, IV and V of Table 11, respectively. The 2,4-dinitrophenyl- hydrazones, having low R, values in system I11 (Fig. 2 A), show less reproducible Rs values in this system, as a second front is developed just above the start, the position of which is difficult to control.This second front did not influence the higher Rs values. The region in which this second front appears has been indicated in Fig. 2 A.July, 19671 CARBONYL MIXTURES BY THIN-LAYER CHROMATOGRAPHY 435 The use of acetylcellulose for the separation of 2,4-dinitrophenylhydrazones has been described by Forss and Ramshaw.8 They studied the separation of a homologous series of aliphatic-type 2,4-dinitrophenylhydrazones by paper chromatography on sheets of acetyl- cellulose. For comparison, the positions of the 2,4-dinitrophenylhydrazones of the saturated aliphatic aldehydes on our acetylcellulose plates are shown in Fig. 3. C 0 6 0 5 3 4 3 3 f3: Fig. 2. Mixture of 2,4-dinitrophenylhydr- azones of some aromatic carbonyl compounds in the three solvents: A, light petroleum - meth- anol - ethyl acetate (9.5 + 1.5 + 1.0); B, light petroleum - dibutyl ether - ethyl acetate - meth- anol (6.5 + 1.0 + 1.5 + 1.0) ; C, light petroleum - dibutyl ether - ethyl acetate - methanol (8.5 + 0.5 + 0.5 + 0.5).The position of the second front for solvent A is indicated: 1, vanillin; 2, phydroxybenzaldehyde ; 3, anisaldehyde : 4, benzaldehyde ; 5, hydratropic aldehyde ; 6, p-isopropyl-a-mcthylhydrocinnamaldehyde 1.00 0.80 u - 0.60 2 0.40 0.20 0 I I I I I I CI c, cs c, c, c,, c,, Chain length Fig. 3. RF values of 2,4-dinitrophenyl- hydrazones of saturated aliphatic aldehydes as a function of chain length on thin layers of acetylated cellulose; curve X, solvent A, light petroleum - methanol - ethyl acetate (9.5 + 1.5 1.0): curve 0, solvent B, light petroleum - dibutyl ether - ethyl acetate - methanol (6.5 + 1.0 + 1.5 + 1.0); curve 0, solvent C, light petroleum - dibutyl ether - ethyl acetate - meth- anol (8-5 + 0.5 + 0-5 + 0.5) 2,4-Dinitrophenylhydrazones derived from terpene carbonyl compounds, if present in the mixture, are usually mixed with the derivatives of the non-terpene aliphatic carbonyl compounds, particularly in the C,, to C,, region. It is hoped that the results of investigations on the separation of mixtures of aliphatic and terpene carbonyl compounds will be published at a later date. 1. 2 . 3. 4. 5. 6. 7. 8. REFERENCES Badings, H. T., and Wassink, J . G., Ned. Melk- e a Zuiveltijdschv., 1963, 17, 132. Schwartz, D. P., and Parks, 0. W., Microchern. J . , 1963, 7, 403. Urbach, G., J . Chvornat., 1963, 12, 196. de Jong, K., Mosterd, K., and Sloot, D., R e d . Tvav. Chirn. Pays-Bas Belg., 1963, 82, 837. Dhont, J. H., and de Rooy, C., Analyst, 1961, 86, 74. Ritter, F. J., and Meyer, G. M., Natuve, 1962, 193, 941. Schwartz, D. P., Johnson, A. R., and Parks, 0. W., Micvocheutz. J., 1962, 6, 37. Forss, D. A., and Ramshaw, E. H., J . Chrounat., 1963, 10, 268. Received August 9th, 1965

ISSN:0003-2654    

DOI:10.1039/AN9679200431

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

436 Analyst, July, 1967, Vol. 92, $9. 436-442 The Determination of Total Organic Matter (Carbon Content) in Aqueous Media Part I. Organic Matter in Aqueous Plant Streams BY F. R. CROPPER, D. M. HEINEKEY AND A. WESTWELL (Imperial Chemical Industries Ltd., Dyestufls Division, Hexagon House, Blackley, Manchester 9) The determination of total organic matter (carbon content) in aqueous liquors that are re-cycled in (or rejected from) organic chemical manufacturing processes is necessary so that the carbon balance can be calculated. A rapid method has been developed that involves the injection of the aqueous solution into a combustion tube containing copper(I1) oxide, where all of the organic matter is oxidised to carbon dioxide; this is swept, with added hydrogen, through a catalyst to reduce it to methane, which is determined by using a flame-ionisation detector.Each test with 2 ,u1 of sample containing 1 to 10 pg of carbon takes about 2 minutes. The reproducibility is estimated to be about -J=5 per cent. of carbon content, which is adequate for the analysis of plant side streams. EXISTING methods for the determination of organic matter (as carbon) in aqueous solutions are lengthy and usually involve wet oxidation to give carbon dioxide, which is distilled out and absorbed in alkali solution, followed by classical methods for the determination of carbonate. The work described in this paper is based on the idea that oxidation to carbon dioxide could be made rapidly on an ultra-micro scale, and that the carbon dioxide could be reduced to methane and measured quantitatively by using a flame-ionisation detector.The reduction to methane is a desirable feature in this method; it would be possible to avoid this stage and to determine the carbon dioxide by using thermal conductivity detectors,l but these detectors are relatively insensitive, and require careful control of flow-rates and pressure, whereas the flame-ionisation detector2 93 is highly sensitive to methane, relatively insensitive to changes in flow-rate and pressure, and gives no signal for water vapour. The method that was therefore established involved the injection of 2 pl of the aqueous sample into a combustion tube containing copper(I1) oxide; the liberated carbon dioxide, with added hydrogen, was then swept through a catalyst to reduce it to methane, and the methane was determined by using a flame-ionisation detector. The scale of operations was adjusted to cover the range 1 to 1Opg of carbon, and samples were diluted so that 2 4 contained an appropriate amount of organic matter.This method gives the total carbon content, i.e., organic matter plus any free carbon dioxide in the original sample. If the amount of free carbon dioxide is likely to be significant, it can be determined separately with the same apparatus but injecting the sample at an auxiliary injection port between the combustion tube and the catalytic reduction tube; in this event, the organic matter (calculated as carbon) in the sample is determined by difference. EXPERIMENTAL APPARATUS- The apparatus (Fig. 1) is described in detail below. Earlier versions of the apparatus differed slightly, and improvements were introduced as the work progressed.The following points are relevant to the design of this apparatus. Combusiio~~ unit-The dimensions of this unit are not critical and the tube was made to fit an available electric furnace. It is, however, important to use M.A.R. copper(TT) oxide and to maintain this at 850" to 900" C.CROPPER, HEINEKEY AND WESTWELL 437 D m i.1 A = Carrier gas (nitrogen) inlet B = Injection port C = Combustion tube (850" t o 900" C) D = Electric furnace E = Glass connecting pieces F 5 Auxiliary injection port G = Heater, 150" C H = Reducer gas inlet J = Reduction tube, 300" t o 350' C K = Silica-gel drying tube L = 18-inch U-Column] containing M = T-piece N = Burner gas inlet P = Flame-ionisation detector R = Air inlet S = Amplifier and recorder silicagel Fig.1. Details of apparatus Auxiliary injection port and heater-The complete apparatus has to be calibrated with known volumes of methane, and this is conveniently achieved by injecting methane with a Hamilton syringe through an auxiliary injection port (fitted with a silicone rubber plug). As it is desirable that the methane peak in the calibration runs should be the same shape as in tests on aqueous solutions, the auxiliary injection port was placed close to the exit from the combustion furnace. This injection port is also convenient for the injection of aqueous samples when it is necessary to determine the free carbon dioxide, in which event the heater containing glass beads at 150" C vaporises the sample.Reducer gas inlet-The effluent from the combustion tube (nitrogen, carbon dioxide and water vapour) must be enriched with hydrogen before the stream enters the reduction tube, and an inlet pipe is therefore necessary at this point; hydrogen from a cylinder could be used, but hydrogen - nitrogen, 1 + 1 v/v, was more convenient in this laboratory as cylinders containing this mixture were already available for gas-chromatographic work. It is maintained at 300" to 350" C for conversion of carbon dioxide to methane; this tube is also used (at 250" to 260" C) for the reduction of nickel oxide to nickel in the preparation of the catalyst packing. Silica-gel drying tube and column-The effluent from the reduction tube (nitrogen, hydrogen, water vapour and methane) could be passed directly to the flame-ionisation detector.It is known, however, that the response of this detector is influenced by excess of water vapour in the gas stream and it is desirable that the bulk of the water vapour should be removed first; in addition, the time taken from the moment of injection of the aqueous solution to emergence of the methane from the reduction tube is so small that the methane would be recorded before the apparatus had recovered from the surges that occur when the sample is injected and the water absorbed (on the silica gel). The 18-inch column of silica gel was inserted to delay the methane and to improve the shape of the peak; in the earlier work it also served to hold the excess of water, but to facilitate regeneration, the short tube of self-indicating Reduction tube-The dimensions of this tube are not critical.438 CROPPER et al.: DETERMINATION OF TOTAL ORGANIC [Analyst, Vol. 92 silica gel, K, was inserted so that the operator would know when this section of silica gel should be regenerated. Detector and electrical equipment-The detector and its associated electrical equipment were of the type described by Cropper and Kamin~ky.~ This is a simple apparatus that can be built in the laboratory, but any commercial flame-ionisation detector and associated equipment would be suitable. METHOD SPECIAL REAGENTS- Hydrogen - nitrogen, 1 + 1 v/v-Cylinders, obtainable from British Oxygen Co. Co$per(II) oxide-Wire form, M.A.R. grade, obtainable from B.D.H.Ltd. Sel f-indicating siEica gel-Coarse grade, obtainable from B.D.H. Ltd. Silica gel, 30 to 50 mesh-This is obtainable from Silica Gel Ltd., London, W.l. Dry before use by heating in an oven at 105" to 110" C for 2 hours. Methane-Cylinders, obtainable from I.C.I. Ltd., Agricultural Division. Attach to the reducing valve on the cylinder a short length of silicone rubber tubing fitted at the end with a spring clip. Blow off some methane to expel air from the tube and close the spring clip. Mrithdraw methane by inserting a hypodermic syringe through the wall of the tubing. Nickel -3rebrick catalyst5-Weigh 10 g of 30 to 60-mesh firebrick (Sil-0-Cel 622 from Johns-Manville) into 50 ml of a saturated aqueous solution of nickel nitrate and mix well; remove the surplus liquor by filtration, with gentle suction, through a Buchner funnel.Dry the filter-cake overnight at 105" to 110" C and then heat in a muffle furnace in a fume cupboard for 5 hours at 400" to 500" C; at this stage oxides of nitrogen are evolved. Pack the dried material into the 13-cm x 6-mm bore reduction tube (see Special appara- tus) to give a 10-cm length of packing held in position by silica-wool plugs. Attach a hydrogen - nitrogen, 1 + 1 v/v, supply line and attach a glass exit pipe at the other end of the reduction tube; adjust the hydrogen-nitrogen flow to 20ml per minute and burn the effluent gas at the exit pipe. Switch on the electrical heater round the reduction tube and adjust the current so that the temperature is 250" to 260" C. Allow the hydrogen - nitrogen to flow overnight, to effect the reduction of nickel oxide to metallic nickel; the catalyst is now ready for use, having been prepared in situ in the reduction tube.Glycerol solzdion-Weigh accurately about 1 g of glycerol into a 100-ml graduated flask, dissolve it in boiled-out water, make up to volume with boiled-out water and mix well. Retain a sample of the boiled-out water for blank tests. 1 p1 of solution = weight of glycerol taken (g) x 3.9 pg of carbon. SPECIAL APPARATUS- The apparatus is shown assembled in Fig. 1, and consists of the following parts- Combustion unit-This is a straight silica tube, C, 30-cm x 7-mm bore (with a side-arm, A), packed with a 19-cm length of copper(I1) oxide held in position by a silica-wool plug. The packed tube is fitted vertically into a 20-cm x 10-mm bore mains electrical furnace, D, so that the copper(I1) oxide packing is held entirely within the furnace; the packing is held at 850" to 900" C by adjusting a variable transformer. The upper end of the combustion tube is fitted with a silicone rubber plug for use as the injection port, B.The side-arm, A, is connected to a supply of nitrogen (carrier gas) from a cylinder, the flow-rate being measured by a capillary flow-meter. The lower end of the combustion tube is drawn out to give an exit pipe of about 6mm 0.d. Auxiliary injectio.iz port and heater-This consists of an 8-mm o.d. Pyrex tube, G, with two side-arms separated by about 5 cm. The tube between the side-arms is fitted with an electrical heater of Nichrome resistance wire wound in the following way: wind a 40-inch length of 30-gauge Nichrome resistance wire (resistance 14 ohms) round a 20-gauge needle to give a tightly coiled spring about 34 inches long, and leave 3 inches of the straight wire at each end.Stretch this spring evenly so as to separate each coil and to give an over-all length of about 5 inches. Wind this coil round the glass tube between the side-arms to give a helix of 4 to 5 turns. Bind each end of the helix in position with fine copper wire. Separate and insulate each turn of the helix by winding on a helix of asbestos string, and then cover the wire with asbestos tape. Insulate the 3-inch straight ends of the heater wire with rubber tubing and connect to a 12-volt supply.July, 19671 MATTER (CARBON CONTENT) IN AQUEOUS MEDIA.PART I 439 Fill the tube between the side-arms with glass beads (2mm diameter) and hold them in place with silica-wool plugs. Adjust the temperature of the glass beads to 150" C and close the top of the heater by a silicone rubber plug, F, which serves as the auxiliary injection port. Reduction tube-This consists of a 13-cm x 6-mm bore Pyrex glass tube, J, wound with an electrical heater in a similar way to that previously described, except that 75 inches of 30-gauge Nichrome resistance wire is used to give a total resistance of 26 ohms. This is connected to a variable 25-volt supply. A temperature of 300" to 350" C is used when reducing carbon dioxide to methane, but the temperature is adjusted to between 250" and 260°C when the catalyst is being prepared.Silica-gel drying tube-This is a 13-cm x 6-mm bore Pyrex tube packed with self- indicating silica gel and held in position with silica-wool plugs. The drying tube is regenerated by heating when about three-quarters of the length of the silica gel has changed colour. Silica-gel column-This is an 18-inch x 4 6 m m bore U-shaped column, L, packed with 30 to 50-mesh silica gel, held in place with silica-wool plugs. Burner gas inlet-This is a T-piece, M, between the silica-gel column and the detector, connected to a supply of hydrogen -nitrogen, 1 + 1 v/v. Flame-ionisation detector and associated equipment-As described by Cropper and K amins k y . The above parts are connected together (Fig. 1) by using short glass connecting pieces (e.g., E, 4.5-mm bore) and silicone rubber tubing as connecting sleeves.PROCEDURE- Connect the nitrogen supply at inlet A, and connect the hydrogen - nitrogen supplies at inlets H and N. Switch on the combustion furnace, D, heater, G, and reduction tube heater, J, and start the flow of nitrogen at 10ml per minute. Allow about 90 minutes for the temperatures to become steady, i.e., 850" to 900" C in the centre of the combustion tube, 150" C in the heater, G, and 300" to 350" C in the reduction tube. During this period the nitrogen will purge the air from the apparatus. Start the flow of reducer gas at H, and of burner gas at N, and start the supply of air to the detector (400 ml per minute). Adjust the gas flows so that the apparatus is receiving 10 ml of nitrogen, 20 ml of reducer gas and 50ml of burner gas per minute.WARNING-The gas flows must be set up in the above order so that,the reducer gas does not pass back Switch on the electrical equipment for operating the detector and adjust the sensitivity controls so that a peak of about two-thirds full-scale deflection is obtained after injecting 20 pl of methane at the auxiliary injection port; use a recorder with a chart speed of 30 inches per hour. into the combustion tube, with consequent risk of explosion. CALIBRATION- With a 25-p1 Hamilton syringe, inject 5, 10, 15 and 20-4 loads of methane successively, via the auxiliary injection port. The methane peaks will be recorded about 45 seconds after each injection. Measure the areas of the peaks by peak height x width at half-height, and plot a graph of area against weight of carbon in micrograms, taking lpl of methane as equivalent to 0-5 pg of carbon.The graph should be a straight line passing through the origin. CHECK ON THE PERFORMANCE OF OXIDATION AND REDUCTION STAGES- With a 10-pl Hamilton syringe, inject successively several portions of 2 - 0 4 of the 1 per cent. glycerol solution via the injection port at the top of the combustion tube, allowing about 3 minutes between each injection; the needle should be inserted so that it penetrates about a t inch into the copper(I1) oxide. The peak given by the 2-p1 load of 1 per cent. glycerol solution should be about two-thirds full-scale deflection. Carry out a blank test by injecting 2 p1 of the boiled-out water used in making the glycerol solution.Measure the areas of the methane peaks and deduct the area (if any) for the "blank" on the boiled-out water. Calculate the mean net area for the replicate tests on 2 p1 of 1 per cent. glycerol, and calculate micrograms of carbon injected for the particular solution in use. Ascertain where this point (mean net area against micrograms of carbon)440 CROPPER et al. : DETERMINATION OF TOTAL ORGANIC [Analyst, Vol. 92 will fall on the calibration graph obtained for methane. If the point is within 5 per cent. of the line for methane, it can be taken that the oxidation and reduction stages are working satisfactorily, in which event the factor- F = net area carbon, pg can be used in calculating the results on plant samples. PREPARATION OF SAMPLE SOLUTIONS- Adjust the strength of plant samples by adding boiled-out water to give 0.2 to 0.5 per cent.solutions, i.e., so that 2 pl of the diluted sample will contain 5 to 10 pg of carbon. Retain a sample of the boiled-out water for the blank test. TESTS ON SAMPLE SOLUTIONS AND CALCULATIONS- (a) Total carbon content-Inject at port, B, 2.0 p1 of the diluted sample solution exactly as described for the 1 per cent. glycerol solution; carry out a blank test on 2 pl of the boiled-out water. Measure the areas under the methane peaks, and deduct the area (if any) for methane given in the blank test. Then, if V , ml of original sample was diluted to I/, ml, the total carbon in the original sample is- Let the net area for the diluted sample solution be A. 0.05 A V2 Percentage w/v of total carbon =- x - =C,.F V l (b) Free carbon dioxide-Inject at port, F, 2.0 pl of diluted sample solution and proceed exactly as in section (a) above. Let the net area for the diluted sample solution be B, Then free carbon dioxide in the original sample is- 0.05 B Vz percentage w/v of free carbon dioxide (calculated as carbon) = ~ x-= c,, Vl and percentage w/v, calculated as carbon dioxide = 3.67C2. (c) Total organic matter-Total organic matter in the original sample = C , - C,. In many instances it may be ascertained that the amount of free carbon dioxide is insignificant compared with the total carbon content, in which event test (b) can be omitted in future. REGENERATION OF THE COMBUSTION TUBE (see Results and discussion)- Turn off the nitrogen and reducer gas supplies, and detach the combustion tube from the rest of the apparatus.Detach the nitrogen line and attach an air line; allow air to flow through the combustion tube (still at 850" to 900" C) for 1 hour at 50 to 100 ml per minute, then detach the air line and reconnect the nitrogen supply. Purge the combustion tube with nitrogen for 5 minutes, and then reconnect the combustion tube to the rest of the apparatus. REGENERATION OF SILICA-GEL TUBE, K- 130" C until the original colour is restored. Remove the silica-gel drying tube, K, from the apparatus and heat it in an oven at RESULTS AND DISCUSSIOK CHOICE OF CONDITIONS- The conditions used in this method were adjusted to suit the detection of 1 to 10 pg of methane (which is well within the capabilities of a flame-ionisation detector) and to give a peak for methane within 2 minutes of injection.Having set up the apparatus and calibrated it by using known volumes of methane, it was necessary to check that injection of carbon dioxide into the reduction tube would give complete conversion to methane. This was carried out by injecting 5 to 20 pl of carbon dioxide into the auxiliary injection port and calculating the percentage conversion by using the methane calibration. With the reduction tube tem- perature below 150" C no methane was produced, and at 200" C the peak was broad and shallow ; as the temperature was increased, the peak became progressively sharper, and above 275" C the peak size and shape remained constant. A reduction tube temperature of 300" to 350" C was therefore adopted.The reducer gas (hydrogen - nitrogen) flow-rateJuly, 19671 MATTER (CARBON CONTENT) IN AQUEOUS MEDIA. PART I 441 also affected the peak size, especially below 10ml per minute, but at the chosen flow-rate of 20 ml per minute the effect was only +_ 1 per cent. in area for variations of f 1 ml per minute; with hydrogen in place of hydrogen - nitrogen, however, the variations were con- siderably greater. Carrier gas flow-rate was fixed at 10 ml per minute (a convenient rate for purging the packed combustion tube several times per minute) and the burner gas flow-rate was fixed at 50 ml per minute, so that the total gas rate at the detector was 80 ml per minute. Variations of +5 ml per minute in carrier gas and burner gas flow-rates had negligible effect (less than 0.5 per cent.) on peak size.Under the chosen conditions, complete conversion to methane was obtained within experimental error of k5 per cent. It was then necessary to check that the oxidation of organic compounds in dilute aqueous solution would be complete; preliminary tests showed that the peak area was independent of the speed of injection of the load from the Hamilton syringe only when it was 2 pl or less, and a load of 2 p1 was therefore fixed for future work. A series of tests on 2 - 4 loads of 0.4 to 1.6 per cent. glycerol showed that peak area was proportional to the amount of carbon injected; the line for peak area against micrograms of carbon passed through the origin, but it lay slightly higher than that for methane, possibly owing to slightly different peak shapes given by injection of 2 pl of aqueous solutions (which give about 3 ml of water vapour) as against injection of 5 to 20 p1 of gas.The tests proved that oxidation to carbon dioxide and reduction to methane were complete. In the routine use of this apparatus it is necessary to check occasionally that the apparatus continues to give complete conversion, by running tests on methane and on glycerol solution; this must be done because there will be a gradual accumu- lation of copper at the top of the combustion tube, and if involatile organic matter is deposited on this copper in subsequent tests there will be incomplete oxidation. For this reason, the combustion tube contents are regenerated (by passage of air) if the results on glycerol are more than 5 per cent.below the corresponding level for methane. For routine tests, however, it is convenient to use the glycerol solution as a “secondary standard,” i.e., to compare peak areas for samples with the peak area given by 2 p1 of 1 per cent. glycerol solution, as given in the Method above. The results shown in Table I were obtained in this way with a series of known solutions of different compounds. TABLE I RESULTS WITH KNOWN AQUEOUS SOLUTIONS Components in aqueous solution Carbon, per cent. w/v r A \ Calculated from weight of substance taken Found Ethanol (about 0.5 per cent.) . . . . .. . . . . 0.26 Isopropyl alcohol (about 1 per cent.) .. .. .. 0.62 Cyclohexanol- cyclohexanone mixture (about 0.5 per cent.) 0.73 Adipic acid (about 1 per cent.) . ... .. .. 0.49 ] 0.64 Mixture of 0.27 per cent. of cyclohexanol, 0.24 per cent. of cyclohexanone, 0.5 per cent. of adipic acid and excess of involatile inorganic matter* .. .. . . .. 0.27 0.61 0.75 0-50 0.65 * A “blank” solution containing only this inorganic matter gave no visible movement on the chart base-line. APPLICATIONS- The method was designed in the first place for application to aqueous liquors containing up to 12 per cent. of organic compounds (containing only carbon, hydrogen and oxygen) with, or without, an involatile inorganic constituent. It was expected that free carbon dioxide would be at a negligibly low level, and the test for total carbon therefore sufficed. As the range on the injected solution was 0 to 0-5 per cent.w/v, calculated as carbon, it followed that stronger solutions were diluted with boiled-out water before testing. In the second application, the aqueous liquors were strongly ammoniacal and contained up to several per cent. of a nitrogen-containing compound, and it was expected that a considerable amount of “free” carbon dioxide would be present (as ammonium carbonate) ; tests on known solutions of adiponitrile with excess of AnalaR “ammonium carbonate”442 CROPPER, HEINEKEY AND WESTWELL (NH4HC0,.NH,COONH4) gave satisfactory results for both total carbon and for “free” carbon dioxide. It was evident that “ammonium carbonate” gave up its carbon dioxide for conversion to methane under the prevailing conditions, and that these amounts of ammonia have no effect on the response of the detector to methane. In the test for free carbon dioxide, it was considered possible that some volatile organic compounds might pass unchanged through the reduction tube and thus give a signal when reaching the detector, but it was ascertained that no interference was caused when 2 per cent.solutions of acetic acid, ethylamine, hexamethylene diamine and adiponitrile were injected at the auxiliary injection port ; these volatile substances are therefore effectively held by the silica-gel column. In the third application, chlorinated amines were present in solution in excess of hydro- chloric acid; as a safeguard, a wad of silver gauze was placed at the exit from the combustion tube, but it is not known if this is necessary. The results with known solutions of chloro- dimethylaniline and chlorodiphenylamine in excess of hydrochloric acid showed complete conversion to methane, with no interference from the hydrochloric acid.The life of the combustion tube packing is not known, but as each test will consume less than 0.3 mg of copper(I1) oxide, it is likely to be long; however, the risk that involatile organic matter might be deposited on copper (or on inorganic residues from previous tests) must not be disregarded, and for this reason occasional check tests with methane and with glycerol solution are obviously desirable. If regeneration of the packing does not eliminate any discrepancy between results for methane and for glycerol, the top portion of copper(I1) oxide should be replaced. CONCLUSIONS It has been shown that total carbon and free carbon dioxide can be determined rapidly in aqueous plant liquors by injection into a combustion tube containing copper(I1) oxide, reduction to methane and measurement with a flame-ionisation detector; the method is applicable to liquors containing organic compounds with carbon, hydrogen, oxygen, nitrogen and chlorine, excess of ammonia or hydrochloric acid and involatile inorganic matter. The range is from 0.05 per cent. w/v upwards, samples that are stronger than 0-5 per cent. w/v being diluted before testing. The reproducibility is estimated to be about +5 per cent. of the carbon figure, which is usually adequate for testing plant re-cycle streams and effluents for calculating the carbon balance in a manufacturing process. The method described in this paper is the subject of U.K. Patent Application No. 48483/66. REFERENCES 1. 2. 3. 4. 5. West, D. L., Analyt. Chem., 1964, 36, 2194. Dewar, R. A., and McWilliams, I. G., British Patent 838,189, July 4th, 1957. -,- , in Desty, D. H., Editor, “Gas Chromatography 1958,” Butterworths Scientific Pub- Cropper, F. R., and Kaminsky, S., Analyt. Chem., 1963, 35, 735. Porter, K., and Volman, D. H., Ibid., 1962, 34, 748. lications, London, 1958, p. 142. Received November 1 lth, 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200436

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Analyst, July, 1967, Vol. 92, $9. 443-449 443 The Determination of Total Organic Matter (Carbon Content) in Aqueous Media Part 11. * Involatile Organic Matter in De-mineralised Water BY F. R. CROPPER, D. M. HEINEKEY AND A. WESTWELL (Imperial Chemical Indwstries Ltd., Dyestuffs Division, Hexagon House, Blackley, Munchester 9) The method described in Part I has been adapted for the analysis of de-mineralised water. The sample is concentrated by evaporation, and the concentrate is injected on to copper(I1) oxide at 850” to 900” C; the organic matter is converted to carbon dioxide, which is reduced to methane and measured with a flame-ionisation detector. The method covers the range 0.1 to 1-5 p.p.m. of involatile organic matter (calculated as carbon) in the original de-mineralised water, and the reproducibility is estimated to be 5 0 .1 p.p.m. DE-MINERALISED water is used in chemical process work and as feed-stock for high pressure boilers, for which purpose it is desirable that the amount of involatile organic matter shall be not greater than 1 p.p.m. Existing methods for the determination of involatile organic matter in de-mineralised water involve evaporating a 2-litre sample to 40 ml, and applying to this concentrate lengthy chemical methods, such as described e1sewhere.l A method is described in Part I of this series2 for the determination of organic matter (above 0.05 per cent.) in aqueous solutions of plant effluents and re-cycle streams. This method involves injection of 2 p1 of sample containing 1 to 10 pg of carbon, i.e., 500 to 5000 p.p.m.To adapt the method to the range 0.1 to 1 p.p.m. it was necessary to investigate several ways of improving the sensitivity, by increasing (a) the sensitivity of detection of methane, ( b ) the amount of solution injected into the combustion tube and (c) the concen- tration by introducing a preliminary small-scale evaporation stage. EXPERIMENTAL The apparatus and method used were essentially as described in Part I, modified in accordance with the results obtained during the work on improving the sensitivity, as detailed below. SENSITIVITY OF DETECTION OF METHANE- The sensitivity of detection could be increased about 20-fold on the apparatus used by adjusting the attenuation of the amplifier, but when an increase much greater than 10-fold was used, even after the delay tube described below was inserted, the base-line showed too much “noise.” A 10-fold change in attentuation was therefore adopted; this had the effect of giving about 80 per cent.full-scale deflection when 2 p1 of methane gas was injected at the auxiliary port. This corresponded, however, to only about a 5-fold increase in sensitivity for water samples because passage of the sample through the combustion tube and the delay tube caused the derived methane peak to be wider and lower in height than it was when methane gas was injected at the auxiliary port. Thus the over-all sensitivity for aqueous samples (as assessed by peak height) was about 5 times that in Part I. AMOUNT OF SOLUTION INJECTED- The “load” size chosen in Part I was 2 pl because this was the maximum volume of sample that could be injected instantly on to the copper(I1) oxide without “blow-back” of water vapour down the nitrogen supply line.It was found, however, that larger loads * For details of Part I of this series, see reference list, p. 449.444 CROPPER et al. : DETERMINATION OF TOTAL ORGANIC [Analyst, Vol. 92 (up to 25 pl) could be used provided the injection is carried out slowly, i.e., so that the flow of nitrogen carrier gas does not fall below 2 ml per minute. A straight-line calibration graph for methane peak area against volume injected was obtained when a standard glycerol solution was used and the load size increased up to 25 pl. Loads of 25 p1, however, took rather too long to inject (15 to 20 seconds) and a load of 10 p1 was chosen for the ultimate method.The injection time is thus about 5 seconds. With this increased load size, it was found that a large proportion of the water that had been injected condensed out from the carrier-gas stream on to the cool wall of the exit from the combustion tube. A T-piece (U, Fig. 1) was therefore inserted in the apparatus to act as a water trap; it was checked that any water held in this trap would not retain detectable carbon dioxide from the carrier-gas stream. A delay tube (E, Fig. 1) was also inserted between the water trap and the reduction tube, so that the methane peak was recorded well after all the disturbance caused by the pressure changes that occurred when the sample was vaporiscd, and when vapour was absorbed in the silica gel.In addition, the silica-gel tube was lengthened to 54 inches. The technique used to inject the lop1 of sample was modified to avoid a relatively large “blank.” This procedure involved filling the syringe with about 12 pl of sample, inserting it through the rubber septum, lowering the plunger to the 10-pl mark, leaving the syringe in position for about 3 minutes, and then depressing the plunger slowly (during 5 seconds) to inject the l o p 1 of sample. The reasons for this procedure are given in the Discussion. CONCENTRATION STAGE- Only a very small quantity of concentrate is required for injection, and the concentration stage can therefore be carried out on a small scale; for quantitative results, however, the volume of concentrate from a known volume of sample must be measured, and hence about 1 ml of concentrate is required.As 10 pl of injected solution should contain 0.2 to 2 pg of carbon (20 to 200 p.p.m.), it follows that a 100-fold concentration would be suitable for the range 0.2 to 2 p.p.m. It was found convenient to evaporate 150 ml of the de-mineralised water sample in a beaker on a hot-plate until only about 1 to 1.5 ml remained, and to measure this volume in a graduated pipette. The detailed procedure is described in the Method below. METHOD APPARATUS- This is described in Part I. The three modifications, as shown in Fig. 1, were as follows. Water trap (U)-This is a T-piece of 4 to 5-mm bore, inserted between the combustion tube and the connecting tube, E, with the lower end closed with a silicone rubber bung.The condensed water collects in this lower end, and at intervals the water is removed by inserting a hypodermic syringe needle up through the silicone rubber bung. Delay tube (E)-This connecting tube, E, in Part I is replaced by an 18-inch length of Pyrex glass tubing folded into a convenient shape to keep the apparatus compact; the folds are not shown in Fig. 1. Silica-gel column (M)-This is as in Part I, but lengthened (and folded) to 54 inches, and packed almost full with 30 to 50-mesh silica gel, held in place with silica-wool plugs. SPECIAL REAGENTS- As described in Part I, with the exception that the 1 per cent. standard glycerol solution is replaced by a series of dilute standard solutions prepared as follows. Glycerol soldions-Weigh accurately about 1 g of glycerol into a 100-ml graduated flask, dissolve it in boiled-out de-mineralised water, make up to volume with boiled-out de-mineralised water and mix well.This is solution A . 1 pl of solution A = weight of glycerol taken x 3.9 pg of carbon. Measure 0.1, 0.2, 0.3, 0.4 and 0.5 ml of solution A into a series of 10-ml graduated flasks, make up to volume with boiled-out dc-mineralised water and mix well. These solutions, S,, S,, S,, S , and S,, correspond to the range from 0.39 to 1.95 x weight of glycerol taken, expressed as pg of carbon in 10 p1. Retain a sample of the boiled-out de-mineralised water for blank tests.July, 19671 MATTER (CARBON CONTENT) IN AQUEOUS MEDIA. PART 11 445 PROCEDURE- Set up the apparatus as shown in Fig. 1, i.e., connected in the order: combustion tube, C, water trap, U, delay tube, E, auxiliary injection port and heater, F, connecting tube, J, eeduction tube, K, silica-gel drying tube, L, silica-gel column, M, T-piece, N, and flame- I onisation detector, R.Connect the gas supplies, i.e., nitrogen, to side-pipe, A, and hydrogen - I itrogen at the lower side-arm, H, and at T-piece, N. B R A = B = c = D = E = F = G = H = K = I = Switch on Carrier gas (nitrogen) inlet Injection port Combustion tube (850" t o 900" C) Electric furnace 18-inch Delay tube Auxiliary injection port Heater (I 50" C) Reducer gas inlet Glass connecting pieces Reduction tube (300" t o 350" C) L = Silica-gel drying tube M = 54-inch folded-column containing silica gel N = 1-piece P = Burner gas inlet R = Flame-ionisation detector S = Air inlet T = Amplifier and recorder U = Water trap Fig.1. Details of apparatus the combustion furnace, D, the heater, G, and the reduction tube heater, I(, and start the flow of nitrogen at 10 ml per minute; allow about 90 minutes for the temperatures to become steady, i e . , at 850" to 900" C in the centre of the combustion tube, 150" to 200" C at the heater, G, and 300" to 350" C in the reduction tube. During this period the nitrogen will purge the air from the apparatus. Start the flow of reducer gas at H, keeping a watch on the flow of nitrogen; then start the flow of burner gas at P, and start the supply of air to the detector at 400 ml per minute. Adjust the flows of gases so that the apparatus is receiving 10 ml of nitrogen, 20 ml of reducer gas and 50 ml of burner gas per minute.WARNING-The gas flows must be set up in the above order so that the reducer gas does not pass back Switch on the electrical equipment for operating the detector and adjust the sensitivity controls so that a peak of about 80 per cent. of full-scale deflection is obtained after injecting 2 pl of methane at the auxiliary injection port; use a recorder with a chart speed of 30 inches per hour. into the combustion tube, with consequent risk of explosion.446 [Analyst, Vol. 92 CALIBRATION- With a 10-p1 Hamilton syringe, inject 0.5, 1.0, 1.5 and 2.0 pl loads of methane (special reagent, Part I) successively, via the auxiliary injection port. The methane peaks will be recorded 1 minute after each injection.Measure the areas of the peaks by peak height times width at half-peak height, and plot a graph of area against weight of carbon in micro- grams, taking 1 p1 of methane as equivalent to 0.5 pg of carbon; the graph should be a straight line passing through the origin. With a 50-pl Hamilton syringe inject lop1 of standard glycerol solutions S,, S,, F3, S , and S , successively, via the injection port at the top of the combustion tube by using the following technique- Fill the syringe with about 12 p1 of the solution and insert the needle through the injection port so that the tip penetrates about a quarter of an inch into the copper oxide. Lower the plunger of the syringe until 10 pl of solution remains in the syringe. Leave the syringe in position until the methane peak has been recorded (about 3 minutes) and then inject the remaining 10 pl of solution slowly and steadily on to the copper(I1) oxide, taking care (by watching the nitrogen gas flow-meter) that the flow of nitrogen does not fall to less than 2 ml per minute; this last stage of the injection should take about 5 seconds. When the 10 pl has been injected, remove the syringe from the septum; the methane peak, which is to be measured, is recorded about a further 3 minutes after the injection of the lop1 of solution.A typical recorder chart is shown in Fig. 2. The peaks recorded just after each injection are caused by flow and pressure fluctuations in the system during injection, and during condensation and absorption of water. CROPPER et al. : DETERMINATION OF TOTAL ORGANIC Minutes Fig.2. Typical chromatogram Carry out a blank test by injecting 10 pl of the bciled-out de-mineralised water used in making the glycerol solutions (see “Blank” on boiled-out de-mineralised water). In addition, carry out check tests to determine any free carbon dioxide (which may have been absorbed from the atmosphere since the boiled-out water was prepared) by injecting 10 pl of the glycerol solutions and the boiled-out de-mineralised water through the auxiliary injection port. Measure the areas of the methane peaks obtained when the 10 pl of glycerol solutions and the boiled-out de-mineralised water were injected into the combustion tube; subtract the corresponding areas of the methane peaks (if any) produced when the 10 pl of glycerol solutions and the boiled-out de-mineralised water were injected through the auxiliary port.The peak given by glycerol solution, S,, should be about 80 per cent. of full-scale deflection. Plot the net area (i.e., for glycerol minus free carbon dioxide) against micrograms of carbon on the same graph sheet as used €or methane. The two lines should coincide or be parallel with the line for glycerol, cutting the carbon axis between 0 and -0-2 pg, in which event the apparatus can be used for tests on samples (see Checks on performance of apparatus); the calibration line for glycerol must be used to give the results on samples (see Discussion).July, 19671 MATTER (CARBON CONTENT) IN AQUEOUS MEDIA. PART 11 447 PREPARATION OF SAMPLE CONCENTRATE- Clean a 100-ml beaker with chromic - sulphuric acid, wash well with water and finally with de-mineralised water; do not dry the inside of the beaker by wiping with a cloth.Measure 50ml of the sample of de-mineralised water into the beaker and boil it down on a hot-plate; add more of the de-mineralised water until a total of 150 ml has been used, taking care not to let the beaker boil dry at any stage. When all of the 150ml of water has been added and the volume remaining in the beaker is about 5 ml, heat more gently until the volume remaining is about 1 to 1.5 ml. Swirl the beaker to wash down the walls, stand the beaker on its edge, let the condensate collect at the bottom, draw it up into a 2-ml graduated pipette and measure its volume. Let this volume be R ml. Transfer the concentrate to a small trident vial.This concentrate is now ready for testing. TEST ON SAMPLE CONCENTRATE AND CALCULATION- Inject 10 p1 of the concentrate through the injection port at the top of the combustion tube by using the same technique as for the glycerol standards; when this test is complete, also inject 10 p1 through the auxiliary injection port. Measure the area under the methane peak given by the 10-pl injection in the test, and deduct the area (if any) for methane given by injection at the auxiliary port (caused by free carbon dioxide picked up after completing the concentration stage). Read from the glycerol calibration the amount of carbon (in micrograms) corresponding to the net area; let this be Tpg. RT 1.5 Total involatile organic matter, calculated as carbon = - p.p.m.(w/v). “BLANK” ON BOILED-OUT DE-MINERALISED WATER- The boiled-out de-mineralised water used in making the standard glycerol solutions may give a peak corresponding to about 0-15 pg of carbon from a 10-pl load. This is not caused by organic matter in the water, but is a “blank” on the apparatus (see Discussion). If when the apparatus is first set up, the blank is much greater than this, it should be reduced by carrying out the following procedure- Detach the combustion tube from the rest of the apparatus and adjust the nitrogen flow to 10 ml per minute. Slowly drip de-mineralised water through the injection port at the top of the combustion tube, taking care that the nitrogen flow-rate does not fall to less than 2 ml per minute. When about 5 ml of water have been passed through the combustion tube, re-assemble the apparatus, inject 10 pl of boiled-out de-mineralised water and measure the area of the methane peak; this should now correspond to less than 0-2 pg of carbon, and, if so, the apparatus can be used.CHECKS ON PERFORMANCE OF APPARATUS- The injection of samples gradually converts the copper(I1) oxide in the combustion tube to metallic copper, and there will be a gradual accumulation of copper at the top of the combustion tube; there will, therefore, be a risk that oxidation will become incomplete as tests proceed, and this risk must be avoided by repeated checks on the performance of the apparatus. It is therefore recommended that, after every 10 injections on sample solutions, the operator should inject 10 pl of the standard glycerol solution, S,, and compare the peak area with that given on the calibration graph; if the area coincides with the calibra- tion line, the apparatus is in good order, but if the area is significantly different from that given on the calibration line, the previous 10 results should be rejected, and the combustion tube regenerated as described in Part I.The silica gel in L should also be regenerated as described in Part I. RESULTS The results on typical samples of de-mineralised water are shown in Tables I and 11. The reproducibility is shown in Table I ; several 150-ml portions of each sample were con- centrated to between 1 and 1.5 ml, and the amount of carbon in each concentrate was deter- mined in duplicate.448 CROPPER et al.: DETERMINATION OF TOTAL ORGANIC TABLE I REPRODUCIBILITY OF RESULTS [A92alyystJ Vol. 92 Concentrate Sample A, Sample B, Sample C, 1 0.65, 0.65 0.60, 0.60 0.55, 0.60 2 0.51, 0-50 0.75, 0.80 0.75, 0.75 3 0.55, 0.60 0.75, 0.70 - 4 0.47, 0-50 - - 5 0.50, 0.55 - - number p.p.m. w/v of carbon p.p.m. w/v of carbon p.p.m. w/v of carbon Mean .. .. 0.55 0.70 0.66 The reproducibility is about kO.1 p.p.m., which is adequate for testing against a limit of 1 p.p.m.; the variation between concentrates is considerably greater than that between duplicates on a given concentrate, and this suggests that, if required, better reproducibility would be obtained by scaling up the concentration stage, e.g., evaporate 1 litre and make up the concentrate to exactly 10ml. Table I1 shows the results on a further set of samples; only one concentrate of each sample was prepared, but each concentrate was tested in triplicate or quadruplicate. The table also gives results by a potassium dichromate - sulphuric acid oxidation method similar to that described e1sewhere.l TABLE I1 RESULTS WITH FURTHER SAMPLES, EXPRESSED AS P.P.M.w/v OF CARBON Water from source A Water from source B w-7 Exit from Exit +from r Exit from bxit from cation anion Clarified cation anion Exit from Type of sample exchanger exchanger water exchanger exchanger filters By this method, i.e., oxi- Individual 1.55, 1.8 0.4, 0.4 1.5, 1.4 1.05, 1.15 0-35, 0.45 2.3 dation and reduction results 1.75, 1.7 0.5, 0.5 1.45, 1.4 1.45, 1.1 0-45 3-0 to methane } Mean 1.7 0.45 1.45 1.2 0.4 2.7 By chemical method similar to that - - - described elsewherel .. . . 1.5 0.3 1.4 The agreement between results by the two methods is regarded as satisfactory. DISCUSSION It is appreciated that the method would be much more attractive if the concentration step were eliminated; it is, however, likely that such a method would require considerably greater care as regards the purity of the copper(I1) oxide, carrier gas and hydrogen, and would necessitate quite considerable expenditure on commercially built apparatus. The method described above has been developed by using relatively cheap equipment and ordinary quality reagents, and will, therefore, probably be more readily acceptable to industrial laboratories. When the method was being developed on boiled-out de-mineralised water (not concen- trated) large and variable methane peaks were obtained; this was found to result from two causes.One arose from contamination by injection of the syringe needle through the silicone rubber septum, either by silicone rubber itself, or by additives in the rubber that would give a film on the outside of the needle. This difficulty was overcome by using the two-stage injection technique described above, so that any Contamination was burnt off the needle (and allowed to emerge as methane) before the 10-p1 load was injected. With fresh copper(I1) oxide packing, the results on injecting boiled-out de-mineralised water (with the two-stage injection technique) were very high (0.3 to 0.5 pg of carbon), and the result decreased with repeated injections with water. This “blank” was brought down to about 0.15 pg for the 10-p1 load by giving the combustion tube packing a preliminary treatment with 5 ml of water, as described in “Blank” on boiled-out de-mineralised water.The “blank” had thus been brought to a low and steady level of about 0.15 pg of carbon per 10 pl, i.e., 15 p.p.m. in the injected load. This value could be caused either by organic matter or carbon dioxide in the de-mineralised water, or to a “blank” given by the apparatus; The other was due to a “blank” value on the combustion tube packing.July, 19671 MATTER (CARBON CONTENT) IN AQUEOUS MEDIA. PART 11 449 injection at the auxiliary port showed that it was not caused by free carbon dioxide, and concentration from 150 ml to between 1 and 1.5 ml and analysis of the concentrate showed it was not caused by involatile organic matter.The water was refluxed overnight with potassium dichromate - sulphuric acid and distilled, but the distillate still gave the same result, and, as this treatment would be likely to destroy volatile organic matter, it was concluded that the blank was caused by the apparatus. The importance of this blank is evident when it is found that the calibration line for glycerol is parallel to the line for methane but cuts the carbon axis at about -0.2 pg. This intercept value (e.g., 0-2 pg) corresponds to the peak area given by 10 pl of the de-mineralised water used in making the glycerol solution, i e . , without added glycerol and without the concentration stage, but corrected for any free carbon dioxide. This de-mineralised water should be checked for involatile carbon content by concentrating and injecting lop1 of concentrate, but the result should be very low, e.g., 1 p.p.m., corresponding to only 0.01 pg when 10 ,ul of the de-mineralised water is injected (without previous concentration).In these circumstances, it is taken that the intercept value (e.g., 0.2 pg) is an apparatus blank. It follows that the results on samples of de-mineralised water (with previous concentration) should be obtained by using the glycerol calibration line and not the methane calibration line. Nevertheless, frequent checks on the apparatus with both glycerol solution and methane are considered desirable, so that it can be seen that the calibration lines are parallel (or deviate by not more than 5 per cent.), thus ensuring that the oxidation and reduction stages continue to work satisfactorily. CONCLUSIONS It has been shown that the amount of involatile organic matter in de-mineralised water can be determined rapidly by concentration of 100-fold and injection of the concentrate into a combustion tube containing copper( 11) oxide, reduction to methane and measurement with a flame-ionisation detector. The range is from 0.2 p.p.m. of carbon upwards, with a reproducibility of about kO.1 p.p.m. ; the method should be suitable for controlling the involatile carbon content to below 1 p.p.m. The method described in this paper is the subject of U.K. Patent Application No. 48483/66. REFERENCES 1. “Recommended Methods for the Analysis of Trade Effluents,” Joint Committee of the Association of Chemical Manufacturers and the Society for Analytical Chemistry, W. Heffer and Sons Ltd., Cambridge, 1958. NOTE-Reference 2 is to Part I of this series. 2. Cropper, F. R., Heinekey, D. M., and Westwell, A., Analyst, 1967, 92, 434. Received November l l t h , 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200443

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

450 Analyst, July, 1967, Vol. 92, $9. 450-455 The Application of Gas Chromatography to the Examination of the Constituents of Cannabis Satiua L. BY L. T. HEAYSMAN, E. A. WALKER AND D. 'r. LEWIS (Ministvy of Technology, Labovatovy of the Government Chemist, Cornwall House, Stamford Street, London, S.E.l) Gas-chromatographic methods for the identification of Cannabis sativa L. (Indian hemp) have been examined. A column of 1 per cent. Carbowax 20M on silanised Chromosorb G has been found to give better separation of the cannabinols than columns previously used. The method may be used to distinguish cannabis in the presence of other vegetable materials that are known to give positive reactions to the usual colorimetric tests. Where doubt exists this may be resolved by gas chromatography of the trimethyl- silyl derivatives.The method is also applicable to the determination of the resinous extract of cannabis in mixtures of cannabis and tobacco. THE world-wide increase of illicit trading in Indian hemp or marihuana (Cannabis sativa L.) has increased the need for reliable methods of identification. Although microscopic examina- tion, coupled with the various colorimetric tests, such as the Beam1 test, and its modification,2 the Ghamrawy test,3 the official A.O.A.C. Duquenois* - Levine test and the Wasicky test,5 may be regarded as unambiguous when dealing with vegetable forms of the drug, this cannot be said to apply equally to the resinous extract that contains only small traces of recognisable plant fragments. In this instance reliance has been placed on the colorimetric tests.It has, however, been shown6 that an appreciable number of other vegetable extracts also respond to one or more of these tests, all of them contributing to a diminution in the certainty of identification. This element of doubt has led several authors to examine alternative methods, a review of which has been made by GrliL6 Among the possible alternatives gas - liquid chromatography appears to offer a means of positive identification that could be adapted to quantitative measurement. Previous applications of gas chromatography to the investigation of cannabis have involved the use of relatively non-polar, low-loaded columns of 1 to 3 per cent. methyl silicone gum7,8y9 and cyanosilicone gumlo for the separation of the three principal constituents, namely, cannabinol (CBN) , cannabidiol (CBD) and tetrahydrocannabinol (THC) , of which THC is the physiologically active member.6 THC was not completely separated from other cannabinols when using silicone gum as a stationary phase.It was therefore considered desirable to investigate the value of more polar stationary phases, at the relatively high temperature and low loading required, in order to obtain retention times of practical duration. One per cent. of Carbowax 20M sup- ported on silanised Chromosorb G was found to fulfil this r61e satisfactorily. A preliminary investigation had shown that the relative retention times of the con- stituents when chromatographed on S.E.30 agreed with those of Davis, Farmilo and Osadchuk.7 Identification of the peaks produced from a Carbowax column was achieved by using thin-layer chromatography in the manner described by Korte and Sieperll and Korte, Sieper and Tira.12 These workers have described a method for separation of the constituents of cannabis in which silica gel impregnated with dimethylformamide is used.HEAYSMAN, WALKER AND LEWIS 451 By using pure and synthetic cannabinols as standards, they found that distinctive colours were produced when their thin-layer plates were sprayed with o-dianisidine - tetrazolium chloride, and established the order of elution of the main constituents of cannabis extracts.We have extended this method to a preparative scale, and by extracting the individual zones (which were identified by their colours with o-dianisidine - tetrazolium chloride) we obtained sufficient of each constituent for use as standards. These were used to establish the identity of each peak when using a Carbowax 20M column, the relative retention times with respect to an internal standard and the linearity of detector response to each compound.Gas chromatography of the trimethylsilyl ethers of the cannabinols has also been investi- gated. Satisfactory separation of these derivatives was obtained at a temperature 50" C lower than that necessary for the separation of the original compounds, thereby increasing the operating life of the column. EXPERIMENTAL APPARATUS- A Pye Argon Chromatograph was used with a p-ionisation detector operated at 1250 volts and amplifier sensitivity of times 10. A glass column, 4 feet x 0.15 inch, was packed with 1 per cent.Carbowax 20M on DMCS-treated, 70 to 80-mesh, Chromosorb G obtained from Johns-Manville. The column was conditioned overnight at 250" C and its operating tem- perature was 230" C for the original extract and 180" C for the silanised extract. The flow-rate of the carrier gas was 70 ml per minute. Thin-layer separations were carried out on 20 x 40-cm glass plates coated with silica gel G (Merck) by means of a Shandon preparative-layer apparatus. Samples were extracted in chromatographic columns, 20 cm x 1 cm i.d., fitted with a glass sinter disc, porosity 2. EXTRACTION- About 1 g of ground hemp sample was generally sufficient to produce an extract of suitable concentration. A known weight of the ground sample was placed in an extraction tube that had previously been filled to a height of 4 cm with silica gel (Hopkin and Williams MFC grade).Benzene was poured into the tube and allowed to flow through the column until 25ml of extract had been collected. This was found to be sufficient to remove the cannabinols. Further washing with benzene produced no significant amounts. The silica gel retained some of the non-cannabinolic constituents that cause tailing of the solvent peak with resulting difficulties in quantitative measurements. The solvent was removed with gentle heat in a current of air, and the residue dissolved in a minimum quantity of a mixture of cyclohexane and ethanol (60 + 40) containing 5 per cent. of dibenzyl phthalate. (This mixture was used because the residue appeared to dissolve in it most readily.) A volume of 0.2 to 0.5 p1 of this solution was injected into the gas chromatograph. The dibenzyl phthalate served as an internal standard and as a reference material for the measurement of relative retention times (Fig.1 and Table I). TABLE I RELATIVE RETENTION TIMES Cannabinols (230" C) Cannabinol trimethylsilyl ethers (1 80" C) Cannabidiol . . . . . . . . 0.33 Cannabidiol . . . . . . . . 0-43 Tetrahydrocannabinol . . . . 0.37 Tetrahydrocannabinol . . . . 1-63 Cannabinol . . . . . . . . 0.71 Cannabinol . . .. .. . . 2.96 Dibenzyl phthalate . . . . . . 1.00 Anthracene , . . . . . . . 1.00 THIN-LAYER CHROMATOGRAPHY- Glass plates, 40 x 20 cm, were coated with silica gel G (Merck) to a thickness of 1 mm by using a slurry of a mixture of 50g of silica gel and 50ml of water.The plates were impregnated with dimethylformamide in the manner described by Korte and Sieper,ll and Korte, Sieper and Tira.12452 HEAYSMAN et d. : APPLICATION OF GAS CHROMATOGRAPHY TO [AadySt, VOl. 92 The residue from a benzene extraction of cannabis was dissolved in cyclohexane and applied to an impregnated plate in a streak 3 mm wide and 36 cm long, at a distance of 3 cm from the longer edge. The plate was developed with cyclohexane until the solvent front had reached the level previously reached by the dimethylformamide. Saturation of the mobile phase with dimethylformamide was not found to be necessary. Identification of the zones was made by spraying strips, 2 cm wide at each end of the plate, with o-dianisi- dine - tetrazolium chloride.Zone boundaries were then marked across the plate with the help of ultraviolet light, which revealed their precise contours as dark bands, and the zones were removed and extracted. The various THC isomers were included in one extract. Dibenzyl phthalate THC ;r. I I I I I I I Time, minutes 60 50 40 30 20 10 1 Fig. 1. Chromatogram of cannabis constituents on 1 per cent. Carbowax on silanised 70 t o 80-mcsh Chromosorb G ; temperature 230" C; flow-rate 70 ml per minute ; /3-ionisation detector CALIBRATION OF CHROMATOGRAPHS- The relation between peak area and concentration of the cannabinols was investigated by using the thin-layer extracts. These extracts were diluted with increasing amounts of solvent, the sample vial being weighed before and after the removal of injections.Linear relationship was found for each cannabinol over a range of peak areas corresponding to peak heights of 7 to 60 per cent. full-scale deflection of the recorder, which agrees with the findings of Davis, Farmilo and Osadchuk.7 Peak areas per unit weight were measured for each cannabinol and for dibenzyl phthalate, thus enabling the latter compound to be used as an internal standard (Table 11). TABLE I1 RELATIVE RESPONSE FACTORS FOR CANNABINOLS AGAINST DIBENZYL PHTHALATE Cannabidiol against dibenzyl phthalate . . .. . . 1.49 Tetrahydrocannabinol against dibenzyl phthalate . . 0.67 Cannabinol against dibenzyl phthalate . . .. . . 0.42 PREPARATION OF TRIMETHYLSILYL ETHERS- The prolonged use of a loaded Carbowax 20M column at a temperature of 230" C is liable to limit the efficient life of the column.There is, therefore, an advantage in beingJuly, 19671 THE EXAMINATION OF THE CONSTITUENTS OF Cannabis Sativa L. 453 able to work at a lower temperature, and this can be done if the cannabinol extract is silanised. The cannabinols were converted into trimethylsilyl ethers which could be separated at 180" C on the same column. This reaction was carried out in a pyridine solution of the extract by using the method of Sweeley, Bentley, Marketa and Wells.13 After removal of the solvent the benzene extract was dissolved in 1 ml of anhydrous pyridine containing a little anthracene in a small, stoppered glass tube; 0.1 ml of hexamethyldisilazane was then added, followed by 0.1 ml of trimethyl- chlorosilane.The tube was shaken for 30 seconds and then allowed to stand for 10 minutes. A sample of 0.5 to 1.0 p1 was removed from the mixture and injected into the gas chromato- graph. Gas chromatography at 230" C showed no peaks for unreacted cannabinols. From this evidence and from the comparable size of the peaks from the trimethylsilyl ethers, it was assumed that the reaction with the cannabinols was almost complete. It is eluted between CBD and THC, and is not affected by the silanising reagents, It may therefore be dissolved in the pyridine before the extract solution is made, 0.5 per cent. being a suitable concentration. A typical chromatograph of silyl derivatives is shown in Fig. 2 and relative retention times of the silyl ethers are included in Table I.Anthracene is a suitable reference material. 7 Time, minutes Fig. 2. Chromatogram of trimethyl- silyl derivatives of Cannabis sativa L. extract on 1 per cent. Carbowax on 70 to 80-mesh DMCS-treated Chromosorb G; temperature 180" C; flow-rate 70 ml per minute; p-ionisa- tion detector Plants reported to give positive reactions with the usual cannabis colorimetric reagents12 were also examined under conditions of extraction and gas chromatography identical with those used for the analysis of the cannabinols. The results obtained are shown in Fig. 3, in which retention times relative to dibenzyl phtlialate are indicated. Any confusion that might arise from the proximity of peaks from plant extracts to those of the cannabinols, e.g., Datura Straunonium L. to THC, and Rosmayinus oficinalis L.to CBN, could be resolved by gas - liquid chromatography of the trimethylsilyl derivatives, when the differences in retention times make distinction certain. Although tobacco has not been included in Fig. 3, mixtures of cannabis and tobacco were examined and it was found that, by using the extraction method described, as little as 1 per cent. of cannabis could be satisfactorily detected.454 HEAYSMAN et al. : APPLICATION OF GAS CHROMATOGRAPHY TO [Analyst, Vol. 92 nus niger 1. inus officinalis L. IS vulgaris L. dula officinalis Chaix ramus niger 1. Bellodonna L. i officinalis 1. arinus officinalis L. ra Stramonium L. nus vulgaris 1. scyamus niger L. iandula officinolis Chaix osmarinus officinalis L. yoscyamus niger L.ialvia officinalis L. h t u r a Strarnonium L. Thymus vulgaris L. Hyoscyamus niger L. I Lavandula officinalis BD Papaver somniferum L. Atropa Belladonna L. 1i::ura Stramonium L. halx Salvia officinolis L. I. I 0.2 0.3 0.4 cus elasticus Roxb. a 0.5 0.6 0.7 iN Rosmarinus officinalis L. Ficus elasticus Roxb. :also 1-18 1.73 2.60) Rosmarinus officinalis L. L Dibenzyl hthalate Papaver 0.8 0.9 I .o 9mniferum L. Relative retention times Fig. 3. Relative retention times of cannabis constituents compared with those of other vegetable extracts DISCUSSION Our results demonstrate the improvement obtained when a polar column of 1 per cent. Carbowax 20M is used for the separation of CBN, CBD and THC, compared with the results obtained on the less polar silicone elastomers when gas chromatography is used for the identification of Cannabis sativa L.The question of cannabidiolic acid in the samples has not been considered. It has already been shown7 that this is converted to CBD below 200" C. We have confirmed the presence of the acid, both by methylations and by the coloured zone at the origin in thin-layer plates after spraying with o-dianisidine - tetrazolium chloride.ll 9 l 2 Altogether, twenty different samples of cannabis were examined, most of which had been seized by H.M. Customs as illegal imports. Their geographical origin was, therefore, not known with certainty. If, however, the ports of origin of the ships from which the seizures were made could be taken as a reasonable indication of the area of origin, the bulk of the samples came from Asia and a few from the West Indies.One sample, however, was known with certainty to have been grown in England. No outstanding difference in the proportions of CBD, CBN and THC were found between the samples from the two principal source areas. On the other hand, the English sample was found to contain predominantly CBD, with only relatively small amounts of CBN and THC. This was in general agreement with the findings of Davis, Farmilo and Osadchuk.' Most of this work was carried out with a p-ionisation detector, but both flame ionisation and thermal conductivity have also been used. It was found essential to use on-column injection, or, where an instrument was equipped with a heated injection port, to use a glass liner. Failure to do this resulted in complete disappearance of the CBN, CBD and THC, presumably owing to decomposition on the walls of the injection port. Stainless-steel columns appeared to be satisfactory, but no quantitative evaluation was carried out with them. The authors thank the Ministry of Technology for permission to publish this paper.July, 19671 THE EXAMINATION OF THE CONSTITUENTS OF Cannabis Sativa L. REFERENCES 455 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Beam, W., Wellcome Tropical Research Laboratory, 4th Report, Khartoum B, 25, 191 1. Nickolls, L. C., Analyst, 1936, 61, 605. Ghamrawy, M. A., J . Egypt. Med. Ass., 1937, 20, 193. Duquenois, P., and Moustapha, H. N., Ibid., 1938, 21, 224. Simmon, J . S., and Gentzow, C. J., Editors, “Medical & Public Health Lab. Methods,” H. Kimpton, Grlic, L., Bull. Narcot., 1964, 16, No. 4, 29. Davis, T. W. M., Farmilo, C. G., and Osadchuk, M., Analyt. Chem., 1963, 35, 751. Kingston, C. R., and Kirk, P. L., Ibid., 1961, 33, 1794. Lerner, M., Mills, A. L., and Mount, S. F., J . Forens. Sci., 1963, 8, 126. Lerner, M., Science, N.Y., 1963, 140, 175. Korte, F., and Sieper, H., J . Chromat., 1964, 14, 178. Korte, F., Sieper, H., and Tira, S., Bull. Narcot., 1965, 17, No. 1, 35. Sweeley, C. C., Bentley, R., Marketa, M., and Wells, W. W., J . Amer. Chem. SOC., 1965, 85, 2497. Received J u l y l l t h , 1966 London, 1965.

ISSN:0003-2654    

DOI:10.1039/AN9679200450

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

456 AnaZyst, July, 1967, Vol. 92,pp. 456-462 The Determination of the Oxidisable Nitrogen Oxides Present in Cigarette Smoke BY G. A. L. SMITH," P. J. SULLIVAN AND W. J. IRVINE The use of brucine for the colorimetric determination of the oxidisable oxides of nitrogen in cigarette smoke is described. The oxides of nitrogen in the cigarette smoke are converted to nitrate with dilute aqueous hydrogen peroxide, and the nitrate is determined by using a modification of the method of Robinson, Allen and Gacoka. Results are presented for the determination of the oxidisable nitrogen oxides present in the smoke of various types of cigarettes. METHODS commonly used in the determination of nitrate include (a) determination of ammonia formed after reduction with metallic reducing agents1 9 2 ; (b) determination of nitrite formed after reduction with z i n ~ , ~ , ~ , ~ hydrazine6 9 7 or cadmium amalgams; (c) spectro- photometric determination of the reaction products of the nitrate ions and organic aromatic compounds, such as diphenylben~idine,~ diphenylamine,1° resorcino1,ll phenol-2,4-disulphonic acid,l2$l3 9 1 4 2,4-xylen01,~~9~~ 2,6-xyleno1,17 ,18 3,4-xyleno1,19 b r ~ c i n e , ~ ~ , ~ ~ chromotropic acid,22 l-amin~pyrene,~~ ~trychnidine~~ and phenarsazinic acid25 ; and (d) a combined enzymatic and colorimetric determination of nitrate in plants with Pseudomo%as oleovoram.26 Phillipe and Hackney27 showed qualitatively the presence of nitrous oxide in cigarette smoke. Tada28 stated that the major oxide of nitrogen present was nitric oxide, and Norman and Keith29 confirmed this.However, Bokheven and Niessen30 suggested that approximately equal amounts of nitric oxide and nitrogen dioxide were present in cigarette smoke. The amount of nitrogen oxides in the smoke of commercial cigarettes was found by Haagen-Smit31 to be 145 to 665 p.p.m. and by Eokheven and Niessen30 to be 170 to 210 p.p.m. However, W e ~ t c o t t ~ ~ reported much lower levels of nitrogen oxides in the smoke of British cigarettes in comparison with those manufactured in the United States, and with this we are in full agreement. As there was some doubt as to which was the most plentiful oxide of nitrogen present, it was decided to pass the cigarette smoke through dilute aqueous hydrogen peroxide, which oxidised the oxides of nitrogen present to nitrate, and to express the final result as micrograms of nitrogen dioxide per cigarette.Recent workers20J3 have improved the brucine method and preliminary tests showed it to be suitable for determinations involving tobacco. ABSORPTIOK MAXIMUM OF THE PRODUCT WHEN BRUCINE REACTS WITH NITRATE IONS IN A 2-ml aliquot of a solution containing S pg per ml of nitrate ions was trans- ferred by pipette into a dry 10-ml calibrated flask. To this was added from a pipette 0.5 ml of 5 per cent. w/w brucine in glacial acetic acid. From a burette 4 ml of concentrated sulphuric acid were added, 0.2ml at a time, the flask being shaken after each addition of the acid for about half a minute. Finally the flask was placed in an ice-bath at 0" C for 15 minutes. I t was then removed and allowed to attain room temperature by standing it on the bench for 30 minutes. The absorption spectrum was then plotted automatically with a Perkin-Elmer 137 ultraviolet spectrophotometer and a broad maximum was observed at 410 mp.Varying the concentration of the brucine reagent in glacial acetic acid from 2.5 to 10 per cent. produced only a small increase in the optical density, and no wavelength shift of the absorption maximum of the final solution. EFFECTS OF TEMPERATURE ON THE COLOUK FORMATION- Two temperature effects were examined; the first, the temperature at which the addition of the sulphuric acid was made; the second, the rate of cooling after the addition of the acid had been completed. (Carreras Risearch Division, Nevendon Road, Basildon, Essex) The coefficient of variation of the method is 7-7 per cent.EXPERIMENTAL SULPHURIC ACID- * Present addrcss : Hopkins & Williams Ltd., Freshwater Road, Chadwvell Heath, Essex.SMITH, SULLIVAN AND IRVINE 457 For the examination of the first effect, solutions were prepared as described in the first paragraph of the "Experimental," with the exception of the addition of sulphuric acid, which was added 0.2 ml at a time to individual flasks immersed in water-baths at 0", 25", 50" and 100" C. On completing the addition, each flask was removed from its water-bath, and allowed to attain room temperature by standing it on tlie bench for 30 minutes. Measure- ment of the optical densities at 410 mp showed that the lower the reaction temperature on adding sulphuric acid, the higher was the final colour intensity obtained.The exception was the series of flasks immersed in the water-bath at 0" C, when, instead of the characteristic yellow colour, a pink colour was formed. Examination of the second effect was made by repeat- ing the above, but before allowing them to attain room temperature by standing them on the bench for 30 minutes, the flasks were placed in an ice-bath for 15 minutes. It was shown that with a faster rate of cooling, the final solution again had a higher colour intensity. From this it was decided that during the addition of sulphuric acid, the flasks should be kept in a water-bath at 25" C, and that immediately after the addition was completed the flask should be immersed in an ice-bath for 15 minutes, and then brought to room temperature by standing it on the bench for 30 minutes.EFFECT OF TIME AND SODIUM SULPHATE CONCENTRATION ON COLOUR FORMATIOX- The optical density of the solution was determined immediately after the period allowed for attaining room temperature and then measured repeatedly for up to 3 hours. The intensity of the colour dropped slightly during the first half-hour, but then remained constant. This decrease in optical density for the first half-hour made it necessary to take the quantitative reading 1 hour after removal from the ice-bath. With solutions up to 1 M with respect to sodium sulphate a negligible reduction in the colour intensity was produced; this occurred when the optical density was measured at temperatures between 20" and 30" C.PREPARATION OF CALIBRATION GRAPH- A series of test solutions of differing nitrate concentrations (in the range 2 to 1Opg per ml) was used to show that the colour formed obeys Beer's law. The optical densities of the test solutions were measured in l-cm cells at 410mp against a reagent blank. (The observed molecular extinction coefficient of the coloured compound was 1.34 x lo4). In this way amounts down to 0.5pg of nitrate ion per ml could be determined. PREPARATION OF THE ION-EXCHANGE COLUMN- Pour a mixture of 2 N hydrochloric acid and De-Acidite FF (SRA 61) resin beads into a glass column (part of a broken burette suffices) so that beads occupy a volume equal to 1 ml. Wash the beads with de-ionised water until a negative reaction is obtained with silver nitrate solution.If a greater volume of beads is used, the recovery of nitrate ions under the prescribed conditions is poor. Further, the recovery of nitrate ions should be determined for each batch of resin used. DETERMINATIOX OF COLUMN EFFICIENCY- Pass a suitable volume (50ml) of solution containing 200 to 500pg of nitrate ions through the column allowing the eluate to come through no faster than 1 drop every 5 seconds. Wash the resin with 10ml of de-ionised water. Recover the nitrate ions from the resin by passing 40 ml of M sodium sulpliate through the column at the rate prescribed in the previous paragraph. Collect the eluate in a 50-ml calibrated flask. Wash the column again with 10 ml of de-ionised water and collect the washings in the flask. Determine the nitrate content of the eluate spectrophotometrically with the brucine reagent.The results for the recovery of tlie nitrate ion from the ion-exchange column are as follows- Nitrate Mean nitrate Range of nitrate Coefficient added, Number of recovered, recovered, Standard of variation, PFLg determinations per cent. per cent. deviation per cent. 500 20 96.4 90.0-100.4 2.6 2.7 200 5 95.2 93.3- 97.3 - - Discard the eluate. Make up to the mark with de-ionised water.458 SMITH, SULLIVAN AND IRVINE : DETERMINATION OF THE DETERMINATION OF THE EFFICIENCY OF THE TRAPPING SYSTEM USED- during use. [Analyst, Vol. 92 Both traps are continuously agitated The trapping system is as shown in Fig. 1. Nitrogen I Excess pressure vent To smoking A = 25-mI Generator containing acidified sodium nitrite solu- tion.TI and T,= Trap I and Trap 2, each containing 20ml of 0-3 per cent. aqueous h yd ro- gen peroxide. Fig. 1. Determination of trapping efficiency The nitrogen oxides were generated by adding 5ml of 2 N hydrochloric acid to the sodium nitrite solution in the generator. The system was connected to a C.S.M.10 smoking machine setlso that a 35-ml puff, of two seconds’ duration, was taken once a minute on each channel. A cylinder of “White Spot” nitrogen was used to flush the oxides from the generator into the traps. A total of 350 ml was flushed through the system. The aqueous hydrogen peroxide in the traps containing the nitrate ion was passed through the column of De-Acidite FF (SRA 61) to remove all the nitrate ions formed, and these were then eluted with 40 ml of M sodium sulphate and determined spectrophotometrically with brucine.The residual nitrite left in the generator was oxidised to nitrate and determined as previously described. The trapping efficiency of the system shown in Fig. 1 is as follows- Nitrate added to Residual nitrate in Nitrate in Trapping Number of generator generator, traps 1 and 2, efficiency, determinations (as nitrate, pg) Pg P.g per cent. 16 450 45 to 258 175 to 378 88.5 to 101.8 Mean . . .. .. . . 95.4 Standard deviation . . . . 3-84 Coefficient of variation . . . . 4.03 The trapping efficiency is given by- 100 Ion-exchange column efficiency, per cent. ( Nitrate in traps 1 and 2, pg (Nitrate added - nitrate I remaining in generator) pg METHOD REAGENTS- All reagents should be of analytical-reagent grade unless otherwise specified.Sodium nitrate. Sodium sulphate, M, aqueous. Sulphuric acid, concentrated. Acetic acid, glacial. Hydrochloric acid, 2 N. Brucine-Obtainable from Hopkin & Williams Ltd., should be re-crystallised from dilute Brucine reagent, 5 per cent. w / v in glacial acetic acid-This solution should be colourless De-Acidite FF (SRA 61) resin (Permutit Ltd.). Hydrogen peroxide, 0.3 per cent. w l v , aqueous. aqueous acetic acid by precipitation with dilute alkali. or nearly so, and should be kept in the dark for not longer than 1 week.July, 19671 APPARATUS- OXIDISABLE NITROGEN OXIDES PRESENT IN CIGARETTE SMOKE Smoking unit, C.S.M. 10-Obtainable from Cigarette Components Ltd. Cambridge Jilters and holders-Fitted to the smoking unit.Mechanical shaker. S~ectrophotometer-Unicam SP500 or an E.E.L. Spectra. c- machine 459 Cambridge filter holder and cigarette PREPARATION OF CALIBRATION GRAPH- Transfer by pipette 2-ml aliquots of the standard nitrate solutions (containing 0 to 20 pg of nitrate ion) into clean, dry 10-ml calibrated flasks. Into each flask transfer by pipette 0.5 ml of brucine reagent. Immerse each flask in turn in a water-bath at 25" C, then slowly add 4 ml of concentrated sulphuric acid from a burette, 0.2 ml at a time, agitating the flask for half a minute after each addition. On completing the addition of sulphuric acid, place each flask in an ice-bath for 15 minutes. Remove the flasks and then allow to stand for 1 hour. Prepare the reagent blank in the same way, omitting the nitrate solution and substituting for it 2 ml of de-ionised water.Measure the optical density of each solution in a 1-cm cell against the reagent blank solution at 410mp. A straight-line relationship should be obtained. The calibration graph should be checked at least once a week. The slope of the straight line obtained varies with each batch of brucine used. Plot a graph of optical density versus micrograms of nitrate ions. CONDITIONING AND SELECTION OF CIGARETTES- humidity of 60 per cent. and at a temperature of 20" C. Condition the cigarettes to be used in the determination for 48 hours at a relative SMOKING PROCEDURE- 100-ml round-bottomed flasks. each arm. Transfer by pipette 40ml of 0-3 per cent. w/v hydrogen peroxide into each of two Place glass helices in the U-tube to a depth of 2 inches in C.S.Set up the trapping system as shown in Fig. 2 omitting the cigarette. Attach it to the C.S.M.10 smoking machine. Set up the whole unit so that a 2-second puff of 35 ml is taken every minute. The volume of the puff must be calibrated by using a bubble-meter before each smoking operation. Mark the weight-selected cigarettes at a butt length of 23 cm and smoke 5 cigarettes in the usual manner. After smoking, take two clearing puffs and dismantle the trapping system.460 [Analyst, Vol. 92 DETERMINATION OF THE OXIDES OF NITROGEN PRESENT IN CIGARETTE SMOKE- Pass the contents of the round-bottomed flasks through the ion-exchange column, carefully washing out each flask with de-ionised water, and passing the washings through the column.Wash the glass helices in the U-tube with de-ionised water and pass the washings through the column. Allow the eluate to flow from the column no faster than 1 drop every 5 seconds. Finally wash the resin with 10 ml of de-ionised water. Discard the eluate. Recover the nitrate ions from the resin by passing 40 ml of M sodium sulphate slowly through the column. Wash the column with 10 ml of de-ionised water. Collect the eluate and washings in a 50-ml calibrated flask and make up to the mark with de-ionised water. Transfer by pipette a 2-ml aliquot of the extract into a clean, dry 10-ml calibrated flask, and follow the procedure in the paragraph describing the preparation of the calibration graph. From the optical density of the sample solution, obtain from the calibration graph its equivalent in micrograms of nitrate ions.Let this be B pg, and C be the number of cigarettes SMITH, SULLIVAN AND IRVINE : DETERMINATION OF THE smoked. Therefore, oxides of nitrogen B C where 0.74 is the conversion factor correction factor for the recovery this instance C = 5. - - as micrograms of nitrogen dioxide per cigarette x 25 x 1-10 x 0.74, for nitrate ion to nitrogen dioxide, and 1.10 the combined and trapping efficiencies of the oxides of nitrogen. In The nitrogen dioxide content per cigarette = 4.06B pg. DETERMINATION OF THE WATER-SOLUBLE NITRATE CONTENT OF TOBACCO- The water-soluble nitrate was removed from the tobacco by an aqueous extraction, and removed from the extract by passing through a column of ion-exchange resin, as previously described.The nitrate ion was recovered and determined spectrophotometrically with brucine reagent. RESULTS AND DISCUSSION The statistical analysis of a series of replicate determinations of the percentage of water-soluble nitrate present in tobacco was performed on a single sample of ground leaf and the results are as follows- Mean Range, Coefficient of Number of water-soluble nitrate. water-soluble nitrate. Standard variation. determinations per cent. w/w 20 0.089 THE WATER-SOLUBLE NITRATE Type of Cigar leaf, Cameroon Cigar leaf, Brazil . . Cigar leaf, Sumatra Cigar leaf. Holland . . per cent. w/w deviation per cent. 0.079 to 0.097 0.0064 7-2 TABLE I PRESENT IN VARIOUS TYPES OF TOBACCO LEAF tobacco . . . .. . .. . . . . . . . . Ciiar leaf; Jamaica Virginia type leaf 1 . . 2 . . 3 .. 2 .. 3 . . American cigarette blend 1 . . 2 . . 3 . . 4 ,. 5 . . 6 . . 7 . . Turkish cigarette blend 1 . . 2 . . French cigarette blend 1 . . British cigarette blend 1 . . . . .. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Water-soluble nitrate, per cent. w/w 0-24 0.11 0.12 0.30 0.14 0.13 0.024 0.008 0.073 0.078 0.089 0.51 0-53 0.38 0.38 0.54 0.14 0.95 0.30 0.19 1.70July, 19671 OXIDISABLE NITROGEN OXIDES PRESENT IN CIGARETTE SMOKE 461 A series of replicate determinations of the nitrogen dioxide concentration in cigarette smoke was performed on a single brand of plain 70-mm virginia-blend cigarettes, and the results are shown below- Number of Mean, Range, Standard deviation, Coefficient of variation, determinations p g per cigarette p g per cigarette pg per cigarette per cent.20 63 55 to 70 4.8 7.7 To the same brand of plain cigarettes were attached three commercial filter tips. The results given below show the reduction in oxides of nitrogen content of the cigarette smoke achieved by attaching these filters. The figures given are the means of four determinations. Filter material Yield of nitrogen dioxide, Filtration eficienc y, (15-mm length) pg per cigarette per cent. Paper . . . . .. . . . . . . 58 8 Cellulose acetate . . . . . . . . 49 22 Activated charcoal on cellulose acetate . . 35 44 From this it can be seen that the oxides of nitrogen can be removed by filtration.A comparison of the oxides of nitrogen content of smoke from various cigarettes was made. In addition, a correlation between the water-soluble nitrate content of the cigarette blends and the oxides of nitrogen content of the cigarette smoke was obtained. COMPARISON OF THE Cigarettes American type 1 . . . . 2 .. . . 3 . . . . 4 .. . . British 1 . . . . . . 2 . . . . . . 3 . . . . . . Turkish 1 . . . . . . 2 . . .. . . French 1 . . .. . . TABLE I1 OXIDES OF NITROGEN CONTENT OF VARIOUS CIGARETTES Length, mm 60 + 25 filter 62 + 17 filter 69 69 70 70 70 70 69 69 Water-soluble nitrate, mg per cigarette 7.5 1-4 8.9 2.7 0.9 0.8 0.8 3.1 2-5 19.1 SMOKE FROM Oxides of nitrogen, as nitrogen dioxide, pg per cigarette 350 110 400 110 65 45 50 140 80 580 The results in Table I1 confirm the findings of W e ~ t c o t t ~ ~ that smoke produced by American cigarettes contains appreciably more oxides of nitrogen than the smoke from British cigarettes.DISCUSSION The water-soluble nitrate content of a cigarette and the oxides of nitrogen concentration in the smoke produced by it are generally related proportionally. Oxides of nitrogen can be removed selectively from the smoke with an activated charcoal filter (and doubtless larger proportions could be removed if charcoal of a higher activity were used). From Table I, it can be seen that, in general, the tobacco blends used in American cigarettes contain appreciably more water-soluble nitrate than their British counterparts, this probably arising from the much larger proportion of burley tobacco34 that is used in American cigarettes. We thank the directors of Carreras Limited for permission to publish this work, and Mr.J. Ostrowski for diagrams of the apparatus in Figs, 1 and 2. REFERENCES 1. 2. Macdonald, A. M. G., Ind. Chemist, 1955, 31, 515. 3. 4. 5. Chow, T. I., and Tohnstone, M. S., Analytica Chim. Acta, 1962, 27, 441. Bremner, J. M., Analyst, 1955, 80, 626. Foyn, E., Rep. Norw. Fishery Mar. Invest., 1951, 9, No. 14. Bray, R. H., Soil Sci., 1945, 60, 219.462 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. SMITH, SULLIVAN AND IRVINE Mullin, J. B., and Riley, J. P., Ibid., 1955, 12, 464. Strickland, J. B. H., and Porsens, T. R., Bull. Fish.Bes. Bd Can., 1960, No. 125. Morris, A. W., and Riley, J. P., Analytica Chim. Acta, 1963, 29, 272. Atkins, W. R. G., J . Mar. Biol. Ass. U.K., 1932, 18, 167. Trofimov, A. V., Zh. Prikl. Khim. Leningr., 1935, 9, 756. Costa, R. L., Boln Inst. ESP. Oceanogr., 1951, No. 43. Hora, F. B., and Webber, P. J., Analyst, 1960, 85, 567. Chamot, E. M., Pratt, D. S., and Redfield, H. W., J . Amer. Chem. SOC., 1911, 33, 366. Taras, M. J., Analyt. Chem., 1950, 22, 1020. Ruckett, J., Duffield, W. D., and Milton, R. F., Analyst, 1955, 80, 141. Swain, J. S., Chem. & Ind., 1957, 479. Hartley, A. M., and Asai, R. I., Analyt. Chem., 1963, 35, 1207. Montgomery, H. A. C., and Dymock, J. F., Analyst, 1962, 87, 374. Holler, A. C., and Huch, R. V., Analyt. Chem., 1949, 21, 1385. Robinson, J. B. D., Allen, M. de V., and Gacoka, P., Analyst, 1959, 84, 635. Hosne, R. W., and Denmead, A. T., J . Aust. Inst. Agric. Sci., 1955, 21, 34. West, P. W., and Lyles, G. L., Analytica Chim. Acta, 1960, 23, 227. Sawicki, E., Johnson, H., and Stanley, T. W., Analyt. Chem., 1963, 35, 1934. Westland, A. D., and Langford, R. R., Ibid., 1956, 28, 1996. Pietsch, R., Mikrochim. Acta, 1956, 1672. Hill, R. M., Pivnick, H., Englehand, W. E., and Bogard, M., J . Agric. Fd Chem., 1959, 7, 261. Phillipe, R. J., and Hackney, E. J., Tob. Sci., 1959, 3, 139. Tada, O., Rep. Inst. Sci. Lab., 1962, 60, 7. Norman, V., and Keith, C. H., Nature, 1965, 205, 915. Bokhoven, C., and Niessen, H. J., Ibid., 1961, 192, 458. Haagen-Smit, A. J., Brunelle, M. F., and Hara, J., Amer. Med. Ass., Arch. of Ind. Health, 1959, Westcott, D. F., Paper presented a t the 18th Tobacco Chemists’ Research Conference, Rayleigh, Greenberg, A. E., Rossun, J. R., Moskowitz, N., and Villoruz, R. A., J . Amer. V a t . WAS Ass., Broaddus, G. M., York, J. E., jun., and Moseley, J. M., Tob. Sci., 1965, 9, 149. 20, 399. North Carolina, 1964. 1958, 50, 821. Received March 25th, 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200456

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

Analyst, July, 1967, Vol. 92,pp. 463465 463 A Spectrophotometric Method for the Micro Determination of Glycolaldehyde BY R. A. BASSON AND T. A. DU PLESSIS (Chemistvy Division, Atomic Eneygy Boavd, Private Bag 256, Pvetovia, Republic of South A fvica) A method is given for the determination of glycolaldehyde by oxidation with perchloric acid and reaction with 2,4-dinitrophenylhydrazine. The resulting compound is extracted with benzene, made alkaline with sodium ethoxide and determined spectrophotometrically at 560 mp. The compound is stable for a t least 1 hour and has a large extinction coefficient (4-17 x lo4 litres per mole per cm). SOME years ago a brief description1 of a method for the micro determination of glycolaldehyde based on the 2,4-dinitrophenylhydrazone method of Johnson and Scholes2 was given.We have investigated this method in some detail and, in view of its high sensitivity and relative simplicity, consider that it might be of value to publish our findings. The determination involves the oxidation of glycolaldehyde, in dilute aqueous solutions, to glyoxal by perchloric acid and the formation of the glyoxal bis-2,4-dinitrophenylhydrazone. The hydrazone is subsequently extracted with benzene and made alkaline with a solution of sodium ethoxide in ethanol. The optical density of the resulting blue colour is then deter- mined spectrophotometrically at 560 mp. EXPERIMENTAL REAGENTS- 45 per cent. perchloric acid. on prolonged standing. 2,4-Dinitrophenylhyd~a~ine-Prepare this by dissolving 0.5 g of the solid in 1 litre of The solution is stable for up to several weeks but darkens Sodium ethoxide solution-Prepare by dissolving 10 g of sodium in 1 litre of ethanol.Ethanol and benzene were of analytical-reagent grade and used without further purification. PROCEDURE- Add 5 ml of the dinitrophenylhydrazine - perchloric acid reagent to a 1-ml aliquot of the aqueous glycolaldehyde solution containing between 0.5 and 5 pg of the aldehyde. Dilute the mixture with 20 ml of water and heat the solution in a stoppered flask on a boiling water bath for 1 hour. Then remove the flask from the bath, cool it and transfer the contents to a separating funnel. Extract the hydrazone with 20 ml of benzene, followed by a second extraction with 5 ml of benzene. Combine the extracts, treat with 5 ml of sodium ethoxide solution and make up the resulting solution to 50 ml with ethanol.The optical density of this solution is determined after 5 minutes on a Zeiss RPQ 20 spectrophotometer, or similar instrument, at 560mp and the concentration of glycolaldehyde is determined from a calibration curve. EFFECT OF HEATING TIME- The oxidation of glycolaldehyde by perchloric acid in dilute solution is slow at room temperature, and the optical density of the final solution of samples that were not heated took several days to reach a maximum value. The optical density of samples heated at 95" C increased rapidly, however, with heating time, and reached a maximum value within 1 hour, as will be seen in Fig. 1. The completion of oxidation was checked by comparing identical concentrations of glyoxal and glycolaldehyde which gave the same optical density for heating times of over half an hour.464 .- % 0.4.- 0 0.2 ' - [Analyst, Vol.92 I Pig. 1. Effect of heating time on the oxidation of glycolaldehyde by perchloric acid EFFECT OF SODIUM ETHOXIDE- The concentration of sodium ethoxide in the final solution does not appear to be an important parameter as solutions containing between 5 and 25 g of sodium per litre of ethanol gave identical results. It was found advisable to prepare fresh ethoxide solutions every few days. COLOUR DEVELOPMENT AND STABILITY- The blue colour develops immediately on addition of the ethanolic ethoxide solution, and initially appears to fade slightly. After 5 minutes, however, the optical density becomes constant and is stable for at least 60 minutes.700 500 400 Wavelength, mu Fig. 2. Absorption spectrum for glyoxal bis-2,4-dinitrophenyl- hydrazone SPECTRUM, CALIBRATION AND ABSORPTIVITY- The spectra of the final solutions obtained from glycolaldehyde and from glyoxal samples were identical giving a well defined peak in the visible with a maximum at 560 mp, as shown in Fig. 2. Similarly, calibration curves constructed for known concentrations of the twoJuly, 19671 FOR THE MICRO DETERMINATION OF GLYCOLALDEHYDE 465 aldehydes were identical. These curves obey the Beer - Lambert law over the concentration range studied, and from them the molar absorptivity was determined as 4.17 x 104 litres per mole per cm. The temperature dependency of the molar absorptivity is small and no difference was found in the optical densities of samples between 20" and 30" C.The precision of the method is satisfactory and the optical density of identical samples was reproducible to within 4 per cent. D I s c u s s I o N The description of the method by Clay, Johnson and Weissl is somewhat misleading, as it indicates that the essential difference between this and the conventional 2,4-dinitro- phenylhydrazone method for mono-aldehydes lies in the extraction of the hydrazone with benzene instead of with carbon tetrachloride. If the determination is carried out in this way, the characteristic blue colour of the di-aldehyde hydrazone is not obtained when the solution is made alkaline but, instead, a red - brown solution with an absorption maximum at 420 mp, identical to that obtained with mono-aldehydes, is found.Under these conditions the presence of mono-aldehydes in the sample leads to serious interference. It is thus essential to allow sufficient time for the oxidation of the hydroxyl group and, as this is slow at room temperature, to accelerate the reaction by heating. If this is done, mono-aldehydes do not interfere, as was shown by the analysis of 1 0 - 4 ~ solutions of formaldehyde, acetaldehyde and propionaldehyde, which had optical densities at 560 mp that were not significantly higher than the blank. Glyoxal, and probably other di-aldehydes, will, however, interfere with the method. As previously mentioned, the hydrazone of glyoxal is much less soluble in carbon tetra- chloride than the hydrazones of mono-aldehydes.We have established that very little of the glyoxal hydrazone is extracted by carbon tetrachloride while the extraction with benzene is quantitative. Extraction with benzene does have the disadvantage that blank values are much higher than in carbon tetrachloride extractions, probably due to extraction of unconverted 2,4-dinitrophenylhydra~one.~ In view of the high value of the molar absorp- tivity, however, this is not a serious drawback. The use of sodium ethoxide solutions instead of ethanolic sodium hydroxide is preferred on the grounds that finely dispersed precipitates are occasionally obtained with the latter. Sodium ethoxide has also been used in the development of chromatograms of di-aldehyde hydra zone^,^ which suggested its use in the present method. REFERENCES 1. 2. 3. 4. Clay, P. G., Johnson, G. R. A., and Weiss, J., J . Chem. SOC., 1958, 2175. Johnson, G. R. A., and Scholes, G., Analyst, 1954, 79, 217. Basson, R. A., Apzalyt. Chem., 1966, 38, 637. Schmitt, W. J., Moricani, E. J., and O'Connor, W. F., Ibid., 1956, 28, 249. Received Novevnbev 22nd, 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200463

出版商:RSC    

年代:1967

数据来源: RSC

摘要:

466 Analyst, July, 1967, Vol. 92, $9. 466-467 A Coagulation Method for Determining Silica (Without Dehydration) in Silicate Materials BY H. BENNETT AND R. A. REED ( T h e British Ceramic Research Association, Queens Road, Penkhull, Stoke-on-Trent) Silica may be coagulated from acidic solutions by any of a series of “Polyox” resins. This technique enables silica to be determined in silicate materials without dehydration. The procedure fits into the method of B.S. 1902 : Part 2A : 1964 for the analysis of aluminosilicates and aluminous materials, without interference, giving results of equal accuracy. THE present British Standard Method for the analysis of high silica and aluminosilicate materials (British Standard 1902 : Part 2A : 1964) relies on the determination of silica by a single dehydration, thus saving a considerable amount of time as compared with more traditional methods.The sample is decomposed by fusion in alkali carbonate, the melt dissolved in hydrochloric acid and the solution evaporated to dryness. The silica precipitate is filtered off, ignited, weighed, treated with hydrofluoric acid and the residue again ignited and weighed. The difference in weights yields the amount of “gravimetric” silica. The residue is then fused in potassium pyrosulphate and the melt dissolved in the solution from the silica filtration. The combined solution is made up to volume and aliquots are used for the colorimetric determinations of residual silica with ammonium molybdate, iron(II1) oxide with 1 ,lo-phenanthroline, and titania with hydrogen peroxide.A further aliquot is used for the determination of alumina by removing iron and titanium by means of a cupferron - chloroform extraction, adding excess of EDTA and titrating the excess with a zinc solution, with dithizone as indicator. Alkalis are determined on a separate portion of the sample after decomposing it with hydrofluoric acid, and further portions of this solution are used for the determination of lime and magnesia by EDTA titration. EXPERIMENTAL To increase the speed of analysis of these types of sample still further an investigation was made of the gelatin precipitation method for silica, in the hope of avoiding even the single dehydration. It was found that, although gelatin coagulated silica formed after fusion and dissolution of the melt in hydrochloric acid without dehydration, colorimetric deter- mination of the 1 to 2 per cent.of silica remaining in solution was not possible because of the formation of a precipitate by the reaction of gelatin and molybdate. Several synthetic industrial coagulating agents were tried as replacements for gelatin, and, of these, a range of ethylene oxide polymers distributed under the name of “Polyox” (by Union Carbide Ltd.) proved effective. Seven of these materials were tested (Polyox coagulant and Polyox resins: WSR 35, WSR-N-80, WSR 205, 301, WSR-N-750 and WSR-N-3000*). All were comparably effective. Although experimentation is incomplete, it is clear that silica can be almost quantitatively precipitated without previous dehydration, and that the silica remaining in solution can be determined colorimetrically by reaction with molybdate.Subsequent colorimetric deter- minations of iron( 111) oxide with 1 ,lo-phenanthroline and titania with hydrogen peroxide, and the complexometric determination of alumina by the Wanninen and Ringboml technique are free from interference. In the procedure adopted a 1-g sample is mixed with a suitable flux in a 7-cm platinum dish, the dish covered with a lid and the mixture fused gently at first over a burner, and finally at 1200” C in a furnace. For clay type and high silica materials a flux of 3 g of fusion mixture (an equimolecular mixture of potassium and sodium carbonates) and 0.4 g of boric acid was found suitable, with a final fusion at 1200” C for 10 minutes. For materials con- taining 45 to 85 per cent.of alumina, 2 g of fusion mixture and 0.4 g of boric acid are used as a flux, with a final fusion period of 15 to 30 minutes, depending on the alumina content of the sample. After cooling the dish, 15 ml of hydrochloric acid (sp.gr. 1-18> and 10 ml of * Ridsdale & Co., Newham Hall, Middlesbrough, are hoping to make I’olyox WSK 35 available in laboratory amounts.BENNETT AND REED 467 water are added, and the dish gently warmed. After a few minutes the melt will have dissolved, and at this point the lid should be removed and any spray adhering rinsed back into the dish with the minimum volume of water. During dissolution, or after a short evaporation, samples with higher silica contents (about 30 per cent.) should produce a stiff gel, although those with lower silica contents may not apparently form a gel.It has been found, however, that after evaporating for about 30 minutes sufficient polymerisation has occurred to enable the procedure to be continued. After the formation of a stiff gel or, for samples with low silica content that cannot be expected to form a gel, evaporation for 30 minutes, the dish is removed from the source of heat and half of a Whatman ashless tablet stirred into the gel, followed by 5 ml of the coagulant solution (0-25 per cent.) and, finally, by 10 ml of water to render the mixture less viscous. The solution filters rapidly through a 12.5-cm Whatman No. 42 paper and, after being washed a few times with dilute hydrochloric acid, the precipitate is washed free from chlorides with hot water. The determination is completed by following the procedure given by the British Standard 1902 : Part 2A : 1964.Care should be taken during the “burning off” stage; very slight traces of carbon may remain, but these appear to be negligible in weight. The silica remaining in solution may be determined with ammonium molybdate. Usually it has been found satisfactory to use the solution without adjusting the acidity, although a slight increase in acidity may occasionally be necessary to avoid a slight cloudiness in the solution. Provided care is taken to ensure correct preparation of the solution before the coagulant is added, filtration is fast, but if coagulation is attempted too soon, i.e., before the silica has polymerised sufficiently, filtration will fail.A wide range of samples has been analysed by this technique with success, but more results have been obtained with BCS 315 firebrick. The results for precision and accuracy are shown in Table I. It is then allowed to stand for 5 minutes. TABLE I RESULTS WITH BCS 315 FIREBRICK Accepted value, Found, Constituent per cent. per cent. (mean) Standard deviation Number of results SiO, 51-2 51.29 0.06, 47 141,03 42.4 42-36 0.03 28 Fez03 3-01 3.01 0.02 14 TiO, 1.23 1.22 0.01, 14 The amount of residual silica remaining in solution after removal of the “gravimetric” silica is rather higher by this technique than after evaporation. If the sample is fused in fusion mixture alone, about 0-5 per cent. of silica remains in solution after dehydration; for a borate-containing flux, this figure rises to about 0-8 per cent.In the coagulation method the figure appears to be roughly proportional to the amount of silica in the sample; at the 99 per cent. level the figure may rise to about 2 per cent.; at the 55 per cent. level it averages about 1.5 per cent.; at 30 per cent., slightly less than 1 per cent.; and for material such as bauxite, with silica contents at about 6 per cent., the figure drops below 0.5 per cent. Ex- perience has shown that the precision of the colorimetric determination is usually of the order of 2 per cent. of content, so that given care in the maintenance of conditions maximum errors from this source should not exceed 0.05 per cent. absolute. Spectrographic analysis of the precipitated silica with a direct-reading spectrometer has shown the amount of borate trapped to be less than 0.02 per cent. and, generally, only 0.01 per cent. This compares with a level of about 0.05 per cent., after dehydration of a solution from a borate-containing melt. Work is to be continued to improve the use of the reagent, and to exploit its use over the widest range of ceramic materials. The present procedure, although probably not the best possible, has shown great advantages in speed without loss of accuracy and, as such, is offered for the guidance of other workers in silicate analysis. The authors thank Dr. N. F. Astbury, Director of Research of the British Ceramic Research Association, for permission to publish this paper. REFERENCE 1. TVanninen, E., and Ringbom, A., Analytica Chiwz. Acta, 1955, 12, 308. Received December 9th, 1966

ISSN:0003-2654    

DOI:10.1039/AN9679200466

出版商:RSC    

年代:1967

数据来源: RSC