摘要:

JANUARY 1976 The Analyst Vol. 101 No. 1198 Sample Preparation in the Micro-determination of Organic Compounds in Plasma or Urine A Review" Eric Reid Wolfson Bioanalytical Centre, University of Surrey, Guildford, G U2 6XH Summary of Contents Introduction Ionisability and other chemical features Risk of adsorptive losses Initial treatment of plasma samples Hydrolysis of conjugates General considerations Extraction apparatus and ease of phase separation Development of solvent methods Alumina and charcoal Ion-exchange resins Amberlite XAD-2 resin Other agents Generalisations and examples The advent of high-pressure liquid chromatography (HPLC) Methods of handling conjugates Solvent extraction Separation with a solid additive or column material Choice of approach Trends in methods of sample preparation Concluding comments Introduction When trace amounts of therapeutic substances or of endogenous compounds of comparatively low relative molecular mass have to be determined, possibly as metabolites, an isolative procedure of variable complexity usually has to be applied to the sample initially, whether it is of plasma or urine.As listed for drugs1 the concentrations of the organic substances may vary from 10 to less than 10-1 mg 1-l. The portion of sample taken is typically 1 ml for plasma samples and 10 ml for urine, although the latter may be richer in the substance(s) to be assayed. Removal of adventitious fluorogens may be necessary with plasma (e.g., reference 2) and especially with urine (e.g., reference 3). In general, the sample has to be partially purified.These generalisations do not apply to radioimmunoassay, which is not to be considered here, although it is notable for its high sensitivity, if not assured ~pecificity,~ and is of growing importance in drug a ~ s a y . ~ ~ ~ In fact, a sample preparation step may be desirable even for radioimmunoassay (e.g., reference 6) and other affinity (saturation analysis) methods (e.g., references 7 and 8). General guidance on ways of preparing samples for assay is scarce in the literature and the rationale for particular methods is often unstated. A classical paper by Brodie et aL9 on the assay of basic substances repays study, as do articles by Titus1* and Trevor et a2.l1 and a book on separation science.12 The present survey of approaches is prompted by our analytical involvement in bioavailability trials and in method-development studies on metadrenalines (metanephrines) , i.e., 3-0-methyl derivatives of catecholamines.Some examples that illustrate principles have been taken from the latter field as well as the literature on drug assay. For details see summaries in advertisement pages. 1 * Reprints of this paper will be available shortly.2 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, VoZ. 101 Toxicological literature, concerned with the screening of urine or blood samples for drugs of abuse, is largely unhelpful, partly because the drug concentrations are often high.1 The emphasis has often been more on rapid detection than on exact determination, and indeed many papers lack results on percentage recovery.Moreover, a multi-purpose pre- paration procedure [e.g. , repeated extraction with a mixture of acetone and diethyl ether (1 + 1)13] has commonly been favoured, although with awareness of its possible inefficiency for some of the drugs that might be present. Ionisability and Other Chemical Features Drugs and other organic compounds that are determined in body fluids may be non- ionisable (e.g., glutethimide and sulphonylureas), anionic except below a certain pH value (e.g., barbiturates and aspirin) and cationic except when dissociated at alkaline pH (e.g., morphine and catecholamines). Alternatively, the classes can be termed acidic, basic and inherently neutral. The term acidic usually connotes acidity greater than that of phenolic hydroxyl groups.Amphoteric compounds comprjse a special class, having a charged group virtually irrespective of the pH. In contrast with ion-exchange chromatography, solvent extraction is best performed with compounds in un-ionised form, taking account clf differences in polarity that govern the choice of solvent (see Development of Solvent Methods, below). Whatever the approach, the object of the analytical exercise will be frustrated if interfering compounds are carried through to the final step, or if autoxidation is allowed to occur. For phenolic compounds, which in alkaline solution are especially prone to autoxidation, 2-mercaptoethanol is a favoured antioxidant although it can jeopardise the use of fluorimetry. Risk of Adsorptive Losses The following remarks are concerned not with possible chromatographic losses due to incomplete elution, as referred to below (in the section on Separation with a Solid Additive or Column Material), but with possible adsorption of the sample on to cuvettes or other vessels in micro-determinations. This phenomenon, perhaps better known in the context of very dilute solutions of purified proteins, can lead to erratically low assay values (e.g., reference 16).While the presence of a detergent helped to obviate this trouble in the instance of tetra- hydrocannabinol,6 a more common preventive measure is to silanise the glassware or to apply a very thin coat of PTFE.17 Such adsorption, as recognised by Brodie et can occur both with aqueous media, particularly at neutral or alkaline pH values for organic bases, and with organic solvents, especially if of weak polarity.The adsorptive loss they observed with chloroquine at a concentration of 70 pg 1-1 in heptane had risen to 25 per cent. after 30min, but was nil if ethanol had been added (to 1.3 per cent. V/V). In extracting quinacrine with light petroleum,18 they used isoamyl alcohol (mainly, if not entirely, 3-methylbutan-l-o1*) to minimise adsorption. In sampling from heptane extracts, Dill et aZ.19 used pre-wetted pipettes. Material lost on to extraction tubes may re-appear in subsequent assays if the tubes are re-used, as was observed by Spirtes20 with chlorpromazine; silanisation did not prevent this problem, but treatment with acidic dichromate and ammonia solution did do so. In scintillation counting studies the adsorption of benzoic acid was found to be small, at least with present-day vials, com- pared with that of dicarboxylic acids.21 With oestrogens in aqueous buffers there were losses due not to adsorption but to creepage; these could be minimised by silanisation.22 Use of a scavenger may also help.17 Inadvertent adsorption is not a universal occurrence.Initial Treatment of Plasma Samples When blood rather than urine is the starting material there are special problems, which will only be touched on here. With serum, U.S.-manufactured devices to aid separation from the clot may give rise to drug-like analytical artifact^.^^^^^ Compounds that are present in samples of plasma (or serum, which is virtually synonymous in the present context) normally have t o be obtained free from protein, yet without loss of any protein-bound moiety.Dis- * In addition to this isomer, the “isoamyl alcohol” used by many investigators may contain 16-30 per cent. V / V of 2-methylbutan-1-01 as stated in some catalogues; however, some suppliers claim high purity.January, 1976 OF ORGANIC COMPOUNDS I N PLASMA OR URINE 3 sociation of the latter may be favoured by diluting the plasma. For this and other reasons, five-fold dilution of the plasma at the outset is advantageous.ll Overnight ultrafiltration, as in an assay method for acetaminophen (paracetamol) ,2K is a possible way of deproteinising the plasma, although it would not itself encourage disso- ciation of a bound drug. A more common way is to use an acidic protein precipitant, e.g., tungstate followed by sulphuric acid.3 Some worked4 use perchloric acid, added to give a concentration of 5 per cent.m/V; this acid has the advantage that the perchlorate anion can, if desired, be removed subsequently as its insoluble potassium salt. Trichloroacetic acid is sometimes used (e.g., reference 9), but it may contain impurities that can give a high assay blank,14 and it may itself be carried through the assay if the next step is extraction with a solvent such as diethyl ether, in which it is soluble. Nevertheless, it was an effective first step in a gas - liquid chromatographic assay2G (with an electron-capture detector) for biguanides in plasma at levels as low as 1 pg 1-1. Alternatively, the compound can be extracted from the plasma with an organic solvent, after acidification if it is basic and strongly protein-bound.g Indeed, classical protein pre- cipitation procedures are seldom obligatory.They may even be undesirable if the compound is relatively soluble in water (a possible difficulty with heavy-metal reagents being co- precipitation of the test compound with the protein) ; with use of a suitable pH and, if neces- sary, with salt addition to the aqueous phase, an effective water-immiscible organic solvent can usually be found, even for coumarin anticoagulants or other drugs that are extensively protein-bound in plasma (A. Bye and R. H. Nimmo-Smith, personal communications). The extraction of serum, half-saturated with sodium dihydrogen orthophosphate, with hexane worked well for methaq~alone.~' If 1,2-dichloroethane is used, notwithstanding the risk of a high assay blank,ll it may be advantageous carefully to layer the plasma on to the solvent, so as to obviate gel formation after mixing.9 In choosing a suitable pH for the particular compound that is the subject of the assay, it should be re- membered that an acidic pH will encourage precipitation of protein. This is particularly desirable when a direct assay is to be carried out on the organic phase, e.g., an isoamyl acetate extract in the colorimetric determination of sulphonylureas2* and a chloroform extract in the gas - liquid chromatographic determination (after evaporating down the extract) of anti-epileptic dr~gs.~9 Chloroform extraction at an acidic pH value is effective even for phenytoin, which is strongly protein-b~und.~~ Two examples of water-miscible solvents, which, when present in excess, precipitate the proteins and hopefully liberate the bound compound, are acetone and ethanol.',ll These solvents have been used in a d m i ~ t u r e .~ ~ The choice of a water-immiscible solvent is considered later. Hydrolysis of Conjugates With urine, and even with plasma (e.g., reference 25), there can arise the difficulty of the pre- sence of conjugates, particularly glucuronides and sulphates, that should not be neglected in the assay. Approaches that could obviate hydrolysis are considered later (under Methods of Handling Conjugates). Commonly, however, conjugates are first hydrolysed by enzymic means or with hot acid (say 0.1 M hydro- chloric acid for 30 min at 100 "C).Acid hydrolysis may give rise to high blanks14 and poor thin-layer chromatographic patterns,32 and has to be rigorous for some sulphates and even for some glucuronides30 (again see Methods of Handling Conjugates). The enzymic approach, possibly with a glucuronidase - sulphatase preparation such as glusulase, is less likely to cause degradation of the compound to be determined or of concomitants, although enzymic treatment can have disadvantages in addition to expense and slowness. Thus, catechol- amine assays15 may be prejudiced by the presence of dopa and dopamine as contaminants in glusulase preparation^.^^ Moreover, mucoprotein carried through from the enzyme prepara- tion may cause emulsions or column clogging unless it is removed by an agent such as tungstic acid, and the urine may have to be freed initially from enzyme-inhibiting a n i ~ n s .l * , ~ ~ Conjugates are not amenable to solvent extraction. General Considerations The processing steps subsequent to any initial splitting-off of bound or conjugated sample compound are the main theme of this review. The detailed processing is governed partly4 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, VoZ. 101 by the nature of the final analytical step, which is chosen from a range that has been well surveyed by Clifford and in the context of benzodiazepines. When the final step is not a discriminative analysis, such as differential spectrophotometry, or a separatory approach, such as gas - liquid chr~matography,~~ the processing may have to be complex and selective.If, however, a drug to be administered in vivo is isotopically labelled, and it is by radioactivity that body-fluid levels of the drug and/or of known metabolites are to be measured, the processing or the final analysis can be less rigorous. Radioisotopically labelled compounds used in uitro rather than in vivo are valuable in developing and applying assay methods. When they are added initially in trace amounts to plasma or urine samples, they allow the over-all recovery of the endogenous material in the assay to be established merely by liquid scintillation counting (e.g., reference 32). However, their use can have the disadvantage of necessitating a costly synthesis if they are not available off the shelf.Moreover, if a sample preparation procedure has been devised through the use of radioactivity measurements, and then “real” test specimens (unlabelled) are put through the procedure, the processed sample may prove to be too impure to undergo the classical assay method for determining the endogenous compound. Traditionally, extra test samples spiked with unlabelled authentic material are put through the sample preparation steps for reference purposes. Separate standards not subjected to the full processing procedure then serve merely to check that parts of the method, for ex- ample a final fluorimetric measurement, are working satisfactorily. Processed aqueous standards, which are not actually added to the plasma or urine samples, are of limited value except for pilot work.They may even show a different response to concentration relation- ship, as observed in a fluorimetric assay for furosemide.16 Whatever the calibration procedure, losses as high as 50 per cent. can be tolerated if re- producibility is good. When the final step is chromatographic, the labour of running spiked samples alongside test samples can be obviated by adding to the actual test samples, at the outset, a reference compound that travels in a different position. For example, chloroform has been used for this purpose in the determination of plasma halothane by gas-liquid chr~matography.~~ Such a short-cut is invaluable if used warily. Some authors (e.g., ref- erence 26) have taken care to verify the assumed similarity and consistency in respect of processing losses.If the reference compound is merely added just before the chromatographic step its value lies mainly in checking the volume that was chromatographed. Solvent Extraction Even in the above assay procedure for halothane, in which plasma can be analysed directly, it is advantageous to perform a traditional solvent extraction with heptane or carbon tetra- chloride initially, as by this means the gas - liquid chromatographic column lasts longer and the injection port need not be cleaned so often.37 Ilowever, in most assays a sample prepara- tion procedure is essential rather than optional and solvent extraction is the traditional approach to preparation.= The optimisation of conditions for particular assays is discussed in the section on Develop- ment of Solvent Methods.General points to be taken into consideration include solvent purity, as found when using ethyl acetate39; inadvertent exposure to air and light rendered the solvent detrimental to the spectrophotometric determination of a catecholamine meta- bolite (VMA). Chlorinated solvents should be pre-washed with water if phosgene is detri- mental to the compound. It is well known that impure solvents may give rise to high blanks.11~1~ Adsorbents such as certain aluminas can constitute an effective means of purify- ing solvent~.~0~41 When specifying isoamyl alcohol, authors should state its composition, For plasma cortisol, chloroform is as satisfactory as dichloromethane and is cheaper.*% A more important factor is toxicity, which in the case of benzene is so high as to warrant its substitution with toluene.From the viewpoint of laboratory safety, diethyl ether freed from peroxides can be regarded as an acceptable risk. A possible alternative is diisopropyl ether, pre-washed successively with alkali (in order to remove any antioxidant), acid and water; any evaporation of extracts should be carried out in a stream of nitrogen near to room temperature (R. H. Nimmo-Smith, personal communication). Even so, the peroxide hazard is a serious one. The following remarks mainly concern the equipment needed for the task, especially for tedious multi-step extractions, and practical suggestions. In choosing a solvent, cost is a marginal consideration.January, 1976 OF ORGANIC COMPOUNDS I N PLASMA OR URINE 5 Extraction Apparatus and Ease of Phase Separation Traditional separating funnels can be a handicap rather than an asset, particularly when a large number of samples have to be put through successive extractions, each with a centri- fugation step.It suffices to extract the samples in stoppered tubes that will withstand centrifugation. If not completely full, they allow efficient extraction when clamped near to the horizontal position and shaken longitudinally ( e g . , reference 43) at, say, 120 oscillations per minute for 15 min. It should not be taken on trust that phase mixing (e.g., in an alkane - water system) will be as efficient with a tumbling or rolling device as with vigorous oscillation or even conscientious shaking by hand. Additionally, the possible influence of shaking time, as shown in the extraction of N-hydroxyphentermine (for gas - liquid chromatographic assay44) and, in our own laboratory, of indomethacin should not be disregarded.Neverthe- less Trevor et aZ.ll give a warning “concerning the time and vigor of extractions which in the reported literature is often far longer than necessary and appears to coincide strongly with the gustatorial habits of a particular laboratory.” With plasma, there are instances of the best yield being obtained with less than 1 min of shaking3s A roller or tilting device will work well for some solvents and entails a minimum of risk of persisting em~lsification.~~,45 Stirring rather than shaking is stipulated if blood is to be extracted with acetone - ether.13 Scattered information exists (e.g., reference 11) on the avoidance of emulsions that are intractable to centrifugation.Diethyl ether is less emulsion-prone than chloroform. A re~ornmendation~~ that the organic solvent to aqueous solvent phase ratio be 10: 1 cannot sen- sibly be applied in assay work. A ratio as low as 1 : 40 gave good extraction of drugs such as amphetamine from urine at an alkaline pH with a dense solvent such as chloroform, obviating the usual need to use a purified solvent and to concentrate the extra~t.~C The ratio was only 1 : 80 in drug-screening studies4’ with a chloroform - propan-2-01 mixture (4 + 1) at an alkaline pH, the urine sample first being saturated with a carbonate - hydrogen carbonate mixture added as a solid; centrifuge tubes with a sharply tapered tip facilitated removal of the heavy organic phase.The authors believed that the addition of salts helped to avoid the formation of emulsions, besides facilitating extraction into a volume of solvent so small that subsequent concentration (for gas - liquid chromatography) was unnecessary. Avoidance of emulsion troubles as an incidental benefit of using a salt -solvent pair has likewise been noted by other authors, e.g. , by Horning et aL4* who added a saturating amount of ammonium carbonate to urine or plasma samples. An example of salt addition in the steroid field (neutral steroids) is the addition of ammonium sulphate solution at pH 1, prior to the extraction with ethyl acetate.34 Extraction at an acidic pH has the attendant advantage of lessening the risk of emulsion formation. An example of practical information in the literature concerns the extraction of plasma with 1,2-di~hloroethane~~ ; the organic phase sometimes formed a gelatinous solid emulsion.If its formation was not prevented by the layering procedure outlined above under Initial Treatment of Plasma Samples, an effective remedy was to stir the contents of the tube vigorously with a glass rod and then re-centrifuge them. Another example comes from our own laboratory; A. D. R. Harrison sometimes noted a haze, prejudicial to fluorimetry, in the alkaline, aqueous phase from the back-extraction of heptane extracts. This haze, attributable to heptane droplets, could be obviated by centrifuging at 5 “C rather than a t room tempera- ture. Similar haze problems following back-extraction from chloroform extracts can often be removed by bubbling nitrogen through the aqueous phase.With some systems, mere standing without centrifugation rapidly brings about a clean separation of the phases , uncontaminated by denatured protein. Thus, centrifugation can be omitted when plasma is extracted with ben~ene,~ chloroform29 or a mixture of nonane and propan-2-01, as in the assay of triglycerides. Such systems are essential for automatic or semi-aut omatic extractors, which lack provision for centrifugation. Stockwell and Sawyer50 (see also reference 51) remark: “Solvent extraction systems may be used to effect concentration, clean up or to provide both functions. There is no universally available automatic equipment which will combine the features required to carry out both functions on widely varying volumes of original sample.” Semi-automatic apparatus, devised by Brown et aZ.52 and somewhat resembling a counter-current distribution apparatus, is available commercially, but the fact that it is tailored to urinary oestrogen assays limits its versatility.This criticism is less applicable to another commercially available instrument (designed by P. F. Dixon), which is based on a sampler turntable with a stirrer and dipping6 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, Vol. 101 probes. To achieve accurate control in automatically removing a phase, an interface- sensing device (eg., reference 53) is needed, but this is seldom warranted in routine assays. It is still common practice, albeit a tedious one, to remove a portion of one phase manually with the aid of a bulb-controlled pipette.Hydrophobic phase-separating filter-paper can help,54 although it may introduce contaminants such as tin.55 Automatic handling by the discrete approach is more likely to gain in popularity than is the continuous-flow approach, although the latter can be adapted (e.g., reference 56) to handle two-phase s y ~ t e m s . ~ ~ * 5 ~ Flow through a coil is also the key feature of an ingenious counter- current chromatographic apparatus that might justify the optimism of its originators in which a coil is continuously subjected to gentle ~entrifugation.~~ Development of Solvent Methods Certain aspects, including relative volumes, have been touched on in the sections onIon- isability and Other Chemical Features, Initial Treatment of Plasma Samples and Solvent Ex- traction (see also references 38 and 58).The so-called salt - solvent pair approach59 allows even a water-miscible solvent such as ethanol to be used in small amount, although the efficiency may be poor. The choice of solvent may have to take account not only of safety and other aspects already considered, but also of any manipulative preference for a low or high density compared with water. With 1,2-dichloroethane a useful lowering of density may be achieved by admixing a non-polar solvent snch as cyclohexane.11 Except with amphoteric compounds,11 the pH is normally set so that the sample compound is un-ionised, unless it is to be back-extracted from an organic phase for the purpose of puri- fication or of obtaining an aqueous solution for analysis.When a basic drug is unstable at an alkaline pH, a near-neutral pH may be suitable for its extraction into a solvent (e.g., a mixture of chloroform and propan-2-01) if an acidic agent, such as bromocresol purple, is added so as to yield a salt ; this salt may conveniently dissociate when a thin-layer chromato- gram is subsequently perf0rmed.~9,~~ With urinary metadrenalines at an alkaline pH, Coward and Smith61 found that addition of an aldehyde, such as dodecanal, with cyclohexanone as solvent, gave improved extraction, presumably as a result of complex formation ; however, A. A. A. Aziz, in our laboratory, could not achieve efficient extraction thereby, and these authors62 abandoned solvent extraction in favour OF an ion-exchange approach.Useful though it may be on occasion to choose an extractant that will interact chemically with the desired trace substance, solvent interactions occurring inadvertently can be trouble- some. Examples of such interactions include imine formation from primary amines with acetone, oxazolidine formation from /I-hydroxylated secondary amines with ketonic solvents and decomposition of some drug metabolites by ethyl acetate; interaction can also occur between halogenated hydrocarbons and many basic compounds.44 Chin and FastlichZ7 list a number of drugs that axe extractable (in the presence of sodium dihydrogen orthophosphate) with diethyl ether but. not with hexane. Ethyl methyl ketone at pH 2.0 was notably effective for q ~ e r c e t i n .~ ~ The stratagem of adding a trace amount of an alcohol to a hydrocarbon extractant, mentioned in the section on adsorptive losses,g can also overcome problems of inefficient and variable extraction, as encountered with pronethalol and pr~pranolol,~~ for which heptane containing 1 per cent. of ethanol was highly effective at pH 10.2. This stratagem has been successful not only with weakly basic drugs but also with weakly acidic drugs, such as atromid, where 2,2,4-trimethylpentane containing 5 per cent. of ethanol works well at acidic pH values; other mixtures that have been chosen empirically for various drugs include cyclohexane containing 2 per cent. of isoamyl alcohol (mainly 3-methyl- butan-1-01) or 5 per cent.of butan-1-01, and heptane containing 1.5 per cent. of isoamyl alcohol (B. Scales, personal communication). Consideration may have to be given to the question of selectivity.12 The main variables that can be manipulated are pH and solvent polarity. For barbiturates in plasma, in the context of spectrophotometric determination, Bush and Sanders-Bushs4 (see also reference 9) write: “The least polar solvent that will give a reasonable recovery of drug will also extract minimal amounts of ‘impurities’ which often interfere in subsequent steps of an analysis.”* * In terms of polarity,1a hydrocarbons and carbon tetrachloride can be grouped together, followed by a group that includes chloroform, diethyl ether and, a t its upper end, 1,2-dichloroethane, and then by a group containing 1,4-dioxan, ethyl acetate and pentan-1-01 ; the propanols have particularly high polarity, com- parable with that of ethanol.January, 19 76 OF ORGANIC COMPOUNDS IN PLASMA OR URINE 7 When the nature of interfering impurities is unclear, reliance has to be placed on an empirical approach. The following examples may be informative.For the gas - liquid chromatographic determination of glutethimide and a more polar active metabolite, Hansen and F i ~ c h e r ~ ~ achieved efficient extraction by use of diethyl ether (or chloroform, but not light petroleum) at an acidic pH, but had to remove impurities by subsequent extraction into hexane from an ethanolic solution, with aqueous hydrochloric acid present so as to give two phases. The separation of fluphenazine from its metabolites can be achieved by an approach that entailed a quantitative study of extraction behaviour in relation to pH with a solvent of appropriate polarity.66 For urinary 5-hydroxyindol-3-ylacetic acid, elaborate solvent-extraction steps to ensure specificity in the final colorimetry can be re- placed by a simple ether extraction (with hydrochloric acid and sodium chloride present) if 2-mercaptoethanol is finally added to suppress any adventitious colour.67 Care in choosing the first extractant can reduce the need for subsequent clean-up extractions.In the assay of serum for quinidine, benzene extraction conveniently left irrelevant metabolites in the aqueous p h a ~ e . ~ ~ ~ ~ For various drugs in carbonate-saturated solutions, Horning et aL4* found that propan-2-01, although as efficient as ethyl acetate, extracted impurities that proved to be confusing in subsequent chromatography.That butan-1-01 may be too powerful an extractant (and thus give high blanks) was demonstrated in a thorough study by Anton and S a ~ r e , ~ ~ who relied on carefully chosen solvent treatments in order to extract metadrenalines free from the parent catecholamines. It is illuminating to consider their elaborate series of extractions, each followed by centri- fugation, as applied at a near-neutral pH to urine, or, after removing protein centrifugally at an acidic pH, to plasma or a tissue homogenate. (1) Add isoamyl alcohol in order to remove contaminants (this is used together with a judiciously chosen buffer); remove the aqueous layer free from interface pigment and then (2) adjust to pH 10.0 with borate solution and saturate with dipotassium hydrogen orthophosphate, added as a solid.(3) Add diethyl ether; save the ether layer and (4) use 0-1 M hydrochloric acid to back-extract. (5) Next repeat steps (2) and (3), then (6) back-extract into 0.01 M hydrochloric acid in order to obtain a solution for assay by differential fluorimetry. Over-all recoveries as high as 50 per cent. were reported for this technique, but Weil-Malherbe68 had less success. Step (1) above illustrates the not uncommon practice (e.g., references 3 and 41) of pre- extracting so as to remove potentially interfering substances into an organic phase, which is then discarded. pre-extracted with ethyl acetate (three times) at pH 1.In this method the subsequent extraction of metadrena- lines from the aqueous phase (into toluene) is preceded by an oxidation step, in order to obtain an ultraviolet light absorbing product (vanillin), which is then back-extracted from toluene into carbonate solution before the spectrophotometric determination, the over-all outcome being high specificity and, it is claimed, quantitative recovery. The extraction of reaction products, as in the foregoing example (see also reference 70), can be a very useful approach, when applicable. Thus, urine that has been subjected to treatment with borohydride and periodate can be extracted first with 2-methylbutane and then with ether to extract different derived 17-oxosteroids for determination of the ll-oxy- genation index.71 Quercetin can be converted into a fluorescent complex that is extractable into diisopropyl ether.4l In the presence of a barbiturate and of calcium ions tetracyclines can form a fluorescent complex which, even if present in trace amount, is extractable into a solvent such as chl~roform.~~ Dill et aZ.19 have done notable work on the fluorescent dye method.In principle, it may be possible to achieve selectivity merely by chemical masking of unwanted compounds, e.g., of primary amines when a secondary amine is to be assayed by derivative formation from isotopes.70 It is illuminating to read Glazko's de~cription'~ of how a simplified method for the spectro- photometric determination of diphenylhydantoin (phenytoin) was developed.As a prelude to alkaline oxidation, in which chloroform would interfere, the drug (a weak acid) was extracted from plasma at pH 6.8 with 1,2-dichloroethane and then back-extracted into 1 N sodium hydroxide solution. Whereas other authors had steam-distilled the benzophenone formed by oxidation at 100 OC, Glazko collected it in an overlying layer of alumina-purified 2,2,4- tsimethylpentane, which projected above a gasket so that it could be air-cooled without exposure to steam heat. Glazko74 discusses the specificity of the method, its merits compared Thus, in order to assay metadrenalines Gupta et8 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION AnaZyst, VoZ. 101 with a gas - liquid chromatographic approach and the high sensitivity attainable by using fluorimetric measurement.Separation with a Solid Additive or Column Material To some extent, logic can guide the choice of a solid phase and of the correct working con- ditions to select a desired solute from aqueous solutions or remove unwanted solutes.12 This process is usually carried out with a column, although batch operation is sometimes practised. However, there is an empirical element even when the work is in the hands of knowledgeable investigators. The terms adsorption, partition and ion exchange do not imply mutual exclusiveness in the separation mechanism. Whatever the solid phase, there is a risk that a small absolute amount of the solute: may be retained, apparently irreversibly; the percentage loss can be serious when the solute is a trace constituent of a body fluid.Even with ion-exchange resins, there can be inadvertent absorptive losses, although as mentioned below, they can be minimised by adding an organic solvent or urea. Alumina and Charcoal Alumina represents a classical example of the use of an adsorbent, having long been re- cognised as an effective adsorbent for adrenaline or other aromatic compounds with vicinal hydroxyl groups when used a t a pH in the region of 8614975 Metadrenalines lack vicinal hydroxyl groups and should not be adsorbed; thus, when it is these metabolites rather than the parent catecholamines that are to be determined, alumina treatment may still be advan- tageous.14368 With catecholamines, Wong et aZ.15 found that they had to keep the amount of alumina very small as, on following published procedures, recoveries as low as 10 per cent.were sometimes obtained; interference in gas - liquid chromatography due to impurities in the alumina was also observed. One advantage possessed by alumina, and by adsorbents in general, is that comparatively high concentrations of salt, as found in urine, can. be tolerated even when augmented as a result of acid hydrolysis followed by ne~tra1isati~~n.l~ Nevertheless, adsorbents have not come into widespread use in the present context of sample preparation. There may be lingering suspicions about batch-to-batch variability and, at least with urine, about between- sample variations, which might jeopardise the hope of adsorbing the desired constituent quantitatively with minimum adsorption of unwanted constituents.This hope is most likely to be realised if the adsorption isotherm is steep at low concentrations of the con- stituent. It is immaterial that it may flatten off at high concentrations that will never be encountered in practice. The isotherms for adsorption of certain fatty axids on to charcoal (tested as a column) were shown by Hagdahl et aZ.76 (working in Uppsi&.) to depend on the medium. Whereas there was a steep initial portion with 50 per cent. m/V ethanol as the medium, with 95 per cent. ethanol the graph was linear (and shallow for hexanoic and octanoic acids, although not for butyric acid). The isotherm for phenylalanine was slightly depressed, although unaltered in shape, when 0.1 M hydrochloric acid was used !in place of water, but there remained the problem of achieving efficient elution.This led the Uppsala group76 to investigate the use of charcoal that had been desaturated (deactivated) suitably by treatment with hexan-1-01 (0.1 per cent. in water). This technique gave lower isotherms, still non-linear for phenylalanine and tryptophan: “It is not certain until the isotherms have been very considerably depressed, that desorption equilibrium will be readily established, as adsorption is very strong at low concentrations (less than 10 mM) .” Isotherms for various amino-acids were sufficiently different to allow chromatographic separation. In order to adsorb urinary aromatic constituents, while not attempting to resolve them, Asatoor and Dalgliesh77 used charcoal that had been deactivated with a long-chain aliphatic compound such as stearic acid and usually obtained efficient elution by use of water con- taining 5-10 per cent.m/V of phenol. For compounds with a basic (amino) group, these authors recommend that the elution be at an acidic pH, or that the deactivator be octadecyl- amine. In our own studies of urinary metadrena’lines, which can be adsorbed at a neutral pH, A. A. A. Aziz has obtained good recoveries with stearic acid (in ethanol) as deactivator, but not with octadecylamine ; acidified methanol was a good eluting agent, whereas aqueous phenol was not. Separation from urinary pigments was readily achieved. A. J. Winter hasJanuary, 1976 OF ORGANIC COMPOUNDS I N PLASMA OR URINE 9 noted that corticosterone and acidic compounds, in contrast to metadrenalines, are held rather tenaciously.Charcoal that has been stored exposed to air may lose its adsorptive efficiency, as observed in the pioneer studies in U p p ~ a l a , ~ ~ which studies are informative concerning the properties and processing of charcoal. Active charcoal was used to adsorb a urinary imipramine metabolite, which was then eluted with aqueous ethanol, apparently non-q~antitatively.~~ This metabolite was shown to be a glucuronide. Meola and Vankoao describe a drug screening procedure that is based on adsorption of the drug from urine at pH 11 with Norit A charcoal (neutral, pharmaceutical grade), followed by elution with diethyl ether or a mixture of chloroform and propan-2-01, but give no information on recoveries. However, in the drug assay field the literature on the use of charcoal is sparse, and hardly takes cognisance of the pre-1960 Uppsala literature cited above, or of potentially relevant radioimmunoassay literature, describing methods in which the carbon particles are pre-coated with a protein or other compound of high relative molecular mass and the question of elution does not arise.Here the aim is merely to remove unbound compound selectively after ligand treatment ; with urine there may be poor adsorp- tion owing to interference by urea and creatinine.al Ion-exchange Resins As the voluminous literature testifies,14 urinary catecholamines and their metabolites have been a focus for the trial of cation-exchange resins in sample preparation. The resin treatment may be preceded by a procedure such as adsorption on to alumina, or the resin may be applied directly to a urine sample if the problem of interfering cations is circumvented.A simple procedures2 which, in our experience, works better than more complex procedures, consists in passing a diluted urine sample at near-neutral pH, after acid hydrolysis, through a short Dowex 50 (ammonium) cation-exchange column, in the presence of urea so as to minimise adsorption. Metadrenalines are eluted with dilute ammonia solution and survive subsequent concentration, whereas catecholamines are destroyed (as we have confirmed). The de-salted concentrate of metadrenalines thereby obtained, although free from amino-acids, is still too impure for fluorimetric analysis. It is therefore subjected to cation-exchange chromato- graphy with carboxymethylcellulose under carefully defined conditions, although this method is partly empirical,82 or with cellulose phosphate, which is a more refined method in that the two metadrenalines can be separated,s3 as had previously been claimed for a carboxylic resin of methacrylate type.84 A Dowex 50 (sulphonic resin) step performed with more elaborate conditions than those outlined abovea2ya3 has been described with several variations, e.g., the use of hydrochloric or formic acid to allow unwanted compounds, such as catecholamines, to be removed.Despite meticulous descriptions of resin preparation and running conditions, as in a proceduree4 in which amino-acids were largely removed by a sodium acetate washing step, few published procedures have been readily repeatable outside the originator’s laboratory.This contention also holds for procedures in which a carboxylic rather than a sulphonic cation exchanger is advocated14ya5 for the initial treatment of the urine, one advantage of this refinement being the greater purity of the resin as purchased. Stratagems such as the use of borate to elute catecholamines as complexesa4 likewise may not be readily repeatable. Therefore, one should not be too sanguine in expecting published cation-exchange procedures to work in one’s own hands without special care and should be alert to possible batch variations in the resin itself. Notwithstanding these reservations, the potential selectivity of ion-exchange resinsS6 is a strong incentive to use them in sample preparation despite the tedium of pre-washing and of re-attaining equilibrium when the composition of the inflowing liquid is changed.With weak cation exchangers (such as Biorex 70) the risk of impaired loading due to urinary electro- lytes is greater than with strong (sulphonic) cation-exchangers. Weil-Malherbe6s therefore advocated an electrodialytic de-salting procedure, but later14 he accepted this risk, stipulating merely that the urine sample must not be so large as to saturate the column with inorganic cations. While he questioned the usefulness of diluting the urine, this step does seem to be a useful p r e c a u t i ~ n . ~ ~ ~ ~ ~ Difficulties due to salts can, of course, arise not only with urine but also with resin eluates that are to be re-run in a second ion-exchange column, unless only volatilisable components have been used for elution.Chalmers and co-~orkers8~-89 use pyridinium acetate, removable by freeze-drying?O in separating acidic compounds from urine (which can be loaded directly)10 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, Vol. 101 by anion-exchange chromatography with DEAE - Sephadex. With polystyrene-based anion exchangersJ8 such as Dowex 1, some acidic compounds may decompose or be retained too strongly.88 Nevertheless , anion exchangers such as Dowex 1 (which may conveniently remain wet if the inflow of eluting agent is inadvertently interrupted) deserve to be more widely used for the removal of unwanted acidic constituents. Investigators using strong anion or cation exchangers for sample preparation might take cognisance of early literatures6 on the successful use of organic solvents as eluting agents, the solvents acting partly in a partition mode.One example of the use of aqueous - organic systems is the addition of lJ4-dioxan to minimise adsorptive retardation with organic acids.91 Ethanol can improve the elution of dopamine and 5-hydroxytryptamine from Dowex 50 resin with dilute hydrochloric acid, good recoveries being obtained even with very low loads; aqueous methanol is a useful first eluting agent.92 It may, however, be technically difficult to achieve a smooth, bubble-free transition from an aqueous to an organic system with an ion exchanger, or with the non-ionic resin XAD-2. Amberlite XAD-2 Resin This hydrophobic polystyrene resin (Rohm and Haas Co., Amberlite series) lacks ionisable groups and acts as a weak adsorbent; it is potentially valuable for removing trace consti- tuents that possess a hydrophobic moiety and must be isolated from an aqueous medium, such as urine.Once adsorbed on to a column of the coarse, porous beads, the compound can be eluted with an agent such as methanol. On isolating morphine (with a good recovery) and its metabolites from urine, Fujimoto and Wang93 found that the methanol used to elute substances from the XAD-2 column first re- moved urinary pigments, which therefore served as a useful advance signal. Subsequent thin- layer chromatography gave a pattern which was much cleaner than that for material which had been solvent-extracted from hydrolysed urine.Mu16 et aLg4 examined the behaviour of XAD-2 resin towards urine that was spiked with various drugs of abuse. The eluting agent selected was chloroform - propan-2-01. Whereas some drugs, such as phenobarbital, were efficiently adsorbed and eluted, with others there was poor adsorption and/or elution. With morphine (recovery only 64 per cent.) and especially with its glucuronide (recovery nil) the main difficulty was incomplete adsorption, which was not remedied by adjusting the pH to 9. Kullberg and Gorodet~ky~~ treated urine at pH 8.5 with XAD-2 and eluted with methanol - chloroform , which gave a clean extract for subsequent thin-layer chromatography. They obtained 60-80 per cent. recovery for morphine arid various other drugs of abuse, except for aspirin (which was poorly adsorbed).They describe important points of technique (see also references 95 and 96). Elution with organic solvents entails technical difficulties that are associated partly with the displacement of residual water from the column, and can be mini- mised by prior passage of a small volume of acetone. This water may form, in the collecting tube, an upper phase which, advantageously to thin-layer chromatography, contains un- wanted urinary pigments but which would also contain much of the eluted drug if it was not deliberately saturated with ammonium chloride solution at pH 10 (see also reference 94). Adsorptive resins, such as XAD-2, are evidently not a panacea for sample preparation. XAD-2 scarcely adsorbs purine analogues (A.Bye, personal communication) , and in our laboratory it has not consistently given a good recovery of both metadrenalines. Its efficacy may vary with different urines, perhaps reflecting erratic selectivity for unwanted urinary constituents in addition to the compound under XAD-2 resin is not normally used to separate one drug from another, although potentially it could achieve Moreover, the efficacy of XAD-2 in dealing with morphine glucuro- nide is controver~ial.3~~~~~9~ The stratagem of hydrolysing the glucuronide on the XAD-2 column gave only a 40 per cent. over-all recovery, although the thin-layer chromatographic patterns were cleaner than for material separated by XAD-2 from hydrolysed urine.32 How- ever, good recovery of a sulphate conjugate by using XAD-2 has been reported in the isolation of bile acids from urine.98 Other Agents For aromatic constituents of urine (not quantified) , Vallon et aLQQ used Sephadex G-10 in an adsorption mode at pH 1, and on raising the pH they achieved selective elution, apparentlyJanuary, 1976 OF ORGANIC COMPOUNDS I N PLASMA OR URINE 11 less facile with conjugates.Oestrogen conjugates can be isolated, perhaps with some in- advertent selectivity, by the aid of Sephadex.loOJO1 Chromatography on lipophilic Sephadex1O2 with organic solvents has served as an intermediate step in cleaning up extracted material for thin-layer chromatography. Thus Makino et aLg8 (cj. reference 102) used Sephadex LH-20 to chromatograph a chloroform - methanol solution of sulphated and non-sulphated bile acids that had been isolated from plasma or urine by XAD-2.This chromatography, which in the case of the sulphate fraction was followed by hydrolysis, gave bile acids, which in turn were converted into methyl esters and purified in benzene solution on an aluminium oxide column. Another study in the steroid field has entailed the use of Sephadex rendered rich in hydroxy- alkoxypropyl r e ~ i d u e s . ~ ~ ~ J ~ * In the isolation of homovanillic acid from rat urine, lipophilic Sephadex facilitated the final clean-up (unnecessary with human urine) after chromatography on silicic acid of a solution in chloroform.105 Dry silica gel, when added to acidified urine, selectively extracts weak carboxylic acids, which can then be eluted from the gel (as a column) by organic solvents.lm Silica gel can also adsorb from a dichloromethane extract.lo7 For the removal of coloured impurities, an aqueous solution of metadrenalines isolated from urine for fluorimetric measurement was treated with Florex (an aluminium silicate) in the presence of ureaa2; significant losses were later encountered, h0wever.~3 Choice of Approach In addition to the various pre-separation procedures surveyed above, there remains the option of chromatography on a thin layer, as applied to chlorpropamide and its metabolites,los or on suitably pre-treatedlo9 paper.Chromatography on a conventional film or sheet may, however, be inappropriate because only a low load of the salt-free urine or other test fluid can be applied, even if streaked rather than spotted.This possible limitation also applies to sheet electrophoresis, as used with pre-concentrated urine for metabolites of 3H-noradre- naline, in which the metabolites are separated on paper in a borate buffer at 300 V.l10 Another conceivable approach, hardly justifiable in practice, is preparative gas - liquid chromato- A possible variant of the solvent approach, carried out in the interests of efficiency if not selectivity, is liquid - solid extraction after freeze-drying the ample.^^^^^^ Freeze-drying under carefully chosen conditions has proved to be a useful intermediate step in the analysis of urine for organic acids,s7 but seems unlikely to be widely adopted. If the final analytical step entails chromatography with a small loading volume, the sample preparation approach must be appropriate for obtaining the sample in a concentrated form.Semi-purified samples may have to be evaporated, taking all necessary precautions; thus, an organic base may have to be dried as its hydrochloride in order to obviate volatilisation. Even when solvent removal can be carried out at atmospheric pressure, suitably with the aid of a manifold and a stream of nitrogen,l12 it is a tedious operation. However, the need to use only a minute volume of solvent for meticulously dissolving the residue can be obviated, when a chromatographic plate has to be loaded, if an automatic device for repeated spotting is available, thus allowing the use of a larger v01ume.l~~ As already mentioned in connection with ion-exchange chromatography, difficulty arises if involatile salts are present. This problem does not apply if solvent-extraction procedures have been used. graphy- Generalisations and Examples Solvent extraction is generally the ideal method if, as happens all too rarely, a single primary extraction (followed, if necessary, by one washing step and a back-extraction) provides a good recovery (in terms of amount and condition) of the desired compound.Conditions can be manipulated to give selectivity of extraction (see the section on Develop- ment of Solvent Methods). Clifford and S r n ~ t h ~ ~ give an illuminating survey in the context of 1,4-benzodiazepines with different pK values and note that diethyl ether at pH 7-9 is often an effective solvent. The statement that amphoteric or neutral compounds do not lend themselves to solvent extractions8 is rather sweeping.In the toxicological field, alternative approaches such as chromatography were still rare when the applicability of classical chro- matographic methods was reviewed by Kirk13 in 1968. (HPLC has only recently achieved recognition, as is discussed below; it is hardly a ready means of sample preparation.)12 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, VoZ. IOI Among the few direct comparisons of solvent extraction with alternative procedures is the report by Miller et whose XAD-2 results with urinary morphine (see also reference 32) were “comparable to or better than those obtained by liquid - liquid extraction techniques. In addition, the studies show that morphine extracted from urine by XAD-2 resin can be stored on the resin for 14 d without degradation.” In his survey of methods for determining urinary catecholamines and related compounds, Weil-Malherbe14 emphasises the need for a selective solid-phase step (adsorptive or ion-exchange), and sees no advantage in having a prior solvent-extraction step.Indeed, as indicated in the section on Development of Solvent Methods (see also reference 62), metadrenalines are not readily obtained in good yield and purity by solvent extraction. Because, however, adsorption on alumina does not itself guarantee specific fluorescence measurements in the instance of catecholamines from acid- hydrolysed urine, the eluate from the alumina column should be put through an ion-exchange step.6s Ion-exchange column procedures, while being tedious and sometimes difficult, are evidently advantageous for discriminatory separations.Whien partition coefficients are unfavourable for solvent extraction (as with some organic acidsss) or when the eluate can be assayed directly by means of a photometric procedure, the ion-exc hange approach is particularly attractive. The narrower the column, the slower will be the loacling step but the better will be the prospect of eluting the desired constituent in a small volume. In order to obtain good resolution it may be helpful to use a long column, to use unorthodox eluting agents (e.g., reference 92) and also to use a gradient (smooth or stepped), ;t stratagem that has understandably not found favour in connection with body-fluid assays.Poor recoveries with low loads (see section on Ion-exchange Resins) are a particular h.azard, which may go undetected if model experiments are performed only with loads of the trace substance that are unrealistically high. Once the conditions have been optimised for a chromatographic separation, ion-exchange or otherwise, it should be possible to dispense with a fraction collector and to collect stan- dardised portions of eluate, particularly if the separation does not have to be rigorous and a short column suffices. It may even be feasible to operate in a batch mode rather than with a column, or to use ion-exchange paper (see reference 32). However, so-called liquid ion exchange (as used in a steroid study114) is really a refinement of the solvent-extraction approach that depends on ion pairing; use of the term has been strongly condemned.12 Ion pairing is in fact a powerful aid to efficiency and discrimination both in partition chromatography and in solvent e~tracti0n.l~~ Notable difficulties in choosing conditions for sample preparation are exemplified by the literature in two fields.The steroid field abounds with examples of mu1 ti-stage techniques,loS e.g., the successive use of solvent extraction, ion- exchange and thin-layer chromatography for urinary oestrogens in trace amounts as a prelude to their determination (as esters) by gas - liquid chromatography.l16 AdlercreutzlOl considers that, for pregnancy oestrogens, sample preparation needs to be more rigorous with gas - liquid chromatography than with colorimetry or fluorimetry.The biogenic amine field has likewise proved to be a test of patience and judgement : “values (for metadrenalines) obtained from 24-h samples of urine collected from normal individuals vary greatly according to the method used, even to the point where values analyzed in the same laboratory at different times seem to differ.”E2 Procedures entailing solvent extraction, ion exchange or adsorption on alumina have already been discussed. When the two metadrenalines have to be measured individually in the same sample, the risks of differential determination by a non-separatory method, such as fluori- metry, may have to be accepted. Preferably, however, the compounds should be separated, and this can be achieved for metadrenalines by use of cellulose phosphate chromatography= or by gas - liquid chromatography after derivative formation. Thin-layer chromatography has also been used, for the challenging problem of determining adrenaline and noradrenaline in plasma.117 Following the enzymic introduction of radioisotopically labelled methyl groups, the pair of metadrenalines were selectively solvent extracted, with rather low efficiency, before the thin-layer chromatographic separation, after which each was eluted and oxidised to vanillin, which was counted following solvent-extraction steps. The efficacy of gas - liquid chromatography in tirug analysis has been well reviewed by Riedmann,36 who considered the sample preparation aspect and the derivative-formation step that is usually unavoidable.If a nitrogen-selective detector is used, a simplified sample preparation procedure may suffice.ll* Whenever sample preparation demands a sequence ofI Solvent extraction Add salt? or a complexing agent? I lon-exchange resin First de-salt? i Adsorption procedure e.g., Alumina XA D -2 resin Charcoal Gel, e.g., Sephadex Apply procedures such as the above (concentrate if necessary) to clean up for determination I TLC (or paper chromatography) Determine Elute in situ I Determination (differential?) off-line in dilute solution Colorimetry U V absorption Fluorimetry I Der ivat ise , Column chroma- tography, liquid then GLC FID mobile phase ECD Specific element Conven- +l HPLC detector tional UV de- I I tec tor Fig.1. Possible stages in the assay of plasma or urine for organic compounds of low relative molecular mass (e.g., drugs or catecholamine metabolites) present in trace amounts.Trends in Methods of Sample Preparation Difficult though it is to generalise, some salient points can now be stated. (1). In connection with bioavailability assessment and therapeutic monitoring, there is an increasing need to obtain reliable values (say within &5 per cent.) for drug levels, particularly in plasma, at concentrations that may be below 100 pg 1-l. When present-day assay methods purport to achieve the above, with some assurance of specificity, successful operation may call for experience, dexterity and patience, and may be costly in terms of man-hours. In order to achieve the requisite sensitivity and discrimination by use of approaches other than radioimmunoassay or related affinity methods, an adequate sample-preparation procedure (Fig. 1) must be used, which takes account of any need to measure conjugates (see section on Methods of Handling Conjugates, below) or other metabolites.This procedure may have to be rather elaborate [cf. (2), and section on Concluding Comments, below], possibly in the interests of obtaining low blanks as well as good discrimi- nation when there is a classical photometric end-measurement (colorimetry, fluorimetry or ultraviolet absorptiometry ; Fig. 1) that is not concomitant with a chromatographic separation. For such a separatory measurement it has become traditional to use gas-liquid chromatography. While improved techniques for forming derivatives and detection are helping to keep gas - liquid chromatography in the forefront of methods, its protagonists need to be alert to two alternative approaches that do not entail derivative formation.One is thin-layer chromatography, which can readily give tolerably good precision and productivity (2). (3). (4). (5).14 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, VoZ. 101 now that automatic instruments for applying spots and for scanning them in sitzc have become a~ailable.1~~ The other is high-pressure liquid chromatography. The Advent of High-pressure Liquid Chromatography (HPLC) When Michaelis et al.l19 reviewed this powerful technique and its pharmaceutical applica- tions in 1973, many trace constituents were already amenable to analysis by HPLC.Even disregarding mere model separations of unproved applicability to plasma and urine, the list has become increasingly impressive.120 However, in a recent HPLC study of phenothiazine metabolites that had been isolated from urine by ion-exchange and solvent-extraction pro- cedures, Landgraf121 was not greatly impressed by his HPLC results, particularly the quan- titative aspects (depending on measurements made at a wavelength of 254nm). This contrasts with the enthusiasm of the Oak Ridge group, who have been ambitiously applying HPLC (not really high speed) to “profile” whole urine in the context of screening.122 Recently25 they have applied anion-exchange HFLC to a serum or urine ultrafiltrate as a means of following the metabolic fate of acetaminophen, when given in high dosage.HPLC with a pellicular cation exchanger has been employed to follow the fate of ingested furo- semidele; with unprocessed serum or urine a concentration of the order of 5 mg 1-1 could be measured r e a d i l ~ . l ~ ~ When comparatively high concentrations are being analysed, it may be valid to use raw samples, preferably protecting the column with a disposable pre-column. In general, how- ever, it is advisable to apply sample-preparation steps initially.10**124 In scanning urine by HPLC, which otherwise would be performed directly on the urine, batchwise pre-treatment with a cation exchanger served to reduce noise due to ultraviolet-absorbing basic constituents of urine.125 In attempting to carry out the determination by cation-exchange HPLC of urinary metadrenalines at concentrations below 100 pg F , J.P. Leppard, in this laboratory, found it essential to reduce noise by a sample-preparation step such as treatment with Dowex 50.s2 In determining theophylline, Manion et ~ 1 . ~ ~ found that elimination of caffeine interference from plasma (by solvent extraction) was less important with HPLC than with gas - liquid chromatography; an XAD-2 column step had to be adopted in the HPLC analysis of urine. Usually, the sample-preparation procedure need not be elaborate as HPLC allows good discrimination, if not sensitivity. HPLC is, unfortunately, less sensitive than gas - liquid chromatography because of the present-day reliance on ultraviolet detection. Fluorimetric detection may gain in popularity when a reagent-addition facility becomes a standard feature of HPLC apparatus, enabling fluorescent derivatives to be formed after the Electrochemical detection is a promising technique but is not without difficulties, among which is the constraint of having to use acidic, de-gassed eluting agents.Already, it has, in conjunction with a cation exchanger, enabled L-dopa to be determined in serum a t a level of lOOpgl-l, following deproteinisation (with perchloric acid) and an alumina column step, which achieved a 20-fold concentration.126 Even this level of sensitivity, ELS obtained by use of a concentrate, is far less impressive than that of gas - liquid chromatographic analysis. Chemical ionisation mass spectrometry used on-line with HPLC could give good sensitivity, perhaps with some sacrifice of HPLC resolution.This mass-spectrometric approach has even been applied directly to drugs and metabolites in plasma, with no sample preparation other than extraction into benzene; deuterated derivatives of the compounds were added as internal standards.12’ Field desorption mass spectrometry is, in principle, well suited to HPLC eluates, but it is rather insensitive; it could, however, be advantageous in enabling intact conjugates to be examined without chromatographic separation.l= Methods of Handling Conjugates Although conjugates are rather involatile for gas - liquid chromatography, they may be amenable to analysis by HPLC. The above HI?LC analysis of urine after ingestion of a~etaminophen~~ gave sulphate conjugates late in the elution process (near 21 h).In a preliminary report,129 which, however, hardly inspires confidence, the glucuronide and the sulphate of +-hydroxyacetanilide, of which the ingested dose was high, gave good HPLC peaks on an anion exchanger with a 1-p1 urine sample, but other sulphates were difficult to elute. Before the advent of HPLC, conventional. ion-exchange columns were shown to beJanuary, 1976 OF ORGANIC COMPOUNDS IN PLASMA OR URINE 15 capable of handling sulphate conjugates of urinary metadrenalines (at acidic pH) andg49l3O glucuronide conjugates of dopamine metabolites.131 Evidently, then, initial hydrolysis is not obligatory when urinary conjugates have to be determined. For sample preparation organic solvents are likely to be ineffective88 unless the extraction is effected at pH 1 (inapplicable to basic d r ~ g ~ ) ~ ~ ~ ~ ~ ~ ~ ~ or an ion-association ~ y ~ t e m ~ ~ J ~ ~ ~ ~ ~ ~ can be found.Isolated observations cited in the section on Separation with a Solid Additive or Column Material indicate that charcoal, Sephadex and XAD-2 may some- times be effective in extracting conjugates. It is sometimes of intrinsic interest to ascertain what proportion of a metabolisable com- pound is converted into conjugates in '0iz)0.103~130 Usually, however, the advantage of extracting intact conjugates lies rather in tidier analyses, insofar as initial hydrolysis of urine may sometimes create problems later (see Hydrolysis of Conjugates). Acid hydrolysis is particularly likely to create noise problems when the conjugate needs rigorous conditions to ensure complete hydrolysis, as exemplified by the sulphates of certain steroids including bile acids.Such conditions may be analytically innocuous if applied to semi-purified con- jugates, rather than to the original urine, as in a studys8 cited under Other Agents. In this study, bile acid sulphates were split solvolytically by standing them for 2 d in ethanol - acetone a t pH 1, then neutralising and drying them; the residue was heated for 4 h at 120 "C in 15 per cent. sodium hydroxide solution. In an infant steroid inve~tigation,1~~J~~ incubation of an ethyl acetate extract obtained from urine at pH 1 led to solvolysis of sulphates but not of glucuronides. had some success with mild procedures for acid hydrolysis in the presence of an organic solvent. Evidently, techniques are potentially available for handling intact conjugates, to be hydro- lysed either unconventionally (not at the outset) or not at all. The feasibility of hydro- lysing conjugates while they are adsorbed on XAD-2 or charcoal deserves investigation, in spite of one disappointing re~ult.3~ Concluding Comments In the metadrenalines field, Fecher et The next few years are likely to see piecemeal advances rather than major innovations in assay techniques for plasma and urine.HPLC is now being absorbed into the repertoire of approaches and will be rendered increasingly useful by advances in detector instrumentation120 (see the section on HPLC above) and in separation systems. Ion-pair partition systems, as recently tried for biogenic amines and acidic metabolites,132 offer selectivity in solvent extrac- tion115 as well as in chr~matography.~~~ Such a system has enabled extracted nucleotides to be freed from trichloroacetic or perchloric acid.134 Advances in sample preparation, relevant even to HPLC, will come largely from the skilful use of solid phases for isolative steps that are not readily effected with solvents.As indicated in the sections on Alumina and Charcoal and Amberlite XAD-2 Resin, even a simple adsorptive procedure may be convenient for concentrating or cleaning up, perhaps with modest selectivity. With ion exchangers, which are especially useful for basic compounds, judicious choice of conditions, particularly pH, can give selectivity in the loading step as well as in elution (eg., references 82 and 131). Ion-exchange membrane fdt.erP5 warrant investigation for their possible usefulness in the present context of organic trace analysis.There is a particular need for adsorptive procedures that will effectively extract compounds from plasma even if strongly protein-bound. This has already been achieved with thyroxine by use of Sephadex G-25 in a strongly alkaline medium.136 A useful corollary would be ways in which to determine free compounds and protein-bound compounds ~eparate1y.l~~ Radio- immunoassay literature is relevant to the choice of adsorptive step, although not to the requisite desorption. Neglect to check how bond compounds behave is common. Conventional sample-preparation procedures that have to be run repetitively warrant automatic, or at least work-simplified, appro ache^.^^,^^^ Here the cost-effectiveness aspect has to be considered, as it should be also in comparing different ways of assaying a particular compound.Account has to be taken of operator time, instrument cost and work-load, resolution attainable in the final measurement step and the desired precision and ~pecificity.~ Bushlog gives a warning about possible tribulations that could be encountered with un- conventional methods, such as double isotope dilution. The setting up of a radioimmuno- assay or other affinity method should not be lightly embarked on. Cost effectiveness may rule out the use, in multi-sample routine work, of lengthy methods with a low throughput,16 REID : SAMPLE PREPARATION IN THE MICRO-DETERMINATION Analyst, VoZ.101 good though they may be in respect of specificity. as with metadrenalines,*2P it may be worth having available both a routine method, possibly not very specific even when only a single drug has been administered, and a more elaborate method that meets the rigorous requirements of research investigations. Thus, an.ticonvulsants to be gas chromatographed can be isolated and methylated by a simplified procedure that obviates solvent evaporation.138 Within this Bioanalytical Centre, funded by the Wolfson Foundation to be of service t o industry, Drs. A. D. R. Harrison and J. P. Leppard deserve particular thanks for enthusiastic work and for comments on a draft of this review. Other useful comments were received from Drs. A.Bye and R. H. Nimmo-Smith of the Wellcome Foundation. The work on meta- drenalines carried out at the Centre was supported by the Medical Research Council. With 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. References Bassett, R. G., Wright, J. A., and Cravey, R. H., Clin. Chem., 1975, 21, 44. Cram&, G., and Isaksson, B., Scand. J . Clin. Lab. Invest., 1963, 15, 553. Moss, G. F., Jones, K. M., Ritchie, J. T., and Cox, J. S. G., Toxic. Appl. Pharmac., 1971, 20, 147. Muld, S. J., Bastos, M. L., and Jukofsky, D., Clin. (Them., 1974, 20, 243. Butler, V. P., Metabolism, 1973, 22, 1145. Teale, J. D., Forman, E. J., King, L. J., and Marks, V., Proc.SOC. Analyt. Chem., 1974, 11, 219. Murphy, B. P., and Pattee, C. J., J . Clin. Endow. Metab., 1964, 24, 187. Murphy, B. P., J . Lab. Clin. Med., 1965, 66, 161. Brodie, B. B., Udenfriend, S., and Baer, J. E., J . Biol. Chem., 1947, 168, 299. Titus, E. O., in La Du, B. N., Mandel, H. G., and Way, E. L., Editors, “Fundamentals of Drug Metabolism and Drug Disposition,’’ Williams and Wilkins, Baltimore, U.S.A., 1972, p. 419. Trevor, A., Rowland, M., and Way, E. L., in La Du, B. N., Mandel, H. G., and Way, E. L., Editors, “Fundamentals of Drug Metabolism and Drug 13isposition,” Williams and Wilkins, Baltimore, U.S.A., 1972, p. 369. Karger, B. L., Snyder, L. R., and Horvath, C., “An Introduction to Separation Science,” John Wiley and Sons, New York, 1973. Kirk, P.L., Adv. Chromat., 1968, 5, 79. Weil-Malherbe, H., in Glick, D., Editar, “Methods of Biochemical Analysis,” Supplementary volume, “Analysis of Biogenic Amines and their Related Enzymes,” Interscience Publishers, New York, 1971, p. 119. Wong, K. P., Ruthven, C. R. J., and Sandler, M., Clinica Chim. Acta, 1973, 47, 216. Forrey, A. W., Kimpel, B., Blair, A. D., and Cutler, R. E., CEin. Chem., 1974, 20, 152. O’Brien, J., Zazulak, W., Abbey, V., and Hinsvark, O., J . Chromat. Sci., 1972, 10, 336. Brodie, B. B., Udenfriend, S., Dill, W., and Downing, G., J . Biol. Chem., 1947, 168, 311. Dill, W. A., Chucot, L., Chang, T., and Glazko, A. J., Anesthesiology, 1971, 34, 73. Spirtes, M. A.. Clin. Chem., 1972, 18, 317. Wigfield, D. C., Analyt. Biochem., 1975, 63, 286. Kushinsky, S., and Anderson, M., Clin.Chem., 1974., 20, 1528. Brown, H. H., Vanko, M., and Meola, J. M., Clin. Chem., 1974, 20, 919. Deome, A. J., Clin. Chem., 1974, 20, 1383. Mrochek, J. E., Katz, S., Christie, W. H., and Dinsmore, S. R., Clin. Chem., 1974, 20, 1086. Matin, S. B., Karam, J, H., and Forsham, P. H., Analyt. Chem., 1976, 47, 646. Chin, D., and Fastlich, E., Clin. Chem., 1974, 20, 1382. Carmichael, R. H., Clin. Chem., 1959, 5, 597. Gardner-Thorpe, C., Parsonage, M. J., and Toothill, C., Clinica Chim. Acta, 1971, 35, 39. Goldberg, V. D., Rosewarne, C., and Lascelles, P. ‘I., Proc. SOC. Analyt. Chem., 1974, 11, 288. Hartel, G., and Harjanne, A., Clinica Chim. Acta, 1969, 23, 289. Kullberg, M. P., and Gorodetzky, C. W., Clin. Chenz., 1974, 20, 177. Anton, A.H., Serrano, A., Tjandramaga, T. B., and Goldberg, L. I., Science, N.Y., 1973, 182, 60. Shackleton, C. H. L., Charro-Salgado, A. L., and Mitchell, F. L., Clinica Chim. Acta, 1968, 21, 105. Clifford, J. M., and Smyth, W. F., Analyst, 1974, 90, 241. Riedmann, M., Xenobiotica, 1973, 3, 411. Daley, R. D., in Florey, K., Editor, “Analytical Profiles of Drug Substances,” Volume 1, Academic Press, New York, 1972, p. 119. Jones, C. R., in Reid, E., Editor, “Methodological Developments in Biochemistry. 5. Assay of Drugs and Other Trace Organics in Biological Fluids,” ASP Biological and Medical Press B.V., Amsterdam, 1976, in the press. Farrand, R. J., J. Clin. Path., 1966, 19, 92. Dill, W. A., and Glazko, A. J., Clin. Chem., 1972, 18, 675. Gugler, R., and Dengler, H. J., Clin.Chem., 1973, 19, 36. Sammons, H. G., Almond, S. J., and Botterill, V. &I., Ann .Clin. Biochem.. 1969, 6, 169. Anton, A. H., and Sayre, D. F., J . Pharmac. Exp. They., 1966, 153, 15. Beckett, A. H., in Porter, R., and O’Connor, M., Editors, “The Poisoned Patient,’’ Ciba Foundation Symposium No. 26, Elsevier Publishing Co., Amsterdam, 1974, p. 57. Jackson, J. V., in Clarke, E. G. C., Editov, “Isolation and Identification of Drugs,” Pharmaceutical Press, London, 1969, p. 16.January, 1976 OF ORGANIC COMPOUNDS IN PLASMA OR URINE 17 46. 47. 48. 49. 50. 51. 62. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 66. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 86. 86. 87. 88. 89. 90. 91. 92. 93. 94. 96. 96. 97. 98. 99. 100.101. 102. Ramsey, J., and Campbell, D. B., J . Chromat., 1971, 63, 303. Aggarwal, V., Bath, R., and Sunshine, I., Clin. Chem., 1974, 20, 307. Horning, M. G., Gregory, P., Nowlin, J., Stafford, M., Lertratanangkoon, K., Butler, C., Stillwell, Brodie, B. B., and Udenfriend, S., J . Biol. Chem., 1945, 158, 705. Stockwell, P. B., and Sawyer, R., Lab. Equip. Dig., 1972, 10 (ll), 51. Foreman, J. K. F., and Stockwell, P. B., “Automatic Chemical Analysis,” Ellis Horwood, Chichester, Brown, J. B., MacLeod, S. C., Macnaughtan, C., Smith, M. A., and Smyth, B., J . Endocr., 1988, Porter, D. G., Jackson, C. J., and Bunting, W., Lab. Pract., 1974, 23, 111. Manion, C. V., Shoeman, D. W., and Azarnoff, D. L., J . Chromat., 1974, 101, 169. Meakin, J. C., Clin. Chem., 1973, 19, 141.Blackmore, D. J., Curry, A. S., Hayes, T. S., and Rutter, E. R., Clin. Chem., 1971, 17, 896. Ito, Y., Hurst, R. E., Bowman, R. L., and Achter, E. K., Separ. Purif. Methods, 1974, 3, 117. Blanke, R. V., in Tietz, N. W., Editor, “Fundamentals of Clinical Chemistry,” Saunders Publishers, Bastos, M. L., Kananen, G. E., Young, R. M., Monforte, J. R., and Sunshine, I., Clin. Chem., 1970, Stoner, R. E., and Parker, C., Clin. Chem., 1974, 20, 309. Coward, R. F., and Smith, P., Chemy Ind., 1966, 210. Coward, R. F., and Smith, P., Clinica Chim. Acta, 1966, 14, 672. Black, J. W., Duncan, W. A. M., and Shanks, R. G., Brit. J . Pharrnac., 1965, 25, 577. Bush, M. J., and Sanders-Bush, E., in Woodbury, D. M., Penvy, J. K., and Schmidt, R. P., Editors, Hansen, A. R., and Fischer, L.J., Clin. Chem., 1974, 20, 236. Whelpton, R., and Curry, S. H., in Reid, E., Editor, “Methodological Developments in Biochemistry. Assay of Drugs and Other Trace Organics in Biological Fluids,” ASP Biological and Medical W. G., and Hill, R. M., Clin. Chem., 1974, 20, 282. 1974. 42, 5. Philadelphia Pa., 1970, p. 833. 16, 931. “Antiepileptic Drugs,” Raven Press, New York, 1972, p. 293. 5. Press B.V., Amsterdam, 1976, in the press. Goldenberg, H., Clin. Chem., 1973, 19, 38. Weil-Malherbe, H., Meth. Biochem. Analysis, 1968, 16, 293. Gupta, R. N., Price, D., and Keane, P. M., Clin. Chem., 1973, 19, 611. Chamberlain, J., in Reid, E., Editor, “Methodological Developments in Biochemistry. Few, J. D., J . Endocr., 1968, 41, 213. Kohn, K. W., Analyt. Chem., 1961, 33, 862.Glazko, A. J., in Woodbury, D. M., Penry, J. K., and Schmidt, R. P., Editors, “Antiepileptic Drugs,” Raven Press, New York, 1972, p. 103. Glazko, A. J., Clin. Chem., 1974, 20, 915. Drell, W., Analyt. Biochem., 1970, 34, 142. Hagdahl, L., Williams, J. P., and Tiselius, A., A r k . Kemi, 1951, 4, 193. Asatoor, A., and Dalgliesh, C. E., J . Chem. SOG., 1956, 2291. Claesson, S.. A r k . Kemi Miner. Geol., 1946, 23A (l), 1. Herrman, B., and Pulver, R., Archs. I n t . Pharmacodyn. Thbr., 1960, 126, 454. Meola, J. M., and Vanko, M., Clin. Chem., 1974, 20, 184. Mortensen, E., Clin. Chem., 1974, 20, 1146. Kahane, Z., and Vestergaard, P., J . Lab. Clin. Med., 1967, 70, 333. Kahane, Z., and Vestergaard, P., Clinica Chim. Acta, 1969, 25, 453. Taniguchi, K., Kakimoto, Y., and Armstrong, M.D., J . Lab. Clin. Med., 1964, 64, 469. Sandhu, R. S., and Freed, R. M., Stand. Meth. Clin. Chem., 1972, 7, 231. Morris, C. J. 0. R., and Morris, P., “Separation Methods in Biochemistry,” First edition, Pitman Chalmers, R. A., and Watts, R. W. E., Analyst, 1972, 97, 958. Chalmers, R. A., in Reid, E., Editor, “Methodological Developments in Biochemistry. 5. Assay of Drugs and Other Trace Organics in Biological Fluids,” ASP Biological and Medical Press B.V., Amsterdam, 1976, in the press. and Sons, London, 1963. 5 . Assay of Drugs and Other Trace Organics in Biological Fluids,” ASP Biological and Medical Press B.V., Amsterdam, 1976, in the press. Chalmers, R. A., and Lawson, A. M., Chemy Britain, 1975, 11, 290. Chalmers, R. A., Lawson, A. M., and Watts, R.W. E., unpublished work. Davies, C. W., and Owen, B. D. R., J . Chem. Soc., 1956, 1681. Atack, C. V., and Magnusson, T., J . Pharm. Pharmac., 1970, 22, 625. Fujimoto, J. M., and Wang, R. I. H., Toxic. Appl. Pharmac., 1970, 16, 186. MulC, S. J., Bastos, M. L., Jukofsky, D., and Saffer, E., J . Chromat., 1971, 63, 289. Hetland, L. B., Knowlton, D. A., and Couri, D., Clinica Chim. Acta, 1972, 36, 473. Miller, W. L., Kullberg, M. P., Banning, M. E., Brown, L. D., and Doctor, B. P., Biochem. Mcd., Chu, C.-H., and Pietrzyk, D. J., Analyt. Chem., 1974, 46, 330. Makino, I., Shinozaki, K., Nakagawa, S., and Mashimo, K., J . Lipid Res., 1974, 15, 132. Vallon, J. J., Badinand, A., and Bichon, C., Clinica Chim. Acta, 1972, 36, 397. Beling, C. J., Adv. Obstet. Gynecol., 1967, 1, 88. Adlercreutz, H., Clinica Chim. Acta, 1971, 34, 231. Sjovall, F., Nystrom, E., and Haahti, E., Adv. Chrornat., 1968, 6, 119. 1973, 7, 145.18 103. 104. 106. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 126. 126. 127. 128. 129. 130. 131. 132. 133. 134. 136. 136. 137. 138. REID Anderson, R. A., Chambaz, E. M., Defaye, G., Madani, C., Baillie, T. A., and Brooks, C. J. W., Anderson, R. A., Chambaz, E. M., and Brooks, C. J. W., J . Chromat., 1974, 99, 485. Dailey, J. W., and AnggArd, E., Biochem. Pharmac., 1973, 22, 2591. Kesner, L., Aronson, F. L., Silverman, M., and Chan, P. C., Clin. Chem., 1975, 21, 353. O’Garman, L. P., Clinica Chim. Acta, 1969, 23, 247. Taylor, J . A., Clin. Pharmac. Ther., 1972, 13, 710. Bush, I. E., Adv. Clin. Chem., 1969, 12, 57. Fecher, R., Chanley, J . D., and Rosenblatt, S., Andyt. Biochem.. 1964, 9, 64. Broich, J. R., Hoffman, D. B., Goldner, S. J., Andryauskas, S., and Umberger, C. J., J . Chromat., Yeh, S. Y., Flanary, H., and Sloan, J., Clin. Chem., 1973, 19, 687. Reid, E., in Spencer, B., Editor, “Industrial Aspects of Biochemistry,” North-Holland Publishing Mattox, V. R., Litwiller, R. D., and Goodrich, J. E., Biochem. J., 1972, 126, 633. Schill, G., in Marinsky, J . A., and Marcus, Y., Editors, “Ion Exchange and Solvent Extraction,” Morreal, C. E., Dao, T. L., and Lonergan, P. A., Stevoids, 1972, 20, 383. Passon, P. G., and Peuler, J . D., Analyt. Biochem., 1973, 51, 618. Cameron, J. D., Clinica Chim. Acta, 1974, 56, 307. Michaelis, A. F., Cornish, D. W., and Vivilecchia, R., J . Pharm. Sci., 1973, 62, 1399. Leppard, J. P., and Reid, E., A . Rep. Prog. Chem. B , 1974, 71, 44. Landgraf,,, W. C., in Carr, C . J., and Usdin, E., Editors, “The Phenothiazines and Structurally Related Scott, C. D., Chilote, D. D., Katz, S., and Pitt, W. ’SV., jun., J . Chromat. Sci., 1973, 11, 90. Blair, A. D., Forrey, A. W., Meijsen, B. T., and Cutler, R. E., J . Pharm. Sci., 1975, 64, 1334. Dell, D., in Reid, E., Editor, “Methodological Developments in Biochemistry. J . Chromat. Sci., 1974, 12, 636. 1971, 63, 309. Co., Amsterdam, 1974, p. 913. Volume 6, Marcel Dekker, New York, 1974, p. 1. Drugs, Raven Press, New York, 1974, p. 367. 5. Assay of Drugs and Other Trace Organics in Biological Fluids,” ASP Biological and Medical Press B.V., Amsterdam, 1976, in the press. Stevenson, R. L., and Burtis, C. A., Clin. Chem., 1971, 17, 774. Kissinger, P. T., Felice, L. J., Riggin, R. M., Pachid, L. A.. and Wenke, D. C., Clin. Chem., 1974, Garland, W. A., Trager, W. F., and Nelson, S. D., Biomed. Mass Spectrosc., 1974, 1, 124. Schulten, H. R., and Games, D. E., Biomed. Mass Sjbectrosc., 1974, 1, 120. Anders, M. W., and Latorre, J. P., J . Chromat., 1971, 55, 409. Rosen, L., and Goodall, McC., Am. J . Physiol., 1962, 202, 883. Tyce, G. M., Sharpless, N. S., and Owen, C. A., jun., Biochem. Pharmac., 1972, 21, 2409. Persson, B.-A., and Karger, B. L., J . Chromat. Sci., 1974, 12, 621. Eksborg, S., Lagerstrom, P.-O., Modin, R., and Schill, G., J . Chromat., 1974, 83, 99. Khym, J. X., Clin. Chem., 1975, 21, 1245. James, H., Analyst, 1973, 98, 274. Bauer, R., Clin. Chem., 1974, 20, 917. Booker, H. E., and Darcey, B., Epilepsia, 1973, 14, 77. Joern, W. A., Clin. Chem., 1975, 21, 1548. 20, 992. Received July 16t12, 1975 Accepted August Zlst, 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100001

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

ANALAO 101 (1198) 1-72 (1976)ISSN 0003-2654January 1976THE ANALYSTTHE ANALYTICAL JOURNAL OF THE CHEMICAL SOCIETYCONTENTSREVIEW PAPERI193239444955626770Sample Preparation in the Micro-determination o f Organic Compounds in Plasmaor Urine-Eric ReidORIGINAL PAPERSThe Automatic Determination o f Silicate Dissolved in Natural Fresh Water byMeans of Procedures Involving the Use o f Either cc- or P-MolybdosilicicAcid-Victor W. Truesdale and Christopher J. SmithThe Determination o f Strontium-90and Strontium-89 in Water without Separationof Strontium from Calcium-J. G. T. Regan and J. F. C. TylerThe Determination o f Tellurium in Leaded Free-cutting Steels by Atomic-absorption Spectrometry-W. D. Cobb, W. W. Foster and T. s. HarrisonColorimetric Determination o f Phosphorus in Silicates Following Fusion withLithium Metaborate-J.B. BodkinDetermination of Benzoic and Sorbic Acids in Orange Juice-Tamar Gutfinger,Rina Ashkenazy and A. LetanApplication of the Tubular Graphite Electrode in the Measurement o f ReactionKinetics: Development o f an Automatic Technique-L. R. Sharma, V. P. Soi,J. D. Sharma and Ramesh Kumar KaliaIREPORT BY THE ANALYTICAL METHODS COMMITTEEThe Use o f 50 per cent. Hydrogen Peroxide for the Destruction o f Organic Matter(Second Report)CO M M U N ICATIO NChanges in Absorbance Values o f Solutions of Brilliant Green After Initial Dis-solution o f the Solid: Precaution t o be Taken in i t s Use as a ColorimetricReagent-A. G. Fogg, Anne Willcox and D. Thorburn BurnsBook ReviewsSummaries o f Papers in this Issue-Pages iv, v, v i i iPrinted by Heffers Printers Ltd Cambridge, EnglandEntered as Second Class at New York, USA Post Offic

ISSN:0003-2654    

DOI:10.1039/AN97601BX003

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

...Vlll SUMMARIES OF PAPERS I N THIS ISSUEDetermination of Benzoic and Sorbic Acids in Orange JuiceA procedure has been devised for the separate determination of benzoic andsorbic acids in orange juice. It is based on the steani distillation of thesepreservatives from an acidified juice into an alkaline trap. The distillate isdivided into two portions, each of which is treated with potassium dichromateand sulphuric acid, but under different conditions. Drastic oxidationleads to the destruction of sorbic acid and, after re-distillation, allows thedetermination of benzoic acid a t 225 nm (with no interference from sorbicacid). Milder oxidation selectively converts sorbic acid into malonaldehyde,and the latter is subsequently converted into a coloured compound anddetermined a t 532 nm (with no interference from benzoic acid).TAMAR GUTFINGER, RINA ASHKENAZY and A.LETANDepartment of Food Engineering and Biotechnology, Technion (Israel Institute ofTechnology), Haifa, Israel.Analyst, 1976, 101, 49-54.January, 1976Recoveries of both preservatives from orange juice were satisfactory.Application of the Tubular Graphite Electrode in theMeasurement of Reaction Kinetics : Development of anAutomatic TechniqueThe unique feature of the tubular electrode, that it permits continuous analysisof an electrolyte solution passing through it, has been made use of in develop-ing an automatic voltammetric technique for studying the kinetics ofreactions involving one or more electro-active reactants or products. Thetechnique provides for automatic dilution of the reaction mixture with thesupporting electrolyte in order to bring the concentrations within thevoltammetric range, automatic freezing of the reaction when required, auto-matic stirring of the reaction mixture solution and, especially, automaticrecording of the current -time graph, from which the kinetics of the reactionscan easily be studied.The technique has been standardised by studying voltammetrically thereaction kinetics of the oxidation of potassium iodide with hydrogen peroxidein aqueous acidic solutions and comparing the data with those obtainedtitrimetrically. The results appear to show that the automatic techniquehas a high dependability.Encouraging results have been obtained in thestudy of the reaction kinetics of the Oxidation of potassium hexacyano-ferrate(I1) with hydrogen peroxide and of sodium formate with potassiumpermanganate in acidic media.L.R. SHARMA, V. P. SOI, J. D. SHARMA and RAMESH KUMAR KALIADepartment of Chemical Engineering and Technology, Punjab University, Chandi-garh 160014, India.Analyst, 1976, 101, 55-61.The Use of 50 per cent. Hydrogen Peroxide for theDestruction of Organic Matter (Second Report)Report prepared by the Metallic Impurities in Organic Matter Sub-committee.ANALYTICAL METHODS COMMITTEEThe Chemical Society, Burlington House, London, W1V OBN.Analyst, 1976, 101, 62-66.Changes in Absorbance Values of Solutions of Brilliant GreenAfter Initial Dissolution of the Solid: Precaution to beTaken in its Use as a Colorimetric ReagentCommunicationA. G. FOGG, ANNE WILLCOX and D. THORBURN BURNSChemistry Department, University of Technology, Loughborough, Leicestershire,LEll 3TU.Analyst, 1976, 101, 67-69

ISSN:0003-2654    

DOI:10.1039/AN97601BP005

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, January, 1976, Vol. 101, pp. 19-31 19 The Automatic Determination of Silicate Dissolved in Natural Fresh Water by Means of Procedures Involving the Use of Either a- or P-Molybdosilicic Acid Victor W. Truesdale and Christopher J. Smith Institute of Hydrology, Maclean Building, Crowmarsh Gifloord, Wallingfovd, Oxfordshire, OX10 8BB This paper describes the practical application of Truesdale and Smith‘s (1975) fundamental re-appraisal of the conditions that lead to the formation of a- and 8-molybdosilicic acid in aqueous solutions. Two procedures, one using the a-acid and the other the p-acid, are proposed. Both procedures are designed so as to enable silicate-silicon concentrations of between 0 and 1.0 mg 1-1 to be determined, although slight adjustments to the manifold of the Technicon AutoAnalyzer I1 system would make both suitable for other concentration ranges.The appropriate “molybdenum blues” derived from the yellow acids are used. Tests of the precision of both procedures were made at five concentration levels between 0 and 1.0 mg 1-1 of silicate-silicon in distilled water. The maximum value of the coefficient of variation was 1.44 per cent. In other tests 27 samples of natural water from the upland area surrounding the sources of the rivers Wye and Severn were analysed by both methods. The results suggest that for these types of water the methods yield the same value for silicate concentration. Stricklandl found that two forms of molybdosilicic acid can be formed in mixtures of molybdate, acid and silicate. His discovery led to improvements in methods for silicate analysis; in particular, it demonstrated that conditions that allow mixtures of the acids to form should be avoided.Nevertheless, Stricklandl did not supply an adequate description of the conditions under which the two acids are formed. Recently,2 his belief that the critical factor that determines the form produced in any solution is the ratio of the concentrations of acid and molybdate used has been shown to be erroneous; it is now recommended2 that the primary factors to be considered are the pH and molybdate concentration of the reaction mixture. The use of an inappropriate fundamental framework with which to describe the behaviour of a system must eventually hinder, or even mislead, further investigation of the system. We believe that acceptance of Strickland‘s acid to molybdate ratio concept1 has seriously retarded progress in several branches of molybdosilicic acid chemistry, including the analytical branch.2 Perhaps the clearest example of this effect is that in which Hargi~,~ in the studies of the formation of the a-molybdosilicic acid, describes how he was compelled to adopt a contorted experimental design so as to satisfy the acid to molybdate ratio constraint.He stated that the ratio concept introduced a problem that is not encountered in otherwise analogous studies of the formation of molybdophosphoric and bismuthomolybdophosphoric acids. We believe that the evolution of methods for silicate analy~isl+-~* has been impeded for three reasons. Firstly, it has not been possible reliably to produce solutions that contain only one acid.2 Secondly, it has been impossible to describe, and perhaps even investigate, the kinetics of molybdosilicic acid in a manner that would assist the analyst.H a r g i ~ ~ ~ seems to be the only worker to have attempted such a study. However, his work was restricted to a narrow band of conditions that covered only /3-molybdosilicic acid formation and conse- quently the results are of limited direct application to analytical problems. Thirdly, studies as yet unpublished26 show that existing knowledge of the spectrophotometric characteristics of aqueous solutions of a- and p-acids is generally incorrect. In particular, in contrast to previous workers,1,4,5,23 we find that the absorptivities of the yellow acids are not equal at any wavelength between 290 and 400 nm when examined under identical conditions.It is20 TRUESDALE AND SMITH : AUTOMATIC DETERMINATION OF SILICATE Analyst, VoZ. 101 not possible, therefore, to develop an analytical procedure in which use is made of the yellow acids and a wavelength between these limits, and which becomes independent of the com- position of a mixture of the a- and ,&acids in the manner described by Garrett and Walkere4 In this paper two automatic procedures for the analysis of dissolved silicate are presented; one procedure depends upon the formation of a-molybdosilicic acid, the other on that of /3-molybdosilicic acid. An automatic a-acid method for application to fresh waters does not appear to have been proposed before.Both procedures are free from significant interference from phosphate-phosphorus at the concentrations likely to be encountered in most natural fresh waters, and both are designed to operate at room temperatures above 17.0 "C. The a-acid procedure is therefore freed from two major disadvantages that Morrison and Wilson13 encountered with their manual a-acid procedure, namely, high temperatures and phosphate interference. This study began because, in our experience, an existing automatic procedure20 for silicate analysis (by the p-acid procedure) was prone to erratic behaviour, which could not be accounted for in terms of AutoAnalyzer manifold operation. Instead, it seemed that the poor reproducibility resulted from a lack of understanding of the chemistry involved, and therefore, that further development work was required.It now seems that the study will also assist in the eventual objective of obtaining a standard method for silicate analysis. Of course, a prerequisite of a truly standard method is the knowledge that it is accurate. The accuracy of silicate procedures must remain in doubt, however, until more is known about the reactivity to acid and molybdate reagents of both interfering substances and the possible silicate species present in natural waters. The latter aspect still requires clarification, especially as it has been shown that condensation of silicate monomers can occur in some natural waters, particularly during free~ing.~',28 At this stage, therefore, we suggest that there is a real need to test the concordance of two (or more) procedures on batches of samples, whenever possible.Such a comparison is presented in this paper. Experimental Apparatus The approach described previously2 was used to study the formation of the two molybdo- silicic acids under various conditions. Absorbance was measured by means of a Hilger and Watts Uvichem 1600 spectrophotometer ; for kinetic studies the output from this instrument was displayed on a Smith's Servoscribe recorder. An EIL, Model 23A, direct-reading pH meter, standardised at pH 1.0 and 4.0, was used. The pH 1.0 standard was prepared by mixing 25 ml of 0.20 M potassium chloride solution with 67 ml of 0.20 M hydrochloric acid; the pH 4.0 standard was prepared from buffer ta,blets (Burroughs Wellcome & Co.).No corrections have been applied for differences of ionic strength between standards and other solutions. AnalaR reagents were used throughout the study and distilled water was prepared by distilling tap water in a Manesty still. Automatic analysis was performed on a Technicon AutoAnalyzer I1 system that incorporated a digit a1 read-out. Results Time Required for Formation of Molybdosilidc Acid GePzeral consideration of the kinetics of molybdosilicic acid formation For the reasons given above it should now be possible to make use of studies of the kinetics of molybdosilicic acid formation in order to improve methods of silicate analysis. Whereas the appropriate time of formation has previously had to be determined e m p i r i ~ a l l y , ~ J ~ , ~ ~ it might soon be possible to apply a rate equation that describes the reaction time precisely in terms of the relevant variables: pH, molybdate concentration, temperature, etc.In an attempt to achieve this end we have examined the kinetics of the formation process. Although the detailed results of these studies will be publkhed elsewhere, those of direct analytical application are presented here. Preliminary investigations showed that the formation curves for both acids are given by A , = A , (1 - e-kt) . . where A , and A , are the absorbances of the solution at time t and 00, respectively, and k is an apparent rate constant (k > 0). Thus, A4t approaches an asymptotic value A ,Junuury , 1976 IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS 21 exponentially, as t tends to infinity.We have tested the appropriateness of this equation more than 100 times by forming the molybdosilicic acids in mixtures containing silicate-silicon (0-3.00 mg 1-l) and molybdate-molybdenum (0-0-050 M) at appropriate pH values2 between 0.8 and 4.8. In each instance the graph depicting log,, ( A , - A,) veYsZts time was linear with a correlation coefficient greater than 0-98. From equation (1) it follows that the precise time (if) required for a given fraction (f) of reaction is - 2.303 - . (2) - 2.303 Support for the validity of equation (1) can be drawn from two sources that were not available to us at the time we performed our work. Firstly, from studies of the initial formation rates of molybdosilicic acids, H a r g i ~ ~ ~ concluded that the formation reactions are first order with respect to silicate concentration, an observation compatible with equation (1).Secondly, Garrett and Walker4 found in their investigations that graphs of log ( A , - A,) versus time were linear. In both sets of work it appears that the 13-acid was formed, although some doubt must remain as Strickland's acid to molybdate ratio was used to describe the conditions. Also, these earlier reports, as well as being limited to @acid formation, covered a much narrower range of pH and molybdate concentrations than has been tested in this work. Sfiecifc considerations In order to obtain rapid formation of the molybdosilicic acids during analysis, the pH of the reaction mixtures used in the 13- and a-acid procedures should be kept as close as possible to 1.8 and 4.0, respectively.There are two reasons for this requirement. Firstly, the pH limits for the a- and ,%acid production are 4.0 and 5-0 and 0.8 and 1.8, respec- tively.2 Secondly, as our unpublished studies of the formation kinetics have shown, between the aforementioned limits the formation rate of the /3-acid increases with increasing pH and the formation rate of the a-acid decreases with increasing pH. In order to accommodate the required buffering capacity as well as a rapid formation of molybdosilicic acid, we have chosen to allow the pH of the reaction mixture to vary between 1.4 and 1.8 in the p-acid procedure, and between 4.0 and 4.2 in the a-acid procedure. Accord- ingly, the maximum time required for formation of molybdosilicic acid has been found by using equation (2) and the apparent rate constants obtained at pH values of 1.4 and 4.2 for /3- and a-acid procedures, respectively.A molybdate concentration of 0 . 0 5 0 ~ in a reaction mixture has been adopted in the recommended procedures so as t o take advantage of a high rate of molybdosilicic acid forma- tion. Our unpublished studies show that the rate constant for 13-molybdosilicic acid formation rises rapidly over the range of molybdate concentrations between 0 and 0.025 M. There- after, the rise in rate constant for each increment of molybdate becomes progressively smaller; at 0.050 M it is less than 1/16th of the rise obtaining at low molybdate concentrations. The rate of formation of the a-acid behaves in a similar manner when the molybdate concentration is changed in this way.We have not used still higher concentrations of molybdate in the reaction mixture because Stricklandl has shown that they can produce difficulties during the reduction of the molybdosilicic acid with tin( 11) chloride. For our purposes a procedure that could accommodate a concentration range of 0-1 mg 1-1 in the sample was required. In this range the maximum resolution that can be expected from the AutoAnalyzer is approximately 2 in 1000 pg 1-l. (This constraint is imposed by both the thickness of the line the recorder draws and the size of the chart used.) Accordingly, a reaction time that yields 99.8 per cent. of the total yield of molybdosilicic acid (f= 0.998) was adopted. Measurements of the kinetics of molybdosilicic acid formation at pH values 1.4 and 4.2, temperature 17-0 "C (lowest likely operating temperature) and molybdate con- centration equal to 0 . 0 5 0 ~ showed that the minimum reaction periods for a- and p-acid procedures operating under these conditions are 10.6 and 4.7 min, respectively.The effect of sodium chloride upon the rate of formation of the molybdosilicic acids has been studied. The presence of 100mg1-1 of sodium chloride in the reaction mixtures in which the molybdosilicic acids are formed (molybdate-molybdenum concentration 0.050 M) had an insignificant effect on the rate constants (Table I). The small differences of up to 6 per cent. can be accounted for by slight variations in pH of the reaction mixtures. After22 TRUESDALE AND SMITH : AUTOMATIC DETERMINATION OF SILICATE Analyst, Vol.101 allowing for the amount that the sample contributes to the reaction mixture it is found that the a- and P-acid procedures will tolerate at least 363 and 535 mg 1-1 of sodium chloride, respectively, in the sample. TABLE I EFFECT OF EITHER SODIUM CHLORIDE OR PHOSPHATE ON THE RATE OF FORMATION OF THE M0LYE;DOSILICIC ACIDS Reaction mixture composition : 0.050 M molybdate-molybdenum (pH, 3.9-4.0 for the a-acid ; 1.5-1.6 for the p-acid); the u-acid mixture was 0.1 N 1.n sodium acetate and 0.1 N in acetic acid. Concentration of Concentration of Apparent rate constantlmin-1 sodium chloride phosphate-phosphorus -7 added/mg I-' added/mg 1-l 0 0 0.94, 0.93 1.58, 1-57 100 0 0.91, 0.89 1-55, 1.50 0 4-0 0.89, 0.91 1-49, 1.49 Buffering A fixed set of reagents and mixing regimes is used in each automatic method. It is essential, therefore, to ensure that likely variations in the acidity (alkalinity) of the samples does not give rise to intolerable pH levels in the reaction mixture.Buffering is desirable and probably essential in the a-acid procedure in which the concentration of acid in the reaction mixture is low (pH approximately 4.0). The j?-acid procedure is inherently less sensitive to changes in acidity or alkalinity of the samples because the acid concentration of the reaction mixture is high (pH approximately 1.5). Therefore, with most natural waters a pH buffer is not necessary in the j?-acid procedure. However, if for any reason a greater range of acidity (alkalinity) is anticipated additional buffering might be necessary; dichloroacetic acid (pKa = 1.5) seems likely to be suitable.In the present work, the reaction mixture of the a-acid procedure is buffered at pH 4-0. There are two reasons for choosing this pH value, which is at the lower end of the range appropriate for production of the a-acid at room temperature. Firstly, it is possible to take advantage of the higher rates of formation of the a-compound that occur at this pH. The reaction period of 10.6 min can easily be accomrnodated on an AutoAnalyzer and its use circumvents the need for the heating that has been recommended in some manual a-acid methods.5A13 Secondly, both alkaline and acidic natural waters can be accommodated in a single procedure if the pH of the buffer is biased towards increased acidity in this way.The buffering capacity of 1 1 of the reaction mix:ture with the recommended composition is such that addition of 1.4 x lob2 equiv of acid or alkali changes the pH by 0.2 unit. In the recommended automatic procedure, this pH change wiU therefore be produced by a sample containing 5.1 x loe2 equivl-1 of acid or alkali. This buffering capacity appears to be sufficiently large to cope with the variable acidity (alkalinity) of most natural waters. Titrations of 0.050 M molybdate solution with 0.5 N hydrochloric acid showed that 4.8 x 10-2 equiv of acid are required to change the pH of 1 1 of the reaction mixture for the j?-acid procedure from 1.4 to 1.8. The /3-acid procedure recommended here can therefore tolerate natural waters with alkalinities (to pH 1-4) of approximately 0.24 equiv 1-1.This buffering capacity seems to be sufficiently large for most natural fresh waters. Reduction Several different reducing agents have been used in conjunction with molybdosilicic acids. We found that tin(I1) chloride is an excellent reagent for automatic analysis provided that its exposure to air is limited, and for this purpose we prefer it to the metol(9-methylamino- phenol sulphate) - sulphite reducing agent used by Brewer and Riley15 because the latter requires high temperatures and a lengthy reaction period. It is also preferable to the 1- amino-2-naphthol-4-sulphonic acid reagent because, as Morrison and Wilsonll have shown, the absorptivity of the a-acid product reduced with tin(I1) chloride is 20 per cent.higher. N in the final reaction mixture is satisfactory, and the yield of reduction products obtained by the use of 1.0 x 10-3 and 2.0 x 1 0 - 3 ~ concentrations is probably similar (Table 11). Although with these concentrations the absorbances (at 660 nm for a- and P-acids, 740 nrn for the a-acid and 790 nm for the P-acid) of the blank solutions (0.0 mg 1-1 of silicate-silicon) were identical, the reaction mixtures A tin(I1) chloride concentration of 1.5 XJanuary, 1976 23 were of different colours. Whereas with 2.0 x N tin(I1) chloride the blank for /3-molybdo- silicic acid was an intense golden colour, that with 1.0 x N tin(I1) chloride was only faintly brown. Similarly, the blanks for a-molybdosilicic acid were of noticeably different shades of green.IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS TABLE I1 EFFECT UPON THE ABSORBANCE OF THE FINAL MIXTURE OF CHANGING THE TIN (11) CHLORIDE CONCENTRATION Distilled water was used in the reference beam of the spectrophotometer. Reduced 8-molybdosilicic acid Reduced a-molybdosilicic acid - r Tin(I1) chloride concentration*/N 1.0 x 2.0 x 10-3 1.0 x 10-3 2.0 x 10-3 Absorbance at 660 nm (l-cm cell) Absorbance at 790 nm (l-cm cell) Absorbance at 740 nm (l-cm cell) Incrementt 0.365 0.366 0.52 1 0.619 Blank 0.016 0.023 0,011 0.019 Incrementt 0.712 0.701 - - Blank 0.028 0.031 - - Incrementt - - 0.624 0.620 Blank - - 0.010 0-018 * In each instance, a 1.00- or 2.00-ml aliquot of 0.050 N tin(I1) chloride solution was added to a mixture of 26-0 ml of a molybdosilicic acid solution (2.00 or 0 mgl-1 of silicate-silicon) and 26.0 ml of 2.0 N sulphuric acid, which had stood for 2 min.When necessary extra distilled water was added to compensate for differences in volume. t The increment in absorbance for 2.00 mg 1-1 of silicate-silicon. With the recommended conditions the reduction of /3-molybdosilicic acid is almost complete within a few seconds at 17 "C. Thereafter, a slight decrease in absorbance (at 660 or 790 nm) of approximately 1 per cent. min-1 is observed for at least 5 min. In contrast, the reduction of the a-molybdosilicic acid is a slower process. Nevertheless, after 2 min the rate of increase in absorbance (at 660 or 740nm) was less than 1 per cent. min-l. Therefore, the periods allowed for reduction, 1-5 and 2.8 min in the Is- and a-acid procedures, respectively, are long enough.Further, slight variations in these times produced by the AutoAnalyzer will not affect the reproducibility significantly. The stability of the tin(I1) chloride reagents has been investigated, and the stock 1.0 N solution was found not to deteriorate during storage for 6 months in a stoppered bottle. Although 0-01 N tin(I1) chloride solution, prepared by diluting a stock of 1.0 N solution [in concentrated hydrochloric acid (11-2 N)] with distilled water, did not deteriorate during storage for 50 h in stoppered bottles, it deteriorated markedly when stored in an open beaker for 24 h. The tests were conducted using both ct- and /3-molybdosilicic acids.During the experiments sub-samples of a reservoir (1 1) of yellow a-molybdosilicic acid solution (approxi- mately 2 mg 1-1 of silicon) that were treated with 2 N hydrochloric acid and reduced with tin(I1) chloride from a stoppered bottle returned a mean absorbance (740nm, l-cm cell) of 0.618 with a standard deviation of 0.002. On the 50th hour, tests were also performed on a /?-molybdosilicic acid solution using both the 50 h old batch of 0.01 N tin(I1) chloride solution as well as a batch prepared from a fresh 1.0 N stock solution. The absorbances (810 nm, l-cm cell) of the reduced products were identical. The above results show that the 0.01 N tin(I1) chloride reagent must not be stored in an open beaker. Therefore, in order to avoid deterioration of the reagent during routine work we store it in a partially sealed 100-ml standard flask.The solution is drawn through a glass tube that passes through a rubber bung to the bottom of the flask; compensating air enters the flask through a second fine-bore tube. This storage regimen has proved to be reliable during extensive usage of both a- and /3-acid procedures; the response obtained from standard silicate solutions has not decreased systematically during any one day's analysis. Interference by Phosphate Phosphate-phosphorus can combine with molybdate to produce molybdophosphoric acid. This reaction takes place spontaneously under the conditions used in the procedures recom-24 TRUESDALE AND SMITH : AUTOMATIC DETERMINATION OF SILICATE Anulyst, VoZ. 101 mended here and can interfere (Table 111).With the /3-acid procedures this interference has previously been overcome by destroying the molybdophosphoric acid by addition of oxalic or tartaric acid. However, Morrison and Wilson13 state that they were unable to overcome the interference in their manual a-acid method by this means. In contrast, Grasshoff21 succeeded in overcoming the interference in his cc-acid procedure for sea water by adding oxalic acid. Both of the procedures presented here have been freed from most of the inter- ference from up to 7.0 mg 1-1 of phosphate-phosphorus in the samples. TABLE I11 SUPPRESSION OF PHOSPHATE INTERFERENCE BY MEANS OF OXALIC ACID: THE OVER-ALL EFFECT In each instance 3.0 ml of 50.0 g 1-l oxalic acid solution (or 3.0 ml of distilled water) were added t o a mixture of 26.0 ml of molybdosilicic acid (2.00 or 0.00 mg 1-1 of silicon) and 25.0 ml of 2.0 N sulphuric acid.After 2 min 1-00 ml of 0.050 N tin(I1) chloride solution was added to give the molybdenum blue. Aliquots (25.0 ml) of molybdophosphoric acid (2.00 or 0 mg 1-1 of phosphorus) were treated similarly. Absorbance (l-cm cell) I 7 u.-Acid procedure 8-Acid procedure 660 nm 740 nm 660 nm 790 nm Sample identity - * Blank . . .. .. .. . . 0.008 0.007 0.009 0.009 Silicate solution (2.00 mg 1-1 of Si) . . 0.502 0.602 0.323 0.688 Increments Phosphate solution (2.00 mg 1-l of P) 0.558 0.588 0.547 0.483 Silicate + phosphate . . .. .. 1.110 1.218 0.858 1.155 Blank .. . . .. .. .. 0.016 0.012 0.002 0.001 Silicate solution (2.00 mg 1-1 of Si) .. 0.475 0.588 0.32 1 0.690 Silicate + phosphate . . .. . . 0.471 0.589 0.320 0.681 Without oxalic acid With oxalic acid { Phosphate solution (2.00 mg 1-l of P) 0.014 0.012 0.004 0.00 1 In our exploratory experiments on the suppression of phosphate interference, approximately 1.0 mg 1-1 of phosphate-phosphorus was maintained in the complete reaction mixture used for producing the blue, reduced molybdosilicic .acids. The molybdosilicic and molybdo- phosphoric acids were allowed 10 min to develop at either pH 4.0 or 1.7 and room temperature (approximately 20 "C). Whereas the addition of 3.0 g 1-1 of oxalic acid (in the final mixture) changed the blanks only slightly, it affected the increment in absorbance yielded by a given amount of silicate-silicon. The increment in abs'orbance obtained in the a-acid procedure decreased by approximately 6 per cent.(Table 111). The reason for this effect is not known although several experiments, including some in which the absorbance (at 390 nm) of the yellow a-molybdosilicic acid was studied, were performed. Nevertheless, the experiments showed that the effect, such as it is, occurs rapidly after the oxalic acid is added and that afterwards the increment does not change significantly (Table IV). The behaviour in the ,%acid procedure is different. The increment in absorbance yielded by a given amount of silicate varies with the period allowed to elapse between the addition of oxalic acid and tin(I1) chloride. However, it appears (Table IV) that the addition of oxalic acid does not affect the magnitude of the initial value of the increment:. As the absorbance, at both 660 and 790 nm, of the blue derivative decreases with increas'ed time of exposure to oxalic acid (Table IV), it seems that some of the /3-molybdosilicic acid decomposed; transformation of P-acid into a-acid would have led to an increase in the absorbance at 660nm.Further investi- gation of these phenomena was suspended as they appear to be under control in the automatic procedures recommended here ; the procedures yield reproducible results as well as linear calibration graphs. Notwithstanding the existence of the above effects, the results in Tables I11 and IV suggested that oxalic acid can suppress the tested phosphate interference. Other tests showed that with only 2.0 g 1-1 of oxalic acid in the final mixture, some interference remained after 2 min in both methods.It seemed advisable, therefore, to use the higher concentration of 3.0 g 1-l. Although the above experiments suggested that :phosphate interference had been suppressed satisfactorily, tests with the automatic a-acid and /3-acid procedures demonstrated a residualJanuary, 1976 IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS 25 TABLE IV SUPPRESSION OF PHOSPHATE INTERFERENCE BY MEANS OF OXALIC ACID : EFFECT OF VARYING THE TIME OF EXPOSURE OF THE MOLYBDOSILICIC ACIDS TO OXALIC ACID In each instance 10.0 ml of 100 mg 1-' silicate-silicon.solution (or 10 ml of distilled water in the case of a blank) were added to 500 ml of acidified molybdate (0.050 M molybdenum) solution at either pH 1.6 (for p-acid) or pH 4.0 (for cc-acid).Each 510-ml aliquot was allowed to stand for 10 min before 500 ml of 2 N sulphuric acid were added and the solution divided into two 500-ml portions. One portion was treated with 30.0 ml of oxalic acid (50 g 1-I) and the other with 30.0 ml of distilled water. After various periods of time, 26.5-ml aliquots of each mixture were treated with 0.5 ml of 0.05 N tin(I1) chloride solution and the absorbance (l-crn cell) of each solution was measured. The mixtures that were not treated with oxalic acid are controls; with the p-acid they show the rate of transformation to the a-acid. a-Acid I A \ Absorbance Exposure timelmin 660 nm 740 nm With oxalic acid- 4 0.440 0.539 10 0.442 0.539 15 0.443 0.542 20 0.442 0.539 25 0.442 0.540 30 0.444 0.539 35 0.444 0.540 40 0.446 0.540 45 0-448 0.542 50 0.451 0.539 Without oxalic acid- 2 0.467 0.552 23 0.462 0-548 42 0.468 0-55 1 p-Acid I L v Exposure I-A-> timelmin 660 nm 790 nm Absorbance 5 12 15 20 25 30 35 40 45 50 55 60 83 0.327 0.322 0.320 0.318 0.315 0.315 0-311 0.3 10 0.310 0.309 0.307 0.305 0.298 0.673 0.653 0.644 0.634 0.62 1 0.613 0.600 0.591 0.582 0.582 0.570 0.559 0.522 10 0.333 0.667 18 0.330 0.663 42 0.333 0.641 0.628 62 0.332 interference when the above amounts of oxalic acid were used. The interference was greater in the cc-acid procedure, where samples of distilled water containing 0, 1, 2, 4 and 7 mg 1-1 of phosphate-phosphorus produced a response equivalent to 0, 120, 170, 190 and 200 pg 1-1 of silicate-silicon, respectively.The non-linear relationship between response and concentration excludes the possibility of silicate contamination of the phosphate reagents. Moreover, that argument is also untenable as phosphate solutions prepared from potassium dihydrogen orthophosphate, dipotassium hydrogen phosphate and sodium hexametaphosphate gave similar responses. Impurities of another type, however, cannot be dismissed. It is possible, therefore, that the behaviour observed is limited to certain phosphate reagents and that the phosphate in natural waters would not create the problem during analysis. Nevertheless, as the possibility of an interference by naturally occurring phosphate could not be excluded, it became essential to suppress the residual interference as far as possible. This suppression was accomplished by adding phosphate to the molybdate reagent and increasing the oxalic acid concentration above that used in the exploratory tests.The increase is approximately 6-fold for the a-acid procedure and 2-5-fold for the /3-acid procedure. With this arrangement, up to 7 mg 1-1 of phosphate-phosphorus in a sample produces a response in both procedures equivalent to less than 3 pg 1-1 of silicate-silicon. We also found that our earlier routine use of a detergent (Ultra-wet 60-L) with the a-acid procedure was neither necessary nor desirable with the later arrangement of reagents. The later arrangement of reagents seemed to possess its own detergent action and very smooth recorder traces were obtained. The inclusion of Ultra-wet 60-L detergent in the molybdate reagent led to intolerable disruption of the recorder traces. The nature of the residual compound responsible for the phosphate interference is still unknown to us.Kinetic studies showed that under the conditions used in the procedures the rate of formation of cc- and 13-molybdosilicic acids is not changed significantly by the presence of26 TRUESDALE AND SMITH : AUTOMATIC DETERMINATION OF SILICATE Analyst, VoZ. 101 the amount of phosphate-phosphorus used in the above experiments. In each instance (Table I) the rate constants agreed to within 6 per cent., the difference being probably due to slight variations in the pH of the reaction mixtures. Therefore, competition between phosphate and silicate for molybdate, as apparently observed by Wilson,14 does not occur in these procedures.Presumably this is so because, here, a relatively high concentrationlo of molybdate-molybdenum has been used in the reaction mixture. Only fresh solutions of phosphate, prepared from disodium hydrogen orthophosphate, were used in the above tests. It was found that w:hen stock phosphate solutions are stored in glass bottles they become contaminated with silicate. In studies of these interferences such contamination can lead to erroneous conclusioiis if its presence is not suspected. Sample Blank When samples contain different amounts of background material it is essential to make a correction. Ideally, for a given sample this correction would be obtained by analysing a second aliquot from which all silicate had been removed but was otherwise of exactly the same composition as the parent material; in this way the correction would compensate for chemical interferences as well as background coloration.Unfortunately, such treatments for the removal of silicate do not exist and alternative approaches, which compensate only for background coloration, have to be used. Here, it is recommended that sample blanks be obtained by repeating the analysis under exactly the same conditions, except that distilled water is used in place of the molybdate reagent. By this means, the formation of molybdo- silicic acid is prevented, and the conditions under which background coloration is measured closely resemble those under which the reduced molybdosilicic acid concentration is measured.Filtration In order to avoid introducing a significant turbidity blank into the analysis samples should be filtered. In this work samples were filtered through Millipore membrane filters, of 0.8 pm average pore diameter, which were mounted on a IMillipore Swinnex filter holder, the top of which had been modified to accept a polypropylene tube that was 3.5 cm in diameter and 12 cm long, into which the sample was poured. Provided that the membranes were washed before filtration commenced, standard silicate solutions (0 and 1-00 mg 1-1 of silicon) filtered through this apparatus did not suffer signiiicant deterioration ; the apparatus seems to be free from the kind of contamination problem that Liss and Spencer22 encountered with some other types of filtration apparatus.However, when the membrane filters were not washed, 15-ml aliquots of the standard solutions became contaminated with approximately 10 pg 1-I of silicate-silicon. Distilled Water If the distilled water used for calibration purposes contains a significant amount of dissolved silicate-silicon the results for the analyses will be systematically too low. In order to prevent such errors the distilled water should be repeatedly re-distilled in an all-metal still until the results for the analyses carried out with water from consecutive distillations are in agree- ment. We have applied this approach here and have found that a single distillation of the relatively hard Wallingford water was sufficient. Precision The precision of each procedure was tested at five concentration levels between 0 and 1-00 mg 1-1 by analysing five sets of 11 replicate samples of silicate-silicon in distilled water.The sets were taken in order of increasing concentration and were not randomised; each estimate of precision was therefore obtained under conditions which probably slightly favour a low standard deviation. The tests for the procedures were made on consecutive days, while the tests for any one procedure were completed within 3 h. The results obtained (Table V) show that the precision of the two procedures is similar and within the levels usually expected of automatic methods that involve the use of the Technicon AutoAnalyzer I1 system. The maximum value of the coefficient of variation, and therefore the worst precision, was obtained during an analysis of the 200 pg 1-1 sample by the a-acid procedure.Jannzcary, 1976 27 It is unlikely that the procedures would be operated with maximum range expansion as greater sensitivity could be more reliably obtained by a slight increase in the sample flow in the manifold. Nevertheless, the tests conducted here, with maximum range expansion (Table V), show that both procedures will tolerate range expansion without a serious loss of precision, Calibration Graph The calibration graphs for the range 0-1-0 mg 1-1 are linear (Table V).With the 660-nm interference filters used in this work, the absorbance range extends between 0 and approxi- mately 0.25. The use of 740-nm and 810-nm interference filters for the a- and P-acid pro- cedures, respectively, would allow a greater absorbance range to be spanned.2 In general, the use of these higher wavelengths has the further advantage that “sample blank’’ values are reduced.TABLE V PRECISION OF a- AND )6-ACID PROCEDURES IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS (a) At recommended scale expansion (100 chart divisions = 1000 pg 1-l) ; absorbance approximately 0 to 0.25. a-Acid procedure ,!?-Acid procedure Sample concentration/pg 1-1 M e a n m n t of Mean-nt of of silicate-silicon (chart divisions) variation, per cent. (chart divisions) variation, per cent. 200 400 600 800 1000 19.4 39.2 59.4 79.3 97.3 1.44 0.66 0.45 0.33 0.51 19.5 39-1 59-2 79.4 99.0 0.72 0.46 0.27 0.15 0-43 Linear regression of the calibration graphs gave Y a = 0.098~ + 0.16 where Y a and Yp are the responses (chart divisions) for a- and ,!?-acid procedures, respectively, and x is the silicate-silicon concentration (pg 1-l). In both instances analysis of variance showed no evidence for non-linear components in the linear regression.Yp = 0.100~ - 0.55 (b) With maximum scale expansion, i.e., absorbance approximately 0 to 0.05. a-Acid procedure ,!?-Acid procedure Full scale equivalent to . . .. . . .. 215 pg1-1 254 pg 1-1 Mean response for 200 pg 1-1 (chart divisions) 93.0 78.7 Coefficient of variation, per cent. . . .. 2.4 0.56 Base-line stability With the a-acid procedure we have been unable to detect any systematic drift in the base-line. However, the base-line and peaks of the @-acid procedure were found to drift at approximately 3 chart units h-l (in absorbance, 0.006 per hour).We believe that the drift is caused by changes in the acidified molybdate solution; tests showed that the other reagents were not responsible for the drift. It seems that it could be eliminated by using separate acid and molybdate reagents. However, we prefer to tolerate this low rate of drift and thereby take advantage of the precise pH control afforded by a single reagent. Experience has shown that it is inadvisable to substitute hydrochloric acid for the recom- mended sulphuric acid. When we made this substitution, it was found that both the base-line and the peaks drifted excessively for 1 h or more. We are unable to offer a complete ex- planation for this phenomenon but believe that it is caused by an interaction between hydro- chloric acid and the pump-tube material; earlier work2 had shown that this change in mineral acids does not change the amount of molybdosilicic acid formed., Comparison of the Procedures The a- and /3-acid procedures were compared by analysing 27 filtered samples of water from the upland area that surrounds the sources of the rivers Severn and Wye; each sample was analysed in triplicate.The pH values of the samples ranged between 3.2 and 5.0 and some samples that were taken from peat bogs contained appreciable amounts of humic materials (about 50 mg 1-1). A typical analysis of the samples gave the following results:28 TRUESDALE AND SMITH : AUTOMATIC DETERMINATION OF SILICATE Analyst, VoZ. 101 sodium 3.0 mg l-l, potassium 0.5 mg l-l, calcium 0.35 mg l-l, magnesium 0.44 mg l-l, chloride 2.0 mg 1-1 and conductivity 70 pi2-l cm-? In order to accommodate the samples, full-scale deflection of the AutoAnalyzer recorder was fixed at 2.0 mg 1-1 of silicate-silicon.The results of the silicate analyses are given in Table VI. (As no attempt was made to “round-off” TABLE VI RESULTS AND ANALYSIS OF VARIANCE OF RESULTS OBTAINED DURING COMPARISON OF THE a- AND ,&ACID PROCEDURES USING UPLAND RIVER WATERS (a) Individual results Sample No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Procedure & &Acid 78 76 76 1474 1472 1476 1080 1074 1076 1054 1058 1042 616 622 626 1388 1374 1378 1524 1540 1560 104 98 96 1188 1200 1200 818 824 826 402 400 392 1442 1438 1460 278 260 2 64 870 880 886 j3-Acid 70 72 70 1438 1450 1452 1066 1064 1058 1030 1038 1038 620 626 626 1386 1390 1390 1546 1558 1562 124 118 118 1204 1214 1216 840 838 836 412 410 406 1452 1456 1458 284 274 270 862 872 872 (b) Analysis of variance Variation Samples .. .. .. Procedures . . .. .. Concentrations in pg 1-1. Difference, between means, Sample per cent. N 0. 4.0 0.94 0-65 0.76 0.2 1 0.31 0.23 9.4 1.3 0.92 1.4 0.30 1.6 0-57 .. .. .. .. Procedures x samples interaction Residuals (between triplicate variation) . . Total .. .. .. .. .. 15 16 17 18 19 20 21 22 23 24 25 26 27 Degrees of freedom .. 26 .. 1 .. 26 .. 108 .. 161 Procedure & &Acid 634 632 632 922 930 928 1230 1230 1230 1026 1026 1024 550 544 542 994 1002 1006 606 598 596 1356 1352 1360 648 636 634 1384 1392 1390 1530 1548 1542 1158 1152 1146 1282 1284 1282 P-Ac‘id 632 62 8 626 922 922 924 1224 1224 1220 1022 1022 1020 550 648 548 1000 1000 1000 600 596 592 1362 1356 1356 638 634 628 1382 1384 1386 1632 1542 1544 1146 1138 1134 1276 1274 1276 Mean squares 1 136 768 25.284 183.22 31.46 Difference between means, per cent.0.32 0.22 0.30 0.20 0.30 0.03 0.33 0.06 0.47 0.17 0.02 0-66 0.29 F-ratio 6204 0.14 5.82 -January, 1976 29 the figures until statistical analysis had been completed, the results given in the table must not be taken to imply extraordinarily good precision.) IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS A “mixed effects’’ statistical model of the form Yijh = p + si + “j + (sm)ij + Eijk was used to describe Yijh, the value of the kth determination on the ith sample by the j t h procedure in terms of an over-all mean p, sample and procedure effects Si, mj (i = 1 , 2 .. .27; j = 1,2), an interaction term (Sm)ij, and a residual expressing the deviation appropriate to the kth determination. Samples and procedures were regarded as random and fixed factors, respectively, in the senses used by S~heffd.,~ The analysis of variance is shown in Table V where the appropriate test of the hypothesis that procedures yield identical determinations is given by the variance ratio (procedures mean square divided by samples x procedures mean square). This analysis shows no significance so that the data are consistent with the hypothesis of no difference between results returned by the procedures. An F-test of the procedures x samples mean square against the mean square between triplicate analyses shows that there is significant (P < 0.001) procedures x samples inter- action, thus demonstrating that the difference between means of triplicate determinations, obtained for each procedure, varies from sample to sample.Inspection of the differences between triplicate means shows, however, that 25 of the 27 samples gave very similar means for silicate content by the two procedures (differing by less than 1.6 per cent.); the bulk of the significant interaction is caused by two “outlier” samples, the differences between the triplicate means of which.amount to 4.0 and 9.4 per cent. As there was no objective reason for discounting them, apart from their low silicate concentration, they were included in the whole analysis although their contribution to the interaction is disproportionately large.Taken over all samples, the mean silicate determinations by the tc- and /3-acid procedures were 950 3 pg 1-1 and 949 3 pg l-l, respectively. Analytical Methods Reagents for or-Molybdosilicic Acid Procedure Standard silicate solution (0.500 g I-1 of silicon). Fuse 1.0696 g of silica (SiO,) with 5-0 g of sodium carbonate in a platinum crucible. Dissolve the cooled melt in distilled water and make the total volume up to 11. This stock solution should be stored in a polythene vessel and aliquots diluted with distilled water to give working solutions. Bufered molybdate reagent. To 12-2 g of AnalaR ammonium molybdate [(NH,), Mo702,. 4H,O] add 138 ml of a solution containing 136.0 g 1-1 of sodium acetate trihydrate and 58 ml 1-1 of glacial acetic acid (this latter solution is 1.0 M in both sodium acetate and acetic acid). Add distilled water to make the volume up to 950 ml.By adding 5-0 N sulphuric acid (approximately 12 ml) adjust the pH of the mixture to 4.0 using a pH meter. Add 25.0 ml of 100 mg 1-1 phosphate-phosphorus solution (0.439 g 1-1 of potassium dihydrogen orthophos- phate). Filter the solution through a Millipore membrane (average pore diameter 0.8 pm). As a precipitate forms in the solution after about 24 h, it should be prepared immediately before use. Sulphwic acid - oxalic acid reagent. Dissolve 50.0 g of oxalic acid in approximately 900 ml of distilled water. Add 78.5ml of concentrated sulphuric acid (sp. gr. = 1.84) and, after cooling, add distilled water to make the volume up to 1 1.Prepare 100 ml of 1.0 N stock solution by dissolving 11.0 g of tin(I1) chloride dihydrate (SnC1,.2H20) in concentrated hydrochloric acid. This solution is stable for at least 6 months when stored in a stoppered bottle. When required, prepare 100 ml of a 0.01 N working solution by diluting 1 ml of the stock solution with distilled water. This working solution is stable for at least 3 d when stored in a stoppered bottle. Nevertheless, we recommend that it should be prepared freshly each day and stored in the flask described above. Add distilled water to make the volume up to 11. Tin(l1) chloride reducing agent. Reagents for ,&Molybdosilicic Acid Procedure Acidi$ed molybdate reagent. Dissolve 10.8 g of ammonium molybdate [ (NH,),Mo70,,.4H,0] in approximately 950 ml of distilled water.Using a pH meter, adjust the pH of the solution to 1.40 by adding 5 N sulphuric acid (approximately 20 ml). Add 20.0 ml of 100 mg 1-130 TRUESDALE AND SMITH: AUTOMATIC DETERMINATION OF SILICATE Analyst, VoZ. 101 phosphate-phosphorus solution (0.439 g 1-1 of potassium dihydrogen orthophosphate). Add distilled water to give 1 1 of solution. Filter the solution through a Millipore membrane (average pore diameter 0.8 pm) before use. As a precipitate forms in this solution after a few days, it should be prepared freshly each day. Oxalic acid - sulphuric acid reagent. Dissolve 35-0 g of oxalic acid in approximately 700 ml of distilled water. Add 140 ml of concentrated sulphuric acid (sp.gr. 1434)) cool and make the volume up to 1 1 with distilled water. Tin(I1) chloride reducing agent and silicate standaitd. As described above for the a-molyb- dosilicic acid procedure. 1 Procedure Assemble the AutoAnalyzer I1 manifold according to the design given in Fig. 1 and specifications shown in Table VII. Ensure that the temperature of reagents and samples is higher than 17.0 "C. While pumping the reagents and distilled water as sample (sampler L. Waste Q Sampler v - Mixing Mixing Mixing coil C coil B coil A I ;t-t I Waste +-- Wash Sample Acidified molybdate reagent Air Sulphuric acid-oxalic acid Tin(l I) chloride reagent Colorimeter 5-cm flow cell 660-nm filters Recorder Printer Fig. 1. Design of AutoAnalyzer I1 manifold (see Table VII for specifications). TABLE VII THE AUTOANALYZER 11 MANIFOLD DESIGN USED I N THE a- AND /3-ACID PROCEDURES Pump tube (clear standard) diameterlin Wash .. .. .. .. * . Sample . . . . . . .. .. Acidified molybdate reagent . . Tin(I1) chloride reagent . . .. .. Air . . . . .. .. .. Sulphuric acid - oxalic acid reagent . . Return line from colorimeter . . Samples per hour .. .. .. .. .. Sample to wash ratio Time of mixing in coilslrnin . . A .. .. .. .. .. B .. . . .. .. .. C .. .. .. .. .. A .. .. .. .. .. B .. .. . . . . .. C .. .. . . . . . . Glass coils used in this study a-Acid procedure .. 0-073 .. 0.020 .. 0.035 .. 0.025 . . 0.030 .. 0.020 .. 0.045 .. 8 : 1 .. 40 .. 10.6 .. 1.3 .. 2.8 . . 6 x 10turns . . 1 x 10 turns . . 2 x 10 turns + 1 x 5turns P-Acid procedure 0.073 0.025 0-051 0.025 0.030 0.025 0.056 5 : 1 60 4.7 1.8 1.5 5 x 10turns 2 x 10 turns 1 x 10 turnsJanuary, 1976 31 probe in wash position), set the absorbance at zero. When the base-line hasstabilised (approxi- mately 30 min are required in each instance), introduce samples and standards at the rate of 40 and 60 per hour for the a- and /3-acid procedures, respectively.The sample to wash ratios, in the same order, are 8: 1 and 5: 1. A few minutes after the last sample has entered the AutoAnalyzer, substitute distilled water for the molybdate reagent and wait for a new setting of the base-line to be established; adjust the recorder to zero absorbance. Pass the samples through for a second time in order to obtain a peak that gives a measure of the sample blank. Read off the peak heights of the standards, samples (Ps) and sample blanks ( P b ) after correcting for any slight drifting of the base-line in either of the two traces obtained. Calculate the gradient of the calibration graph (m) and then obtain each silicate-silicon concentration as follows IN FRESH WATER BY PROCEDURES INVOLVING MOLYBDOSILICIC ACIDS [Si] = (PS - Pb) pg 1-1 m where P, and P b are expressed in chart units and m in chart units p g l 1.Thanks are due to Mr. A. G. P. Debney for his encouragement and support, Mr. R. T. Clarke for assistance with the statistical treatments and Dr. J. S . G . McCulloch for his criticism of the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. References Strickland, J. D. H., J . Am. Chem. Soc., 1952, 74, 862, 868 and 872. Truesdale, V. W., and Smith, C. J., Analyst, 1975, 100, 203. Hargis, L. G., Analyt. Chem., 1970, 42, 1497. Garrett, H. E., and Walker, A. J., Analyst, 1964, 89, 642. Ringbom, A., Ahlers, P. E., and Siitonen, S., Analytica Chim. Acta, 1959, 20, 78. Anderson, L. H., Acta Chem. Scand., 1958, 12, 495. Luke, C. L., Analyt. Chgm., 1953, 25, 148. Hill, U. T., Analyt. Chem., 1949, 21, 589. Baker, P. M., and Farrant, B. R., Analyst, 1968, 93, 732. Webber, H. M., and Wilson, A. L., Analyst, 1964, 89, 632. Morrison, I. R., and Wilson, A. L., Analyst, 1963, 88, 88. Morrison, I. R., and Wilson, A. L., Analyst, 1963, 88, 100. Morrison, I. R., and Wilson, A. L., Analyst, 1963, 88, 446. Wilson, A. L., Analyst, 1966, 90, 270. Brewer, P. G., and Riley, J. P., Analytica Chim. Acta, 1966, 35, 514. Govett, G. J. S., Analytica Chim. Acta, 1961, 25, 69. Abdullah, M. I., and Riley, J. P., “Automation in Analytical Chemistry,” Technicon Symposium, Chalmers, R. A., and Sinclair, A. G., Analytica Chim. Acta, 1965, 33, 384. Shapiro, L., Res. U.S. Geol. Surv., 1974, 2, 357. Technicon AutoAnalyzer 11, Industrial Method 66-71E, Technicon AutoAnalyzer Corporation, Grasshoff, K., Deep Sea Res., 1964, 11, 597. Liss, P. S., and Spencer, C. P., J . Mar. Biol. Ass. U.K., 1969, 49, 589. Langer, K., 2. Analyt. Chem., 1969, 245, 139. Mullin, J. B., and Riley, J. P., Analytica Chim. Acta, 1955, 12, 162. Hargis, L. G., Analyt. Chem., 1970, 42, 1494. Truesdale, V. W., and Smith, C. J., Analyst, 1975, 100, 797. Koyayashi, J., in Golterman, H. L., and Clymo, R. S., Editors, “Chemical Environment in the Aquatic Habitat,” N.V. Noord-Hollandsche Uitgevers Maatschappij, Amsterdam, 1967, pp. Burton, J. D., Leatherland, T. M., and Liss, P. S., Limnol. Oceanogr., 1970, 15, 473. Scheffk, H., “The Analysis of Variance,” John Wiley & Sons, New York and London, Second Received April ‘tth, 1976 Accepted July 22nd, 1976 Paris, 1966. Basings toke. 41-55. Impression, March, 1961, p. 261.

ISSN:0003-2654    

DOI:10.1039/AN9760100019

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

32 Analyst, January, 1976, Vol. 101, pp. 32-38 The Determination of Strontium-90 and Strontium-89 in Water without Separation of Strontium from Calcium J. G. T. Regan and J. F. C. Tyler Department of Industry, Laboratory of the Government Chemist, Cornwall House, Stamford Street, London, SE1 9NQ A method has been developed for measuring strontium-90 and strontium-89 activities in water without separating the strontium from calcium. After evaporation of the sample to a convenient volume, the strontium plus calcium is purified radiochemically and an aqueous solution of their chlorides counted in a liquid scintillation spectrometer, which records the Cerenkov radiation produced. After a suitable period so as to allow the ingrowth of yttrium-90, the source is re-counted. From the two counts the activities of both strontium- 90 and strontium-89 are calculated. Activities as low as 0.1 pCi 1-1 in natural waters as well as the higher levels in effluents from nuclear installations can be measured by this method. Strontium-90 is among the most important radionuclides for which there is a need for ex- tensive and precise surveillance, and easy and accurate methods for its determination are required.The assessment of strontium-89 levels of activity is of use in the tracing and dating of nuclear weapon tests and for measuring the age of fission products discharged from nuclear installations. j?, 0.54 MeV j?, 2.27 MeV j?, 1-47 MeV Tt = 52-7 d The two radionuclides are pure beta-emitters and decay as follows : --f 90Y + {loZr 89Sr + 89Y Tt = 27.7 years Tt = 64.0 h They are most frequently counted by end-window Geiger counting and so the source has to be of low mass, and separation of the strontium carrier from calcium is essential.The most effective and commonly used method for achieving this separation involves the use of fuming nitric acid. This method can be used for calcium to strontium ratios as high as 200: 1, giving good yields of strontium and at the same time effecting a high degree of decontamination from radionuclides. Its great disadvantage is the hazardous nature of the fuming nitric acid, litre volumes of which have to be used for samples high in calcium.1 Strontium-90 has been determined by Cerenkov c o ~ n t i n g , ~ ? ~ using the modern liquid scintillation counter for measuring the Cerenkov r a d i a t i ~ n .~ , ~ The application of this technique to the determination of strontium-89 and strontium-90 as described below has several advantages : the modern liquid scintillation counter is fully automatic and very reliable; the maximum beta-energy of calcium-45 is below the Cerenkov threshold in water (0.27 MeV), so there is no need -to separate calcium from strontium; the source mass can be as high as 3.5 g in 20 ml; and the efficiency of counting is high, The main disadvantage is the high background count-rate. It should be noted that the diff- erence in the Cerenkov spectra of the two radionuclides is not sufficient to enable either to be determined in the presence of the other at low levels of activity and, for this purpose, recourse is had to the ingrowth of yttrium-90 from strontium-90.Also, the determination of the two radionuclides without separating the strontium and calcium is possible only if the mixture of calcium and strontium can be adequately decontaminated from other radionuclides. This step is achieved by a series of scavenges and the resulting precipitates that contain most of the gamma-emitters can, if determinations are required, be compressed into a disc, in the way that potassium bromide discs are prepared for infrared spectroscopy, so as to form a compact and precise source for counting on a lithium-drifted germanium detector. Ruthen- Crown Copyright.REGAN AND TYLER 33 ium must be scavenged the correspondingly efficiently because of the high energy (Emax. = 3.5 MeV) and response of the ruthenium-106 - rhodium-106 pair of radio nuclides.Experimental Apparatus Packard “Tricarb” liquid scintillation spectrometer. Perkin-Elmer 306 atomic-absorption spectrophotometer. Centrifuges, centrifuge bottles of 250-ml capacity and centrifuge tubes of 40-ml capacity were used. Reagents All reagents should be of analytical-reagent grade. Strontium carrier solution. Yttrium carrier solution. Ruthenium carrier solution. Sodium chromate solution, 10 g per 100 ml. Barium chloride solution, 4.5 g per 100 ml. Sodium hydroxide solution, 50 g per 100 ml. Hydrochloric acid (1 + 1). Ammonium carbonate, solid. Hydroxylammonium chloride, solid. Ammonia solution, sp. gr. 0.88. A standardised solution of strontium chloride containing 50 A 20 mg ml-1 solution of yttrium oxide in diluted hydrochloric A solution of ruthenium( 111) chloride in diluted hydrochloric mg ml-l of strontium.acid. acid containing 5 mg ml-l of ruthenium. Procedure for Natural Waters Take sufficient sample to meet the needs of accuracy and convenience; up to 100 1 can be taken. Acidify with hydrochloric acid, add 5 ml of strontium carrier solution and 5 ml of barium chloride solution, then boil the mixture until the volume is reduced to 350 ml (if salts precipitate it is necessary to process a larger volume). Precipitate the hydroxides and car- bonates by the addition of ammonia solution and solid ammonium carbonate, centrifuge the mixture and wash the precipitate twice with water. Dissolve the precipitate in hydro- chloric acid and dilute the solution to 300 ml (silica may appear at this stage but will be re- moved later with the barium chromate).Boil the solution so as to expel carbon dioxide and make it just alkaline to bromothymol blue with sodium hydroxide solution. (The presence of iron may mask the indicator, but at the desired pH the pale yellow colour of the iron will darken to brown.) Heat the solution to boiling, add 5 ml of sodium chromate solution and, after boiling for 1 min in order to coagulate the precipitate, filter it through a Whatman No. 30 filter-paper and wash the precipitate twice with water. Heat the combined filtrate and washings to boiling, add 5 ml of barium chloride solution, boil the solution, filter it and wash the precipitate twice with water. To this combined filtrate and washings, add 2 ml each of ruthenium and yttrium carrier solutions and sufficient sodium hypochlorite solution (about 3 ml) to oxidise the ruthenium and produce a straw colour.Heat to just boiling, add sufficient solid hydroxylammonium chloride to reduce the ruthenium and chromate, then precipitate the ruthenium and the chromium with sodium hydroxide solution at pH 8 and boil the solution in order to coagulate the precipitate. Allow to settle, filter the mixture on a Whatman No. 541 filter-paper and wash the precipitate and filter-paper once with water. Make the solution just acid by the addition of hydrochloric acid, add 1 ml of ruthenium and 2 ml of yttrium carrier solutions and repeat the procedure from “and sufficient sodium hypochlorite . . . .” Add 1 ml of yttrium carrier solution, followed by 1 ml of sodium hydroxide solution and sufficient solid sodium carbonate to precipitate the hydroxides and the carbonates.Centri- fuge, wash the precipitate twice with water, dissolve it in the minimum amount of hydro- chloric acid and dilute the solution to 70ml. Boil the solution so as to expel the carbon34 REGAN AND TYLER: DETERMINATION OF STRONTIUM-90 AND -89 IN Amlyst, Vol. 101 dioxide, add a few crystals of hydroxylammonium chloride and adjust the pH to 8 with sodium hydroxide solution. Boil the solution in order to coagulate the precipitate, filter it on a Whatman No, 541 filter-paper so as to remove the yttrium hydroxide and record the time at this stage, which is that at which yttrium-90 ingrowth starts.Wash the precipitate twice with water, acidify the filtrate with a few drops of hydrochloric acid and reduce the volume of the solution to about 22 ml by boiling it. Cool and pipette 20 ml into a plastic liquid scintillation vial. Count and record the time at which counting is started. After a suitable ingrowth period of 1-10 d, depending on the activity present and the accuracy desired, re-count and again note the time at which the counting started. Chemical yield Determine the strontium and the calcium in the final counting solution, and if it is present in significant amounts, as is found with large volumes of samples of natural waters, also determine the strontium in the carrier-free solution. Make the determination by means of atomic-absorption spectrophotometry, using 1 per cent.lanthanum as a releasing agent. Calculate the strontium yield and from the calcium content of the counting solution assess the counting efficiencies of the three radionuclides from the counting efficiency graphs. Cownting eficiency graphs The variation of the counting efficiency with the calcium content of the counting solution is small but may need to be taken into account. Analytical-reagent grade calcium salts contain trace amounts of iron, which colour con- centrated solutions and cannot be removed by direct precipitation as the hydroxide unless a carrier is present. Use yttrium as a carrier, then precipitate the calcium as carbonate, dissolve the precipitate in the minimum amount of hydrochloric acid and boil the solution in order to remove carbon dioxide.Prepare three sets of solutions containing 0, 1 , 2 and 3.5 g of calcium per 20 ml of solution and dose one set with strontium-90, another with strontium-89 and the third with yttrium-90. Use high levels of activity for dosing so that counting times are short and ingrowth of yttrium-90 is slight. Count these solutions and then draw the counting efficiency graphs. Typical ranges of efficiency, per cent., for 0-3.5 g of calcium are: strontium-90, 0-7-1.7; strontium-89, 35-8-38-4; and yttrium-90, 56.5-60.3. The average strontium yield is 60 per cent. Calculations pCi 1-1 CF - CID k x R x V x 2.22 Strontium-90 activity = and pCi 1-1 1 R x 'v x 2.22 ['I(& + "k") - cF f] Strontium-89 activity = (at the time of analysis) where CI = Initial counts per minute corrected for background.CF = Final counts per minute corrected for background. GI = GF = D = Emsr = Egos, = Egoy = Fractional counter efficiency for yttrium-90. R = Fractional chemical yield of strontium. V = Volume of sample in litres. k Ingrowth factor for yttrium-90 at the mid-point of the first count. Ingrowth factor for yttrium-90 at the mid-point of the final count. Strontium-89 decay factor for the time elapsed between the mid-points of the initial and final counts. Fractional counter efficiency for strontium-89. Fractional counter efficiency for strontium-90. = Egos, (1 - D) + E90y (GF - GID)January, 1976 WATER WITHOUT SEPARATION OF STRONTIUM FROM CALCIUM E g o s r + GIEooy K = E89Sr Procedure for samples high in calcium 35 If the calcium content of the sample exceeds 3.5 g carry out a partial separation of the strontium from the calcium, modifying the method as follows.Dissolve the first carbonate precipitate in 1 + 1 acetic acid, dilute the solution to 300 ml with water and boil it in order to expel carbon dioxide. Adjust the pH of the solution to 6 by the addition of sodium hydroxide solution, precipitate calcium sulphate by addition of 10 ml of saturated sodium sulphate solution and heat the mixture to boiling. Cool, centrifuge it and, after washing the precipitate twice with water, add to the latter 20 ml of water and 3 g of sodium carbonate and heat the mixture on a boiling water bath for 30 min. Centrifuge it, wash the solids with water and then dissolve them in hydrochloric acid.Continue as under Procedure for Natural Waters, from “Dissolve the precipitate in hydrochloric acid . . . .I’ This procedure has been applied successfully to samples containing 10 g of calcium. Procedure for Effluents from Nuclear Installations Usually a volume of between 10 and 400ml is sufficient for these samples and it is not necessary to use the procedure for samples high in calcium. Suitably scale down the procedure for natural waters and do not use less than 50mg of strontium carrier in order to keep the activity of the strontium radionuclides to an acceptable level during the atomic-absorption spectrophotometric determination of the strontium. Also, for ease of handling and for effective scavenging, use an amount of ruthenium carrier of not less than 5 mg and of yttrium carrier of not less than 10 mg.Results Limits of Detection The limits of detection of strontium-89 and strontium-90 are interdependent and the pre- sence of either will raise the limit of detection of the other. In the absence of strontium-89 the limit of detection of strontium-90 is 0.8 pCi and in the absence of strontium-90 that of strontium-89 is 1.0 pCi. These limits have been calculated on the assumption that stront- ium-89 will be detected if the count-rate of the source is greater than three standard de- viations above the background count-rate and that strontium-90 will be detected if the net ingrowth count-rate is greater than three standard deviations of the background count-rate. They relate to a counting time of 500 min, a chemical recovery of 60 per cent., a background count-rate of 13 counts min-1, an ingrowth period of 8 d, a strontium-89 counting efficiency of 36 per cent.and a strontium-90 counting efficiency of 56 per cent. Decontamination Factors decontamination experiments were carried out. In order to determine the level of contamination of the final source by other radionuclides, Simulated water samples containing known TABLE I DECONTAMINATION FACTORS Decontamination factor as Nuclide added Uranium-238 to uranium-234 of uranium series Radium-226 to lead-210 portion of uranium series Lead-210 plus equilibrated daughters . . Thorium-232 series . . .. .. .. Ruthenium-106 . . .. .. .. .. Cerium-144 . . .. .. .. .. Antimony-124 . . .. .. .. .. Antimony-125 . . .. .. .. .. Cobalt-60 .... .. .. .. Caesium-137 . . .. .. .. .. Barium-140 . . . . .. .. .. strontium-90 .. x 104 .. 9.0 x 103 .. 4.6 x 105 . . >2.4 x 104 .. 3.8 x 104 . . >4.9 x 105 . . >2.0 x 105 . . >i.o x 105 .. 2.6 x 105 .. 1.9 x 105 . . >1.0 x 106 - 3 strontium-89 ~ 8 . 2 x 104 1.6 x 105 1.9 x 105 ~ 2 . 2 x 104 5.0 x 103 2.1 x 105 7.0 x 104 3.6 x 105 3.2 x 104 5.7 x 103 >1-6 x losTABLE I1 COMPARISON OF RESULTS FOR NATURAL WATERS Results are expressed in pCi 1-l. Calcium/ Period 1 Period 2 Period 3 Period 4 100 1 Procedure Strontium-90 Strontium-89 Strontium-90 Strontium-89 Strontium-90 Strontium-89 Strontium-90 Strontium-89 I A A g Per If \- I- 7 0.13 Described 2.19 f 0.04 <0-29 2.41 f 0.04 <0*13 1.87 f 0.02 <0*15 2.43 f 0.07 <0*22 <0.22 2-04 & 0.04 Previous 2.12 -f 0.03 < 0.36 2.31 f 0.05 <0*19 1-74 f 0.08 <0.16 0.87 Described 4.3 Described 9.8 Described 11.2 Described Previous Previous Previous Previous 1-14 f 0.08 0.94 f 0.03 0.07 f 0.03 0.11 f 0.01 0-38 f 0.03 0.40 f 0.02 0.19 f 0.05 0.21 f 0.01 < 0.53 <0*66 <om21 < 0.24 < n.22 <0*29 <0*38 <0*18 1-12 f 0.05 1.08 f 0.04 0.14 f 0.04 0.23 f 0.01 0.39 f 0.03 0.44 f 0.03 0.19 f 0.07 0.13 f 0.01 < 0.20 <0-34 <0*15 <0*19 t 0 .1 4 < 0.09 <0*34 < O * l l 1.00 f 0.05 0.82 f 0.03 0.07 & 0.03 0.09 f 0.01 0.35 -& 0.03 0.33 f 0-02 0.08 f 0.06 0.13 0.01 < 0.30 <0*34 <0*17 <0*15 <0*27 (0.15 <0*36 <0*18 1.14 f 0.08 1.01 f 0.04 0.16 & 0.04 0-07 f 0.01 0-36 f 0.04 0.29 f 0.01 0.19 f 0.03 0.14 f 0.01 <0*31 <om20 <0*16 <0*20 <0-15 <0*20 <0*15 <Om16 The confidence limits are for counting statistics only and are 1-64 x standard deviation for the plus or minus figures given under strontium-90 and 3 X standard deviation for the figures given under strontium-89.January, 1976 WATER WITHOUT SEPARATION OF STRONTIUM FROM CALCIUM 37 activities of possible contaminants were taken through the procedure.The final source was counted and re-counted after 15 d, and decontamination factors were calculated from the ratio of the added activity of the parent nuclide only to the apparent activities of stront- ium-90 or strontium-89 in the sample. The results obtained are listed in Table I. Natural Waters Samples from five sites with a range of calcium concentrations were examined over four different periods by the described method and the method previously used in this laboratory.’ The latter method consists in effecting separation by use of fuming nitric acid and Geiger counting an yttrium source for strontium-90 and a strontium source for strontium-89.The sample volumes were approximately 100 1, one quarter to one third of which was examined by the described method and the remainder by the previous method. The results obtained are listed in Table 11. Effluents from Nuclear Installations Effluent samples were examined by the described method and by the method previously used for such samples in this laboratory, the latter consisting in carrying out separations with fuming nitric acid and, after removing the yttrium-90, Geiger counting a strontium source for comparison with a similarly prepared strontium-90 standard source, no allowance being made for the possible presence of strontium-89.The results obtained are listed in Table 111. TABLE I11 COMPARISON OF RESULTS FOR NUCLEAR INSTALLATION EFFLUENTS Results are expressed in nCi 1-l. Previous method’ : total radiostrontium activity as strontium-90 35.3 0.91 0.4 1 45.7 1-51 2-2 Described method Strontium-90 Strontium439 activity activity 0.91 0.02 0.39 0.04 44.3 1.5 1.32 0.02 1.7 0.1 A > 35.1 0.6 Neutron-irradiated Uranium Natural uranium oxide was irradiated in a reactor in order to provide mixed fission products and was then dissolved in nitric acid. The activity of caesium-137 was measured so as to enable the strontium-90 and strontium-89 activities of the solution to be calculated from fission yield data. An aliquot of the solution was examined by the described method; activities were : strontium-90, calculated 0.61, described method 0.78 &- 0.24 ; strontium-89, calculated 62.4, described method 65-4 & 0-1 nCi 1-l. Time of Analysis Processing times for natural waters, after evaporation, are very much reduced, a batch of two samples plus a blank taking 14 h as opposed to 14 h for each soft and 30 h for each hard water by the fuming nitric acid method. The maximum counting time that can conveniently be used in practice is 500 min, so two samples and a blank are counted in a 24-h period and, of course, it is best to count in cycles of 100 min per counting vial. This paper is published by permission of the Government Chemist. References 1. 2. 3. “The Radiochemistry of Barium, Calcium and Strontium,” National Academy of Sciences, National Randolph, R. B., Int. J . Appl. Radiat. Isotopes, 1975, 26, 9. Narrog, Von J., Atompraxis, 1965, 11, 373. Research Council, Nuclear Science Series, NAS-NS 3010, Washington, D.C., 1960.38 REGAN AND TYLER 4. 5. 6. 7. Ross, H. H., Analyt. Chem., 1969, 41, 1260. Parker, R. P., and Elrick, R. H., in Bransome, E. D., Editor, “The Current Status of Liquid Scintil- “The Radiochemistry of Ruthenium,” National Academy of Sciences, National Research Council, Wood, R., and Richards, L. A., Analyst, 1965, 90, 606. Received July 24th, 1975 Accepted September 15th, 1975 lation Counting,” Grune and Stratton, New York, 1970. Nuclear Science Series, NAS-NS 3029, Washington, D.C., 1961.

ISSN:0003-2654    

DOI:10.1039/AN9760100032

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, January, 1976, Vol. 101, pp. 39-43 39 The Determination of Tellurium in Leaded Free- cutting Steels by Atomic-a bsorption Spectrometry W. D. Cobb, W. W. Foster and T. S. Harrison British Steel Corporation, Scunthorpe and Lancashire Group, P.O. Box No. 1, Scunthorpe, South Humberside, DN16 1BP A procedure is presented for the determination of tellurium within the range 0.005-0.08 per cent. , with solvent conditions that permit its incorporation in a composite scheme that includes lead. In the preferred fuel-rich flame the background absorbance of iron and inter-element interferences are at a minimum. Good recoveries of tellurium when added to solutions of leaded and stainless steels, together with satisfactory reproducibilities for samples and good agreements with results by the slower turbidimetric procedure, are shown.Tellurium, up to 0.05 per cent., is added to leaded free-cutting steels in order to improve further their machinability. Gravimetric and titrimetric methods for the determination of small amounts of tellurium are relatively insensitive and presuppose no interferences. Photometric determinations based on the formation of complexes with thiourea, iodide and sodium diethyldithiocarbamate are more precise,l while reduction with tin(I1) chloride to colloidal tellurium in an acidic medium is the basis of a turbidimetric procedure that is commonly used within the steel industry.2 Previous workers1 applied atomic-absorption spectrophotometry to the determination of tellurium within the range 0.0005-0.030 per cent.in steel and reported that the sensitivity was improved by extracting it into amyl acetate. Other worker^,^ using aqueous solutions, preferred its preliminary conversion into potassium hexaiodotellurate or tellurium diethyl- dithiocarbamate and extraction into isobutyl methyl ketone. In more recent years the continued improvement in instrument and lamp design has enabled a more stable signal and superior sensitivity to be obtained for tellurium so that it can now be determined directly in acidic solution within the range 0-0-10 per cent. in steel. The object of the present work was therefore to develop a suitable procedure that would, at the same time, extend the composite scheme devised earlier for other elements.* Consequently, its determination is often required.Experimental The method developed for application to plain carbon, leaded free-cutting and stainless steels containing from 0.005 to 0.08 per cent. of tellurium is as follows. Apparatus A Techtron, Model AA4, atomic-absorption spectrophotometer, up-dated to the AA5 instrument by incorporation of the later type of read-out and burner units, was used with the following conditions. Wavelength . . .. .. .. Lamp current . . .. .. Slit width . . .. .. .. Burner .. .. .. .. Observation height . . .. .. Fuel setting . . .. .. .. Support setting . . .. .. Damping .. .. .. .. Scale- expansion . . .. .. 214.3 nm 8 mA 100 pm Air - acetylene, 10-cm path length Light beam 1 mm above burner top Acetylene cylinder pressure, 10 lb in-2 and fuel flow adjusted to give minimum flame absorbance Air pressure, 30 lb in-2; gauge reading, 18 lb in+ B x 240 Reagents COBB et al.: DETERMINATION OF TELLURIUM IN LEADED Analyst, Vol. 101 H Y ~ Y O C ~ ~ O Y ~ C acid, sp. gr. 1.16-1.18. Nitric acid, sp. gr. 1.42. Stock iron solution. Transfer 50 g of pure iron sponge into a 1.5-1 beaker. Add 300 ml of Cautiously Replace any loss of acid Cool the solution, dilute it to 1 1 with water Dissolve 0.1 g of Specpure tellurium metal in 20 ml of nitric hydrochloric acid (sp. gr. 1.16-1.18) and heat it gently so as to dissolve the iron. oxidise it with nitric acid (sp. gr. 1.42) and boil off nitrous fumes. with hydrochloric acid and add a further 200 ml. and filter. acid (50 per cent. V / V ) and dilute the solution to 11. Standard tellurium solution.1 ml of solution = 100 pg of tellurium. Preparation of sample solution Transfer 2 g of sample into a 150-ml conical beaker and add 40 ml of hydrochloric acid (50 per cent. V / V ) . Cover the beaker with a watch-glass and allow the mixture to digest until the sample has dissolved, then oxidise the iron with nitric acid (sp. gr. 1-42), adding one or two drops in excess. Boil the solution so as to expel nitrous fumes, cool it and dilute it to 100 ml in a calibrated flask. Preparation of calibration solutions Add the tellurium fractions according to the following scheme, dilute each solution to the calibra- tion mark and mix. Filter if necessary. Into each of a series of 50-ml calibrated flasks transfer 20 ml of stock iron solution. Standard tellurium solution/ml 0 1 2 4 6 8 Tellurium, p.p.rn. 0 2 4 8 12 16 Tellurium, per cent.0 0.01 0.02 0.04 0.06 0.08 Determination of Tellurium Set the instrument according to the instrument conditions given above and spray the appropriate calibration solutions followed by the sample solution. Spray water between each test and set the zero while spraying water each time. Repeat this procedure. Prepare a calibration graph by plotting the absorbance readings against the equivalent percentages of tellurium. Convert the absorbance reading of the sample into tellurium percentage by reference to this graph. Discussion The choice of instrument parameters arose from the following considerations. Flame Conditions The absorbance by the flame at a wavelength of 214.3 nm was found to be dependent on the fuel to air ratio.The minimum absorbance was obtained by adjusting the acetylene flow while observing the meter needle, which resulted in a fuel-rich flame. Although the use of a weaker mixture was found not to affect the sensitivity or noise level, the background absorb- ance caused by iron(II1) chloride was found to have its lowest value with the recommended flame conditions. Sensitivity The sensitivity in aqueous solution was found to be 0.18 p.p.m. and in the presence of iron (2 per cent. m/V), as in the method, to be 0-20 p.p.m. In order to cover the desired range of up to 0-080 per cent. of tellurium for a 2 per cent. m/V sample solution, a concentration of up to 16 p.p.m. was required. For tellurium contents of up to 0.040 per cent. a scale expansionJanuary, 1976 FREE-CUTTING STEELS BY ATOMIC-ABSORPTION SPECTROMETRY 41 of x 2 was used.The noise level under these conditions was acceptable, requiring a low damping setting. The lamp current, wavelength and slit width recommended by the manu- facturer were found to give the optimum results. Matrix Effects The apparent absorbance at 214.3 nm of a solution containing 2 per cent. m/V of iron, with the flame adjusted to give minimum absorbance, is equivalent to 1 p.p.m. of tellurium. This effect is attributed to light scattering by solid particles in the flame, which resulted from spraying solutions of higher salt concentrations. In the proposed method this effect is overcome by including 2 per cent. m/V of pure iron in the calibration solutions. For leaded steels this apparent absorbance is identical with that found with pure iron for the same concentration.For stainless steels, however, the apparent absorbance is equivalent to 1.2 p.p.m. of tellurium. As the recovery of tellurium (see Interference Tests) is within the range of reproducibility of the method (i.e., no change in sensitivity), this slight increase in apparent absorbance can be allowed for by processing a stainless-steel sample containing no tellurium and deducting any absorbance greater than the nil-point absorbance from that due to the sample solution, or by preparing calibration solutions containing element concentrations that are similar to those of the sample under test. Interference Tests As there is little published information on interferences in the direct determination of tellurium in solutions of steel samples, the effect of adding tellurium equivalent to 0.020 and 0.040 per cent.to solutions of leaded steels and also to stainless steels was studied. The analytical values for these samples are given in Table I. For the stainless-steel samples the TABLE I ANALYTICAL VALUES FOR SAMPLES USED IN RECOVERY TESTS Sample T w e C P Mn S Si Pb 1759 Leaded steel 0.08 0.087 1.12 0.27 <0*006 0.25 8332 0.13 0.128 1-18 0.42 0.26 BCS 235/2 18/9 Stainless 0.073 0.020 0.020 0.018 0-82 BCS 335 Stainless steel 0.093 0.018 0.94 0.023 0.67 0.0015 US 14 0.081 0.022 0.61 0.019 0.51 0.0005 steel + Ti Sample Type Cr cu V co Ti Ta 1759 Leaded steel 8332 BCS 235/2 18/9 Stainless 18.60 0.12 0.04 0.056 0.32 BCS 335 Stainless steel 18-45 0.11 0.04 0.034 0.46 0.0017 US 14 18.22 0.046 steel + Ti Ni 9.38 9.47 7-95 Al 0.048 increase in background absorbance was equivalent to 0.001 per cent.of tellurium in each instance, which was allowed for in the recovery tests. The results shown in Table I1 indicate that the method is free from inter-element inter- f erence. TABLE I1 TELLURIUM FOUND, PER CENT., AFTER ADDING (A) 0.020 AND (B) 0.040 PER CENT. TO SAMPLES OF LEADED AND STAINLESS STEELS Sample A B 1759 0.020, 0.020 0.040, 0.040 8332 0.020, 0.0205 0.040, 0.0405 BCS 235/2 0.021, 0*020 0.0405, 0.0405 BCS 335 0.021, 0.020 0.0405, 0.040 US 14 0.021, 0.020 0.0405, 0.040542 Application and Reproducibility Tests and reproducibility tests were carried out on samples of “tellurised” leaded steels.results obtained are given in Tables I11 and IV. COBB et aZ. : DETERMINATION OF TELLURIUM IN LEADED Analyst, VoZ. 101 As no standardised samples were available for determining tellurium in steel, application The All atomic-absorption spectrophotometric TABLE I11 RESULTS OF REPRODUCIBILITY TESTS FOR LEADED FREE-CUTTING STEEL SAMPLES Tellurium found by chemical method, Sample per cent. 3189 BB13 0.005 3189 B13 0.010 1993 A5 0.035 1993 B5 0.034 1993 C5 0.029 Mean Reproducibility value Tellurium found by AAS, per cent. (95 per cent. f 2 s) 0-006, 0.0055, 0.005, 0-0055, 0.0045, 0.0045, 0.005, 0.004, 0.0055, 0.006 0.005 0.0012 0.012, 0.011, 0.012, 0.011, 0.009, 0.0115, 0.012, 0.0115, 0.0095, 0.0115 0.011 0.0021 0-033, 0.034, 0.0345, 0.034. 0.0345, 0.0345, 0.034, 0.0335, 0.0335, 0.033 0.034 0.0012 0.033, 0.034, 0.0355, 0.034, 0.034, 0-0345, 0.034, 0.035, 0.0335, 0.034 0.034 0.001 4 0.028, 0.030, 0.030, 0.0295, 0.029, 0.029, 0,029, 0.030, 0.029, 0.029 0.029 0.0013 results given were obtained by following the complete procedure on separate occasions.The chemical values, provided by independent laboratories, were obtained by photometric mea- surement of the turbidity due to colloidal suspension of tellurium following reduction with tin(I1) chloride.2 The good agreement of the results with those obtained by the chemical procedure confirms that the method is accurate and gives adequate reproducibility. TABLE IV COMPARISON OF RESULTS OBTAINED BY CHEMICAL AND ATOMIC-ABSORPTION SPECTROPHOTOMETRIC METHODS Samples of leaded free-cutting steel 3140 2206 A4 2206 A10 2206 A14 2206 A15 4572 B1 4572 B3 4572 B9 3958 M2 3992 M2 3992 M3 3355 4 BOT 1 BOT 2 Tellurium, per cent.I A \ Chemical value Value found by AAS 0.018 0.017, 0-017 0.053 0.0515, 0.0515 0.041 0.0405, 0.042 0.036 0.037, 0.037 0.040 0.0405, 0.0415 0.047 0.046, 0.046 0.039 0.0395, 0.0405 0.048 0.046, 0.046 0.042 0.042, 0.042 0.036 0.0365, 0.0355 0.036 0.036, 0.036 0.040 0.040, 0.040 0.024 0.024, 0.023 0.027 0-026, 0-025 Conclusions This rapid, convenient atomic-absorption spectrophotometric method for the determination of tellurium in the range 0*005-0-080 per cent. in plain and leaded free-cutting steels gives results that are in close agreement with those given by the existing turbidimetric method and good reproducibility. Inter-element interference is very slight and the method is also applicable to stainless steels. Six determinations can be completed in about 1.5 h but this time is proportionately reduced by incorporating the procedure in the composite scheme, which includes lead, to considerable practical advantage. The authors thank Messrs. L. Kidman and R. Merry of the Rotherham Works, and Messrs. A. K. Wright and G. Stamp of the Normanby Park Works, for the supply of additional steelJanuary, 1976 FREE-CUTTING STEELS BY ATOMIC-ABSORPTION SPECTROMETRY 43 samples and analytical data. They also thank the Management of the Scunthorpe and Lancashire Group, British Steel Corporation, for permission to publish this work. References 1. 2 . 3. 4. Marcec, M. V., Kinson, K., and Belcher, C . B., AnaZytica Chim. Acta, 1968, 41, 447. Inland Steel Co., U.S.A., personal communication, 7th March, 1962. Chakrabarti, C. L., Analytica Chim. Acta, 1967, 39, 293. Hamson, T. S., Foster, W. W., and Cobb, W. D., Metallurgia Metal Forming, 1973, 40, 361. Received July 24th, 1976 Accepted September I d , 1976

ISSN:0003-2654    

DOI:10.1039/AN9760100039

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

442 Analyst, January, 1976, Vol. 101, pp. 44-48 Colorimetric Determination of Phosphorus in Silicates Following Fusion with Lithium Metaborate J. B. Bodkin Mineral Constitution Laboratories, The Pennsylvania State University, University Park, Pennsylvania 16802, U.S.A. The determination of phosphorus in silicate materials by use of a colorimetric technique following fusion with lithium metaborate is described. The method is rapid and the results obtained compare favourably with the accepted values for standard rock and mineral samples and also with gravimetric results. Reliable values of u p to 1.75 per cent. for the phosphorus(V) oxide content have been obtained by this method. Although phosphorus is rarely present in amounts greater than those of minor constituents, it is nearly always determined in analyses of silicate rocks.Phosphorus content is an im- portant component in chemical calculations involving rock analyses and is significant in petrological models devised to explain the origin and subsequent history of basaltic magmas. Basic silica-undersaturated rocks, such as trachybasalts and basanitoids, may contain several per cent. m/m of phosphorus(V) oxide, whereas in granitic rocks and in minerals the content is much lower and often is in the parts per million range. The wide range of phosphorus(V) oxide contents has been a factor in the use of various colorimetric methods that have been proposed for its determination .1--3 The introduction of the lithium metaborate - nitric acid solution technique4 has led to the rapid determination of the major and minor constituents in rocks and minerals by atomic- absorption ~pectrophotometry.~~6 The need for a reliable and rapid determination of phos- phorus has resulted in the development of a colorimetric procedure using the same sample solution as that analysed for the major constituents. The most sensitive methods for determining phosphorus have involved the reduction of molybdenum(V1) to molybdenum(V) and determination of the intensity of the colour of the molybdenum or heteropoly blue formed by this species.Ingamells' proposed a method for the rapid colorimetric determination of phosphorus in the presence of silica following dis- solution of the sample using a lithium metaborate - nitric acid solution technique. We obtained erratic results with this method, however, and attempts to improve it by removal of the silica also proved unsuccessful.The method of Fogg and Wilkinson* was found to be reliable for the determination of phosphorus and readily adaptable to a sample solution obtained by the lithium metaborate - nitric acid solution technique, but it is necessary to treat the sample solution with hydro- fluoric and sulphuric acids in order to remove the silica. Experimental Reagents Lithium metaborate. Anhydrous LiBO,. Nitric acid, 4 per cent. VlV. 4-Nitrophenol solution, 0.2 per cent. m/V. Sulphuric acid, 1 N. Sodium hydroxide solution, 1 N. Ascorbic acid. Ammonium molybdate reagent. Dissolve 10 g of (NH,)6Mo,0,4.4H,0 in 75 ml of water. Heat in order to effect dissolution and then allow to cool to room temperature, Filter into a 100-ml calibrated flask and dilute to volume.Add the solution with stirring to 300ml of 50 per cent. V/V sulphuric acid contained in a polyethylene bottle. Cool the resulting solution to room temperature. This reagent is stable for approximately 1 month but should be discarded either on the formation of a precipitate or on the development of a blue colour. Standard phosphate solution. Dissolve 0.7669 g of potassium dihydrogen orthophosphate in distilled water and dilute to 11. 1 ml of solution = 400 pg of phosphorus(V) oxide.BODKIN 45 Dilute 25.0 ml of the standard phosphate solution to Dilute standard phosphate solution. 1 1. 1 ml of solution = 10 pg of phosphorus(V) oxide. Instrumentation used for all determinations.A Bausch and Lomb Spectronic 70-6 spectrophotometer and 20-mm absorbance cells were Procedure The lithium metaborate - nitric acid solution technique has been reported previously4 but slight modifications have since been incorporated. The method is included here for com- pleteness. A 200-mg amount of sample, ground to pass a 200-mesh sieve (ASTM), is mixed with 800 mg of anhydrous lithium metaborate. The mixture is transferred to a pre-ignited graphite fusion crucible, which is placed in a muffle furnace at 1050 “C for 10 min. While molten, the fused material is poured into a polypropylene beaker containing 80.0 ml of 4 per cent. V/V nitric acid. The beaker is covered and placed on a magnetic stirrer until the sample is com- pletely dissolved (usually 5-10 min).This sample solution is used without further dilution. One millilitre of this solution is equivalent to 0.0025 g of sample. A 5-00-ml aliquot (0.0125g of sample) of the solution is placed in a 55-ml platinum or PTFE dish. Five millilitres of hydrofluoric acid and 5 drops of 50 per cent. V/V sulphuric acid are added. The dish is placed on a hot-plate and the solution heated slowly to fumes of sulphur(V1) oxide. The fuming is continued nearly to dryness. After cooling, the residue is moistened with water and a further 5 drops of 50 per cent. V/V sulphuric acid are added, the solution being taken to dryness. The dish is again cooled and 5 drops of 50 per cent. V/V sulphuric acid and 5 ml of water are added. The dish is heated in order to effect complete dissolution and the solution is then transferred to a 100-ml beaker with water, washings from the dish being added to the beaker. The acidity is adjusted by adding 2 drops of 0.2 per cent.m/V 4-nitrophenol indicator solution followed by the dropwise addition of 1 N sodium hydroxide solution until a deep yellow colour is obtained. Sulphuric acid (1 N) is added dropwise until the solution is colour- less. After diluting the solution to 40 ml ammonium molybdate reagent (4.0 ml) is added, followed by 0.2 g of ascorbic acid. The solution is heated to boiling, boiled for 1 min and then cooled to room temperature before transferring it to a 50-ml calibrated flask and diluting to volume. The absorbance of the solution at 810nm is measured, using a 20-mm cell, and with Specpure silica being carried through the entire procedure as a blank solution.The phosphorus(V) oxide concentrations are determined from a calibration graph prepared by using the dilute standard phosphate solution. Up to 0.40 per cent. m/m of phosphorus(V) oxide can be determined using a 50-ml calibrated flask as described. Higher phosphorus concentrations were determined by using larger calibrated flasks and adjusting the blank value accordingly. Preparation of Calibration Graph phosphate solution. Procedure, starting with the paragraph “The acidity is adjusted by . . . .” Into a series of 100-ml beakers place 0,0.5, 1.0, 2.0,3.0, 4.0 and 5.0 ml of the dilutestandard Follow the same steps for developing the colour as described under Discussion The method chosen for the colour development is essentially that of Fogg and Wilkinson,B who applied their procedure to the determination of phosphorus in boiler water, effluents and salt deposits.The colour development takes place rapidly on boiling the solution for 1 min and is stable for an indefinite period. Silicon and arsenic are the only elements that interfere, both elements reacting to form the molybdenum-blue colour under the same experimental conditions. Although use of this procedure8 appears to overcome the interference by silica, attempts to determine the phosphorus in the presence of silica in lithium metaborate - nitric acid solution were unsuccessful. The removal of silica by volatilisation with hydrofluoric acid and sulphuric acid overcomes this problem. The developed colour faded immediately.46 BODKIN : COLORIMETRIC DETERMINATION OF PHOSPHORUS IN Analyst, VoZ.101 TABLE I COMPARISON OF RESULTS OBTAINED FOR PHOSPHORUS(V) OXIDE CONTENT FOR VARIOUS STANDARDS BY PROPOSED METHOD WITH ACCEPTED VALUES Phosphorus(V) oxide content, per cent. I 1 Accepted value r > A Source Material Designation This work Value Reference USGS* . . . . Granite G- 1 0.069 0.09 9 Diabase w- 1 0.129 0.14 9 Granite G-2 0.133 0.14 10 Granodiorite GSP- 1 0.286 0.28 10 Basalt BCR-1 0.363 0.36 10 Andesite AGV- 1 0.494 0.49 10 Granite GA 0-121 0.12 11 Basalt BR 1.05 1.04 11 CRPGt .. . . Granite GH 0.007 0.01 11 CSRMS . . . . Syenite s- 1 0.212 0.22 12 * United States Geological Survey, Reston, Va., U.S.A. t Centre de Recherches Petrographiques et Geochimiques, Nancy, France.Canadian Standard Reference Materials, Mines Branch, Ottawa, Ontario, Canada. Arsenic is not usually found in significant amounts in most rock and mineral samples. Arsenic present as arsenate interferes, whereas arsenite does not. Reduction of arsenate to arsenite should overcome this interference, although this has not been attempted. The proposed method permits the determination of phosphorus on the same sample solution as that used for the determination of the major elements and the phosphorus is easily and rapidly determined. The method has been applied to a wide range of materials and no problems have been encountered. TABLE I1 COMPARISON OF RESULTS OBTAINED FOR PHOSPHORUS(V) OXIDE CONTENT BY THE PROPOSED METHOD AND BY GRAVIMETRIC METHODS Phosphorus(V) oxide content, per cent.f A I Material Proposed method Gravimetric method* Granite Gneiss .. .. 0.100 0.097 0.108 Granite Aplite . . .. .. 0.036 0.034 Granite Aplite . . . . . . 0.042 0.036 Granodiorite Gneiss . . . . 0.093 0.092 0.098 Granite Aplite . . .. .. 0.048 0.053 0.043 0.044 Granite .. .. .. 0.022 0.023 0.029 0.023 Granite Gneiss .. .. 0.091 0.096 0.092 Amphibolite . . .. .. 0.247 0.239 Trachybasalt . . .. .. 0.4 69 0.451 0.455 Trachybasalt . . .. .. 0.767 0.765 0.755 Basanitoid . . .. .. 0.797 0.798 0.799 Trachybasalt . . .. .. 1.49 1.52 1.51 Trachybasalt . . .. . . 1.75 1.76 1.77 * Gravimetric determinations by S. S. Goldich, Geology Department, Northern Illinois University, DeKalb, Ill., U.S.A.January, 1976 SILICATES FOLLOWING FUSION WITH LITHIUM METABORATE 47 The results obtained for phosphorus(V) oxide in some standard rock samples are compared with the accepted values in Table I. Agreement with the accepted values is within 0.01 per cent.in all instances, with the exception of Granite G-1, for which the difference is 0.02 per cent. However, an independent gravimetric determination of phosphorus on a sample from the same bottle as that used in the colorimetric determination yielded a value of 0.072 per cent. of phosphorus(V) oxide, which is in agreement with the value obtained (0.069 per cent.) by the proposed colorimetric method. Examination of the data given by Fleischer9 shows many other values of about 0.07 per cent. for colorimetric determinations of phosphorus. A number of samples were independently analysed by two different gravimetric methods and also by the proposed colorimetric procedure.The gravimetric methods for these separate determinations were : (1), a single precipitation and weighing as ammonium molybdophos- phate ; and (2), a preliminary precipitation as ammonium molybdophosphate followed by precipitation as ammonium magnesium phosphate and ignition to give magnesium pyrophos- phate (Mg2P20,). These results are compared in Table I1 and indicate the range of the method. It is possible to extend the range to include at least twice the highest value given in the table (1.75 per cent.) by a combination of increasing the volumes and selecting smaller absorbance cells. Table I11 contains values for phosphorus(V) oxide in a number of new standard rock and mineral samples for which very few or tentative results have been published.TABLE I11 VALUES OBTAINED BY THE PROPOSED METHOD FOR PHOSPHORUS(V) OXIDE CONTENT OF NEW STANDARD SAMPLES Phosphorus(V) oxide Source Material Designation content, per cent. CRPG .. . . Biotite Mica-Fe 0.420 Diorite DR-N 0.249 Disthene DT-N 0.087 Granite GS-N 0-282 Bauxite BX-N 0-142 CSRM .. .. Gabbro MRG- 1 0.054 Syenite SY-2 0.441 Syenite SY-3 0.560 NIM* .. . . Granite NIM-G 0.012 Dunite NIM-D 0.020 Norite NIM-N 0.015 P yroxenite NIM-P 0.024 Luj avrite NIM-L 0-049 Phlogopite Mica-Mg 0.002 Serpentine UB-N 0.010 Feldspar FK-N 0.009 Syenite NIM-S 0.119 * National Institute for Metallurgy, Aukland Park, South Africa. Mean values, standard deviations and coefficients of variation for four USGS standards are listed in Table IV.These values were obtained from individual determinations carried out routinely over a period of 2 years. The data indicate that the proposed method is capable of good reproducibility on a routine basis. The mean values reported in Table IV are slightly different from the values reported in Table I, which were obtained from single determinations carried out as a special project. TABLE IV Sample w- 1 G-2 BCR- 1 AGV- 1 PRECISION DATA Mean phosphorus(V) Standard oxide content, deviation, Number of 0.131 0.0044 7 0.133 0.0043 12 0.360 0.0036 10 0.496 0-0105 13 per cent. per cent. determinations Coefficient of variation, per cent. 3.36 3.22 1.00 2.1248 BODKIN The author is indebted to S. S. Goldich for supplying the samples listed in Table 11. This work was supported in part by the Earth Sciences Section, National Science Foundation, NSF Grant DES74-13305. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Baadsgaard, H., and Sandell, E. B., Analytica Chim. A d a , 1954, 11, 183. Maxwell, J . A., “Rock and Mineral Analysis,” Interscience Publishers, New York, 1968, p. 394. Riley, J. P., Analytica Chim. Acta, 1958, 19, 425. Suhr, N . H . , and Ingamells, C. O., Analyt. Chem., 1966, 38, 730. Medlin, J . H., Suhr, N. H., and Bodkin, J. B., Atom. Absorption Newsl., 1969, 8, 25. Medlin, J. H., Suhr, N. H., and Bodkin, J. B., Chem. Geol., 1970, 6, 143. Ingamells, C. O., Analyt. Chem., 1966, 38, 1228. Fogg, D. N., and Wilkinson, N. T., Analyst, 1958, 83, 406. Fleischer, M., Geochim. Cosmochim. Acta, 1969, 33, 65. Flanagan, F. J., Geochim. Cosmochim. Acta, 1973, 37, 1189. de la Roche, H., and Govindaraju, K., Mkth. Phys. Analyse, 1971, 7, 314. Sine, N. M., Taylor, W. O., Webber, G. R., and Lewis, C. L., Geochim. Cosmochim. Acta, 1969,33, 121. Received June 19th, 1975 Accepted August 20th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9760100044

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, January, 1976, Vol. 101,pp. 49-54 49 Determination of Benzoic and Sorbic Acids in Orange Juice Tamar Gutfinger, Rina Ashkenazy and A. Letan Department of Food Engineering and Biotechnology, Technion (Israel Institute of Technology) , Haifa, Israel A procedure has been devised for the separate determination of benzoic and sorbic acids in orange juice. It is based on the steam distillation of these preservatives from an acidified juice into an alkaline trap. The distillate is divided into two portions, each of which is treated with potassium dichromate and sulphuric acid, but under different conditions. Drastic oxidation leads to the destruction of sorbic acid and, after re-distillation, allows the determination of benzoic acid at 225 nm (with no interference from sorbic acid).Milder oxidation selectively converts sorbic acid into malonaldehyde, and the latter is subsequently converted into a coloured compound and determined at 532 nm (with no interference from benzoic acid). Recoveries of both preservatives from orange juice were satisfactory. A direct spectrophotometric method of analysis for the determination of both sorbic and benzoic acids has been described by various investigat~rs.l-~ Generally, the methods are based on measurements of light absorption at two wavelengths (approximately 225 and 255 nm) of either an ether extract or a distillate containing the above acids. The procedure is fairly rapid, but its accuracy is rather poor. In the region of 225 nm (the wavelength of maximum absorbance of benzoic acid) the spectrum given by a sorbate has a steep slope and its absorbance is high in comparison with that of a benzoate.This coincidence leads to erroneous readings when both benzoate and sorbate are present in the same solution. In addition, determinations in the ultraviolet region may yield results that are too high, owing to the presence of various light-absorbing substances from the examined food product in the extract or the distillate. Several workers reported better results by using a blank which they derived from a sample without preservative^.^,^ Such a blank is not usually obtainable, because with a commercial sample it is difficult to have knowledge of all of its ingredients ; the presence and concentrations of these ingredients will depend also on technological processes and/or on conditions of storage.Another group of workers6 used a blank that was obtained by distillation of a neutralised sample (from which the investigated preservatives will not volatilise). Such a blank has drawbacks, however, as it does not reflect those interfering components which will steam distil from an acidified sample. Several investigators7-10 successfully determined sorbic acid in various food products by the method of Schmidt,' which is based on the colorimetric measurement of the condensation product of malonaldehyde (a sorbic acid derivative) with thiobarbituric acid (TBA) . The proposed procedure for the determination of benzoic and sorbic acids in orange juice (see Fig. 1) calls for the determination of each of these preservatives under conditions that render the solution devoid of interferences caused by the other preservative and by inter- fering substances in the juice itself.Benzoic and sorbic acids are distilled off from an acidified juice. A portion of the distillate is subjected to oxidation with potassium dichromate under drastic conditions. Sorbic acid and volatile compounds from the juice are destroyed by this process and the unaffected benzoic acid is re-distilled and determined at 225 nm. The re- mainder of the distillate is subjected to oxidation under much milder conditions (also with potassium dichromate) . Malonaldehyde is formed and subsequently converted into a highly coloured component; the latter is determined at 532 nm,' while neither benzoic acid nor volatile compounds from the juice interfere. Experimental Apparatus Distillation apparatus.See Fig. 2. Spectrophotorneter. Visible and ultraviolet ranges.50 GUTFINGER, ASHKENAZY AND LETAN : DETERMINATION OF Analyst, Vol. 101 Juice (contains NaB and KS) Acidification L Distillation I (contains NaB and KS) 4 1 Condensation of malonaldehyde with TBA I I Colourz; givative Spectro photometric determination (at 532 nm) of the coloured condensation product (sorbic acid derivative) Distil lati on contains only NaB) I .c Spectrophotometric determination (at 225 nm) of NaB Fig. 1. Scheme for the determination of sodium benzoate (NaB) and potassium sorbate (KS) in orange juice. (a), Pathway of the blank distillate (blank 1) (see Experimental); ( b ) , pathway of the main distillate (and of blank 2) (see Experi- mental) ; (c), oxidation with 0.2 N K,Cr,O, and 4 N H,SO,.Sorbic acid and volatiles from the juice are oxidised to substances that do not absorb a t 225 nm. (d), Oxida- tion with 0.01 N K,Cr,O, and 0.3 N H,SO,. Malonaldehyde is produced by oxidation of sorbic acid. Reagents Analytical-reagent grade reagents are to be used throughout the determination. Phenolphthalein solution. Dissolve 1 g of phenolphthalein in 100 ml of ethanol (95 per cent.). Sodium suiphate, anhydrous, granular. Orthophosphoric acid, 85 per cent. Sodium hydroxide solution, 0-1 N. Dissolve 2 g of sodium hydroxide in 500 ml of water. Sodium hydroxide solution, 0.01 N. Dilute 10 ml of 0.1 N sodium hydroxide solution to 100 ml with water. Potassium dichromate solution, 0.2 N.Dissolve 4.9 g of potassium dichromate in water and dilute to 500 ml with water. Potassium dichromate solution, 0-01 N. Dilute 5 ml of 0.2 N potassium dichromate solution to 100 ml with water. Sulphuric acid, 4 N. Dilute 66.5 ml of sulphuric acid (concentrated) to 500 ml with water. Sulphuric acid, 0-3 N. Dilute 7.5 ml of 4 N sulphuric acid to 100 ml with water, Thiobarbduric acid solution, 0.5 per cent. Dissolve 0-5 g of thiobarbituric acid in 20 ml of water plus 10 ml of 1 N sodium hydroxide solution, add 11 ml of 1 N hydrochloric acid and dilute the mixture to 100 ml with water. Sodium benzoate stock solution, 100 j5.p.m. Dissolve 100 mg of sodium benzoate (which has has previously been dried at 105 "C for 2 h) in water in a 1000-ml calibrated flask.Add 100 ml of 0.1 N sodium hydroxide solution and dilute to 1000 rnl with water. Dilute 20 ml of the stock solution to 100 ml with water. Prepare the solution fresh daily. Sodium benzoate working solution, 20 j5.p.m.January, 19 76 has previously been dried at 105 "C for 2 h) in water in a 1000-ml calibrated flask. of 0.1 N sodium hydroxide solution and dilute to 1000 ml with water. sorbate to 100 ml with water. BENZOIC AND SORBIC ACIDS I N ORANGE JUICE 51 Potassium sorbate stock solution, 100 p.p.m. Dissolve 100 mg of potassium sorbate (which Add 100 ml Potassiwn sorbate working solution, 2 $.p.m. Dilute 2 ml of the stock solution of potassium Preliminary Experiments In these experiments blank 1 was prepared from a neutralised juice according to the method described under Preparation of blank distillate (blank 1) and blank 2 from an acidified juice according to the method described under Preparation of main distillate.Several experiments were performed on juice containing no preservatives. Procedure Preliminary operations Place an accurately weighed sample of about 5 g of juice into a 100-ml calibrated flask. Add 50 ml of water to the juice and neutralise the sample with 0.1 N sodium hydroxide solu- tion until the appearance of a pink colour in the presence of phenolphthalein. Dilute the solution to 100 ml with water. (In our experiments distilled water or freshly pressed and pasteurised orange juice was used for preparation of the model solutions of benzoic and sorbic acids.) Preparation of main distillate Pipette 20ml of the neutralised sample solution (above) or a portion containing about 200-1000 pg of sodium benzoate and about 100-500 pg of potassium sorbate into a 250-ml distillation flask (A, Fig.2). Next add 1 ml of orthophosphoric acid, 20 g of anhydrous sodium sulphate and about 30 ml of water and heat the solution vigorously over a gas flame. Adjust the rate of heating so as to obtain 35 ml of distillate in about 10 min and collect the distillate in a 50-ml calibrated flask (C, Fig. 2) that contains 5 ml of 0.1 N sodium hydroxide solution. Then rinse the condenser (B, Fig. 2) and dilute the distillate to 50 ml with water. Fig. 2. Apparatus for distillation of benzoic and sorbic acids.ll A, distilla- tion flask, 250 ml; B, condenser; C, calibrated flask, 50 ml.Dimensions are in centimetres. Preparation of blank distillate (blank 1) acid. Repeat the above procedure using the neutralised juice without addition of orthophosphoric52 GUTFINGER, ASHKENAZY AND LETAN : DETERMINATION OF Analyst, VoZ. 101 Determination of benzoic acid Preparation of standard graph. Transfer 1-, 2-, 3-, 4- and 5-ml portions of the standard sodium benzoate working solution into 10-ml calibrated flasks, each containing 1 ml of 0.1 N sodium hydroxide solution. Dilute the solution in each flask to the mark with distilled water and mix. Read the absorbance of each solution at 225 nm against 0.01 N sodium hydroxide solution and plot a graph of the concentration of sodium benzoate (in p.p.m.) versus absorbance (at 225 nm). Place 25 ml of each distillate (main and blank) in separate distillation flasks of 250-ml capacity. Add 25 ml of 0.2 N potassium dichromate solution and 6.5 ml of 4 N sulphuric acid to each.Heat in a boiling water bath for exactly 10 min, then cool and proceed as described under Preparation of main distillate, commencing at “Next add 1 ml of orthophosphoric acid . . . .” Measure the absorbance values of the two solutions at 225 nm against 0.01 N sodium hydroxide solution. Subtract the absorbance of the blank from that found for the main solution, then from the standard graph find the concentration of sodium benzoate corre- sponding to the corrected absorbance. Calculate the content of sodium benzoate in the juice as follows : Determination in distillate. Sodium benzoate, p.p.m. = 100 M/W where M p.p.m.is the concentration of sodium benzoate in the portion of distillate (corres- ponding to the value found from the calibration graph), W g is the mass of sample taken for distillation and 100 is the dilution factor. Determination of sorbic acid Preparation of standard graph. Transfer 0-, 1-, 2-, 3-, 4- and 5-ml portions of the working solution of potassium sorbate into separate 10-ml calibrated flasks and dilute them to 5 ml with water. To each flask add 1 ml of 0.01 N potassium dichromate solution and 1 ml of 0.3 N sulphuric acid and mix. Heat the solutions in a boiling water bath for exactly 5 min, add 2 ml of thiobarbituric acid solution and heat them again in a boiling water bath for an additional 10 min (exactly). Cool and dilute the solutions to 10 ml with water, then read their absorbance values at 532 nm against a reagent blank.Plot a graph of the concentration of potassium sorbate (in p.p.m.) versus absorbance (at 532 nm). Place 1 ml of each distillate (main and blank) in separate 10-ml calibrated flasks. Dilute each to 5ml with water and proceed as described under Preparation of standard graph, commencing at “add 1 ml of 0.01 N potassium dichromate . . . .” Subtract the absorbance of the blank from that found for the main solution and from the standard graph find the concentration of potassium sorbate corresponding to the corrected absorbance. Calculate the potassium sorbate content in the juice as follows: Determinatiorz in distillate. Potassium sorbate, p.p.ni. = 500 M/W where M p.p.m. is the concentration of potassium sorbate in the portion of distillate (corre- sponding to the value found from the calibration graph), W g is the mass of sample taken for distillation and 500 is the dilution factor.Results and Discussion It can be seen from Table I that when the single-distillation procedure2 was used, blank 1 (a distillate from neutralised juice) did not have contributions to the absorbance of light in the ultraviolet region from some of the volatile substances in the juice which were present in the distillate from acidified juice. Higher absorptions were always recorded for blank 2 (a distillate from acidified juice of the same origin, but without preservatives). In fact, the latter blank should be used in the determinations of benzoic and sorbic acids described, but this is, of course, impossible as no samples of juice without preservatives of the same origin and with the same history of processing and storage will be available at the time of deter- mination of the preservatives in a commercial sample.An analyst operating according to the procedure of Monselise2 is compelled to use blank 1 and will therefore obtain results that are too high (owing to the lower absorbance of that blank, see above). The proposed method overcomes the difficulty by concluding the procedure in two stagesJanzlary, 1976 BENZOIC AND SORBIC ACIDS I N ORANGE JUICE 53 TABLE I ABSORBANCE OF BLANKS 1 AND 2 OBTAINED FROM ORANGE JUICE WITH NO PRESERVATIVES Determinations were made in triplicate. Mass of juice taken for Sample A Sample B Procedure distillation/ c-------A-------I I A 1 g Blank 1* Blank 2t Blank 1* Blank 2 t h = 225 nm- Single distillation2 .... 5 0.1 14 0.204 0,169 0.191 Proposed .. .. .. 5 0.088 0.096 0.126 0.130 Single distillation* .. .. 1 0.040 0-097 0.034 0.105 Proposed .. .. .. 1 0.049 0.054 0.033 0.029 Proposed .. .. .. 6 0~000 0.008 0.012 0.017 0.006 Proposed .. .. .. 1 0.000 0~000 * From neutralised juice, see Preliminary experiments. t From acidified juice, see Preliminary experiments. h = 632 nm- 0.006 (see Fig. 1). (i) In one portion of the distillate from the acidified juice, interfering volatile compounds are eliminated by drastic oxidation ; this is followed by re-distillation of benzoic acid. Sorbic acid is also destroyed during the oxidation, which is advantageous as the ben- zoic acid in the re-distilled liquid can be determined without interference from the other preservative.(ii) In another portion of the distillate from the acidified juice, sorbic acid is converted first into malonaldehyde and then into a coloured derivative ; the absorbance of the latter is read in the visible-light region, at a wavelength at which the presence of neither the volatile compounds from the juice nor benzoic acid can interfere. It has been shown (Table I) that in the proposed procedure the respective readings at 225 nm (determination of benzoic acid) and 532 nm (determination of sorbic acid) were virtually the same for blanks from either neutralised or acidified juice with no preservatives. Linear graphs of absorbance versus concentration were obtained for sodium benzoate and potassium sorbate up to 10 p.p.m.(with absorbances of up to 0.6) and 1 p.p.m. (with absorbances of up to 0.3), respectively. TABLE I1 RECOVERY OF SODIUM BENZOATE AND POTASSIUM SORBATE FROM MODEL SOLUTIONS IN WATER OR IN ORANGE JUICE* Determinations were made in duplicate. Mass of sample taken Concentration, for distillation/ Solution p.p.m. g I n watev- Sodium benzoate .. .. 40 5 Potassium sorbate . . .. 40 Sodium benzoate .. .. 400 Potassium sorbate . . .. 400 I n orange juice- Sodium benzoate Potassium sorbate Sodium benzoate Potassium sorbate Sodium benzoate Potassium sorbate Sodium benzoate Potassium sorbate Sodium benzoate Potassium sorbate . . .. .. .. .. .. .. .. 40 40 400 150 300 300 500 500 1000 500 1 5 2 1 1 1 Recovery of each component, per cent.97 98 97 96 99 104 96 109 98 99 104 92 92 9754 GUTFINGER, ASHKENAZY AND LETAN The results obtained for the determination of benzoate and sorbate in model solutions are summarised in Table 11. The distillation (or re-distillation) time required for recovery of the preservatives was 10 min. The recoveries of both preservatives were satisfactory (92-104 per cent. for benzoate and 92-109 per cent. for sorbate). In conclusion, the proposed method facilitates the determination of benzoic and sorbic acids with no interference from each other or from light-absorbing (in the ultraviolet region) volatile compounds in the juice. While being more elaborate than the other meth~ds,l-~ it remains fairly simple, as each step can be easily and rapidly carried out. The results obtained by use of this method are satisfactory and reproducible. The method was tested with orange juice but the authors believe that it can also be used. for other fruit and vegetable juices. References 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. Lorenzen, W., and Sieh, R., 2. Lebensmittelunters. u. -Forsch., 1962, 118, 223. Monselise, J. J., Israel J . Technol., 1969, 7 , 263. Gantenbein, W. M., and Karasz, A. B., J , Ass. Off- Analyt. Chem., 1969, 52, 738. Luckman, F. H., and Melnick, D., Fd Res., 1966, 20, 649. English, E., Analyst, 1969, 84, 466. Monselise, J. J., Friedman, Y., and Fathi, R., Israel J . Technol., 1970, 8, 437. Schmidt, H., 2. Analyt. Chem., 1960, 178, 173. Nury, F. S., and Bolin, H. R., J . Fd Sci., 1962, 27, 370. Wilamowski, G., J . Ass. Off. Analyt. Chem., 1971, 54, 663. Caputi, A., and Slinkard, K., J . Ass. Off. Analyt. Chem., 1976, 58, 133. Israeli Standard No. 862, 1973. Received August 7th, 1976 Accepted September llth, 1975

ISSN:0003-2654    

DOI:10.1039/AN9760100049

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

Analyst, January, 1976, VoL. 101, fip. 55-61 55 Application of the Tubular Graphite Electrode in the Measurement of Reaction Kinetics: Development of an Automatic Technique L. R. Sharma, V. P. Soi, J. D. Sharma and Ramesh Kumar Kalia Department of Chemical Engineering and Technology, Punjab University, Chandigarh 160014, India The unique feature of the tubular electrode, that it permits continuous analysis of an electrolyte solution passing through it, has been made use of in develop- ing an automatic voltammetric technique for studying the kinetics of reactions involving one or more electro-active reactants or products. The technique provides for automatic dilution of the reaction mixture with the supporting electrolyte in order to bring the concentrations within the voltammetric range, automatic freezing of the reaction when required, auto- matic stirring of the reaction mixture solution and, especially, automatic recording of the current - time graph, from which the kinetics of the reactions can easily be studied.The technique has been standardised by studying voltammetrically the reaction kinetics of the oxidation of potassium iodide with hydrogen peroxide in aqueous acidic solutions and comparing the data with those obtained titrimetrically. The results appear to show that the automatic technique has a high dependability. Encouraging results have been obtained in the study of the reaction kinetics of the oxidation of potassium hexacyano- ferrate(I1) with hydrogen peroxide and of sodium formate with potassium permanganate in acidic media.The tubular graphite ele~trodel-~ can be conveniently used for the continuous analysis of an electro-active solution passing through it at a given flow-rate. It was thought of interest to use this unique property in the study of kinetics of reactions involving one or more electro- active reactants or products, and particularly to develop and standardise an automatic technique for this purpose. It is proposed that the reaction should proceed in a vessel such as a gravity feed with constant head with the reactants at any desired concentration. The reaction mixture coming out of the vessel is then diluted to the concentrations required for the voltammetric measurements by mixing it with a pre-chilled solution of the supporting electrolyte before it passes through the electrode.The dilution process also has the essential function of freezing the reaction at that instant. Experimental Apparatus The complete assembly, designed and fabricated for the present work, is shown in Fig. 1 (a). The main component of the assembly is the mixer, which is required for the instantaneous and thorough mixing of the reaction mixture with the supporting electrolyte solution. The mixer is shown separately in Fig. 1 (b). It consists of a cylindrical glass tube about 5 cm in diameter and about 13 cm in length, with a slanting plate, P, across the centre and a delivery tube, D, near the top. The plate has a hole of about 7 mm diameter towards its lower end. The reaction under investigation is allowed to proceed in the reservoir of the double-jacketed gravity feed GI and the supporting electrolyte solution is placed in the reservoir of the gravity feed G,, which is surrounded by an ice jacket.The reaction mixture solution, flowing out of the jet, J, at a very small flow-rate that is controlled by adjusting the length of the tube TI of the gravity feed GI, is picked up by the pre-cooled supporting electrolyte solution, which flows from the delivery tube, D, at a much higher rate (about 100 times higher). This latter flow-rate is controlled by the length of the tube T, of the gravity feed G, and also by the height of G, with respect to the mixer. The two solutions are mixed fairly well in this way and the reaction is simultaneously brought almost to a standstill. Greater opportunity for mixing is provided as the solutions pass through the mixer, as illustrated in Fig.1. It is56 SHARMA et aZ. : APPLICATION OF THE TUBULAR GRAPHITE Analyst, VoZ. 101 expected, therefore, that the solution which comes out of the mixer through the side tube, S, is composed of a completely uniform mixture of the two solutions. As only a limited amount of the solution is to be passed through the tubular electrode, a major portion of the solution is by-passed through the overflow tube, Q, which is fixed in such a way as to allow only a negligible retention of the solution [Fig. 1 (b)]. Electrolytic solution from gravity feed GI Supporting electrolyte gravity feed G2 from - To thermostat t II Fig. 1. methods. text. (a), Apparatus designed for use in the study of reaction kinetics by automatic ( b ) , Mixer on an enlarged scale.Lettered components are identified in the T.G.E. = tubular graphite electrode. The efficiency of the mixer in bringing about instantaneous and uniform mixing of the solutions was tested by mixing a solution of methylene blue from gravity feed G, with a large excess of water from gravity feed G,, collecting portions of the mixture coming out of the side tube, S, at different intervals of time and subjecting them to colorimetric analysis. The fact that the concentration of dye solution in the first portion was found to be exactly the same as that in subsequent portions, shows that the mixer is capable of effecting a satisfactory mixing of the two solutions. Selection of the Reaction for Kinetic Measurements The oxidation in acidic medium of potassium iodide by hydrogen peroxide 2KI + H,O, -+ 2KOH + I, was considered to be a suitable reaction with which to study the kinetics by voltammetric methods, as the concentration of one of the reactants, namely with the iodide ions, in theJanuary, 1976 ELECTRODE IN THE MEASUREMENT OF REACTION KINETICS 57 reaction mixture could easily be determined by carrying out its electro-oxidation at the tubular graphite elect rode.Blank Experiment M in potassium iodide in the gravity feed G, and 0.1 M sulphuric acid in the gravity feed G,. The iodide solution was maintained at 25 "C by circulating water from the thermostat at 25 "C through the outside jacket of the gravity feed G,. The supporting electrolyte solution was cooled to 5 "C by surrounding the gravity feed G, with crushed ice.The flow-rate of the iodide solution was adjusted to 5 ml min-l by altering the length of the tube T, dipping into the iodide solution. Similarly, the flow-rate of the supporting electrolyte solution (measured by the rotameter, R) was adjusted to 495 ml min-l by altering the height of the gravity feed G, and the length of the tube T, dipping into the solution. The iodide solution coming out of the mixer and then passing through the electrode had a concentration of 10-4~. A complete voltammogram for the oxidation of iodide ions was then scanned by following the procedure described earlier., The flow-rate of the solution passing through the electrode was measured accurately and the limiting current at a flow-rate of 10.8 ml min-l was found to be 19.5 PA.Voltammograms were also run for mixtures of potassium iodide and iodine solutions of appropriate concentrations. In each instance, an extremely well defined voltammogram with a well defined diffusion plateau was obtained, irrespective of the presence or absence of iodine in solution. The limiting current was found to be directly proportional to the concen- tration of the iodide ions. A blank experiment was run by placing 0.1 M sulphuric acid that was Automatic Kinetic Measurements The following procedure was adopted for the study of the kinetics of the reaction selected. Equal volumes of 0.02 M potassium iodide and 0.02 M hydrogen peroxide solutions, both kept thermostatically at 25 "C, were mixed in a beaker and at the same instant the chart drive of the polarograph was switched to the "on" position.The mixture was immediately transferred to the reservoir of the gravity feed G, and the tube T, introduced into the vessel. Water at 25 "C was circulated around the reservoir continuously. The stopcock S, of the gravity feed G, was kept open and the supporting electrolyte solution, 0.1 M sulphuric acid cooled to 5 "C, was kept flowing through the mixer at a constant rate of 495 ml min-1. The stopcock S, of the gravity feed G, was then opened and concurrently (at time t,) a mark was put on the time axis of the recorder chart. As the retention of the solution by the capillary jet below the stopcock S, was extremely small, the reaction mixture, flowing a t a rate of 5 ml min-1, took only a negligible time to clear the jet.As soon as the solution came out of the jet, J, it met the current of pre-chilled electrolyte solution, which resulted in the requisite dilution as well as in the freezing of the reaction at that instant. After passing through the mixer the mixed solution entered the tubular graphite electrode, which was maintained at a potential of 0.75 V versus S.C.E. This potential lies well within the range for the diffusion plateau of the voltammogram obtained for the oxidation of iodide ions. As soon as the reaction mixture passed through the electrode a limiting current proportional to the concen- tration of the iodide ions in the mixture was recorded on the chart. It is evident that the limiting current thus recorded corresponded to that concentration of iodide ions present in the solution at the instant of opening the stopcock S,.With the progress of the reaction, there was a continuous decrease in the concentration of the iodide ions and a corresponding decrease in the magnitude of the limiting current, recorded automatically on the recorder chart. Thus a current - time graph (which is equivalent to a concentration - time graph) was recorded automatically during the oxidation of potassium iodide by hydrogen peroxide. This graph is shown in Fig. 2. The horizontal portion, from zero to A of the graph represents the residual current when only the supporting electrolyte solution from the gravity feed G, is passing through the electrode. The portion AB represents the sudden increase in current as a result of the oxidation of the iodide ions present in the reaction mixture.The maximum current represented by the point B corresponds to the concentration of iodide ions at the instant of opening the stopcock S,, as already discussed. The portion BCE of the curve represents the decrease in the concen- tration of the iodide ions with the progress of the reaction.58 SHARMA et al. : APPLICATION OF THE TUBULAR GRAPHITE Analyst, Vol. 101 It should be kept in mind that the maximum current represented by the point B corres- ponds to time tl and not to time t,. The time interval t, - t, corresponds to the time period between the opening of the stopcock S, and the recording of the steady value of the limiting current for that iodide concentration on the recorder chart.This time period must be elimin- ated when working out the kinetics of the reaction. 22 20 18 16 4 14- 1 2 - a 8 - 6 - 4 - 2 - 2 E 10- B - - - A I I I I Fig. 2. Automatically recorded current - time graph for the oxidation of 0.02 M potassium iodide (100 ml) with 0.02 M hydrogen peroxide (100 ml) in acidic medium. It should also be noted that the gravity feed G, provides a constant-pressure head only when the air, after displacing the liquid contained in the tube TI, starts to bubble out of its lower end. Before that stage the pressure head is not constant. This accounts for the somewhat irregular portion BC of the curve BCE. It is thus only the portion CE of the graph that represents values of limiting currents obtained at a constant flow-rate of the solution. Hence it is this portion which should be taken into account when studying the kinetics of the reaction.The extent of the irregular portion of the curve can easily be minimised by decreasing the diameter of the glass tube T,. The smaller the diameter, the less is the amount of solution contained in it and hence the time required for displacement of that solution. The bubbling of air into the solutions serves an additional but important purpose. It keeps the solution in the reaction vessel properly stirred. By decreasing the diameter of the tube, the rate of bubbling increases and this provides for better stirring. Thus, the gravity-feed device not only provides a constant-pressure head and thereby a constant flow-rate of the solution but it also provides for automatic stirring of the reaction mixture, which is essential in the study of kinetics. Comparison of the Kinetic Data Obtained Voltammetrically with Those Obtained Titrimetrically It was thought of interest to compare the kinetic data obtained voltammetrically for the reaction under investigation with those obtained titrimetrically. For this purpose, the current - time data of the graph CE were converted into concentration - time data, making use of the relationship between current and concentration obtained in the blank experiment and also of the Levich equation for the tubular e l e ~ t r o d e , ~ ~ ~ ~ ~ so that differences in the working parameters (e.g., flow-rate and electrode length) in the two experiments would be taken into account.The concentration - time data thus obtained are recorded as a graph in Fig.3. The kinetics of the selected reaction were also studied by following the usual titrimetric procedure. The concentration - time data obtained in this way are also shown in Fig. 3. It is interesting to note that the concentration - time graphs obtained are almost superimpos- able. As the initial concentrations of both of the reactants were the same in both experi-January, 1976 ELECTRODE IN THE MEASUREMENT OF REACTION KINETICS 59 X Lc 0 0.40 1 I I I I 1 00 200 300 400 Timels Fig. 3. Comparison of the kinetic data for the oxidation of potassium iodide with hydrogen peroxide obtained voltammetrically (curve A) by use of auto- matic methods with those obtained titrimetrically (curve B) in the usual way.ments, the rate of reaction would be the same in both instances. Therefore, the concen- tration of potassium iodide at different intervals of time, whether determined voltammetric- ally or titrimetrically, should be the same provided that the mixing and dilution of the reaction mixture in the mixer is perfect. The almost complete superimposition of the two concentration - time graphs clearly shows that this is so. These observations confirm the dependability of the technique that has been developed for the automatic recording ofthe current - time graph (which is equivalent to the concentration - time graph) in the study of kinetics of reactions involving electro-active species. TABLE I KINETICS OF THE OXIDATION OF POTASSIUM IODIDE WITH HYDROGEN PEROXIDE I N 0.1 M SULPHURIC ACID Volumes of KI and H,O,, 100 ml; potential applied to the electrode, 0-75 V versus S.C.E.; automatic dilution of the reaction mixture with the supporting electrolyte, 100 times; flow-rate of the diluted reaction mixture solution through the tubular electrode, 15 ml min-l. Concentration of KI solution Concentration passing through of KI in reservoir Limiting the electrode/ GI a t time Time (t)/s current/pA M x 10-4 t / M X lo-' With initial concentrations of KI and H,O, of 0.02 M- 60 16.6 0.874 90 15.1 0.795 120 14.1 0.742 160 12.8 0.674 200 11.9 0.626 250 10.9 0,574 300 10.0 0.526 400 8.6 0.447 600 7.3 0.384 0.874 0.795 0.742 0.674 0.626 0.5 74 0.526 0.447 0.384 With initial concentrations of KI and H202 of 0.01 M- 60 8-8 0-463 0.463 90 8.4 0.442 0.442 120 8.0 0.42 1 0.42 1 150 7.7 0.405 0.405 200 7.3 0.384 0.384 300 6.5 0.342 0.342 400 5.8 0.305 0.305 Rate constant for second- order reaction/ 1 mol-1 s-l 0.240 0.287 0.290 0-302 0.299 0.297 0.300 0.309 0.32 1 0-266 0.292 0.313 0.313 0.302 0.308 0.32060 SHARMA et al.: APPLICATION OF THE TUBULAR GRAPHITE Analyst, VoZ. 101 Analysis of the Current - Time Graph The portion CE of the current - time graph (Fig. 2) was analysed in order to study the kinetics of the reaction under investigation. The oxidation of potassium iodide with hydrogen peroxide in an acidic medium is known to be a second-order reaction.8 This order is confirmed by the constancy of the values for the second- order rate constants obtained in the present investigations and given in the last column of Table I.The average value of the rate constant is 0.30 1 mol-l s-l. The oxidation of potassium iodide was also studied for initial concentrations of potassium iodide and hydrogen peroxide of 0.01 M (with 0.1 M sulphuric acid). The voltammetric current - time graph was recorded automatically, as before, and then analysed. The data obtained are given in Table I. It can be seen that the values of the rate constants corres- ponding to different time intervals are close to the average value of 0.30 1 mol-l s-l. For a further standardisation of the automatic technique, the current - time graphs were obtained for different initial concentrations of potassium iodide and hydrogen peroxide. On analysis of each of the curves the order of the reaction was found to be 2 and the average value for the rate constant to be 0.30 1 mol-1 s-1.The relevant data are given in Table I. TABLE I1 KINETICS OF THE OXIDATION OF POTASSIUM HEXACYANOFERRATE(II) WITH HYDROGEN PEROXIDE IN 0.1 M SULPHURIC ACID Volume of each solution, 100 ml; potential applied to the electrode, 0.4 V versus S.C.E. : automatic dilution of the reaction mixture with the supporting electrolyte, 100 times; flow-rate of the diluted reaction mixture solution through the tubular electrode, 15 ml min-l ; initial concentration of hexacyanoferrate(I1) solution, 0.02 M ; initial concentration of H,O, solution, 0.02 M. Concentration of hexacyano- ferrate(I1) Concentration solution passing of hexacyano- Rate constant through the ferrate(I1) in for second- Limiting electrode/ reservoir G, at order reaction/ Time ( t ) / s current/pA M x lo-* ti met/^ x 10-2 l r n o l - l ~ - ~ 60 12.2 90 10.8 120 9.9 160 9.1 200 8.0 250 7.2 300 6.6 0-813 0.720 0.660 0.607 0.533 0.480 0.433 0.813 0.383 0.720 0.432 0.660 0.429 0.607 0.432 0.533 0.438 0.480 0.433 0.433 0.437 TABLE I11 KINETICS OF THE OXIDATION OF SODIUM FORMATE WITH POTASSIUM PERMANGANATE IN 0.1 M SULPHURIC ACID Volume of each solution, 100 ml; potential applied to the electrode, 0.40 V versus S.C.E.; automatic dilution of reaction mixture with the supporting electrolyte, 100 times; flow-rate of the diluted reactioh mixture through the tubular electrode, 16 ml min-l; concentration of sodium formate solution, 0.2 M ; concentration of potassium permanganate solution, 0.004 M.Concentration of KMnO, Concentration solution passing Concentration of sodium Rate constant through the of KMn0,in formate in for second- Limiting electrode/ reservoir G, at reservoir G, at order reaction/ Time ( t ) / s current/pA M x loA5 time t / ~ x time t / ~ x loda 1 mol-1 s-l 60 90 150 200 300 400 600 600 13.4 12.2 10.3 8.9 6.5 4.5 3.2 2.2 1.489 1.356 1.144 0.989 0.720 0.500 0,356 0.244 1.489 1.356 1.144 0-989 0.720 0-500 0.356 0.244 98-7 98.04 97-85 97-48 96-80 96-25 95-87 95.6 0.0208 0.0180 0-01 60 0.01 50 0-0146 0-0149 0.0151 0.0153JanGary, 1976 ELECTRODE IN THE MEASUREMENT OF REACTION KINETICS 61 The automatic technique described above was also tried in studying the kinetics of the oxidation of potassium hexacyanoferrate( 11) with hydrogen peroxide and of the oxidation of sodium formate with potassium permanganate, both reactions taking place in 0.1 M sulphuric acid. The relevant portions of the automatically recorded current - time graphs were analysed as before and the corresponding data are recorded in Tables I1 and 111. In both instances the results obtained are encouraging. It was found that both reactions were of second order with a rate constant of 0.43 1 mol-l s-l for the oxidation of the hexacyanoferrate(I1) and of 0.015 1 mol-l s-1 for the oxidation of sodium formate. References 1. 2. 3. 4. 5. 6. 7. 8. Sharma, L. R., and Jatinder Dutt, Indian J . Chem., 1968, 6, 593. Sharma, L. R., and Jatinder Dutt, Indian J . Chem., 1968, 6, 697. Sharma, L. R., and Jatinder Dutt, Indian J . Chem., 1969, 7, 485. Sharma, L. R., and Jatinder Dutt, Indian J . Chem., 1970, 8, 173. Sharma, L. R., Jatinder Dutt, and Inderjit Nirdosh, BY. Chem. Engng., 1969, 14, 1220. Levich, V. G., “Physico-Chemical Hydrodynamics,’’ Prentice Hall, Englewood Cliffs, New J ersey, Blaedel, W. J., Olson, C. L., and Sharma, L. R., Analyt. Chem., 1963, 35, 2100. Glasstone, S., “Textbook of Physical Chemistry,” Second Edition, Macmillan & Co., London, 1974. Received July 17th, 1975 Accepted September 2nd, 1976 1962.

ISSN:0003-2654    

DOI:10.1039/AN9760100055

出版商:RSC    

年代:1976

数据来源: RSC

摘要:

62 Analyst, Janzcary, 1976, Vol. 101, pp. 62-66 Analytical Methods Committee REPORT PREPARED BY THE METALLIC IMPURITIES IN ORGANIC MATTER SUB-COMMITTEE The Use of 50 per cent. Hydrogen Peroxide for the Destruction of Organic Matter (Second Report) The Analytical Methods Committee has received and approved for publication the following Report from its Metallic Impurities in Organic Matter Sub-committee. Report The constitution of the Sub-committee responsible for the preparation of this Report was: Dr. L. E. Coles (Chairman), Mr. W. Cassidy, Mr. P. M. Coleman, Mr. G. Collett, Dr. W. H. Evans, Mr. S. Greenfield, Mr. W. H. Hill, Mr. D. A. Lambie, Dr. R. F. Milton, Mr. B. E. Pearce, Mr. W. L. Sheppard and Mr. C. A. Watson, with Mr. P. W. Shallis as Secretary and Mr. J. J. Wilson as Assistant Secretary.Introduction In 1967 the Sub-committee published its first Report1 on the use of 50 per cent. hydrogen peroxide for the destruction of organic matter. With the increase in demand for methods of analysis involving wet digestion as a preliminary stage, it has been considered desirable to update the original report, drawing on the collective experience of the members of the Sub-Commit tee. The first Report claimed certain advantages for the use of 50 per cent. hydrogen peroxide over the use of the more conventional oxidising acids, although it was realised that the use of this material was not without some hazard. The hazards involved and the safety pre- cautions to be observed were detailed in the Report, and are not reproduced here in full. The Report gave specific procedures for the use of 50 per cent.hydrogen peroxide for the wet oxidation of A, plastics; B, readily oxidisable materials; C, soft drinks, etc. ; D, syrups; and E, herbs, spices, gums, etc. It must be emphasised that 50 per cent. (approximately 160-volume) hydrogen peroxide is a potentially dangerous chemical (see section on Hazards and Safety Precautions in the first Report1). Peroxide that has been splashed on to the skin can be “neutralised” with dilute ammonia solution, followed by a thorough rinse with water. This procedure should also be followed if there is any spillage of the hydrogen peroxide. It is a statutory require- ment in the UK that eye protection must be worn by all workers using materials such as 50 per cent.hydrogen peroxide.2 The advantages of the use of 50 per cent. hydrogen peroxide have been confirmed in practice, and the damage to fume cupboards has been found to be noticeably less than when the more conventional oxidising acids are used. Most of the work done by the Sub-committee in recent years has been on the application of atomic-absorption spectroscopy to the determination of metallic impurities in organic matter, and methods of wet oxidation have had to be adapted in order to analyse different types of organic material. Difficulties have been encountered in the analysis of fats and oils, foods with a high fat content, syrups and fruit concentrates. This Report deals with methods of wet digestion that have been developed by members of the Sub-committee for the analysis of these types of materials.It must be stressed that in no circumstances should workers attempt to apply either of the procedures detailed in this Report to the wet oxidation of materials of types other than those specified. It is also extremely important that the amounts of sample and reagents specified and the details of the procedures given should be adhered to strictly.ANALYTICAL METHODS COMMITTEE 63 Fats and Oils The direct presentation to the flame, in atomic-absorption spectroscopy, of fats dissolved in organic solvents is useful only when the sensitivity for and the concentration of the metal under examination come within the ranges that can be measured by use of the instrument. Flameless techniques for fats and oils can give improved sensitivity, but often with less accuracy and precision.Background absorption problems occur with both techniques, particularly at wavelengths of less than 250 nm. It is therefore difficult to determine accu- rately the trace-metal content of, for instance, margarine that contains salt, or milk products containing large amounts of calcium and phosphorus. The analyst therefore requires a method for destroying the fats before extracting the trace metals with reagents such as 1,5-diphenylthiocarbazone (dithizone) or ammonium pyrrolidine dithiocarbamate. Dry ashing as normally carried out on fats and oils results in ignition of the molten degraded material with consequent loss of trace metals. Ignition can be avoided, however, by pre- charring the samples in test-tubes held vertically on a hot-plate at 350 "C, followed by a gradual increase in temperature up to 550 "C in a closed (free from air turbulence) muffle furnace.Use of this procedure is restricted to samples of about 0.5 g, which does not improve the effective sensitivity of the analysis. Wet-digestion techniques allow a larger mass of sample to be handled, but reactions can be violent. It must be emphasised that the use of perchloric acid for the wet digestion of fats and oils is particularly dangerous, and if nitric acid is used for the wet digestion there is the possibility that it will react with any glycerol liberated from the comparatively large sample of fat to form nitroglycerine. The usual practice of digestion in a Kjeldahl flask is wasteful of nitric acid, the bulk of which is expelled into the fume cupboard with deleterious effects on both the cupboard and the operator unless stringent precautions are taken.In addition, large volumes of acids are undesirable in determinations of trace metals as they introduce high blank values and tend to leach trace metals from the glassware. The major problems normally encountered in wet digestions occur when too large a sample is contained in the reaction vessel. The method described here for fats and oils will effect a rapid and efficient wet oxidation based on the addition of small amounts of the sample followed by small amounts of peroxide to heated sulphuric acid. Discussion The method is based on the fact that molten fats and oils will form a thin layer on the surface of hot sulphuric acid.Many conical flasks have slightly domed bases, but it is ad- visable to use the minimum amount of sulphuric acid and therefore it is advantageous to select for the reaction a flask with the flattest possible base rather than a Kjeldahl flask, in order to extend the layer of acid. A degree of heating control is necessary, and this is best achieved by heating the flask on a hot-plate consisting of a sheet of steel plate mounted on a tripod over a Bunsen burner, as this allows the flask to be moved as necessary from a hot region to a cooler one. As stressed in the first Report, peroxide must always be used in the presence of sufficient sulphuric acid to prevent violent oxidation. The first indication that too little sulphuric acid is present is that a dry froth advances up the sides of the flask, then small light flashes occur, followed by a rapidly accelerating reaction that expels vapours upwards.Safe conditions for the reaction are an excess of sulphuric acid, a limited amount of sample and just sufficient hydrogen peroxide to oxidise the sample; the use of a wide-necked flask is an additional safety measure. The best conditions for rapid and efficient oxidation are those in which the acid is maintained at gentle fuming during the digestion of each small increment of sample, but care must be taken to avoid heating the acid too strongly as oxidation by peroxide is less efficient at higher temperatures and becomes relatively ineffective if charring occurs. Oxidation of a charred mass can, however, be assisted by adding a few drops of concentrated nitric acid and carefully heating until all nitrous fumes have been expelled.No problems have been experienced in the digestion of 3 4 g of margarine with about 25 ml of hydrogen peroxide by the method described below. It is, however, desirable and sometimes essential that all residual hydrogen peroxide is removed by the addition of a suitable reducing agent such as sulphur dioxide.64 Method A ANALYTICAL METHODS COMMITTEE: USE OF 50% HYDROGEN Analyst, Vol. 101 Caution. Strict adherence to the conditions of the procedure is essential, otherwise the method is potentially hazardous. A@aratus Conical $asks. Wide-necked, volume 500 ml. Hot-plate. A plate of 5-mm sheet steel (20 x 20 cm), mounted on a tripod over a gas Pasteur pipettes.Glass rods. Measuring cylinder. burner is particularly suitable. Reagents Sulphuric acid, concentrated. Any grade of known low metals content can be used. Hydrogen peroxide, 50 per cent. m/V. Store in a refrigerator. Procedure From a measuring cylinder transfer 5 ml of sulphuric acid to the conical flask, place the flask on the hot-plate, and apply heat. Adjust the position of the flask on the hot-plate so that the acid just evolves fumes. For an oil sample place a Pasteur pipette in the sample container or for a fat sample place two glass rods in the sample container, and weigh the container and contents. Transfer about 0-log of the sample on the surface of the acid in the flask, and swirl the flask until a film is formed on the surface of the acid.Partially fill a long Pasteur pipette with hydrogen peroxide, insert the pipette into the flask so that its tip is about 3 cm above the surface of the acid, and add hydrogen peroxide dropwise to different areas of the surface until the acid is colourless. Continue heating until all bubbling ceases and until acid fumes are evolved. Add approximately another 0.10 g of sample and digest in the same manner. (If, during the reaction, there is any evidence of insufficient sulphuric acid being present in the flask, add more acid before making further additions of either sample or peroxide,) Continue digesting small portions of the sample until sufficient has been oxidised for the analysis. Re-weigh the sample container and contents. Allow the digest to cool, and carefully dilute it with water.Foods, Syrups and Fruit Concentrates One of the samples used in a recent collaborative test carried out by the Sub-committee was a cheese savoury baby food, which was specially chosen for its behaviour during wet oxidation as it is more difficult to oxidise than most foods. Similarly, problems arise in the digestion of syrups and fruit concentrates with high solids contents. If these samples are treated by the procedures previously recommended,l and especially if several samples are dealt with at the same time, there is a high risk of excessive charring with consequent wastage of time and reagents. The method recommended in this Report for fats and oils could be used for these materials, but the addition of essentially aqueous samples is more difficult to effect than the addition of fats and oils, which float on the surface of the acid.The method described below involves the same basic principle of small and frequent additions, but in this instance the sample and peroxide are added together. Discussion The method depends on the miscibility of the sample with peroxide, which enables the reactants to be brought together in the reaction vessel simultaneously. Concern has been expressed about the mixing of peroxide with sample and the possible hazard of ignition or explosion. It is a fact that contact between 50 per cent. hydrogen peroxide and anhydrous organic fibres or powders is liable to result in spontaneous com- bustion, and examples have been quoted of fires starting on paper tissues, cloths, mops and brooms that are used in dealing with spillage of the material.The samples for which this procedure is recommended, however, have a high proportion of water and, in the experience of one member of the Sub-committee who has used this method for the last 7 years on aboutJanuary, 1976 PEROXIDE FOR THE DESTRUCTION OF ORGANIC MATTER 65 5000 samples each year, there is no danger provided that the temperature of the mixture is not allowed to rise above ambient. The safe conditions for the reaction are the same as for Method A, and the risk of violent oxidation is again minimised by making only small additions of the mixture at a time to the acid in the reaction vessel. Method B Caution. Strict adherence to the conditions of the procedure is essential, otherwise the method is potentially hazardous.Apparatus Beakers. Borosilicate glass, volume 100 ml. Conical $asks. Borosilicate glass , volume 250 ml. Pasteur pipettes. Measuring cylinder. Hot-plate. As for Method A. Reagents Procedure Place the weighed or measured sample into a beaker (usually 20 g or 20 ml will be sufficient, but a larger sample will improve the effective sensitivity of the method). Add an equal volume of hydrogen peroxide to the contents of the beaker, and mix well. Do not heat this mixture or leave it unattended. From a measuring cylinder transfer 5 ml of sulphuric acid to a conical flask, and place the flask on the floor of a fume cupboard. Partially fill a Pasteur pipette with the sample mixture and add about 2 ml of it to the acid in the flask.The heat evolved on addition of the sample mixture will usually initiate the oxidation reaction; if it does not, place the flask on the hot-plate and warm gently until the reaction begins, and then return the flask to the floor of the fume cupboard. Continue to add portions of about 2 ml of the sample mixture to the acid at regular intervals so as to maintain the reaction. After about five or six additions the volume of the reactants in the flask may have increased significantly. If this happens, transfer the flask to the hot-plate and complete the digestion with dropwise addition of peroxide alone until no further discoloration occurs. Return the flask to the fume cupboard floor, allow to cool, and then resume additions of the sample mixture. (If, during the reaction, there is any evidence of insufficient sulphuric acid being present in the flask, add up to a further 5 ml of acid before continuing with additions of the sample mixture.) When all of the sample mixture has been added, place the flask on the hot-plate and complete the digestion with dropwise addition of peroxide until the fuming sulphuric acid is colourless.Allow the digest to cool, and carefully dilute it with water. Conclusions Since the publication of the first Report the experience gained by members of the Sub- Committee in the use of 50 per cent. hydrogen peroxide has been of great value in planning the collaborative experiments carried out. The revised procedures recommended in this report provide quick, efficient and clean methods of preparing difficult samples for analysis, the skills required being easy to acquire.The methods can be applied with confidence by suitably instructed laboratory staff. Pressure-digestion Vessels The use of pressure-digestion vessels for the wet oxidation of organic matter with 50 per cent. hydrogen peroxide cannot be recommended. These vessels are PTFE-lined steel bombs that can be securely closed either by screw-threaded barrels or with high-tensile steel bolts. They are designed primarily for dissolving ores and metals in oxidising acids. An experiment designed to check the efficiency of oxidation with hydrogen peroxide in such a container resulted in violent fracture of the eight securing bolts. A sample of milk (4 ml) was placed66 ANALYTICAL METHODS COMMITTEE in the vessel and dried overnight at 98 "C; 10 ml of sulphuric acid were added, the vessel and contents were cooled to -10 "C, and 20 ml of 50 per cent. hydrogen peroxide, also at -10 "C, were added. The bomb was sealed and shaken at intervals while the temperature rose to ambient; when opened, it was found that no digestion had occurred. The bomb was re-sealed and placed in an oven at 60 "C for 2 h; when opened after cooling it was found that digestion had not reached completion. A further 10ml of peroxide were added, the bomb was re-sealed, placed in a muffle furnace and heated only gradually from ambient. At about 80 "C an explosion occurred. The eight securing bolts had fractured and the lid of the vessel had been projected into the lining of the furnace. References 1. 2. Analytical Methods Committee, Analyst, 1967, 92, 403. "The Protection of Eyes Regulations 1974," S.I. 1974 No. 1681, H.M. Stationery Office, London,

ISSN:0003-2654    

DOI:10.1039/AN9760100062

出版商:RSC    

年代:1976

数据来源: RSC