ISSN:0003-2654    

DOI:10.1039/AN97499FX005

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499BX007

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499FP013

出版商:RSC    

年代:1974

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97499BP019

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

FEBRUARY, 1974 THE ANALYST Editorial Volume 99, No. 11 75 Hazardous Chemicals as Reagents WE live in an age of increasing anxiety about “the quality of life,” and its impairment by technological advances. Fears are expressed about pollution of the atmosphere, pesticide residues in food, the dumping of toxic wastes on land or into rivers and oceans, the safety of industrial processes and transport systems, the carnage and nuisance factors arising from high traffic densities, noise levels in cities, and so on. Legislation relevant to various such factors has been enacted or is under consideration in many countries. The tide of opinion in this direction, and the consequent legislation, are of considerable import for the analyst. It is he who has to measure pollution levels in the atmosphere, the amounts of toxic materials in industrial effluents and in river and other waters, and the concentrations of toxic vapours to which personnel may be exposed in industrial processes.In these and many similar ways the analyst has a vital r81e to play in seeking to contain within reasonable bounds the many hazards and nuisances than can arise in present-day life. The analytical work required in such contexts can call for expertise of a high order, particularly when pollutants or contaminants need to be measured at very low levels of concentration. Against this background it is similarly important for the analyst to ensure that his own house is in order. Clearly the analyst himself should seek to avoid unnecessary hazards in his work either by eliminating the use of dangerous substances or procedures or, when this is im- practicable, by ensuring that the recognised safety precautions are taken.A particular prob- lem for the analyst can arise when a long established and widely used reagent is belatedly found to have hazardous properties; examples are the amines benzidine and o-tolidine, now known to be carcinogenic. This type of problem extends further, into the realms of publication of scientific papers in primary journals. Thus, is it ethical to accept for publication a paper advocating the use of a reagent known to be a hazardous substance, particularly when alternative reagents of a safer nature may be applicable? This dilemma has recently been faced by the Executive Committee for The Analyst, in connection with a paper in this issue (p. 128) concerned with the use of o-tolidine. The decision reached by the Committee was that it had no right to censor or prevent the publica- tion of any scientific paper solely on the grounds of its advocating the use of a reagent known to be a hazardous substance. However, the Committee has deemed it essential that any such paper published in The Analyst should make clear the hazardous nature of the reagent used, and should include information on or references to the appropriate safety precautions to be adopted. This policy has been put into effect in relation to the particular paper on p. 128 of this issue, and will be applied to any future paper in The Analyst involving substances or procedures recognised to be abnormally hazardous. H. J. CLULEY Chairman, Analyst Executive Committee 81

ISSN:0003-2654    

DOI:10.1039/AN9749900081

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

82 Analyst, February, 1974, Vol. 99, pp. 82-92 Ionic Polymerisation as a Means of End-point Indication in Non-aqueous Thermometric Titrimetry Part V.* The Iodimetric Determination of Organic Bases, Hydrazine Derivatives and Water BY E. J. GREENHOW AND L. E. SPENCER (Department of Chemistry, Chelsea College, University of London, Manresa Road, London, S . W.3) A thermometric titration method has been evaluated in which organic bases, hydrazines, phosphines and quaternary ammonium halides, and also water, have been titrated with iodine in non-aqueous solutions containing alkyl vinyl ethers. The latter polymerise with the excess of iodine evolving heat, which marks the end-point. The ratio of the reactants in titrations of most of the amines examined, namely 3.6 to 4.6 atoms or 1.8 to 2.3 molecules of iodine to 1 molecule of amine, depending on the amine, is favourable to the titration.With hydrazine derivatives, the ratio ranges from 4-2 to less than 1 atom of iodine to 1 molecule of the hydrazine, depending on the hydrazine derivative. Water can be titrated with iodine in the presence of alkyl vinyl ethers, about thirteen molecules of water consuming one atom of iodine. The end-point in titrations of most of the compounds examined is marked by a sharp inflection in the titration graph when an automatic procedure is used. Precisions are usually better than 1 per cent. with 0-05 M and 2 per cent. with 0-01 M titrant solutions. Sample sizes down to about 0-0005 mmol, depending on the iodine consumed in the reaction, can be determined with 0.01 M titrant solution.Calibration graphs show that, except in the titration of water, the volume of titrant and amount of sample are linearly related in the range 0 to 1 ml of titrant. The curvatures of calibration graphs for water depend on the rates of addition of iodine to the sample; linearity can almost be achieved a t an appropriate titration rate. It is suggested that the stoicheiometry, i.e., the iodine consumed per molecule of sample, is a quantitative measurement of the basic properties of the compounds investigated. The different stoicheiometries for different compounds make the iodimetric method useful for the selective determination of the constituents of binary mixtures of bases and hydrazine derivatives, but unsuitable for the determination of the total basic or hydrazine function in more complex mixtures.THE determination of organic bases by iodimetric titration has not previously been considered, although reactions between pyridine and iodine at ambient temperature were reported as early as 1895.l Pyridinium bromide perbromide (C,H,N.HBr.Br,) and trimethylphenyl- ammonium tribromide are available as laboratory reagents but the simple amine - halogen addition compounds appear to have received little attention. Indeed, in a recent edition of one of the standard works on organic chemistry,2 it is noted that the reaction between halogens and primary and secondary amines “does not give useful products.” The titration of organic hydrazine derivatives with iodine, bromine or binary inter- halogen compounds of iodine, bromine and chlorine, is an established assay pr~cedure.~ The oxidation reaction is usually, but not always, accompanied by the evolution of nitrogen.With hydrazides, four atoms, i e . , two molecules, of iodine are normally consumed by one molecule of hydrazide, but it has been pointed out3 that reactions of halogens with hydrazine derivatives in general may be complicated by there being more than one possible route and, consequently, variations in the stoicheiometry. It has been stressed that strict adherence to experimental conditions is necessary in order to obtain reproducible results. * For Part IV of this series, see Analyst, 1973, 98, 485. @ SAC and the authors.GREENHOW AND SPENCER 83 Iodimetry, in the form of the Karl Fischer reaction, is the most important chemical method for the determination of water in organic solvents.In this reaction, two atoms of iodine are, in theory, consumed by one molecule of water. Siggia4 has shown that iodine reacts with vinyl ethers in the presence of an alcohol according to the following equation : ROCH=CH, + I, + R’OH -+ ROCH(OR’)CH,I + HI It would seem likely, therefore, that water should undergo a similar reaction in which two atoms of iodine would correspond to one molecule of water. The reaction product in this instance, ROCH(OH)CH,I, could conceivably react with further iodine and vinyl ether, raising the reaction ratio to four atoms of iodine to one molecule of water. The use of catalytic thermometric titration for the determination of organic bases is described in Part I.5 More recently, a catalytic thermometric procedure has been reported6 for the determination of hydrazine derivatives and xanthates and, less effectively, for thiols. In this procedure iodine was used as the titrant and as a catalyst for the cationic polymerisa- tion process that indicates the end-point, and ethyl vinyl ether was used as the monomer.A detailed investigation of this procedure, which forms the basis of the present paper, has shown that in non-aqueous solution iodine can be used as a titrant not only for certain oxidisable compounds, such as hydrazine derivatives, but also for organic bases and water. A number of organic solvents for the titrant (iodine) and titrands have been assessed in an effort to achieve the optimum determination conditions.In addition, iodine bromide and iodine chloride have been examined as alternative titrants and some alkyl vinyl ethers other than ethyl vinyl ether have been examined as alternative monomers. EXPERIMENTAL REAGENTS- Laboratory-reagent grade dimethylformamide, toluene, pyridine, NN-dimethylacetamide, 1,2-dichloroethane, tetrahydrothiophene 1,l-dioxide (sulpholane) and dimethyl sulphoxide were dried over molecular sieve 4A before use. Iodine and acetic acid (both AnalaR grade) were used as received. Ethyl vinyl ether, n-butyl vinyl ether, 2-chloroethyl vinyl ether, isobutyl vinyl ether, other organic bases, benzoic acid, iodine bromide, iodine chloride and iodine trichloride were laboratory-reagent grade materials and were used without further purification.Divinyl ether was extracted with distilled water and dried over alumina before use. Benzenesulphonohydrazide, 4,4’-oxybis (benzenesulphonohydrazide) and 2-h ydroxyethyl- hydrazine were gifts from Fisons Limited, Agrochemicals Division. Other hydrazine deriva- tives were laboratory-reagent grade materials. Solutions of iodine, 0.05 and 0.01 M in organic solvents, were standardised by adding to 20 ml of the solutions 50 ml of water, 1 g of potassium iodide and 3 ml of 1 M sulphuric acid, and titrating with 0.1 and 0.01 M sodium thiosulphate solutions, respectively, with starch as indicator. Solutions of iodine chloride, iodine trichloride and iodine bromide in dimethylformamide, all 0.05 M, were standardised by the method used for 0.05 M iodine solution.APPARATUS- For thermometric titration-Use the automatic apparatus described in Part III7 with an 8-ml tit r at ion flask. For gasometric determinations-Use a 50-ml gas burette, connected to a 50-ml, mag- netically stirred reaction vessel with a side-arm that is fitted with a serum cap. Details of the apparatus were given by Dixon8 PROCEDURE- Thermometric titration-Prepare a solution of the sample in dimethylformamide or in another appropriate solvent. The concentration will depend on the stoicheiometry of the reaction with iodine and on the titrant concentration; thus 1 ml of the solution should contain about 0.1 mequiv of the sample compound when the 0.05 M titrant is used and about 0.02 mequiv with 0.01 M titrant.84 GREENHOW AND SPENCER: IONIC POLYMERISATION FOR END-POINT [Analyst, Vol.99 Transfer by pipette 1 ml of the sample solution into the reaction flask, add 2 ml of ethyl vinyl ether, stir the solution, then add titrant at the rate of about 0.1 ml min-1 from the motor-driven syringe. The titrant volume at the end-point is taken to be the volume corresponding to the point of inflection in the titration graph. When this inflection is in- distinct, the end-point is taken to be the point where the tangent to the main heat rise leaves the graph at its lower temperature end. Carry out a blank titration by using an equal volume of the same batch of solvent, with the same water content, as that used for the sample titration. The chart recorder of the automatic titration apparatus is conveniently operated at a chart speed of 600 mm h-l and in the range 0 to 100 mV with 0.05 M titrant and 0 to 50 mV with 0.01 M titrant.Gasometric determinations-Transfer by pipette 5 ml of a 0.8 M solution of iodine in dimethylformamide into the reaction vessel, sweep out the air in the vessel with dry nitrogen and connect the vessel to the gas burette. Stir the iodine solution and add, by injection from a syringe through the serum cap, 1-ml aliquots of a 0.18 M solution of the hydrazine derivative in dimethylformamide at 5-minute intervals, reading the gas burette immediately before each injection. Correct the observed increases in volume to volumes at S.T.P. and subtract from these values the volume (1 ml) of the sample solution injected. For the reverse of the above procedure, i e ., addition of aliquots of iodine to an excess of the hydrazide, carry out the above operations, but with 4 mmol of hydrazide dissolved in 10 ml of dimethylformamide in the reaction flask and with 1-ml aliquots of 0.5 M iodine. RESULTS AND DISCUSSION A further investigation of the catalytic thermometric titration of isonicotinoylhydrazine with iodine in dimethylformamide solution, reported briefly in an earlier paper,6 has revealed that: (a), addition of an equimolar amount of acetic or benzoic acid to the hydrazide prior to titration has a negligible effect on the titration value; (b), addition of small amounts, e.g., 1 per cent., of pyridine or water to the hydrazide prior to titration significantly increases the titration value but does not reduce the sharpness of the end-point; and (c), benzoyl- hydrazine requires much less titrant (1.4 atoms of iodine per molecule) than does isonicotinoyl- hydrazine (4.2 atoms of iodine per molecule).Observation ( b ) above indicates that both pyridine and water are titrated iodimetrically by the catalytic thermometric procedure. By using the procedure, a number of pyridine derivatives, primary, secondary and tertiary aliphatic and alicyclic amines, quaternary ammonium halides, organophosphorus derivatives and heterocyclic compounds containing two nitrogen atoms, have been deter- mined. In Table I the compounds are listed and the reactivities in terms of the number of atoms of iodine combining with one molecule of the compound are given. Typical titration graphs are shown in Figs.1 and 6. It can be seen that, if the inflection point of the iodimetric titration is taken as the basis for the calculation, pyridine combines with 4.1 atoms of iodine in dimethylformamide solution. This result agrees well with the formula C,H5N.2I, obtained by Prescott and Trowbridge1 for the crystalline adduct of iodine with pyridine. Most of the aliphatic amines and pyridine derivatives required from 3.6 to 4.6 atoms of iodine for each molecule in the titration. It is interesting to note that the secondary amines are less reactive towards iodine than are the tertiary amines, including the pyridine deriva- tives, although the dissociation constants of aliphatic bases show the secondary amines to be stronger bases than the tertiary amines in aqueous solution.Aniline and the alkylanilines, as might be expected from their weakly basic character, show much lower reactivity towards iodine. Substitution of electron-withdrawing groups on the aniline and pyridine rings causes a reduction in the reactivity towards iodine. Thus, +-nitroaniline does not combine with iodine and 2,6-pyridinedicarboxylic acid shows very low reactivity. Diphenylamine and triphenylamine are unreactive for the same reason. In contrast with triphenylamine, triphenylphosphine combines with iodine in the ratio of about 1 atom of iodine to 1 molecule of triphenylphosphine but the inflection at the end- point is not sharp (Fig. 1). The N-oxides of pyridine and 3-picoline were found to react with iodine in an approxi- mate ratio of 2 atoms of iodine to 1 molecule of the N-oxide but, again, there was no distinctFebruary, 19741 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY.PART v 85 TABLE I ORGANIC BASES AND QUATERNARY AMMONIUM HALIDES TITRATED WITH 0-05 M IODINE IN DIMETHYLFORMAMIDE Conditions: 0.025 mmol of base or halide in 1 rnl of dirnethylformamide added to 2 ml of ethyl vinyl ether and titrated by the thermometric procedure Aliphatic and alicyclic amines- n-Butylamine (3.6) ; benzylamine (3.7) ; 1,2-dianiinoethane (6.75) ; morpholine (3.6) ; piperidine (3.6) ; triethylamine (4.2) ; tris(hydroxymethy1)methylamine (4.3) ; 2-NN-diethylaminopropio- nitrile (4.4) ; N-methylmorpholine (4.6) ; and N-ethylpiperidine (4-4) Pyridine (4.1) ; a-picoline (3.8) ; 2,6-lutidine (4.0) ; 2,6-pyridinedicarboxylic acid (0.09) ; pyridine N-oxide (2.2) ; 3-picoline N-oxide (2.2) ; quinoline (4.1) ; 4-methylquinoline (4.3) ; 8-hydroxy- quinoline (2.1) ; and 2-hydroxyquinoline (0) Aniline (1.3) ; o-toluidine (1.5) ; p-toluidine (1.8) ; p-nitroaniline (0) ; diphenylamine (0) ; and triphenylamine (0) Hexamethylenetetramine (5.0) ; phthalazine (3.5) ; quinoxaline (0) ; and benzimidazole (4.0).Tetra-n-butylammonium bromide (1 -8) : tetra-n-butylammonium iodide (1-5) ; cetyltrimethyl- ammonium bromide (1.4) ; cetylpyridinium bromide (1.8) ; and benzyldimethylmyristyl- ammonium chloride (2.4) Triphenylphosphine (1.1) ; and triphenylphosphine oxide (0) Figures in parentheses following the name of the compound denote the number of iodine atoms combining with one molecule of the compound on the basis of the iodimetric titration. Pyridine derivatives- Aniline derivatives- Heterocyclic nitrogen compounds- Quniernary ammonium halides- Phosphorus compounds- inflection point in the titration graph and the end-points were difficult to establish (Fig.1). Triphenylphosphine oxide did not react with iodine. Quinoline and 4-methylquinoline are similar to pyridine and the methylpyridines in their reactivities towards iodine, but 8-hydroxyquinoline is far less reactive and 2-hydroxy- quinoline shows no reaction. Presumably hydrogen bonding in the last two compounds inhibits or prevents formation of adducts with iodine. a h 1 I 1 I 0.05~ iodine reagent/ml (1 division = 0.5 ml) Fig. 1. Catalytic thermometric titration of organic bases, water and quaternary ammonium halides with 0.05 M iodine reagent. Compounds/mg : a, n-butylamine, 0.84; b, tris(hydroxymethyl)methylamine, 1-2 ; c, pyridine, 1.3 ; d, hexamethylenetetramine, 1.6; e, morpholine, 1-89: f, 8-hydroxyquinoline, 2.9; g, pyridine N-oxide, 2.3; h, tri- phenylphosphine, 7.4 ; j, cetyltrimethylammonium bromide, 4.0 ; k, tetra-n-butylammon- ium iodide, 9.3 ; m, benzyldimethylmyristylammonium chloride, 5.0 ; and w, water, 10.086 GREENHOW AND SPENCER: IONIC POLYMERISATION FOR END-POINT [Analyst, Vol.99 A number of heterocyclic compounds containing more than one ring nitrogen atom have been titrated. Hexamethylenetetramine combines with only five atoms of iodine instead of the possible sixteen, calculated on the basis of four atoms of iodine for each nitrogen atom.Quinoxaline proved to be unreactive but phthalazine, an isomer of quinoxaline, com- bined with 3-5 atoms of iodine per molecule. Presumably the more remote nitrogen atoms in phthalazine are less affected by the benzene ring. Benzimidazole, with one of its two nitrogen atoms unconjugated to the benzene ring, is even more reactive than phthalazine. Quaternary ammonium halides, including alkylpyridinium halides, can be titrated by means of this iodimetric method. With the compounds examined it was found that from 1.4 to 2.4 atoms of iodine combined with one molecule of the halide. There appears to be a tendency to form trihalides, which are known to be stable systems, and for the halide ion to influence the reactivity. The possibility of water reacting with iodine in the presence of vinyl ethers was discussed in the introduction.It has been confirmed that water can be titrated in solution in dimethyl- formamide and that a sharp end-point is obtained (Figs. 1 and 6). There is some indication that the presence of water increases the sharpness of the end-point in the titration of other compounds. In the titration of water, the consumption of the iodine titrant at the indicated end-point is dependent on the rate of addition of the titrant, as shown in Fig. 2. It can be seen that there is an almost linear relationship between the water content of the sample and the titrant required when titrant is added a t a rate of 0.06ml min-l. This rate would, therefore, be the recommended rate of addition of titrant in the determination of water and, at this rate of addition, about one atom of iodine is consumed by thirteen molecules of water.Clearly, the method is not highly sensitive for the determination of water because 23.4 mg of water would require only 1 ml of 0.05 M iodine solution on the basis of the above reaction ratio. 0.05~ iodine reagent/ml Fig. 2. Effect of the rate of addition of titrant on calibration graphs in the thermometric titration of water with 0.05 M iodine reagent. Rate of addition of titrantlml mi@: a, 0.02; b, 0.032; c, 0.06; d, 0.12; e, 0.20; and f, 0.60 The reaction of water, iodine and ethylvinylether in dimethylformamide does not appear to be simple and, if it is similar to the reaction described by Siggia4 for alcohols, obviously does not proceed to completion during the course of the titration.It is possible, however, that water is being titrated merely as a weak base. As water is titrated, an allowance must be made when organic solvents containing trace amounts of water are used in the titration of bases, hydrazine derivatives, etc. A convenient procedure is to carry out a blank titration on an equal volume of the same solvent, takenFebruary, 19741 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART v 87 from the same batch as that used for dissolving the sample. With dimethylformamide dried over molecular sieve 4A, the blank titration with 0-05 M iodine is about 0.15 ml. It is important, of course, that the sample itself should be dry. In addition to isonicotinoylhydrazine, a number of other hydrazine derivatives have been determined, including hydroxyalkyl-, dialkyl-, aryl- and diarylhydrazines, arylhydra- zides, arylsulphonohydrazides and semicarbazides. In Table 11, the reactivities of these compounds towards iodine are given in terms of iodine atoms per molecule.TABLE I1 HYDRAZINE DERIVATIVES TITRATED WITH 0.05 M IODINE I N DIMETHYLFORMAMIDE Conditions: Sufficient compound to give a titration value of about 0.5 ml is dissolved in 1 ml of dimethylformamide, added to 2 ml of ethyl vinyl ether and titrated by the thermometric procedure NiV-Dimethylhydrazine (3.4) ; 2-hydroxyethylhydrazine (2.2) ; phenylhydrazine (1 -0) ; 4-nitro- phenylhydrazine (1.1) ; 2,4-dinitrophenylhydrazine (0.18) ; NN-diphenylhydrazine (0.86) ; benzoyl- hydrazine (1.4) ; isonicotinoylhydrazine (4.2) ; benzenesulphonohydrazide (0.33) ; 4,4’-oxybis(ben- zenesulphonohydrazide) (0.63) ; semicarbazide hydrochloride ( 1.9) ; and thiosemicarbazide (3.1) Figures in parentheses following the name of the compound denote the number of iodine atoms combining with one molecule of the compound on the basis of the iodimetric titration.The alkylhydrazines, i.e., NN-dimethylhydrazine and 2-hydroxyethylhydrazine, were the most reactive of the hydrazine derivatives examined, if one excepts thiosemicarbazide and isonicotinoylhydrazine, which possess reactive groups other than the hydrazine group. Titration graphs for some of the hydrazine derivatives are shown in Fig. 3. It can be seen that in the titration of benzenesulphonohydrazide, 4,4‘-oxybis(benzenesu1phonohydrazide) a h” 0.05~ iodine reagent/ml (1 division = 0.5 ml) Fig.3. Catalytic thermometric titration of hydrazine derivatives with 0.05 M iodine reagent. Compounds/mg : a, 2-hydroxyethylhydrazine, 1.1 ; b, NN-dimethyl- hydrazine, 0.65; c, NN-diphenylhydrazine, 9-2 ; d, 4-nitrophenylhydrazine, 5.1 ; e, 2,4- dinitrophenylhydrazine, 20.2 ; f, benzoylhydrazine, 3.4; g, isonicotinoylhydrazine, 0.85 ; h, * phenylhydrazine, 11.0; j, 4,4’-oxybis(benzenesulphonohydrazide), 26.6 ; and k, * benzenesulphonohydrazide, 28.9. * Titrant, 1 division = 0.83 ml; temperature, 1 division = 2.0 “C88 GREENHOW AND SPENCER: IONIC POLYMERISATION FOR END-POINT [Analyst, Vol. 99 and phenylhydrazine, iodine is consumed after the first sharp inflection in the titration graph, which has been taken as marking the end-point of the titration.It is probable that a slow evolution of nitrogen occurs after the initial reaction with iodine and, ultimately, four atoms iodine would be required for each hydrazine group. Any further reaction with nitrogen evolution would, of course, be accelerated by the rising temperature caused by the ionic polymerisat ion. Observations (b) and (c) above suggest that, in dirnethylformamide, iodine reacts with the heterocyclic nitrogen as well as with the hydrazine group of the isonicotinoylhydrazine. If the reaction of the 4.2 atoms of iodine with each molecule of isonicotinoylhydrazine occurred only at the hydrazine group, the reaction could be explained essentially by an equation similar to that which applies to the reaction in aqueous solution: In this instance a molecule of nitrogen would be liberated from each molecule of hydrazide.Gasometric experiments with hydrazides have been carried out to determine whether the iodine reactant is consumed in reactions involving the evolution of a gas. Aliquots (1 ml) of a solution of isonicotinoylhydrazine in dimethylformamide were added to an excess of iodine in 5 ml of dimethylformamide at ambient temperature (25 “C in this experiment) and the volume of nitrogen evolved after the addition of each aliquot was measured. The experiment was then repeated with benzoylhydrazine. The results of the experiment are summarised in Fig. 4. With both hydrazides, the rate of evolution of nitrogen after addition of the second aliquot was almost constant until five aliquots had been added, when it began to decrease.During the “steady state,” the rate of nitrogen evolution was about the same for both hydrazides, 0.6 mmol of nitrogen being released from each millimole of hydrazide, which is equivalent to a reaction ratio of 2.4 atoms of iodine to 1 molecule of hydrazide if two molecules of iodine are required for the release of one molecule of nitrogen. 0.1 8 M hydrazide in .dimethyl f ormamide/mmol 5 0 1 2 3 4 5 0.5 M iodine in dirnethylforrnarnidefmrnol Fig. 4. Gasometric measurements in the reaction of isonicotinoylhydrazine and benzoylhydrazine with iodine in dimethylformamide solution : a, addition of 0.18- mmol aliquots of isonicotinoylhydrazine to 4 mmol of iodine; b, addition of 0-18-mmol aliquots of benzoylhydrazine to 4 mmol of iodine; c, addition of 0.5-mmol aliquots of iodine to 4 mmol of benzoylhydrazine; and d.addition of 0.5-mmol aliquots of iodine to 4 mmol of isonicotinoylhydrazine The experimental conditions are not the same as those obtaining in the iodimetric titration because the ethyl vinyl ether is omitted for obvious reasons. However, it is reasonable to assume that the reactivities of the two hydrazides are similar when measured by the amount of nitrogen evolved and that the much greater consumption of iodine by isonicotinoylhydrazine,February, 19741 INDICATION IN NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART v 89 4.2 atoms compared with 1.4 atoms per molecule, is caused by a reaction involving the heterocyclic nitrogen of the latter compound.When the reverse procedure is carried out, i.e., when l-ml aliquots of iodine are added to an excess of the hydrazide in dimethylf onnarnide, nitrogen is evolved immediately following addition of the first aliquot in both instances (Fig. 4). Thus, there is no evidence from this experiment that with isonicotinoylhydrazine a reaction with the heterocyclic nitrogen takes precedence over a reaction with the hydrazide group. With all of the samples examined, except water, calibration graphs proved to be linear in the range 0 to 1 ml of titrant when titrant was added at rates of 0.05 to 0.2 ml min-l. The effect of the rate of titrant addition on the calibration graph for isonicotinoylhydrazine is shown in Fig. 5. Increasing the rate of addition of the titrant can be seen to displace the linear calibration, but, because the lines are parallel, the calibration factor remains constant 0-75 0.05~ iodine reagedm1 Fig.5. Effect of the rate of addition of titrant on calibration graphs in the thermometric titration of isonicotinoyl- hydrazine with 0-05 M iodine reagent. Rate of addition of titrantlml min-1: a, 0.02; b, 0.032; c, 0.06; d, 0.12; and e, 0-20 The sharp inflection at the end-point with most of the titrations gives this catalytic thermometric method a slightly higher precision than the corresponding acid - base titrations described in Parts I5 and 11,9 in which the 0.05 M titrant was used. With this titrant, coefficients of variation of less than 1 per cent. have been obtained in most of the titrations. With 0.01 M titrant, sharp end-point inflections were still obtained but the precision was of the order of 2 per cent.Some precision values are given in Table 111. Titration graphs obtained with 0.01 and 0.005 M titrants are shown in Fig. 6. A number of solvents have been examined as possible alternatives to dimethylformamide. Toluene, NN-dimethylacetamide, 1,2-dichloroethaneJ tetrahydrothiophene 1, l-dioxide (sul- pholane), propylene carbonate and 2-methoxypropionitrile can be used as solvents for the iodine titrant or for the sample. Toluene gives less sharp end-points than does dimethyl- formamide with some compounds and the latter solvent has been preferred in the present study. Several alkyl vinyl ethers were assessed before ethyl vinyl ether was chosen as the monomer for this detailed investigation.The alternatives, n-butyl vinyl ether, isobutyl vinyl ether, 2-chloroethyl vinyl ether and divinyl ether, all proved to be inferior to the chosen monomer. No discernible end-points were obtained with the last two ethers but n-butyl and isobutyl vinyl ethers gave acceptable end-points in the titration of amines (Fig. 7). As interhalogen compounds, such as iodine bromide and iodine chloride, have been used as titrants in oxidation - reduction reactions, some of these compounds, namely iodine bromide, iodine chloride and iodine trichloride, have been tried as alternatives to iodine90 GREENHOW AND SPENCER : IONIC POLYMERISATION FOR END-POINT {ArtdySt, VOl. 99 TABLE I11 RESULTS FOR PRECISION FROM THE THERMOMETRIC TITRATIOX OF ORGANIC BASES, HYDRAZINE DERIVATIVES AND WATER WITH 0.05 AND 0.01 M IODINE SOLUTION IN DIMETHYLFORMAMIDE" Compound Pyridine * ... .. Benzoylhydrazine . . . . Tris(hydroxymethy1) methylamine NN-Dimethylhydrazine . . Isonicotinoylhydrazine . . Benzenesulphonohydrazide . . Thiosemicarbazide . . .. Water . . .. . . .. Tris (hydroxymethyl) methylamine Isonicotinoylhydrazine . . a-Picoline . . .. .. Amount takenlmg . . 1-03 . . 1.20 . . 0.86 . . 6.64 . . 1-70 . . 28.9 . . 2.34 . . 17.0 . . 0.24 . . 0.34 . . 0.28 Ti tran t Mean molarityt n$ titre/ml 0.05 0.05 0.05 0.05 0-05 0.05 0.05 0.05 0.0 1 0.01 0.01 4 3 4 3 4 3 3 4 3 3 3 0.49 0.62 0.44 0.66 0.38 0-55 0.70 0.66 0.55 0.54 0-59 Standard deviation 0.0025 0-0054 0.0026 0*0040 0.0024 0.0058 0.0082 0.0045 0.010 0.0035 0,0064 Coefficient of variation, per cent.0.51 0.88 0-59 0.61 0-63 1.06 1.18 0-68 1-82 0.64 1.07 * By thermometric procedure; titrant added a t 0.12 ml min-1. t Nominal value. $ Number of determinations. in solution in dimethylformamide for the thermometric titrations. All three of these inter- halogen compounds gave inferior titration graphs to that of iodine in the titration of pyridine, although all gave rises in temperature of the same order as that obtained with iodine. Although the iodimetric titration method is suitable for the assay of single compounds when interfering substances are absent, the method is obviously unsatisfactory for the determination of the total amount of a particular functional group in a complex mixture because different compounds combine with iodine in different molar ratios.However, Iodine reagent/ml (1 division = 0.5 ml) Fig. 6. Catalytic thermometric titration of organic bases, hydrazine derivatives and water with 0.01 and 0-005 M iodine reagents. a b C d e f Compound/mg . . A,O.54 B,O*24 C,0-28 D,O-14 E,2.0 E,l-O Titrant/M . . . . 0.01 0.01 0.01 0.005 0.01 0-005 Recorder/mV range 50 100 100 50 100 100 Compounds : A, isonicotinoylhydrazine ; B, tris(hydroxymethy1) methylamine; C, cc-picoline; D, pyridine; and E, waterFebruary, 19741 INDICATION I N NON-AQUEOUS THERMOMETRIC TITRIMETRY. PART V 91 differences in reaction ratios make it possible to determine selectively the components of binary mixtures. A calibration graph for the analysis of a mixture of benzoylhydrazine and isonicotinoylhydrazine is shown in Fig.8. d I I 1 1 0.05~ iodine reagent (1 division = 0.5 mi) Fig. 7. Effect of monomer on end-point sharpness in blank titrations and in titrations of pyridine with 0.05 M iodine reagent. a, b and c, blank titrations (1 ml of dimethylformamide) ; d, e and f, titrations of pyridine (0.01 mmol). Monomers: a and d, ethyl vinyl ether; b and e, n-butyl vinyl ether; and c and f, isobutyl vinyl ether The results of titrations carried out with organic bases suggest that the number of iodine atoms that combine with one molecule of the base can be used as a measure of the strength of the base in the solvent used. In addition, the widely differing reactivities of the various compounds examined should make it possible to use the iodimetric titration in order to elucidate the structure of more complex heterocyclic nitrogen compounds and hydrazine derivatives. With bases, such a procedure for investigation of structure could be used in conjunction with that proposed in Part L5 t 5 0-25 0 __-___----------------- Blank titration 0.8 0.6 0.4 0.2 o 0.025~ reagent I/ml 0.2 0.4 0.6 0.8 I 0.025~ reagent Il/ml Fig.8. Calibration graph for the determination of benzoyl- hydrazine and isonicotinoylhydrazine in binary mixtures by catalytic thermometric titration with 0.05 M iodine reagent. Reagent I, isonicotinoylhydrazine in dimethylformamide ; reagent 11, benzoyl- hydrazine in dimethylformamide92 GREENHOW AND SPENCER The fact that, for many compounds, this iodimetric method gives results which indicate that iodine combines with the titrand in fractional molar amounts should not detract from its value as an analytical method. The requirements of reproducibility of results and linearity of calibration graphs can be met. Fisons Limited, Agrochemical Division, are thanked for gifts of chemicals and Mr. S. F. George is thanked for the construction of apparatus. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. Note-References 5, 7 and 9 are to Parts I, I11 and I1 of this series, respectively. Prescott, A. B., and Trowbridge, P. F., J . Amer. Chem. SOL, 1895, 17, 859, Coffey, S., Editor, “Rodd’s Chemistry of the Carbon Compounds,” Second Edition, Volume 1, Cheronis, N. D., and Ma, T. S., “Organic Functional Group Analysis by Micro and Semirnicro Siggia, S., Analyt. Chem., 1948, 20, 762. Greenhow, E. J., and Spencer, L. E., Analyst, 1973, 98, 81. Greenhow, E. J., Chew. G. Ind., 1973, 697. Greenhow, E. J., and Spencer, L. E., Analyst, 1973, 98, 98. Dixon, J. P., “Modern Methods of Organic Microanalysis,” Van Nostrand Co. Ltd., London, Greenhow, E. J., and Spencer, L. E., Analyst, 1973, 98, 90. Part B, Elsevier Publishing Company, London, 1965, p. 128. Methods,” Interscience Publishers, New York, 1964, p. 289. 1968, p. 211. Received July 19th, 1973 Accepted August 17th, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900082

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

Analyst, February, 1974, Vol. 99, p p . 93-107 93 Problems in the Determination of Carbon in Steel by a Precision Coulometric Method BY J. D. HOBSON AND H. LEIGH (Dunford Hadfields Ltd., East Hecla Works, Shefield, S9 1TZ) An apparatus for the coulometric determination of carbon in steel with a precision approaching 1 part in 1000 at the 1 per cent. level has been con- structed. The measures needed t o overcome sources of error in gas absorption, pH measurement, control and coulometric measurement are described. The design avoids the use of a gas proportionating pump and can be used with a 0.2-g sample containing up to 2 per cent. of carbon, and proportionately larger amounts up to 2 g for lower carbon contents. Results on a series of standard steels are given. A RECENT paper by Boniface and Jenkins1 has summarised the choice of methods currently available for the determination of carbon in steel and given reasons for the choice of the coulometric principle for development as a reference method.A similar need arose in this laboratory from the tendency towards narrower specification limits for carbon in low-alloy steels intended for automated heat treatment in the engineering industry. From private information about the work of the Carbon Study Group of the Analysis Committee of B.I.S.R.A., later described by Scholes a t the 23rd Chemists’ Conference,2 it became clear in 1968 that the coulometric method was capable of producing results of excellent precision and had the advantage of being an absolute method, independent of chemical standard- isation.Visits to laboratories in Germany confirmed these impressions but although two types of apparatus have been described in the literat~re,~-~ which are now in commercial production, they have not become popular in the United Kingdom, possibly because of their relatively high cost. However, a simple manual coulometric procedure has been given6 as a German standard procedure for steels with less than 0.1 per cent. of carbon.5 A later version dealing with all classes of iron, steel and ferro-alloys is now described in a supplementary v o l ~ m e , ~ but this standard refers only to the use of commercially available apparatus operated according to the maker’s instructions. The rapidly decreasing cost of modern integrated circuits and digital displays makes electronic integration an attractive basis for precision measurement.It was therefore decided to develop our own automatic coulometric carbon-in-steel analyser, and this paper describes some of the little known problems encountered, and their solution. CHOICE OF EXPERIMENTAL CONDITIONS Difficulties with retention of carbon dioxide in the coulometric cell have caused previous workers to limit the carbon content of the specimen to a maximum of 1 mg1,4,5 and often to as little as 0.2 or 0.3 mg.3,6 An expensive precision proportionating pump, usually of 1 : 9 ratio has been needed to deal with higher carbon levels3~~~~ ; indeed Thomich* has advocated the use of an additional 1:4 ratio pump for carbon contents greater than 1.5 per cent. In the manual titration methods of Boniface and Jenkins and of the Verein Deutscher Eisenhiittenleute, timing of a fixed coulometric titration current of 20 mA is used.One of the commercial instruments3 uses currents up to 321.2 mA integrated by a low-friction current motor in earlier versions, and electronic integration later; the other5 uses a count of variable frequency pulses of 100-5 mil from a known capacitor. Our aim was to develop a method with a precision of 1 part in 1000 at the 1 per cent. of carbon level, i.e., a sensitivity of better than 2 pg, and capable of dealing with up to 2 per cent. of carbon without the need to use a precision proportionating pump. A typical analytical balance has a tolerance of 0.1 mg, so the amount of sample could not be reduced below 0.2 g without using a micro-balance, which would have been an undesirable restriction.The proposed apparatus had therefore to be capable of handling up to 4mg of carbon, and it was obvious that there would be a serious problem in ensuring retention of carbon dioxide 0 SAC and the authors.94 HOBSON AND LEIGH: PROBLEMS IN THE DETERMINATION OF CARBON [Analyst, VOl. 99 in the absorption cell. Much higher coulometric titration currents would also be required, because 4mg of carbon would require 64C, equivalent to an average of just over 0 6 A , if generated steadily for 2 minutes. The absorption - coulometric titration system chosen for study was based on that described by Abresch and ~o-workers~~~ in which the reagent used is alkaline barium per- chlorate, rather than barium chloride, which avoids the generation of chlorine as a side reaction.End-point detection was chosen to be a combined glass- calomel rather than a platinum electrode in weak hydrogen peroxide s ~ l u t i o n ~ , ~ because the former allowed a much simpler glass cell assembly to be constructed, although it was known that this choice might introduce problems from electrical interference in the high-impedance circuits. In order to develop a method capable of measuring down to 1 part in 1000, the first experiments comprised investigations into the precision that would be needed in measuring or maintaining variables of the absorption - coulometric titration cycle. On the assumption of nine randomly distributed contributory errors, each would have to be kept below about 1 part in 3000.Following German p r a ~ t i c e , ~ ~ ~ an experimental cell was set up containing 140ml of 5 per cent. m/V barium perchlorate solution, and 0.033 N sodium hydroxide solution was added from a 10-ml burette, the changes in pH monitored by a conventional pH meter being noted. Each 1 ml of standard alkali solution was equivalent to 200 pg of carbon or 0.1 per cent. in a 0.2-g sample. The curve obtained by plotting carbon equivalent against pH, and against millivolts output from the combined electrode, allowed the rate of change of pH and millivolts to be determined. In similar experiments, barium hydroxide was added or generated electrolytically and a graph of pH or electrode millivolts against millicoulombs was estab- lished. Fig. 1 shows the form of the curves.It was anticipated that the coulometric cell would generate heat, necessitating temperature control, and the titrations showed the curves to be of similar shape at room temperature and a t 37 "C. , , , , f 1 , I , 1 3 0 0 7~ 400 800 1200 1600 2000 Carbon/pg Relationship between pH, electrode potential and equivalent carbon content in a 140-ml coulometer cell using an electrode with E , a t pH 2 Fig. 1. From the curve it is clear that the inflection and true end-point are close to pH 7, but in this region absorption of carbon dioxide would be expected to be very incomplete. The relevant literature suggested that other workers had found a pH of about 10 to 11 to be necessary for complete absorpti~n.~?~ Reading from the graph, for the cell containing 140 ml of 5 per cent.m/V barium perchlorate solution, at pH 10.2, the determination of carbon with a precision of 2 pg required the operating end-point error to be less than 0.0046 pH unit, equivalent to an error of less than 0.268mV from the electrode system in the absence of any other variables.February, 19741 IN STEEL BY A PRECISION COULOMETRIC METHOD 95 Experiments with this cell, initially with manual titration and later with automatic titration, produced satisfactory results with 1.4 mg of carbon (0.7 per cent. on a 0.2-g sample), but with higher carbon contents reproducibility and precision of results were unsatisfactory. During these early experiments, a number of problems were encountered that would have to be resolved in the determination of carbon in steel up to the 4-mg level without the use of fractionating devices.The necessary solutions were found concurrently during the development period, but for clarity each problem, and the modification leading to the final design of the apparatus, is described separately. FACTORS TO BE CONSIDERED Chemistry and electrochemical equivalence Temperature control Absorption of carbon dioxide pH electrode Coulometer cell Coulometer Combustion conditions Other factors affecting the precision of the method CHEMISTRY AND ELECTROCHEMICAL EQUIVALENCE- The steel is burnt in an excess of oxygen and its carbon content converted into carbon dioxide, which is absorbed in the cathode compartment of the coulometric cell, thus precipitat- ing barium carbonate and liberating the equivalent amount of hydrogen ions.The original pH is restored by electrolysis. Many reactions are possible but the following indicate the electrical equivalents and formation of the final products. Cathode compartment- Absorption : Ba2+ + CO, + H,O + BaCO, + 2Hf (precipitate) Electrolytic titration : 2H+ + 2e- --+ H, (gas) Anode compwtment- Electrolysis : H,O + 2H+ + i(0,) + 2e- (gas) Chemical neutralisation : 2Hf + BaCO, -+ Ba2+ + CO, + H,O (sas) Thus in this process, one carbon atom =- 2e- 12.010 g of carbon = 2 F sz 2 x 96 495 C by using the value for the Faraday obtained by Craig and Hoffman,lo from which 4 mg of carbon correspond to 2-0 per cent. in a 0.2-g sample and 64.276 C. TEMPERATURE CONTROL- The change in potential from a pH-sensing electrode for a change of 1 pH unit is given by 2.303 RT F E = From this equation, the electrode e.m.f.depends upon the absolute temperature ; sub- stitution of the appropriate constants gives a rate of change of e.m.f. of 58.16 mV per pH unit at 20 "C, but 62.13 mV per pH at 40 "C. For older electrodes, the output is usually zero at about pH 2 ; for modern electrodes the potential is usually arranged to be zero at about pH 7.96 HOBSON AND LEIGH: PROBLEMS I N THE DETERMINATION OF CARBON [AnabSf, VOl. 99 The effect of temperature on the apparent pH is illustrated in Fig. 2. With a combined electrode that has E, at pH 2, to maintain the potential within 0.268mV (equivalent to 2 pg of carbon in a 140-ml cell) at pH 10.2 requires the temperature stability to be better than 0.165 "C; with an electrode with E, at pH 7 the temperature problem is eased to main- t aining temperature stability within 0.422 "C.However, each of the contributory errors must be one third of these values, equivalent to 2/3pg of carbon, and as the cell volume was subsequently increased to 350 ml, the required temperature stability with an electrode with E, at pH 7 had to be better than 0.055 "C (see Absorption of carbon dioxide). 0 2 4 6 8 10 12 14 PH Fig. 2. Effect of E,, pH and temperature on the electrode e.m.f. Some rough measurements showed that the resistance of the cell might be expected to be about 100 62. Assuming the worst case of 2 per cent. of carbon from a 0.2-g sample of steel, and the electrolysis spread over about 2 minutes at an average current of just over 0.5 A, the heat input to the cell would be 825 cal, causing a rise of about 5-9 "C in 140 ml of electrolyte or 2.4 "C in 350 ml.In order to avoid the added complication of a water-cooling system, we decided to run the cell above room temperature. The heat dissipated in the cell during the high-current phase of a titration is removed by a vertical current of air at room temperature. The air enters the case of the magnetic stirrer motor from a small blower, and leaves via a ring of holes surrounding the base of the glass coulometer cell. During standby conditions and at the end of a titration the cell is maintained above room temperature by a small electrical resistance heater immersed in the electrolyte, and controlled by a sensing device also in the cathode compartment.A practical difficulty was to produce a heater of sufficient power output, yet of sufficiently low thermal capacity and small physical size to give rapid control, and efficiently insulated to withstand the rather aggressive alkaline barium perchlorate solution. The original design was a 25-W wire-wound heater immersed in silicone oil in a glass envelope. This heater had a large time constant and was unable to maintain the cell tem- perature within the specified limits when absorbing the carbon dioxide from 4 mg of carbon. A new design with a printed circuit heater enclosed in a shrinkable silicone sheath had a smaller thermal equivalent, better conductivity and a much smaller time constant. It also gave a higher power output with little change in physical size.The improvement in control with the new heater is shown in Fig. 3, which also illustrates the advantage in operatingFebruary, 19741 I N STEEL BY A PRECISION COULOMETRIC METHOD 97 -0.02 c i 1 2 3 4 5 6 7 8 9 Timelm i nu t es Fig. 3. Effect of heater design on temperature control of coulometer : (a), original heater (room tempera- ture 20.5 "C, cell a t 32.9 "C; ( b ) , redesigned heater (cell at 32.9 "C) ; and ( G ) , redesigned heater (cell at 40 "C) the cell with a greater cooling differential. From a study of cooling curves of the cell elec- trolyte it was concluded that a temperature differential of at least 5 "C between the cell operating temperature and ambient room temperature was desirable. Eventually, it was decided that the optimum operating temperature would lie in the range 35 to 37 "C.The temperature control system that has been developed monitors by means of a thermistor in the input circuit of a transistorised amplifier. The maximum output of about 2-5 A into the heater of low therrnal capacity enables up to 42 W to be applied to the cell. The time constant is very small and the temperature is maintained steady within &O-015 "C, which is within the range required by other parameters. ABSORPTION OF CARBON DIOXIDE- Gas dispersion-A major problem with cells for the absorption of carbon dioxide is to produce a very fine gas dispersion, and in the past this has been achieved by the use of high-speed reciprocating or spinning gas distributor^,^ or by other complicated mechanical means.The need is for either very small bubbles or for a long path length in the electrolyte,98 [Analyst, vol. 99 or a convenient compromise. Our cell is designed to use a simple encapsulated stirrer, driven magnetically from below the cell, together with gas distribution through a porous glass frit. The problem of producing fine gas bubbles was investigated photographically by using electronic flash. It was found that grade 3 glass sinters (pore size 20 to 30 pm) produced no improvement over grade 2 (40 to 6Opm) because the greater resistance to flow raised the gas pressure and the bubbles expanded to the same size on liberation. The addition of an agent for causing frothing and reducing surface tension was found to be helpful. Methanol, ethanol, butanol and pentanol were tried, and also commercial wetting agents, but the traditional addition of propan-2-01~~~ was shown to be the most effective in reducing mean bubble size and suppressing the occasional formation of large bubbles.Additions from 0.1 per cent. to the normal 1 per cent. were tried; no advantage was found with concentrations greater than 0.3 per cent. Fig. 4 shows portions of typical high-speed photographs, from which it can be seen that most of the bubbles have a diameter of about 0.01 inch with the selected conditions. Efficient absorption of carbon dioxide is improved as the period in contact with the absorption solution is increased, therefore a long bubble path is desirable. A stroboscope was used to investigate the action of the magnetic stirrer.Although the manufacturer gave a maximum speed of 1250 r.p.m., it was discovered that eddy currents induced in its steel cover-plate were in fact reducing the true speed to 750 r.p.m. When a non-magnetic top-plate was fitted to the stirrer, speeds up to 2100 r.p.m. could be obtained. An immediate improvement in carbon dioxide absorption was found, and the stirrer is now run a t 1800 r.p.m. Improvements in stirring speed and cell arrangement also allowed the use of a grade 1 glass sinter gas distributor, with a corresponding drop in back-pressure in the pumping system, together with a reduced tendency to blockage of the distributor by precipitated barium carbonate. Cell eZectmZyte-In the original cell containing 140 ml of 5 per cent. m/V barium per- chlorate solution, with a maximum current of 0.7 A, not more than 2 mg of carbon could be retained completely during a rapid combustion, and an improved circuit capable of delivering 1.8 A was still inadequate for titrating 4 mg of carbon because of an excessive reduction in pH before recovery during coulometric titration.The possibility of slowing down the introduction of carbon dioxide was rejected because of the resulting long titration times; after much experimentation the volume of the cell was increased to 350 ml. During this work, various combinations of barium perchlorate elec- trolyte at 5, 10,20, 30 and 40 per cent. m/V concentration were tried as anolyte and catholyte. The slope of the titration curve was greatly reduced at high concentration, which corre- spondingly increased the problems of temperature control and pH measurement.Low concentrations in the anode compartment appeared to be unable to supply the barium-ion migration required by high currents, and the use of different concentrations for anolyte and catholyte caused problems due to diffusion. Eventually the use of 20 per cent. m/V barium perchlorate solution was adopted as the best compromise, allowing high currents, and giving complete absorption of carbon dioxide with adequate sensitivity at the end-point. The effect of barium perchlorate concentration on the sensitivity of the pH - coulomb equivalent curve can be seen in the two typical graphs of Fig. 5, in which small changes in choice of end-point cause large differences in the sensitivity of detection required.Raising the cell operating temperature, with other conditions remaining constant, also decreases the slope of the pH - coulomb equivalent curve, two typical examples being shown in Fig. 6. Here the millivolt output from the sensing electrode alters both its absolute value and its temperature - pH coefficient and these variables also affect the sensitivity required for end-point detection. A few of the results accumulated in the experiments to assess the minimum pH required for complete carbon dioxide absorption are given in Table I, and are for a cell containing 20 per cent. barium perchlorate electrolyte operating at 37 "C and with propan-2-01 additions to aid the formation of fine bubbles. There was no significant loss of reproducibility until the end-point had been lowered below pH 9.35.The improved circuits and solenoid valve described in other sections now prevent the sensing electrode potential from falling more than 10 mV, equivalent to 0.16 pH unit, even with very rapid absorption of carbon dioxide equivalent to 4 mg of carbon, thus allowing the choice of pH 9.5 as the operating point with the cell conditions described above. HOBSON AND LEIGH: PROBLEMS I N THE DETERMINATION OF CARBONa , . a- iis [To face p. 98February, 19741 7 I N STEEL BY A PRECISION COULOMETRIC METHOD 1 1 I 1 1 1 I 1 1 1 0. 400 800 1200 1600 2000 I i Equivalent carbon conten t/pg Fig. 5. Effect of barium perchlorate concentra- tion on the relationship between pH and coulometric carbon sensitivity a t 37 "C. Barium perchlorate con- centration: A, 10 per cent.m/V; and B, 30 per cent. m/V. Points I to IV- Point pH mV m V r 1 p g o f C I 9-52 195 0.037 I1 9.69 205 0.024 I11 9.44 190 0,022 IV 9.52 195 0.018 99 PH ELECTRODE- It is obvious t h t the precision of the instrument is not only dependent on th ability to detect small changes in the potential of the electrolyte, but also on the stability of the mV-pH output of the pH electrode. Therefore, much investigation into the choice of the most suitable electrode has been needed. 7 0 5 10 35 20 25 30 Coulombs 1 I I I I 1 0 400 800 1200 1600 2000 Equivalent carbon content/pg Fig. 6. Effect of operating temperature on the relationship between pH and coulometric carbon sensitivity. Barium perchlorate concentration, 30 per cent. m/V.Temperature: A, 22 OC; and B, 37 "C. Points I to IV- Point pH mV m V r 1 p g o f C I 9.70 195 0.025 I1 9.86 205 0.020 I11 9.44 190 0-022 IV 9.52 195 0.018100 HOBSON AND LEIGH: PROBLEMS IN THE DETERMINATION OF CARBON [Analyst, VOl, 99 TABLE I EFFECT OF pH ON THE REPRODUCIBILITY OF CARBON DIOXIDE ABSORPTION End-point Carbon determinations - r A \ mV 205 195 185 175 165 157 A precision equivalent reauired the electrode and PH No. Mean, per cent. Standard deviation (u) 9.67 10 1*8239 9.51 10 1.820, 9-35 10 1.821, 9.02 10 1.803, 8-89 10 1.7859 9.19 10 1.811, to 2/3 pg of carbon at the detection circuit to respond 0.0021, 0-0025, 0.0022, 0.0058, 0.0083, 0.0048, end-point in 350 ml of electrolyte to 0.02 mV. Ideally, an electrode wiih immediate response to this order of change, and an electrolytic current generator capable of maintaining the end-point within this limit, are desirable.In practice, it is im- possible to maintain the end-point pH so exactly because of the rapid combustion and evolution of carbon dioxide, a finite mixing time in the cell and the limits of current available. Therefore, some deviation in pH is inevitable, and as it is impossible to construct an electrode with an immediate response, a time constant is always present in the titration. Early development work was carried out with a combined electrode with a robust cylindrical membrane, and porcelain-frit liquid junction to the salt bridge, which had zero output at pH 2-0. These features tended to give a rather slow response, and it was found that when high-carbon steels were analysed, with a consequent excursion of 20 to 30mV from the end-point, the electrode system required several minutes to recover equilibrium.Study of the characteristics of the electrode, by imposing a sudden large change of pH [Fig. 7 ( a ) ] , showed that the high-resistance cylindrical glass membrane was responsible for the initial 3 minutes of the time constant, but thereafter the salt bridge was responsible. The glass half of the combined electrode has an overshoot characteristic, which in the cell results in the recording by the electrode of a higher pH than that of the electrolyte and the electrolysis current is proportionately reduced, producing a very long tail to the titration cycle. This delay in completing the titration is progressively aggravated as the excursion from the end-point is increased [Fig.8 (a)]. -495 I 1 -490 jf - ' O l -95 0 5 i o 15 Tim e/m i nu tes 0 5 10 15 Time/minutes Fig. 7. Time constant of glass - calomel and salt bridge sections of combined electrode with reference to separate calomel electrode, when transferred from buffer solution of pH 7-0 to buffer solution of pH 9.0. (a), Robust cylindrical glass membrane, porcelain frit liquid junction : A, glass - calomel, E , at pH 2; B, differential; and C, salt bridge. ( b ) , Low-resistance spherical membrane, taper-sleeve liquid junction: D, glass - calomel, E, at pH 7 ; E, differential; and F, salt bridgeFebruary, 19741 IN STEEL BY A PRECISION COULOMETRIC METHOD 101 It was obvious from these investigations that an electrode with a sensitive glass mem- brane and salt-bridge junction with a faster response was desirable. From the many types available, an electrode with a low-resistance glass spherical membrane and a sleeve-type liquid junction to the salt bridge, with zero output at pH 7, appeared the optimum choice.In order to reduce the electrical interference from the coulometer and neighbouring apparatus, it was found necessary to connect the pH electrode by double-screened co-axial cable. Tests with an electrode specially made to this specification showed a remarkable improvement in response time, Figs. 7 (b) and 8 ( b ) , and the change in E, reduced the effect of temperature fluctuation. Time/minutes Timelminu tes Fig. 8. pH 2 electrode; and (b), E , a t pH 7 electrode One of the drawbacks of using a lower-resistance glass membrane is its apparent memory effect when subjected to larger changes of 2 to 3 pH units.This effect becomes apparent when the cell solution is replenished from stock solution at approximately pH 7, or if the apparatus has long periods of disuse and the pH of the electrolyte falls to 7-0. When the pH is increased to the operating pH end-point the electrode requires 1 to 2 hours to become stabilised. During the recovery period the apparatus shows a spuriously high background count. For example, after 24 hours of continuous use the background is equivalent to less than 0.0002 per cent. of carbon per minute, whereas with a newly filled cell, which has had the pH raised rapidly from 7 to 9.5, a background equivalent to between 0.0010 and 0-0020 per cent.of carbon per minute is typical. Thus it is desirable that the electrolyte should be maintained continuously at the desired end-point, even when carbon determinations are not being made. Long-term use of this type of electrode at this pH does not appear to have any detri- mental effect other than a very slow change in E,, but provided that the end-point is checked with a buffer periodically this change is unimportant. Instead of the conventional potassium chloride, 3 M sodium chloride solution is used in the electrode salt bridge in order to prevent the formation of comparatively insoluble potassium perchlorate at the liquid junction. The possibility of eliminating the salt bridge and calomel half-cell by use of a silver - silver chloride reference electrode is now being investigated.Approach of titration to end-point after excursions of increasing amplitude: (a), E, at THE COULOMETRIC CELL- Details of the coulometric cell are shown in Fig. 9. Many such cells have been described, but the anode and cathode compartments of the electrolysis cell, and the reference and glass halves of the pH electrode, are usually separate entities joined by salt bridges and one or more diffusion d i a p h r a g m ~ . ~ - ~ ~ ~ This arrangement necessitates an expensive and fragile con- struction and results in high resistance of the electrolysis cell with subsequent low electrolysis currents or exaggerated heating effects. These problems are avoided in our design by the102 HOBSON AND LEIGH: PROBLEMS I N THE DETERMINATION OF CARBON [Analyst, VOl.99 use of a concentric structure, which is largely self supporting, and by the extensive use of the O-ring seals commonly used in vacuum techniques. The diaphragm that separates the anode and cathode compartments of the electrolysis cell was the subject of considerable experimentation. In addition to glass frits, experimental generators with anionic or cationic resins, and double glass frits with an intermediate salt bridge were tried. However, all the variants except a single glass frit had to be rejected on account of high resistance. A grade 3 frit (pore size 20 to 30 pm) of 25 mm diameter was eventually chosen as the best compromise between the needs of conductivity and retention of barium carbonate in the anode compartment.A = Anode B = BaC03 C = Cathode E = pH electrode F = Fan G = Gas distributor H = Head with O-rings M = Motor S = Stirrer Fig. 9. Diagram of the 350-ml capacity coulometric cell (not shown-heater and tem- perature sensor) The platinum electrolysis electrodes are positioned close to the diaphragm so as to keep the resistance as low as possible, but a layer of barium carbonate must be maintained between the anode and the diaphragm in order to prevent back-diffusion of acid ions into the cathode compartment. The gas dispenser is placed close to the bottom of the cell in order to achieve a long immersion time of the carbon dioxide bubbles in the electrolyte.February, 19741 IN STEEL BY A PRECISION COULOMETRIC METHOD 103 Rapid recognition of any change in pH of the electrolyte is vital, as slow sensing of carbon dioxide absorption would allow a serious fall in pH before the electrolysis current could take effect, and slow reaction to the generation of hydroxyl ions would allow the cell to overshoot towards excessive alkalinity before current cut-off.For these reasons the positioning of the pH electrode is fairly critical and has been determined by experiment. Gas flow solenoid valve THE COULOMETER- pH sensing unit, comparator, and electrolysis drive zcnit-In the initial design the pH of the cell was sensed by a combined glass - calomel electrode. Its output was amplified by an operational amplifier of very high input impedance, and was compared in a second amplifier with a standard voltage pre-set to a millivolt potential corresponding with the desired end- point pH. The signal from the comparator was fed via a transformer to the power circuit, which supplied current to the coulometer cell.For some time, satisfactory operation was achieved with this design, but eventually a breakdown in insulation occurred between the input and output circuits, causing first interference and then serious damage to the pH electrode. Because of the high impedance involved, this circuit was subject also to a.c. pick-up. The current design illustrated in Fig. 10 involves the use of the combined glass - calomel pH electrode described above, the outputs of the two halves being fed to a matched pair of operational amplifiers, with the generator cathode electrode as the common mode. This arrangement completely isolates the electrode from the output circuit and reduces the effect of a.c.pick-up. The differential output from the amplifiers is compared in the following stage with a high-stability standard voltage pre-set to a millivolt potential corrresponding to the desired end-point pH. A . T Electrode or comparator voltmeter r 1 I L t Glass pH and reference electrodes . 2 Logar it hm ic Comparator am pi if ier Differential - amplifier I (combined) L Fig. 10. Block diagram of the coulometer circuits 11 Ai The third stage synchronously rectifies and amplifies the signal from the comparator and drives a power transistor supplying current to the coulometer cell. The current can reach 1.8 A when the pH of the cell deviates from the desired value.Integrator-The potential difference developed by the electrolytic current across a high- stability resistor in the cell circuit drives an integrator circuit. In the earlier model, an 1 power meter Integration resistor electrodes . = Calibration - detector and heater L Digital Integrator display resistor . ii l i Heater 1 a::iYK!r , Power supplies Weston Timer and standard crystal osci I la tor cel I104 HOBSON AND LEIGH: PROBLEMS IN THE DETERMINATION OF CARBON [Analyst, Vol. 99 analogue integrator was used based on an operational amplifier, the output of which gave a direct reading of the percentage of carbon on the 2-V range of a 0 to 20.00-digital voltmeter. For experimental purposes, determination of the blank, or for higher sensitivity at low carbon contents, the voltmeter could be switched to a more sensitive range and measurements made to the nearest 0-0001 per cent.of carbon. Coulometric standardisation was facilitated by measurement, with the same voltmeter, of the potential developed across a precision resistor carrying a standardising current. Pre- cision of standardisation and results was dependent on the sensitivity and accuracy of the digital voltmeter; in the later stages of development, the 0 to 1999 output proved to have too little resolution. Another fault in this design was a stress effect in the integrating capacitor, which caused a “memory” error when low results followed high results. The present design avoids both of these problems by using integration by means of a capacitor in which the current produces a positive charge, which is removed by a series of negative pulses of known amplitude and duration.The pulses are counted and the total is shown on a five-digit display electronically calibrated to read directly the percentage of carbon, based on a 0.2-g sample. Standardisation-The coulometric standardisation is checked by integration of the standardising current for an accurately measured time of 200 s derived from a crystal- controlled oscillator ; the standardising current is continuously checked by comparing the voltage it develops across a calibrated resistor with the potential of a permanently incor- porated Weston cadmium cell at a known temperature. The calibration resistor has a temperature coefficient of 0.001 per cent.per “C and an accuracy guaranteed to be better than 0.01 per cent. Consideration of the possible variables, for example, ambient temperature and heat dissipation, gives an estimated stability of better than 1 part in 5000, and the accuracy of calibration is dependent solely on the long-term stabilities of the resistor and Weston cell, together expected to be much better than 0.1 per cent. The calibration count is recorded on the digital display, and should have the following values calculated from the known e.m.f. and measured temperature of the Weston cell within the electronic unit : Weston cell r 1 Tempera- ture/”C V 15 1.018 79 20 1.018 61 25 1.018 39 30 1.018 12 35 1.017 81 40 1.017 46 Standard current/mA 101-879 101.861 101,839 101.812 101.781 101.746 Equivalent carbon - C mg per cent.20.3758 1.268 01 0.6340 20.3722 1.267 99 0.6339 20-3678 1.267 51 0,6338 0.6336 20.3624 1-267 18 20.3562 1.266 80 0.6334 20-3492 1.266 36 0.6332 In practice, the result has been stable within 1 part in 3000 over a period of several weeks. Any fault or drift in the integrator would become known from the display of an incorrect value when checking standardisation. Auxiliary circuits-An auxiliary circuit from the comparator is used to operate a solenoid valve situated in the combustion gas line from the furnace, so as to cut off the flow of com- bustion gases to the cell if the pH of the electrolyte deviates by more than 3 mV from the end-point. This precaution is required in order to prevent large potential changes in the pH electrode; much larger excursions can be tolerated without loss of efficiency of absorption of carbon dioxide.The gas flow is reconnected when the pH of the electrolyte has recovered to the pre-set level. A second auxiliary circuit from the comparator is used to indicate on a meter the amount of deviation from the end-point during titration. Facilities are provided in the comparator to re-set the end-point by using a buffer solution should the mV - pH relationship of the pH electrode change, or when a new electrode is fitted. A second meter displays the instan- taneous coulometric current on a logarithmic scale that covers the range 100 pA to 1.0 A. COMBUSTION CONDITIONS- The furnace is a conventional unit heated by three resistor rods and operated at 1320 “C. The temperature is controlled automatically by a thyristor device in the circuit of one rod.February, 19741 IN STEEL BY A PRECISION COULOMETRIC METHOD 105 The combustion tube is chosen to be of 15.5 mm i d .so as to take specially made low-blank, small refractory boats with the minimum of dead-space. Oxygen at 5 p.s.i. from a cylinder is passed through a pre-combustion tube and a soda- lime tower before being supplied to the combustion tube, near the open end, at the rate of 1-5 1 min-1. We have encountered trace amounts of hydrocarbons in a cylinder of oxygen on one occasion, and pre-combustion is incorporated as a precaution against possible errors from this source. Some of the oxygen and all of the carbon dioxide are withdrawn at the closed end of the furnace by a small diaphragm pump, and are delivered at the rate of 350 ml min-1 to the coulometer cell via the solenoid valve mentioned previously, after the oxides of sulphur have been removed with manganese dioxide.When this gas circuit is closed, oxygen, at the rate of 100 ml min-l, reaches the cell via a by-pass line. A third tube in the furnace is used to pre-burn boats at 1320 "C in an unpurified oxygen flow of 500 ml min-1. To ensure low and reproducible blanks, it is essential that boats be cooled in oxygen and removed only to receive the steel sample immediately prior to the carbon determination. The train, shown diagrammatically in Fig. 11, is constructed as far as possible from 8 inch bore metal pipe in order to rninimise dead-space. Great care is needed in the elimina- tion of small leaks.Ingress into the suction side of the system of as little as 4 ml of air containing 0-03 per cent. of carbon dioxide would be equivalent to 0.6 pg of carbon, which is the tolerable error of the coulometer. For this reason much reconstruction of the seals on such items as the pump, flow gauges and needle valves was required. As the method is equally sensitive to the gaseous products of sulphur combustion it is important that the removal of sulphur gases be extremely efficient. Oxygen a t 5 p.s.i.g. cel I Fig. 11. Diagram of furnace train and oxygen flow system (flow-rates, ml min-l) OTHER FACTORS AFFECTING THE PRECISION OF THE METHOD- Change in pH of the cell electrolyte by loss of water-In a cell containing 350 ml of elec- trolyte at 37 "C, assuming a dry gas flow of 400 ml min-1 for a period of 6 minutes, the loss of water in saturating the gas would be 0.12 ml.This loss would produce a maximum change in pH of 0.0002 unit, equivalent t o 0.0001 per cent. of carbon on a 0.2-g sample. From the chemistry of the cell, the titration of the carbon dioxide by electrolysis results in a further loss of water: 2 x 96 495 C = 12.01 g of carbon = 18 g of water; and 4 mg of carbon (2.00 per cent. on a 0.2-g sample) = 6 mg of water per determination, i.e., 0.006 ml, which is one twentieth of the loss by evaporation and therefore negligible. Back-difusion in the electrolysis cell-The pH of the solution in the anode compartment depends on the solubility of barium carbonate in an acidic solution buffered by barium perchlorate and cannot be higher than about 7-0.Diffusion of this solution into the cathode106 HOBSON AND LEIGH: PROBLEMS IN THE DETERMINATION OF CARBON [Analyst, Vol. 99 compartment would result in a reduction of the alkalinity. It is therefore convenient, although not essential, to pass a few microamperes continuously through the coulometric cell; this action serves to keep the cell polarised, produces a smoother response of the coulo- meter integration system near the end-point and slightly decreases the background blank. Life of celZ electrolyte-The condition of the electrolyte in the cathode compartment does not change during many weeks, apart from very small losses by evaporation, but the effective- ness of the propan-2-01 is limited. Increasing the amount of propan-2-01 has no advantage and addition of 1 ml in 350 ml of electrolyte at intervals of 2 to 3 days appears satisfactory.However, it is usual to change the electrolyte when the amount of precipitated barium car- bonate becomes excessive, perhaps once per week, but this depends upon the type and number of steels analysed, and satisfactory results have been obtained over a period of a month with regular additions of propan-2-01. In the anode compartment it is necessary only to maintain a $-inch thick layer of barium carbonate in order to prevent back-diffusion of acid ions. Blank-Investigation into the use of fluxes to aid combustion was carried out. While low-alloy steels gave comparable results with or without flux, some alloy steels required the use of flux in order to ensure complete recovery of carbon.It was decided to standardise combustion by using 0-3g of tin, which contributed the equivalent of 0.003 per cent. of carbon to the blank. If the apparatus is left continuously self-balancing on standby, the residual chemical and electrochemical blank can be kept below the equivalent of 0.4 pg of carbon per minute, i.e., about 0.001 per cent. on 0 4 g of steel titrated in 6 minutes. The total reagent and electrochemical blank for a 6-minute determination, with proper care in preparation and handling of boats, selection of flux, etc., is highly reproducible, and is usually equivalent to 0.004 to 0.007 per cent. of carbon; long-term studies have given a between-day standard deviation equivalent to 0.001, per cent.of carbon. RESULTS OBTAINED, FUTURE DEVELOPMENTS AND COMMENTS The B.I.S.R.A. Carbon Study Group mentioned in the introduction tested methods that are in use in twenty laboratories and depend on the following seven different principles: non-aqueous titration ; thermal conductivity; low-pressure gasometry ; electrical conductivity ; infrared absorption ; coulometry ; and gravimetry. Five laboratories used the coulometric method to produce one result on six separate days from each of a series of eight steels. The original report contains a statistical survey carried out to derive within- and between-laboratory standard deviations for each steel and method principle. To test our coulometric method, we analysed samples of the same steels, and additional samples, once per day on ten different days, and calculated the means and standard deviations of the results.Results from the improved system are compared in Table I1 with those obtainedll by the Carbon Study Group. It can be seen that our results are in every instance more reproducible than the typical results derived from the work of the Study Group, which had already demonstrated that coulometry and low-pressure gasometry were the most reproducible method principles of the seven principles tested. 2 The reproducibilities are also much better than those given by Thomich8 for a range of steels analysed in quadruplicate in six German laboratories using the coulometric method for arbitrational analysis issued in 1971.' On the other hand, it must be admitted that we have not yet reached our objective of attaining a long-term standard deviation of 0.001 per cent.of carbon at the 1 per cent. level, and further development work is in progress. This work is mainly devoted to reducing elec- trical noise and minute drifts in the high-impedance electronic circuits, to improving the electrode system and to speeding up the determination in order to minimise the effects of background noise and apparatus blanks. The present operational cycle of 5 to 6 minutes makes the apparatus too slow for routine bath-sample analysis, but it is ideal for referee work and standardisation. It has already proved to be an extremely useful apparatus for the investigation of problems involving carbon heterogeneity in ingots and billets, in studies of case carburisation, and in measuring de- carburisation caused in heat treatment.Its accuracy depends solely on electrochemical equivalents, the precision and stability of a high-grade calibrated resistor and a Weston cell.February, 19741 IN STEEL BY A PRECISION COULOMETRIC METHOD 107 Hence, it gives a completely independent check upon results obtained by the other traditional methods mentioned earlier. TABLE I1 CARBON RESULTS BY COULOMETRIC TITRATION Carbon, per cent. Amount of Steel No. Type samplelg B.C.S. 260/3 H.P. iron 1.0 U.S.C. 11 - 1.0 B.C.S. 265/2 Mild steel 1.0 B.C.S. 333 18 Cr 8 Ni 1.0 B.C.S. 237/1 EN 32 1.0 B.S.C. (U.S.A.) (Pins) 1.0 DH 1 EN 8 0.2 B.C.S. 240/2 EN 8 0.2 DH 2 EN 42 0.2 B.C.S. 220/1 7W 5 Mo 0.2 DH 3 EN 44 0.2 B.C.S. 16311 EN 44 0.2 DH 4 EN 44E 0.2 B.C.S.247/4 White iron 0.1 B.C.S. 247/4 White iron 0.2 B.C.S. 247/3 White iron 0.1 Standardising count, measured a t 29 to 34 “C Certificate or accepted value 0*001, - 0.048 0.066 0.105 0 . 4 6 0.385 0.665 0.93- 1-005 1.2 1- 1.26- 3.05 3.05 3.0 0.63355 a t 30 “C 0.41- B.I.S.R.A. group r-in Mean standard (six results) deviation laboratory (2) (4 - - 0.0227 0.0005 0.0491 0.0006 0.0673 0.0008 0.1066 0*0009 - - - - 0.4136 0.0032 0.9271 0.0064 1.216 0.0055 - - - - - - - - 3.012 0.0146 Observations corrected to 30 “C D.M. coulomet& Mean (ten results) 0.0017 0.0228, 0.0493, 0.0653, 0.1062 0.4396 0.3789 0.4106 0.6659 0.9277 1.006, 1.209, 1,269, 3.051 3.053 0.6336, (2) - Standard deviation 0~0001 ( 0) 0~0002, 0~0002, 0.0002, 0-0005 0.0009 0.0014 0.0012 0.0013 0.0015 0.001 1 0.0016 0.0027 0-0034 0.0041 - 0~0001, The measuring circuits can readily be re-set to deal with other coulometric titrations, for example of sulphur dioxide or ammonia, and applications to the determination of sulphur, oxygen and nitrogen in steel are under consideration. The authors are grateful to Mr. G. F. Smith, Technical Director, Dunford Hadfields Ltd., for permission to publish this paper. They also express their thanks to their colleagues, Mr. T. W. Lomas and Mr. M. Swift, who have been responsible for the design and construction of several versions of the electronic sections of the apparatus described. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES Boniface, H. J., and Jenkins, R. H., Analyst, 1971, 96, 37. Scholes, P. H., “Proceedings of the Twenty-Third B.I.S.R.A. Chemists’ Conference,’, 1970, p. 37. Abresch, K., and Claassen, I., translated by Leveson, L. L., “Coulometric Analysis,” 1961, Chapman Metalluygia, 1969, 79, 265. Buchel, E., “Internationales Jahrbuch Chemische Industrie,” Vogt-Schild, Solothum, 1962-63. ‘‘Handbuch fur das Eisenhutten Laboratorium,” Band 4, Schiedanalysen, Verein Deutscher Eisen- “Handbuch fur das Eisenhiitten Laboratorium,” Band 5 (Supplement), Verein Deutscher Eisen- Thomich, W., Arch. EisenhiittWes., 1972, 43, 239. Abresch, K., and Lemm, H., Arch. EisenhiittWes., 1959, 30, 1. Craig, D. N., and Hoffman, J. I., “Proceedings of the National Bureau of Standards Semicentenial Symposium on Electrochemical Standards,” September 19th-21stJ 1951, p. 13. Bagshawe, B., and Scholes, P. H., Report CAC/122/73 of the Joint B.S.C./B.I.S.R.A. Chemical Analysis Committee, Determination of Carbon in Steel Study Group; Greenfield, A., and Scholes, P. H., Study Group Report MG/DA/400/70. Received August 20th, 1973 Accepted October 8t12, 1973 and Hall, London, 1965, pp. 154, 177 and 178. huttenleute, Dusseldorf, 1955, p. 73. hiittenleute, Diisseldorf, 1971.

ISSN:0003-2654    

DOI:10.1039/AN9749900093

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

108 Amdyst, February, 1974, Vol. 99, PP. 108-113 A Highly Selective Direct Colorimetric Procedure for the Determination of Zirconium in Steel with Arsenazo I11 by Using a Pressure-digestion Technique for Sample Dissolution BY A. ASHTON, A.G. FOGG AND D. THORBURN BURNS (Department of Chemistry, Loughborough University of Technology, Loughborough, Leicestershire, LE 11 3 T U ) Zirconium in steel is determined colorimetrically with arsenazo I11 in 7.5 M perchloric acid solution. Chemically inert forms of zirconium, which remain unchanged after conventional treatment with acid, are dissolved by means of a pressure-digestion technique. Results obtained with six British Chemical Standard steel samples show excellent precision and accuracy. Interference studies with fifteen metals commonly found in steel indicate that the procedure is applicable to a wide range of steels.Concentrations of zirconium in excess of 0.001 per cent. nz/m in steel samples can be determined. ARSENAZO I11 is an excellent colorimetric reagent for the determination of zirc0nium.l The arsenazo I11 - zirconium complex is formed at very high acid concentrations (e.g., 9 M hydro- chloric acid), at which concentrations most other metal complexes of arsenazo I11 are dis- sociated; hafnium(IV), thorium(IV), titanium(IV), lanthanum(II1) and uranium(V1) are the only metals reported to interfere.l The sensitivity is high, the zirconium complex having a molar absorptivity of 1.2 x lO51mol-1cm-l under the conditions used by Dedkov, Ryabchikov and Savvin.1 A major difficulty in determining zirconium colorimetrically is caused by the extensive formation of unreactive polynuclear hydrolysis products by zirconium( IV) .2 Pakalns3 has reported in detail on the depolymerisation of zirconium solutions prior to determining the zirconium with arsenazo 111.Zirconyl chloride octahydrate solutions (1000 pg ml-l of Zr) in 1 M perchloric acid, for example, were shown to react completely with arsenazo 111, after being allowed to depolymerise at room temperature for 24 hours before use. Minimum concentrations of strong acids that are required in order to depolymerise zirconium solutions of several concentrations at their boiling-point and to maintain the zirconium in a monomeric condition, after cooling the solutions, are given. The results given for perchloric acid solutions have been used in this work in preparing standard zirconium solutions and have been found to give excellent results.In a second paper, Pakalns4 reported the determination of zirconium in steel with arsenazo 111. His investigation of direct acid-dissolution procedures confirmed earlier work by CechovA5 that separation of zirconium (with cupferron) was necessary when even small amounts of other metals that give precipitates due to hydrolysis were present; silica and other precipitates due to hydrolysis co-precipitate zirconium. In the present work, a PTFE-lined steel pressure vessel was used to complete the acid dissolution of steel samples. A direct colorimetric finish with arsenazo I11 was then applied without prior separation of zirconium.Excellent results were obtained with this procedure, whereas low recoveries of zirconium were always obtained when dissolution was effected a t atmospheric pressure alone. PTFE-lined pressure vessels have been used previously for the determination of trace elements in siliceous materials6-8 and foodstuff^.^,^^ They have also been used in determinations of nitrogen11 and total aluminium12 in steel. EXPERIMENTAL If full advantage is to be taken of the fact that arsenazo I11 is a highly sensitive and selective reagent, then the colorimetric determination must be carried out in strongly acidic @ SAC and t h e authors.ASHTON, FOGG AND THORBURN BURNS 109 media. As high concentrations of ~ulphate,~ phosphate3 and 11itratel3.l~ interfere, the corre- sponding strong acids cannot be used.Kammori and co-workers15 found that interference from nitrate could be overcome by the addition of urea; absorbance measurements were made in nitric acid solution, for which they claim that the absorbance is less dependent on the acidity of the solution. In the present work, this procedure was found to be unsatisfactory: the mean of six determinations at the 0.4pgml-1 level of zirconium gave an absorbance of 0.449 at 665 nm with a standard deviation of 0.039. The choice of strong acid was therefore limited to hydrochloric and perchloric acids. With hydrochloric acid maximum absorbance occurs at 9 M concentration and with perchlori c acid at 7.5 M concentration (see Fig. 1). Close to this maximum the acid concentration was slightly less critical with perchloric acid. Furthermore, 9 M hydrochloric acid is more con- centrated than its azeotrope, which makes it difficult to use.For these reasons, perchloric acid was used as the reagent of choice. B 1 6 Acid concentrationh Fig. 1. Effect of hydrochloric and perchloric acid concentrations on the absorbance of the zirconium - arsenazo I11 complex : A, hydrochloric acid; and B, perchloric acid. Zirconium con- centration 0.6 pg ml-l ; and arsenazo I11 concentra- tion 0.002 per cent. m/V The procedure for dissolution of steels by using a PTFE-lined pressure vessel is given in detail below. The results given in Table I indicate the necessity for the use of a pressure vessel. The inclusion of hydrofluoric acid in the dissolution mixture was found to be necessary even when the pressure vessel was used, and nitric acid was also included in the mixture in order to decompose carbides.The efficient removal of hydrofluoric and nitric acids by evaporation to fumes prior to colorimetric determination was found to be very important. Even low concentrations of fluoride interfere with the reaction of zirconium with arsenazo 111. By optimising reagent concentrations and volumes, and by examination of colour development and stability times, the procedure outlined below was developed and is recom- mended for general use. Ascorbic acid was added to sample solutions at the colorimetric stage in order to mask iron(II1); it was not necessary to add it to standard solutions in the preparation of the calibration graph. The pressure-digestion vessel used in this work was obtained from S.and J. Juniper and Co., Harlow, Essex. REAGENTS- Aqua regia-Mix equal volumes of concentrated hydrochloric acid (analytical-reagent grade, sp. gr. 1-16 to 1.18) and concentrated nitric acid (analytical-reagent grade, sp. gr. 1.42). HydroJ7uoric acid, 40 per cent. m/V-Analytical-reagent grade. Perchloric acid, 7.5 M-Dilute 1300 ml of analytical-reagent grade concentrated per- chloric acid (70 per cent. m/V) to 2 litres in a calibrated flask. Ascorbic acid-Analytical-reagent grade.110 [Analyst, Vol. 99 Arsenazo 111 solution, 0-1 per cent. m/V-Dissolve 0.1 g of arsenazo I11 plus about 0.5 g of sodium hydroxide in about 50 ml of water. Add concentrated hydrochloric acid dropwise with stirring, until the colour of the solution just changes to red -violet, and dilute the solution to 100ml with water in a calibrated flask.Concentrated standard zirconium solution in 1 M perchloric acid, 100 pg ml-1 of Zr-Dissolve 0.177 g of zirconyl chloride octahydrate in about 200 ml of 1 M perchloric acid by boiling the mixture under reflux for 1 hour. Cool and dilute the solution to 500 ml with 1 M perchloric acid in a calibrated flask. Dilute standard xirconiuna solution in 7.5 M $erchloric acid, 10 pg ml-I of Zr-By pipette, transfer 10 ml of concentrated standard zirconium solution into a 100-ml calibrated flask, add 62 ml of 70 per cent. m/V perchloric acid solution and dilute to 100 ml with water. TABLE I RECOVERIES OF ZIRCONIUM FROM B.C.S. STEEL 271 BY USING DIFFERENT ASHTON et al.: DIRECT COLORIMETRIC DETERMINATION DISSOLUTION PROCEDURES Zirconium found, Coefficient of Recovery, t Dissolution procedure per cent. m/m variation,* per cent. per cent. Boiling 7.5 hi HC10, for 30 minutes , . .. 0-0033 7 7.5 Preliminary dissolution in HC1 and HF, fol- lowed by evaporation to fumes with 7.5 M HC10, so as to drive off excess of HF . . 0.0395 6 88 Recommended procedure with pressure diges- tion vessel . . .. .. .. .. 0-044 2 100 * Six determinations. t Based on results given in Table 111. PREPARATION OF CALIBRATION GRAPH- By pipette, introduce aliquots of the dilute standard zirconium solution (0 to 3.0 ml) into dry 50-ml calibrated flasks. To each flask add, by pipette, 1 ml of arsenazo I11 solution and dilute the mixture to 50 ml with 7.5 M perchloric acid.Measure the absorbance of the solution at 664 nm within 15 minutes of preparation against water, using l-cm cells. Subtract the absorbance of the solution that contains no zirconium. PROCEDURE- Remove the PTFE liner from the pressure vessel and weigh into the liner an appropriate amount of steel (Note 1 and Table 11). Add 20 ml of aqua regia and allow the steel sample to dissolve without heating. When evolution of hydrogen has ceased, add 5 ml of 40 per cent. m/V hydrofluoric acid solution, place and seal the liner in the pressure vessel (Note 2 and instruction manual where appropriate) and leave the pressure vessel in an oven a t 200 "C overnight (Note 3). TABLE I1 SUGGESTED SAMPLE SIZES AND ALIQUOTS FOR USE WITH VARIOUS STEELS Expected zirconium content, per cent.mlm 0.001 to 0.006 0.006 to 0.01 0.01 to 0-02 0-02 to 0.04 0-04 to 0.05 0.05 to 0-10 0.10 to 0.20 Amount of steel to be takenlg Aliquot takenlml 0.5 25 0.5 15 0-5 10 0.5 5 0.1 15 0.1 10 0.1 5 Cool the pressure vessel to room temperature and open it carefully. Transfer the contents quantitatively into a suitable PTFE vessel with 25ml of 7-5 M perchloric acid. Reduce the volume to less than 10 ml by heating the vessel on a hot-plate (Note 4), add 25 ml of 7.5 M perchloric acid and warm the mixture, if necessary, to dissolve any crystalline material (Note 5). Transfer the solution into a 50-ml calibrated flask and dilute it to 50ml with 7.5 M perchloric acid. With a pipette, transfer an appropriate aliquot of the solution (see Table 11) into a second 50-ml calibrated flask, add 1 g of ascorbic acid, dilute to about 30 ml with 7.5 M perchloricFebruary, 19741 OF ZIRCONIUM IN STEEL WITH ARSENAZO I11 111 acid and swirl the flask to dissolve the ascorbic acid without heating.Continue the deter- mination as described under Preparation of calibration graph beginning at “add, by pipette, 1 ml of arsenazo I11 solution. . .” For steels containing large amounts of metals that give coloured ions, determine the absorbance at 664nm of an aliquot of steel solution in the absence of arsenazo I11 and deduct this absorbance together with the arsenazo I11 reagent blank from the absorbance value obtained with the sample. Although in many instances the zirconium content of the steel will not be known even approximately, the sample sizes and aliquots recommended in Table I1 will give an indication of a suitable amount of sample to take as a trial.The sample size and aliquots recommended for each range of zirconium content give absorbance values between 0.3 and 0.6, except for steel samples containing less than 0.004 per cent. wzlm of zirconium. NOTES- 1. Caution-Perchloric acid must NOT be included in the digestion mixture in the sealed pressure vessel. The liner should be perfectly clean and dry, and care should be taken to ensure that no traces of perchloric acid from the previous determination remain in the liner. 2. Corrosion products tend to form a t the surfaces between the metal and the liner. These surfaces should be cleaned regularly in order to avoid difficulties arising in removing the liner after digestion.3. For some steels this digestion period can be reduced. Complete recovery of zirconium was made from British Chemical Standards Steels Nos. 271, 272, 274 and 275 after only 2 hours’ digestion. 4. This evaporation stage to remove hydrogen fluoride and oxides of nitrogen, although time consuming, is very important. Trace amounts of fluoride, in particular, interfere with the reaction of zirconium with arsenazo 111. 5. With steels that have a high chromium content, a red chromium(V1) compound may crystallise out on evaporating the solution to 10 ml. This precipitate dissolves more readily if about 1 g of ascorbic acid is added after the 7.5 M perchloric acid, which treatment should be carried out without warming as ascorbic acid is rapidly decomposed by warm perchloric acid solutions.Results obtained with the procedure described for the preparation of the calibration graph with the dilute standard zirconium solution had a coefficient of variation of 1-5 per cent. for ten determinations at the 0.4 pg ml-l of zirconium level. The calibration graph was linear in the range 0.1 to 0.6 pg ml-l of zirconium, and its slope corresponded to a molar absorptivity for the complex of 1-32 x lo5 1 mol-1 cm-1 at 664 nm. Results obtained with six British Chemical Standards steel samples by using the recommended procedure are given in Table 111. These results show good reproducibility and compare well with those obtained by other methods. TABLE I11 DETERMINATION OF ZIRCONIUM IN BRITISH CHEMICAL STANDARD STEELS Zirconium content, per cent.mlm B.C.S. values* r I Steel No. Chemical Spectrographic 27 1 0.04, 0.04, 272 ‘0.03, 0.03, 274 0.00, 0~01, 275 0.01, 0.02, 276 0.00, 0.00, 277 0.04, 0.05, Xylenolt orange procedure 0.044 0.031 0.012 0.02 1 0.008 0.051 X-ray$ fluorescence 0.043 0.029 0.010 0.019 0.006 0.050 Recommended procedures 0.044, 0.045, 0.043 0.031, 0.033, 0.033 0.010, 0.011, 0.011 0.020, 0.020, 0.021 0.006, 0-006, 0.005 0.050, 0.048, 0.047 * Zirconium is a non-standardised element in these steels. t Results of Keller and Hennesen.16 $ Results of Klima and Scholes.17 5 Results obtained from three different dissolutions of the steel sample. INTERFERENCES- The effects of fourteen other metals on the determination of zirconium are shown in Table IV.The metal to zirconium ratios chosen for study are well above those normally112 ASHTON et al. : DIRECT COLORIMETRIC DETERMINATION [Analyst, Vol. 99 found in steels. Of these metals only titanium interferes and then only when present above about a thirty-fold ratio of titanium to zirconium. Metal Antimony .. Arsenic . . Manganese . . Copper . . Titanium . . Aluminium . . Tin . . .. Lead . . .. Beryllium . . Molybdenum Chromium .. Vanadium .. Cobalt .. Nickel .. TABLE IV INTERFERENCE RESULTS Amount of zirconium added 20.0 pg .. .. .. . * .. .. .. .. .. .. . . . . . . . . . . . . .. .. .. .. .. .. .. .. . . . . Metal to zirconium ratio 10 100 10 000 10 000 30 50 100 300 1000 100 100 1000 1000 10 000 1000 10 000 3000 Zirconium found/pg 19-3, 19-2 19-6, 19.5 20.0, 20.2 20.7, 20.9* 19.3, 19.3 18.8, 19.0 17-5, 17.9 11.3, 9.3 19-8, 19.2 19.8, 19.0 20.3.19.5 20.5, 19.5 19.3, 19.3 18.8, 19.3* 19.5, 19-3* 19.3, 19.0 19.3, 19*0* * These metals produced ions in the dissolution process, which absorbed a t 664 nm. A correction was made for this effect (see text). Tungsten, which has not been included in Table IV, is the only metal that interferes seriously. A colloidal precipitate of yellow tungstic acid forms when the digestion mixture is evaporated to 1 O m l . This precipitate cannot readily be filtered or centrifuged, and it is probable that zirconium is co-precipitated. Consequently, the procedure in its present form cannot be applied to high-tungsten steels. Of the steels studied here, B.C.S. 276 and 277 contained the largest amount of tungsten (0.20 and 0.12 per cent.w/m, respectively), and this amount of tungsten did not interfere in the determination. DISCUSSION The procedure given above is recommended as it gives highly satisfactory results. No prior separation of zirconium is required owing to the use of a PTFE-lined pressure-digestion vessel that eliminates most interferences. Pakalns4 has suggested that the low results obtained with conventional acid-dissolution procedures are caused by co-precipitation of zirconium with silica and other precipitates that are due to hydrolysis. In the present work, the necessity for an intense pressure treatment of the steel sample with a digestion mixture containing hydrofluoric acid in order to release the zirconium has been demonstrated.It seems more probable to the present authors that the intense treatment is required so as to dissolve refractory zirconium compounds present in the steel rather than to avoid co- precipitation. We thank B.1.S.R.A.-the Corporate Laboratories of the British Steel Corporation for financial support. Mr. P. H. Scholes. In particular, we gratefully acknowledge the experienced guidance of REFERENCES 1. Dedkov, Yu. M., Ryabchikov, D. I., and Savvin, S. B., Zh. Analit. Khim., 1965, 20, 574. 2. Blumenthal, W. B., Talanta, 1968, 15, 877. 3. Pakalns, P., Analytica Chim. Acta, 1969, 44, 73. 4. - , Ibid., 1971, 57, 51. 5. CechovQ, D., Chemist Analyst, 1967, 56, 94. 6. Bernas, B., Analyt. Chem., 1968, 40, 1682. 7. Schnetzler, C. C., and Nava, D. F., Earth Planetary Sci. Lett., 1971, 11, 345. 8. Nava, D. F., Walter, L. S., and Doan, A. S., jun., J . Geophys. Res., 1971, 76, 4067.February, 19741 OF ZIRCONIUM IN STEEL WITH ARSENAZO 111 113 9. 10. 11. 12. 13. 14. 15. 16. 17. Nelson, G., and Smith, D. C., Proc. SOC. Analyt. Chem., 1972, 9, 168. Holack, W., Krinitz, B., and Williams, J. C., J . Ass. OH. Analyt. Chem., 1972, 55, 741. Menis, O., N.B.S. Technical Note 454, National Bureau of Standards, Washington, D.C., 1968. Headridge, J. B., and Sowerbutts, A., Analyst, 1973, 98, 57. Sawin, S. B., Talanta, 1961, 8, 673. Savvin, S. B., Kadaner, D. S., and Ryabova, A. S., Zh. Analit. Khim., 1964, 19, 561. Kammori, O., Taguchi. I., and Komiya, R., Bunseki Kagaku, 1965, 14, 106. Keller, H., and Hennesen, K., Awh. EisenhiittWes, 1968, 39, 921. Klima, Z., and Scholes, P. H., Analyst, 1973, 98, 351. Received August Gth, 1973 Accepted August 23rd, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900108

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

114 Analyst, February, 1974, Vol. 99, p p . 114-119 Determination of an Isomeric Impurity in Samples of Morantel Tartrate by Gas - Liquid Chromatographic Analysis of the Products of a Controlled Degradation Procedure RY E. ADDISON, E. DAVISON AND P. F. WADSWORTH (P’zer Central Research, Pjzer Ltd., Sandwich, Kent) A procedure is described for the determination of the tartrate salt of the 4-methylthienyl isomer of morantel, which is present in samples of morantel tartrate. The procedure involves a controlled degradative oxidation of the ethylene bond that converts morantel and the 4-methylthienyl isomer of morantel into 3-methylthiophene-2-carbaldehyde and 4-methylthiophene- 2-carbaldehyde, respectively. Extraction of the two aldehydes is followed by the determination of their relative concentrations by a gas - liquid chro- matographic procedure. It is shown that the ratio of the aldehydes given by gas - liquid chromatography is not significantly different from the ratio of morantel to the 4-methylthienyl isomer of morantel present in prepared mixtures of the two isomers, The method is shown to have good precision and to be applicable to morantel samples containing 0.10 to 7-00 per cent.m/vn of the 4-niethylthienyl impurity. MORANTEL tartrate is a broad-spectrum anthelmintic agent used for the treatment of a wide variety of infestations in domestic animals. The chemical structure of the compound is given below (I), COOH i I I H-C-OH HO-C-H COOH The chemical name of morantel tartrate is 1,4,5,6-tetrahydro-l-methyl-2-[trans-2-(3-methyl- 2-thieny1)vinyllpyrimidine hydrogen tartrate.The synthesis of morantel involves the condensation of 3-methylthiophene-2-carbalde- hyde with 1,4,5,6-tetrahydro-1,2-dimethylpyrimidine in the presence of methyl formate, as shown in 11. @ SAC and the authors.ADDISON, DAVISON AND WADSWORTH 115 The 3-methylthiophene-2-carbaldehyde used in this step always contains an appreciable proportion (approximately 11 per cent. m/m) of the isomeric impurity 4-methylthiophene- 2-carbaldehyde. It is not possible to remove the impurity at this stage and hence, during this reaction sequence, it reacts with 1,4,5,6-tetrahydro-1,2-dimethylpyrimidine according to the scheme shown (111) to give 1,4,5,6-tetrahydro-l-methyl-2- [trans-2-(4-methyl-2-thienyl)- vinyllpyrimidine, which is referred to as the 4-methylthienyl isomer.Crystallisation of crude morantel tartrate reduces the level of the 4-methylthienyl isomer considerably, but it is necessary quantitatively to monitor the amount of this isomeric impurity present in the finished product. The method described in this paper will determine the 4-methylthienyl isomer down to a level of 0.10 per cent. m/m. Morantel and the 4-methylthienyl isomer of morantel have very similar physical and chemical properties, governed largely by the polar, highly basic, tetrahydropyrimidine ring. In order to accentuate the difference in molecular structure for the purposes of chromato- graphic separation, a degradation procedure has been devised that oxidises the ethylenic bond1 connecting the thiophene and tetrahydropyrimidine ring systems.This selective oxidation of morantel and its 4-methylthienyl isomer produces corresponding amounts of 3-methyl- thiophene-2-carbaldehyde and 4-methylthiophene-2-carbaldehyde. The use of alkaline permanganate solution in conjunction with rapid extraction of the carbaldehydes into toluene avoids over-oxidation of the ethylenic bonds to the corresponding carboxylic acids. The ratio of the two carbaldehydes produced is determined by a gas-liquid chromato- graphic procedure. It is essential to verify the absence of these two carbaldehydes in samples before carrying out the determination. EXPERIMENTAL AND RESULTS REAGENTS AND MATERIALS- Anhydrous sodium carbonate-AnalaR grade. Potassium permanganate, 1-00 per cent. m/V aqueous solution-Dissolve AnalaR grade potassium permanganate (1.00 0.01 g) in distilled water and make the volume up to 100 ml.Potassium perrnanganate solutions are not stable indefinitely; the solution should therefore be stored in a dark bottle and replaced 3 days after preparation. TolNevze-AnalaR grade. Gas - liquid chromatographic column packing-Gas-Chrom Q (100 to 120 mesh), coated with 5 per cent. m/m polyethylene glycol 400, is a suitable packing and can be prepared in the laboratory or purchased from Perkin-Elmer Ltd., Beaconsfield, Bucks. APPARATUS- Se$arating funnel, 50-ml capacity. Safety pipettes, 2-ml and 10-ml capacity. Phase-separatingJilter-papers, Type 1 PS, of 9.0 cm diameter (Whatman Ltd.). Gas - liquid chromatograph eq%ti$ped with a jlame-ionisation detector.Glass column for gas - liquid chromatography-The column should be 3.0 m long, with an internal diameter of 1-5 mm. The configuration of the column is dependent on the model of gas chromatograph used. For the Varian, Model 204, gas chromatograph used by us, the column has a helical configuration. Potcntiometric recorder with a range of 2.5 mV.116 ADDISION et al. : DETERMINATION OF AN ISOMERIC IMPURITY [Analyst, Vol. 99 Syringe for gas - liquid chromatography, capable of injecting 1.0 p l liquid samples. Graticule, 25 mm with divisions every 0.1 mm-This was a Type T1 from Graticules Ltd., 18-20 Garrick Street, London, W.C.2. Cl-iitUMAIOGltAl?HlC LONUlllOHS- The following conditions are used: glass column, 3.0 m by 1.5 mm (the column used in this laboratory had an outside diameter of Q inch); packing, Gas-Chrom Q (100 to 120 mesh) coated with 5 per cent.m/m polyethylene glycol 400; column temperature, 105 "C isothermal (see NOTE) ; injection temperature, 150 "C; detector temperature, 150 "C; carrier gas, nitrogen at a flow-rate of 30 ml min-l, measured at room temperature (see NOTE) ; and sample size, Under these conditions, the two components of interest had the following retention times from injection : 4-methylthiophene-2-carbaldehyde, 29.0 minutes ; and 3-methylthio- phene-Zcarbaldehyde, 31.8 minutes (separation factor = 1.10). In order to facilitate accurate measurement of peak areas, a column is required that has approximately 4000 theoretical plates.2 The work described here was carried out on a column with 5000 theoretical plates, measured by using the 3-methylthiophene-2-carb- aldehyde peak.1.0 p1. NOTE-The temperature of the column and flow-rate of the carrier gas may require adjustment when different chromatographs and different column configurations are used. PROCEDURE- Add the morantel tartrate sample (0.10 & 0.01 g) to 10 ml of distilled water. Ensure that dissolution is complete and then transfer the solution to a 50-ml separating funnel. Add 1.Og of sodium carbonate to the separating funnel and shake the funnel until the carbonate is dissolved. Add 10 ml of the potassium permanganate solution to the separating funnel with a pipette over a period of 1 minute while swirling the contents and allow the mixture to stand for 2 minutes before adding toluene (2.0 m1) to the funnel, again by means of a pipette. Extract the aldehydes into the toluene by shaking the separating funnel for 2 minutes, then discard the aqueous phase and transfer the organic phase to a small vial by filtering it through a phase-separating filter-paper.Inject 1.0 pl of the toluene solution B Fig. 1. Gas - liquid chromato- gram showing the separation of (4- methylthiophene-2-carbaldehyde (A) from 3-methylthiophene-2-carbalde- hyde (B) . Chromatographic conditions are given in the textFebruary, 19741 IN MORANTEL TARTRATE BY GAS - LIQUID CHROMATOGRAPHY 117 on to the gas-liquid chromatographic column and record the chromatographic peaks due to the 4-methylthiophene-2-carbaldehyde and 3-methylthiophene-2-carbaldehyde at suitable attenuation settings.Measure the peak heights to the nearest 0.5 mm and peak widths to the nearest 0.1 mm by using a ruler and the graticule, respectively. Calculate the area of each peak by multiplying the peak height by the width at half-height. A typical chromato- gram is shown in Fig. 1. CALCULATION- Let A represent the area of the 4-methylthiophene-2-carbaldehyde peak, B the area of the 3-methylthiophene-2-carbaldehyde peak, a the attenuation of the 4-methylthio- phene-2-carbaldehyde peak and b the attenuation of the 3-methylthiophene-2-carbaldehyde peak. Then, the percentage m/m of the 4-methylthienyl isomer in the sample of morantel tartrate is Aa/(Aa + Bb) x 100. For examples of this calculation, see Table I. TABLE I ANALYSES OF A SYNTHETIC BLEND OF MORANTEL TARTRATE AND THE TARTRATE Ten replicate analyses SALT OF THE 4-METHYLTHIENYL ISOMER Peak Peak height of Peak height of Peak Area of Area of 4-isomer*/ width at 3-isomer*/ width at 4-isomer* 3-isomer* Content of mm half-height of mm half-height of peak/mm2 peak/mma 4-isomer, (attenua- 4-isomer*/ (attenua- 3-isomer*/ (attenua- (attenua- per cent.Run tion x 1) mm tion x 16) mm tion x 1) tion x 16) m/m A 69.5 5.0 178.0 5.0 347.5 890 2.38 B 82.0 4.8 214.0 5.0 393-6 1070 2-25 C 70.5 4.7 185.0 5.0 331.4 925 2.19 D 82.5 4.9 218.0 5.1 404.3 1112 2.22 E 78.5 4-8 215.0 5.2 376.8 1118 2.06 F 80.0 4.9 212.0 5.1 392.0 1081 2-22 G 86.0 4.8 227-0 5.1 412.8 1158 2.1 8 H 90.0 4.8 225.5 5.1 432.0 1150 2.29 70-5 4.9 189-0 5.1 345.5 964 2-19 62.0 4.7 159.5 5.0 291.4 798 2.23 J K 4-methylthiophene-2-carbaldehyde, respectively.variance = [S(,,)I2 = 0.0067; standard deviation = S(,,) = 0.082. * The abbreviations 3-isomer and 4-isomer refer t o 3-methylthiophene-2-carbaldehyde and Using the above results, standard statistical calculations give : mean value = 2.221 ; BLANK DETERMINATION- It is essential to verify the absence of the two carbaldehydes in samples analysed by the above method. This verification can be conveniently accomplished by use of a blank determination. Follow the described procedure with a second sample of morantel tartrate but omit the addition of potassium permanganate solution. The gas - liquid chromatographic examination of the toluene extract must show no evidence of either carbaldehyde. PRECISION- In order to evaluate the precision of the method, a solution was prepared that contained approximately 1.0 g of morantel tartrate and 0.02 g of the 4-methylthienyl isomer per 100 ml of distilled water, corresponding to an impurity level of approximately 2.0 per cent.m/m of the 4-methylthienyl isomer with respect to morantel. This level was chosen because of its relevance to quality control limits. Ten replicate analyses were carried out on 10-ml aliquots of the solution by following the given procedure. The results and calculations are given in Table I. The standard deviation calculated from the results, S(lo), is 0.082. Thus, the result of a single determination can be quoted as being &0.19 per cent. m/m with 95 per cent. confidence. In order to evaluate the error involved in the gas - liquid chromatographic determination of the aldehydes, a toluene solution containing 4-methylthiophene-2-carbaldehyde and118 ADDISON et al.: DETERMINATION OF AN ISOMERIC IMPURITY [Arzalyst, Vol. 99 3-methylthiophene-2-carbaldehyde was chromatographed seven times. The relative amounts of the two isomers were calculated by the method given under Prodecure and Calculation. The standard deviation for the seven replicates was calculated to give a value for S(7) of 0.057 per cent. m/m. Results are given in Table 11. TABLE I1 RESULTS OF GAS - LIQUID CHROMATOGRAPHY RUNS ON THE SAME TOLUENE SOLUTION OF 3-METHYLTHIOPHENE-2-CARBALDEHYDE AND 4-NETHYLTHIOPHENE-2-CARBALDEHYDE Seven replicate analyses Peak Peak height of Peak height of Peak Area of Area of 4-isomer*/ width a t 3-isomer*/ width at 4-isomer" 3-isomer* Content of mm half-height of mm half-height of peak/mma peak/mm2 4-isomer, (attenua- 4-isomer*/ (attenua- 3-isomer*/ (attenua- (attenua- per cent.Run tion x 1) mm tion x 16) mm tion x 1) tion x 16) mlm 1 73.0 4.8 180.5 5.1 350.4 92 1 2.32 2 7 4 4 4.9 177.5 5.1 362-6 905 2.44 3 66.5 4.7 162.5 5-1 312.6 829 2-30 4 73.0 4.7 180.0 5.0 343.1 900 2-33 5 69.5 4-8 173.5 5.1 333-6 885 2-29 6 70.0 4.7 174.0 5-0 329.0 870 2-30 7 79.0 4.7 188.5 5.0 371.3 943 2.40 * The abbreviations 3-isomer and 4-isomer refer to 3-methylthiophene-2-carbaldehyde and Using the above results, standard statistical calculations give : mean value = 2.341 ; 4-methylthiophene-2-carbaldehyde, respectively. [S(,)f2 = 0.0032; standard deviation S(,) = 0.057.Application of the F test shows no significant difference between the variance for the gas - liquid chromatographic analysis and the variance associated with the whole procedure. Therefore, if increased precision is required it is advisble to replicate the gas - liquid chromato- graphic analysis. As the standard deviation for the whole procedure, S(lo) is 0.082, the 95 per cent. confidence limit for the mean of n replicates is given by the term 0*19rt-+ at the 2 per cent. m/m level. ACCURACY- The accuracy of the procedure was evaluated by making standard additions of the tartrate salt of the 4-methylthienyl isomer to a sample of morantel tartrate, followed by analysis of the resulting mixtures. A standard solution of morantel tartrate was prepared by weighing 0.9761 g into a 100-ml calibrated flask, dissolving it in, and making the solution up to volume with, distilled water.This is referred to as solution A. A standard solution of the 4-methylthienyl isomer was prepared by weighing 0.1241 g into a 100-ml calibrated flask, again dissolving it in, and making the solution up to volume with, distilled water. Aliquots of solution A (10.0 ml) were transferred by pipette into 50-ml separating funnels together with various aliquots (0, 1-0,2.0, 3.0,4.0,5.0 and 6-0 ml) of solution 13. The resulting solutions thus contained known amounts of morantel and its 4-methylthienyl isomer. The solutions were then taken through the remainder of the procedure, commencing at "Add 1.0 g of sodium carbonate. . . ." The results are presented in Table 111.The graph of the percentage (m/m) of impurity added versus the percentage (m/m) of the impurity found is linear and has a slope close to unity, proving that the ratio of 4-methylthiophene-2-carb- aldehyde to 3-methylthiophene-2-carbaldehyde produced in the oxidation step is consistent with the ratio of the isomeric 4-methylthienyl impurity to morantel present in samples of morantel tartrate. The fact that the line diverges slightly from a slope of unity is indicative of a small bias in the method. The source of this bias is not known but it is small in com- parison with the precision of the method. Thus, a t a 4-methylthienyl isomer content of 2-00 This is referred to as solution B. The results show that the accuracy of the method is acceptable.February, 19741 IN MORANTEL TARTRATE BY GAS - LIQUID CHROMATOGRAPHY 119 per cent.m/m, the bias is 0.08 per cent. m/m, whereas the precision of the determination is 50.19 per cent. mlm with a 95 per cent. confidence limit. TABLE I11 ANALYSES OF SYNTHETIC BLENDS OF MORANTEL TARTRATE AND THE TARTRATE SALT OF THE 4-METHYLTHIENYL ISOMER OF MORANTEL Solu- tion Run A/ml 1 10 2 10 3 10 4 10 5 10 6 10 7 10 4-isomer” Solu- added, tion per cent. B/ml m/m 0 0 1.0 1-26 2-0 2.48 3.0 3.67 4.0 4.84 5.0 5.98 6.0 7.09 Peak height of 4-isomer* / mm (attenua- tion) 69.5 (A2) 132 (A2) 202.5 (A2) 122.5 (A4) 149 (A4) 182 (A4) Single analyses Peak Peak Peak width height of width Content at half- 3-isomer*/ at half- Area of Area of of height of mm height of 4-isomer* 3-isomer* 4-isomer*, 4-isomer*/ (attenua- 3-isomer*/ peak/ peak/ per cent.mm tion) mm mm2 mm2 m/m - - - Nil 60094 Not detected 4.5 163 (A64) 4.8 626 50074 1.23 4-3 167 (A64) 4.7 1125 47 226 2.33 4.3 159 (A64) 4.7 1742 47 827 3.51 4.3 143.5 (A64) 4.7 2107 43 166 4.65 4.3 141.5 (A64) 4.6 2563 41 658 5.80 4.3 142 (-464) 4.7 3130 42 714 6.83 * The abbreviations 3-isomer and 4-isomer refer to 3-methylthiophene-2-carbaldehyde and In addition, regression analysis of the results in Table I11 shows that the intercept is not significantly different from zero. In view of the peak heights given in this table, the demonstrated linearity of the results and the lack of any response when nil 4-methyl isomer is added, the stated lower limit for this method of 0.1 per cent. m/m is a conservative estimate. It is concluded that the method is adequate for the quantitative determination of 1,4,5,6- t e trahydro-1 -me thyl-2- [trans-2- (4-methyl-2-thienyl)vinyll pyrimidine hydrogen tartrate, an isomeric impurity present in samples of morantel tartrate. The method is rapid, easily carried out, makes use of readily available laboratory apparatus and can be used for quality control purposes. 4-methylthiophene-2-carbaldehyde, respectively. REFERENCES 1. Migrdichian, V., “Organic Synthesis,” Volume 2, Reinhold Publishing Corporation, New York, 1957, P. 896. 2. Keulemins, A. I. M., “Gas Chromatography,” Reinhold Publishing Corporation, New York, 1959. Received July 4th, 1973 Accepted September 3rd, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900114

出版商:RSC    

年代:1974

数据来源: RSC

摘要:

120 Analyst, February, 1974, Vol. 99, $9. 120-127 Determination of Thiabendazole in Citrus Fruits by Ultraviolet Spectrophotometry” BY ANNA RA JZMAN (Division of Fruit and Vegetable Storage, Institute f o r Technology and Storage of Agriculturai Products, Agricultural Research Organization, The Volcani Center, P.O. Box 6, Bet Dagan, Israel) A method for the ultraviolet spectrophotometric determination of thiabendazole in citrus fruit is described. It is extracted from the fruits by equilibrating the mashed peels or pulps with chloroform, the extract is acidi- fied and the mixture concentrated in order to eliminate the chloroform and filtered. After making the filtrate slightly alkaline the thiabendazole is re-extracted with chloroform and determined by measuring the absorption of the extract a t 302 to 303 nm.The procedure enables the satisfactory elimination of interfering substances to be achieved, and small amounts of thiabendazole of the order of 0.1 p.p.m. in the peel and the pulps and 0.03 p.p.m. in the whole fruit to be determined. The recovery of thiabendazole added to the peel or pulp of fruit varies between 94.1 and 103.0 per cent. THIABENDAZOLE [2- (4’-thiazolyl) benzimidazole, C,,H,N,S] , has been used for some time as a fungicide for the post-harvest treatment of citrus fruits so as to reduce the incidence of rot in stored fruits.1 There is a need for a routine inexpensive method for determining minute amounts of thiabendazole, on or in the citrus fruits (often less than 1 p.p.m. in the whole fruit). In some of the procedures previously described, it is extracted from citrus fruit with ethyl a ~ e t a t e ~ - ~ and, after elimination of the interfering substances, determined in hydrochloric acid solution by spectrofluorimetry2 or by ultraviolet ~pectrophotometry,~ in methanolic solution by spectrofl~orimetry~ and in ethyl acetate solution by gas chromatography3; in other pro- cedures, the thiabendazole is extracted froin the fruit with dichloromethane and determined either c~lorimetrically~ or in methanolic solution by ultraviolet spectrophotometry.6 Whichever method is used, the main problems in its determination in citrus fruits occur in its quantitative extraction from the fruit and in the elimination of interfering substances.Ethyl acetate is used for the extraction of thiabendazole in most methods because of the very slight solubility of this compound in most other organic solvent^.^ The extractives interfering in the determination are eliminated by subjecting the ethyl acetate extract to liquid - liquid extraction^.^^ Those which interfere in the fluorimetric determination are not always completely removed* and the liquid - liquid extraction is combined with the separation of thiabendazole by thin-layer chr~matography.~ The recovery of thiabendazole added to whole-fruit blends is about 85 per cent.4 or 90 per cent.3 and is lower than when it is added to surface-stripping solutions of intact fruit^.^ In the method described herein , thiabendazole is extracted from the fruit with chloroform and determined in the chloroform solution by ultraviolet spectrophotometry.The procedure described for its extraction and for the clean-up of interfering substances is based on the solubility of thiabendazole in various solvents and on its distribution behaviour between chloroform and various aqueous solutions. The procedure is simple, inexpensive and enables the determina- tion of thiabendazole in the fruits to be made with satisfactory accuracy. METHOD APPARATUS- All glassware must be scrupulously clean and all joints made of ground glass. Ultraviolet spectrophotometer-A Perkin-Elmer spectrophotometer, Model 137, and stop- Blenders-A Vir-Tis “45” homogeniser with a 500-ml stainless-steel container, and a pered rectangular silica cells, of 10 and 40-mm path length, were used. Waring-type blender with a 1000-ml jar, were used.* Contribution from the Agricultural Research Organization, The Volcani Center, Bet Dagan, Israel. @ SAC and the author. 1973 Series, No. 154-E.RAJZMAN 121 Concentration afiparatus-Flat-bottomed 350-ml flasks with short necks with ground-glass stoppers, a Vigreux reflux column of effective length 15 cm, a condenser and a connection tube. Separating funnels, 100 and 500-ml capacity. Microburette, 2-ml capacity, with 0.01-ml divisions. All reagents should be of recognised analytical grade. Thiabendaxole-As supplied by Merck, Sharp & Dohme. Chloroform. Sodium hydrogen carbonate solution, 5 per cent. m1V. Hydrochloric acid, 0.1 N. Sodium hydroxide solution, 1 N. Trisodium citrate dihydrate. Sodium sulphate, anhydrous. pH indicator papers-Whatman - BDH indicator papers, narrow range pH 4 to 6, 8 to 10, REAGENTS- and 10 to 12, were used. ADJUSTMENT OF pH- A 2-ml volume of 1~ sodium hydroxide solution added with a microburette to 20 ml of 0 .1 ~ hydrochloric acid gives a mixture with a pH between 10.5 and 11, which is determined with a pH meter. Addition of 0.5 ml of the 5 per cent. sodium hydrogen carbonate solution depresses the pH of the mixture to about 9.5 which is also determined with a pH meter. In the procedure described below the required pH is checked with pH indicator paper and this check can be achieved by placing on the paper a drop of the above solution before and after addition of the sodium hydrogen carbonate and noting the colour and the nature of the developed stains corresponding to pH 10.5 to 11 and 9.5, and making the necessary compari- sons.GENERAL PROCEDURE Thiabendazole, which is localised mainly in the peel of the fruit, is determined separately in the peel and in the pulp and calculated, as necessary, for the whole fruit. It can be deter- mined directly in the whole fruit by following the procedure described for the pulp. It is extracted from mashed peel or pulp by equilibration' of the latter with chloroform; the chloroform extract is concentrated and hydrochloric acid is added to the concentrate, which is then concentrated, filtered and the filtrate made slightly alkaline. The thiabendazole is extracted from this solution with chloroform and transferred into 0-1 N hydrochloric acid solution. After making the latter slightly alkaline, the thiabendazole is re-extracted with chloroform and determined by measuring the absorption of the extract at 302 to 303 nm.PREPARATION OF FRUIT EXTRACTS- Preparation of peel samples-Weigh an average sample of fruit (for example, ten fruits) and remove carefully the peel including the albedo 5 3 '*c; to avoid contaminating the pulp with trace amounts of thiabendazole from the peel. During peeling, prepare an average sample of peel by setting aside from each fruit an average sample of peel corresponding to about one third of the fruit surface. Chop the peel on a plate, mix it and take for analysis an accurately weighed amount not exceeding 100 g. Preparation of pulp samples-Weigh the pulps and calculate the total amount of peel. By taking an aliquot from each pulp prepare a mixed sample of pulp not exceeding 500 g.Weigh the sample accurately and place it in the 1000-ml jar of the Waring blender. For each 100 g of orange or grapefruit pulp add 7 g of trisodium citrate, and for each 100 g of lemon pulp 20 g of trisodium citrate. With a pH indicator paper, check the pH of the mixture, which should be between 4.5 and 5. Take for analysis an accurately weighed portion of the pulp mixture containing not more than 100 g of pulp (107 g of orange or grapefruit pulp mixture and 120 g of lemon pulp mixture correspond to 100 g of fruit pulp). Extraction of thiabendaxole-Place in the 500-ml stainless-steel container of the homogeniser the accurately weighed amount of chopped peel or pulp mixture. Add exactly 3 ml of Blend the mixture for 2 minutes.122 RAJZMAN : DETERMINATION OF THIABENDAZOLE IN [Analyst, Vol.99 chloroform per gram of peel or pulp. Locate the shaft with the cutting blades in the container. Weigh the latter with its contents, note the total amount and mince for 15 minutes. Mince the peels at high speed to an impalpable purde and the pulps at moderate speed. Do not remove the cutting blades from the container. Adjust, by adding chloroform, the mass of the container with its contents to that previously noted so as to compensate for the loss of chloroform that may have occurred during mincing. Avoid any change in the volume of the extract during the subsequent operations. Transfer the chloroform extract through a fine sieve into a 500-ml separating funnel and compress slightly the residue of peel extract on the sieve in order to obtain the maximum amount of extract.Allow the phases to separate. Filter the chloroform phase through an adequate, folded filter-paper (Whatman No. 2) containing about 25 g of anhydrous sodium sulphate. Collect the filtrate, which should be clear, in a ground-glass stoppered graduated cylinder. Note the volume of the filtrate (generally about 85 per cent. of the volume of chloroform used for the extraction), each 3 ml of which represents 1 g of peel or pulp taken for analysis. Transfer the filtrate quantitatively into a 350-ml flat-bottomed flask with a short neck, connect it to the condenser via the Vigreux column and concentrate the extract at atmospheric pressure to a volume of about 10 to 15 ml.Separation of thiabendazole from the bulk of extractive substances-Add to the extract 20 ml of water and exactly 5 ml of 0 . 1 ~ hydrochloric acid, and boil to eliminate completely the remaining chloroform. Detach the flask from the concentration apparatus and boil the acidic mixture while exposed, concentrating it to approximately 10 ml. Cool and transfer the mixture quantitatively into a 25-ml calibrated flask, rinsing the flask with distilled water, and make the volume up to the mark with distilled water. Mix the solution and filter care- fully through a folded filter-paper (Whatman No. 2). The filtrate is almost colourless and should be sufficiently clear to avoid the formation of an emulsion during subsequent extraction with chloroform. CEean-up of interfering substances-Place exactly 20 ml of the filtrate, which contains 4 ml of 0 .1 ~ hydrochloric acid, into a 100-ml separating funnel. Wash the filtrate several times (usually four or five times) with 1-ml portions of chloroform, shaking the separating funnel gently to avoid formation of emulsions, until two consecutive washes remain colourless. Decant the washes carefully into a separating funnel, extract them with exactly 5 ml of 0 . 1 ~ hydrochloric acid and discard them. Transfer the acidic phase quantitatively into the separ- ating funnel containing the filtrate, rinse the empty funnel with a few drops of distilled water and add the rinsings to the combined acidic solutions. The solution obtained contains 9 ml of 0 . 1 ~ hydrochloric acid and a certain amount of interfering substances that react with sodium hydroxide, some of which change on neutralisation from colourless to yellow.Make the solution alkaline to pH 10.5 to 11 by adding, from a microburette, 0.9 ml of 1~ sodium hydroxide solution, continuing the addition carefully until the colour of the solution turns yellow. Check the pH of the mixture with pH indicator paper and adjust it to between 10.5 and 11, as described under Adjustment of pH. Then add 0.5 ml of 5 per cent. sodium hydrogen carbonate solution and check the pH with the pH indicator paper; the final pH should be about 9.5. Extract the mixture four times with 5-ml portions of chloroform and decant the chloroform phases carefully into a 100-ml separating funnel, discarding the aqueous phase.Extract the combined chloroform phases with exactly 10, 10 and 5 ml of 0 . 1 ~ hydro- chloric acid. After the first two extractions, decant the chloroform phase into a clean separa- ting funnel and discard it after the third extraction. Add quantitatively the second and the third acidic extracts to the first extract, rinsing the empty funnel with a few drops of water. Wash the combined acidic extracts with three 1-ml portions of chloroform, collect the washes in a separating funnel and extract them with 5 ml of 0 . 1 ~ hydrochloric acid. Discard the chloroform phase and quantitatively transfer the acidic phase into the funnel containing the combined acidic extracts. With a microburette, add 3.0 ml of 1~ sodium hydroxide solution and check the pH of the mixture, which should be between approximately 10.5 and 11.Then add 0-5 ml of 5 per cent. sodium hydrogen carbonate solution and check the pH of the mixture, which should be about 9.5. Extract it with five 4 to 5-ml portions of chloroform, collect the chloroform phases in a 25-ml calibrated flask and make the volume up to the mark. Finally, add anhydrous sodium sulphate, mix, allow the mixture to stand and determine the thiabendazole content of the chloroform extract.February, 19741 CITRUS FRUITS BY ULTRAVIOLET SPECTROPHOTOMETRY 123 DETERMINATION OF THE THIABENDAZOLE CONTENT OF THE EXTRACT- A solution of thiabendazole in chloroform gives a characteristic ultraviolet absorption spectrum with a wide absorption band between 255 and 330 nm, with a maximum at 302 to 303 nm and two plateaux, one slight at 295 nm and the other distinct at 310 to 315 nm (Fig.1). Preparation of standard graphs-Prepare two standard graphs by using the 10 and 40-mm path length cells. Dissolve 20 mg of thiabendazole in a few millilitres of chloroform. and make the volume up to 100 ml. From this solution, prepare, by suitably diluting with chloroform, several solutions containing increasing amounts of thiabendazole ranging from 0.1 to 13 pg ml-1 and record their ultraviolet spectra against chloroform in the reference cell, by using the 40-rnm cells for solutions containing 0-1 to 3-2 pg ml-l and the 10-mm cells for solutions containing 0-5 to 13 pg d-l. Determine the absorbances at 302 to 303 nm and prepare the standard graphs, each of which is a straight line.The absorptivity, E$$, value, for thiabendazole in chloroform solution was found to be 1080. Determiwation of thiabendaxole-Record the spectrum of the chloroform extract against chloroform in the reference cell, by using the 40-mm cell for extracts containing less than 1 to 2 pg ml-1 of thiabendazole and the 10-mm cell for more concentrated extracts. If the absorbance at 302 to 303 nm in the 10-mm cell exceeds 1.4, dilute the extract adequately and note the degree of dilution. Measure the absorbances at 302 to 303 nm and determine from the standard graph the concentration of thiabendazole in the extract, correcting for dilution if necessary. p.p.m., where apg ml-1 is the Thiabendazole content in peel or pulp = concentration of this compound in the chloroform extract, and b ml is the volume of the chloroform filtrate taken.Calculate the thiabendazole content in the whole fruit if necessary. Calculate the content of the peel or pulp taken for the analysis as follows: a x 25 x 25 x 3 20 x b t V 200 250 300 Wave1 engt h/n m 350 390 Fig. 1. Ultraviolet spectra recorded in 10-mm path length cells: a, thiabendazole in chloroform solution; b, extract from fruit treated with thiabendazole; and c , extract from untreated fruit The ultraviolet spectrum for the treated fruits is shown in Fig. 1. If, accidentally, the extract still contains trace amounts of interfering substances, the absorbance at 330 nm may increase and exceed 0.1 and the slope of that portion of the spectrum between 330 and 390 nm increases.In this event purify the extract once more as follows: transfer into a 100-ml separating funnel an aliquot of the chloroform extract and proceed as described above for the combined chloroform phases, extracting the chloroform phase with exactly 10, 10 and 5 ml of 0 . 1 ~ hydrochloric acid. Continue as described and determine the thiabendazole content of the chloroform extract. SENSITIVITY- DISCUSSION The method permits the determination of small amounts of thiabendazole of the order124 UJZMAN : DETERMINATION OF THIABENDAZOLE IN [Analyst, VOl. 99 of 0.10 p.p.m. in 100 g of peel or pulp and of 0.03 p.p.m. in the whole fruit. The sensitivity can be increased by taking larger samples of peel or pulp. ACCURACY- Thiabendazole was determined in the peel and pulp (Table I) of untreated fruits to which known amounts of the compound had been added: 1 ml of a solution of thiabendazole in chloroform was dispersed on 100 g of chopped peels or pulps, placed in the container of the homogeniser, for 10 to 15 minutes before the addition of the required amount of the solvent to the peels and before the addition of the trisodium citrate to the pulps.The recovery of added thiabendazole varies between 94-1 and 103.0 per cent. (Table I). It was also determined in several samples originating from the same mixture of finely chopped peel of the treated fruit (Table 11). The results show that the deviations are slight. TABLE I RECOVERY OF THIABENDAZOLE ADDED TO PEELS AND PULP OF NON-TREATED FRUITS Thiabendazole added to 100 g of peel or pulp/ Peels- 0 20 40 100 250 1250 2500 Pulp- 0 20 50 Recovery, per cent. Shamouti oranges Grapefruit Lemons 0 96.0 100.6 101.1 101-5 94.4 99.2 96.0 97.0 94.1 99.2 0 102.0 98.5 94.1 99.6 0 103-0 98-2 98.5 100.2 96.0 99.0 97.3 99.9 - 0 95.6 102.0 - - 98.0 99.2 95.6 100.8 96.0 98-3 0 0 97.0 94-8 95.8 96.7 97.5 96.5 98.3 100.1 VALIDITY OF THE METHOD- SoZvent-The solubility of thiabendazole at 25 "C in various solvents was determined.For this purpose a saturated solution in a given solvent was filtered, the ultraviolet spectrum TABLE I1 100-g samples taken from the same mixture of finely chopped peel DETERMINATION OF THIABENDAZOLE I N PEELS OF TREATED FRUITS Thiabendazole Mixture Sample No. found, pg A 1 179 2 180 3 182 B 1 271 2 286 3 286 C 1 920 2 939 3 965 Average 180.3 Average 280.7 Average 941.3 Deviation from average value, per cent. - 0.72 -0.16 + 0.94 - 3.4 + 1.5 + 1.9 - 2.2 - 0.24 + 2.5February, 19743 CITRUS FRUITS BY ULTRAVIOLET SPECTROPHOTOMETRY TABLE I11 SOLUBILITY OF THIABENDAZOLE IN VARIOUS SOLVENTS AT 25 “c 125 Solvent Chloroform ..Ethyl acetate . . Hexane . . .. Cyclohexane . . Hydrochloric acid/N 0.1 0.05 0.02 0.01 0.001 0*0001 Sodium hydroxide solution/N 1 0.1 0.05 0.01 Citric acid solution, per cent. 5.0 2-5 1.0 Solubility of thiabendazolelmg ml-1 .. 3-2 .. 1.8 .. 0.001 .. 0.00 1 20.5 7.9 3.2 3.0 1.0 0.05 2.0 0.25 0.12 0.05 6.3 3.6 2-8 of the filtrate, adequately diluted if necessary, was recorded against this solvent in the reference cell and the maximum absorbance of the characteristic peak noted. The concentration of this compound in the filtrate was calculated according to the absorbance of a known amount dissolved in the respective solvent.It was found that thiabendazole was more soluble in chloroform than in ethyl acetate (Table 111). It has been reported3y4 that during the extraction of thiabendazole with ethyl acetate the fruit pulp or whole-fruit blend forms emulsions with the solvent, which have to be broken by centrifugation or some product added in order to reduce3 or prevent their formation. It is to be noted that no emulsion formation was observed with the use of the Vir-Tis “45” homogeniser for mincing the pulp or the whole fruit with chloroform. Extraction of thiabendaxole f r o m citrus peels and pulps-Citrus fruits contain considerable amounts of organic acids (chiefly citric acid) and their salts, which are mainly localised in fruit juices.The total acidity, expressed in grams of citric acid per 100 ml of citrus juice, corresponding to free and combined organic acids, is about 1 to 2 g for orange and grapefruit juice and 5 to 7 g for lemon juice. Citrus fruit peels contain a comparatively small amount of organic acids. The pH of orange and grapefruit juice is about 3, of lemon juice about 2, and of citrus peel about 5. In order to establish the conditions for the quantitative extraction of thiabendazole from citrus fruit, the solubility of this compound in aqueous solutions of citric acid was determined as described above (Table 111), and its distribution behaviour between chloroform and various aqueous solutions of pure citric acid, and of mixtures of citric acid and trisodium citrate, was studied (Table IV).The distribution behaviour was expressed as the fraction of total solute which distributes itself in the non-polar phase for equal volumes of solvent.8 The fraction found in the chloroform phase was calculated as a percentage of the total amount of thiabendazole. The fraction found in the chloroform phase depends upon the pH of the aqueous solution (Table IV). When the pH is between 4.5 and 5, about 98 per cent. of the thiabendazole partitions into the chloroform, thus enabling almost all of it to be extracted by equilibrating one volume of the aqueous phase with three volumes of chloroform, as described in the procedure. At least 3.06 g of trisodium citrate dihydrate per 1 g of citric acid (2 mol of trisodium citrate per 1 mol of citric acid) should be added in order to increase the pH of the solution to between 4.5 and 5, which justifies the addition of trisodium citrate to the pulp of the fruit.Owing to the relatively high pH of citrus peel, the addition of trisodium citrate to the peel is unnecessary. The presence of an excess of trisodium citrate does not affect the distribution behaviour of thiabendazole.126 RAJZMAN : DETERMINATION OF THIABENDAZOLE IN [Analyst, Vol. 99 Se@aration of thiabendaxole from the bulk of other extractives-This compound is soluble in hydrochloric acid solutions (Table 111). Complete elimination of chloroform from the acidic mixture is necessary in order to precipitate the extractive substances, which are insoluble in hydrochloric acid.The acidic mixture should be heated to dissolve the thiabendazole, which is eventually incorporated in the insoluble bulk of other extractives, and concentrating the mixture contributes to the elimination of a large proportion of the steam-volatile extractives, TABLE IV DISTRIBUTIONS OF THIABENDAZOLE BETWEEN CHLOROFORM AND VARIOUS AQUEOUS SOLUTIONS FOR EQUAL VOLUMES OF SOLVENTS 200 pg dissolved in 20 ml of chloroform and solution shaken for 10 minutes with 20 ml of aqueous solution; pH of aqueous solutions determined by pH meter before partition with chloroform Aqueous solution Citric acid, per cent. m/V 5 2 1 0.5 0.1 0.01 0.005 0.002 0.001 Citric acid, 1 per cent. m/V (a) + trisodium citrate dihydratelmg per 100 ml (b) 0 191 765 1531 3062 3828 Hydrochloric acid Sodium hydroxide PH 1.8 2.05 2.2 2.4 2.8 3-4 3.7 4.0 4.4 2.2 2-8 3.8 4.4 4.5 5.1 1.0 2.0 3.0 4.0 5.0 6.6 9.4 1.07 11.8 12.7 14.0 Molar ratio bla 0: 1 0.12 : 1 0.5 : 1 1.0 : 1 2.0: 1 2.5: 1 Thiabendazole, per cent.of total, found in chloroform phase 11 19 26 36 60 86 94 98 99 26 53 89 97 99-100 99-100 3 19 65 93 98-99 98-99 98-99 98 95 84 48 Clean-up of the remaining interfering substances-In order to establish the conditions that pennit the separation of thiabendazole from the remaining interfering substances, its solubility in hydrochloric acid and sodium hydroxide solutions was determined (Table 111), and its distribution behaviour between chloroform and the various aqueous solutions was studied (Table IV).The results show that under the conditions of the procedure described, it is possible to extract it almost quantitatively from the chloroform phase with 0 . 1 ~ hydrochloric acid, and from the aqueous phase, at pH 5 to 10, with chloroform. It was found that in order to eliminate most of the interfering substances, the aqueous phase must be made slightly alkaline, to a pH of 10.5 to 11. The subsequent addition of the sodium hydrogen carbonate buffer solution decreases the pH to about 9.5. The ultraviolet spectrum for untreated fruit (Fig. 1) shows that the interfering substances are practically eliminated.February, 19741 CITRUS FRUITS BY ULTRAVIOLET SPECTROPHOTOMETRY 127 The low recoveries of thiabendazole39* added to the fruit may have been caused in part by the extraction and clean-up procedures used by these authors. Its partition behaviour between various aqueous solutions on the one hand and ethyl acetate or chloroform on the other was found to be similar for the two solvents. Biphenyl and 2-phenylphenol, currently used for post-harvest treatment of citrus fruit, did not interfere in the determination, as these products are eliminated during the various stages of the procedure. On the other hand, methyl l-(butylcarbamoy1)-2-benzimidazole carbamate (benomyl), an experimental fungicide that may replace thiabendazole, and its degradation product methyl 2-benzimidazole carbamate, interfere in the determination by the method described. However, it is unlikely that thiabendazole and benomyl would be used simultaneously for citrus fruit treatment. This work was carried out with the technical assistance of Mrs. Hana Heller and Mrs. Sofia Abramowicz. 1. 2. 3. 4. 5. 6. 7. 8. REFERENCES Eckert, J. W., Wld Rev. Pest Control, 1969. 8, 116. “Pesticide Analytical Manual,” Section 2, Food and Drug Administration, Washington, D.C., 1969, Mestres, R., Campo, M., and Tourte, J., Trav. SOG. Pharm. Montpellier, 1970, 30, 193. Norman, S. M., Fouse, D. C., and Craft, C. C., J . Ass. 08. Analyt. Chern., 1972, 55, 1239. Kroller, E., Dt. LebensmittRdsch., 1971, 67, 229. Hey, H., 2. Lebensmittelunters, u.-Forsch., 1972, 149, 79. Gunther, F. A., Residue Rev., 1969, 28, 83. Beroza, M., Inscoe, M. N., and Bowman, M. C., Ibid., 1969, 30, 1. pp. 120 to 242. Received April llth, 1973 Accepted August 3rd, 1973

ISSN:0003-2654    

DOI:10.1039/AN9749900120

出版商:RSC    

年代:1974

数据来源: RSC