摘要:

Critical Review Analytical Chemistry of Fruit Bioflavonoids A Review Kevin Robards* and Michael Antolovich Charles Sturt University Riverina, P.O. Box 588, Wagga Wagga 2678, NSW Australia Summary of Contents Introduction Chemical Structure Significance Function in Plants Plant resistance Photoprotection Properties and Physiological Activity in Animals Uses Chemotaxonomy Fruit Quality Biosynthesis Sources Anthocyanic Pigments Flavonols Flavan-3-ols Flavanones Flavones Other Flavonoids Analysis Total Phenols Sample Preparation Sample clean-up Hydrolysis and glycoside analysis Thin-layer Chromatography Gas Chromatography High-performance Liquid Chromatography Detection Coupled Methods Other Techniques References Keywords: Review; fruit; bioflavonoids; phenolics; flavonoids Introduction The bioflavonoids are aromatic secondary plant metabolites belonging to the class of plant phenolics.Many of the latter were characterized during the classical period of organic chemistry but their recognition as a discrete group of genetically related plant metabolites came in 1957 with the formation in England of the Plant Phenolics Group, later the Phytochemical Society of Europe.The terms ‘phenolic’, ‘polyphenol’ and ‘phenols’ can be precisely defined chemically but, in the present context of plant phenols, such a definition is not entirely satisfactory since it would inevitably include compounds such as oestrone, the female sex hormone which is principally terpenoid in origin. For this reason, a definition based on metabolic origin is preferable, the plant phenols being regarded as those substances derived from the shikimate pathway and phenylpropanoid metabolism (Fig. 1). The major classes of plant phenols are listed in Table 1. Also included is an example of a fruit in which the constituents of the particular class are abundant. The fruit juice industry has become one of the worldAs major agricultural businesses with world trade in fruit juices annually exceeding $10 billion.1 Further, it is estimated that fruit juice consumption accounts for 25–30% of dietary intake of flavonoids. The single most important group of phenolics is the flavonoids.They are highly diverse, in both their chemical structure and proposed biological functions.2–13 Their metabolic pathways are particularly complex with, in many cases, multiple alternative metabolic fates. Profiling of the components of a pathway over time provides a dynamic view of the metabolic events occurring in the plant. Nevertheless, establishing a biological function for such compounds is often very Kevin Robards is Associate Professor of chemistry at Charles Sturt University Riverina.He obtained his Ph.D. in analytical chemistry in 1979 from the University of New South Wales after working as a chemist in industry. His research interests are focused on the application of analytical chemistry to the food and environmental sciences.Michael Antolovich is a lecturer at Charles Sturt University Riverina. He completed his Ph.D. in 1989 at the University of New South Wales, followed by a postdoctoral position at Princeton University working on metalloporphyrins. He returned to Australia on a Research Fellowship. His current research interests are in computational chemistry and its application to food and environmental sciences. Analyst, February 1997, Vol. 122(11R–34R) 11Rdifficult and complicated by the fact that the alternative products for a given metabolite may vary from tissue to tissue, from one growth condition to another and in response to environmental stimuli.Interest in the bioflavonoids is related to their diversity, biological significance as secondary plant metabolites and ecological role,9 use as chemotaxonomic markers,14 impact on fruit quality,15 physiological effects,11,16,17 and industrial applications.18 In consequence, commercial interest in these compounds is considerable.19 Their pharmacological properties also account for recent interest in measuring dietary intakes of bioflavonoids which ranged between 23 mg d21 estimated in The Netherlands and 170 mg d21 estimated in the USA.Major dietary sources of flavonoids determined from studies and analyses conducted in The Netherlands include tea, onions, apples, and red wine.20 However, current estimates of daily consumption of flavonoids differ considerably.21 All of these aspects justify the intense interest in bioflavonoids which has been manifested over several decades and accounts for the many reviews and monographs1–9,22 devoted to various aspects of these compounds.This review critically examines the analytical chemistry of the bioflavonoids in fruits with emphasis on work of the last decade, although earlier studies are included where they are relevant. Methods used for the analysis of samples other than fruits will be discussed where these illustrate current applications which can be extended to include fruits or which may emerge as important advances over existing methods.The terms bioflavonoid and flavonoid are used interchangeably to encompass all compounds derived from the basic structures in Fig. 2. Where a distinction between the various flavonoids is intended, the name of the specific class of compound or derivative is used. Thus, flavonoid glycoside is used generically whereas flavanoid glycoside denotes a particular class of glycoside.There are many flavonoids present in low concentrations which remain unidentified but whose significance may far outweigh their concentration level. Isolation and structure elucidation of these compounds are the initial steps to understanding their significance and action. Information on their biosynthesis is essential to understanding the interaction between plants and the environment. Methods of characterization and identification follow those in general use for natural substances. Hence, preparation of an extract, biological screening, bioguided fractionation, isolation and structure elucidation constitute the usual approach.For structural elucidation, physical methods based on spectral characteristics feature prominently, although older chemical and biochemical approaches should be considered particularly as adjuncts to spectral analysis. These methods, which include NMR spectrometry, have been treated elsewhere6,23 and are not considered in the present discussion.On the other hand, MS is included because of its role as an on-line detection device in GC and, more recently, HPLC. Commercial development of coupled LC–NMR as a complementary technique to LC–MS will provide enormous benefits and stimulate rapid developments in flavonoid chemistry. Fig. 1 Biosynthetic pathway leading to the various flavonoid classes. Table 1 Major classes of phenolics in fruits Basic skeleton Class Examples Common fruit source C6 Simple phenols Catechol, Hydroquinone, resorcinol Benzoquinones C6–C1 Phenolic acids p-Hydroxybenzoic acid, salicylic acid Strawberry C6–C2 Acetophenones Phenylacetic acids p-Hydroxyphenylacetic acid C6–C3 Hydroxycinnamic acids Caffeic acid, ferulic acid Apple Phenylpropenes Eugenol, myristicin Coumarins Umbelliferone, aesculetin, scopolin Citrus Isocoumarins Chromones Eugenin C6–C4 Naphthoquinones Juglone Walnut C6–C1–C6 Xanthones Mangostin, mangiferin Mango C6–C2–C6 Stilbenes Resveratrol Grape Anthraquinones Emodin C6–C3–C6 Flavonoids Quercetin, cyanidin Cherry (C6–C3)2 Lignans Pinoresinol Neolignans (C6–C3–C6)2 Biflavonoids Agathisflavone (C6–C3)n Lignins Stone fruits (C6)6 Catechol melanins (C6–C3–C6)n Condensed tannins (flavolans) Persimmon 12R Analyst, February 1997, Vol. 122Chemical Structure The flavonoids are built upon a C6–C3–C6 flavone skeleton in which the three-carbon bridge between the phenyl groups is commonly cyclized with oxygen.Several classes (Fig. 2) are differentiated according to the degree of unsaturation and degree of oxidation of the three-carbon segment. Within the various classes, further differentiation is possible based on the number and nature of substituent groups attached to the rings. The range of known flavonoids is therefore vast, currently exceeding 5000.5 These are frequently referred to by trivial names which generally relate in some way to the plant origin.For example, quercetin was originally isolated from Quercus and tricin from Triticum. The range of trivial names is immense and can be confusing. Fortunately, excellent summaries have been compiled by Swain24 and Wollenweber and Dietz.25 Additional structural complexity is introduced by the common occurrence of flavonoids as the O-glycosides in which one or more of the flavonoid hydroxyl groups is bound to a sugar or sugars by an acid-labile hemiacetal bond.A common flavonoid such as kaempferol may be found to occur in nature in any one of 214 different glycosidic forms.9 In principle, any of the hydroxyl groups can be glycosylated but certain positions favour glycosylation; for example, the 7-hydroxyl in flavones, isoflavones and dihydroflavones, the 3- and 7-hydroxyl in flavonols and dihydroflavonols and the 3- and 5-hydroxyl in anthocyanidins. Glucose is the most commonly encountered sugar with galactose, rhamnose, xylose and arabinose not uncommon and mannose, fructose, glucuronic and galacturonic acids being rare.Disaccharides and even higher are also found in association with flavonoids, the more common being rutinose (6-O-a-l-rhamnosyl-d-glucose) and neohesperidose (2-O-a-lrhamnosyl- d-glucose). Acylation of the glycosides in which one or more of the sugar hydroxyls is derivatised with an acid such as acetic or ferulic is occasionally observed. Glycosylation has a profound effect on the flavonoid rendering it more water soluble, permitting storage of the flavonoid in the cell vacuole where they are commonly found.Glycosylation may also occur via direct linkage of the sugar to the benzene nucleus (flavonoid C-glycoside) by an acid-resistant carbon–carbon bond. This is much less common with a more restricted range of sugars and flavonoid aglycone types. Flavonoid sulfates containing one or more sulfate residues attached to a phenolic or sugar hydroxyl are even less common. Various groupings are identified within the various classes of flavonoids because of structural similarities.Flavanoids or dihydroflavonoids are so-called because the C-2 and C-3 of their skeleton are hydrogenated. Thus, the flavanoids include flavanones (or dihydroflavones), flavanonols (also called 3-hydroxyflavanones or dihydroflavonols) and the dihydrochalcones. Strictly, flavanols and flavans are also included but they are usually treated separately because they do not possess a carbonyl group in their heterocyclic ring.There are only a few naturally occurring flavans (see Fig. 2),26,27 although the term is sometimes used collectively to include the flavan-3-ols and flavan-3,4-diols. Significance It is estimated28 that about 2% of all carbon photosynthesized by plants, amounting to about 1 3 109 t per annum, is converted into flavonoids or closely related compounds. The flavonoids constitute one of the largest group of naturally occurring phenolics.Quantitative data are provided in Table 2 for various classes of phenolics in fresh fruit. Such data should be handled very cautiously for a number of reasons, the most important being that results display great variability for different cultivars of the same species and also between different authors. Nevertheless, the data do provide a guide to concentration levels and distribution between fruits. The number of polar, water-soluble flavonoids is considerable, but there are also many lipophilic compounds such as the flavones.The former are sequestered in the vacuole and are always present in conjugated form, frequently in glycosidic linkage, but may be released as the free aglycone during fungal infection or insect grazing. In such cases, they are likely to be considerably more toxic than the bound form, to the invading organism. This is seen in the case of quercetin, which is an effective inhibitor of enzyme activity of many types,9 whereas the related quercetin O-glycosides (e.g., rutin) cause negligible inhibition.Function in Plants In some cases, flavonoid function may well be related to primary metabolism. Some flavonoids may have an indirect effect on plant growth while others may protect the more vulnerable cell constituents from damaging radiation by virtue of their strong UV absorption.31 In general, however, the search for a function for these compounds has focused on the interaction that may take place between the plant and other living organisms and, in particular, on the effects of flavonoids on microorganisms which may infect plants and on animals which graze on plants.Fig. 2 Structures of the various classes of flavonoids. Analyst, February 1997, Vol. 122 13RPlant resistance The early literature contained many suggestions that the flavonoids were involved in disease resistance in plants. Further, the association of increased endogenous flavonoid synthesis with the early stages of infection was recognized but the precise nature of their involvement in disease resistance remained unknown.The development of the phytoalexin theory dramatically altered this situation. Muller32 defined phytoalexins as ‘compounds produced after infection under the influence of two metabolic systems, that of the host and that of the parasite, and inhibitory to the parasite.’ Most literature reports are concerned with the role of phytoalexins in resistance to disease caused by fungi and fewer relate phytoalexin accumulation to resistance to disease caused by bacteria.33 By the mid- 1960s it became apparent that phytoalexins were produced not only in response to infection but also in response to various forms of physiological stress.This has been an area of intense research activity over the last few decades and has been reviewed by Kuc.33 A common response of plant cells to stress such as wounding, infection or elicitation is the induced incorporation of phenylpropanoids into the cell wall.However, as stated by Matern and Grimmig,10 ‘The precise role of the phenolic cell wall reinforcement for the protection of plants has . . . remained ill-defined due to limited analytical knowledge and the complexity of the cell wall architecture.’ The discovery of new analytical methods underpins scientific progress and nowhere is this more evident than in the study of the flavonoids. Photoprotection The hypothesis of the protective role of flavonoids against harmful UV radiations is supported by the enhanced levels of flavonoids observed in plants exposed to strong UV radiation. 34,35 Moreover, photocontrol of flavonoid biosynthesis is induced or increased by UV irradiation.36,37 Monici et al.35 evaluated the role of kaempferol and pelargonidin as photoprotectors, the compounds being chosen as representative of the widely distributed flavonols and anthocyanidins. They concluded that both compounds contribute to plant protection but via distinctly different mechanisms.The action of pelargonidin seems more likely to derive from its strong radical scavenger action whereas kaempferol can be considered a good screen against UV radiation. Their conclusions are supported by in vivo experiments.34 Properties and Physiological Activity in Animals The flavonoids are potent antioxidants, free radical scavengers38 and metal chelators; they inhibit lipid peroxidation20 and exhibit various physiological activities,39–44 including antiinflammatory, 45 anti-allergic, anti-carcinogenic, antihypertensive and anti-arthritic activities.46 The biological properties12,17,18,21 of quercetin, the most frequently studied flavonoid, are certainly consistent with these activities.Thus, quercetin and other flavonoids modify eicosanoid biosynthesis (anti-prostanoid and anti-inflammatory responses), protect lowdensity lipoprotein from oxidation (prevent atherosclerotic Table 2 Food sources and properties of phenolic substances in fruits.Where concentration data are available, they are quoted as extreme values in ripe fruits3,5,6,29,30 Class of phenolic Specific example Food source Biological properties Concentration/mg kg21 Hydroxybenzoic acids Widely distributed 1–200 Hydroxycinnamic acid derivatives Widely distributed 1–500 Coumarins Citrus flavedo 2000–7000 Flavones Apigenin and luteolin Found mainly in citrus Co-pigments in flowers; UV fruits, red grapes and protectants in leaves green beans Polymethoxylated flavones Sinensitin Citrus flavedo Citrus peel oils 1–10 g l21 Flavonols Quercetin, kaemferol Found in many fruits and Co-pigments in flowers; UV and myricetin vegetables such as kale, protectants in leaves spinach, onions, parsley, French beans, endive and apples Flavanones Hesperitin, naringenin Usually found in citrus fruits such as grapefruit, oranges and lemons Flavonol glycosides Rutin Widely distributed 2–300 Flavanone glycosides Hesperidin, Citrus Some have bitter tastes 1000–5000 neohesperidin and naringin Anthocyanins Glycosides of Coloured berries Red to blue pigments 3000–5000 (i.e., anthocyanidin glycosides) pelargonidin and Other fruits 20–500 delphinidin Flavan-3-ols (catechins) +Catechin, These are the main 10–300 epigallocatechin, polyphenols in green epicatechin gallate, tea.Fruits such as epicatechin and apples, cherries and epigallocatechin pears also contain gallate limited amounts of catechins Chalcones, aurones Yellow pigments Ellagic acid* Abundant in most berries, especially cranberries, red raspberries and some nuts * Not a true flavonoid, but closely related. 14R Analyst, February 1997, Vol. 122plaque formation) as a result of their potent antioxidant action,13,47 prevent platelet aggregation (anti-thrombic effects) and promote relaxation of cardiovascular smooth muscle (antihypertensive and anti-arrhythmic effects).Free radical formation is considered to play a key role in the development of cancer and coronary heart disease by attack on biomolecules (lipids, proteins, DNA) or the biomembrane. The flavonoids may provide protection48 of membranes (including mitochondrial membranes) against the oxidative damage implicated either directly or indirectly via DNA damage in neurodegeneration, ageing, and malignant progression to cancer. In vitro testing has shown that many flavonoids, most notably quercetin, one of the most widespread in the human diet, are mutagenic. However, systematic studies also in vivo have produced no clear evidence of carcinogenicity.Moreover, the same compounds which proved mutagenic in the Ames test subsequently proved to inhibit tumour development in several experimental animal models. Quercetin in particular was demonstrated to be a potent anticarcinogen in rodents against skin, colonic and mammary cancers.It also inhibited the induction and progression of human cancers.49 These conflicting biochemical activities are very puzzling and it was suggested that phenols could test positive as a mutagen because of artefactual errors due to the test conditions. The role of flavonoids in carcinogenesis and genotoxicity has been extensively studied43,50–53 with particular interest in the flavonoids in wines and other beverages.54,55 Hence attention has generally focused on the anti-tumour activity of the flavones luteolin and apigenin and the flavonols, quercetin, kaempferol and myricetin.56–58 Solimani et al.49 used flow linear dichroism to study the interaction between DNA and a series of flavonoids whose biological activity spanned a wide range of potency, namely quercetin (the most active), morin, rutin, naringin and 2,3-dihydroquercetin (inactive).The biologically active flavonoids bound DNA by intercalation and their affinity for DNA followed the same sequence of the potency of their activity.The hypotensive effects of flavonoids have been demonstrated in animal experiments. Thus, of four flavonoid glycosides59 extracted from orange peel, two decreased the blood pressure of rats. Potentially pharmacologically active agents were also studied60 in the peel of Citrus unshiu. Flavonoid glycosides were isolated from an aqueous extract of peel by precipitation with ethanol and sequential extraction with hexane and butanol.Six flavonoids were isolated from the crude extracts by column chromatography on silica gel as (i) limocitrin-3-b-d-glucose, (ii) limocitrin-3-a-l-rhamnose, (iii) 3,6-di-c-glucosylapigenin, (iv) narirutin, (v) rutin and (vi) narcissin. Compounds (iii) and (v) exhibited hypotensive effects in rats. These findings are supported by epidemiological studies.20,61 Inhabitants of certain regions of France show62 increased longevity and decreased incidence of cardiovascular disease compared with the USA despite consumption of fats at comparable levels, which are correlated with increased risk of heart attacks.This observation has been termed the French paradox and is attributed, in part, to the routine consumption of wine63 and other flavonoid-rich foods in the Mediterranean area. Flavonoids are found ubiquitously in plants and thus are part of the human diet (Table 2). Dietary flavonoids consist64 mainly of anthocyanidins, flavonols, flavones, catechins and flavanones.They are absorbed from the gastrointestinal tracts of humans and animals and are excreted either unchanged or as flavonoid metabolites in the urine and faeces. Accurate data on population-wide intakes of flavonoids are not available but important dietary sources are vegetables, fruits and beverages, the last accounting for at least 25–30% of the total daily flavonoid intake.65 Early estimates of flavonoid ingestion were based mainly on food analyses using techniques of doubtful accuracy.For example, Kuhnau65 estimated the dietary intake of flavonoids as 1 g d21 in the USA. More recent estimates are considerably lower but significant variations are expected reflecting different diets. Using data of the Dutch National Food Consumption Survey 1987–88,66 the average total intake of the flavonoids quercetin, kaempferol, myricetin, apigenin and luteolin among 4 112 adults was determined as 23 mg d21. This intake exceeded that of the antioxidants b-carotene and vitamin E and thus flavonoids represented an important source of antioxidants in the human diet.The most important flavonoid was the flavonol quercetin (mean intake 16 mg d21) with the most important sources of flavonoids being tea (48% of total intake), onions (29%) and apples (7%). Flavonoid intake did not vary between seasons and did not correlate with total energy intake, but was weakly correlated with the intake of vitamin A, dietary fibre and vitamin C.However, data on flavonoid absorption from the gut are limited, as are data on the effects of degradation by intestinal microorganisms. The amount of flavonoid that remains biologically available67 may not be of sufficient concentration to explain the beneficial effects seen with the Mediterranean diet. Moreover, synergistic and antagonistic effects may occur. More research is needed for further elucidation68 of the mechanisms of flavonoid absorption, metabolism and biochemical action.Large-scale epidemiological evaluations of the effects of the flavonoids on chronic diseases are needed to assess the findings from experimental studies. Uses Certain flavonoids are used for their pharmacological properties whilst others have important industrial applications. For example, naringin and neohesperidin can be converted into their corresponding dihydrochalcones with a strong sweetening capacity.18 There are, however, unexplored opportunities for exploiting these compounds.The diversity of the flavonoids in agronomic lines and their role in plant resistance69 suggest that they could be exploited for increased food production. Also, the widespread distribution of flavonoids in plants could be explored for their increased use12 in medicine and disease control. The full exploitation of this potential will depend on enhancing the levels of these metabolites in plants18 by regulating the associated processes of growth and cell differentiation.For instance, dietary modulation of the susceptibility to disease appears to be a possibility. Chemotaxonomy Flavonoid data on various plants demonstrate their value in appreciating plant variability and its phylogenetic significance at species or infraspecific level. Flavonoid compounds are particularly convenient for this purpose as they are widely distributed among plants and are chemically stable. They show a structural diversity due to differences in oxygenation, methylation and glycosylation processes.Moreover, flavonoid profiles are now relatively easy to establish using highperformance techniques discussed later. Thus, flavonoids have been widely used70–75 in chemotaxonomy particularly on the genera Ribes (e.g., blackcurrant, gooseberry), Rubus (e.g., blackberry, raspberry), Vaccinium (e.g., cowberry) and Vitis (grape). Historically, the root of classical genetics is that of the flavonoid compounds, in particular, of the anthocyanins.76 The studies of Torre and Barrit,77 which examined the quantitative distribution of anthocyanin pigments in numerous species of Rubus led to the publication73 of fundamental work on ‘anthocyanin variation in the g. Rubus.’ Quantitative variations depended on three major genes which determined the quantity of pigments and the colour of the fruit.The large number of cultivars and species of Vitis complicate the situation but two groups were distinguished78 dependent on the presence Analyst, February 1997, Vol. 122 15Ror absence of acylated anthocyanins. Factorial analysis of anthocyanin data in different grape varieties has been used79 to identify the criteria to be retained to enable characterization of a cultivar. The total anthocyanin content made possible80 the characterization of species of the genus Rubus. In the case of citrus, polymethoxylated flavones have been measured81 by HPLC in peel oils of orange, mandarin, tangerine and clementine.Standardized principal component analysis was conducted on seven parameters, namely individual concentration of six flavones plus the sum of these concentrations. Factorial discriminant analysis showed the taxonomic significance of the data in differentiation of orange and mandarin groups, the latter characterized by higher total flavone concentrations, particularly tangeretin and nobiletin. However, most interest has focused on the use of flavanone glycosides as taxonomic markers for citrus sytematics,14,82 as discussed below.Apart from its fundamental significance, chemotaxonomy has been applied to the authentication of products such as juices and jams and detecting product substitutions.83 Flavonoid profiles are ideal for this purpose; they are complex and most flavonoids are not commercially available. As an illustration, pattern recognition of flavanone glycoside profiles has been used84 to detect the addition of grapefruit juice to orange juice. Fruit Quality Flavonoids may contribute to fruit quality in a number of ways, e.g., by contributing to sensory attributes such as colour and flavour and through the contribution of some specific flavanoids to bitterness of certain fruits.85–88 In fruits such as apples, the flavonoids contribute89,90 to the texture of the fruit.They are also involved in the formation of undesirable brown pigments in fresh fruit following bruising or cutting and/or during storage91 as the result of enzymatic oxidation of endogenous phenolics into quinones which then polymerize into brown products.Apart from economic considerations, which are considerable, browning reactions are responsible for some phenomena affecting colour, taste and nutritional value of the plant products. In addition, this reactivity can affect the resistance characteristics of fruits and vegetables against storage fungi. The susceptibility of apples to browning illustrates the complex interactions between polyphenol oxidase activity and phenolic content.In some instances, enzymatic activity has been identified as the main factor in browning whereas in others the phenolic content has been highlighted.92 Amiot et al.93 distinguished soluble and insoluble brown pigments, the latter correlating with the flavan-3-ol content of the apple. In fruit juices, flavonoids may contribute to sediment formation by combination of polyphenols such as catechins with proteins to form sediments which cause undesirable hazes in products including wine and fruit juices.93 Biosynthesis An understanding of the essential features of flavonoid biosynthesis is important to understanding their diversity and to the design of sound analytical procedures.Flavonoid biosynthesis involves the interaction of at least five different pathways, namely the glycolytic pathway, the pentose phosphate pathway, the shikimate pathway that synthesizes phenylalanine, the general phenylpropanoid metabolism that produces activated cinnamic acid derivatives (4-coumaroyl- CoA) and also the plant structural component lignin and finally the diverse specific flavonoid pathways.The last three should be viewed as segments of a single unit, that of aromatic metabolism. An overview is presented in Fig. 1 with detailed discussions available elsewhere.76,94,95 Enzymes responsible for the formation of different flavonoid classes and for structural modifications, such as hydroxylation, methylation, glycosylation and acylation, have been identified.95 Moreover, amino acid and nucleic acid sequences are now available for several of these enzymes.Sources Flavonoids are characteristic constituents of green plants with the possible exception of algae and hornworts. They occur in virtually all parts of the plant but the quantitative distribution varies between different organs of the plant96,97 and within different populations of the same plant species.This variability is largely controlled by genetics but other factors include maturity, climate, position on the tree, rootstock and agricultural practices. In the case of processed products, technological processes to which fruits are exposed may impact significantly on the flavonoid content. For example, a fivefold increase in phloretin glycosides was measured98 in diffusion-extracted apple juices relative to pressed juices; whereas quercetin glycosides were barely detected in pressed juices, a range from 29.9 to 51.8 mg l21 was found in the diffusion-extracted juices.The effects of processing on the flavonoid content of numerous other juices have been reported. These include the content of flavonols in red raspberry juice99 and polymethoxylated flavones in commercial versus hand-squeezed orange juice.96 Despite the complexity and diversity of the genetically controlled flavonoid distributions of plants, some general observations are possible.An important and interesting feature of these distributions is the strong tendency for taxonomically related plants to produce similar types of flavonoids. Thus, three of the numerous classes of flavonoids are widespread and quantitatively dominant: flavonols, anthocyanins and flavan- 3-ols, the last present as both monomers and in condensed forms (tannins). The other classes, notably flavones, flavanones, flavanonols, chalcones and dihydrochalcones, are important only in particular fruits.For example, the most important flavonoids in citrus are the flavanones. Table 3 summarizes the major flavonoids found in various fruits; data are for ripe fleshy fruits. Common and systematic names for the more frequently encountered flavonoid aglycones are given in Table 4. It needs to be borne in mind that compounds found in various parts of the fruit may differ; in particular, compounds associated with the leaves may well be different from those present in flowers, stems, roots or fruits.Indeed, the type of phenolic associated with the surface of plants, e.g., in leaf waxes, is usually different from that occurring within the plant. Surface flavonoids are usually highly methylated and lack sugar substitution.102 Modern techniques reveal that most plants are likely to contain 5–15 major flavonoids with anything from 20 upwards of minor flavonoids; newer, higher resolution techniques may reveal this to be an underestimate.Nevertheless, a limited number of flavonoids are usually characteristic of most fruits. Hence Fernandez de Simon et al.101,103 established the phenolic composition of juices or nectars of orange, apple, pineapple, peach, apricot, pear and grape. They concluded that hydroxycinnamic acid esters with tartaric acid are typical of grape, phloridzin is typical of apple and isorhamnetin glycosides are typical of pear. Myricetin is only found in peach and luteolin and apigenin glucosides are found only in orange.Apricot could be detected by the presence of two coumarins and pineapple by the presence of sinapic acid and the absence of phenolics (benzoic acids and aldehydes, flavan-3-ols, flavonols, chalcones, cinnamic acids and their derivatives). Anthocyanic Pigments The anthocyanins are pigments which give most fruits their red, violet and blue colour,104,105 although the red colour of some 16R Analyst, February 1997, Vol. 122fruits (e.g., orange, tomato) is caused by carotenoid pigments rather than anthocyanins. The latter are glycosides which release the anthocyanidin aglycone by hydrolysis.29 The aglycones exist in cationic form in acidic medium with numerous mesomeric forms. Interest in the anthocyanic pigments and their stability106 can be attributed to their contribution to the colour of many processed products, including jams and juices. Nonetheless, anthocyanins have not been used extensively as additives in the food industry107 because of their instability towards a variety of chemical and physical factors (e.g., pH and light), the difficulty of purification and their limited commercial availability.However, the discovery of more stable acylated anthocyanins will probably see the realization of their considerable potential108 as safe food additives. Six anthocyanidins are widespread and commonly contribute to the pigmentation of fruits. Cyanidin is the most common and, in terms of frequency of occurrence,109 is followed in decreasing order by delphinidin, peonidin, pelargonidin, petunidin and malvidin.There are few fruits which do not contain cyanidin and, in a number of fruits, e.g., peach and pear, it is the single dominant aglycone. In other fruits, two aglycones are characteristic, e.g., cyanidin and pelargonidin in raspberry cultivars. Lowbush and highbush blueberries are unusual in containing significant amounts of the five aglycones110 delphinidin, cyanidin, petunidin, peonidin and malvidin.These were present as non-acylated glucosides and galactosides with the corresponding acetylated anthocyanin also occurring in several cultivars. The anthocyanidin glycosides are characteristic of a fruit. Quantitative variations occur in the anthocyanin content of ripe fruits111,112 in response to climatic factors, light and temperature in particular. Fruits, in contrast to other plant organs, are characterized2 by a relatively large amount of monoglycosides compared to diglycosides.Glycosylation of anthocyanidins almost always occurs at the 3-position with glucose, arabinose and galactose the most common sugar moieties. Hence the most common anthocyanins in fruit consist of the 3-monoglucosides of cyanidin, delphinidin, peonidin, pelargonidin and petunidin, cyanidin 3-galactoside, cyanidin 3-arabinoside plus the single diglycoside cyanidin 3-rutinoside. Among these pigments, cyanidin 3-glucoside is the most Table 3 Major flavonoids found in selected fruits100,101 Major flavonoids Flavan-3-ols*/ Fruit Flavonols Anthocyanins flavanones Chalcones Others Apple Quercetin, quercetin Cyanidin glycosides (+)-Catechin, Phloretin derivatives, glycosides including including acylated (2)-epicatechin notably phloridzin rutin, kaempferol derivatives Citrus— Sweet orange Glycosides of Hesperidin, narirutin, Flavones: sinensetin, pelargonidin, eriocitrin, narirutin- nobiletin, tangeretin peonidin, 4A-glucoside isosinensitin delphinidin, petunidin Grapefruit Naringin, narirutin, Flavones: tangeretin, hesperidin, polymethoxylated neohesperidin flavones Lemon Rutin, limocitrol, Hesperidin, eriocitrin Flavones: diosmin, limocitrin, luteolin-7-rutinoside isolimocitrol Grape Quercetin, kaempferol, Glycosides of (+)-Catechin, Flavanonols: glycosides glycosides of quercetin cyanidin, peonidin, (2)-epicatechin, of dihydroquercetin kaempferol, delphinidin, (+)-gallocatechin, and myricetin, petunidin, malvidin (2)-epigallocatechin dihydrokaempferol isorhamnetin, including rutin Pear Quercetin, glycosides of Cyanidin glycosides (+)-Catechin, Arbutin, phloretin quercetin, isoquercetin (2)-epicatechin glucoside kaempferol and isorhamnetin Stone fruit— Peach Myricetin, quercetin, Cyanidin glycosides (+)-Catechin, kaempferol, quercetin (2)-epicatechin and kaempferol glycosides, including rutin Plum Glycosides of Glycosides of (+)-Catechin, kaempferol and cyanidin, peonidin (2)-epicatechin quercetin Apricot Quercetin, quercetin 3-O-glucoside, rutin, kaempferol glycoside Sweet cherry Glycosides of Cyanidin (+)-Catechin, kaempferol and glycosides (2)-epicatechin quercetin Tomato Kaempferol and Naringenin, Chalconaringenin quercetin glycosides naringenin glycosides * Associated mainly with skin and seeds.Analyst, February 1997, Vol. 122 17Rabundant in both fruits and other plant organs.Other anthocyanins are limited in their distribution. Anthocyanins identified in red raspberry, for example, were113,114 cyanidin- 3-sophoroside, cyanidin-3-glucorutinoside, cyanidin-3-glucoside, cyanidin- 3-rutinoside, pelargonidin-3-sophoroside and pelargonidin- 3-glucorutinoside. These pigments were relatively unstable and degraded during storage and fermentation.115 The major anthocyanin in raspberry, cyanidin-3-sophoroside, is also the most stable pigment whilst cyanidin-3-glucoside is considerably less stable.The anthocyanin patterns of raspberry cultivars were distinguished by quantitative rather than qualitative differences. Polymerized pigments in raspberry juices indicated a history of processing or storage abuse. Anthocyanins are the major phenolics in dark-coloured cherry genotypes116 with total anthocyanin content ranging from 82 to 297 mg per 100 g of pitted cherry. Total anthocyanin is considerably less at 2–41 mg per 100 g of pitted fruit in light-coloured cherries.The 3-rutinoside and 3-glucoside of cyanidin are the major anthocyanins with the same glycosides of peonidin as minor anthocyanins. Another minor anthocyanin is pelargonidin 3-rutinoside. In addition to glycosylation, acylated anthocyanins are found fairly often in fruits, the situation being particularly complex in grapes75,117–121 where the 3-monoglucosides corresponding to the five aglycones can all be acylated by acetic or p-coumaric acid.Flavonols These compounds are very widespread in higher plants where they occur usually as O-glycosides in the leaves and outer parts of the plant, while only trace amounts are found in parts of the plant below the soil surface. There are far fewer detailed quantitative studies on fruit flavonols than on anthocyanins particularly in relation to genetic and environmental variability. Nevertheless, over 200 flavonol aglycones have been identified in plants, although only four of these, quercetin, kaempferol, myricetin and isorhamnetin, are common in fruits.Glycosylation occurs preferentially at the 3-hydroxyl group122,123 and the predominant types in fruits are 3-O-monoglycosides in the following order: 3-glucosides > 3-galactosides > 3-rhamnosides > 3-glucuronides. The only diglycosides observed with any frequency in fruit are the 3-rutinosides of quercetin and kaempferol. Complete characterization of a flavonol monoglycoside requires a knowledge of whether the sugar–aglycone bond is an a or b linkage and whether the sugar is in furanose or pyranose form.In general, it has been found that sugars with a d-configuration, namely glucose, galactose, xylose and glucuronic acid, are usually linked to the aglycone by b bonds whilst a linkages occur to l-arabinose and l-rhamnose. This is illustrated by the characteristic flavonoid glycosides of apple,98,124,125 which include quercetin a-l-arabinofuranoside, b-d-galactopyranoside, b-d-glucopyranoside, a-l-rhamnopyranoside and b-d-xylopyranoside.The occurrence of flavones and flavonols has been thoroughly reviewed64,126 The flavonol content of Rosaceae fruits, e.g., strawberry, raspberry and blackberry,127,128 is dominated by quercetin and kaempferol and their glycosides. The flavonol glycoside profile does not differ greatly between the various Rosaceae fruits, but does differ from flavonol glycoside profiles of stone fruits. Apricots, plums and peaches contain129,130 kaempferol and quercetin glycosides.The main glycoside in apricots is rutin, followed by kaempferol-3-rutinoside present at a considerably lower concentration (by a factor of 10). Small amounts of 3-glucosides and 3-galactosides of kaempferol and quercetin and quercetin-3-rhamnoside are also present with traces of more highly glycosylated flavonols, although with considerable varietal variation in amount present. The main glycosides in peaches are 3-glucosides and 3-galactosides of kaempferol and quercetin.In addition, kaempferol-3-rutinoside, quercetin-3-rutinoside and quercetin-3-rhamnoside are present. Traces of triglycosides, present in very small amounts, have been incompletely identified as kaempferol- and quercetin- 3-glucosyl-7-diglucoside and 3-galactosyl-7-diglucoside. Flavan-3-ols The monomeric flavan-3,4-diols, referred to as leucoanthocyanidins, are frequently found in the woody tissues including the bark of trees but are not major compounds in fruits.On the other hand, the flavan-3-ols are important constituents of fruits in oligomeric or polymeric forms as proanthocyanidins or condensed tannins. However, the monomers are also important natural products in their own right.93 Among these, (+)-catechin, (2)-epicatechin, (+)-gallocatechin and (2)-epigallocatechin are found in fruits. These are generally found in fruits in free rather than glycosylated forms, which distinguishes them from other flavonoids.The catechins are important in so far as they are the natural substrates of polyphenol oxidases and are therefore involved in browning phenomena. They are also the monomer units for the procyanidins. Flavanones In most of the plant kingdom, flavanones occur in small amounts compared with other flavonoids, yet they are the predominant flavonoid in citrus. In terms of its flavonoid composition, citrus is exceptional, some citrus flavanoids being found nowhere else.Four aglycones are common, namely naringenin, eriodictyol, isosakuranetin and hesperetin. Further, citrus flavanones usually occur as glycosides whereas in other plants flavanones are seldom found in glycosidic form.15 Glycosylation occurs at position 7 either by rutinose or neohesperidose, disaccharides formed by a glucose and rhamnose molecule differing only in the type of linkage 1 ? 6 or 1 ? 2. This has formed the basis for classification of citrus.Table 4 Common and systematic names of selected flavonols, flavones and flavanones Common name Systematic name Flavanols— Quercetin 3,3A,4,A,5,7-Pentahydroxyflavonol Kaempferol 3,4A,5,7-Tetrahydroxyflavonol Myricetin 3,3A,4A,5,5A,7-Hexahydroxyflavonol Isorhamnetin 3A-Methylquercetin Quercetagetin 3,3A,4A,5,6,7-Hexahydroxyflavonol Flavones— Tangeretin 4A,5,6,7,8-Pentamethoxyflavone Heptamethoxyflavone 3,3A,4A,5,6,7,8-Heptamethoxyflavone Nobiletin 3A,4A,5,6,7,8-Hexamethoxyflavone Sinensetin 3A,4A,5,6,7-Pentamethoxyflavone Scutellarein 4A,5,6,7-Tetramethoxyflavone Isosinensetin 3A,4A,5,7,8-Pentamethoxyflavone Quercetogetin 3,3A,4A,5,6,7-Hexamethoxyflavone Chrysin 5,7-Dihydroxyflavone Apigenin 4A,5,7-Trihydroxyflavone Luteolin 3A,4A,5,7-Tetrahydroxyflavone Diosmetin 4A-Methylluteolin Tricetin 3A,4A,5A,5,7-Pentahydroxyflavone Flavanones— Naringenin 4A,5,7-Trihydroxyflavanone Eriodictyol 3A,4A,5,7-Tetrahydroxyflavanone Hesperetin 3A,5,7-Trihydroxy-4A-methoxyflavanone Dihydroquercetin 3,3A,4A,5,7-Pentahydroxyflavanone Dihydrofisetin 3,3A,4A,7-Tetrahydroxyflavanone Dihydrorobinetin 3,3A,4A,5A,7-Pentahydroxyflavanone 18R Analyst, February 1997, Vol. 122Thus, most commercial citrus cultivars19,131,132 contain only the non-bitter rutinosides whereas sour orange and pummelo have only bitter flavanone neohesperidosides.133 Grapefruit are considered as hybrids because they contain both flavanone rutinosides and neohesperidosides.134 Further distinction is possible on the basis of the predominant flavanone; in the case of sweet oranges, mandarins, lemons and citrons this is hesperidin, whereas naringin is the major flavanone in grapefruit and pummelo.It is now accepted that naringin is absent from sweet orange varieties. Nevertheless, the evidence is contradictory135 and a recent publication136 reported concentration data for naringin in a number of sweet orange varieties. The latter data were obtained by HPLC using coulometric array detection.In our experience, naringin is generally absent but in some sweet oranges a peak co-elutes with naringin using typical reversedphase conditions. Our preliminary results suggest that this peak is not naringin, a fact noted also by Ooghe et al.,137 but the true identity of the peak remains unknown. Nevertheless, it does account for the confusion about the presence of naringin in sweet orange. Resolution of this issue is important as naringin is used as a chemotaxonomic marker in distinguishing sweet orange from other citrus cultivars.Data on flavanones in other fruits are fragmentary. Of the phenolics identified in the fruit cuticles of tomato cultivars,138 free naringenin, naringenin 7-glucoside (prunin) and the corresponding chalcone, chalconaringenin, were abundant. These were synthesized mainly during the climacteric and were largely bound to the cutin matrix. The composition of the flavonoid fraction was controlled by the spectral quality of incident radiation, red light favouring the formation of chalconaringenin.Flavones These compounds are not common in fruits and are never predominant. Citrus is again a special case containing a number of polymethoxylated flavones as minor flavonoids. Some of these, e.g., nobiletin and sinensetin (sweet orange peel) and tangeretin (tangerine oil), have been known for some time. Concentrations are very high in the flavedo and they are readily isolated from the essential oil of citrus fruits but are also identifiable in the juice.139 Other Flavonoids Other flavonoids are either not widely distributed in fruits or are present as minor components of the total flavonoid content.This section is intended to provide an indication of flavonoid diversity rather than a comprehensive treatise on flavonoid distribution. Nevertheless, the flavanonols, particularly dihydrokaempferol and chalcones, should be noted. The latter are important in relation to stability of anthocyanins since the chalcone form results from an endergonic reaction.140 Analysis Many analytical procedures have been developed for flavonoid compounds reflecting the varied reasons for undertaking the analysis.In some instances, profiling of the flavonoid content is necessary to examine process-related variability in flavonoid composition, whereas quantification is the ultimate goal in many cases. In other instances, isolation and identification of unknown flavonoid compounds is demanded.The design of the analytical procedure will depend very much on the intent of the analysis. For example, in both profiling and quantification studies, the most successful approaches have been based on chromatography. Indeed, a sophisticated high-resolution technique is mandatory because of the number and diversity of flavonoids. For this reason and because it avoids the need for derivatization, HPLC has found widespread acceptance in this role.Where identification is required, spectrometric methods are likely to assume more importance, particularly in coupled modes with HPLC or GC. The methods which follow are for soluble flavonoids and specifically exclude lignins and condensed tannins, which raise special problems. Total Phenols As the amount of individual flavonoids in fruits is usually low, they have often been recorded unspecifically as ‘total phenolics.’ Various enzymatic methods have been reported141 for this purpose.Historically, however, total phenolics were most conveniently assessed by spectrophotometric measurement142 on a simple extract of the plant or fruit material. There are a number of difficulties associated with such measurements and their continued value is debatable. First, exhaustive extraction with alcoholic and aqueous alcoholic solvents is likely to leave behind much tannin and other phenolics bound at the cell wall in which case measurement of ‘total phenolics’ is in reality confined to the soluble fraction.Second, the diversity of phenolics means that the selection of a reagent and/or absorbing wavelength will be a compromise, although this will be less of a problem where a single class of phenolics predominates. Results are expressed in terms of molar equivalents of a commonly occurring flavonoid, e.g., hesperidin143 or an appropriately chosen mixture of flavonoids.144 Colorimetric procedures rely on the reaction of the flavonoid with one of a number of reagents143 of varying selectivity.Folin–Ciocalteu reagent,63,121,145 which has been used before and after precipitation of flavonoids in acidic methanol,146 and vanillin are the classic reagents. Swain and Goldstein147 have reviewed methods relating to direct measurement and recommended the Folin reagent for total phenols and vanillin where catechins and proanthocyanidins are the major substances. The Folin reagent is widely used but will react with compounds other than the target phenols and interfering reductants must be removed prior to assay.Newer reagents include Prussian Blue,148 4-(dimethylamino)cinnamaldehyde and a rapid browning test in oxidative medium.146 With few exceptions, the inherent problems of direct spectrophotometric measurement relegate such methods to one of historical interest only. Nevertheless, a highly selective spectrophotometric method has been developed149 for quercetin based on its oxidation reaction in neutral solution with N-bromosuccinimide in the presence of phenol to give a violet chromogen measurable at 510 nm.All phenols absorb radiation in the UV region (Table 5). For flavonoids, the spectrum typically consists of two absorption Table 5 UV absorption maxima30 Absorption maxima/nm Compound class* Band II Band I Simple phenols 266–295 Phenolic acids 235–305 Hydroxycinnamic acids 227–245, 310–332 Hydroxycoumarins ca. 210, 250–260, 280–303 312–351 Flavonoids— Flavones, biflavones 250–280 310–350 Isoflavones 245–275 310–330 Flavonols 250–280 350–385 Flavanones 275–295 310–330 Chalcones 240–260 365–390 Aurones 240–270 390–430 Anthocyanins 265–275 465–560 * Usual solvent is methanol with the exception of methanolic HCl for anthocyanins.Analyst, February 1997, Vol. 122 19Rmaxima in the ranges 240–285 nm (band II) and 300–550 nm (band I). The precise positions and relative intensities of these bands provide valuable information on the nature of the flavonoid and its oxygenation pattern.150 Flavonoids such as 6- and 8-hydroxyflavonols, chalcones and aurones are characterized by a band extending into the visible region with longwave maxima from 380 to 430 nm whereas the anthocyanins absorb in the visible region, usually at wavelengths above 500 nm.Ionization with alkali normally causes a bathochromic shift of 15–50 nm with an increase in absorbance. The limited use of direct spectrophotometric measurements, whether in the visible or UV region, can be attributed in part to the lack of specificity of such methods.In general, they lead to an overestimation of ‘flavanoid’ content.146 Specificity can be enhanced in direct spectrophotometric methods by derivative spectrometry. For instance, chrysin and quercetin were determined151 spectrophotometrically using first- and second-derivative spectra in a method that requires no preliminary separation of the flavanoids.Continued interest in UV measurements can be attributed largely to the widespread popularity of this technique as a detection method in HPLC. The practising flavonoid chromatographer may gain much useful information from an examination of the earlier literature and application of its lessons. For example, Hostettmann et al.152 demonstrated the use of HPLC with UV and post-column derivatization for the characterization of phenolics. An excellent treatise on such approaches including the use of shift reagents153 was given by Markham.23 Sample Preparation Sample preparation encompasses a number of steps in the overall analytical scheme, from selection of a sample through extraction of the flavonoid to clean-up or purification.The ultimate goal of these procedures is the preparation of a sample extract uniformly enriched in all components of interest and free from interfering matrix components. Various procedures have been used at each stage reflecting the range of sample types and physico-chemical properties of the various flavonoids.The need for analyte recovery is ultimately related to the limited specificity and sensitivity of analytical procedures. The procedure must allow quantitative recovery of the flavonoids whilst avoiding any chemical modifications in the analytes which result in artefacts and unnecessarily complicate subsequent steps. For example, heat-sensitive components such as coumarins may be degraded by use of elevated temperatures and labile glycosides may be hydrolysed in some situations.Apart from its analytical implications, the extraction and recovery of phenolics120,121 can be of critical technological interest in processes such as wine production. Flavonoid aglycones are polyphenols and as such share the properties of phenols such as solubility in alkali resulting from their slightly acidic nature. However, if left in alkali in the presence of oxygen many will degrade.Historically, recovery of flavonoids by liquid extraction of the fruit154 has been common. Flavonoids are generally stable compounds and may be extracted from the dried, ground plant material with cold or hot solvents. Suitable solvents for this purpose are aqueous mixtures with ethanol, methanol, acetone and dimethylformamide. Extractions have been performed on freeze-dried extracts of the fruit or, alternatively, by maceration of the fresh, undried fruit with the extracting solvent.124 In the latter case, the required proportion of water is lower.The above procedure is unsuitable for anthocyanins and the less polar aglycones such as flavanones, polymethoxylated flavones, isoflavones and flavonols. The latter are more soluble in solvents such as chloroform, ethoxyethane and ethyl acetate– methanol, although flavonols have been successfully extracted with aqueous alcohol.155 In contrast, the anthocyanins are traditionally recovered as the flavylium cation by extraction with cold methanol containing hydrochloric acid.156 Caution is necessary with acylated anthocyanins, which are frequently labile in solutions containing mineral acid, and this is one of the reasons why the relatively common acylated pigments have been overlooked in earlier studies.157 Replacement of hydrochloric acid with weaker acids, either formic or acetic acid, allows the recovery of these compounds.75,158,159 Care must be exercised to ensure that the acetylated derivatives are in fact natural and not an artefact of the extraction process.160 With the most labile anthocyanins, the use of non-acidified solvents is probably a sensible precaution.Alternatively, solid-phase extraction on C18 cartridges has been used.161 When the adsorbed anthocyanins are subsequently eluted with an alkaline borate solution, a class separation is achieved. It appears that those anthocyanins possessing o-dihydroxy groups (cyanidin, delphinidin, petunidin) form a charged borate complex, resulting in a more hydrophilic species.This complex is preferentially eluted from the reversed-phase cartridge while those anthocyanins not containing o-dihydroxy groups (pelargonidin, peonidin, malvidin) are enriched on the cartridge. On the other hand, elution with 0.01% HCl in methanol produces no fractionation. A more exhaustive clean-up on polyvinylpolypyrrolidone was also examined. The relative proportions of the anthocyanins was different for the two procedures.Thus, for quantitative analysis the extraction and/or clean-up procedure should be thoroughly checked.162 Jackman and Smith163 have discussed factors such as pH, temperature, oxygen, light, enzymes, nucleophilic agents, sugar derivatives and co-pigments which affect anthocyanin stability. Solvent extraction has also been used for flavonoid recovery from fruit juices. Thus, polarity differences in citrus juice components have been exploited84 in a comprehensive recovery scheme for (carotenoids), polymethoxylated flavones and flavanone glycosides based on extraction with solvents of graded polarity.After dilution with methanol, the juice was extracted with hexane to remove the carotenoids. Further extraction of the juice with dichloromethane isolated the polymethoxylated flavones, which were chromatographed by reversed-phase LC with an acetonitrile–methanol–water mobile phase and detection at 280 nm.The flavanone glycosides remaining in solution were chromatographed on a C18 column with an acetonitrile–water mobile phase and detection at 280 nm. Isolation of polymethoxylated flavones in citrus juices has been performed by extraction with organic solvent following addition of sodium hydroxide164,165 to eliminate the possible interfering lactones. A comparative study of several solvents with regard to their effectiveness in extracting polymethoxylated flavones from intact and NaOH-treated juices has been published.165 In terms of total flavones, methyl isobutyl ketone was only slightly less efficient than benzene but was more effective for specific flavones.This data demonstrate the need to consider carefully any recovery problem101,103 on an individual basis. It was also concluded that the addition of NaOH to the juice leads to degradation of the polymethoxylated flavones and artefact formation. Ideally, fruit juices require minimal sample preparation beyond filtration.166 For example, flavonol glycosides and phenolic acids were determined121 in grape juice directly after filtering.Ultracentrifugation followed by filtration has been employed for flavonoid recovery from citrus.167,168 Poor recoveries can be attributed to low solubility of certain flavonoids21 and also to sorptive losses on the filtration medium. These effects have been thoroughly investigated by Widmer and Martin.169 In other instances, some form of preliminary sample processing has been deemed desirable.132 Solid-phase extraction on mini-cartridges has been employed97,170,171 in an attempt to minimize the effects of sample 20R Analyst, February 1997, Vol. 122preparation on extract integrity. Flavonoids were recovered by elution with methanol from a Sep-Pak C18 cartridge132 following elution of sugars with aqueous methanol. Recoveries compared favourably with those achieved by simple filtration. However, with cloudy juices in particular, both filtration and solid-phase extraction may be ineffective in recovering flavonoids located in suspended juice solids, even though these may represent a large fraction of the total flavonoids present.Under these circumstances, solvent extraction may be a preferable alternative. Hesperidin, the major flavonoid of sweet orange, presents a specific problem because of its low solubility in aqueous media. Addition of dimethylformamide to the juice has been used137,172 in an effort to improve solubility, but in this case some early eluting peaks are lost in the chromatogram. This also results in sample dilution with decrease in sensitivity.Heating of the juice has been used137 to increase hesperidin solubility, although it remains unclear whether total or soluble hesperidin is important. Buffering of the sample in the pH range 4.5–5.0 prior to extraction has been recommended173 to overcome problems of the pH dependence of flavanone glycoside recovery. With relatively few exceptions (e.g., refs. 21 and 169), methods of sample preparation have not been systematically investigated. In these circumstances, it becomes very important when comparing results from different methods to consider the extraction procedure.174 Sample clean-up The need for this step varies depending on the sample type, method of extraction and subsequent procedure. In some instances, recovery and clean-up will have been combined in a single step.Nevertheless, the original extract, particularly if aqueous alcohol has been used as extracting solvent, will generally contain numerous non-flavonoid substances which can interfere in later stages of the analysis. Carotenoid and chlorophyll pigments can be removed by liquid–liquid extraction with hexane of the aqueous extract after removal of organic solvents.17,176 Potential loss of lipophilic flavonoids must be monitored. Column chromatography has been widely investigated for preliminary fractionation of sample extracts.Suitable stationary phases include silica gel, alumina, microcrystalline cellulose and DEAE-cellulose, magnesium silicate, polyamide and Sephadex. Silica is especially suited to the separation of less polar isoflavones, flavanones and (usually highly alkylated) flavone and flavonol aglycones. For instance, 5,7-dihydroxyflavanone was isolated177 by elution from silica with chloroform–ethyl actetate (1 + 1).However, traces of iron(iii) in the adsorbent may cause irreversible sorption of flavonoids. In contrast to silica, polyamide is a good general-purpose phase suited to separation of flavonoids of varying polarity although ideal for glycosides124,178,179 Polyamide 6 minicolumns were used128 for the fractionation of flavonols in red raspberry juices. Quercetin glycosides, quercetin and kaempferol were eluted with methanol whilst a second fraction eluted with 0.5% ammonia in methanol contained three flavonol glucuronides, two flavonol forms, aglycones, ellagic acid and its derivatives.Polyamide was also chosen124 in conjunction with Sephadex resins for the isolation of flavonoid glycosides of Spartan apple peel. Microcrystalline cellulose is ideal for separation of glycosides from one another or from aglycones and for the separation of the less polar aglycones. Separation of flavanoid glycosides from flavonol and/or flavone glycosides is usually difficult but can be achieved on cellulose columns using water as eluent.The more water-soluble flavanoids are less strongly retained and elute from the column first. Sephadex LH-20 is useful for final cleanup of flavonoid extracts180,181 but can also be applied to initial fractionations.180 Hydrolysis and glycoside analysis Hydrolysis of glycosides182 is used as an aid to structural elucidation and characterisation as discussed by Markham.23 Three types of hydrolytic treatment are used for this purpose, acidic, enzymatic and alkaline.Hydrolysis has also been used to minimize interferences in subsequent chromatography179 and as an aid to simplifying chromatographic data,183–186 particularly in instances where the appropriate standards are unavailable. In this role, chemical treatment has been more common because it is less selective and more exhaustive. There is considerable variation in the lability of the glycosidic bond under hydrolytic conditions.Hydrolysis methods when used for purposes other than characterization/ structural elucidation of unknown flavonoids result in a reduction in information content. Hence, a sample extract containing, say, five O-glucosides of a single aglycone plus the free aglycone will produce after acid hydrolysis a single HPLC peak. The advantages in terms of simplicity of interpretation and quantification are apparent. Hence, acid and base hydrolysis simplified187 the complex HPLC profile of red raspberry juice phenolics dramatically.Minor differences were observed in the profiles resulting from the two treatments following sample preparation on Sep-Pak C18 cartridges. The rate of acid–base hydrolysis of glycosides depends on acid–base strength, the nature of the sugar and the position of attachment to the flavonoid nucleus. For example, glucuronides resist acid hydrolysis whereas by comparison glucosides are cleaved rapidly. C-Glycosides generally remain intact although structural rearrangements can occur in presence of hot acids188 owing, for example, to a Wessely–Moser rearrangement which has the effect of interconverting 6- and 8-C-glycosides.189 The five major flavonoid aglycones quercetin, kaempferol, myricetin, luteolin and apigenin were determined190 in freeze-dried vegetables and fruits after acid hydrolysis of the parent glycosides.The aglycones were separated by reversed-phase HPLC, the identity of the eluted compounds being confirmed by UV photodiode-array detection.Completeness of hydrolysis and extraction were optimized by systematically testing different conditions such as acid concentration, reaction period and methanol concentration in the extraction solution using samples containing various types of flavonoid glycosides. Optimum hydrolysis conditions were presented for flavonol glucuronides, flavonol glucosides and flavone glycosides. Recoveries of the flavonols quercetin, kaempferol and myricetin ranged from 77 to 110% and of the flavones apigenin and luteolin from 99 to 106%.Thin-layer Chromatography Prior to the advent of chromatography, analysis of flavonoids was a difficult proposition. The advent of paper chromatography revolutionized the analysis of natural products and many paper chromatographic (PC) methods were developed for flavonoids in the 1950s and 1960s. Excellent compilations are available but are largely of historical interest, having been supplanted by TLC.TLC in its simplest form is inexpensive and is especially useful as a method for the simultaneous analysis of several samples. The usual advantages of TLC, namely speed and an open-bed technique,191 are realized in flavonoid analysis.192,193 The selection of a suitable stationary phase and solvent depends on the class(es) of flavonoid to be examined. Hydrophilic flavonoids, for example, can be readily separated by TLC on polyamide194 or microcrystalline cellulose layers.122 Silica gel layers have traditionally been used for the less hydrophilic flavonoids, including methylated flavones and isoflavones.Thus, polymethoxylated flavones were isolated195 from orange juice concentrates by extraction with benzene and separated on silica gel layers. On the other hand, flavanone glycosides of citrus juices were separated196 on polyamide Analyst, February 1997, Vol. 122 21Rlayers. The glycosides were extracted from the juices with ethyl acetate following an initial extraction with ethoxyethane to remove less polar substances.The ethyl acetate extract was subjected to clean-up on a Sephadex G-25 column prior to TLC. Applications of layer chromatography to anthocyanins have been reviewed by Strack and Wray.197 Detection can be achieved by direct examination in the case of anthocyanins. More generally, direct viewing of the plates under ultraviolet radiation101,198 provides a sensitive means of spot location or, alternatively, various spray reagents are available.Useful reagents are aqueous aluminium chloride,196 sodium borohydride,82 diazotised sulfanilic acid and vanillin.23 Colour reactions which permit the determination of the type of flavonoid have been tabulated.199 Thus, for example, flavanones and flavanonols are reduced by magnesium and HCl in methanol solution to flavylium ions, which yield intensively orange, red and violet colours. Flavones and flavonols show almost no reaction.Flavanones and flavanonols, in turn, may be distinguished by substitution of magnesium by zinc in the test since only flavanonols are reduced to blue–violet compounds. 200 More recently, the advantages of diphenyltin dichloride have been demonstrated201 for both the qualitative and quantitative analysis of flavones and flavonols on thin-layer plates by forming fluorescent complexes of different colour. Two-dimensional TLC101,127 can be used for difficult separations such as the resolution of critical pairs.Alternatively, the excellent resolutions achieved with TLC can be improved further with high-performance layers202 which are available for a range of adsorbents and reversed-phase materials. HPTLC has been applied to flavanones and dihydroflavonols on silanized silica gel.203 Gas Chromatography As originally practised, quantification was difficult with both PC and TLC. The introduction of GC overcame this difficulty and it was inevitable that the new technique be applied to flavonoid analysis.Nevertheless, flavonoids were not an ideal application area for GC, which never assumed the importance achieved in other areas. The main difficulty is the limited volatility of many flavonoids, notably the glycosides, which must therefore be derivatized prior to GC analysis. In a typical analysis, recovery of flavonoids204 from orange juice by elution from polyamide columns was followed by the formation of triimethylsilyl ethers. The results obtained by this method were in good agreement with those for the spectrophotometric determination of rutin at 358 nm and hesperidin and naringin at 284 nm after separation of components by TLC and elution of flavonoids from the layer material.The polymethoxylated flavones are exceptional in that GC provides a viable alternative for their analysis. For this purpose, packed columns205 employing relatively non-polar phases such as OV-17 on Chromosorb W are unsuitable.On the other hand, high-efficiency open-tubular columns are ideal, producing139,206,207 excellent separations of the flavones in orange peel oil. A stationary phase is now commercially available208 which offers improved retention and selectivity for these compounds. The main advantage of such columns, however, is the low stationary phase bleed, which permits operation at elevated temperatures with minimal interference in the detection process. This greatly facilitates the use of coupled GC–MS.High-performance Liquid Chromatography HPLC combines the advantages of simultaneous separation and quantification without the need for preliminary derivatization, in most cases. Progress in this technique over the last decade is evident from an examination of the excellent text by Markham23 published in 1982, which devoted 10 pages to describing uses of PC for flavonoid analysis but dismissed HPLC in less than 1 page.This situation is now unthinkable. The merits of HPLC are seen in the resolution of the 3-glucoside and 3-galactoside of cyanidin in a 20 min separation. Before the advent of HPLC, this separation required 2 d to achieve by PC.6 The data in Table 6 summarize chromatographic methods for flavonoid analysis. Normal-phase chromatography has been used for the separation of flavonoids in orange juice and skins of ripe tomato.213,214 Non-polar components were removed from the plant material by extraction, following which the aqueous phase was subjected to clean-up on a polyamide column.Flavonoids including flavone, flavonol and flavanone aglycones and their glycosides were eluted with methanol prior to acetylation. The recovered flavonoid acetates were separated isocratically on LiChrosorb Si60 using benzene–acetonitrile, benzene–ethanol or isooctane –ethanol–acetonitrile solvent systems and detected at either 312 or 270 nm. Similarly, polymethoxylated flavones were separated211 on LiChrosorb Si60 following extraction from orange and tangerine peels.Capacity factors of the flavones were correlated212 with the position of the methoxy groups on the flavone skeleton. As an alternative to HPLC on a bare adsorbent, supercritical fluid chromatography allowed223 excellent separations of polymethoxylated flavones of citrus oils. Carbon dioxide modified with methanol gave rapid elution of the compounds as sharp, well resolved peaks.For these normal-phase systems, there is the concern217 that highly polar materials may be retained irreversibly on the column, with the result that the separation characteristics could be gradually altered. Thus, reversed-phase chromatography (RPC) has invariably been the method of choice217,239–241 for the separation of the flavonoids, usually on C8 165,213 or C18 columns101,166,186,215,234,242,243 used in conjunction with aqueous mobile phases and methanol, acetonitrile or, less commonly, tetrahydrofuran as organic modifier.The benefits of RPC have been realized for all classes of flavonoids and, indeed, phenolics in general, but RPC has particularly enhanced the separation performance of anthocyanins. In an early application,244 18 anthocyanins in a mixture of 20 were separated by RPC in 2 h. For the separation of anthocyanin pigments of red raspberry108,242 on a C18 column, 15% acetic acid was used as organic modifier. Both the retention properties and the spectral properties obtained with photodiode array detection were used for characterization.More generally, small amounts of acetic acid, formic acid or phosphate buffers (e.g., 50 mm)234 incorporated in the mobile phase tend to improve separations markedly. Recently, a branched-chain, fluorocarbonaceous, silane-bonded silica gel material has been developed245 for HPLC. This phase is characterized by its ability to resolve polyphenols such as flavonoids as sharp peaks and excellent durability under extreme eluting conditions, in contrast to ordinary hydrocarbon-bonded silica gel columns.Rommel and Wrolstad128 found that polymer columns or endcapped C18 columns with high carbon loads were ineffective for the separation of flavonoids of red raspberry juices. However, a C18 column, not end-capped and with a low carbon load, gave a good separation. Stationary and mobile phase effects on retention and selectivity have been examined in several studies.209,215,217 Thus, Pietrogrande and Kahie246 compared the retention behaviours of several flavonoid compounds on reversed-phase HPLC columns including phenyl, cyano and octadecyl phases.The selectivity properties of methanol, acetonitrile and tetrahydrofuran as organic modifiers were also reported on each stationary phase. Selectivity depended on both phases but specific stationary phase selectivity effects were more pronounced with methanol; in particular, the phenyl phase showed a greater selective retention for unsaturated flavonoids while octadecyl proved more selective for glycosides. Retention data 22R Analyst, February 1997, Vol. 122Table 6 Conditions used for the chromatographic analysis of flavonoids Analyte Sample Recovery Column Mobile phase Detection Comments Ref. Flavanones, flavanonols Plant material Extraction, hydrolysis and polyamide column TLC, silica gel Dichloromethane–acetic acid–water Spray reagents 200 Cinnamic and benzoic acids, flavones and glycosides N.A.N.A. Bondapak C18, 43300 mm Isocratic : water–acetic acid (methanol) UV, 280 nm Substituent effects on elution order 209 (Umbelliferone, scopoletin), naringin, hesperidin, bergaptol Grapefruit juice Heat, centrifuge, acidify and ethyl acetate– methanol extraction Zorbax ODS, 4.63250 mm Gradient : acetonitrile– acetate buffer UV, 280 nm and fluorescence, ex 350 nm, em 450 nm 210 Polymethoxylated flavones (35) Orange and tangerine juices Filter, add NaOH then chloroform extraction Zorbax C8 4.63250 mm Microbondapak C18 Isocratic : THF–water Isocratic : acetonitrile–water UV, 313 nm 164 Polymethoxylated flavones (35) Orange concentrates Filter, add NaOH then benzene extraction Zorbax C8, 4.63250 mm Isocratic :THF– acetonitrile–water Stopped flow UV, 313 nm and fluorescence, ex 360 nm, em 415 nm Five major polymethoxylated flavones in juice 165 Polymethoxylated flavones (316) Peel extracts Soxhlet extraction with benzene LiChrosorb Si60, 43250 mm Isocratic : heptane–ethanol or propan-2-ol UV, 280 nm 211, 212 Coumaric acids, flavanones Tomato skins Reflux, extraction and hydrolysis TLC on cellulose or silica gel.GC of derivatives Various Spray reagent for TLC; GC–MS 138 Aglycones and flavonoid glycosides (328) Tomato skins, orange juice Centrifuge, elution from polyamide followed by precolumn acetylation LiChrosorb Si60, 33200 mm Isocratic : benzene– acetonitrile UV, various (270 and 300 nm) Acetylated derivatives 213, 214 Aglycones and flavonoid glycosides (334) N.A.N.A. Bondapak C18, 3.93300 mm Isocratic: methanol–acetic acid–water UV, 254 nm Substituent effects on elution order 215 Catechins, procyanidins Cider apple Homogenized, filtered Spherisorb 5 Hexyl, 53100 mm Ternary gradient UV, 280 nm 216 Aglycones and flavonoid glycosides (3141) N.A. N.A. LiChrosorb RP-18, 4.63250 mm Gradient : formic acid– water–methanol UV, 280 nm Substituent effects on elution order 217 Anthocyanins Elderberry Heat and pH adjustment Nucleosil C18, 4.63150 mm Gradient : phosphoric acid– water–tetrahydrofuran UV/VIS, 254, 340 and 510 nm 218 Anthocyanins Fruits Methanol extraction (and silylation for GC) Aquapore RP-300 Gradient : water–methanol– acetonitrile VIS, 530 nm 219 Polymethoxylated flavones and glycosides Citrus peel Extraction with water at alkaline pH Ultrasphere C8, 4.63250 mm Gradient : methanol– acetonitrile–water–acetic acid UV, 240–400 nm Structure effects on retention; data on hesperedin complex 19 Flavonol glycosides (40) and sulfates N.A.N.A. Partsil 5 CCS/C8, 53250 nm Gradient : methanol–acetic acid–water UV, 365 nm Structure effects on retention; interference effects 220 Naringin Orange and tangerine juices Dimethylformamide (improves solubility but loss in sensitivity and some early eluting peaks < 10 min are absent) Hypersil ODS, 4.63250 mm Isocratic : ammonium acetate–acetonitrile UV 172 Anthocyanidins Cranberry juice Diluted juice eluted from SPE C18 cartridge and filtered Micropak MCH-10 C18 Isocratic: water–acetic acid–methanol– acetonitrile VIS, 530 nm 105 Catechins, procyanidins Grape seeds Maceration and extraction Brownlee Labs C18, 4.63250 mm Gradient : acetonitrile–water UV, 280 nm 120 Narirutin, naringin, hesperidin, neohesperidin Citrus juices Centrifuge and filter or SPE Zorbax ODS, 4.63250 mm Isocratic: water– acetonitrile–acetic acid UV, 280 nm Data for 52 cultivars 132 Analyst, February 1997, Vol. 122 23RTable 6—continued Analyte Sample Recovery Column Mobile phase Detection Comments Ref. Flavonoid glycosides, particularly of quercetin Apple peel (Extraction with ethyl acetate, fractionation on polyamide and Sephadex) Radial Pak RP (C18) Gradient :tetrahydrofuran in trifluoroacetic acid UV, 270 nm 124 Polymethoxylated flavones Orange and mandarin peel oils Oil expressed in a hydraulic press LiChrosorb Si60 Heptane–propan-2-ol UV, 280 nm 81 Anthocyanins Bog whortleberry Filtration and droplet counter-current chromatography Supelcosil LC18, 4.63250 mm Gradient : water–methanol– acetic acid VIS, 515 ± 25 nm 221 (Carotenoids), polymethoxylated flavones, flavanone glycosides Orange juice Comprehensive recovery scheme Zorbax ODS, 4.63250 mm Gradient : acetic acid– water–methanol UV, 280 nm 84 Quercetin and phloretin glycosides Apple skins Homogenize with methanol, extract nonpolar components with hexane and clean-up on C18 Radial Pak C18, 83100 mm UV, photodiode array Preparative-scale HPLC; MS data 175 Naringin, neohesperidin Citrus juices Centrifuge and filter Supelco C18, 3.63125 mm Isocratic : water– acetonitrile–acetic acid UV, 280 nm (photodiode array) 166 Phenolic acids, aglycones, flavonoid glycosides Bark and leaf of prunus N.A.LiChrosorb C18, 4.03250 mm Various UV, 280 nm Detailed retention data 222 Polymethoxylated flavones (39) Juices Filter then SPE Hypersil ODS, 2.13200 mm Gradient : water– acetonitrile–THF UV, photodiode array Concentration in juice (fresh versus concentrate) 170 Polymethoxylated flavones (35) Orange juice Addition of dimethylformamide, heat and centrifuge then SPE on a C18 cartridge Hypersil ODS, 4.63250 mm Isocratic : acetonitrile–water UV (photodiode array), fluorescence, ex 330 nm, em 430 nm Juice, peel and pulpwash data 96 Flavonoid glycosides Citrus Centrifuge and filter Superspher, 4.63250 mm Gradient : phosphate buffer in acetonitrile UV, various wavelengths Flavonoid content of various citrus 167 Cinnamic acids, flavonol glycosides, procyanidins Pear juice Filtration or Sephadex LH- 20 Supelcosil LC-18, 4.63250 mm Gradient : phosphate buffer in methanol UV, 280, 320 nm 123 Flavonol glycosides, phenolic acids, procyanidins Grape juice Filtration or Sephadex LH- 20 Supelcosil LC-18, 4.63250 mm Gradient : phosphate buffer in methanol UV, 280, 320 nm 121 Flavonol glycosides, procyanidins Apple juice Filtration or Sephadex LH- 20 Supelcosil LC-18, 4.63250 mm Gradient : phosphate buffer in methanol UV, 280, 320 nm Treatment effects 125 Anthocyanins Cranberry and strawberry Elution from SPE C18 cartridge Supelcosil ODS, 53250 mm, PLRP-S, 4.63250 mm Isocratic : acetic acid– acetonitrile Gradient : phosphoric acid– acetonitrile UV/VIS, 260, 520 nm Polymer-based column for low pH 108, 161 Anthocyanin Red raspberry juice and wine Filtration Supelcosil LC-18, 4.63250 mm Gradient : acetic acid– acetonitrile VIS, 520 nm 115 Polymethoxylated flavones (38) Citrus oils N.A.Zorbax Si, 4.63250 mm Supercritical fluid chromatography : methanol–carbon dioxide UV, 313 nm 223, 224 Polymethoxylated flavones (327) Orange peel oil Oil diluted with ethyl acetate OV-1, 0.32 mm350 m GC FID 206, 207 Naringin, narirutin, hesperidin, neohesperidin, prunin Citrus juice Methanol elution from polyamide cartridge Cyclobond 1, 4.63250 mm Gradient : water–methanol– acetic acid UV, 280 nm (diode array) Diastereomer separation 225 Cinnamic acids Orange and grapefruit (Hydrolysis), ethyl acetate extraction and silica gel clean-up LiChrospher RP18, 43250 mm Isocratic : acetic acid– water–methanol UV, 300 nm Distribution in orange and grapefruit sections 226 24R Analyst, February 1997, Vol. 122Table 6—continued Analyte Sample Recovery Column Mobile phase Detection Comments Ref. Narirutin, hesperidin, didymin, narirutin-4A-glucoside Citrus Centrifuge and filter Supelcosil LC-18, 4.63150 mm Gradient : phosphate buffer in acetonitrile UV, 280 nm 168 Naringin, neohesperidin, neoeriocitrin Citrus Centrifuge and filter Supelco LC18, 150 mm Isocratic : water– acetonitrile–acetic acid or phosphate buffer in acetonitrile UV, 280 nm Interference effects due to sorptive losses and coelution; extensive report on filter types 169 Cinnamic derivatives, flavonols, flavan-3-ols, dihydrochalcones Apple Freeze-dried material extracted with methanol, dried and extracted with ethyl acetate Rosil C18, 4.63150 mm Gradient : water– acetonitrile–methanol containing phosphoric acid UV, 280, 320 nm 93 Cinnamic and benzoic acids, flavonols, flavan-3-ols, chalcones and glycosides Juices of orange, apple, pineapple, peach, apricot, pear, grape Diethyl ether–ethyl acetate extraction after concentration Novapak C18, 3.93300 mm or 3.93150 mm Gradient : acetic acid– methanol–water UV, 280, 340, 254, 365 nm Concentration data for a range of fruits 101 Cinnamic acids Orange juice Ethyl acetate extraction of alkaline centrifugate Adsorbosphere C18, 4.63250 mm Isocratic : various UV, 300 nm Solvent optimization 227 Flavanones Citrus juice Ethyl acetate extraction, hydrolysis and TMS derivatization RSL 200BP, 0.25 mm350 m GC MS 228 Flavonols (quercetin, kaempferol, myricetin), flavones (luteolin, apigenin) Strawberry, apple, red currant, apricot, pear, cherry, plum, peach, grape Extraction of freeze-dried material followed by hydrolysis Nova Pak C18, 3.93150 mm Isocratic : acetonitrile– phosphate buffer or methanol–phosphate buffer UV, 370 nm 186, 190 Anthocyanins Apples Homogenization and extraction Spheri 10-RP18 Gradient : formic acid– methanol–water UV, 260, 280, 325 nm 229 Anthocyanins Blackberry juice and wine Filtration Supelcosil LC-18, 4.63250 mm Gradient : acetonitrile– acetic acid VIS, 520 nm 104 (Amino acids), cinnamic acids, narirutin, naringin, hesperidin Apple, orange and grapefruit juices Centrifuge and filter M.S.Gel C18, 4.63150 mm Gradient : phosphate buffer in acetonitrile–methanol Coulometric array 136 Phenolic acids, flavonols and glycosides Red raspberry SPE and hydrolysis Spherisorb ODS1, 4.63250 mm (not end-capped, low carbon load phase) Gradient : acetonitrile– acetic acid UV, 260, 360 nm 128, 187, 230 Flavonoid aglycones (349) N.A.N.A. Permabond OV1, 0.25 mm325 m GC MS 231 Anthocyanins, anthocyanidins Raspberry and blackberry juices Filtration and SPE Polymer Labs.PLRP-5, 53250 mm Gradient : acetic acid– acetonitrile VIS, 520 nm 114 Flavonols (quercetin, kaempferol, myricetin); flavones (apigenin, luteolin) Juices of apple, grape, tomato, grapefruit, lemon and orange Acid hydrolysis and extraction Nova-Pak C18 Isocratic : acetonitrile– phosphate buffer UV, 370 nm Quantitative data for various juices 54 Flavan-3-ols, procyanidins N.A. N.A. Hypersil ODS, 43250 mm Gradient : formic acid– methanol UV, 280 nm vs. PCR, 640 nm 232 Eriocitrin, hesperidin, naringin, narirutin, neohesperidin Orange and grapefruit juice Sample heated, centrifuged and filtered Novapak RP18, 3.93150 mm Gradient : phosphate buffer in acetonitrile UV, 280 nm Extraction efficiency 137 Anthocyanins Red raspberry juice SPE Spherisorb ODSII, 4.63250 mm Gradient : acetic acid– methanol VIS, 515 nm Effect of storage conditions 113 Analyst, February 1997, Vol. 122 25RTable 6—continued Analyte Sample Recovery Column Mobile phase Detection Comments Ref.Eriocitrin, hesperidin, naringin, narirutin, neohesperidin, neoeriocitrin Lemon, lime, grapefruit, orange Dilution in dimethylformamide– ammonium oxalate plus heat/centrifuge and filter Alltech RP18, 4.63300 mm Isocratic : water– acetonitrile–THF–acetic acid UV, 280 nm Data for various citrus; buffer necessary to ensure adequate extraction 173, 233 Cinnamic acid derivatives, flavonoids, anthocyanins Wine and musts Filtration Novapack C18, 3.93150 mm Phosphate buffer in acetonitrile UV/VIS, photodiode array Wine characterization 234 Aglycones, flavonoid glycosides (325) Citrus fruit Centrifuge and SPE LiChrospher RP18, 4.03250 mm (325) Gradient : phosphoric acid in methanol UV, photodiode array Data for distribution in fruit tissues 97 Polymethoxylated flavones (36) Orange juice Extraction with benzene Novapak RP18, 3.93150 mm Gradient :THF– acetonitrile–water UV, photodiode array, 340 nm 139 Anthocyanins Blueberries Homogenize and filter SuperPac Pep-S, 43250 mm Gradient : formic acid– water–methanol VIS, 525 nm 110 Anthocyanins Cherry Homogenize and filter SuperPac Pep-S, 43250 mm Gradient : formic acid– water–methanol UV, 280 nm; VIS, 525 nm 235 Cinnamic derivatives, flavonol glycosides Pear Freeze-dried material homogenized with methanol, washed with hexane and SPE Adsorbosphere C18, 4.63150 mm Gradient : acetonitrile– methanol–water– phosphoric acid UV UV and MS 122 Narirutin, naringin, hesperidin, neohesperidin, (nootkatone) Grapefruit, pummelo, tangelo DMSO extraction of dried fruit Bondapak C18, 43250 mm Isocratic : water–methanol– acetonitrile–acetic acid UV, 280 nm 18, 236 Hesperidin, neohesperidin, narirutin Orange juice Dilution, centrifuge and filter Uncoated fused-silica capillary tubings Borate buffer in acetonitrile (capillary electrophoresis) UV/VIS 237 Catechins, flavonol glycosides Apple juice Centrifuge and ethyl acetate extraction Spherisorb ODS-2, 4.63250 mm Novapak C18 3.9 3 300 mm Gradient : aqueous phosphate buffer UV 238 Hesperidin, naringin, narirutin, rhoifolin Orange, grapefruit Extraction with methanol heat, centrifuge and filter Alltima C18, 4.63250 mm Isocratic : water– acetonitrile–propan-2-ol– formic acid or water– tetrahydrofuran UV, photodiode array Data for orange and grapefruit; extraction efficiency 21 26R Analyst, February 1997, Vol. 122for flavonoids on a LiChrospher ODS-5 phase with 10 noncongeneric eluents were analysed247 and capacity factors quantitatively related to structural information found from 20 molecular descriptors which included physico-chemical parameters.Structural features reflecting positive charge distribution within a molecule were identified as being of most importance for retention. On a more practical note, a quaternary mobile phase afforded233 greater control of selectivity than a ternary mixture. Under the usual reversed-phase conditions, the more polar compounds are generally eluted first.209 Thus, diglycosides precede monoglycosides, which precede aglycones.The elution pattern for flavonoids containing equivalent substitution patterns19 is flavanone followed by flavonol and flavone. This elution pattern holds for both aglycones and glycosides. For isomeric compounds which differ in the structure of the saccharides attached at the 7-position, the rutinoside eluted ahead of the neohesperidoside.This relationship was established for the isomeric pairs hesperidin–neohesperidin, naringin– narirutin and eriocitrin–neoeriocitrin. For anthocyanins, molecular structure–retention characteristics have been demonstrated, e,g., in the work of Vande Casteele et al.248 and Takeda et al.249 Goiffon et al.250 established rules governing the chromatographic behaviour of anthocyanins on a reversed-phase column. The overall polarity and stereochemistry of the compound are key factors.In particular, the following factors are well documented: substitution of the anthocyanidin B-ring, the position, nature and number of sugar moieties attached to the anthocyanidin and the sugar acylation. Thus, substitution of the B-ring gives the elution order delphinidin < cyanidin < petunidin < pelargonidin < peonidin < malvidin with hydroxyl groups decreasing and methoxyl groups increasing retention. Peonidin–malvidin and cyanidin–petunidin are critical pairs depending on the choice of mobile phase.Glycosylation generally decreases retention in the order 3,7-diglycosides < 3,5-diglycosides < 3-glycosides and 3-galactosides < 3-glucosides < 3-rutinosides, subject to modification by the nature of the sugar moiety. Acylation decreases mobility. The determination of phenolics in apple juices and cider by the usual technique of RPC involving an aqueous methanol gradient and detection at 280 nm resulted in interference due to co-elution of phenolic acids and procyanidins.216 A preliminary fractionation of the phenolic compounds of wine into neutral and acidic groups215 before injection into the HPLC column eliminated the problem.Alternatively, co-elution was overcome216 by operating the reversed-phase column initially at pH 7, when the phenolic acids were eluted rapidly in an ionized form, followed by a decrease in pH to 2.5 for the remainder of the run. Co-elution and potential interference problems are also possible220 with certain flavonol glycosides and sulfates. This difficulty was readily resolved by the addition of an ion-pair reagent to the reversed-phase system.Tetrabutylammonium phosphate proved to be suitable and this shifted the elution of the flavonol sulfates to longer retention times whereas the glycosides were unaffected. Flavanone glycosides can exist as a pair of diastereoisomers by virtue of the carbohydrate attached at the chiral C-2 atom.The benzoylated derivatives252 have been resolved into their diastereoisomers on a silica gel stationary phase by HPLC. Separation of the underivatized diastereomeric forms is also theoretically achievable but has not been generally realized in practice by RPC on alkylsilane phases. With a cyclodextrinbonded stationary phase the separation has been achieved in the reversed-phase mode225 with a water–methanol–acetic acid gradient increasing in methanol content.Despite the obvious successes of RPC, the separations between glycosides and aglycones are not adequate for the resolution of a complex mixture containing many compounds of each group. For this reason, it has become common practice to separate the various classes in a preliminary extraction step. For example, poylmethoxylated flavones were determined in orange juice96 following extraction with dimethylformamide and clean-up on a C18 cartridge. Flavones were chromatographed on a Gynkochrom ODS-Hypersil column with UV detection.The amount and distribution of polymethoxylated flavones in juice, peel and pulp wash varied considerably with the content of orange peel being, about 1003 higher than that of juice. As an alternative to preliminary fractionation, column flushing has been used135 particularly in the analysis of glycosides to remove unwanted groups of compounds at the conclusion of a chromatogram. For example, in a typical flavanone glycoside analysis of whole filtered citrus juice, substances removed during the column wash cycle include free aglycones and carotenoids plus some phenolic acids.Isocratic elution has been used particularly where members of an individual class of flavonoid are to be determined. For example, polymethoxylated flavones were separated isocratically in the reversed-phase mode165 using a C8 column and tetrahydrofuran–acetonitrile–water as mobile phase. Fisher253 applied an isocratic system of water–acetonitrile (4 + 1 v/v) for the resolution and quantification of flavanone glycosides in citrus juice.However, gradient elution229 is more common in recognition of the complex flavonoid profiles of many fruits. With isocratic elution, the separation between glycoside and parent aglycone is small,209 but is increased significantly by employing a solvent gradient.210 In a typical application, flavonol glycosides and phenolic acids were determined in grape juice121 by HPLC on a Supelcosil LC-18 column with a methanol gradient in a phosphate buffer.The eluate was monitored at 280 and 320 nm with a photodiode- array detector. The only sample preparation involved was preliminary filtration, although determination of procyanidins required preliminary isolation on a Sephadex LH-20 column. It is not surprising that total phenols as determined colorimetrically with Folin– Ciocalteau reagent at 765 nm showed no correlation (r2 = 0.141) with HPLC results. Detection Detection of the eluted species has been based, most commonly, on measurement of absorption of radiation186,190,235 at characteristic wavelengths (Tables 5 and 6).Hence 515–520 nm has been widely used for anthocyanidins and anthocyanins whereas various wavelengths in the UV region have been used for other flavonoids, e.g., 280 nm for flavanone glycosides and 313 nm for the polymethoxylated flavones. For comparison, detection wavelengths used for non-flavonoid components include 450 and 465 nm for carotenoids, 200 nm for amino acids, 210 nm for limonoids, 245 nm for ascorbic acid and 214 nm254 for organic acids.Post-column derivatization has received little attention232 but offers a number of advantages, including enhanced selectivity. Hostettmann et al.152 devised a post-column reactor in which UV spectra and shifts could be measured by a photodiode-array detector, although an improved system has now been described. 255 Anthocyanins are usually separated in the typical reversed-phase system involving acidic mobile phases as their flavylium cations.In this situation, they can be selectively detected at their longer wavelength absorption maxima of 500–550 nm where other interfering phenolics show no absorption. Anthocyanins can also be detected in the UV region as the colourless chalcone, an equilibrium form256 favoured at pH 3.5. This was illustrated with malvidin-3-glucoside,257 which showed a peak at 280 nm with a retention time slightly less than that of the corresponding peak of the cationic form.Developments in photodiode-array detection97,98,243 facilitate such studies. Analyst, February 1997, Vol. 122 27RThe extensive use of photodiode-array detection can be attributed to the ability to collect on-line spectra258 without using stopped-flow techniques. This is illustrated by the measurement and characterization of flavonols128 in red raspberry juices by photodiode-array spectral techniques. Polymethoxylated flavones in citrus juices were separated170 by RPC following isolation by SPE on a C18 cartridge.The eluting substances were detected with a photodiode array in the region 230–400 nm. The UV spectra of eluting peaks obtained at the apex and both inflection points of the peak were compared and exact coincidence of the three spectra after due allowance for the background absorption was taken as an indication of purity.In most instances, a photodiode array UV detector has been chosen but operation in the visible region234 provides enhanced scope for the characterization of pigments. For example, anthocyanin pigment profiles of commercial food colouring products were characterized108 by HPLC with photodiode-array detection. Both the retention properties on reversed-phase HPLC and the spectral properties by photodiode-array detection were used for characterization. Fluorescence detection is an obvious means of improving sensitivity and selectivity in flavonoid analysis.It is interesting that one of the earliest papers on HPLC210 employed this means of detection as an adjunct to conventional UV detection. Stopped-flow scans were used to obtain excitation and emission spectra of the eluted species. This early paper demonstrated the complementary nature of the two methods of detection. The hydroxycoumarins scopoletin and umbelliferone exhibit strong emission at 450–460 nm.The limited stability and light sensitivity of several phenolics were noted and should serve as a warning. A subsequent paper165 demonstrated widely varying fluorescence intensities for five common polymethoxylated flavones found in orange juice. Discrepancies between quantitative data obtained by HPLC and TLC were noted and an explanation was presented for this variation. Electrochemical detection of phenolics has been described. 154,259 Eluted species were characterized by retention data and current–potential responses.In an approach analogous to the more familiar photodiode-array detectors, 16 serial coulometric detectors maintained at different potentials were used136 for on-line resolution of co-eluting phenolic and flavonoid compounds and the generation of voltammetric data. Resolution with the detector array was based on ease of oxidation and may be related to differences in structure where availability of electrons and the capacity for charge stabilization differed.The magnitude of the reduction potential for each class of compound corresponded to specific substitution patterns in the sequence catechol > methoxycatechol > monohydroxyl > methoxyl. Twenty-seven standard compounds were resolved in a 45 min run and limits of detection were in the low-ng ml21 range. The utility of the technique in generating multivariate data for differentiation of juices and juice mixtures was shown. Coupled Methods The on-line coupling of methods is of enormous potential because the selectivity can then be tuned in an optimum way, which in turn can be translated to either a faster analysis260 or an improved determination limit.It is difficult to provide an adequate definition of a coupled technique which is valid in all situations. Hence the distinction between coupled and other techniques is based more on historical development than on any fundamental considerations. The combinations of chromatography with either mass or Fourier transform IR spectrometry currently provide the best methods of qualitative and quantitative analysis.Of these techniques, GC–MS has become routine and is generally carried out with either electron impact (EI) ionization or chemical ionization (CI) sources, since these are appropriate for the introduction of volatile compounds. Comprehensive information about the mass spectra of the flavonoids has been published.261,262 In general, the EI mass spectra of flavonoid aglycones are characterized263,264 by intense molecular ion peaks plus significant fragments from both A and B rings that have been classified according to systems developed by Mabry and Markham265 and Grayer.262 The fragmentations often provide sufficient information to determine molecular mass, elemental formula, substitution patterns in the A and B rings and the class of flavonoid.The EI and CI mass spectra of glycosides266 are dominated by the same ions as for the corresponding aglycones;266 the protonated aglycone invariably being the base peak in the CI mass spectra.A distinction with the mass spectra of the aglycones263 is the relatively weak fragments from fission of the A and B rings. Because of limited volatility, analysis of flavonoids and their glycosides by GC and thus GC–MS has not generally found favour. However, GC–MS has been applied267 to the TMS-ether derivatives of flavone and flavonol glycosides.Flavanones have also been examined in citrus juices by GC–MS using EI228 following derivatization and sample hydrolysis. With anthocyanins, derivatization is an essential step for GC–MS.219,268 In contrast, the polymethoxylated flavones possess suitable volatility for GC. Thus, Berahia et al.269 analysed 39 polymethoxylated flavones by GC–MS. In addition to the common behaviour of flavones under EI conditions, such as a retro- Diels–Alder reaction which gives a characteristic fragment from the phenyl group of the flavone skeleton, new fragmentation pathways were identified and proposed.Ions characteristic of various substitution patterns were also identified. Improvements in GC column technology have increased the range of flavonoids amenable to GC–MS as the underivatized compounds. For instance, Schmidt et al.231 analysed 49 flavones, flavonols, flavanones and chalcones without derivatization by GC–MS. Compared with direct inlet mass spectra, the GC–MS data exhibited the same typical fragmentation patterns but with slight differences in intensities.Classical mass spectrometric gas-phase ionization techniques such as EI and CI are generally less suitable260 for polar, nonvolatile compounds such as the flavonoids. Soft ionization techniques such as laser desorption,270 field desorption (FD),271 plasma desorption and particularly fast atom bombardment (FAB or secondary ion emission)272 revolutionized the MS analysis of these compounds. For instance, FD–MS has been applied to the study of flavanone glycosides273 and FAB-MS to flavanone and dihydroflavonol glycosides.274 These methods also facilitate the coupling of HPLC and MS.The major obstacles to coupled LC–MS275 were the problems of dealing with liquid solvent and of producing gas-phase ions, particularly intact molecular ion species, without the application of heat. The newer soft ionization methods overcome lack of volatility by direct formation or emission of ions from the surface of a condensed phase.Hence they eliminate the need for neutral molecule volatilization prior to ionization and generally minimize thermal degradation of the molecular species. First-generation LC–MS instruments such as the moving wire or belt interface276 overcame incompatibility between high vacuum and the introduction of solvent by removing the liquid. Early approaches to LC–MS never achieved widespread acceptance.In the second generation, e.g., continuous-flow FAB (or dynamic FAB), soft ionization techniques were coupled with liquid introduction. When introduced, continuousflow FAB-MS rapidly superseded all other ionization methods for flavonoids and, in particular, anthocyanin studies, as it provided an ideal technique for the analysis of highly polar compounds, without the need for derivatization. It has the advantage of producing a molecular ion plus various fragmentation ions which provide structural information.Nowadays, 28R Analyst, February 1997, Vol. 122interfacing and ionization have merged in third-generation instruments such as the thermospray277 and electrospray278 types. With spray ionization as distinct from field desorption, ions are generated in an excess of ambient bath gas, not in a vacuum, and are therefore ideal for coping with an LC effluent. The chromatographic eluate passes through a resistively heated stainless-steel capillary tube located in the thermospray probe.A supersonic jet of vapour is created by adjusting the temperature of the capillary to a level where the solvent is partially vaporized. The vapour jet contains an entrained ‘mist’ of small, statistically generated electrically charged droplets. The droplets continue to vaporize and shrink as they traverse the source until a point is reached where free ions are repelled from the droplet surface and leave the thermospray source through an orifice in a sampling cone.The process is greatly enhanced if the analyte is itself ionic or by the presence of a volatile electrolyte, such as 0.1 m ammonium acetate. Thermospray emerged as a practical LC–MS technique279 applicable to nonvolatile samples in aqueous effluents at conventional flow rates, whereas continuous-flow FAB has developed as a technique compatible with microbore LC. However, thermospray is not without its difficulties; in particular, the efficiency of ion production varies widely with compound type and the flow rate and temperature of the inlet tube must be optimized for each different compound class.Moreover, each class of compound requires different conditions for optimum ionisation, and this is further complicated by gradient elution. Thermospray continues to find applications despite its limitations280,281 and has recently been reviewed.282 Plasmaspray283 overcomes a number of these limitations, producing more fragmentation than thermospray and ionizing a wider range of compounds.Electrospray and its several variations284–286 are a newer development using atmospheric pressure ionization mass spectrometers. The type of analytes and, more specifically, their polarity, are the prime factors in the choice of the LC–MS interface. Nonetheless, electrospray and ionspray are undoubtedly the fastest developing approaches287 and have largely replaced thermospray. The power of electrospray as an alternative, highly sensitive soft ionization technique for the investigation of polar, non-volatile and thermolabile molecules such as anthocyanins has been demonstrated.288,289 Electrospray operates with flow rates at less than 10 ml min21 using microbore columns or a conventional column equipped with an effluent splitter. Ionspray ionization allows higher flow rates of 2–200 ml min21,268,290 when used with narrow-bore (e.g., 100 3 2 mm id) reversed-phase columns, as illustrated by the structure determination of anthocyanins.289 The chromatographic eluate enters the ion source through a capillary maintained at a high potential. The strong electric field generated by this potential causes the eluate to be expelled from the capillary as a plume of charged droplets.As solvent evaporates from the small droplets, a critical size is reached where repulsion between the charged entities in the droplet (mobile phase, electrolyte and sample ions) becomes greater than the surface tension forces holding them into the droplet and they are ejected into the surrounding gas.The ionized species enter the mass analyser through a skimmer cone. Mass spectra generated by all three techniques have a similar form, i.e., dominated by pseudo-molecular ions with little or no fragmentation. Hence, the ES mass spectrum typically shows the molecular cation M+, aglycone ion and ions associated with the solvent, although fragmentation can often be induced by raising the cone voltage.Acid (acetic or formic) is often added to mobile phases in positive ion electrospray as a source of protons to assist ionization. Sensitivity is improved when the organic content in the mobile phase exceeds 20%. Atmospheric pressure chemical ionization (APCI) is a further development of electrospray in which the nebulized sample is ionized through a corona discharge and analytes become electrically charged by chemical ionization. APCI, which is compatible with flow rates of up to 2 ml min21, was used291 to determine various isoflavones.The negative ion mode provided quality mass spectra which gave not only the molecular mass of the isoflavones, but also their molecular structures. Deuterium oxide was used to induce peak shifts in the mass spectra to determine the number of exchangeable hydrogen atoms in each molecule. On rare occasions, MS can provide data sufficient for full flavonoid structure analysis, but more generally it is used to determine molecular mass and to establish the distribution of substituents between the A- and B-rings.This situation will improve with new developments. Tandem mass spectrometry (MS– MS)292,293 has been applied successfully to problems involving trace analysis of citrus flavanones and metabolite identification.294 Positive CI-MS–MS was superior to EI-MS– MS for the detection of a common fragment ion for flavanones at m/z 153. Using this approach, the flavanones, naringenin and hesperitin were detected in human urine after citrus ingestion.Glycosides were labile under the experimental conditions, probably during ionization. MS–MS, particularly in combination with LC and soft ionization techniques,281 can be expected to improve significantly the separations of complex samples. The information available from such methods can be expected to increase with the development of newer technologies, including collisionally induced dissociation spectra.281 For the latter, alternating low and high orifice voltages are used in which no fragmentation occurs at low voltage and fragmentation is induced at the high orifice voltage.This permits the simultaneous measurement of molecular mass and structural characterization. A multi-dimensional chromatographic system combining on-line HPLC with high-resolution GC–MS has been described295 for the analysis of complex mixtures. The system was not applied to flavonoid analysis but gave excellent results for the characterization of neroli oil.Other developments still in their infancy, at least in applications to flavonoid analysis, are the coupling of GC and FTIR spectrometry and capillary electrophoresis (CE) with MS. Trimethylsilyl derivatives of several flavonoids were studied296 by GC–FTIR. The correlation between retention and gas-phase IR data was successfully used in structural identification of compounds having very similar chromatographic behaviour.The shift of the carbonyl frequency gave information on the presence of substituting agents. In combination with the inherent sensitivity and selectivity of mass spectrometry, coupled CE–MS becomes a very powerful technique. The correspondence between CE and electrospray ionization flow rates provides the basis for an extremely attractive technique. Hence, various isoflavones were separated297 on an uncoated fused-silica column using 25 mm ammonium acetate buffer and negative electrospray ionization MS detection.This approach permitted the determination of molecular mass of the isoflavones and also the presence of various functional groups according to observed losses from the [M 2 H]2 ion during collision-induced dissociation effected by adjusting the MS parameters. Other Techniques Several other techniques have been examined for application to flavonoid analysis. For example, a quantitative method using radioimmunoassay131,298,299 gave the naringin concentrations in orange flavedo and albedo.Chemiluminescence has also been studied. Hence, a number of flavonoids, when excited by hydroxyl radical, emit light with an intensity consistent with that of the radical-scavenging activities of these compounds. Thus, chemiluminescence was reported300 to decrease in the Analyst, February 1997, Vol. 122 29Rorder rutin > myricetin = isoquercitrin > quercetin > kaempferol > isorhamnetin for the major flavonols and nasunin > rubrobrassicin > delphinidin > cyanidin = malvin > malvidin for anthocyanins.Based on these data, chemiluminescence warrants closer examination, particularly as an on-line detection technique for HPLC where the inherent sensitivity and selectivity will be optimally exploited. Capillary (zone) electrophoresis (CE or CZE) is a highperformance technique301 in which separation is achieved on short, uncoated, fused-silica capillary tubes (100 cm 50–100 mm id).In a typical application, orange juice was analysed237,302,302 by CE using a 35 mm sodium borate buffer containing 5% acetonitrile at 21 kV with high-speed scanning detection for UV and visible spectra. The injection volume is critical as a 1 m capillary of 75 mm id contains about 5 ml of buffer, hence, the sample volume must be less than 50 nl to avoid overload. In this instance, a 10 s hydrodynamic injection produced optimum results. The only sample preparation required was dilution and filtration.The method permitted the simultaneous analysis of a broad range of charged water-soluble molecules whilst non-polar compounds such as carotenoids moved with the electroosmotic flow without being separated. As the flavonoids are charged molecules in alkaline media, they were separated in this study. Capillary electrophoresis is not applicable to the separation of uncharged solutes. Despite a superficial resemblance to HPLC, separation in CE depends on differences in electrical properties of the analytes rather than differences in solute distribution between two phases.Thus, most forms of CE are not considered as chromatography. The exception is a technique developed by Terabe and co-workers304,305 called micellar electrokinetic capillary chromatography (MECC), which is a hybrid of electrophoresis and chromatography and is a true chromatographic process. This technique involved introduction to the buffer of a surfactant [e.g., sodium dodecyl sulfate (SDS)]306 at a concentration exceeding that of the critical micelle concentration.At this point the surfactant ions begin to aggregate and form spherical particles whose hydrocarbon tails are in the interior of the sphere with the charged ends exposed to water. The micelles constitute a stable second phase that is capable of solubilizing non-polar solutes into the hydrocarbon interior of the micelles. With SDS as surfactant, the surface of the micelles has a large negative charge, giving them a large electrophoretic mobility towards the positive electrode.However, most buffers are characterized by a high electroosmotic flow rate toward the negative electrode such that the micelles are carried towards that electrode also, but at a much reduced rate. Hence, the system consists of a faster moving aqueous phase with a slower moving micellar phase and solutes will distribute themselves between the aqueous phase and the interior hydrocarbon phase of the micelles.Solute polarity determines the position of the resulting equilibria and hence the migration rate. The migration behaviour of flavonoids in MECC has received little attention.307 Factors affecting resolution and selectivity have been identified103,307 as applied voltage, capillary temperature, electrolyte concentration and nature (complexing or non-complexing buffers), buffer pH, micelle concentration and nature (SDS or cetyltrimethylammonium bromide)308 and the addition of organic modifiers to the running buffer (organic solvents, cyclodextrins, urea, cholate). Organic solvents modified the interaction between micelles and solutes thus altering retention and resolution.307 When methanol was used,309 either as a sample solvent or as a constituent of the buffer, the most hydrophobic flavones appeared as double or triple peaks in the electropherograms.These double peaks disappeared when acetonitrile was used instead of methanol.Micelles in MECC provide both ionic and hydrophobic interactions, the extent of each depending on the buffer pH. The separation of selected flavonoids was improved by SDS at pH 8.3 but there was little or no effect at higher pH. At pH 10.5, the separation was mainly regulated by ionization of the hydroxyl groups and borate complexation of the carbohydrate residues. Similar effects have been reported310 for the separation of flavonoid glycosides.A correlation was generally observed309,311 between the migration order in MECC of flavone aglycones and the elution order previously reported for reversed-phase HPLC. Applications of both CE and MECC to the analysis of secondary plant metabolites have been reviewed.312,313 The advantages and criticisms of CE were examined in relation to HPLC, it being concluded that CE will become an indispensable tool, together with HPLC and GC, in phytochemical laboratories since these techniques are in many ways complementary,311 and problems that are difficult to solve by HPLC, can often be solved using CE.

ISSN:0003-2654    

DOI:10.1039/a606499j

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Chemiluminescence Determination of Tiopronin by Flow Injection Analysis Based on Cerium(IV) Oxidation Sensitized by Quinine Yining Zhaoa, Willy R. G. Baeyens*a, Xinrong Zhanga, Anthony C. Calokerinosb, Kenichiro Nakashimac and Guido Van Der Wekena a Laboratory of Drug Quality Control, Faculty of Pharmaceutical Sciences, University of Ghent, Harelbekestraat 72, B-9000 Ghent, Belgium b Laboratory of Analytical Chemistry, Chemistry Department, University of Athens, Panepistimiopolis 157 71, Athens, Greece c School of Pharmaceutical Sciences, Nagasaki University, 1-14 Bunkyo-machi, Nagasaki 852, Japan A flow injection analysis method is proposed for the determination of tiopronin based upon the oxidation by cerium(iv) in dilute sulfuric acid medium and sensitized by quinine. With the peak height as a quantitative parameter applying optimum working conditions, tiopronin is determined over the 1–400 mm range (150 ml per injection, n = 10, r = 0.9994) with a detection limit of 0.34 mm and an RSD (n = 10) less than 2% at 20 and 50 mm.The proposed method, combining the advantages of speed and sensitivity, was applied to the routine determination of tiopronin in a pharmaceutical preparation. Keywords: Tiopronin; chemiluminescence; flow injection analysis; pharmaceutical analysis; thiols Tiopronin (N-2-mercaptopropionylglycine) is an effective and safe drug in the treatment of cystinuria, rheumatoid arthritis as well as liver disorders and, in addition to penicillamine, has been utilized for several years.1–3 Tiopronin and penicillamine both react with the symmetric cysteine disulfide by a thiolate disulfide interaction, which gives the more soluble mixed disulfide that can be quantified in biological fluids.2 Although tiopronin determinations have been widely reported in pharmaceuticals by spectrophotometric,4–6 GC–MS,7 catalytic titrimetric8 and voltammetric9 procedures and by preor post-column derivatization in HPLC with fluorimetric detection,10–14 only a few methods were developed employing chemiluminescence (CL) detection coupled with flow injection analysis (FIA).They are mainly based on indirect CL detection techniques, e.g., the classical luminol oxidation system,15,16 on the inhibitory effect upon the oxidation of thiamine to thiochrome by HgII 17 as well as on the complex formation with NiII,18 or PdII.19 These FIA methods to a certain extent suffer from narrow linear ranges and unsatisfactory detection limits, e.g., for the luminol system the linear range is around 0.1 to 100 mm15,16 and for the PdII system 0.01 to 0.6 mm.19 Normally the detection limit for the determination of tiopronin in biological samples as in plasma should reach down to the mm level, which has already been achieved by HPLC post-column derivatization with fluorimetric detection mode as reported previously.12 The main purpose of the present investigation is to offer a direct CL reaction system sensitized by a suitable fluorophore, which offers some characteristic advantages of relatively good sensitivity and selectivity as compared to the existing indirect CL determinations of thiolic compounds. Furthermore this paper can be regarded as a basis for the development of an HPLC–CL determination of tiopronin and its metabolite in biological fluids. Experimental Reagents All solutions and chemicals were prepared from analytical reagent grade materials using distilled, deionized water.Tiopronin was purchased from Cassenne (Paris, France). A stock solution of 5 mm standard sample was prepared by dissolving 81.6 mg of N-2-mercaptopropionylglycine (Sigma, St. Louis, MA, USA) in water and diluting to 100 ml with water. It was kept in a black flask at 4 °C. The working solutions of the 0.75 mm cerium(iv)-ion and of 0.1 quinine were prepared daily by dissolving 151.6 mg of cerium(iv) sulfate (UCB, Brussels, Belgium) and 32.3 mg of quinine sulfate (BUFA, Uitgeest, Holland) in 500 ml of 0.1 m sulfuric acid, respectively.Instruments The FIA system was installed in a conventional mode, the reaction reagents were pumped to the three-line manifold by a peristaltic pump (Minipuls 2, Gilson, Villiers-Le-Bel, France) at a total 8 ml min21 flow rate through PTFE tubes (Tygon, 2 mm id). The sample solution was injected (150 ml per injection) into a carrier stream (water) pumped by an SP8770 isocratic pump (Spectra-Physics, San Jose, CA, USA) at a flow rate of 1 ml min21 and mixed with the reagent streams in a T-piece positioned 2 cm before entering the flow cell, in which a PTFE tubing coil (Tygon, 1 mm id, 10 cm long) serving to retain the solution was placed directly in front of a photomultiplier (PMT), to generate CL emission, measured by a luminometer (Bio-Orbit 1250, Turku, Finland) linked with a computer (PC Systems, Lokeren, Belgium).An aluminium foil-covered cell is ready to collect maximum reflection of the emitted light.Extreme precautions were taken to ensure that the cell compartment and the PMT tubing were strictly light-tight. Results and Discussion Preliminary Work Preliminary work focusing on the quantitative detection of thiol-containing drugs, such as captopril and penicillamine, in a Analyst, February 1997, Vol. 122 (103–106) 103CeIV CL reaction set-up applying a fluorescent sensitizer in acidic medium was carried out in an earlier stage as previously reported.20,21 Initially, studies on some related thiol-containing drugs, such as cysteine, acetylcysteine, homocysteine and tiopronin were tentatively conducted applying an acidic potassium permanganate CL reaction system; however, no significant emission signals were produced.From four thiolcontaining drugs detected in a CeIV–quinine–sulfuric acid system under their respective optimized conditions, only tiopronin generated the most significant emission signals.Various fluorescent reagents, such as eosine, riboflavine, lucigen, rhodamine 6G and rhodamine B were employed as sensitizers and were compared with quinine for seeking a more effective alternative. The results indicated that quinine sulfate is still the best sensitizer in the present system. Method Development A series of experiments were conducted in order to establish optimum analytical conditions for the measurement of the induced CL-signals. The parameters included concentration of CeIV, quinine and of sulfuric acid, surfactants, sample volume and flow rate.Effects of concentration of CeIV, quinine and sulfuric acid The effects of the concentration of CeIV, quinine and of H2SO4 upon the CL intensity were studied and have been summarized in Figs. 1–3. When varying the concentration of each of the latter, the other parameters were held constant as illustrated in the legends for each figure. The final optimum working concentrations were chosen as CeIV 0.75 mm, quinine 0.1 mm and H2SO4 0.1 m.Effect of surfactants Surfactants, amongst other organized systems, had been investigated previously as possible enhancers of CL intensity by Zhang et al.21 in the determination of penicillamine; no significant results as to emission enhancement were observed, neither was the case in the present instance. Fig. 4 illustrates the effects of surfactants and of beta-cyclodextrin. As can be seen, no interesting enhancing effect could be noticed. On the other hand, adverse effects of these compounds upon HPLC separations may occur as noticed when using surfactants during tentative experiments.Effect of sample volume After elaborating the CL reaction conditions, the injection volume and flow-rate parameters were investigated. The variation of CL emission with the injected sample volume in the 50–150 ml range was studied similarly. Fig. 5 indicates that higher CL intensity occurs with increasing loop volumes, as expected.In the present system, a 150 ml loop was selected for the subsequent investigations. When higher CL intensity is needed, which may occur in some cases, increased injection volumes may be applied; obviously, peak broadening will then have to be dealt with. Effect of flow rate The flow rate is an important parameter in the CL reaction because the time taken to transfer the excited product into the flow cell is critical for maximum collection of the emitted Fig. 1 Effect of CeIV concentration. Quinine 0.1 mm, sulfuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 2 Effect of quinine concentration. CeIV 0.75 mm, sulfuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 3 Effect of sulfuric acid concentration. CeIV 0.75 mm, quinine 0.1 mm, flow rate 4 ml min21, tiopronin 50 mm. Fig. 4 Effect of surfactants and of beta-cyclodextrin. CeIV 0.75 mm, quinine 0.1 mm, sulfuric acid 0.1 m, flow rate 4 ml min21, tiopronin 50 mm in 1 mm of the various reagents; Triton X-100 10% (v/v). 104 Analyst, February 1997, Vol. 122light,22 too low or too high flow rates resulting in the absence of CL in the flow cell. The flow rate of the reagent solutions was initially optimized after fixing the carrier stream (water) flow rate at 2.3 ml min21 and simultaneously increasing the individual flow rates from 2 to 8 ml min21. The highest emission was obtained at 4 ml min21 for each reagent, which suggested that more light is emitted per unit of time under relatively high flow rates; this optimum flow rate apparently produces better dispersion and mixing of the reagent, higher rates may lead to both high pressures in the connector and to excessive reagent consumption.Increasing the flow rate of the carrier stream induced significant enhancement of the CL emission intensity. However, once over 2.3 ml min21 the instrumental noise increases; moreover such flow rates can not be adopted for HPLC purposes owing to an unacceptably high column pressure.As a result, 1 ml min21 of carrier stream flow rate was chosen for the calibration and pharmaceutical application. The effects of reagent flow rates in, respectively, 2.3 and 1 ml min21 carrier flow rates are shown in Fig. 6. Determination of tiopronin With the described manifold and under the optimum experimental conditions (0.75 mm CeIV and 0.1 mm quinine all in 0.1 m sulfuric acid medium; flow rate of 4 ml min21 for the reaction reagents and 1 ml min21 for the carrier stream), a linear concentration of tiopronin versus CL intensity calibration graph was obtained.A total of 10 standards were involved in the calibration process and three replicate injections of tiopronin were made per sample, the regression equation being h = 0.30 [tiopronin] + 1.068, where h is the peak height in mV and the concentration of tiopronin being expressed in mm; the correlation coefficient was 0.9994; the detection limit 0.34 mm (150 ml per injection); the RSD for 10 replicate injections of 20 and 50 mm solutions were all less than 2%.Interferences In order to assess the possible analytical applications of the described CL method, the effect of some common excipients used in pharmaceutical preparations was studied by analysing synthetic sample solutions containing 50 mm of tiopronin together with various excess amounts of excipients. The recovery results are shown in Table 1.Sorbitol shows a modest effect of CL enhancement at a concentration of 1 mm sorbitol (1 : 20 ratio), practically no emission signal is produced in the blank test. However, once the concentration of sorbitol exceeds 1 mm up to ten times, no significant increase in intensity is noticed, hence no effective emission enhancing effects can be attributed to this compound. Polyvidone and sodium carboxymethylcellulose also provide slight CL-enhancement effects, with no practical analytical applications.Adaptation of the present detection method to the analysis of the pharmaceutical preparations may be influenced Fig. 5 Effect of loop size. CeIV 0.75 mm, quinine 0.1 mm, sulphuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 6 Effect of flow rate. CeIV 0.75 mm, quinine 0.1 mm, sulfuric acid 0.1 m, tiopronin 50 mm, loop size 100 ml. Carrier stream flow rate: A, 1 ml min21; B 2.3 ml min21. Table 2 Determination of tiopronin in a pharmaceutical formulation applying the proposed CL method Sample Amount/mg Added/mg Recovered Recovery (%) Acadione* Label Found ± s (n = 10) 250 247.2 ± 2.04 250 485.8 97.7 500 730.5 97.8 750 1021.1 102.4 * Acadione is the trade mark of tiopronin manufactured by Cassenne Laboratory (Paris, France).Table 1 Recovery of tiopronin (50 mm) from various additives used as excipients Concentration ratio (additive to Recovery (%) Addition tiopronin, m/m) (n = 3) Arabic gum 1000 97.1 Lactose 1000 95.5 Galactose 1000 96.1 Saccharose 1000 98.7 Starch 1000 92.5 Carbowax 1000 100.5 Cellulose acetylphthalate Saturation 99.4 Ethyl cellulose 5 103.2 Polyvidone 5 106.1 Carboxymethylcellulose sodium 5 108.1 Dibutylphthalate 5 107 Magnesium stearate Saturation 99.2 CaHPO4 Saturation 92.4 Sorbitol 20 107 Mg(NO3)2 1000 100.9 K2SO4 1000 101.6 NaCl 10 92.5 Analyst, February 1997, Vol. 122 105by these factors, hence transfer of the developed FIA method to an HPLC system seems appropriate.Application to Pharmaceutical Preparation The proposed method was applied to the analysis of the commercial tiopronin formulation. In order to evaluate the validity of the proposed method for the determination of tiopronin in pharmaceuticals, recovery studies were carried out on samples to which known amounts of tiopronin standards were added. The results compared with the labelled contents (Table 2) and demonstrate that the method may be considered for routine analysis of the pharmaceutical preparation.However, partial interference from non-active or other active compounds should always be considered. Therefore, the HPLC coupled CL system should be envisaged to cope with the abovementioned drawbacks, and also to aim at pharmacokinetic studies of tiopronin in biological fluids. Possible CL Mechanism According to the investigation of CL properties of the fluorophore-sensitized CeIV reaction system by Zhang et al.,20 it is assumed that the possible CL mechanism is to be explained.CeIV + Tiopronin (thiol)Red ? CeIII* + TioproninOx CeIII* ? CeIII + light and/or CeIII–Tiopronin complex* ? CeIII + Tiopronin + light Where Red is reduced form; Ox is oxidixed form; and * denotes excited state. In the presence of a fluorophore (quinine), the energy resulting from the redox reaction can be effectively transferred to quinine which in turn generates CL emission: CeIII* CeIII and/or + Quinine ? and/or + Quinine* CeIII–Tiopronin complex* CeIII +Tiopronin Quinine* ? Quinine + light It is clear that in the applied CL system, based upon chemiexcitation and the use of a key sensitizer, the fluorophore plays an important role in the energy-transfer process.Conclusion The method for the determination of the thiol-containing drug tiopronin by means of CeIV oxidation sensitized by the quinine fluorophore has been successfully established as a sensitive and selective direct CL determination technique when compared with indirect CL techniques. Additionally, the method does not require sophisticated instruments.Routine drug quality control may be achieved although possible drawbacks, such as interference from formulation excipients, should be considered for each specific formulation. Further HPLC techniques coupled to the CL–FIA system are currently under investigation for the determination of tiopronin and its metabolite in human urine in pharmacokinetic studies. The authors express their gratitude to Cassenne Laboratory (Paris, France) for kindly providing the commercial formulation Acadione (tiopronin) for this work.References 1 Remien, A., and Kallistratos, G., J. Eur. Urol., 1975, 1, 227. 2 Hautman, R. E., and Robertson, W. G., Clinical and Basic Research, Plenum Press, New York, 1989, pp. 139–143. 3 Denneberg, T., Jeppson, J. O., and Stenberg, P., Proc. EDTA, 1983, 20, 427. 4 Raggi, M. A., Cesaroni, M. R., and Di-Pietra, A. M., Farmaco, Ed.Part., 1983, 38, 312. 5 Raggi, M. A., Cavrini, V., and Di-Pietra, A. M., J. Pharm. Sci., 1982, 71, 1384. 6 Raggi, M. A., Nobile, L., Cavrini, V., and Di-Pietra, A. M., Boll. Chim. Farm., 1986, 125, 295. 7 Matsuura, K., and Takashina, H., J. Chromatogr. B, Biomed. Appl., 1993, 127, 229. 8 Vinas, P., Cordoba, M. H., and Sanchez-Pedreno, C., Analyst, 1990, 115, 757. 9 Cassassas, E., Arino, C., Esteban, M., and Redondo, A., Anal. Lett., 1991, 24, 1183. 10 Kagedsal, B., Andersson, T., Carlsson, M., Denneberg, T., and Hoppe, A., J.Chromatogr. B, Biomed. Appl., 1987, 61, 261. 11 Leroy, P., Nicolas, A., Gavailoff, C., Matt, M., Netter, P., Bannwarth, B., Hercelin, B., and Massa, M., J. Chromatogr. B, Biomed. Appl., 1991, 102, 258. 12 Marzo, A., Martelli, A. E., Bruno, G., Nava, D., Mignot, A., Vidal, R., and Lefebvre, M. A., J. Chromatogr., 1991, 536, 327. 13 Kagedal, B., Carlsson, M., and Denneberg, T., J. Chromatogr. B, Biomed. Appl., 1986, 53, 301. 14 Cavrini, V., Gatti, R., Di-Pietra, A. M., and Raggi, M. A., Chromatographia, 1987, 23, 680. 15 Vinas, P., and Garcia, L., J. Pharm. Biomed. Anal., 1993, 11, 15. 16 Vinas, P., and Garcia, L., Fresenius’ J. Anal. Chem., 1993, 345, 723. 17 Perez-Ruiz, T., Martinez-Lozano, C., Tomas, V., and Lambertos, G., J. Microchem., 1991, 44, 72. 18 Pagan, A., Anal. Ciencias, University Murcia, XLVII, 1988, 29–32. 19 Garcia, M. S., Sanchez-Pedreno, C., Alberto, M. I., and Rodenas, V., J.Pharm. Biomed. Anal., 1993, 11, 633. 20 Zhang, X. R., Baeyens, W., Calokerinos, A. C., Imai, K., and Van Der Weken, G., Anal. Chim. Acta, 1995, 303, 121. 21 Zhang, Z. D., Baeyens, W., Zhang, X. R., Calokerinos, A. C., and Van Der Weken, G., Biomed. Chromatogr., 1995, 9, 287. 22 Zhang, Z. D., Baeyens, W., Zhang, X. R., Van Der Weken, G., J. Pharm. Biomed. Anal., 1996, 14, 939. Paper 6/05703I Received August 14, 1996 Accepted October 30, 1996 106 Analyst, February 1997, Vol. 122 Chemiluminescence Determination of Tiopronin by Flow Injection Analysis Based on Cerium(IV) Oxidation Sensitized by Quinine Yining Zhaoa, Willy R.G. Baeyens*a, Xinrong Zhanga, Anthony C. Calokerinosb, Kenichiro Nakashimac and Guido Van Der Wekena a Laboratory of Drug Quality Control, Faculty of Pharmaceutical Sciences, University of Ghent, Harelbekestraat 72, B-9000 Ghent, Belgium b Laboratory of Analytical Chemistry, Chemistry Department, University of Athens, Panepistimiopolis 157 71, Athens, Greece c School of Pharmaceutical Sciences, Nagasaki University, 1-14 Bunkyo-machi, Nagasaki 852, Japan A flow injection analysis method is proposed for the determination of tiopronin based upon the oxidation by cerium(iv) in dilute sulfuric acid medium and sensitized by quinine.With the peak height as a quantitative parameter applying optimum working conditions, tiopronin is determined over the 1–400 mm range (150 ml per injection, n = 10, r = 0.9994) with a detection limit of 0.34 mm and an RSD (n = 10) less than 2% at 20 and 50 mm.The proposed method, combining the advantages of speed and sensitivity, was applied to the routine determination of tiopronin in a pharmaceutical preparation. Keywords: Tiopronin; chemiluminescence; flow injection analysis; pharmaceutical analysis; thiols Tiopronin (N-2-mercaptopropionylglycine) is an effective and safe drug in the treatment of cystinuria, rheumatoid arthritis as well as liver disorders and, in addition to penicillamine, has been utilized for several years.1–3 Tiopronin and penicillamine both react with the symmetric cysteine disulfide by a thiolate disulfide interaction, which gives the more soluble mixed disulfide that can be quantified in biological fluids.2 Although tiopronin determinations have been widely reported in pharmaceuticals by spectrophotometric,4–6 GC–MS,7 catalytic titrimetric8 and voltammetric9 procedures and by preor post-column derivatization in HPLC with fluorimetric detection,10–14 only a few methods were developed employing chemiluminescence (CL) detection coupled with flow injection analysis (FIA).They are mainly based on indirect CL detection techniques, e.g., the classical luminol oxidation system,15,16 on the inhibitory effect upon the oxidation of thiamine to thiochrome by HgII 17 as well as on the complex formation with NiII,18 or PdII.19 These FIA methods to a certain extent suffer from narrow linear ranges and unsatisfactory detection limits, e.g., for the luminol system the linear range is around 0.1 to 100 mm15,16 and for the PdII system 0.01 to 0.6 mm.19 Normally the detection limit for the determination of tiopronin in biological samples as in plasma should reach down to the mm level, which has already been achieved by HPLC post-column derivatization with fluorimetric detection mode as reported previously.12 The main purpose of the present investigation is to offer a direct CL reaction system sensitized by a suitable fluorophore, which offers some characteristic advantages of relatively good sensitivity and selectivity as compared to the existing indirect CL determinations of thiolic compounds.Furthermore this paper can be regarded as a basis for the development of an HPLC–CL determination of tiopronin and its metabolite in biological fluids. Experimental Reagents All solutions and chemicals were prepared from analytical reagent grade materials using distilled, deionized water.Tiopronin was purchased from Cassenne (Paris, France). A stock solution of 5 mm standard sample was prepared by dissolving 81.6 mg of N-2-mercaptopropionylglycine (Sigma, St. Louis, MA, USA) in water and diluting to 100 ml with water. It was kept in a black flask at 4 °C. The working solutions of the 0.75 mm cerium(iv)-ion and of 0.1 quinine were prepared daily by dissolving 151.6 mg of cerium(iv) sulfate (UCB, Brussels, Belgium) and 32.3 mg of quinine sulfate (BUFA, Uitgeest, Holland) in 500 ml of 0.1 m sulfuric acid, respectively.Instruments The FIA system was installed in a conventional mode, the reaction reagents were pumped to the three-line manifold by a peristaltic pump (Minipuls 2, Gilson, Villiers-Le-Bel, France) at a total 8 ml min21 flow rate through PTFE tubes (Tygon, 2 mm id). The sample solution was injected (150 ml per injection) into a carrier stream (water) pumped by an SP8770 isocratic pump (Spectra-Physics, San Jose, CA, USA) at a flow rate of 1 ml min21 and mixed with the reagent streams in a T-piece positioned 2 cm before entering the flow cell, in which a PTFE tubing coil (Tygon, 1 mm id, 10 cm long) serving to retain the solution was placed directly in front of a photomultiplier (PMT), to generate CL emission, measured by a luminometer (Bio-Orbit 1250, Turku, Finland) linked with a computer (PC Systems, Lokeren, Belgium).An aluminium foil-covered cell is ready to collect maximum reflection of the emitted light.Extreme precautions were taken to ensure that the cell compartment and the PMT tubing were strictly light-tight. Results and Discussion Preliminary Work Preliminary work focusing on the quantitative detection of thiol-containing drugs, such as captopril and penicillamine, in a Analyst, February 1997, Vol. 122 (103–106) 103CeIV CL reaction set-up applying a fluorescent sensitizer in acidic medium was carried out in an earlier stage as previously reported.20,21 Initially, studies on some related thiol-containing drugs, such as cysteine, acetylcysteine, homocysteine and tiopronin were tentatively conducted applying an acidic potassium permanganate CL reaction system; however, no significant emission signals were produced.From four thiolcontaining drugs detected in a CeIV–quinine–sulfuric acid system under their respective optimized conditions, only tiopronin generated the most significant emission signals.Various fluorescent reagents, such as eosine, riboflavine, lucigen, rhodamine 6G and rhodamine B were employed as sensitizers and were compared with quinine for seeking a more effective alternative. The results indicated that quinine sulfate is still the best sensitizer in the present system. Method Development A series of experiments were conducted in order to establish optimum analytical conditions for the measurement of the induced CL-signals.The parameters included concentration of CeIV, quinine and of sulfuric acid, surfactants, sample volume and flow rate. Effects of concentration of CeIV, quinine and sulfuric acid The effects of the concentration of CeIV, quinine and of H2SO4 upon the CL intensity were studied and have been summarized in Figs. 1–3. When varying the concentration of each of the latter, the other parameters were held constant as illustrated in the legends for each figure. The final optimum working concentrations were chosen as CeIV 0.75 mm, quinine 0.1 mm and H2SO4 0.1 m.Effect of surfactants Surfactants, amongst other organized systems, had been investigated previously as possible enhancers of CL intensity by Zhang et al.21 in the determination of penicillamine; no significant results as to emission enhancement were observed, neither was the case in the present instance. Fig. 4 illustrates the effects of surfactants and of beta-cyclodextrin. As can be seen, no interesting enhancing effect could be noticed.On the other hand, adverse effects of these compounds upon HPLC separations may occur as noticed when using surfactants during tentative experiments. Effect of sample volume After elaborating the CL reaction conditions, the injection volume and flow-rate parameters were investigated. The variation of CL emission with the injected sample volume in the 50–150 ml range was studied similarly. Fig. 5 indicates that higher CL intensity occurs with increasing loop volumes, as expected.In the present system, a 150 ml loop was selected for the subsequent investigations. When higher CL intensity is needed, which may occur in some cases, increased injection volumes may be applied; obviously, peak broadening will then have to be dealt with. Effect of flow rate The flow rate is an important parameter in the CL reaction because the time taken to transfer the excited product into the flow cell is critical for maximum collection of the emitted Fig. 1 Effect of CeIV concentration. Quinine 0.1 mm, sulfuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 2 Effect of quinine concentration. CeIV 0.75 mm, sulfuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 3 Effect of sulfuric acid concentration. CeIV 0.75 mm, quinine 0.1 mm, flow rate 4 ml min21, tiopronin 50 mm. Fig. 4 Effect of surfactants and of beta-cyclodextrin. CeIV 0.75 mm, quinine 0.1 mm, sulfuric acid 0.1 m, flow rate 4 ml min21, tiopronin 50 mm in 1 mm of the various reagents; Triton X-100 10% (v/v). 104 Analyst, February 1997, Vol. 122light,22 too low or too high flow rates resulting in the absence of CL in the flow cell. The flow rate of the reagent solutions was initially optimized after fixing the carrier stream (water) flow rate at 2.3 ml min21 and simultaneously increasing the individual flow rates from 2 to 8 ml min21. The highest emission was obtained at 4 ml min21 for each reagent, which suggested that more light is emitted per unit of time under relatively high flow rates; this optimum flow rate apparently produces better dispersion and mixing of the reagent, higher rates may lead to both high pressures in the connector and to excessive reagent consumption.Increasing the flow rate of the carrier stream induced significant enhancement of the CL emission intensity. However, once over 2.3 ml min21 the instrumental noise increases; moreover such flow rates can not be adopted for HPLC purposes owing to an unacceptably high column pressure. As a result, 1 ml min21 of carrier stream flow rate was chosen for the calibration and pharmaceutical application.The effects of reagent flow rates in, respectively, 2.3 and 1 ml min21 carrier flow rates are shown in Fig. 6. Determination of tiopronin With the described manifold and under the optimum experimental conditions (0.75 mm CeIV and 0.1 mm quinine all in 0.1 m sulfuric acid medium; flow rate of 4 ml min21 for the reaction reagents and 1 ml min21 for the carrier stream), a linear concentration of tiopronin versus CL intensity calibration graph was obtained.A total of 10 standards were involved in the calibration process and three replicate injections of tiopronin were made per sample, the regression equation being h = 0.30 [tiopronin] + 1.068, where h is the peak height in mV and the concentration of tiopronin being expressed in mm; the correlation coefficient was 0.9994; the detection limit 0.34 mm (150 ml per injection); the RSD for 10 replicate injections of 20 and 50 mm solutions were all less than 2%.Interferences In order to assess the possible analytical applications of the described CL method, the effect of some common excipients used in pharmaceutical preparations was studied by analysing synthetic sample solutions containing 50 mm of tiopronin together with various excess amounts of excipients. The recovery results are shown in Table 1.Sorbitol shows a modest effect of CL enhancement at a concentration of 1 mm sorbitol (1 : 20 ratio), practically no emission signal is produced in the blank test. However, once the concentration of sorbitol exceeds 1 mm up to ten times, no significant increase in intensity is noticed, hence no effective emission enhancing effects can be attributed to this compound. Polyvidone and sodium carboxymethylcellulose also provide slight CL-enhancement effects, with no practical analytical applications.Adaptation of the present detection method to the analysis of the pharmaceutical preparations may be influenced Fig. 5 Effect of loop size. CeIV 0.75 mm, quinine 0.1 mm, sulphuric acid 0.1 m, tiopronin 50 mm, flow rate 4 ml min21. Fig. 6 Effect of flow rate. CeIV 0.75 mm, quinine 0.1 mm, sulfuric acid 0.1 m, tiopronin 50 mm, loop size 100 ml. Carrier stream flow rate: A, 1 ml min21; B 2.3 ml min21. Table 2 Determination of tiopronin in a pharmaceutical formulation applying the proposed CL method Sample Amount/mg Added/mg Recovered Recovery (%) Acadione* Label Found ± s (n = 10) 250 247.2 ± 2.04 250 485.8 97.7 500 730.5 97.8 750 1021.1 102.4 * Acadione is the trade mark of tiopronin manufactured by Cassenne Laboratory (Paris, France).Table 1 Recovery of tiopronin (50 mm) from various additives used as excipients Concentration ratio (additive to Recovery (%) Addition tiopronin, m/m) (n = 3) Arabic gum 1000 97.1 Lactose 1000 95.5 Galactose 1000 96.1 Saccharose 1000 98.7 Starch 1000 92.5 Carbowax 1000 100.5 Cellulose acetylphthalate Saturation 99.4 Ethyl cellulose 5 103.2 Polyvidone 5 106.1 Carboxymethylcellulose sodium 5 108.1 Dibutylphthalate 5 107 Magnesium stearate Saturation 99.2 CaHPO4 Saturation 92.4 Sorbitol 20 107 Mg(NO3)2 1000 100.9 K2SO4 1000 101.6 NaCl 10 92.5 Analyst, February 1997, Vol. 122 105by these factors, hence transfer of the developed FIA method to an HPLC system seems appropriate.Application to Pharmaceutical Preparation The proposed method was applied to the analysis of the commercial tiopronin formulation. In order to evaluate the validity of the proposed method for the determination of tiopronin in pharmaceuticals, recovery studies were carried out on samples to which known amounts of tiopronin standards were added. The results compared with the labelled contents (Table 2) and demonstrate that the method may be considered for routine analysis of the pharmaceutical preparation. However, partial interference from non-active or other active compounds should always be considered.Therefore, the HPLC coupled CL system should be envisaged to cope with the abovementioned drawbacks, and also to aim at pharmacokinetic studies of tiopronin in biological fluids. Possible CL Mechanism According to the investigation of CL properties of the fluorophore-sensitized CeIV reaction system by Zhang et al.,20 it is assumed that the possible CL mechanism is to be explained.CeIV + Tiopronin (thiol)Red ? CeIII* + TioproninOx CeIII* ? CeIII + light and/or CeIII–Tiopronin complex* ? CeIII + Tiopronin + light Where Red is reduced form; Ox is oxidixed form; and * denotes excited state. In the presence of a fluorophore (quinine), the energy resulting from the redox reaction can be effectively transferred to quinine which in turn generates CL emission: CeIII* CeIII and/or + Quinine ? and/or + Quinine* CeIII–Tiopronin complex* CeIII +Tiopronin Quinine* ? Quinine + light It is clear that in the applied CL system, based upon chemiexcitation and the use of a key sensitizer, the fluorophore plays an important role in the energy-transfer process.Conclusion The method for the determination of the thiol-containing drug tiopronin by means of CeIV oxidation sensitized by the quinine fluorophore has been successfully established as a sensitive and selective direct CL determination technique when compared with indirect CL techniques.Additionally, the method does not require sophisticated instruments. Routine drug quality control may be achieved although possible drawbacks, such as interference from formulation excipients, should be considered for each specific formulation. Further HPLC techniques coupled to the CL–FIA system are currently under investigation for the determination of tiopronin and its metabolite in human urine in pharmacokinetic studies. The authors express their gratitude to Cassenne Laboratory (Paris, France) for kindly providing the commercial formulation Acadione (tiopronin) for this work.References 1 Remien, A., and Kallistratos, G., J. Eur. Urol., 1975, 1, 227. 2 Hautman, R. E., and Robertson, W. G., Clinical and Basic Research, Plenum Press, New York, 1989, pp. 139–143. 3 Denneberg, T., Jeppson, J. O., and Stenberg, P., Proc. EDTA, 1983, 20, 427. 4 Raggi, M. A., Cesaroni, M. R., and Di-Pietra, A. M., Farmaco, Ed. Part., 1983, 38, 312. 5 Raggi, M. A., Cavrini, V., and Di-Pietra, A. M., J. Pharm. Sci., 1982, 71, 1384. 6 Raggi, M. A., Nobile, L., Cavrini, V., and Di-Pietra, A. M., Boll. Chim. Farm., 1986, 125, 295. 7 Matsuura, K., and Takashina, H., J. Chromatogr. B, Biomed. Appl., 1993, 127, 229. 8 Vinas, P., Cordoba, M. H., and Sanchez-Pedreno, C., Analyst, 1990, 115, 757. 9 Cassassas, E., Arino, C., Esteban, M., and Redondo, A., Anal. Lett., 1991, 24, 1183. 10 Kagedsal, B., Andersson, T., Carlsson, M., Denneberg, T., and Hoppe, A., J. Chromatogr. B, Biomed. Appl., 1987, 61, 261. 11 Leroy, P., Nicolas, A., Gavailoff, C., Matt, M., Netter, P., Bannwarth, B., Hercelin, B., and Massa, M., J. Chromatogr. B, Biomed. Appl., 1991, 102, 258. 12 Marzo, A., Martelli, A. E., Bruno, G., Nava, D., Mignot, A., Vidal, R., and Lefebvre, M. A., J. Chromatogr., 1991, 536, 327. 13 Kagedal, B., Carlsson, M., and Denneberg, T., J. Chromatogr. B, Biomed. Appl., 1986, 53, 301. 14 Cavrini, V., Gatti, R., Di-Pietra, A. M., and Raggi, M. A., Chromatographia, 1987, 23, 680. 15 Vinas, P., and Garcia, L., J. Pharm. Biomed. Anal., 1993, 11, 15. 16 Vinas, P., and Garcia, L., Fresenius’ J. Anal. Chem., 1993, 345, 723. 17 Perez-Ruiz, T., Martinez-Lozano, C., Tomas, V., and Lambertos, G., J. Microchem., 1991, 44, 72. 18 Pagan, A., Anal. Ciencias, University Murcia, XLVII, 1988, 29–32. 19 Garcia, M. S., Sanchez-Pedreno, C., Alberto, M. I., and Rodenas, V., J. Pharm. Biomed. Anal., 1993, 11, 633. 20 Zhang, X. R., Baeyens, W., Calokerinos, A. C., Imai, K., and Van Der Weken, G., Anal. Chim. Acta, 1995, 303, 121. 21 Zhang, Z. D., Baeyens, W., Zhang, X. R., Calokerinos, A. C., and Van Der Weken, G., Biomed. Chromatogr., 1995, 9, 287. 22 Zhang, Z. D., Baeyens, W., Zhang, X. R., Van Der Weken, G., J. Pharm. Biomed. Anal., 1996, 14, 939. Paper 6/05703I Received August 14, 1996 Accepted October 30, 1996 106 Analyst, February 1997, Vol. 122

ISSN:0003-2654    

DOI:10.1039/a605703i

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

p C Relative potential/mV 0 300 250 200 150 100 50 0 1 2 3 4 5 6 DE 30 cm h–1 j i h g f e d c b a D C B A D C B A 0 0 –20 –40 –60 –80 –120 –140 –100 1 2 3 4 5 pC Relative potential/mV C B A 0 0 160 140 120 100 80 60 40 20 1 2 3 4 5 6 pC Relative potential/mV } v = 30 cm h–1 D E [Hg2+] [Ag+] 10–1 10–2 10–3 10–4 10–5 10–3 meter (PM) and an OH-814/1 recorder (RC). The manifold was made of polyethylene tubing (0.3 mm id) and the samples were injected with a Hamilton syringe. The tubular membranes (WE, RE) were made of pressed pellets15 with a channel (3 ml volume) drilled lengthways through the middle of the pellets. Solid contact (Ag) was ensured for transmission of the potential signal to the pH/mVmeter.The potentiometric differential detector is shown in Fig. 1B. Working Procedures The carrier stream was pumped at a constant flow rate. Two working procedures were used to introduce the working standard solutions or samples: the first procedure involved injection of the solutions into the carrier stream for transport to the working electrode and using the same carrier solution for the reference channel.The second procedure involved passing the solutions directly through the measuring cell. Results and Discussion In order to obtain the best response of the detector, each ion was investigated separately. The following factors were studied: (a) the ionic strength of the carrier stream (jcarrier solution); (b) the injection volume (Vsample); and (c) the ionic strength of the sample (jsample).The ionic strength of the carrier stream was varied between 1 and 0.03; a value of 0.1 was found to be the optimum for all three ions. In Fig. 2 the experimental results are presented for the AgI ion. They were obtained under conditions where the injected sample volume and ionic strength were optimum. The injected volumes were varied between 30 and 500 ml. It was found that the optimum injection volume, for all three ions, was 500 ml.Fig. 3 shows the results for the S22 ion. The ionic strength of the samples was varied between 1 and 0.03 and the following optimum values were obtained: 0.1 for AgI and S22; 0.03 for CuII. In Fig. 4 the experimental results for the CuII ion are presented. The best lower concentration limits (Cinf) of the linear response range are 1026 mol dm23 for AgI, 5 3 1026 mol dm23 for CuII and 1025 mol dm23 for S22 in the injection procedure. The lower concentration limit is shifted towards higher concentrations with increasing ionic strength of the solutions.For example, at j = 1, Cinf = 5 3 1026 mol dm23 AgI as opposed to 1026 mol dm23 AgI at j = 0.1. A statistical analysis of the calibration data obtained by the injection procedure gave the following results: E = 349.28 2 57.13 pCAgI standard deviation: 0.01 for A and 0.003 for B, E Fig. 2 Silver(i) calibration graphs. Injection procedure: A, jcarrier solution = 0.1; Vsample = 500 ml; jsample = 0.1.Aspiration procedure: B, jcarrier solution = 0.1; C, jcarrier solution = 0.2; D, jcarrier solution = 1. Reference carrier solution: 0.1 mol dm23 KNO3 containing 1026 mol dm23 AgI. a = 5 3 1026; b = 1025; c = 5 3 1025; d = 1024; e = 5 3 1024; f = 1023; g = 5 3 1023; h = 1022; i = 5 3 1022; and j = 1021 mol dm23. Fig. 3 Sulfide calibration graphs. Injection procedure: A, Vsample = 500 ml; D, Vsample = 300 ml; C, Vsample = 30 ml. Aspiration procedure: B, jsample = 0.1.Reference carrier solution: sulfide antioxidant buffer reagent containing 1026 mol dm23 S22; j = 0.1. Fig. 4 Copper(ii) calibration graphs. Injection procedure: A, jsample = 0.03; C, jsample = 0.3; Vsample = 500 ml. Aspiration procedure: B, jsample = 0.03. Reference carrier solution: 1021 mol dm23 KNO3 containing 1026 mol dm23 CuII; j = 0.1. Fig. 5 Effect of HgII (injection procedure) on the detector response to 1023 mol dm23 AgI. 108 Analyst, February 1997, Vol. 122= 204.47 2 33.14 pCCuII; standard deviation: 0.195 for A and 0.051 for B; E = 2 179.84 + 28.89 pCS22; standard deviation: 0.017 for A and 0.005 for B (A and B are the parameters of the calibration equation). The response of the detector in the injection procedure is virtually instantaneous because the distance between the injector and the membrane is minimal. Even though the detector does not attain a steady state, the calibration graph obtained with the injection procedure is parallel to that obtained with the aspiration procedure.Because the detector contains two similar membranes, all other influences are eliminated (temperature, medium, etc.). Also, very good stability of the calibration graph is ensured. The detector can be used for a long period (5–6 months) without any special treatment. The response of the detector was evaluated in the presence of various foreign ions, viz., CuII, CdII, HgII, PbII, Cl2, Br2, I2 and SCN2, for the Ag2S membrane and AgI, CdII and PbII for the CuS–Ag2S membrane, and the results were compared with those of the respective classical electrodes used under static conditions. The results did not differ significantly.The only interference was that of HgII on the Ag2S detector (Fig. 5). At concentrations up to 1023 mol dm23 HgII the membrane was strongly poisoned; however, this can be reversed by washing the membrane for 10 min with 1 mol dm23 KNO3 containing 1026 mol dm23 AgI. The accuracy of the detector for analytical applications is illustrated by the results given in Table 1.Practical results (columns 4–6) were obtained by using the calibration equations given above. A good agreement with the theoretical values (columns 1–3) was obtained (see data from columns 7–9). The throughput was 60 samples h21 for AgI; 60 samples h21 for CuII in the concentration range 1026–1023 mol dm23 and 40 samples h21 when the concentration of CuII was > 5 3 1023 mol dm23; and 60 samples h21 for S22.The samples are not affected by each other. They can be injected in any order. References 1 R°u�zi�cka, J., and Hansen, E. H., Anal. Chim. Acta, 1975, 78, 145. 2 Van der Linden, W. E., and Ostervink, R., Anal. Chim. Acta, 1978, 101, 419. 3 Mascini, M., and Palleschi, G., Anal. Chim. Acta, 1978, 100, 215. 4 Alegret, S., Alonso, J., Bartroli, J., Paulis, J. M., Lima, J. L. F. C., and Machado, A. A. S. C., Anal. Chim. Acta, 1984, 164, 147. 5 Van Staden, J.F., Anal. Chim. Acta, 1986, 179, 407. 6 Van Staden, J. F., Anal. Lett., 1986, 19, 1407. 7 Van Staden, J. F., Analyst, 1986, 111, 1231. 8 Van Staden, J. F., Fresenius’ Z. Anal. Chem., 1987, 328, 68. 9 Van Staden, J. F., Analyst, 1987, 112, 595. 10 Van Staden, J. F., Fresenius’ Z. Anal. Chem., 1989, 333, 226. 11 Najib, F. M., and Othman, S., Talanta, 1992, 39, 1259. 12 Ferreira, I. M. P. L. V. Q., and Lima, J. L. F. C., Analyst, 1994, 119, 209. 13 Hop�ýrtean, E., Cosma, V., and Coroian, A., Rom.Pat., 98 786, 1989. 14 Van Staden, J. F., Analyst, 1988, 113, 885. 15 Hop�ýrtean, E., and Olenic, L., Rom. Pat., 108 504, 1996. Paper 6/05410B Received August 2, 1996 Accepted October 31, 1996 Table 1 Accuracy of AgI, CuII and S22 determinations by direct potentiometry Results obtained by injection Concentration taken/mg dm23 procedure/mg dm23 Recovery (%) Sample No. CuII AgI S22 CuII AgI S22 CuII AgI S22 1 0.32 0.54 0.64 0.33 0.56 0.66 103.0 103.7 103.1 (0.01)* (0.01)* (0.01)* 2 1.27 2.16 1.60 1.25 2.21 1.62 99.4 102.3 101.2 (0.02)* (0.04)* (0.03)* 3 3.18 5.40 6.41 3.14 5.49 6.37 99.7 101.6 99.4 (0.05)* (0.08)* (0.08)* 4 12.71 21.58 16.03 12.25 21.25 15.41 96.4 98.5 96.1 (0.15)* (0.20)* (0.15)* 5 31.77 53.95 64.14 30.85 53.57 62.33 98.1 99.3 97.2 (0.25)* (0.50)* (0.50)* 6 127.10 215.80 160.35 123.34 218.80 155.31 97.0 97.3 96.6 (0.90)* (1.50)* (1.20)* * Values in parentheses are standard deviations (s) obtained with repeated sample injections (n = 5).Analyst, February 1997, Vol. 122 109 p C Relative potential/mV 0 300 250 200 150 100 50 0 1 2 3 4 5 6 DE 30 cm h–1 j i h g f e d c b a D C B A D C B A 0 0 –20 –40 –60 –80 –120 –140 –100 1 2 3 4 5 pC Relative potential/mV C B A 0 0 160 140 120 100 80 60 40 20 1 2 3C Relative potential/mV } v = 30 cm h–1 D E [Hg2+] [Ag+] 10–1 10–2 10–3 10–4 10–5 10–3 meter (PM) and an OH-814/1 recorder (RC). The manifold was made of polyethylene tubing (0.3 mm id) and the samples were injected with a Hamilton syringe. The tubular membranes (WE, RE) were made of pressed pellets15 with a channel (3 ml volume) drilled lengthways through the middle of the pellets.Solid contact (Ag) was ensured for transmission of the potential signal to the pH/mVmeter. The potentiometric differential detector is shown in Fig. 1B. Working Procedures The carrier stream was pumped at a constant flow rate. Two working procedures were used to introduce the working standard solutions or samples: the first procedure involved injection of the solutions into the carrier stream for transport to the working electrode and using the same carrier solution for the reference channel.The second procedure involved passing the solutions directly through the measuring cell. Results and Discussion In order to obtain the best response of the detector, each ion was investigated separately. The following factors were studied: (a) the ionic strength of the carrier stream (jcarrier solution); (b) the injection volume (Vsample); and (c) the ionic strength of the sample (jsample).The ionic strength of the carrier stream was varied between 1 and 0.03; a value of 0.1 was found to be the optimum for all three ions. In Fig. 2 the experimental results are presented for the AgI ion. They were obtained under conditions where the injected sample volume and ionic strength were optimum. The injected volumes were varied between 30 and 500 ml.It was found that the optimum injection volume, for all three ions, was 500 ml. Fig. 3 shows the results for the S22 ion. The ionic strength of the samples was varied between 1 and 0.03 and the following optimum values were obtained: 0.1 for AgI and S22; 0.03 for CuII. In Fig. 4 the experimental results for the CuII ion are presented. The best lower concentration limits (Cinf) of the linear response range are 1026 mol dm23 for AgI, 5 3 1026 mol dm23 for CuII and 1025 mol dm23 for S22 in the injection procedure.The lower concentration limit is shifted towards higher concentrations with increasing ionic strength of the solutions. For example, at j = 1, Cinf = 5 3 1026 mol dm23 AgI as opposed to 1026 mol dm23 AgI at j = 0.1. A statistical analysis of the calibration data obtained by the injection procedure gave the following results: E = 349.28 2 57.13 pCAgI standard deviation: 0.01 for A and 0.003 for B, E Fig. 2 Silver(i) calibration graphs. Injection procedure: A, jcarrier solution = 0.1; Vsample = 500 ml; jsample = 0.1. Aspiration procedure: B, jcarrier solution = 0.1; C, jcarrier solution = 0.2; D, jcarrier solution = 1. Reference carrier solution: 0.1 mol dm23 KNO3 containing 1026 mol dm23 AgI. a = 5 3 1026; b = 1025; c = 5 3 1025; d = 1024; e = 5 3 1024; f = 1023; g = 5 3 1023; h = 1022; i = 5 3 1022; and j = 1021 mol dm23. Fig. 3 Sulfide calibration graphs. Injection procedure: A, Vsample = 500 ml; D, Vsample = 300 ml; C, Vsample = 30 ml.Aspiration procedure: B, jsample = 0.1. Reference carrier solution: sulfide antioxidant buffer reagent containing 1026 mol dm23 S22; j = 0.1. Fig. 4 Copper(ii) calibration graphs. Injection procedure: A, jsample = 0.03; C, jsample = 0.3; Vsample = 500 ml. Aspiration procedure: B, jsample = 0.03. Reference carrier solution: 1021 mol dm23 KNO3 containing 1026 mol dm23 CuII; j = 0.1. Fig. 5 Effect of HgII (injection procedure) on the detector response to 1023 mol dm23 AgI. 108 Analyst, February 1997, Vol. 122= 204.47 2 33.14 pCCuII; standard deviation: 0.195 for A and 0.051 for B; E = 2 179.84 + 28.89 pCS22; standard deviation: 0.017 for A and 0.005 for B (A and B are the parameters of the calibration equation). The response of the detector in the injection procedure is virtually instantaneous because the distance between the injector and the membrane is minimal. Even though the detector does not attain a steady state, the calibration graph obtained with the injection procedure is parallel to that obtained with the aspiration procedure.Because the detector contains two similar membranes, all other influences are eliminated (temperature, medium, etc.). Also, very good stability of the calibration graph is ensured. The detector can be used for a long period (5–6 months) without any special treatment. The response of the detector was evaluated in the presence of various foreign ions, viz., CuII, CdII, HgII, PbII, Cl2, Br2, I2 and SCN2, for the Ag2S membrane and AgI, CdII and PbII for the CuS–Ag2S membrane, and the results were compared with those of the respective classical electrodes used under static conditions.The results did not differ significantly. The only interference was that of HgII on the Ag2S detector (Fig. 5). At concentrations up to 1023 mol dm23 HgII the membrane was strongly poisoned; however, this can be reversed by washing the membrane for 10 min with 1 mol dm23 KNO3 containing 1026 mol dm23 AgI.The accuracy of the detector for analytical applications is illustrated by the results given in Table 1. Practical results (columns 4–6) were obtained by using the calibration equations given above. A good agreement with the theoretical values (columns 1–3) was obtained (see data from columns 7–9). The throughput was 60 samples h21 for AgI; 60 samples h21 for CuII in the concentration range 1026–1023 mol dm23 and 40 samples h21 when the concentration of CuII was > 5 3 1023 mol dm23; and 60 samples h21 for S22.The samples are not affected by each other. They can be injected in any order. References 1 R°u�zi�cka, J., and Hansen, E. H., Anal. Chim. Acta, 1975, 78, 145. 2 Van der Linden, W. E., and Ostervink, R., Anal. Chim. Acta, 1978, 101, 419. 3 Mascini, M., and Palleschi, G., Anal. Chim. Acta, 1978, 100, 215. 4 Alegret, S., Alonso, J., Bartroli, J., Paulis, J. M., Lima, J.L. F. C., and Machado, A. A. S. C., Anal. Chim. Acta, 1984, 164, 147. 5 Van Staden, J. F., Anal. Chim. Acta, 1986, 179, 407. 6 Van Staden, J. F., Anal. Lett., 1986, 19, 1407. 7 Van Staden, J. F., Analyst, 1986, 111, 1231. 8 Van Staden, J. F., Fresenius’ Z. Anal. Chem., 1987, 328, 68. 9 Van Staden, J. F., Analyst, 1987, 112, 595. 10 Van Staden, J. F., Fresenius’ Z. Anal. Chem., 1989, 333, 226. 11 Najib, F. M., and Othman, S., Talanta, 1992, 39, 1259. 12 Ferreira, I. M. P. L. V. Q., and Lima, J. L. F. C., Analyst, 1994, 119, 209. 13 Hop�ýrtean, E., Cosma, V., and Coroian, A., Rom. Pat., 98 786, 1989. 14 Van Staden, J. F., Analyst, 1988, 113, 885. 15 Hop�ýrtean, E., and Olenic, L., Rom. Pat., 108 504, 1996. Paper 6/05410B Received August 2, 1996 Accepted October 31, 1996 Table 1 Accuracy of AgI, CuII and S22 determinations by direct potentiometry Results obtained by injection Concentration taken/mg dm23 procedure/mg dm23 Recovery (%) Sample No. CuII AgI S22 CuII AgI S22 CuII AgI S22 1 0.32 0.54 0.64 0.33 0.56 0.66 103.0 103.7 103.1 (0.01)* (0.01)* (0.01)* 2 1.27 2.16 1.60 1.25 2.21 1.62 99.4 102.3 101.2 (0.02)* (0.04)* (0.03)* 3 3.18 5.40 6.41 3.14 5.49 6.37 99.7 101.6 99.4 (0.05)* (0.08)* (0.08)* 4 12.71 21.58 16.03 12.25 21.25 15.41 96.4 98.5 96.1 (0.15)* (0.20)* (0.15)* 5 31.77 53.95 64.14 30.85 53.57 62.33 98.1 99.3 97.2 (0.25)* (0.50)* (0.50)* 6 127.10 215.80 160.35 123.34 218.80 155.31 97.0 97.3 96.6 (0.90)* (1.50)* (1.20)* * Values in parentheses are standard deviations (s) obtained with repeated sample injections (n = 5). Analyst, February 1997, Vol.

ISSN:0003-2654    

DOI:10.1039/a605410b

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Determination of Lignocaine Hydrochloride by Ion-pairing Flow Injection With Piezoelectric Detection Mo Zhihong*, Luo Jie and Li Menglong Department of Applied Chemistry, Sichuan Union University, Chengdu 610065, China A rapid and simple method for the determination of lignocaine hydrochloride by ion-pairing flow injection analysis was developed. It is based on the formation of an ion pair between lignocaine hydrochloride and sodium dodecyl phenylsulfonate and piezoelectric detection.The calibration curve was linear between 0.01 and 2.00 mg ml21, with a detection limit of 8 mg ml21, an RSD of 0.29% (10 replicates) and a sampling frequency of 120 h21. The proposed method was satisfactorily applied to the determination of lignocaine hydrochloride in pharmaceutical preparations. Keywords: Ion pair; flow injection; piezoelectric detection; pharmaceutical analysis; lignocaine hydrochloride Lignocaine (N-diethylaminoacetyl-2,6-xylidine) hydrochloride is a local anaesthetic, widely used for injections and for local application to mucous membranes.At an equal concentration, it is more effective than procaine and gives a greater degree of anaesthesia. Preparations containing 1–4% of lignocaine hydrochloride can be used for epidural anaesthesia and for anaesthetizing the pharynx, larynx and trachea prior to endoscopic examination. Several methods for the determination of lignocaine hydrochloride have been reported in recent years, including liquid chromatography,1,2 spectrophotometry,3,4 polarography and ion-selective electrodes.5–8 Liquid-liquid distribution of ion associates with spectrophotometric detection has been widely applied in pharmaceutical analysis.Since the initial work of Karlberg and Thelander9 on the determination of caffeine in acetylsalicylic acid preparations, a number of applications of ion-pair formation with automated flow injection solvent extraction to drug evaluation have appeared.Recently published methods have used dyes as ion-pairing reagents.10,11 The application of flow injection has resulted in a method which offers the promise of faster and more reproducible assays. However, most of them need solvent extraction, leading to the incorporation of segmentors and consumption of organic solvents. Consequently, their applications have been limited to a great extent. The ability of a piezoelectric shear wave device, commonly referred to as the quartz crystal microbalance (QCM), to respond to small changes in mass at its surface while immersed in liquid has led to numerous fundamental investigations of interfacial phenomena, applications to determination of inorganic species and pharmaceuticals and developments of biosensors.12,13 Recently, ion-pairing flow injection with piezoelectric detection (IPFIP) has been proposed, based on direct mass sensing to the adsorption and desorption of an ion-pairing reagent without solvent extraction, and applied satisfactorily to the determination of catecholamines in pharmaceutical preparations. 14 The purpose of this investigation was to develop a simple, rapid and reliable assay for lignocaine hydrochloride in pharmaceutical preparations using IPFIP. Theory Immersing only one crystal face in a viscous liquid, the resonant frequency can be affected by the mass of the crystal surface and the density and viscosity of the liquid, and much less influenced by the conductivity of the liquid.12 With the addition of surfactant, the density and viscosity of solution vary slightly, but the resonant frequency changes significantly and becomes stable in a time proportional to the concentration of surfactant.Consequently, the frequency change is caused by the small mass change of the quartz electrode surface due to the adsorption of surfactant. The formation of an ion pair between the surfactant and the sample results in desorption of the surfactant, hence leading to a change in frequency. The response of the resonant frequency versus time to sample injection in an IPFIP system has been verified to be as follows:14 fp = K Cs (1a) K = abk1tCR D 1 - k2TCR D -1 D Ê Ë Á � � �(1b) where fp is the peak height of the response; Cs and CR are concentrations of the sample injected and the ion-pairing reagent, respectively; a is the mass sensitivity of the 10 MHz piezoelectric detector, equal to 0.226 Hz cm2 ng21; b is a constant related to the adsorption of the surfactant on the crystal electrode and the flowing state; k1 and k2 are rate constants for the interfacial reaction between sample in solution and surfactant adsorbed on electrode, and the homogeneous reaction between sample and surfactant in the mixing coil, respectively; t and T are times of the interfacial reaction and the homogeneous reaction, respectively; and D is the distribution degree of the sample in the mixing coil.Hence the peak height of the response is linearly related to the concentration of injected sample solutions. This is the basis of the IPFIP method for the determination of lignocaine hydrochloride. Experimental Apparatus The flow injection system and flow-through piezoelectric detector were the same as described previously.14 One face only of the crystal was exposed to a cell volume of 50 ml.A laboratory-made TTL-IC oscillator, described previously,15 was connected to the crystal electrodes by platinum foils pressed to the electrode surfaces and the resonant frequency was monitored with a universal counter (SS7200, Shijiazhuang Electronic Factory No. 4). Solutions were pumped with a speed modulation pump (LP-2A, Xintong Scientific Instrument, Academia Sinica) fitted with Tygon tubes. Samples were injected from a four-way multifunctional automatic valve (V-16A, Xintong Scientific Instrument, Academia Sinica) fitted with a by-pass coil (50 ml). Flow lines were PTFE tubing (0.8 mm id).Reagents All reagents were of analytical-reagent grade and ion-exchanged, distilled water was used throughout. A sodium dodecyl phenylsulfonate (SDPS) solution (0.05 mol l21) and a sodium dodecyl sulfonate (SDS) solution (0.05 Analyst, February 1997, Vol. 122 (111–113) 111mol l21) were prepared in water and from these, dilute carrier solutions were prepared. A dilute hydrochloric acid solution (0.20 mol l21) was prepared to adjust the pH of carrier solutions, checked by a pH meter.Stock solution of lignocaine hydrochloride (10 mg ml21) was prepared in water and stored frozen in a dark bottle. From these, working solutions were prepared as required. General Procedure Volumes of 50 ml of sample and calibration standard solutions were injected into the flow when the base resonant frequency (Fb) was stable, and the peak heights were found to be fp = Fp 2 Fb, where Fp is the maximum frequency during the run. Under the conditions that the flow rate was 2.0 ml min21 and the length of the mixing coil was 300 mm, the recorded peaks were sharp and the baselines were stable at an injection rate of 120 h21.Results and Discussion Selection of Ion-pairing Reagent Fig.1 illustrates the response of frequency versus time to injection of lignocaine hydrochloride into the flowing solution of SDPS and SDS. Sharp peaks occur where the upward slope corresponds to the decrease in adsorption primarily caused by the interfacial ion association between adsorbed surfactants and injected lignocaine hydrochloride.It can be seen that the peak using SDPS as the carrier is higher and sharper than that using SDS as the carrier. The reason is that SDPS is more favourable than SDS towards adsorption on the electrode and the constant b, for SDPS is larger than that for SDS. Hence SDPS was used as the ion-pairing reagent in the investigation. Selection of Operating Parameters The operating parameters studied were flow rate and length of the mixing coil.Increases in flow rate and coil length result in decreases in the reaction times, t and T, and distribution degree, D. The effects of t, T, and D on the response can be derived from eqn. (1), namely that the peak height decreases with increases in T and D and decrease in t. The concentrations used in these experiments were SDPS solution used as the carrier 5 mmol l21 and lignocaine hydrochloride sol1.00 mg ml21. The peak heights were found to increase as the flow rate increased in the range 0.5–1.5 ml min21, which indicates that the effects of decreases of the reaction time, T and distribution degree, D, are greater, and decrease when the flow rate was higher than 1.5 ml min21, which indicates that the effect of the decrease in the reaction time, t, is predominant.In order to increase the throughput, a flow rate of 2.0 ml min21 was adopted as a compromise. With the increase in the length of the mixing coil, the peak heights and widths were found to decrease and increase, respectively.The optimum length of the mixing coil was found to be 300 mm (total volume 0.2 ml) to give the sharpest peak and most stable baseline. Optimization of Reagent Concentration and pH The effect of the concentration of SDPS was studied in the range 0.5–10 mmol l21. Below 5 mmol l21, the peak height was found to be greater and became gradually unchangeable at higher SDPS concentration. Conversely, above 5 mmol l21, the peak height was smaller and decreased to zero.It can be seen from eqn. (1) that at the maximum peak height the concentration of SDPS is equal to D/2k2T(D-1) , which is inversely proportional to the reaction rate and time in the mixing coil. In subsequent work, 5 mmol l21 SDPS solution was used as the carrier. Fig. 2 illustrates the effect of pH on the peak height. As the carrier pH was increased in the range 1–3, the peak height increased significantly.In contrast, as the carrier pH increased in the range 3–7, the peak height decreased and gradually became unchangeable. Because of the protonation of SDPS, the concentration of free SDPS which participates in adsorption and ion association increases as the carrier pH increases. Hence the effect of the carrier pH on the response is similar to that of SDPS concentration as described above. The carrier pH was adjusted to 3 with hydrochloric acid since the dilute SDPS solution was about neutral. Calibration Graph and Reproducibility Using the flow injection system and the conditions SDPS carrier concentration 5 mmol l21, flow rate 2.0 ml min21 and length of the mixing coil 300 mm, a linear calibration graph of the relative peak height with respect to the blank (water) versus the concentration of lignocaine hydrochloride was obtained in the range 0.01– 2.00 mg ml21, with a slope of 800 Hz ml mg21 and a correlation coefficient of 0.9995.The standard deviation of the responses for the blank injection (10 replicates) was 2.55 Hz.The detection limit (3 3 noise), calculated as 3 3 2.55/800, was 8 mg ml21. The RSD 1.00 mg ml21 lignocaine hydrochloride (10 replicates) was 0.29%. The sampling rate was about 120 h21. Interferences The influence of foreign compounds that commonly accompany lignocaine hydrochloride in pharmaceutical preparations was Fig. 1 FIA response of resonant frequency versus time to lignocaine hydrochloride injection using SDPS and SDS as carrier.Concentration of lignocaine hydrochloride, 2.0 mg ml21; concentration of the carrier (pH 3), 5 mmol l21; flow rate, 2.0 ml min21; and length of mixing coil, 300 mm. Fig. 2 Effect of carrier pH on frequency response of ion-pairing flow injection to lignocaine hydrochloride. Concentration of lignocaine hydrochloride, 2.0 mg ml21; concentration of the carrier (SDPS), 5 mmol l21; flow- rate, 2.0 ml min21; and length of mixing coil, 300 mm. 112 Analyst, February 1997, Vol. 122studied by preparing solutions containing 1.00 mg ml21 lignocaine hydrochloride and increasing concentrations of the potential interferents up to 50 mg ml21 or by adding an amount to give an error of ±3%. The errors were determined by comparison with the peak heights given by a solution of analyte containing no foreign substances. Glucose, sucrose, lactose, asparate, citrate and tartrate were tolerated in large amounts (50 mg ml21 was the maximum tested) and 10 mg ml21 of antipyrine, atropine, berberine, sparteine and pilocarpine hydrochloride were also tolerated in the determination of lignocaine hydrochloride.Analysis of Pharmaceutical Preparations The proposed method was applied satisfactorily to the determination of lignocaine hydrochloride in pharmaceutical preparations. Commercially available formulations were analysed and the results obtained are summarized in Table 1. As can be seen, for all formulations the assay results were in good agreement with values for the nominal contents and with those obtained using a non-aqueous titrimetric method.16 The recoveries obtained by adding lignocaine hydrochloride to each pharmaceutical formulation are given in Table 2; they were determined by subtracting the results obtained for the pharmaceutical formulations prepared in a similar manner but to which no lignocaine hydrochloride had been added.Conclusions The proposed method for the determination of lignocaine hydrochloride, based on ion-pairing flow injection with piezoelectric detection, represents an important change to conventional approaches utilizing ion pairing.Usually, ion-pairing alone cannot provide a sufficient change in measurable properties for analysis, leading to the requirement for a separation (an extraction step) between free ion-pairing reagent and ion pair. The proposed method eliminates the separation, the ion-pairing reagent being fixed on a mass-sensing surface.It has distinct features. First, there is greater hydrodynamic and instrumental simplicity resulting from the elimination of the organic phase and segmenter. Second, the sampling frequency, accuracy and precision are superior to those for conventional methods for the analysis of pharmaceutical preparations. Although the sensitivity and selectivity may be relatively poor, they are sufficient for primary component analysis in pharmaceuticals. Improvements may be realized through optimizations of the ion-pairing reagent, operating parameters and detector and the signal processing. References 1 Murakita, H., Hayashi, M., Mikami, H., and Ishida, Y., Bunseki Kagaku, 1986, 35, 236. 2 Zhou, J.-X., Zhang, L.-J., and Wang, E.-K., Electroanalysis (N. Y.), 1993, 5, 295. 3 Korany, M. A., Wahbi, A. M., Elsayed, M. A., and Mandour, S., Anal. Lett., 1984, 17, 1373. 4 Xia, G., and Lui, S., Yaoxue Tongbao, 1986, 21, 25. 5 Huang, T., Pan, D., and Gao, H., Yaowu Fenxi Zazhi, 1984, 4, 21. 6 Shoukry, A. F., Issa, Y. M., El-Sheik, R., and Zareh, M., Microchem. J., 1988, 37, 299. 7 Bouklouze, A. A., El-Jammal, A., Patriarche, G. J., and Christian, G. D., J. Pharm. Biomed. Anal., 1991, 9, 393. 8 Manzo, R. H., Luna, E., and Allemandi, D. A., J. Pharm. Sci., 1991, 80, 80. 9 Karlberg, B., and Thelander, S., Anal. Chim. Acta, 1978, 98, 1. 10 Calatayud, J. M., Sampedro, A. S., and Sarrion, S. N., Analyst, 1990, 115, 855. 11 Sakai, T., Ohta, H., Ohno, N., and Sasaki, H., Fresenius’ J.Anal. Chem., 1994, 349, 475. 12 Buttry, D. A., and Ward, M. D., Chem. Rev., 1992, 92, 1355. 13 Yao, S.-Z., and Nie, L.-H., Chim. J. Anal. Chem., 1996, 24, 234. 14 Mo, Z.-H., Zhang, M.-J., and Xia, Z.-L., Anal. Lett., 1997, 30. 15 Yao, S.-Z., and Mo, Z.-H., Anal. Chim. Acta, 1987, 193, 97. 16 Chinese Pharmacopoeia, Part II, Chemical Industry Press and People’s Press of Hygiene, Beijing, 1995, Appendix 4, p. 402. Paper 6/05796I Received August 20, 1996 Accepted November 4, 1996 Table 1 Determination of lignocaine hydrochloride in pharmaceutical preparations Nominal, value/ Found by titrimetric Found by FI method†/ Sample* mg ml21 method†/mg ml21 mg ml21 1 0.5 0.49 ± 0.004 0.49 ± 0.002 2 1.0 1.01 ± 0.007 0.99 ± 0.003 3‡ 10 9.94 ± 0.10 10.01 ± 0.03 * Supplied by three manufacturers, 1 and 2, injection; 3, capsule. † Average of three determinations ± standard deviation.‡ Prepared by dissolving a 250 mg capsule in 25 ml of water.Table 2 Recovery of lignocaine hydrochloride from pharmaceutical preparations Sample* Added/mg ml21 Found†/mg ml21 Recovery (%) 1 1.00 0.99 99.0 5.00 5.01 100.2 2 1.00 1.01 101.0 5.00 5.04 100.8 3 1.00 1.03 103.0 5.00 5.10 102.0 * See Table 1. † Average of three determinations. Analyst, February 1997, Vol. 122 113 Determination of Lignocaine Hydrochloride by Ion-pairing Flow Injection With Piezoelectric Detection Mo Zhihong*, Luo Jie and Li Menglong Department of Applied Chemistry, Sichuan Union University, Chengdu 610065, China A rapid and simple method for the determination of lignocaine hydrochloride by ion-pairing flow injection analysis was developed. It is based on the formation of an ion pair between lignocaine hydrochloride and sodium dodecyl phenylsulfonate and piezoelectric detection.The calibration curve was linear between 0.01 and 2.00 mg ml21, with a detection limit of 8 mg ml21, an RSD of 0.29% (10 replicates) and a sampling frequency of 120 h21.The proposed method was satisfactorily applied to the determination of lignocaine hydrochloride in pharmaceutical preparations. Keywords: Ion pair; flow injection; piezoelectric detection; pharmaceutical analysis; lignocaine hydrochloride Lignocaine (N-diethylaminoacetyl-2,6-xylidine) hydrochloride is a local anaesthetic, widely used for injections and for local application to mucous membranes. At an equal concentration, it is more effective than procaine and gives a greater degree of anaesthesia.Preparations containing 1–4% of lignocaine hydrochloride can be used for epidural anaesthesia and for anaesthetizing the pharynx, larynx and trachea prior to endoscopic examination. Several methods for the determination of lignocaine hydrochloride have been reported in recent years, including liquid chromatography,1,2 spectrophotometry,3,4 polarography and ion-selective electrodes.5–8 Liquid-liquid distribution of ion associates with spectrophotometric detection has been widely applied in pharmaceutical analysis.Since the initial work of Karlberg and Thelander9 on the determination of caffeine in acetylsalicylic acid preparations, a number of applications of ion-pair formation with automated flow injection solvent extraction to drug evaluation have appeared. Recently published methods have used dyes as ion-pairing reagents.10,11 The application of flow injection has resulted in a method which offers the promise of faster and more reproducible assays.However, most of them need solvent extraction, leading to the incorporation of segmentors and consumption of organic solvents. Consequently, their applications have been limited to a great extent. The ability of a piezoelectric shear wave device, commonly referred to as the quartz crystal microbalance (QCM), to respond to small changes in mass at its surface while immersed in liquid has led to numerous fundamental investigations of interfacial phenomena, applications to determination of inorganic species and pharmaceuticals and developments of biosensors.12,13 Recently, ion-pairing flow injection with piezoelectric detection (IPFIP) has been proposed, based on direct mass sensing to the adsorption and desorption of an ion-pairing reagent without solvent extraction, and applied satisfactorily to the determination of catecholamines in pharmaceutical preparations. 14 The purpose of this investigation was to develop a simple, rapid and reliable assay for lignocaine hydrochloride in pharmaceutical preparations using IPFIP.Theory Immersing only one crystal face in a viscous liquid, the resonant frequency can be affected by the mass of the crystal surface and the density and viscosity of the liquid, and much less influenced by the conductivity of the liquid.12 With the addition of surfactant, the density and viscosity of solution vary slightly, but the resonant frequency changes significantly and becomes stable in a time proportional to the concentration of surfactant. Consequently, the frequency change is caused by the small mass change of the quartz electrode surface due to the adsorption of surfactant.The formation of an ion pair between the surfactant and the sample results in desorption of the surfactant, hence leading to a change in frequency. The response of the resonant frequency versus time to sample injection in an IPFIP system has been verified to be as follows:14 fp = K Cs (1a) K = abk1tCR D 1 - k2TCR D -1 D Ê Ë Á � � �(1b) where fp is the peak height of the response; Cs and CR are concentrations of the sample injected and the ion-pairing reagent, respectively; a is the mass sensitivity of the 10 MHz piezoelectric detector, equal to 0.226 Hz cm2 ng21; b is a constant related to the adsorption of the surfactant on the crystal electrode and the flowing state; k1 and k2 are rate constants for the interfacial reaction between sample in solution and surfactant adsorbed on electrode, and the homogeneous reaction between sample and surfactant in the mixing coil, respectively; t and T are times of the interfacial reaction and the homogeneous reaction, respectively; and D is the distribution degree of the sample in the mixing coil.Hence the peak height of the response is linearly related to the concentration of injected sample solutions. This is the basis of the IPFIP method for the determination of lignocaine hydrochloride.Experimental Apparatus The flow injection system and flow-through piezoelectric detector were the same as described previously.14 One face only of the crystal was exposed to a cell volume of 50 ml. A laboratory-made TTL-IC oscillator, described previously,15 was connected to the crystal electrodes by platinum foils pressed to the electrode surfaces and the resonant frequency was monitored with a universal counter (SS7200, Shijiazhuang Electronic Factory No. 4). Solutions were pumped with a speed modulation pump (LP-2A, Xintong Scientific Instrument, Academia Sinica) fitted with Tygon tubes. Samples were injected from a four-way multifunctional automatic valve (V-16A, Xintong Scientific Instrument, Academia Sinica) fitted with a by-pass coil (50 ml). Flow lines were PTFE tubing (0.8 mm id). Reagents All reagents were of analytical-reagent grade and ion-exchanged, distilled water was used throughout.A sodium dodecyl phenylsulfonate (SDPS) solution (0.05 mol l21) and a sodium dodecyl sulfonate (SDS) solution (0.05 Analyst, February 1997, Vol. 122 (111–113) 111mol l21) were prepared in water and from these, dilute carrier solutions were prepared. A dilute hydrochloric acid solution (0.20 mol l21) was prepared to adjust the pH of carrier solutions, checked by a pH meter. Stock solution of lignocaine hydrochloride (10 mg ml21) was prepared in water and stored frozen in a dark bottle.From these, working solutions were prepared as required. General Procedure Volumes of 50 ml of sample and calibration standard solutions were injected into the flow when the base resonant frequency (Fb) was stable, and the peak heights were found to be fp = Fp 2 Fb, where Fp is the maximum frequency during the run. Under the conditions that the flow rate was 2.0 ml min21 and the length of the mixing coil was 300 mm, the recorded peaks were sharp and the baselines were stable at an injection rate of 120 h21.Results and Discussion Selection of Ion-pairing Reagent Fig.1 illustrates the response of frequency versus time to injection of lignocaine hydrochloride into the flowing solution of SDPS and SDS. Sharp peaks occur where the upward slope corresponds to the decrease in adsorption primarily caused by the interfacial ion association between adsorbed surfactants and injected lignocaine hydrochloride. It can be seen that the peak using SDPS as the carrier is higher and sharper than that using SDS as the carrier. The reason is that SDPS is more favourable than SDS towards adsorption on the electrode and the constant b, for SDPS is larger than that for SDS.Hence SDPS was used as the ion-pairing reagent in the investigation. Selection of Operating Parameters The operating parameters studied were flow rate and length of the mixing coil. Increases in flow rate and coil length result in decreases in the reaction times, t and T, and distribution degree, D.The effects of t, T, and D on the response can be derived from eqn. (1), namely that the peak height decreases with increases in T and D and decrease in t. The concentrations used in these experiments were SDPS solution used as the carrier 5 mmol l21 and lige hydrochloride solution 1.00 mg ml21. The peak heights were found to increase as the flow rate increased in the range 0.5–1.5 ml min21, which indicates that the effects of decreases of the reaction time, T and distribution degree, D, are greater, and decrease when the flow rate was higher than 1.5 ml min21, which indicates that the effect of the decrease in the reaction time, t, is predominant.In order to increase the throughput, a flow rate of 2.0 ml min21 was adopted as a compromise. With the increase in the length of the mixing coil, the peak heights and widths were found to decrease and increase, respectively. The optimum length of the mixing coil was found to be 300 mm (total volume 0.2 ml) to give the sharpest peak and most stable baseline.Optimization of Reagent Concentration and pH The effect of the concentration of SDPS was studied in the range 0.5–10 mmol l21. Below 5 mmol l21, the peak height was found to be greater and became gradually unchangeable at higher SDPS concentration. Conversely, above 5 mmol l21, the peak height was smaller and decreased to zero. It can be seen from eqn.(1) that at the maximum peak height the concentration of SDPS is equal to D/2k2T(D-1) , which is inversely proportional to the reaction rate and time in the mixing coil. In subsequent work, 5 mmol l21 SDPS solution was used as the carrier. Fig. 2 illustrates the effect of pH on the peak height. As the carrier pH was increased in the range 1–3, the peak height increased significantly. In contrast, as the carrier pH increased in the range 3–7, the peak height decreased and gradually became unchangeable.Because of the protonation of SDPS, the concentration of free SDPS which participates in adsorption and ion association increases as the carrier pH increases. Hence the effect of the carrier pH on the response is similar to that of SDPS concentration as described above. The carrier pH was adjusted to 3 with hydrochloric acid since the dilute SDPS solution was about neutral. Calibration Graph and Reproducibility Using the flow injection system and the conditions SDPS carrier concentration 5 mmol l21, flow rate 2.0 ml min21 and length of the mixing coil 300 mm, a linear calibration graph of the relative peak height with respect to the blank (water) versus the concentration of lignocaine hydrochloride was obtained in the range 0.01– 2.00 mg ml21, with a slope of 800 Hz ml mg21 and a correlation coefficient of 0.9995.The standard deviation of the responses for the blank injection (10 replicates) was 2.55 Hz. The detection limit (3 3 noise), calculated as 3 3 2.55/800, was 8 mg ml21.The RSD 1.00 mg ml21 lignocaine hydrochloride (10 replicates) was 0.29%. The sampling rate was about 120 h21. Interferences The influence of foreign compounds that commonly accompany lignocaine hydrochloride in pharmaceutical preparations was Fig. 1 FIA response of resonant frequency versus time to lignocaine hydrochloride injection using SDPS and SDS as carrier. Concentration of lignocaine hydrochloride, 2.0 mg ml21; concentration of the carrier (pH 3), 5 mmol l21; flow rate, 2.0 ml min21; and length of mixing coil, 300 mm.Fig. 2 Effect of carrier pH on frequency response of ion-pairing flow injection to lignocaine hydrochloride. Concentration of lignocaine hydrochloride, 2.0 mg ml21; concentration of the carrier (SDPS), 5 mmol l21; flow- rate, 2.0 ml min21; and length of mixing coil, 300 mm. 112 Analyst, February 1997, Vol. 122studied by preparing solutions containing 1.00 mg ml21 lignocaine hydrochloride and increasing concentrations of the potential interferents up to 50 mg ml21 or by adding an amount to give an error of ±3%.The errors were determined by comparison with the peak heights given by a solution of analyte containing no foreign substances. Glucose, sucrose, lactose, asparate, citrate and tartrate were tolerated in large amounts (50 mg ml21 was the maximum tested) and 10 mg ml21 of antipyrine, atropine, berberine, sparteine and pilocarpine hydrochloride were also tolerated in the determination of lignocaine hydrochloride.Analysis of Pharmaceutical Preparations The proposed method was applied satisfactorily to the determination of lignocaine hydrochloride in pharmaceutical preparations. Commercially available formulations were analysed and the results obtained are summarized in Table 1. As can be seen, for all formulations the assay results were in good agreement with values for the nominal contents and with those obtained using a non-aqueous titrimetric method.16 The recoveries obtained by adding lignocaine hydrochloride to each pharmaceutical formulation are given in Table 2; they were determined by subtracting the results obtained for the pharmaceutical formulations prepared in a similar manner but to which no lignocaine hydrochloride had been added. Conclusions The proposed method for the determination of lignocaine hydrochloride, based on ion-pairing flow injection with piezoelectric detection, represents an important change to conventional approaches utilizing ion pairing. Usually, ion-pairing alone cannot provide a sufficient change in measurable properties for analysis, leading to the requirement for a separation (an extraction step) between free ion-pairing reagent and ion pair.The proposed method eliminates the separation, the ion-pairing reagent being fixed on a mass-sensing surface. It has distinct features. First, there is greater hydrodynamic and instrumental simplicity resulting from the elimination of the organic phase and segmenter.Second, the sampling frequency, accuracy and precision are superior to those for conventional methods for the analysis of pharmaceutical preparations. Although the sensitivity and selectivity may be relatively poor, they are sufficient for primary component analysis in pharmaceuticals. Improvements may be realized through optimizations of the ion-pairing reagent, operating parameters and detector and the signal processing.References 1 Murakita, H., Hayashi, M., Mikami, H., and Ishida, Y., Bunseki Kagaku, 1986, 35, 236. 2 Zhou, J.-X., Zhang, L.-J., and Wang, E.-K., Electroanalysis (N. Y.), 1993, 5, 295. 3 Korany, M. A., Wahbi, A. M., Elsayed, M. A., and Mandour, S., Anal. Lett., 1984, 17, 1373. 4 Xia, G., and Lui, S., Yaoxue Tongbao, 1986, 21, 25. 5 Huang, T., Pan, D., and Gao, H., Yaowu Fenxi Zazhi, 1984, 4, 21. 6 Shoukry, A. F., Issa, Y. M., El-Sheik, R., and Zareh, M., Microchem. J., 1988, 37, 299. 7 Bouklouze, A. A., El-Jammal, A., Patriarche, G. J., and Christian, G. D., J. Pharm. Biomed. Anal., 1991, 9, 393. 8 Manzo, R. H., Luna, E., and Allemandi, D. A., J. Pharm. Sci., 1991, 80, 80. 9 Karlberg, B., and Thelander, S., Anal. Chim. Acta, 1978, 98, 1. 10 Calatayud, J. M., Sampedro, A. S., and Sarrion, S. N., Analyst, 1990, 115, 855. 11 Sakai, T., Ohta, H., Ohno, N., and Sasaki, H., Fresenius’ J. Anal. Chem., 1994, 349, 475. 12 Buttry, D. A., and Ward, M. D., Chem. Rev., 1992, 92, 1355. 13 Yao, S.-Z., and Nie, L.-H., Chim. J. Anal. Chem., 1996, 24, 234. 14 Mo, Z.-H., Zhang, M.-J., and Xia, Z.-L., Anal. Lett., 1997, 30. 15 Yao, S.-Z., and Mo, Z.-H., Anal. Chim. Acta, 1987, 193, 97. 16 Chinese Pharmacopoeia, Part II, Chemical Industry Press and People’s Press of Hygiene, Beijing, 1995, Appendix 4, p. 402. Paper 6/05796I Received August 20, 1996 Accepted November 4, 1996 Table 1 Determination of lignocaine hydrochloride in pharmaceutical preparations Nominal, value/ Found by titrimetric Found by FI method†/ Sample* mg ml21 method†/mg ml21 mg ml21 1 0.5 0.49 ± 0.004 0.49 ± 0.002 2 1.0 1.01 ± 0.007 0.99 ± 0.003 3‡ 10 9.94 ± 0.10 10.01 ± 0.03 * Supplied by three manufacturers, 1 and 2, injection; 3, capsule. † Average of three determinations ± standard deviation. ‡ Prepared by dissolving a 250 mg capsule in 25 ml of water. Table 2 Recovery of lignocaine hydrochloride from pharmaceutical preparations Sample* Added/mg ml21 Found†/mg ml21 Recovery (%) 1 1.00 0.99 99.0 5.00 5.01 100.2 2 1.00 1.01 101.0 5.00 5.04 100.8 3 1.00 1.03 103.0 5.00 5.10 102.0 * See Table 1. † Average of three determinations. Analyst, February 1997, Vol. 122 113

ISSN:0003-2654    

DOI:10.1039/a605796i

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Flow injection Fluorimetric Determination of Ascorbic Acid Based on its Photooxidation by Thionine Blue Tom�as P�erez-Ruiz*, Carmen Mart�ýnez-Lozano, Virginia Tom�as and Ciriaco Sidrach Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, 30071 Murcia, Spain The photooxidation of ascorbic acid sensitized by Thionine Blue was studied. The Leucothionine Blue formed during the reaction is highly fluorescent. A flow injection method using merging zones is proposed for the determination of ascorbic acid over a concentration range from 8 3 1027 to 5 3 1025 mol l21 with a throughput of 80 samples per h.The method was used for the simple and rapid determination of ascorbic acid in pharmaceuticals, fruit juices and soft drinks. Keywords: Flow injection; fluorescence; photooxidation; ascorbic acid; Thionine Blue; pharmaceuticals; fruit juices Photochemical analysis is increasingly used in various fields because of its high sensitivity and selectivity.The combined used of photochemical reactions and other techniques, such as kinetic methods, gravimetry, generation of titrans, etc., has been applied to the determination of a large number of species.1–3 Photochemical reactions have also played an active role in postcolumn reactions in high-performance liquid chromatography and in flow injection (FI) analysis because their characteristics are well suited to reactor technology.4 The combination of FI and photochemical reactions has proved to be a sensitive means for the determination of many analytes.The leuco-forms of Thionine Blue and Thionine generated through the photochemical reaction between these dyes and EDTA have been applied as reductive reagents for spectrophotometrically determining various oxidants using FI systems.5–6 The amperometric determination of oxalate by photochemical decomposition of ferrioxalato in the reaction coil or in the flow cell of the flow system has been reported.7,8 Photochemical reactions coupled to FI systems have also been described for the individual9–12 and simultaneous13 determination of phenothiazine compounds, nitrite,14 hydrogen peroxide, 15 organoarsenicals,16 catecholamines,17 sulfonamides,18 nitrate,19,20 B2-vitamers,21 fluoride and phosphate22 and dissolved organic phosphorus,23 dissolved organic carbon,24 and total dissolved nitrogen25 in natural waters.Vitamin C, a water soluble vitamin, is an important micronutrient and plays many physiological roles26.Fruit and vegetables constitute the principal source of this vitamin in most human diets, where it occurs as l-ascorbic acid (AA) and its oxidized form, dehydro-l-ascorbic acid (DHAA), both of which are biologically active. Since the introduction of FI analysis, many flow systems have been developed for the determination of AA using several detection techniques, e.g., amperometric,27 coulometric,28 voltammetric,29 potentiometric,30 spectrophotometric31 –34 and spectrofluorimetric.35 An alternative for FI determination of AA in complex matrices is to use an enzyme column of ascorbate oxidase to catalyse its oxidation to DHAA;36,37 two measurements are made for each determination, one using the enzyme column and the other a blank column.The decrease in the signal compared with the blank is related to the amount of AA present. AA has been determined by photobleaching Methylene Blue using continuous38 and FI systems,39,40 both procedures involving spectrophotometric detection.The chemiluminometric determination of AA based on its photooxidation sensitized by Toluidine Blue has also been described.41 The aim of the work presented here has been to study the photooxidation of AA by Thionine Blue, in order to develop a simple, sensitive, selective and rapid fluorimetric method for determining AA. The product of this photochemical reaction, Leucothionine Blue, is a highly fluorescent species, which can be easily used for the sensitive determination of AA in a flowinjection system.This procedure is shown to be a good alternative to routine vitamin C analysis in pharmaceutical preparations, fruit juices and soft drinks. Experimental Reagents All solutions were prepared from analytical-reagent grade materials in doubly distilled water. Standard ascorbic acid solution, (1.00 mg ml21). This was prepared daily no more than 3 h prior to use by dissolving 100 mg of ascorbic acid (Merck) in 100 ml of 0.05 mol l21 perchloric acid.Working solutions of lower concentration were prepared by appropriate dilution with water. All solutions were kept in amber-coloured bottles in the dark. Stock solutions of the sensitizers, (1 3 1023 mol l21). Thionine Blue (CI 520245), Thionine (CI 52000), Toluidine Blue (CI 52040), Methylene Blue (CI 52015), Azur B (CI 52010), Azur C (CI 52002) and New Methylene Blue (CI 52030), were prepared by dissolving the appropriate amount of the product (Merck) in water.Solutions of lower concentration were prepared by dilution with water. Phosphate buffers. These were prepared from 0.2 mol ml21 potassium dihydrogenphosphate and sufficient 2 mol l21 potassium hydroxide or hydrochloric acid to give the desired pH. Buffers of lower capacity were prepared by appropriate dilution with water. Apparatus An SLM-Aminco Bowman (Urbana, IL, USA) Series 2 spectrofluorimeter was used for recording spectra and making fluorescence measurements.A Gilson (Villiers le Bell, France) Minipuls-4 peristaltic pump and Omnifit (Cambridge, UK) rotary valves were also used. The irradiation was performed with a tungsten halogen lamp (500 W, 250 V). Except for the pump tube (Tygon) PTFE tubing (0.5 mm id) was used throughout the manifold. A flow-cell Hellma (M�ullheim, Baden, Germany) 176.052 QS (inner volume 25 ml) was also used. The photoreactor was a 200 cm PTFE tubing (0.5 mm id) coiled around a glass tube of 0.5 cm diameter placed inside a Pyrex cyclinder with a double-walled well, through which cooling water continuously flowed at 2 l min21 or more.The lamp was located 20 cm from the reactor. This assembly was housed in a metal box to protect it from light other than that of the lamp. The inside of the box was covered with aluminium foil to permit maximum reflectance of the light from the lamp. Analyst, February 1997, Vol. 122 (115–118) 115Manifolds Three different flow injection configurations were tested for the determination of AA.Fig. 1 is a schematic diagram of the flow configurations employed. In configuration I the sample of ascorbic acid was injected into a buffered Thionine Blue stream. Configuration II is a reverse mode, the sample of AA and the buffer streams being mixed before the injection of the dye. Configuration III shows the merging zones manifold, where AA and Thionine Blue solution were injected simultaneously into two phosphate buffer (pH 3.0) streams with the aid of two rotary valves (with a 250 ml and 150 ml loop, respectively) and then synchronously merged before reaching the irradiated reactor.This stream is directed toward the fluorimetric detector. Configuration III was finally adopted as it provided the greatest sensitivity and the fastest restoration of the baseline. Results and Discussion The photochemical reaction between AA and thiazine dyes involves the formation of DHAA and the reduced form of the dyes, which are colourless and fluorescent.The reaction can be monitored either photometrically by measuring the decrease of the absorbance of the dye or fluorimetrically by measuring the increase of the fluorescence due to the reduced thiazine. This last path is the most sensitive. The thiazine dyes studied for the photooxidation of AA were Thionine, Azur B, Azur C, Methylene Blue, Thionine Blue, Toluidine Blue and New Methylene Blue.Of the dyes tested, Thionine Blue showed the greatest reaction rate and its reduced form had the highest fluorescent quantum yield. The overall reaction between AA and Thionine Blue is: The fluorescence spectra of a solution containing AA (1.5 3 1025 mol l21) and Thionine Blue (5 3 1026 mol l21) illuminated with white light for 1 min and of a Thionine Blue solution (5 3 1026 mol l21) are shown in Fig. 2. As can be see, the measurement of fluoresce nm with excitation at 340 nm makes it possible to monitor the photochemical process with great sensitivity.Configuration Designs The FI configuration used for the determination of ascorbic acid was designed to provide different conditions for magnifying the fluorescence signal. Three different configurations were tested for this purpose, normal, reverse and merging zones (I, II and III of Fig. 1). The reverse mode yielded better results than the normal mode because it resulted in increased sensitivity and a greater decrease in the fluorescence signal of the baseline.Another advantage of the reverse mode was the non-adherence of the dye to the coil walls. This method can be used when abundant quantities of sample are available. However, the merging zones approach was selected because it works with very small quantities of sample and its sensitivity and background are as good as that of the reverse mode. Influence of Manifold Parameters The variables studied were volume injected, flow rate and length of photoreactor. The reagent concentrations used in these experiments were as follows: buffer line, 0.05 mol l21 phosphate buffer of pH 3; AA line, 5 3 1026 mol l21 and Thionine Blue line, 5 3 1025 mol l21.The volumes of AA and Thionine Blue solutions injected varied between 35 and 300 ml. The peak heights increased with increasing volumes up to 200 ml for AA and 140 ml for Thionine Blue, above which they remained virtually constant.A sample volume of 250 ml and a 150 ml volume of Thionine Blue solution were chosen for further experiments. At a constant radiation intensity, the illumination time had a decisive effect on the photochemical reaction and hence on the sensitivity attained. The residence time of the merging zones of AA and Thionine Blue in the photoreactor can be selected by controlling the flow-rate of the two carriers and/or the length of the reactor. An irradiation time of about 20 s was sufficient to achieve the total oxidation of AA in the flow system.This time was obtained using a reactor length of 200 cm and a flow rate of 0.6 ml min21 for each phosphate buffer carrier. Fig. 1 Flow injection manifolds tested for the determination of ascorbic acid. P = peristaltic pump; V = injection valve; R = photoreactor; D = fluorimeter; W = waste. Other details in text. Fig. 2 Excitation (1, 2) and emission (1A, 2A) spectra of Thionine Blue (1, 1A) and Thionine Blue plus ascorbic acid irradiated for 1 min (2, 2A).Concentrations: Thionine Blue 5 3 1026 mol l21 and ascorbic acid 1.5 3 1025 mol l21. 116 Analyst, February 1997, Vol. 122Influence of Reagent Concentration The effect of varying pH and concentration of Thionine Blue solution was tested in the optimized flow system. The rate of photooxidation of AA sensitized by Thionine Blue is very much pH dependent. The peak height was maximal and constant from pH 2.8 to 3.4, and decreased outside this range.Therefore, a 0.2 mol l21 phosphate buffer of pH 3.0 was used as carrier. The effect of Thionine Blue concentration on peak height was studied over the range 1 3 1025–1 3 1024 mol l21. The peak height increased with increasing concentration of the dye solution stream up to 6 3 1025 mol l21, but levelled off at higher concentrations. Enhancement of the fluorescence intensity was a result of an increase in the rate of photooxidation of AA with increasing Thionine Blue concentrations.However, it is worth noting that at dye concentrations higher than 5 3 1024 mol l21 , the filter effect on the radiation emitted was substantial. Therefore, a 8 3 1025 mol l21 Thionine Blue solution was selected. The effect of the oxygen was studied by saturating all solutions with nitrogen or oxygen. Using the photoreaction of air-saturated AA, Thionine Blue and buffer solutions as reference, there was a 28% gain in the amount of leucodye formed when nitrogen was used and a 16% decrease in the amount of leucothionine blue when oxygen saturated solutions were used.Therefore, all solutions should be oxygen-free for maximum sensitivity. However, for routine measurements, airsaturated solutions were satisfactory. The day- to-day reproducibility was very good because the temperature and humidity remain constant in the flow system. Determination of Ascorbic Acid Once chemical and instrumental variables had been selected the flow system was used for the determination of AA.A series of standard solutions of AA were injected into the manifold to test the linearity of the calibration graph. A linear relationship between AA concentration and fluorescence intensity was obtained over the range 8 3 1027–5 3 1025 mol l21 (0.14–8.8 mg ml21). Regression linear analysis of the linear portion of the calibration graph gave a standard deviation of 1.6% for the slope and a correlation coefficient of 0.9993.Interference Studies In order to assess the possible analytical applications of the photochemical fluorimetric method described above, the effect of concomitant species on the determination of AA in real samples was studied by analysing synthetic sample solutions containing 5 31026 mol l21 of AA and various excess amounts of the common excipients used in pharmaceutical preparations, food additives commonly found in fruit juices, and soft drinks and organic acids. A substance was considered not to interfere if the variation in the peak height of AA was less than 3%.The results are shown in Table 1. Analysis of Real Samples The FI method has great potential for the sensitive and rapid determination of AA in real samples. This was confirmed by the results obtained for the determination of AA in pharmaceutical formulations, fruit juices and soft drinks. Several pharmaceutical dosage forms containing AA either in tablet, capsule or sachet form were dissolved and appropriately diluted with water prior to injection into the manifold.The results obtained by the proposed method and the 2,6-DCPIP- (2,6-dichlorophenolindophenol)42 method are given in Table 2. Table 3 gives the AA content of two fruit juices and two soft drinks. The good recovery of AA shown in Table 3 indicates that no significant interference occurred by the additives contained in these samples. Conclusions The photooxidation of AA by Thionine Blue was shown to be suitable for AA determination.The photochemical process was readily automated in a FI system using fluorescent detection with good precision and high sample throughput. The reagent and instrumentation used are inexpensive and the method appears adequate for quality control analysis of AA in pharmaceuticals, fruit juices and soft drinks. Table 1 Tolerance to different species in the determination of ascorbic acid* Maximum tolerable Species added mol ratio Lactose, galactose, sucrose, fructose, citrate, tartrate, acetate, lactate 100† Thiamine, urea, glucose, saccharin, benzoic acid 50 Alanine 10 Acetylsalicilic acid 5 Cysteine, uric acid 1 * Ascorbic acid concentration 5 3 1026 mol l21.† Maximum ratio tested. Table 2 Determination of ascorbic acid in pharmaceutical preparations Ascorbic acid found/g Reference Proposed Preparation* Supplier method42 method† Boi-K B.O.I. 0.270 0.275 ± 0.002 (0.25 g per tablet) (Barcelona, Spain) Citrovit Abell�o 1.126 1.102 ± 0.004 (1 g per sachet) (Madrid, Spain) Vitaendil C.K.P.Wassermann 0.076 0.075 ± 0.002 (0.075 g per tablet) (Barcelona, Spain) Calcium-Sandoz + Sandoz Pharma 0.986 0.974 ± 0.001 vitamin C (Barcelona, Spain) (1 g per tablet) Multibionta granular Merck 0.026 0.025 ± 0.002 (0.025 g per pill) (Barcelona, Spain) Polybion C granular Merck 0.064 0.065 ± 0.002 (0.060 g per pill) * Stated amount of ascorbic acid in the product. † Mean ± s for four determinations.Table 3 Determination of ascorbic acid in natural juice and soft drink Amount of ascorbic acid/ mg per 100 ml Reference Proposed Sample method* method† Natural orange juice 87.9 88.6 ± 0.3 Natural lemon juice 58.4 59.5 ± 0.2 Soft drink 1 10.2 9.9 ± 0.1 Soft drink 2 6.1 6.4 ± 0.2 * Average of two determinations. † Mean ± s for four replicates. Analyst, February 1997, Vol. 2 117The authors express their gratitude to the Direcci�on General de Investigaci�on Cient�ýfica y T�ecnica (Project PB93-1139) and Comunidad Aut�onoma de Murcia (Project PCT 95/15).References 1 Fitzgerald, J. M., Analytical Photochemistry and Photochemical Analysis, Marcel Dekker, New York, 1971, pp. 128–141 and 157–163. 2 P�eter, A., and Cz�anyi, J., Acta Phys. Chem., 1975, 21, 37. 3 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., and Tom�as, V., Quim. Anal., 1987, 6, 119. 4 Birks, J. W., Chemiluminescence and Photochemical Reaction Detection in Chromatography, VCH, New York, 1989, p. 151. 5 Mart�ýnez-Lozano, C., P�erez-Ruiz, T., Tom�as, V., and Yag�ue, E., Analyst, 1988, 113, 1057. 6 M�uller, H., and Hansen, E. H., Chem. Tech. (Leipzig), 1992, 42, 304. 7 Le�on, L. E., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1990, 234, 227. 8 Le�on, L. E., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst 1990, 115, 1549. 9 Chen, D., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst, 1991, 116, 171. 10 Tena, M. T., Luque de Castro, M.D., and Valc�arcel, M., J. Autom. Chem., 1991, 13, 111. 11 Mart�ýnez Calatayud, J., and G�omez, C., Anal. Chim. Acta, 1992, 256, 105. 12 Loassis, B., Aaron, J. J., and Mahedero, M. C., Talanta, 1994, 41, 1985. 13 Chen, D., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Talanta, 1991, 38, 1227. 14 Liu, R.-M., and Liu, D.-J., Analyst, 1991, 116, 497. 15 Genfa, Z., Dasgupta, P., Edgemond, W. S., and Marz, J. N., Anal. Chim. Acta, 1991, 243, 207. 16 Atalian, R.H., and Kalman, D. W., Talanta, 1991, 38, 167. 17 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Val, O., Talanta, 1993, 40, 1625. 18 Mahedero, M. C., and Aaron, J. J., Anal. Chim. Acta, 1992, 269, 193. 19 Liu, R., Liu, D., Sun, A., and Liu, G., Talanta, 1995, 42, 437. 20 Motomizu, S., and Sanada, M., Anal. Chim. Acta, 1995, 308, 406. 21 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Val, O., Analyst, 1994, 119, 1199. 22 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Sanz, A., Analyst, 1996, 121, 477. 23 McKelvie, I. D., Mitri, M., Hart, B. T., Hamilton, I. C., and Stuart, A. D., Anal. Chim. Acta, 1994, 293, 155. 24 Edwards, R. T., McKelvie, I. D., Ferret, P., Hart, B. T., Bapat, J. B., and Koshy, K., Anal. Chim. Acta, 1992, 261, 287. 25 McKelvie, I. D., Mitri, M., Hart, B. T., Hamilton, I. C., and Stuart, A. D., Anal. Chim. Acta, 1994, 293, 155. 26 McCormick, D. B., in: Textbook of Clinical Chemistry, ed. Tietz, N. W., Sanders, Philadephia, PA, 1986, pp. 959–962. 27 Fogg, A. G., Summan, A. M., and Fern�andez-Arciniega, M. A., Analyst, 1985, 111, 341. 28 Curran, D. J., and Tongas, P. T., Anal. Chem., 1984, 56, 672. 29 Fung, Y., and Mo, S., Anal. Chim. Acta, 1992, 261, 375. 30 Karlberg, B., and Thelander, S., Analyst, 1978, 103, 1154. 31 L�azaro, F., R�ýos, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst, 1986, 111, 163 and 167. 32 Sultan, S. M., Abdennabi, A. M., and Suliman, F. E. O., Talanta, 1994, 41, 125. 33 Albero, I., Garc�ýa, S., S�anchez-Pedre�no, C., and Rodr�ýguez, J., Analyst, 1992, 117, 1635. 34 Almuaibed, A. M., and Townshend, A., Talanta, 1992, 39, 1459. 35 Vanderslice, J., and Higgs, D., Micronutr. Anal., 1989, 6, 109. 36 Greenway, G. M., and Ongomo, P., Analyst, 1990, 115, 1297. 37 Bradberry, C. W., and Adams, R. N., Anal. Chem., 1983, 55, 2439. 38 White, V. R., and Fitzgerald, J. M., Anal. Chem., 1975, 47, 903. 39 Sanz-Mart�ýnez, A., R�ýos, A., and Valc�arcel, M., Analyst, 1992, 117, 1761. 40 Le�on, E. L., and Catapano, J., Anal. Lett., 1993, 26, 1741. 41 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., and Sanz, A., Anal. Chim. Acta, 1995, 308, 299. 42 Davies, S. M. R., and Masten, S. J., Anal. Chim. Acta, 1991, 248, 225. Paper 6/06841C Received October 7, 1996 Accepted November 19, 1996 118 Analyst, February 1997, Vol. 122 Flow injection Fluorimetric Determination of Ascorbic Acid Based on its Photooxidation by Thionine Blue Tom�as P�erez-Ruiz*, Carmen Mart�ýnez-Lozano, Virginia Tom�as and Ciriaco Sidrach Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, 30071 Murcia, Spain The photooxidation of ascorbic acid sensitized by Thionine Blue was studied.The Leucothionine Blue formed during the reaction is highly fluorescent. A flow injection method using merging zones is proposed for the determination of ascorbic acid over a concentration range from 8 3 1027 to 5 3 1025 mol l21 with a throughput of 80 samples per h.The method was used for the simple and rapid determination of ascorbic acid in pharmaceuticals, fruit juices and soft drinks. Keywords: Flow injection; fluorescence; photooxidation; ascorbic acid; Thionine Blue; pharmaceuticals; fruit juices Photochemical analysis is increasingly used in various fields because of its high sensitivity and selectivity. The combined used of photochemical reactions and other techniques, such as kinetic methods, gravimetry, generation of titrans, etc., has been applied to the determination of a large number of species.1–3 Photochemical reactions have also played an active role in postcolumn reactions in high-performance liquid chromatography and in flow injection (FI) analysis because their characteristics are well suited to reactor technology.4 The combination of FI and photochemical reactions has proved to be a sensitive means for the determination of many analytes.The leuco-forms of Thionine Blue and Thionine generated through the photochemical reaction between these dyes and EDTA have been applied as reductive reagents for spectrophotometrically determining various oxidants using FI systems.5–6 The amperometric determination of oxalate by photochemical decomposition of ferrioxalato in the reaction coil or in the flow cell of the flow system has been reported.7,8 Photochemical reactions coupled to FI systems have also been described for the individual9–12 and simultaneous13 determination of phenothiazine compounds, nitrite,14 hydrogen peroxide, 15 organoarsenicals,16 catecholamines,17 sulfonamides,18 nitrate,19,20 B2-vitamers,21 fluoride and phosphate22 and dissolved organic phosphorus,23 dissolved organic carbon,24 and total dissolved nitrogen25 in natural waters.Vitamin C, a water soluble vitamin, is an important micronutrient and plays many physiological roles26. Fruit and vegetables constitute the principal source of this vitamin in most human diets, where it occurs as l-ascorbic acid (AA) and its oxidized form, dehydro-l-ascorbic acid (DHAA), both of which are biologically active.Since the introduction of FI analysis, many flow systems have been developed for the determination of AA using several detection techniques, e.g., amperometric,27 coulometric,28 voltammetric,29 potentiometric,30 spectrophotometric31 –34 and spectrofluorimetric.35 An alternative for FI determination of AA in complex matrices is to use an enzyme column of ascorbate oxidase to catalyse its oxidation to DHAA;36,37 two measurements are made for each determination, one using the enzyme column and the other a blank column. The decrease in the signal compared with the blank is related to the amount of AA present.AA has been determined by photobleaching Methylene Blue using continuous38 and FI systems,39,40 both procedures involving spectrophotometric detection. The chemiluminometric determination of AA based on its photooxidation sensitized by Toluidine Blue has also been described.41 The aim of the work presented here has been to study the photooxidation of AA by Thionine Blue, in order to develop a simple, sensitive, selective and rapid fluorimetric method for determining AA.The product of this photochemical reaction, Leucothionine Blue, is a highly fluorescent species, which can be easily used for the sensitive determination of AA in a flowinjection system.This procedure is shown to be a good alternative to routine vitamin C analysis in pharmaceutical preparations, fruit juices and soft drinks. Exilled water. Standard ascorbic acid solution, (1.00 mg ml21). This was prepared daily no more than 3 h prior to use by dissolving 100 mg of ascorbic acid (Merck) in 100 ml of 0.05 mol l21 perchloric acid. Working solutions of lower concentration were prepared by appropriate dilution with water.All solutions were kept in amber-coloured bottles in the dark. Stock solutions of the sensitizers, (1 3 1023 mol l21). Thionine Blue (CI 520245), Thionine (CI 52000), Toluidine Blue (CI 52040), Methylene Blue (CI 52015), Azur B (CI 52010), Azur C (CI 52002) and New Methylene Blue (CI 52030), were prepared by dissolving the appropriate amount of the product (Merck) in water. Solutions of lower concentration were prepared by dilution with water.Phosphate buffers. These were prepared from 0.2 mol ml21 potassium dihydrogenphosphate and sufficient 2 mol l21 potassium hydroxide or hydrochloric acid to give the desired pH. Buffers of lower capacity were prepared by appropriate dilution with water. Apparatus An SLM-Aminco Bowman (Urbana, IL, USA) Series 2 spectrofluorimeter was used for recording spectra and making fluorescence measurements. A Gilson (Villiers le Bell, France) Minipuls-4 peristaltic pump and Omnifit (Cambridge, UK) rotary valves were also used.The irradiation was performed with a tungsten halogen lamp (500 W, 250 V). Except for the pump tube (Tygon) PTFE tubing (0.5 mm id) was used throughout the manifold. A flow-cell Hellma (M�ullheim, Baden, Germany) 176.052 QS (inner volume 25 ml) was also used. The photoreactor was a 200 cm PTFE tubing (0.5 mm id) coiled around a glass tube of 0.5 cm diameter placed inside a Pyrex cyclinder with a double-walled well, through which cooling water continuously flowed at 2 l min21 or more.The lamp was located 20 cm from the reactor. This assembly was housed in a metal box to protect it from light other than that of the lamp. The inside of the box was covered with aluminium foil to permit maximum reflectance of the light from the lamp. Analyst, February 1997, Vol. 122 (115–118) 115Manifolds Three different flow injection configurations were tested for the determination of AA.Fig. 1 is a schematic diagram of the flow configurations employed. In configuration I the sample of ascorbic acid was injected into a buffered Thionine Blue stream. Configuration II is a reverse mode, the sample of AA and the buffer streams being mixed before the injection of the dye. Configuration III shows the merging zones manifold, where AA and Thionine Blue solution were injected simultaneously into two phosphate buffer (pH 3.0) streams with the aid of two rotary valves (with a 250 ml and 150 ml loop, respectively) and then synchronously merged before reaching the irradiated reactor.This stream is directed toward the fluorimetric detector. Configuration III was finally adopted as it provided the greatest sensitivity and the fastest restoration of the baseline. Results and Discussion The photochemical reaction between AA and thiazine dyes involves the formation of DHAA and the reduced form of the dyes, which are colourless and fluorescent. The reaction can be monitored either photometrically by measuring the decrease of the absorbance of the dye or fluorimetrically by measuring the increase of the fluorescence due to the reduced thiazine.This last path is the most sensitive. The thiazine dyes studied for the photooxidation of AA were Thionine, Azur B, Azur C, Methylene Blue, Thionine Blue, Toluidine Blue and New Methylene Blue. Of the dyes tested, Thionine Blue showed the greatest reaction rate and its reduced form had the highest fluorescent quantum yield.The overall reaction between AA and Thionine Blue is: The fluorescence spectra of a solution containing AA (1.5 3 1025 mol l21) and Thionine Blue (5 3 1026 mol l21) illuminated with white light for 1 min and of a Thionine Blue solution (5 3 1026 mol l21) are shown in Fig. 2. As can be see, the measurement of fluorescence at 464 nm with excitation at 340 nm makes it possible to monitor the photochemical process with great sensitivity.Configuration Designs The FI configuration used for the determination of ascorbic acid was designed to provide different conditions for magnifying the fluorescence signal. Three different configurations were tested for this purpose, normal, reverse and merging zones (I, II and III of Fig. 1). The reverse mode yielded better results than the normal mode because it resulted in increased sensitivity and a greater decrease in the fluorescence signal of the baseline. Another advantage of the reverse mode was the non-adherence of the dye to the coil walls.This method can be used when abundant quantities of sample are available. However, the merging zones approach was selected because it works with very small quantities of sample and its sensitivity and background are as good as that of the reverse mode. Influence of Manifold Parameters The variables studied were volume injected, flow rate and length of photoreactor. The reagent concentrations used in these experiments were as follows: buffer line, 0.05 mol l21 phosphate buffer of pH 3; AA line, 5 3 1026 mol l21 and Thionine Blue line, 5 3 1025 mol l21.The volumes of AA and Thionine Blue solutions injected varied between 35 and 300 ml. The peak heights increased with increasing volumes up to 200 ml for AA and 140 ml for Thionine Blue, above which they remained virtually constant. A sample volume of 250 ml and a 150 ml volume of Thionine Blue solution were chosen for further experiments.At a constant radiation intensity, the illumination time had a decisive effect on the photochemical reaction and hence on the sensitivity attained. The residence time of the merging zones of AA and Thionine Blue in the photoreactor can be selected by controlling the flow-rate of the two carriers and/or the length of the reactor. An irradiation time of about 20 s was sufficient to achieve the total oxidation of AA in the flow system. This time was obtained using a reactor length of 200 cm and a flow rate of 0.6 ml min21 for each phosphate buffer carrier. Fig. 1 Flow injection manifolds tested for the determination of ascorbic acid. P = peristaltic pump; V = injection valve; R = photoreactor; D = fluorimeter; W = waste. Other details in text. Fig. 2 Excitation (1, 2) and emission (1A, 2A) spectra of Thionine Blue (1, 1A) and Thionine Blue plus ascorbic acid irradiated for 1 min (2, 2A). Concentrations: Thionine Blue 5 3 1026 mol l21 and ascorbic acid 1.5 3 1025 mol l21. 116 Analyst, February 1997, Vol. 122Influence of Reagent Concentration The effect of varying pH and concentration of Thionine Blue solution was tested in the optimized flow system. The rate of photooxidation of AA sensitized by Thionine Blue is very much pH dependent. The peak height was maximal and constant from pH 2.8 to 3.4, and decreased outside this range. Therefore, a 0.2 mol l21 phosphate buffer of pH 3.0 was used as carrier. The effect of Thionine Blue concentration on peak height was studied over the range 1 3 1025–1 3 1024 mol l21.The peak height increased with increasing concentration of the dye solution stream up to 6 3 1025 mol l21, but levelled off at higher concentrations. Enhancement of the fluorescence intensity was a result of an increase in the rate of photooxidation of AA with increasing Thionine Blue concentrations. However, it is worth noting that at dye concentrations higher than 5 3 1024 mol l21 , the filter effect on the radiation emitted was substantial.Therefore, a 8 3 1025 mol l21 Thionine Blue solution was selected. The effect of the oxygen was studied by saturating all solutions with nitrogen or oxygen. Using the photoreaction of air-saturated AA, Thionine Blue and buffer solutions as reference, there was a 28% gain in the amount of leucodye formed when nitrogen was used and a 16% decrease in the amount of leucothionine blue when oxygen saturated solutions were used.Therefore, all solutions should be oxygen-free for maximum sensitivity. However, for routine measurements, airsaturated solutions were satisfactory. The day- to-day reproducibility was very good because the temperature and humidity remain constant in the flosystem. Determination of Ascorbic Acid Once chemical and instrumental variables had been selected the flow system was used for the determination of AA. A series of standard solutions of AA were injected into the manifold to test the linearity of the calibration graph.A linear relationship between AA concentration and fluorescence intensity was obtained over the range 8 3 1027–5 3 1025 mol l21 (0.14–8.8 mg ml21). Regression linear analysis of the linear portion of the calibration graph gave a standard deviation of 1.6% for the slope and a correlation coefficient of 0.9993. Interference Studies In order to assess the possible analytical applications of the photochemical fluorimetric method described above, the effect of concomitant species on the determination of AA in real samples was studied by analysing synthetic sample solutions containing 5 31026 mol l21 of AA and various excess amounts of the common excipients used in pharmaceutical preparations, food additives commonly found in fruit juices, and soft drinks and organic acids.A substance was considered not to interfere if the variation in the peak height of AA was less than 3%. The results are shown in Table 1.Analysis of Real Samples The FI method has great potential for the sensitive and rapid determination of AA in real samples. This was confirmed by the results obtained for the determination of AA in pharmaceutical formulations, fruit juices and soft drinks. Several pharmaceutical dosage forms containing AA either in tablet, capsule or sachet form were dissolved and appropriately diluted with water prior to injection into the manifold. The results obtained by the proposed method and the 2,6-DCPIP- (2,6-dichlorophenolindophenol)42 method are given in Table 2.Table 3 gives the AA content of two fruit juices and two soft drinks. The good recovery of AA shown in Table 3 indicates that no significant interference occurred by the additives contained in these samples. Conclusions The photooxidation of AA by Thionine Blue was shown to be suitable for AA determination. The photochemical process was readily automated in a FI system using fluorescent detection with good precision and high sample throughput.The reagent and instrumentation used are inexpensive and the method appears adequate for quality control analysis of AA in pharmaceuticals, fruit juices and soft drinks. Table 1 Tolerance to different species in the determination of ascorbic acid* Maximum tolerable Species added mol ratio Lactose, galactose, sucrose, fructose, citrate, tartrate, acetate, lactate 100† Thiamine, urea, glucose, saccharin, benzoic acid 50 Alanine 10 Acetylsalicilic acid 5 Cysteine, uric acid 1 * Ascorbic acid concentration 5 3 1026 mol l21. † Maximum ratio tested.Table 2 Determination of ascorbic acid in pharmaceutical preparations Ascorbic acid found/g Reference Proposed Preparation* Supplier method42 method† Boi-K B.O.I. 0.270 0.275 ± 0.002 (0.25 g per tablet) (Barcelona, Spain) Citrovit Abell�o 1.126 1.102 ± 0.004 (1 g per sachet) (Madrid, Spain) Vitaendil C.K.P. Wassermann 0.076 0.075 ± 0.002 (0.075 g per tablet) (Barcelona, Spain) Calcium-Sandoz + Sandoz Pharma 0.986 0.974 ± 0.001 vitamin C (Barcelona, Spain) (1 g per tablet) Multibionta granular Merck 0.026 0.025 ± 0.002 (0.025 g per pill) (Barcelona, Spain) Polybion C granular Merck 0.064 0.065 ± 0.002 (0.060 g per pill) * Stated amount of ascorbic acid in the product.† Mean ± s for four determinations. Table 3 Determination of ascorbic acid in natural juice and soft drink Amount of ascorbic acid/ mg per 100 ml Reference Proposed Sample method* method† Natural orange juice 87.9 88.6 ± 0.3 Natural lemon juice 58.4 59.5 ± 0.2 Soft drink 1 10.2 9.9 ± 0.1 Soft drink 2 6.1 6.4 ± 0.2 * Average of two determinations.† Mean ± s for four replicates. Analyst, February 1997, Vol. 122 117The authors express their gratitude to the Direcci�on General de Investigaci�on Cient�ýfica y T�ecnica (Project PB93-1139) and Comunidad Aut�onoma de Murcia (Project PCT 95/15). References 1 Fitzgerald, J.M., Analytical Photochemistry and Photochemical Analysis, Marcel Dekker, New York, 1971, pp. 128–141 and 157–163. 2 P�eter, A., and Cz�anyi, J., Acta Phys. Chem., 1975, 21, 37. 3 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., and Tom�as, V., Quim. Anal., 1987, 6, 119. 4 Birks, J. W., Chemiluminescence and Photochemical Reaction Detection in Chromatography, VCH, New York, 1989, p. 151. 5 Mart�ýnez-Lozano, C., P�erez-Ruiz, T., Tom�as, V., and Yag�ue, E., Analyst, 1988, 113, 1057. 6 M�uller, H., and Hansen, E.H., Chem. Tech. (Leipzig), 1992, 42, 304. 7 Le�on, L. E., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1990, 234, 227. 8 Le�on, L. E., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst 1990, 115, 1549. 9 Chen, D., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst, 1991, 116, 171. 10 Tena, M. T., Luque de Castro, M. D., and Valc�arcel, M., J. Autom. Chem., 1991, 13, 111. 11 Mart�ýnez Calatayud, J., and G�omez, C., Anal.Chim. Acta, 1992, 256, 105. 12 Loassis, B., Aaron, J. J., and Mahedero, M. C., Talanta, 1994, 41, 1985. 13 Chen, D., Rios, A., Luque de Castro, M. D., and Valc�arcel, M., Talanta, 1991, 38, 1227. 14 Liu, R.-M., and Liu, D.-J., Analyst, 1991, 116, 497. 15 Genfa, Z., Dasgupta, P., Edgemond, W. S., and Marz, J. N., Anal. Chim. Acta, 1991, 243, 207. 16 Atalian, R. H., and Kalman, D. W., Talanta, 1991, 38, 167. 17 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Val, O., Talanta, 1993, 40, 1625. 18 Mahedero, M. C., and Aaron, J. J., Anal. Chim. Acta, 1992, 269, 193. 19 Liu, R., Liu, D., Sun, A., and Liu, G., Talanta, 1995, 42, 437. 20 Motomizu, S., and Sanada, M., Anal. Chim. Acta, 1995, 308, 406. 21 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Val, O., Analyst, 1994, 119, 1199. 22 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., Tom�as, V., and Sanz, A., Analyst, 1996, 121, 477. 23 McKelvie, I. D., Mitri, M., Hart, B. T., Hamilton, I. C., and Stuart, A. D., Anal. Chim. Acta, 1994, 293, 155. 24 Edwards, R. T., McKelvie, I. D., Ferret, P., Hart, B. T., Bapat, J. B., and Koshy, K., Anal. Chim. Acta, 1992, 261, 287. 25 McKelvie, I. D., Mitri, M., Hart, B. T., Hamilton, I. C., and Stuart, A. D., Anal. Chim. Acta, 1994, 293, 155. 26 McCormick, D. B., in: Textbook of Clinical Chemistry, ed. Tietz, N. W., Sanders, Philadephia, PA, 1986, pp. 959–962. 27 Fogg, A. G., Summan, A. M., and Fern�andez-Arciniega, M. A., Analyst, 1985, 111, 341. 28 Curran, D. J., and Tongas, P. T., Anal. Chem., 1984, 56, 672. 29 Fung, Y., and Mo, S., Anal. Chim. Acta, 1992, 261, 375. 30 Karlberg, B., and Thelander, S., Analyst, 1978, 103, 1154. 31 L�azaro, F., R�ýos, A., Luque de Castro, M. D., and Valc�arcel, M., Analyst, 1986, 111, 163 and 167. 32 Sultan, S. M., Abdennabi, A. M., and Suliman, F. E. O., Talanta, 1994, 41, 125. 33 Albero, I., Garc�ýa, S., S�anchez-Pedre�no, C., and Rodr�ýguez, J., Analyst, 1992, 117, 1635. 34 Almuaibed, A. M., and Townshend, A., Talanta, 1992, 39, 1459. 35 Vanderslice, J., and Higgs, D., Micronutr. Anal., 1989, 6, 109. 36 Greenway, G. M., and Ongomo, P., Analyst, 1990, 115, 1297. 37 Bradberry, C. W., and Adams, R. N., Anal. Chem., 1983, 55, 2439. 38 White, V. R., and Fitzgerald, J. M., Anal. Chem., 1975, 47, 903. 39 Sanz-Mart�ýnez, A., R�ýos, A., and Valc�arcel, M., Analyst, 1992, 117, 1761. 40 Le�on, E. L., and Catapano, J., Anal. Lett., 1993, 26, 1741. 41 P�erez-Ruiz, T., Mart�ýnez-Lozano, C., and Sanz, A., Anal. Chim. Acta, 1995, 308, 299. 42 Davies, S. M. R., and Masten, S. J., Anal. Chim. Acta, 1991, 248, 225. Paper 6

ISSN:0003-2654    

DOI:10.1039/a606841c

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

On-line Flow Injection–Pervaporation of Beer Samples for the Determination of Diacetyl Jos�e M. Izquierdo-Ferrero, Juan M. Fern�andez-Romero and Mar�ýa D. Luque de Castro* Department of Analytical Chemistry, Faculty of Sciences, University of C�ordoba, E-14004 C�ordoba, Spain. E-mail: qa1lucam@uco.es A method for the determination of diacetyl based on the integration of a continuous pervaporation module and a flow injection manifold is presented. The chemical derivatization reaction involes condensation between diacetyl, a-naphthol and creatine in a basic medium to form a coloured compound, which is monitored photometrically at 530 nm.The method affords low detection and quantification limits (5 and 10 ng ml21 of diacetyl calculated as ±3s and ±10s, respectively, for n = 20), a wide linear range (10–2000 ng ml21, r2 = 0.9998, n = 7), good precision (RSD < 3%) and a sampling frequency of 8 h21. The method was applied to the determination of diacetyl in beer from bottled and medium-process samples and the results were in excellent agreement with those given by the conventional method.Keywords: Flow injection; continuous pervaporation; spectrophotometry; diacetyl; beer Diacetyl is an a-diketone responsible for the flavour of several foodstuffs (cheeses, yogurts, beers and wines) and contributes adversely to the taste quality. Diacetyl has been used for both testing food ripeness (e.g., in cheeses and wines) and monitoring of manufacture processes (e.g., in beers).1 Low levels of diacetyl and pentane-2,3-dione are desirable in beer fermentation, as these compounds confer an unpleasant taste similar to butter in the final product.This means that routine control is necessary during the fermentation of ethanol and other compounds in beer production. Beer with an acceptable taste requires the formation of esters and alcohols such as ethyl acetate, isoamyl acetate and isoamyl alcohol and the removal of undesirable compounds (such as a-diketones).Recently, research in this field has been focused on the development of new microorganisms which increase the metabolization of undesirable products during fermentation.2 The low concentration levels of diacetyl in beverages and the presence of potential interferents makes the development of analytical methods with higher sensitivity and selectivity essential. Previous methods for the determination of these species include a separation step, such as distillation, prior to the determination.Usually these methods have been implemented by using optical detection, such as photometry after condensation3,4 or chelating reactions5,6 and fluorimetry.7,8 The individual determination of diacetyl and pentane-2,3-dione has also been proposed using GC–FID,2,9 GC–ECD2,10,11 GC– MS12 or headspace GC.13 Several methods based on HPLC with continuous fluorimetric detection have also been reported.14,16 Recent developments in continuous flow analysis have led to the use of non-chromatographic continuous separation techniques for the determination of one or more analytes, the separation of components into groups and interference removal.Non-chromatographic continuous separation techniques can improve the sensitivity of the method through preconcentration and its selectivity by deleting matrix effects or particular interferents. Membrane-based non-chromatographic separation techniques are particularly useful for increasing selectivity.In some cases, they also allow analyte preconcentration, thereby resulting in improved sensitivity. Continuous gas-diffusion techniques, whether isolated or coupled to other separation techniques such as dialysis or ion exchange, make excellent tools for achieving high selectivity and sensitivity. However, these techniques suffer from serious drawbacks (e.g., clogging of the membrane pores by suspended particles or high molecular mass compounds occasionally present in the sample and deterioration of the membrane by contact with the sample).Both drawbacks can be overcome by using pervaporation.17 This technique has been employed for a long time in industry, in competition with other traditional processes (e.g., distillation, extraction and adsorption); by contrast, it has scarcely been used for analytical purposes. A laboratory-scale pervaporation unit was developed a few years ago in order to overcome problems arising from the use of biosensors in on-line fermentation monitoring.18,19 Several applications of laboratory pervaporation based on different approaches have been developed recently by Luque de Castro and co-workers.20–24 This paper reports the use of an automatic spectrophotometric method for the determination of diacetyl in beer samples based on the on-line coupling of a continuous pervaporation unit to a simple flow injection (FI) manifold.The chemical reaction is based on condensation of the analyte using an excess of anaphthol and creatine in a basic medium as described by Mattessich and Cooper,4 who improved a previous assay for the determination of this analyte.The condensation product is monitored spectrophotometrically at 530 nm. Experimental Instruments and Apparatus A Pye Unicam (Cambridge, UK) SP-500 spectrophotometer furnished with a Hellma (Jamaica, NY, USA) 178.011 QS flow cell and equipped with a Knauer (Bad Homburg, Germany) recorder was used.A laboratory-built pervaporation module equipped with a thermostated magnetic stirrer was also used. Two Gilson (Worthington, OH, USA) Minipuls-3 four-channel peristaltic pumps with rate selector, two Rheodyne (Cotati, CA, USA) Model 5041 injection valves and PTFE tubing of 0.5 mm id were also used. The pervaporation cell, designed in the laboratory was similar to others described by Mattos et al.20 Reagents and Solutions All reagents were of analytical-reagent grade.The donor stream was an aqueous solution with which the sample solution was mixed. A solution containing 125 mmol l21 a-naphthol (Sigma, St. Louis, MO, USA) and 100 mmol l21 creatine (Sigma) prepared in 750 mmol l21 sodium hydroxide (Merck, Darmstradt, Germany) was used as a reagent/acceptor solution. Aqueous diacetyl (Merck) solution was used as standard after suitable dilution. Beer samples provided by Compa�n�ýa Andaluza de Cervezas (C�ordoba, Spain) were used for testing the method.All solutions were prepared in doubly distilled water of Analyst, February 1997, Vol. 122 (119–122) 119high purity obtained from a Millipore (Bedford, MA, USA) Milli-Q Plus system. PTFE membranes of 5.0 mm pore size and 47 mm diameter purchased from Millipore were also used. Manual Procedure A 100 ml volume of beer is subjected to distillation to obtain 20 ml of distillate, of which 10 ml are mixed with 0.5 ml of 1 mg ml21 o-phenylenediamine and kept in the dark for 25 min.Then 2 ml of 4 m HCl are added and the absorbance of the mixture is monitored at 335 nm within 30 min after mixing. A blank containing 10 ml of deionized water instead of distillate is also measured photometrically. Manifold and Procedure Fig. 1 depicts the hydrodynamic system used. It consists of a flow injection manifold divided into two parts. The lower part acts as donor submanifold, which consists of a monochannel system provided with a main injection valve (MIV) that injects the samples into an aqueous stream which circulates through the lower part of the pervaporation cell, which is thermostated at 90 °C.The diacetyl evaporates to the air gap between the donor solution and the membrane and diffuses through the hydrophobic membrane to the upper part of the pervaporation cell where it is collected. The upper submanifold includes the acceptor chamber of the pervaporation cell, which is located in the loop of an auxiliary injection valve (AIV).This valve is filled with the acceptor solution, which contains the reagent mixture (AIV in injection position) and then is switched to the filling position, thus keeping static the loop contents during a preset interval. The analyte which passes through the membrane is collected in the acceptor/reagent solution, with simultaneous development of the condensation reaction. After a presetime (Tp), the AIV is switched again to the inject position and the reaction product is driven to the detector.The peak height (transient signal obtained when the plug containing the coloured product reaches the detector) is proportional to the concentration of diacetyl in the sample. Results and Discussion Optimization of Variables The optimization of the variables involved in the overall process, grouped into physical, chemical and hydrodynamic, was performed using the univariate method. Table 1 shows the ranges over which each variable was studied and the optimum value found.Physical variables Fig. 2(a) shows the influence of temperature in the pervaporation step, which was studied between 30 and 110 °C (boiling point of diacetyl 88 °C). A temperature of 90 °C was sufficient for appropriate transfer of the analyte through the membrane. This temperature was also suitable for the formation of the condensation product. Fig. 1 Manifold for the automatic–photometric determination of diacetyl in beer samples.P, Peristaltic pump; MIV, main injection valve; IV, injection volume; AIV, auxiliary injection valve; PM, pervaporation module; m, hydrophobic membrane; TMS, thermostatic magnetic stirrer; D, detector; w, waste; R, acceptor/reagent solution; C, donor solution; and S, sample. Table 1 Optimization of variables Range Optimum Type Variable studied value Physical Temperature/°C 30–110 90 FI Donor flow rate/ml min21 0.1–0.5 0.3 Acceptor flow rate/ml min21 0.5–2.0 1.8 Volume injected/ml 100–3000 2000 Pervaporation time/s 30–480 360 Chemical Sodium hydroxide concentration/mmol l21 100–1000 750 pH — 12 a-Naphthol concentration/mmol l21 5–200 125 Creatine concentration/mmol l21 5–150 100 Fig. 2 Influence of temperature (a) on the pervaporation process and (b) on the derivatization reaction for concentrations of diacetyl of 20 and 100 ng ml21 (dashed and continuous lines, respectively). 120 Analyst, February 1997, Vol. 122Chemical variables The influence of the pH in both the donor and acceptor solutions was studied.An aqueous solution adjusted to different pH values was used as a donor stream. A higher pervaporation efficiency was achieved by using doubly distilled water as the donor stream. However, the pervaporation efficiency was increased when the acceptor solution contained sodium hydroxide. This was tested by preparing different aqueous solutions containing sodium hydroxide at concentrations between 100 and 1000 mmol l21.A concentration of 750 mmol l21 provided the best medium for collection of the analyte. This solution was also appropriate for the development of the derivatization reaction. Concentrations of 125 mmol l21 a-naphthol and 100 mmol l21 creatine in the acceptor/reagent solution were selected as optimum. Higher concentrations of these reagents caused higher blank signals, resulting in appreciable lack of sensitivity (estimated as 50%). On the other hand, lower concentrations resulted in poor development of the chromogenic reaction.Hydrodynamic variables The flow rate had a dramatic influence on the performance of the system. As the flow rate of the donor solution determined the time during which the analyte was in the pervaporation module, 0.3 ml min21 was selected as optimum for the donor stream (lower values yielded non-reproducible results). On the other hand, the flow rate had no influence on the upper FI subsystem as the collection process occurred under static conditions.When the flow was re-started the reaction plug was driven to the detector. A flow rate of 1.8 ml min21 was appropriate to rinse the flow manifold. In order to obtain the best analytical signal with an acceptable sampling frequency, different stop-times in the range between 30 and 480 s were tried. A pervaporation time of 360 s was chosen as a compromise between the best signal and an acceptable sampling frequency. Injection volumes over 2 ml did not increase the analytical signal significantly. Features of the Method Ten standard solutions of diacetyl with concentrations between 1 and 5000 ng ml21 were prepared and injected in triplicate into the FI–pervaporation assembly using the optimum values of variables found previously.Table 2 summarizes the features of the method (equation, regression coefficient, linear range, detection and quantification limits and RSD). The linear range achieved for the method was suitable for applying it to the determination of diacetyl in beer samples.The sampling frequency achieved under the optimum working conditions was 8 h21. Study of Interferents The study of potential interferents was aimed at those commonly present in beer samples which are structurally similar to diacetyl, such as pyruvic acid, butylene glycol, acetoin (a precursor of diacetyl) and ascorbic acid. All interferents were added to the sample at concentrations higher than usually found in beers.The results showed that pyruvic acid, butylene glycol and ascorbic acid are tolerated at levels up to ten times their content in beer (with a tolerated interferent to analyte ratio of 100 : 1 for pyruvic acid and butylene glycol and 10 : 1 for ascorbic acid). Acetoin interferred at five times its concentration in these samples; nevertheless, acetoin is easily oxidized to diacetyl under atmospheric conditions. Application to the Proposed Method The method was applied to the determination of diacetyl in bottled and medium-process beer.In order to eliminate the gas contained in the original beverage (mainly CO2), all the samples were previously ultrasonicated for 5 min, then injected directly in triplicate. Table 3 summarizes the concentrations found and the results obtained after addition of two amounts of the analyte (50 and 100 ng ml21) to aliquots of the samples. The results provided by the proposed method were compared with those obtained by the conventional method used by Compa�n�ýa Andaluza de Cervezas. This method is based on previous separation of the analyte by distillation (vapour stream) followed by condensation of the analyte with ophenylenediamine, and photometric monitoring of the product at 335 nm.Six beer samples were used for this comparison and the results obtained produced a straight-line graph: y = 0.934x + 1.477 (r2 = 0.9901, n = 6) where y represents the FI–pervaporation method and x the conventional method.As can be seen, an excellent correlation exists between the proposed and the conventional methods. Conclusions The method proposed for the determination of diacetyl in beer samples shows the following features: simplicity in implementing the experimental set-up (i.e., a conventional FI manifold in Table 2 Features of the method Equation* A = 0.0199C + 0.0005 Regression coefficient 0.9998 (n = 7) Linear range/ng ml21 10–2000 Detection limit/ng ml21 5 Quantification limit/ng ml21 10 RSD (%)† — Low level 3.0 High level 2.6 * A denotes absorbance and C analyte concentration in ng ml21.† For 50 and 1000 ng ml21 of diacetyl. Table 3 Application of the method Concentration/ng ml21 Recovery (%)† Sample Conventional Proposed Difference No. Type* method method (%) Addition 1 Addition 2 1 A 18 17 25.5 100 96 2 A 15 16 6.5 94 98 3 A 25 24 24.0 98 97 4 B 30 29 23.3 100 99 5 B 25 27 8.0 107 99 6 B 52 50 23.8 100 100 * A and B denote bottled and medium-process beers, respectively.† 50 and 100 ng ml21 for additions 1 and 2, respectively. Analyst, February 1997, Vol. 122 121which an easy laboratory-built pervaporation unit is included); high selectivity as a result of the nature of the chemical reaction; lower sample consumption (2 ml); and an acceptable sensitivity. These features make it suitable for the routine determination of diacetyl in beer samples, as was demonstrated by comparison with the conventional method. The Comisi�on Interministerial de Ciencia y Tecnolog�ýa (CICYT) is thanked for financial support (project No.PB93- 0827). We gratefully acknowledge Carlos M. G�omez S�anchez of Compa�n�ýa Andaluza de Cervezas for the samples compared with the conventional method. References 1 West, h, A. L., and Becker, K., Am. Soc. Brew. Proc., 1952, 81, 65. 2 Mathis, C., Pons, M. N., Engasser, J. M., and Lenoel, M., Anal. Chim. Acta, 1993, 279, 59. 3 Westerfeld, W. W., J. Biol. Chem., 1945, 161, 495. 4 Mattessich, J., and Cooper, J. R., Anal. Biochem., 1989, 180, 349. 5 Ribereau-Gayon, J., Peynaud, E., Sudraud, P., and Riberau-Gayon, P., Sciences et Techniques de Vin, Dunod, Paris, 1972. 6 Walsh, B., and Cogan, T. M., J. Dairy Res., 1988, 41, 31. 7 Garc�ýa-Vilanova, R. J., and Garc�ýa Estepa, R. M., Talanta, 1993, 40, 1419. 8 Mariaud, M., and Levillain, P., Talanta, 1994, 41, 75. 10 Barbieri, G., Bolzoni, L., Careri, M., Mangia, A., Paroladi, G., Spagnoli, S., and Virgili, R., J.Agric. Food. Chem., 1994, 42, 1170. 11 Otsuka, M., and Ohmori, S. J., J. Chromatogr. B, 1992, 115, 215. 12 Otsuka, M., and Ohmori, S. J., J. Chromatogr. B, 1994, 654, 1. 13 Damiani, P., and Burini, G., J. Assoc. Off. Anal. Chem., 1988, 71, 462. 14 Yamaguisgi, M., Ishida, J., Zhu, X., Nakamura, M., and Yoshitake, T., J. Liq. Chromatogr., 1994, 17, 203. 15 Ulberth, F., J. Assoc. Off. Anal. Chem., 1991, 74, 630. 16 Gilson, T.D., Parker, S. M., and Woodward, J. R., Enzyme Microb. Technol., 1991, 13, 171. 17 Luque de Castro, M. D., and Papaefstathiou, I., in Encyclopedia of Environmental Analysis and Remediation, ed. R. A. Meyers, Wiley, New York, in the press. 18 Prinzing, U., Ogbomo, I., Lehn, C., and Schmidt, H. L., Sens. Actuators B, 1990, 1, 542. 19 Ogbomo, I., Steffl, A., Schumann, W., Prinzing, U., and Schmidt, H. L., J. Biotechnol., 1993, 31, 317. 20 Mattos, I. L., Luque de Castro, M.D., and Valc�arcel, M., Talanta, 1995, 42, 755. 21 Papaefstathiou, I., Tena, M. T., and Luque de Castro, M. D., Anal. Chim. Acta, 1995, 308, 246. 22 Papaefstathiou, I., Luque de Castro, M. D., and Valc�arcel, M., Fresenius’ J. Anal. Chem., 1996, 354, 442. 23 Papaefstathiou, I., and Luque de Castro, M. D., Anal. Lett., 1995, 28, 2063. 24 Papaefstathiou, I., and Luque de Castro, M. D., Anal. Chem., 1995, 67, 3916. Paper 6/06401I Received September 17, 1996 Accepted October 28, 1996 122 Analyst, February 1997, Vol. 122 On-line Flow Injection–Pervaporation of Beer Samples for the Determination of Diacetyl Jos�e M. Izquierdo-Ferrero, Juan M. Fern�andez-Romero and Mar�ýa D. Luque de Castro* Department of Analytical Chemistry, Faculty of Sciences, University of C�ordoba, E-14004 C�ordoba, Spain. E-mail: qa1lucam@uco.es A method for the determination of diacetyl based on the integration of a continuous pervaporation module and a flow injection manifold is presented.The chemical derivatization reaction involes condensation between diacetyl, a-naphthol and creatine in a basic medium to form a coloured compound, which is monitored photometrically at 530 nm. The method affords low detection and quantification limits (5 and 10 ng ml21 of diacetyl calculated as ±3s and ±10s, respectively, for n = 20), a wide linear range (10–2000 ng ml21, r2 = 0.9998, n = 7), good precision (RSD < 3%) and a sampling frequency of 8 h21. The method was applied to the determination of diacetyl in beer from bottled and medium-process samples and the results were in excellent agreement with those given by the conventional method.Keywords: Flow injection; continuous pervaporation; spectrophotometry; diacetyl; beer Diacetyl is an a-diketone responsible for the flavour of several foodstuffs (cheeses, yogurts, beers and wines) and contributes adversely to the taste quality. Diacetyl has been used for both testing food ripeness (e.g., in cheeses and wines) and monitoring of manufacture processes (e.g., in beers).1 Low levels of diacetyl and pentane-2,3-dione are desirable in beer fermentation, as these compounds confer an unpleasant taste similar to butter in the final product.This means that routine control is necessary during the fermentation of ethanol and other compounds in beer production. Beer with an acceptable taste requires the formation of esters and alcohols such as ethyl acetate, isoamyl acetate and isoamyl alcohol and the removal of undesirable compounds (such as a-diketones).Recently, research in this field has been focused on the development of new microorganisms which increase the metabolization of undesirable products during fermentation.2 The low concentration levels of diacetyl in beverages and the presence of potential interferents makes the development of analytical methods with higher sensitivity and selectivity essential. Previous methods for the determination of these species include a separation step, such as distillation, prior to the determination.Usually these methods have been implemented by using optical detection, such as photometry after condensation3,4 or chelating reactions5,6 and fluorimetry.7,8 The individual determination of diacetyl and pentane-2,3-dione has also been proposed using GC–FID,2,9 GC–ECD2,10,11 GC– MS12 or headspace GC.13 Several methods based on HPLC with continuous fluorimetric detection have also been reported.14,16 Recent developments in continuous flow analysis have led to the use of non-chromatographic continuous separation techniques for the determination of one or more analytes, the separation of components into groups and interference removal. Non-chromatographic continuous separation techniques can improve the sensitivity of the method through preconcentration and its selectivity by deleting matrix effects or particular interferents.Membrane-based non-chromatographic separation techniques are particularly useful for increasing selectivity.In some cases, they also allow analyte preconcentration, thereby resulting in improved sensitivity. Continuous gas-diffusion techniques, whether isolated or coupled to other separation techniques such as dialysis or ion exchange, make excellent tools for achieving high selectivity and sensitivity. However, these techniques suffer from serious drawbacks (e.g., clogging of the membrane pores by suspended particles or high molecular mass compounds occasionally present in the sample and deterioration of the membrane by contact with the sample).Both drawbacks can be overcome by using pervaporation.17 This technique has been employed for a long time in industry, in competition with other traditional processes (e.g., distillation, extraction and adsorption); by contrast, it has scarcely been used for analytical purposes. A laboratory-scale pervaporation unit was developed a few years ago in order to overcome problems arising from the use of biosensors in on-line fermentation monitoring.18,19 Several applications of laboratory pervaporation based on different approaches have been developed recently by Luque de Castro and co-workers.20–24 This paper reports the use of an automatic spectrophotometric method for the determination of diacetyl in beer samples based on the on-line coupling of a continuous pervaporation unit to a simple flow injection (FI) manifold.The chemical reaction is based on condensation of the analyte using an excess of anaphthol and creatine in a basic medium as described by Mattessich and Cooper,4 who improved a previous assay for the determination of this analyte. The condensation product is monitored spectrophotometrically at 530 nm. Experimental Instruments and Apparatus A Pye Unicam (Cambridge, UK) SP-500 spectrophotometer furnished with a Hellma (Jamaica, NY, USA) 178.011 QS flow cell and equipped with a Knauer (Bad Homburg, Germany) recorder was used.A laboratory-built pervaporation module equipped with a thermostated magnetic stirrer was also used. Two Gilson (Worthington, OH, USA) Minipuls-3 four-channel peristaltic pumps with rate selector, two Rheodyne (Cotati, CA, USA) Model 5041 injection valves and PTFE tubing of 0.5 mm id were also used. The pervaporation cell, designed in the laboratory was similar to others described by Mattos et al.20 Reagents and Solutions All reagents were of analytical-reagent grade.The donor stream was an aqueous solution with which the sample solution was mixed. A solution containing 125 mmol l21 a-naphthol (Sigma, St. Louis, MO, USA) and 100 mmol l21 creatine (Sigma) prel21 sodium hydroxide (Merck, Darmstradt, Germany) was used as a reagent/acceptor solution. Aqueous diacetyl (Merck) solution was used as standard after suitable dilution. Beer samples provided by Compa�n�ýa Andaluza de Cervezas (C�ordoba, Spain) were used for testing the method.All solutions were prepared in doubly distilled water of Analyst, February 1997, Vol. 122 (119–122) 119high purity obtained from a Millipore (Bedford, MA, USA) Milli-Q Plus system. PTFE membranes of 5.0 mm pore size and 47 mm diameter purchased from Millipore were also used. Manual Procedure A 100 ml volume of beer is subjected to distillation to obtain 20 ml of distillate, of which 10 ml are mixed with 0.5 ml of 1 mg ml21 o-phenylenediamine and kept in the dark for 25 min.Then 2 ml of 4 m HCl are added and the absorbance of the mixture is monitored at 335 nm within 30 min after mixing. A blank containing 10 ml of deionized water instead of distillate is also measured photometrically. Manifold and Procedure Fig. 1 depicts the hydrodynamic system used. It consists of a flow injection manifold divided into two parts. The lower part acts as donor submanifold, which consists of a monochannel system provided with a main injection valve (MIV) that injects the samples into an aqueous stream which circulates through the lower part of the pervaporation cell, which is thermostated at 90 °C.The diacetyl evaporates to the air gap between the donor solution and the membrane and diffuses through the hydrophobic membrane to the upper part of the pervaporation cell where it is collected. The upper submanifold includes the acceptor chamber of the pervaporation cell, which is located in the loop of an auxiliary injection valve (AIV).This valve is filled with the acceptor solution, which contains the reagent mixture (AIV in injection position) and then is switched to the filling position, thus keeping static the loop contents during a preset interval. The analyte which passes through the membrane is collected in the acceptor/reagent solution, with simultaneous development of the condensation reaction. After a preset pervaporation time (Tp), the AIV is switched again to the inject position and the reaction product is driven to the detector.The peak height (transient signal obtained when the plug containing the coloured product reaches the detector) is proportional to the concentration of diacetyl in the sample. Results and Discussion Optimization of Variables The optimization of the variables involved in the overall process, grouped into physical, chemical and hydrodynamic, was performed using the univariate method.Table 1 shows the ranges over which each variable was studied and the optimum value found. Physical variables Fig. 2(a) shows the influence of temperature in the pervaporation step, which was studied between 30 and 110 °C (boiling point of diacetyl 88 °C). A temperature of 90 °C was sufficient for appropriate transfer of the analyte through the membrane. This temperature was also suitable for the formation of the condensation product. Fig. 1 Manifold for the automatic–photometric determination of diacetyl in beer samples.P, Peristaltic pump; MIV, main injection valve; IV, injection volume; AIV, auxiliary injection valve; PM, pervaporation module; m, hydrophobic membrane; TMS, thermostatic magnetic stirrer; D, detector; w, waste; R, acceptor/reagent solution; C, donor solution; and S, sample. Table 1 Optimization of variables Range Optimum Type Variable studied value Physical Temperature/°C 30–110 90 FI Donor flow rate/ml min21 0.1–0.5 0.3 Acceptor flow rate/ml min21 0.5–2.0 1.8 Volume injected/ml 100–3000 2000 Pervaporation time/s 30–480 360 Chemical Sodium hydroxide concentration/mmol l21 100–1000 750 pH — 12 a-Naphthol concentration/mmol l21 5–200 125 Creatine concentration/mmol l21 5–150 100 Fig. 2 Influence of temperature (a) on the pervaporation process and (b) on the derivatization reaction for concentrations of diacetyl of 20 and 100 ng ml21 (dashed and continuous lines, respectively). 120 Analyst, February 1997, Vol. 122Chemical variables The influence of the pH in both the donor and acceptor solutions was studied. An aqueous solution adjusted to different pH values was used as a donor stream. A higher pervaporation efficiency was achieved by using doubly distilled water as the donor stream. However, the pervaporation efficiency was increased when the acceptor solution contained sodium hydroxide. This was tested by preparing different aqueous solutions containing sodium hydroxide at concentrations between 100 and 1000 mmol l21.A concentration of 750 mmol l21 provided the best medium for collection of the analyte. This solution was also appropriate for the development of the derivatization reaction. Concentrations of 125 mmol l21 a-naphthol and 100 mmol l21 creatine in the acceptor/reagent solution were selected as optimum. Higher concentrations of these reagents caused higher blank signals, resulting in appreciable lack of sensitivity (estimated as 50%).On the other hand, lower concentrations resulted in poor development of the chromogenic reaction. Hydrodynamic variables The flow rate had a dramatic influence on the performance of the system. As the flow rate of the donor solution determined the time during which the analyte was in the pervaporation module, 0.3 ml min21 was selected as optimum for the donor stream (lower values yielded non-reproducible results). On the other hand, the flow rate had no influence on the upper FI subsystem as the collection process occurred under static conditions.When the flow was re-started the reaction plug was driven to the detector. A flow rate of 1.8 ml min21 was appropriate to rinse the flow manifold. In order to obtain the best analytical signal with an acceptable sampling frequency, different stop-times in the range between 30 and 480 s were tried. A pervaporation time of 360 s was chosen as a compromise between the best signal and an acceptable sampling frequency.Injection volumes over 2 ml did not increase the analytical signal significantly. Features of the Method Ten standard solutions of diacetyl with concentrations between 1 and 5000 ng ml21 were prepared and injected in triplicate into the FI–pervaporation assembly using the optimum values of variables found previously. Table 2 summarizes the features of the method (equation, regression coefficient, linear range, detection and quantification limits and RSD).The linear range achieved for the method was suitable for applying it to the determination of diacetyl in beer samples. The sampling frequency achieved under the optimum working conditions was 8 h21. Study of Interferents The study of potential interferents was aimed at those commonly present in beer samples which are structurally similar to diacetyl, such as pyruvic acid, butylene glycol, acetoin (a precursor of diacetyl) and ascorbic acid. All interferents were added to the sample at concentrations higher than usually found in beers.The results showed that pyruvic acid, butylene glycol and ascorbic acid are tolerated at levels up to ten times their content in beer (with a tolerated interferent to analyte ratio of 100 : 1 for pyruvic acid and butylene glycol and 10 : 1 for ascorbic acid). Acetoin interferred at five times its concentration in these samples; nevertheless, acetoin is easily oxidized to diacetyl under atmospheric conditions.Application to the Proposed Method The method was applied to the determination of diacetyl in bottled and medium-process beer. In order to eliminate the gas contained in the original beverage (mainly CO2), all the samples were previously ultrasonicated for 5 min, then injected directly in triplicate. Table 3 summarizes the concentrations found and the results obtained after addition of two amounts of the analyte (50 and 100 ng ml21) to aliquots of the samples. The results provided by the proposed method were compared with those obtained by the conventional method used by Compa�n�ýa Andaluza de Cervezas.This method is based on previous separation of the analyte by di (vapour stream) followed by condensation of the analyte with ophenylenediamine, and photometric monitoring of the product at 335 nm. Six beer samples were used for this comparison and the results obtained produced a straight-line graph: y = 0.934x + 1.477 (r2 = 0.9901, n = 6) where y represents the FI–pervaporation method and x the conventional method.As can be seen, an excellent correlation exists between the proposed and the conventional methods. Conclusions The method proposed for the determination of diacetyl in beer samples shows the following features: simplicity in implementing the experimental set-up (i.e., a conventional FI manifold in Table 2 Features of the method Equation* A = 0.0199C + 0.0005 Regression coefficient 0.9998 (n = 7) Linear range/ng ml21 10–2000 Detection limit/ng ml21 5 Quantification limit/ng ml21 10 RSD (%)† — Low level 3.0 High level 2.6 * A denotes absorbance and C analyte concentration in ng ml21.† For 50 and 1000 ng ml21 of diacetyl. Table 3 Application of the method Concentration/ng ml21 Recovery (%)† Sample Conventional Proposed Difference No. Type* method method (%) Addition 1 Addition 2 1 A 18 17 25.5 100 96 2 A 15 16 6.5 94 98 3 A 25 24 24.0 98 97 4 B 30 29 23.3 100 99 5 B 25 27 8.0 107 99 6 B 52 50 23.8 100 100 * A and B denote bottled and medium-process beers, respectively.† 50 and 100 ng ml21 for additions 1 and 2, respectively. Analyst, February 1997, Vol. 122 121which an easy laboratory-built pervaporation unit is included); high selectivity as a result of the nature of the chemical reaction; lower sample consumption (2 ml); and an acceptable sensitivity. These features make it suitable for the routine determination of diacetyl in beer samples, as was demonstrated by comparison with the conventional method.The Comisi�on Interministerial de Ciencia y Tecnolog�ýa (CICYT) is thanked for financial support (project No. PB93- 0827). We gratefully acknowledge Carlos M. G�omez S�anchez of Compa�n�ýa Andaluza de Cervezas for the samples compared with the conventional method. References 1 West, D. B., Lautenbach, A. L., and Becker, K., Am. Soc. Brew. Proc., 1952, 81, 65. 2 Mathis, C., Pons, M. N., Engasser, J. M., and Lenoel, M., Anal. Chim. Acta, 1993, 279, 59. 3 Westerfeld, W. W., J. Biol. Chem., 1945, 161, 495. 4 Mattessich, J., and Cooper, J. R., Anal. Biochem., 1989, 180, 349. 5 Ribereau-Gayon, J., Peynaud, E., Sudraud, P., and Riberau-Gayon, P., Sciences et Techniques de Vin, Dunod, Paris, 1972. 6 Walsh, B., and Cogan, T. M., J. Dairy Res., 1988, 41, 31. 7 Garc�ýa-Vilanova, R. J., and Garc�ýa Estepa, R. M., Talanta, 1993, 40, 1419. 8 Mariaud, M., and Levillain, P., Talanta, 1994, 41, 75. 10 Barbieri, G., Bolzoni, L., Careri, M., Mangia, A., Paroladi, G., Spagnoli, S., and Virgili, R., J. Agric. Food. Chem., 1994, 42, 1170. 11 Otsuka, M., and Ohmori, S. J., J. Chromatogr. B, 1992, 115, 215. 12 Otsuka, M., and Ohmori, S. J., J. Chromatogr. B, 1994, 654, 1. 13 Damiani, P., and Burini, G., J. Assoc. Off. Anal. Chem., 1988, 71, 462. 14 Yamaguisgi, M., Ishida, J., Zhu, X., Nakamura, M., and Yoshitake, T., J. Liq. Chromatogr., 1994, 17, 203. 15 Ulberth, F., J. Assoc. Off. Anal. Chem., 1991, 74, 630. 16 Gilson, T. D., Parker, S. M., and Woodward, J. R., Enzyme Microb. Technol., 1991, 13, 171. 17 Luque de Castro, M. D., and Papaefstathiou, I., in Encyclopedia of Environmental Analysis and Remediation, ed. R. A. Meyers, Wiley, New York, in the press. 18 Prinzing, U., Ogbomo, I., Lehn, C., and Schmidt, H. L., Sens. Actuators B, 1990, 1, 542. 19 Ogbomo, I., Steffl, A., Schumann, W., Prinzing, U., and Schmidt, H. L., J. Biotechnol., 1993, 31, 317. 20 Mattos, I. L., Luque de Castro, M. D., and Valc�arcel, M., Talanta, 1995, 42, 755. 21 Papaefstathiou, I., Tena, M. T., and Luque de Castro, M. D., Anal. Chim. Acta, 1995, 308, 246. 22 Papaefstathiou, I., Luque de Castro, M. D., and Valc�arcel, M., Fresenius’ J. Anal. Chem., 1996, 354, 442. 23 Papaefstathiou, I., and Luque de Castro, M. D., Anal. Lett., 1995, 28, 2063. 24 Papaefstathiou, I., and Luque de Castro, M. D., Anal. Chem., 1995, 67, 3916. Paper 6/06401I Received September 17, 1996 Accepted October 28, 1996 122 Analyst, Feb

ISSN:0003-2654    

DOI:10.1039/a606401i

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Fully Robotic Method for Characterization of Toxic Residues A. Velasco-Arjona and M. D. Luque de Castro* Department of Analytical Chemistry, Faculty of Sciences, University of C�ordoba, E-140004 C�ordoba, Spain A fully automated method for the determination of toxic residues was developed using a robotic station. The robot performs the weighing of the solid sample and applies the standard leaching method based on continuous stirring with discontinuous pH monitoring and control by addition of acid solution for 24 h; it then filters the sample and prepares the different dilutions for application of the luminescence test which is monitored by aspiration of the solutions into the flow cell of a luminescence detector.The data from both the sample mass and dilution step are used together with those from the detector by the computer for calculation of the toxicity. The two manual procedures usually applied in routine laboratories were developed by the robotic station; the results obtained by both methods were compared with those provided by the manual method and showed excellent agreement. The main advantage of the proposed method is that it is fully automated as opposed to the constant human attendance (for at least 28 h) required by the manual method.Keywords: Robotic method; toxic residues; bioluminescence Most of the analytical methods accepted by governmental environmental organizations are manual and involve very slow, time-consuming steps forcing those responsible for routine laboratories to develop and use alternative methods which are faster, more precise and, particularly, with dramatic reduction of human participation.This development leads to laboratory automation, one of the most outstanding trends in analytical chemistry at present.1–4 The environmental field, concerning particularly solid or very complex samples, is very receptive to this trend as the pre-treatment of this type of sample unfailingly requires human participation unless a robotic station is used: there is no other automatic alternative capable of developing basic unitary operations such as weighing, leaching and centrifugation, among others.Conscious of the gap existing in the analytical literature on the preliminary operations of the analytical process concerning solid samples, our main aim is the development of robotic methods which, when developed manually, involve high human participation.With the objective of helping to fill this gap, we have developed methods based on the use of a robotic station alone5–8 or coupled to other automatic alternatives9–11 in order to carry out each analytical step by the most ‘friendly’ alternative. In this context, the work reported here is based on the development of a fully robotic method for the characterization of toxic residues. The present European Union (EU) legislation on this matter12 establishes a very time-consuming method for characterization of toxic residues which involves a leaching step for 24 h with pH control of the leaching solution, followed by a bioassay based on luminescence using Photobacterium phosphoreum, according to which a residue is considered toxic if the solution from the leaching step shows an EC50 (i.e., 50% reduction in luminescence intensity of a sample of bacteria considered to have luminescence of EC100).After weighing the sample manually, the leaching step can be developed by dedicated automated pH-stat devices; then, the user develops the biotest also manually.Experimental Instruments and Apparatus A robotic station consisting of a Zymate II Plus robot, a system V controller, a printer, an all-purpose hand, a syringe hand with a pipette tip rack, two 16 3 100 mm centrifuge tube racks, a Mettler (Highstown, NJ, USA) AE 200 balance, two Master Laboratory Stations (MLS), a Power and Event Controller (PEC), and a Dilute and Dissolve Station were used.In addition, a Shimadzu (New York, USA) RF-1501 spectrofluorimeter furnished with a Hellma (Jamaica, NY, USA) 178-QS flow cell and connected to the ac of the PEC via its analogue output for data acquisition, processing and delivery, a Crison micropH 2001 pH meter, a Gilson (Worthington, OH, USA) Minipuls-3 peristaltic pump, an Agimatic-N stirrer, a laboratory-made small vibrator motor, and a stainless-steel filter (Technicon, Tarrytown, NY, USA) were also used.The system V controller was interfaced to a Netset 286/400 personal computer. A Lumistherm luminescence detector (Neurtek) was used for the manual method. Reagents and Solutions Leaching step. Aqueous solutions (0.5 m) of acetic acid (HOAc) and NaOH, and solid NaCl (all supplied by Merck, Poole, Dorset, UK) were used. Determination step. Photobacterium phosphoreum bacterium and bacterium reactivation solution (both supplied by Neurtek), and 2% aqueous NaCl solution of pH 7.0–7.2 were used.Manual Method Leaching A representative amount (about 100 g) of solid sample is weighed and placed at a stirrer with 16 g of distilled water per gram of sample. Stirring is started and the pH of the solution is monitored. When the pH of the solution is higher than 5.2 it is lowered to 5.0 ± 0.2 by adding 0.5 m HOAc. The pH is adjusted at intervals of 15, 30 and 60 min, passing to the following interval when the pH adjustment is equal to or lower than 0.5 pH units. This sequence is repeated for at least 6 h.The extraction is complete after 24 h. According to the EU legislation12 the amount of acid added must never exceed 4 ml per gram of solid sample. When the extraction step is finished, an amount of distilled water given by the equation below is added V = 20W 2 16W 2 A (1) where V = volume of distilled water to be added (ml), A = volume of 0.5 m HOAc added during the leaching step (ml) and W = mass of sample poured into the extractor (g).Analyst, February 1997, Vol. 122 (123–128) 123After leaching, solid NaCl is added to make the sample 2% in this salt, the pH being finally adjusted to 7.0 ± 0.2 using dilute solutions of HCl or NaOH. Filtration of the treated sample is performed in order to obtain 10 ml of the solution. Biotest The luminescence biotest applied to the solution is as follows: (a) a 2% NaCl solution is prepared and its pH adjusted to 7.0 ± 0.2. (b) A vial of reactivation solution is thawed at room temperature in a water-bath and then placed into the storage wells of the Lumistherm (i.e., a fluorimeter with a thermostated rack where cuvettes divided into three rows, A, B and C, containing ten cuvettes each, are placed).A period of 15 min is allowed for temperature adjustment. (c) Meanwhile, dilutions of the sample are performed in the sequence 1 + 1, 1 + 1.5, 1 + 2, 1 + 3, 1 + 4, 1 + 6, 1 + 8, 1 + 12, 1 + 16. (d) After 15 min, a vial of bacteria is thawed quickly (2 min) in a water-bath to room temperature.(e) A 0.5 ml volume of reactivation solution is added to the bacteria vial and mixed by gentle shaking until homogenization is achieved. The vial is then incubated for 15 min in the Lumistherm rack. (f) The solution in the vial is transferred into the test-tube containing the remaining reactivation solution and the suspension is transferred from the test-tube into the vial and back until homogenization is achieved.(g) The above suspension is pipetted into 20 cuvettes (0.5 ml in each) which are placed in the incubation wells B1–B10 and C1–C10. (h) The luminescence of each cuvette is measured. (i) Portions (0.5 ml) of the diluted sample solutions (row A) are added to the cuvettes B and C. (j) After the exact incubation time (15 min, according to the Spanish legislation13), the luminescence of the cuvettes is measured following the same sequence and time intervals as in blank measurements.In this way blank and sample are always measured in duplicate. Data treatment A straight line is obtained by plotting log(EC100 2 EC)/EC versus log sample % for each tube (EC is the luminescence provided by each dilution of the sample solution). The point at which this line intersects the abscissa corresponds to the per cent. of diluted sample which provides a luminescence of EC50. This per cent. must be multiplied by a factor (F = 00) to obtain the concentration of toxicants expressed in mg l21 from g per 100 ml.Robotic Method The arrangement of the robotic station for development of the method is shown in Fig. 1, and the operational sequence of the overall process can be visualized in the flow chart (see Scheme 1), which can be divided into three general steps as in the conventional method. Leaching step The robot uses the all-purpose hand and catches a 1000 ml precipitate vessel, which is carried to the balance and tared. The robot takes the precipitate vessel containing the residue and sets it, in an inclined position, together with the vibrator motor14 over the precipitate vessel placed in the balance, thus permitting about 20 g of solid residue to be added, before bringing the 1000 ml precipitate vessel to the stirrer and adding 16 ml of distilled water per gram of residue, according to the exact mass of sample.The ac input of the PEC connected to the stirrer is turned on by the controller and stirring starts.The robot takes the electrode, introduces it into a drier tube in order to dry it, turns the stirrer off, introduces the electrode into the precipitate vessel and measures the pH of the solution. Meanwhile, the controller receives the pH signal by means of the ac input of the PEC which is connected to the pH meter. The robot removes the electrode and introduces it into a rinser tube, and then brings it to the diluted and dissolved module for thorough washing with 10 ml of distilled water.The robot returns the electrode to the initial position. If the pH of the solution is higher than 5.2, the robot adds a sufficient volume of HOAc solution to adjust the pH to 5.0 ± 0.2. After this addition, the robot takes the electrode again and checks the pH. The control of pH is repeated for 24 h with the same frequency as in the manual method. After 24 h, the robot adds a volume of distilled water given by: V = 950 2 16W 2 A (2) It adjusts the pH to 7.0 ± 0.2 by adding a sufficient volume of NaOH solution and then introduces the sample aspirater into the leached suspension and filters 10 ml of the treated sample, which is poured into a centrifuge tube with 0.2 g of solid NaCl.Fig. 1 Robotic station arrangement for the characterization of toxic residues. 124 Analyst, February 1997, Vol. 122Biotest The robot adds 1 ml of 2% NaCl solution to each of the eight tubes in row A and prepares the sample dilutions. Then, it adds 0.5 ml of the reactivation solution in the bacteria tube, stirs the tube in the vortex mixer for 2 min (appearance of turbidity) and sets it in the bath for 15 min.Following this, the robot takes the mixer, mixes both solutions, changes its hand and takes volumes of 0.5 ml from the bacteria tube which it pours into the eight empty tubes of row B and, after 15 min, adds 0.5 ml of the diluted sample solutions (one to each tube). The luminescence intensity of each solution is monitored 15 min later by aspiration of the solution into the flow cell placed in the detector. Data treatment The data are treated as described in the manual method, but the factor is now F = 4200 instead of 10 000 because of the different dilution of the leached suspension (as expressed under Calculation of the Dilution Factor). Description of the Peripherals, Their Alteration and Functioning Master laboratory station.The proposed method uses two MLSs, which consist of three syringes each, intended to dispense liquids in conjunction with the dilute and dissolve unit (a dual station that also includes the vortex mixer); however, they can be used for other purposes, as shown below.Vortex mixer. This is a module for stirring of centrifuge tubes. The robot places the tube in the vortex mixer and the controller selects the period and stirring velocity. All-purpose and syringe hands. The all-purpose hand allows the robot to seize test and centrifuge tubes and objects of similar size.For the all-purpose hand to be able to handle a tube of 1.5 mm od, that is, the usual diameter of a probe, an empty cartridge was pierced at the bottom and the aspiration tube inserted through the hole over a length of about 5 cm. The syringe hand is provided with a syringe which allows it to take 1 ml volumes. Controller. This module sends orders to the remainder of the units, including the robot. Power and event controller (PEC). The PEC module acts as an interface between the controller, peripherals and robot.Balance. This module includes two wires which open the balance door on contact and close it on separation. The wires are connected to a PEC switch in such a way that issuing the command bd:open.door brings them into contact (‘on’ position) to open the balance door. The balance plate for tube weighing was modified as shown in Fig. 1, thus allowing the precipitate vessel to be manoeuvred to and from the balance by the robot.Precipitate vessel. In order to make a 1000 ml precipitate vessel suitable for handling by the robot through its all-purpose hand, it was altered according to the detail in Fig. 1. Mixer. This is a Teflon tube of 3 mm id which is used by the robot for mixing the bacteria and reactivation solutions as detailed under Experimental. Drier tubes. These are centrifuge tubes coated on the inside with filter-paper. The robot introduces the electrode into these tubes for drying before pH measurements. Functions of the MLSs and Racks (See Fig. 2) Each of the syringes of the two MLSs is connected to both the reservoir and probe, modified as described above. When not in use, the probes are placed in row 1 of the racks. The robot uses these devices as follows: Water dispenser. The robot seizes and sets the water dispenser above the precipitate vessel and the volume of water required according to the sample mass is added by means of a peristaltic pump connected to the PEC.HOAc and NaOH dispensers. The robot seizes and sets these dispensers over the precipitate vessel and adds the necessary amount of solution to adjust the pH to the required value. Sample aspirater and dispenser. The robot takes the sample aspirater, which has a filter in its tip, and introduces it into the suspension of the leached sample and aspirates 10 ml of clean solution. It then takes the sample dispenser and places it over a tube containing 0.2 g of solid NaCl and adds the filtered solution from the syringe.Aspiration probe. The robot introduces the aspiration probe into a solution, turns the pump on and the solution is driven to the detector. Rinser tubes. These are centrifuge tubes filled with distilled water. The robot introduces the electrode into these tubes in order to remove the coarse particles adhered to it from the sample suspension. To dilute and dissolve. This module, usually employed for addition of liquids to test-tubes, is used here for washing the electrode. The robot places the electrode under the tip of this probe for removing those particles not removed during the rinse step using 10 ml of distilled water.NaCl containers. These are centrifuge tubes which contain 0.2 g of solid NaCl, the required amount of this salt to make the filtered solution 2% in NaCl. Results and Discussion Adaptation of the Manual Method for Robotic Development Some alterations had to be introduced into the manual method in order to make it robot-friendly.These changes were as follows: (a) Homogenization of the bacteria (B) mixture and the reactivation solution (RS) is achieved in the manual method by Scheme 1 Flow chart of the overall robotic process. Analyst, February 1997, Vol. 122 125turning the B tube over the RS tube. The B tube is rinsed three times with the contents of the RS tube. Finally, the whole mixture is gently stirred in the RS tube. As turning one tube over another is a difficult task for the robot, this step was modified as follows: The robot took the tube mixer, introduced it into the RS tube, aspirated a volume of the RS solution, added it to the B tube, aspirated the whole liquid from this tube, dispensed it over the remaining RS solution and rinsed the B tube three times with this mixture.In order to verify the degree of homogenization accomplished in this way, the following test was performed: the robot added 1 ml of the mixture (B + RS) to 11 centrifuge tubes located in rack 2 and measured the luminescence of each by pumping the liquid to the detector.The precision of this operation, expressed as the relative standard deviation (RSD), was 0.66% (see Table 1). (b) The luminescence of each solution in rows A and B was monitored in the manual method before adding the sample solution in order to obtain the signal of each blank (EC100) from which the EC value obtained after adding the diluted sample must be subtracted.The following test was performed in order to determine whether or not this step (not feasible in the robotic method as the liquid must be pumped to the detector for monitoring) could be deleted. Eleven solutions containing 1 ml of mixture solution (B + RS) and 1 ml of 2% NaCl solution were prepared by the robot and their luminescence was monitored by the robot 15 min after preparation. The RSD was 2.03% (see Table 1), showing that there is no significant difference in the preparation of the blanks which justifies the monitoring of each blank.Thus, only one blank was used in order to obtain the EC100 value. (c) It is not clear in the manual method what must be done when more than 4 ml of acid per gram of residue has to be added to the sample in order to maintain the pH at 5.0 ± 0.2 during the 24 h of the leaching step. Some users do not add more than this amount even if the pH is higher than 5.2. In contrast, others prefer to add the necessary volume to maintain the pH within the recommended range.In order to check the influence of each criterion in the final result, one of the residues which was found to require more than 4 ml of acid was treated by the two procedures. The results obtained by the method with pH control (RSD 1.43%) were slightly better than those obtained by the method without pH control (RSD 1.99%). For subsequent experiments the procedure with constant control of the pH within the established range was adopted. This involves the possibility of having to add a volume of solution higher than that given by eqn.(1). To overcome this drawback, a higher final volume (950 ml versus 400 ml of the manual method) of leached sample was obtained by applying eqn. (2). (d) In the manual method, a standard cuvette is used to measure the luminescence of the solutions, whereas in the robotic method, a flow cell is used; therefore, in the robotic method, the signal is lower as the emitting portion reaching the detector beam is smaller.In order to increase the signal, only 5 ml of reactivation solution were added to the bacteria, thus affording a more concentrated bacterial solution. (e) The manual step of filtering the leached suspension in order to have a volume of solution sufficient for the determination was modified in the robotic method by placing a metal filter at the tip of the sample aspiration probe. In this way the clean leached solution filled the body of the syringe, and was subsequently dispensed into the A and B tubes by the sample dispenser.Fig. 2 MLS syringe and rack tube arrangements. Table 1 Repeatability study. (1) Homogenization of the bacteria (B) and the reactivation solution (RS). (2) Solution containing 1 ml of mixture solution (B + RS) and 1 ml of 2% NaCl solution Luminescence Luminescence Check intensity (1) intensity (2) 1 705 230 2 709 223 3 696 225 4 701 218 5 708 219 6 711 233 7 699 225 8 705 226 9 700 230 10 702 224 11 706 228 Average 703.8 225.6 RSD (%) 0.66 2.03 126 Analyst, February 1997, Vol. 122Calculation of the Dilution Factor Taking into account both the possible addition of volumes of HOAc solution higher than 4 ml per gram of sample, which in turn modifies the volume of water added to the suspension after leaching, and the smaller volume of reactivation solution added to the bacteria solution in order to increase sensitivity, a new factor had to be calculated for conversion of the value of the EC50 into mg l21 of toxic residues in the sample.With this aim, and owing to the absence of certified reference materials in our laboratory, the content of toxic residues in one of the samples, as calculated by the manual method, was used as an exact value. This sample was treated six times by the robotic method in order to obtain an average value of EC50 from which a value for the factor F of 4200 was calculated.The alteration of this factor takes into account the final volume of the leached suspension and also losses due to evaporation, which are significant during the 24 h required for the leaching step. Validation of the Method The proposed method was validated by applying it to four samples of industrial organic/inorganic residues (main components given in Table 2) and using procedure B (i.e., maintaining the pH at 5.0 ± 0.2 by addition of the necessary amount of HOAc solution, and exceeding the amount of 4 ml per gram of sample when required).Fig. 3(a) shows a plot of the luminescence intensity provided by the different dilutions of the leaching solution versus the percentage of sample in each tube. The shape of these curves does not permit immediate calculation of the toxic residues in the original sample. The transformation of the data as shown in Fig. 3(b) and the plotting of these values [i.e., log(EC100 2 EC)/EC versus log sample %] allows the content to be calculated easily.The content of toxic substances in the sample was also determined by the manual method. The results obtained with the robotic method were consistent with those provided by the manual method (see Table 3), which confirms the usefulness of the fully automated alternative. Conclusions The method reported here is a representative example of how a robotic station can be used for industrial analysis in order to control the toxicity of waste materials.Both the sample pretreatment and the measurement steps are sufficiently timeconsuming to justify their automation. As there are no other automated alternatives capable of developing sample weighing, pH control, etc., a robotic station is the only way to automate the overall analytical process fully. Auxiliary energy sources such as ultrasound15,16 or microwaves17,18 can be used in order to shorten the leaching step in non-official methods. The authors are grateful to the Spanish Direcci�on General de Investigaci�on Cient�ýfica y T�ecnica (DGICyT) for financial support in the form of Grant PB93/0827.The authors also thank Residues Inertization Plant (EGMASA, Huelva, Spain) for kindly supplying the various samples. References 1 Valc�arcel, M., and Luque de Castro, M. D., Automatic Methods of Analysis, Elsevier, Amsterdam, 1988. 2 Luque de Castro, M. D., and Tena, M. T., Talanta, 1995, 42, 151. Table 2 Main inorganic components (%) and pH of the samples R1 R2 R3 R4 Cd ND* 0.8 0.1 < 0.1 Cr < 0.1 5.2 0.2 < 0.1 Cu < 0.1 0.4 0.3 0.6 Fe 4.3 17.2 19.3 36.2 Mn 0.3 2.4 3.1 0.1 Mo ND 1.3 ND ND Ni < 0.1 2.0 < 0.1 < 0.1 Pb < 0.1 1.8 2.3 0.1 Sn ND ND ND ND Ti 27.5 0.1 ND 0.1 Zn < 0.1 9.5 19.9 1.2 V ND ND ND 4.9 As ND ND ND 4.9 pH 1.7 12.4 8.3 6.9 H2O 39.2 < 0.1 1.2 72.1 NH3 — — — 1.7 *Not detected. Fig. 3 (a) Plot of relative luminescence intensity versus sample per cent. in the measurement tube for four real samples; (b) linearization of the curves in (a) by plotting log(EC1002EC)/EC versus log sample per cent.The point at which this line intersects the abscissa corresponds to the sample per cent. which provides an EC50 (50% of blank luminescence). Table 3 Comparison of the results obtained by the manual and the robotic methods Manual Robotic method/ method/ Sample mg l21 mg l21 Error (%) Residue 1 9 400 9 600 2.13 Residue 2 35 000 34 700 20.86 Residue 3 64 100 63 000 21.72 Residue 4 98 500 97 800 20.71 Analyst, February 1997, Vol. 122 1273 Hurst, W. J., and Mortimer, J. W., Laboratory Robotics: A Guide to Planning, Programming and Applications, VCH, New York, 1987. 4 Tena, M. T., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chem., 1995, 67, 1054. 5 Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Lab. Robot. Autom., 1993, 5, 29. 6 Torres, P., Garc�&yacutuque de Castro, M. D., and Valc�arcel, M., Lab. Robot.Autom., 1994, 6, 229. 7 Torres, P., Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Lab. Robot. Autom., 1994, 6, 233. 8 Torres, P., Garc�ýa-Mesa, J. A., and Luque de Castro, M. D., J. Autom. Chem., 1994, 16, 183. 9 Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chem., 1993, 65, 3540. 10 Torres, P., Ballesteros, E., and Luque de Castro, M. D., Anal. Chim. Acta, 1995, 308, 371. 11 Velasco-Arjona, A., and Luque de Castro, M. D., Anal.Chim. Acta, in the press. 12 Directive of the European Community, 1984, p. 449. 13 Boletin Oficial del Estado, Annex 4, Luminescence Bioassay, 1989, 10 November, 35220. 14 Torres, P., Garc�ýa-Mesa, J. A., and Luque de Castro, M. D., Fresenius’ J. Anal. Chem., 1993, 346, 704. 15 Chen, D., L�azaro, F., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1989, 226, 221. 16 L�azaro, F., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1991, 242, 283. 17 Torres, P., Ballesteros, E., and Luque de Castro, M.D., Anal. Chim. Acta, 1995, 308, 371. 18 Bryce, D. W., Izquierdo, A., and Luque de Castro, M. D., Anal. Chim. Acta, 1996, 324, 69. Paper 6/05048D Received July 22, 1996 Accepted November 7, 1996 128 Analyst, February 1997, Vol. 122 Fully Robotic Method for Characterization of Toxic Residues A. Velasco-Arjona and M. D. Luque de Castro* Department of Analytical Chemistry, Faculty of Sciences, University of C�ordoba, E-140004 C�ordoba, Spain A fully automated method for the determination of toxic residues was developed using a robotic station.The robot performs the weighing of the solid sample and applies the standard leaching method based on continuous stirring with discontinuous pH monitoring and control by addition of acid solution for 24 h; it then filters the sample and prepares the different dilutions for application of the luminescence test which is monitored by aspiration of the solutions into the flow cell of a luminescence detector. The data from both the sample mass and dilution step are used together with those from the detector by the computer for calculation of the toxicity.The two manual procedures usually applied in routine laboratories were developed by the robotic station; the results obtained by both methods were compared with those provided by the manual method and showed excellent agreement. The main advantage of the proposed method is that it is fully automated as opposed to the constant human attendance (for at least 28 h) required by the manual method. Keywords: Robotic method; toxic residues; bioluminescence Most of the analytical methods accepted by governmental environmental organizations are manual and involve very slow, time-consuming steps forcing those responsible for routine laboratories to develop and use alternative methods which are faster, more precise and, particularly, with dramatic reduction of human participation.This development leads to laboratory automation, one of the most outstanding trends in analytical chemistry at present.1–4 The environmental field, concerning particularly solid or very complex samples, is very receptive to this trend as the pre-treatment of this type of sample unfailingly requires human participation unless a robotic station is used: there is no other automatic alternative capable of developing basic unitary operations such as weighing, leaching and centrifugation, among others.Conscious of the gap existing in the analytical literature on the preliminary operations of the analytical process concerning solid samples, our main aim is the development of robotic methods which, when developed manually, involve high human participation. With the objective of helping to fill this gap, we have developed methods based on the use of a robotic station alone5–8 or coupled to other automatic alternatives9–11 in order to carry out each analytical step by the most ‘friendly’ alternative.In this context, the work reported here is based on the development of a fully robotic method for the characterization of toxic residues. The present European Union (EU) legislation on this matter12 establishes a very time-consuming method for characterization of toxic residues which involves a leaching step for 24 h with pH control of the leaching solution, followed by a bioassay based on luminescence using Photobacterium phosphoreum, according to which a residue is considered toxic if the solution from the leaching step shows an EC50 (i.e., 50% reduction in luminescence intensity of a sample of bacteria considered to have luminescence of EC100).After weighing the sample manually, the leaching step can be developed by dedicated automated pH-stat devices; then, the user develops the biotest also manually. Experimental Instruments and Apparatus A robotic station consisting of a Zymate II Plus robot, a system V controller, a printer, an all-purpose hand, a syringe hand with a pipette tip rack, two 16 3 100 mm centrifuge tube racks, a Mettler (Highstown, NJ, USA) AE 200 balance, two Master Laboratory Stations (MLS), a Power and Event Controller (PEC), and a Dilute and Dissolve Station were used.In addition, a Shimadzu (New York, USA) RF-1501 spectrofluorimeter furnished with a Hellma (Jamaica, NY, USA) 178-QS flow cell and connected to the ac of the PEC via its analogue output for data acquisition, processing and delivery, a Crison micropH 2001 pH meter, a Gilson (Worthington, OH, USA) Minipuls-3 peristaltic pump, an Agimatic-N stirrer, a laboratory-made small vibrator motor, and a stainless-steel filter (Technicon, Tarrytown, NY, USA) were also used.The system V controller was interfaced to a Netset 286/400 personal computer. A Lumistherm luminescence detector (Neurtek) was used for the manual method. Reagents and Solutions Leaching step.Aqueous solutions (0.5 m) of acetic acid (HOAc) and NaOH, and solid NaCl (all supplied by Merck, Poole, Dorset, UK) were used. Determination step. Photobacterium phosphoreum bacterium and bacterium reactivation solution (both supplied by Neurtek), and 2% aqueous NaCl solution of pH 7.0–7.2 were used. Manual Method Leaching A representative amount (about 100 g) of solid sample is weighed and placed at a stirrer with 16 g of distilled water per gram of sample. Stirring is started and the pH of the solution is monitored.When the pH of the solution is higher than 5.2 it is lowered to 5.0 ± 0.2 by adding 0.5 m HOAc. The pH is adjusted at intervals of 15, 30 and 60 min, passing to the following interval when the pH adjustment is equal to or lower than 0.5 pH units. This sequence is repeated for at least 6 h. The extraction is complete after 24 h. According to the EU legislation12 the amount of acid added must never exceed 4 ml per gram of solid sample.When the extraction step is finished, an amount of distilled water given by the equation below is added V = 20W 2 16W 2 A (1) where V = volume of distilled water to be added (ml), A = volume of 0.5 m HOAc added during the leaching step (ml) and W = mass of sample poured into the extractor (g). Analyst, February 1997, Vol. 122 (123–128) 123After leaching, solid NaCl is added to make the sample 2% in this salt, the pH being finally adjusted to 7.0 ± 0.2 using dilute solutions of HCl or NaOH.Filtration of the treated sample is performed in order to obtain 10 ml of the solution. Biotest The luminescence biotest applied to the solution is as follows: (a) a 2% NaCl solution is prepared and its pH adjusted to 7.0 ± 0.2. (b) A vial of reactivation solution is thawed at room temperature in a water-bath and then placed into the storage wells of the Lumistherm (i.e., a fluorimeter with a thermostated rack where cuvettes divided into three rows, A, B and C, containing ten cuvettes each, are placed).A period of 15 min is allowed for temperature adjustment. (c) Meanwhile, dilutions of the sample are performed in the sequence 1 + 1, 1 + 1.5, 1 + 2, 1 + 3, 1 + 4, 1 + 6, 1 + 8, 1 + 12, 1 + 16. (d) After 15 min, a vial of bacteria is thawed quickly (2 min) in a water-bath to room temperature.activation solution is added to the bacteria vial and mixed by gentle shaking until homogenization is achieved.The vial is then incubated for 15 min in the Lumistherm rack. (f) The solution in the vial is transferred into the test-tube containing the remaining reactivation solution and the suspension is transferred from the test-tube into the vial and back until homogenization is achieved. (g) The above suspension is pipetted into 20 cuvettes (0.5 ml in each) which are placed in the incubation wells B1–B10 and C1–C10. (h) The luminescence of each cuvette is measured. (i) Portions (0.5 ml) of the diluted sample solutions (row A) are added to the cuvettes B and C.(j) After the exact incubation time (15 min, according to the Spanish legislation13), the luminescence of the cuvettes is measured following the same sequence and time intervals as in blank measurements. In this way blank and sample are always measured in duplicate. Data treatment A straight line is obtained by plotting log(EC100 2 EC)/EC versus log sample % for each tube (EC is the luminescence provided by each dilution of the sample solution).The point at which this line intersects the abscissa corresponds to the per cent. of diluted sample which provides a luminescence of EC50. This per cent. must be multiplied by a factor (F = 10 000) to obtain the concentration of toxicants expressed in mg l21 from g per 100 ml. Robotic Method The arrangement of the robotic station for development of the method is shown in Fig. 1, and the operational sequence of the overall process can be visualized in the flow chart (see Scheme 1), which can be divided into three general steps as in the conventional method.Leaching step The robot uses the all-purpose hand and catches a 1000 ml precipitate vessel, which is carried to the balance and tared. The robot takes the precipitate vessel containing the residue and sets it, in an inclined position, together with the vibrator motor14 over the precipitate vessel placed in the balance, thus permitting about 20 g of solid residue to be added, before bringing the 1000 ml precipitate vessel to the stirrer and adding 16 ml of distilled water per gram of residue, according to the exact mass of sample.The ac input of the PEC connected to the stirrer is turned on by the controller and stirring starts. The robot takes the electrode, introduces it into a drier tube in order to dry it, turns the stirrer off, introduces the electrode into the precipitate vessel and measures the pH of the solution.Meanwhile, the controller receives the pH signal by means of the ac input of the PEC which is connected to the pH meter. The robot removes the electrode and introduces it into a rinser tube, and then brings it to the diluted and dissolved module for thorough washing with 10 ml of distilled water. The robot returns the electrode to the initial position. If the pH of the solution is higher than 5.2, the robot adds a sufficient volume of HOAc solution to adjust the pH to 5.0 ± 0.2.After this addition, the robot takes the electrode again and checks the pH. The control of pH is repeated for 24 h with the same frequency as in the manual method. After 24 h, the robot adds a volume of distilled water given by: V = 950 2 16W 2 A (2) It adjusts the pH to 7.0 ± 0.2 by adding a sufficient volume of NaOH solution and then introduces the sample aspirater into the leached suspension and filters 10 ml of the treated sample, which is poured into a centrifuge tube with 0.2 g of solid NaCl.Fig. 1 Robotic station arrangement for the characterization of toxic residues. 124 Analyst, February 1997, Vol. 122Biotest The robot adds 1 ml of 2% NaCl solution to each of the eight tubes in row A and prepares the sample dilutions. Then, it adds 0.5 ml of the reactivation solution in the bacteria tube, stirs the tube in the vortex mixer for 2 min (appearance of turbidity) and sets it in the bath for 15 min. Following this, the robot takes the mixer, mixes both solutions, changes its hand and takes volumes of 0.5 ml from the bacteria tube which it pours into the eight empty tubes of row B and, after 15 min, adds 0.5 ml of the diluted sample solutions (one to each tube). The luminescence intensity of each solution is monitored 15 min later by aspiration of the solution into the flow cell placed in the detector.Data treatment The data are treated as described in the manual method, but the factor is now F = 4200 instead of 10 000 because of the different dilution of the leached suspension (as expressed under Calculation of the Dilution Factor).Description of the Peripherals, Their Alteration and Functioning Master laboratory station. The proposed method uses two MLSs, which consist of three syringes each, intended to dispense liquids in conjunction with the dilute and dissolve unit (a dual station that also includes the vortex mixer); however, they can be used for other purposes, as shown below.Vortex mixer. This is a module for stirring of centrifuge tubes. The robot places the tube in the vortex mixer and the controller selects the period and stirring velocity. All-purpose and syringe hands. The all-purpose hand allows the robot to seize test and centrifuge tubes and objects of similar size. For the all-purpose hand to be able to handle a tube of 1.5 mm od, that is, the usual diameter of a probe, an empty cartridge was pierced at the bottom and the aspiration tube inserted through the hole over a length of about 5 cm.The syringe hand is provided with a syringe which allows it to take 1 ml volumes. Controller. This module sends orders to the remainder of the units, including the robot. Power and event controller (PEC). The PEC module acts as an interface between the controller, peripherals and robot. Balance. This module includes two wires which open the balance door on contact and close it on separation. The wires are connected to a PEC switch in such a way that issuing the command bd:open.door brings them into contact (‘on’ position) to open the balance door.The balance plate for tube weighing was modified as shown in Fig. 1, thus allowing the precipitate vessel to be manoeuvred to and from the balance by the robot. Precipitate vessel. In order to make a 1000 ml precipitate vessel suitable for handling by the robot through its all-purpose hand, it was altered according to the detail in Fig. 1. Mixer. This is a Teflon tube of 3 mm id which is used by the robot for mixing the bacteria and reactivation solutions as detailed under Experimental. Drier tubes. These are centrifuge tubes coated on the inside with filter-paper. The robot introduces the electrode into these tubes for drying before pH measurements. Functions of the MLSs and Racks (See Fig. 2) Each of the syringes of the two MLSs is connected to both the reservoir and probe, modified as described above.When not in use, the probes are placed in row 1 of the racks. The robot uses these devices as follows: Water dispenser. The robot seizes and sets the water dispenser above the precipitate vessel and the volume of water required according to the sample mass is added by means of a peristaltic pump connected to the PEC. HOAc and NaOH dispensers. The robot seizes and sets these dispensers over the precipitate vessel and adds the necessary amount of solution to adjust the pH to the required value.Sample aspirater and dispenser. The robot takes the sample aspirater, which has a filter in its tip, and introduces it into the suspension of the leached sample and aspirates 10 ml of clean solution. It then takes the sample dispenser and places it over a tube containing 0.2 g of solid NaCl and adds the filtered solution from the syringe. Aspiration probe. The robot introduces the aspiration probe into a solution, turns the pump on and the solution is driven to the detector.Rinser tubes. These are centrifuge tubes filled with distilled water. The robot introduces the electrode into these tubes in order to remove the coarse particles adhered to it from the sample suspension. To dilute and dissolve. This module, usually employed for addition of liquids to test-tubes, is used here for washing the electrode. The robot places the electrode under the tip of this probe for removing those particles not removed during the rinse step using 10 ml of distilled water.NaCl containers. These are centrifuge tubes which contain 0.2 g of solid NaCl, the required amount of this salt to make the filtered solution 2% in NaCl. Results and Discussion Adaptation of the Manual Method for Robotic Development Some alterations had to be introduced into the manual method in order to make it robot-friendly. These changes were as follows: (a) Homogenization of the bacteria (B) mixture and the reactivation solution (RS) is achieved in the manual method by Scheme 1 Flow chart of the overall robotic process.Analyst, February 1997, Vol. 122 125turning the B tube over the RS tube. The B tube is rinsed three times with the contents of the RS tube. Finally, the whole mixture is gently stirred in the RS tube. As turning one tube over another is a difficult task for the robot, this step was modified as follows: The robot took the tube mixer, introduced it into the RS tube, aspirated a volume of the RS solution, added it to the B tube, aspirated the whole liquid from this tube, dispensed it over the remaining RS solution and rinsed the B tube three times with this mixture.In order to verify the degree of homogenization accomplished in this way, the following test was performed: the robot added 1 ml of the mixture (B + RS) to 11 centrifuge tubes located in rack 2 and measured the luminescence of each by pumping the liquid to the detector. The precision of this operation, expressed as the relative standard deviation (RSD), was 0.66% (see Table 1).(b) The luminescence of each solution in rows A and B was monitored in the manual method before adding the sample solution in order to obtain the signal of each blank (EC100) from which the EC value obtained after adding the diluted sample must be subtracted. The following test was performed in order to determine whether or not this step (not feasible in the robotic method as the liquid must be pumped to the detector for monitoring) could be deleted.Eleven solutions containing 1 ml of mixture solution (B + RS) and 1 ml of 2% NaCl solution were prepared by the robot and their luminescence was monitored by the robot 15 min after preparation. The RSD was 2.03% (see Table 1), showing that there is no significant difference in the preparation of the blanks which justifies the monitoring of each blank. Thus, only one blank was used in order to obtain the EC100 value. (c) It is not clear in the manual method what must be done when more than 4 ml of acid per gram of residue has to be added to the sample in order to maintain the pH at 5.0 ± 0.2 during the 24 h of the leaching step.Some users do not add more than this amount even if the pH is higher than 5.2. In contrast, others prefer to add the necessary volume to maintain the pH within the recommended range. In order to check the influence of each criterion in the final result, one of the residues which was found to require more than 4 ml of acid was treated by the two procedures.The results obtained by the method with pH control (RSD 1.43%) were slightly better than those obtained by the method without pH control (RSD 1.99%). For subsequent experiments the procedure with constant control of the pH within the established range was adopted. This involves the possibility of having to add a volume of solution higher than that given by eqn. (1). To overcome this drawback, a higher final volume (950 ml versus 400 ml of the manual method) of leached sample was obtained by applying eqn.(2). (d) In the manual method, a standard cuvette is used to measure the luminescence of the solutions, whereas in the robotic method, a flow cell is used; therefore, in the robotic method, the signal is lower as the emitting portion reaching the detector beam is smaller. In order to increase the signal, only 5 ml of reactivation solution were added to the bacteria, thus affording a more concentrated bacterial solution.(e) The manual step of filtering the leached suspension in order to have a volume of solution sufficient for the determination was modified in the robotic method by placing a metal filter at the tip of the sample aspiration probe. In this way the clean leached solution filled the body of the syringe, and was subsequently dispensed into the A and B tubes by the sample dispenser. Fig. 2 MLS syringe and rack tube arrangements.Table 1 Repeatability study. (1) Homogenization of the bacteria (B) and the reactivation solution (RS). (2) Solution containing 1 ml of mixture solution (B + RS) and 1 ml of 2% NaCl solution Luminescence Luminescence Check intensity (1) intensity (2) 1 705 230 2 709 223 3 696 225 4 701 218 5 708 219 6 711 233 7 699 225 8 705 226 9 700 230 10 702 224 11 706 228 Average 703.8 225.6 RSD (%) 0.66 2.03 126 Analyst, February 1997, Vol. 122Calculation of the Dilution Factor Taking into account both the possible addition of volumes of HOAc solution higher than 4 ml per gram of sample, which in turn modifies the volume of water added to the suspension after leaching, and the smaller volume of reactivation solution added to the bacteria solution in order to increase sensitivity, a new factor had to be calculated for conversion of the value of the EC50 into mg l21 of toxic residues in the sample.With this aim, and owing to the absence of certified reference materials in our laboratory, the content of toxic residues in one of the samples, as calculated by the manual method, was used as an exact value.This sample was treated six times by the robotic method in order to obtain an average value of EC50 from which a value for the factor F of 4200 was calculated. The alteration of this factor takes into account the final volume of the leached suspension and also losses due to evaporation, which are significant during the 24 h required for the leaching step.Validation of the Method The proposed method was validated by applying it to four samples of industrial organic/inorganic residues (main components given in Table 2) and using procedure B (i.e., maintaining the pH at 5.0 ± 0.2 by addition of the necessary amount of HOAc solution, and exceeding the amount of 4 ml per gram of sample when required). Fig. 3(a) shows a plot of the luminescence intensity provided by the different dilutions of the leaching solution versus the percentage of sample in each tube.The shape of these curves does not permit immediate calculation of the toxic residues in the original sample. The transformation of the data as shown in Fig. 3(b) and the plotting of these values [i.e., log(EC100 2 EC)/EC versus log sample %] allows the content to be calculated easily. The content of toxic substances in the sample was also determined by the manual method. The results obtained with the robotic method were consistent with those provided by the manual method (see Table 3), which confirms the usefulness of the fully automated alternative.Conclusions The method reported here is a representative example of how a robotic station can be used for industrial analysis in order to control the toxicity of waste materials. Both the sample pretreatment and the measurement steps are sufficiently timeconsuming to justify their automation. As there are no other automated alternatives capable of developing sample weighing, pH control, etc., a robotic station is the only way to automate the overall analytical process fully.Auxiliary energy sources such as ultrasound15,16 or microwaves17,18 can be used in order to shorten the leaching step in non-official methods. The authors are grateful to the Spanish Direcci�on General de Investigaci�on Cient�ýfica y T�ecnica (DGICyT) for financial support in the form of Grant PB93/0827. The authors also thank Residues Inertization Plant (EGMASA, Huelva, Spain) for kindly supplying the various samples. References 1 Valc�arcel, M., and Luque de Castro, M.D., Automatic Methods of Analysis, Elsevier, Amsterdam, 1988. 2 Luque de Castro, M. D., and Tena, M. T., Talanta, 1995, 42, 151. Table 2 Main inorganic components (%) and pH of the samples R1 R2 R3 R4 Cd ND* 0.8 0.1 < 0.1 Cr < 0.1 5.2 0.2 < 0.1 Cu < 0.1.6 Fe 4.3 17.2 19.3 36.2 Mn 0.3 2.4 3.1 0.1 Mo ND 1.3 ND ND Ni < 0.1 2.0 < 0.1 < 0.1 Pb < 0.1 1.8 2.3 0.1 Sn ND ND ND ND Ti 27.5 0.1 ND 0.1 Zn < 0.1 9.5 19.9 1.2 V ND ND ND 4.9 As ND ND ND 4.9 pH 1.7 12.4 8.3 6.9 H2O 39.2 < 0.1 1.2 72.1 NH3 — — — 1.7 *Not detected. Fig. 3 (a) Plot of relative luminescence intensity versus sample per cent. in the measurement tube for four real samples; (b) linearization of the curves in (a) by plotting log(EC1002EC)/EC versus log sample per cent. The point at which this line intersects the abscissa corresponds to the sample per cent. which provides an EC50 (50% of blank luminescence). Table 3 Comparison of the results obtained by the manual and the robotic methods Manual Robotic method/ method/ Sample mg l21 mg l21 Error (%) Residue 1 9 400 9 600 2.13 Residue 2 35 000 34 700 20.86 Residue 3 64 100 63 000 21.72 Residue 4 98 500 97 800 20.71 Analyst, February 1997, Vol. 122 1273 Hurst, W. J., and Mortimer, J. W., Laboratory Robotics: A Guide to Planning, Programming and Applications, VCH, New York, 1987. 4 Tena, M. T., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chem., 1995, 67, 1054. 5 Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Lab. Robot. Autom., 1993, 5, 29. 6 Torres, P., Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Lab. Robot. Autom., 1994, 6, 229. 7 Torres, P., Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Lab. Robot. Autom., 1994, 6, 233. 8 Torres, P., Garc�ýa-Mesa, J. A., and Luque de Castro, M. D., J. Autom. Chem., 1994, 16, 183. 9 Garc�ýa-Mesa, J. A., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chem., 1993, 65, 3540. 10 Torres, P., Ballesteros, E., and Luque de Castro, M. D., Anal. Chim. Acta, 1995, 308, 371. 11 Velasco-Arjona, A., and Luque de Castro, M. D., Anal. Chim. Acta, in the press. 12 Directive of the European Community, 1984, p. 449. 13 Boletin Oficial del Estado, Annex 4, Luminescence Bioassay, 1989, 10 November, 35220. 14 Torres, P., Garc�ýa-Mesa, J. A., and Luque de Castro, M. D., Fresenius’ J. Anal. Chem., 1993, 346, 704. 15 Chen, D., L�azaro, F., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1989, 226, 221. 16 L�azaro, F., Luque de Castro, M. D., and Valc�arcel, M., Anal. Chim. Acta, 1991, 242, 283. 17 Torres, P., Ballesteros, E., and Luque de Castro, M. D., Anal. Chim. Acta, 1995, 308, 371. 18 Bryce, D. W., Izquierdo, A., and Luque de Castro, M. D., Anal. Chim. Acta, 1996, 324, 69. Paper 6/05048D Received July 22, 1996 Accepted November 7, 1996 128 An

ISSN:0003-2654    

DOI:10.1039/a605048d

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Determination of Ultratrace Amounts of Metallic and Chloride Ion Impurities in Organic Materials for Microelectronics Devices After a Microwave Digestion Method Miyuki Takenaka*a, Shoji Kozukaa, Masaru Hayashia and Hiroshi Endob a Research and Development Center, Toshiba Corporation, 1, Komukai, Toshiba-cho, Saiwai-ku, Kawasaki 210, Japan b Toshiba Research and Consulting Corporation, 1, Komukai, Toshiba-cho, Saiwai-ku, Kawasaki 210, Japan A digestion method was developed for the determination of ultratrace concentrations of sodium, potassium, magnesium, calcium and chloride ions.Using ETAAS and ion chromatography, the digestion method was applied successfully to the determination of ultratrace concentrations of these elements in organic materials for microelectronics devices such as photoresists, epoxy resins, and liquid crystals. Very low contamination levels were maintained throughout the procedure. The blank levels were 0.05 ng for sodium, 0.02 ng for potassium, 0.03 ng for magnesium and calcium, and 20 ng for chloride ion.The method is very effective in measuring the impurity distributions of organic materials whilst preventing their contamination from the surrounding environment and from the reagents used in the procedure. Keywords: Atomic absorption spectrometry; ion chromatography; metallic impurity; chloride ion; microwave digestion; alkali fusion; photoresist; epoxy resin; liquid crystal Organic materials such as photoresists, epoxy resins, and various liquid crystals are employed in the fabrication of VLSIs (very large scale integrated circuits) and LCDs (liquid crystal displays).Impurities in the constituent materials are known to affect the electrical characteristics of these devices adversely. 1–3 The most severe impurities are thought to be radioactive elements, such as thorium and uranium, the presence of results which in so-called soft errors.4 Alkali and alkaline earth elements as well as the anions Cl2 and SO4 22 also degrade device performance because of their high chemical mobility.4 Several analytical methods for the determination of impurities in the above-mentioned organic materials, such as dry ashing methods5,6 and wet digestion methods,7–9 have been reported.The ashing method is a simple technique and gives comparatively accurate values for relatively high concentration levels. Nevertheless, metal losses and sample contamination are likely to occur during chemical treatment.In the wet digestion method, ultratrace concentrations of metals can be obtained reliably, but considerable time and careful sample manipulation are required because of the comparatively long digestion times required and potential contamination from the environment and from sample handling. A knowledge of the chloride ion concentration is also necessary for the evaluation of the effect of impurities on the pertinent electronic characteristics of the above-mentioned devices.However, another technique is required for the measurement of chloride ion concentration because acids such as perchloric acid and hydrochloric acid were used in the analytical methods described in previous papers.10,11 Industrial applications of analytical techniques and procedures often necessitate rapid, accurate analyses of large numbers of samples. As such, the purpose of this investigation was the establishment of sensitive, accurate and rapid determination techniques for the measurement of sodium, potassium, magnesium, calcium and chloride ions in organic materials for microelectronics devices using a digestion method.Experimental Apparatus The digestion vessel for use with a convection oven is illustrated in Fig. 1. The vessel consisted of three nested structures: an innermost PTFE container of 30 ml capacity, an intermediate PTFE container of 100 ml capacity, and an outer stainless-steel shell. For digestion with a microwave oven, an innermost PFA container of 8 ml capacity sealed with a poly(propylene) jacket (designed by Sanai Kagaku, Nagoya, Japan) was used.A conventional microwave oven of 500 W continuous output power was used for microwave digestion. A Perkin-Elmer (Norwalk, CT, USA) Model 5100ZL atomic absorption spectrometer equipped with a Zeeman-effect background correction system was used. Furnace and material parameter values were as listed in Table 1. Regression lines were calculated by plotting peak area as a function of concentration (ng g21).A Dionex (Sunnyvale, CA, USA) 100 ion chromatograph was used for the determination of chloride ion concentration. Instrumental parameters were as listed in Table 2. A Seiko Instruments (Chiba, Japan) Model SPS 4000 inductively coupled plasma atomic emission spectrometer was Fig. 1 Digestion vessel for convection oven. A, Innermost 30 ml PTFE vessel; B, 100 ml PTFE vessel; C, stainless-steel outer shell; D, sample; E, water or dilute acid.Analyst, February 1997, Vol. 122 (129–132) 129used to measure higher concentrations (i.e., ! 1 mg g21) of alkaline earth elements. Reagents and Samples All solutions were prepared using high-purity water. Solutions (1 mg ml21) of the above-mentioned elements (Kanto–Merck, Tokyo, Japan) and chloride ion (Wako Pure Chemicals, Osaka, Japan) were used to prepare the standard solutions used throughout this work. Concentrated nitric acid, hydrochloric acid, perchloric acid (TAMAPURE AA-10, Tama Chemicals, Kawasaki, Japan), and sulfuric acid (TAMAPURE AA-100) were used to digest or extract the impurities.Care was exercised to avoid contamination of samples and solutions by dust, chemicals and handling. Sample manipulation and solution preparation were conducted in a Class 1000-equivalent clean room. Samples such as photoresists, liquid crystals and epoxy resins used in this work were supplied by various manufacturers.Procedure for Digestion With Water or Dilute Acid A 200 mg portion of the organic sample was weighed in the innermost PTFE or PFA vessel. A 5 ml volume of water or dilute acid was then added to the vessel to extract the elements from the sample. Three samples and two blanks were routinely prepared for simultaneous digestion. The five vessels were heated simultaneously for 3 min in the microwave oven. The vessels were then cooled for 15 min and vented for pressure relief after digestion.ETAAS and/or ion chromatography (IC) measurements were performed successively for the five samples after cooling. Procedure for Acid Digestion Approximately 200 mg of the organic sample were weighed in the 100 ml PTFE vessel. Concentrated nitric acid was then added to the vessel to digest the sample at 100 °C on a hot-plate. After 10 min, appropriate amounts of perchloric and sulfuric acids and additional nitric acid were added. The solution was heated to near dryness at 230 °C for 120–300 min on a hot-plate, and then diluted to about 5 ml with water.Procedure for Dry Ashing Approximately 200 mg of the organic sample were weighed in a platinum crucible. The sample was then ashed by gentle continuous heating over a bunsen burner flame until the carbon had been expelled completely. After cooling, 5 ml of 0.1 m hydrochloric acid were added. In cases where a residue was observed, 1 g of sodium carbonate and 0.2 g of boric acid were added to the crucible. The sample was placed in a muffle furnace and fused at 650 °C for 15 min.After cooling, 5 ml of hydrochloric acid were added to the crucible, and the sample was diluted to 100 ml with water. Results and Discussion Microwave Digestion of Sodium and Potassium With Water Sodium and potassium were extracted from photoresists with water using a convection electric oven in previous work.12 The organic samples studied here were similarly analysed with a conventional microwave oven replacing the convection oven.The sodium and potassium concentrations obtained by the convection and microwave oven digestion methods were as shown in Table 3. There was little deviation in the values for the two methods, suggesting that the microwave digestion method is a consistent technique for the determination of ultratrace concentrations of sodium and potassium in the organic materials studied. Virtually complete recovery was also obtained using the standard additions method.Digestion by microwave energy conspicuously shortened the digestion period by a factor of 20; furthermore, the complications of the acid digestion process as described above were obviated. Microwave Digestion of Magnesium and Calcium With Dilute Acid The two digestion methods using water and dilute acid were also applied to the determination of magnesium and calcium. For the procedure using dilute acid, the digests became opaque, suggesting that alkaline earth elements were not extracted completely.The values of the magnesium and calcium concentrations obtained by the aqueous and dilute acid digestion methods were as shown in Table 3. Comparison of the results for the two methods shows that both methods were adequate for the digestion of magnesium and calcium at relatively low concentration levels. In addition, the values of the deviations in their concentrations were also small. However, for comparatively high concentration levels, it can be seen that there was little correlation of the values for the actual concentration levels obtained by the two methods, and that the deviations in the actual concentration values were conspicuously large.Comparison of Microwave Digestion Method With Other Methods Ultratrace level metallic concentration reference standards for electronic devices are not yet available. Therefore, the ashing and acid digestion methods were used to confirm the results of the microwave digestion method.Metallic concentrations in the Table 1 Furnace conditions and instrumental parameters for Zeeman-effect background corrected atomic absorption spectrometry Argon gas Tempera- flow rate/ Step ture/°C Ramp/s Hold/s ml min21 Drying 1 120 1 10 250 Drying 2 130 5 10 250 Ashing 700* 5 10 250 800† 900‡ Atomizing 1600* 0 5 1900† 2400‡ Cleaning 2500 1 2 250 Lamp current/ Slit-width/ Element Wavelength/nm mA nm Na 589.6 8 0.2 K 766.5 12 0.2 Mg 285.2 30 0.4 Ca 422.7 20 0.4 * Na and K.† Mg. ‡ Ca. Table 2 Ion chromatograph parameters Column IonPac AS-4A Effluent 1.8 mm Na2CO3 1.7 mm NaHCO3 Effluent flow rate 1.2 ml min21 Suppressor ASRS Detector Electric conductivity Detector sensitivity 30 ms Injection loop 50 ml 130 Analyst, February 1997, Vol. 122organic materials obtained by the two methods were also as listed in Table 3. The results of the ashing and acid digestion methods for comparatively high metallic concentration values were in good agreement.The high contamination levels obtained using the platinum crucible are thought to have resulted from the platinum crucible itself. Moreover, low level metallic concentration values for the acid digestion method were in good agreement with those obtained by the microwave digestion method. Comparison with the total impurity values shown in Table 3 for the analyses performed using the acid digestion and ashing methods shows that nearly 100% digestion was obtained after digestion for 3 min with the microwave oven.It was also found that the EP-2 and EP-3 samples in Table 3 remained as white precipitates after ashing in the platinum crucible. The precipitate of the EP-3 sample was dissolved easily by 1 ml of concentrated hydrochloric acid, whereas that of the EP-2 sample could not be dissolved by any acid. The alkali fusion method was necessary to decompose the EP-2 sample completely. After decomposition of the EP-2 sample, a high barium concentration (12.8 ± 0.6%) was observed, and is thought to have resulted from the presence of BaSO4.(BaSO4 is typically used as a plasticizer in the manufacture of white plastics.) Consequently, it is considered that the high calcium concentration observed in the EP-2 sample resulted from the high barium plasticizer concentration. A high calcium concentration (4.9 ± 0.1%) was also observed in the EP-3 sample, probably as a consequence of the use of a calcium plasticizer in the manufacture of this sample.It can be concluded that nearly complete microwave digestion was obtained when the deviation in the measured concentration value was comparatively small. Conversely, microwave digestion can be considered to have been incomplete when the measured deviation was large, thus indicating that regardless of their complexities, either of the acid digestion or ashing methods should be used to obtain adequate sample digestion.Digestion of Chloride Ion The digestion method was also applied to the determination of chloride ion impurities in the organic samples. The chloride ion Table 3 Elemental concentrations (mg g21) of impurities in organic materials by several analytical methods* Na Sample Conventional† Microwave‡ Acid digestion Ashing PR-1§ 0.42 ± 0.03 0.46 ± 0.02 0.45 ± 0.01 < 0.5¶ PR-2§ 0.086 ± 0.002 0.088 ± 0.003 0.086 ± 0.004 < 0.5¶ LQ-1· 0.041 ± 0.004 0.038 ± 0.004 0.042 ± 0.002 < 0.5¶ LQ-2· 0.39 ± 0.05 0.38 ± 0.03 0.37 ± 0.03 < 0.5¶ EP-1** 0.12 ± 0.03 0.15 ± 0.03 0.15 ± 0.01 < 0.5¶ EP-2** 2.1 ± 0.3 1.8 ± 0.4 1.8 ± 0.3 NM†† EP-3** 3.1 ± 0.2 2.9 ± 0.2 2.8 ± 0.5 2.5 ± 0.3 K Sample Conventional† Microwave‡ Acid digestion Ashing PR-1§ 0.21 ± 0.02 0.23 ± 0.02 0.23 ± 0.01 < 0.5¶ PR-2§ 0.055 ± 0.003 0.059 ± 0.001 0.059 ± 0.005 < 0.5¶ LQ-1· 0.053 ± 0.002 0.055 ± 0.003 0.056 ± 0.003 < 0.5¶ LQ-2· 0.24 ± 0.03 0.23 ± 0.02 0.23 ± 0.02 < 0.5¶ EP-1** 0.018 ± 0.003 0.017 ± 0.001 0.018 ± 0.002 < 0.5¶ EP-2** 1.0 ± 0.3 1.2 ± 0.2 1.1 ± 0.2 NM†† EP-3** 2.1 ± 0.2 2.3 ± 0.3 2.4 ± 0.1 2.2 ± 0.3 Mg Sample Water‡‡ Dilute acid§Ø Acid digestion Ashing PR-1§ 0.12 ± 0.03 0.15 ± 0.04 0.12 ± 0.02 < 0.3¶ PR-2§ 0.078 ± 0.004 0.077 ± 0.004 0.083 ± 0.002 < 0.3¶ LQ-1· 0.032 ± 0.003 0.039 ± 0.002 0.036 ± 0.003 < 0.3¶ LQ-2· 0.19 ± 0.02 0.25 ± 0.02 0.23 ± 0.01 < 0.3¶ EP-1** 0.082 ± 0.003 0.079 ± 0.002 0.087 ± 0.002 < 0.3¶ EP-2** 0.05 ± 0.04 (%) 0.12 ± 0.07 (%) 0.13 ± 0.03 (%) 2.1 ± 0.1 (%) EP-3** 5.8 ± 4.5 9.5 ± 3.7 15 ± 1 15 ± 2 Ca Sample Water‡‡ Dilute acid§Ø Acid digestion Ashing PR-1§ 0.20 ± 0.03 0.19 ± 0.01 0.18 ± 0.02 < 0.3¶ PR-2§ 0.052 ± 0.004 0.059 ± 0.005 0.053 ± 0.004 < 0.3¶ LQ-1· 0.022 ± 0.002 0.023 ± 0.003 0.022 ± 0.003 < 0.3¶ LQ-2· 0.18 ± 0.03 0.23 ± 0.01 0.18 ± 0.02 < 0.3¶ EP-1** 0.13 ± 0.03 0.12 ± 0.02 0.12 ± 0.01 < 0.3¶ EP-2** 0.03 ± 0.02 (%) 0.04 ± 0.04 (%) 0.05 ± 0.03 (%) 0.10 ± 0.01 (%) EP-3** 0.16 ± 0.12 (%) 1.5 ± 1.3 (%) 4.9 ± 0.1 (%) 4.9 ± 0.1 (%) * Number of samples = 3; ± value is for a single standard deviation; all values listed are blank-corrected.† Convection oven digestion method (60 min); solution is 5 ml of water. ‡ Microwave oven digestion method (3 min); solution is 5 ml of water. § Photoresist samples were supplied by two manufacturers. ¶ < 0.3 and < 0.5 denote values below detection limits. · Liquid crystal samples were supplied by two manufacturers.** Epoxy resin samples were supplied by three manufacturers. †† Because the alkali fusion method was employed, these values could not be measured. ‡‡ Microwave oven digestion method (3 min); solution is 5 ml of water. §§ Microwave oven digestion method (3 min); solution is 5 ml of 0.1 m hydrochloric acid. Analyst, February 1997, Vol. 122 131concentrations obtained by the aqueous digestion method were as shown in Table 4, and were compared with those obtained by the alkali fusion method.13 It can be seen that the values for the two methods were in close agreement. In addition, the values of the deviations in chloride ion concentration showed little variation, indicating that microwave digestion is a consistent technique for the determination of chloride ion concentration as well as those of sodium and potassium.It was found that complete chloride ion digestion was obtained after 3 min. A notable advantage of the proposed digestion method is that the chloride ion concentration can be measured using the same sample used for the measurement of alkali and alkaline earth element concentrations.Blank Levels The blank levels obtained for the three procedures described above were as listed in Table 5. Except for those of the ashing method, minimal values of less than 1 ng per analysis were achieved. The blank levels obtained by the microwave digestion method developed here are the lowest of the three methods. Conclusions The simple microwave digestion method reported here is suitable for the routine determination of the relevant trace impurities of sodium, potassium, calcium and magnesium in organic materials for microelectronics devices at the ng g21 level and chloride ion at the 10 ng g21 level.The time required for analysis was only one-twentieth of that required for digestion with a convection oven. Additionally, the probability of contamination was reduced compared with that inherent in the acid digestion and ashing methods.The proposed method also permits chloride ion determination using the same sample solution as used for the above-mentioned cations. References 1 Neuhaus, H. J., Day, D. R., and Senturia, S. D., J. Electron. Mater., 1985, 14, 379. 2 Gercken, B., Pavel, J., and Reus, G., presented at the 12th International Symposium on Microchemical Techniques, Cordoba, Spain, 1992. 3 Sacher, E., IEEE Trans. Electr.Insul., 1983, EI-18, 369. 4 May, T. C., and Woods, M. H., IEEE Trans. Electr. Devices, 1979, ED-26, 2. 5 Basson, W. D., and B�ohmer, R. G., Analyst, 1972, 97, 482. 6 Koirtyohann, S. R., and Hopkins, C. A., Analyst, 1976, 101, 870. 7 Boer, J. L., and Maessen, F. J., Spectrochim. Acta, Part B, 1983, 38, 739. 8 Friel, J. K., Skinner, C. S., Jackson, S. E., and Longerich, H. P., Analyst, 1990, 115, 269. 9 Pratt, K. W., Kingston, H. M., MacCrehan, W. A., and Koch, W. F., Anal.Chem., 1988, 60, 2024. 10 Bettinelli, M., Baroni, U., and Pastorelli, N., Anal. Chim. Acta, 1989, 225, 159. 11 Fernando, L. A., Heavner, W. D., and Gabrielli, C. C., Anal. Chem., 1986, 58, 511. 12 Takenaka, M., Kozuka, S., and Hashimoto, Y., Bunseki Kagaku, 1993, 42, 71. 13 Hashimoto, Y., Nikkei Microdevices, 1992, 1, 82. Paper 6/04802A Received July 9, 1996 Accepted November 7, 1996 Table 4 Chloride ion concentration in organic materials (mg g21)* Cl Sample Conventional† Microwave‡ Alkali fusion PR-1§ 0.25 ± 0.03 0.27 ± 0.02 0.28 ± 0.05 PR-2§ 4.2 ± 0.2 4.5 ± 0.3 4.3 ± 0.2 LQ-1¶ < 0.02 < 0.02 < 0.1 LQ-2¶ < 0.02 < 0.02 < 0.1 EP-1· 38 ± 5 35 ± 6 40 ± 2 EP-2· 51 ± 3 52 ± 5 55 ± 10 EP-3· 110 ± 20 120 ± 10 120 ± 5 * Number of samples = 3; ± value is for a single standard deviation.† Convection oven digestion method (60 min). ‡ Microwave oven digestion method (3 min); extract was 5 ml of water. § Photoresist samples supplied by two manufacturers.¶ Liquid crystal samples supplied by two manufacturers. · Epoxy resin samples supplied by three manufacturers. Table 5 Blank values found for each procedure (ng per 5 ml per analysis)* Element Digestion† Acid digestion Dry ashing Na 0.05 ± 0.01 0.12 ± 0.03 530 ± 30 K 0.02 ± 0.01 0.05 ± 0.02 510 ± 20 Ca 0.03 ± 0.01 0.13 ± 0.04 320 ± 10 Mg 0.02 ± 0.01 0.17 ± 0.02 280 ± 10 Cl 20 ± 3 NM‡ 110 ± 10§ * Number of samples = 3; ± value is for a single standard deviation. † Aqueous digestion method using microwave oven (3 min); extract was 5 ml of water.‡ NM = Not measured. § This value was obtained by the alkali fusion method. 132 Analyst, February 1997, Vol. 122 Determination of Ultratrace Amounts of Metallic and Chloride Ion Impurities in Organic Materials for Microelectronics Devices After a Microwave Digestion Method Miyuki Takenaka*a, Shoji Kozukaa, Masaru Hayashia and Hiroshi Endob a Research and Development Center, Toshiba Corporation, 1, Komukai, Toshiba-cho, Saiwai-ku, Kawasaki 210, Japan b Toshiba Research and Consulting Corporation, 1, Komukai, Toshiba-cho, Saiwai-ku, Kawasaki 210, Japan A digestion method was developed for the determination of ultratrace concentrations of sodium, potassium, magnesium, calcium and chloride ions.Using ETAAS and ion chromatography, the digestion method was applied successfully to the determination of ultratrace concentrations of these elements in organic materials for microelectronics devices such as photoresists, epoxy resins, and liquid crystals. Very low contamination levels were maintained throughout the procedure.The blank levels were 0.05 ng for sodium, 0.02 ng for potassium, 0.03 ng for magnesium and calcium, and 20 ng for chloride ion. The method is very effective in measuring the impurity distributions of organic materials whilst preventing their contamination from the surrounding environment and from the reagents used in the procedure.Keywords: Atomic absorption spectrometry; ion chromatography; metallic impurity; chloride ion; microwave digestion; alkali fusion; photoresist; epoxy resin; liquid crystal Organic materials such as photoresists, epoxy resins, and various liquid crystals are employed in the fabrication of VLSIs (very large scale integrated circuits) and LCDs (liquid crystal displays). Impurities in the constituent materials are known to affect the electrical characteristics of these devices adversely. 1–3 The most severe impurities are thought to be radioactive elements, such as thorium and uranium, the presence of results which in so-called soft errors.4 Alkali and alkaline earth elements as well as the anions Cl2 and SO4 22 also degrade device performance because of their high chemical mobility.4 Several analytical methods for the determination of impurities in the above-mentioned organic materials, such as dry ashing methods5,6 and wet digestion methods,7–9 have been reported.The ashing method is a simple technique and gives comparatively accurate values for relatively high concentration levels. Nevertheless, metal losses and sample contamination are likely to occur during chemical treatment. In the wet digestion method, ultratrace concentrations of metals can be obtained reliably, but considerable time and careful sample manipulation are required because of the comparatively long digestion times required and potential contamination from the environment and from sample handling.A knowledge of the chloride ion concentration is also necessary for the evaluation of the effect of impurities on the pertinent electronic characteristics of the above-mentioned devices. However, another technique is required for the measurement of chloride ion concentration because acids such as perchloric acid and hydrochloric acid were used in the analytical methods described in previous papers.10,11 Industrial applications of analytical techniques and procedures often necessitate rapid, accurate analyses of large numbers of samples.As such, the purpose of this investigation was the establishment of sensitive, accurate and rapid determination techniques for the measurement of sodium, potassium, magnesium, calcium and chloride ions in organic materials for microelectronics devices using a digestion method. Experimental Apparatus The digestion vessel for use with a convection oven is illustrated in Fig. 1. The vessel consisted of three nested structures: an innermost PTFE container of0 ml capacity, an intermediate PTFE container of 100 ml capacity, and an outer stainless-steel shell. For digestion with a microwave oven, an innermost PFA container of 8 ml capacity sealed with a poly(propylene) jacket (designed by Sanai Kagaku, Nagoya, Japan) was used. A conventional microwave oven of 500 W continuous output power was used for microwave digestion. A Perkin-Elmer (Norwalk, CT, USA) Model 5100ZL atomic absorption spectrometer equipped with a Zeeman-effect background correction system was used.Furnace and material parameter values were as listed in Table 1. Regression lines were calculated by plotting peak area as a function of concentration (ng g21). A Dionex (Sunnyvale, CA, USA) 100 ion chromatograph was used for the determination of chloride ion concentration. Instrumental parameters were as listed in Table 2. A Seiko Instruments (Chiba, Japan) Model SPS 4000 inductively coupled plasma atomic emission spectrometer was Fig. 1 Digestion vessel for convection oven. A, Innermost 30 ml PTFE vessel; B, 100 ml PTFE vessel; C, stainless-steel outer shell; D, sample; E, water or dilute acid. Analyst, February 1997, Vol. 122 (129–132) 129used to measure higher concentrations (i.e., ! 1 mg g21) of alkaline earth elements. Reagents and Samples All solutions were prepared using high-purity water. Solutions (1 mg ml21) of the above-mentioned elements (Kanto–Merck, Tokyo, Japan) and chloride ion (Wako Pure Chemicals, Osaka, Japan) were used to prepare the standard solutions used throughout this work.Concentrated nitric acid, hydrochloric acid, perchloric acid (TAMAPURE AA-10, Tama Chemicals, Kawasaki, Japan), and sulfuric acid (TAMAPURE AA-100) were used to digest or extract the impurities. Care was exercised to avoid contamination of samples and solutions by dust, chemicals and handling. Sample manipulation and solution preparation were conducted in a Class 1000-equivalent clean room.Samples such as photoresists, liquid crystals and epoxy resins used in this work were supplied by various manufacturers. Procedure for Digestion With Water or Dilute Acid A 200 mg portion of the organic sample was weighed in the innermost PTFE or PFA vessel. A 5 ml volume of water or dilute acid was then added to the vessel to extract the elements from the sample. Three samples and two blanks were routinely prepared for simultaneous digestion.The five vessels were heated simultaneously for 3 min in the microwave oven. The vessels were then cooled for 15 min and vented for pressure relief after digestion. ETAAS and/or ion chromatography (IC) measurements were performed successively for the five samples after cooling. Procedure for Acid Digestion Approximately 200 mg of the organic sample were weighed in the 100 ml PTFE vessel. Concentrated nitric acid was then added to the vessel to digest the sample at 100 °C on a hot-plate. After 10 min, appropriate amounts of perchloric and sulfuric acids and additional nitric acid were added.The solution was heated to near dryness at 230 °C for 120–300 min on a hot-plate, and then diluted to about 5 ml with water. Procedure for Dry Ashing Approximately 200 mg of the organic sample were weighed in a platinum crucible. The sample was then ashed by gentle continuous heating over a bunsen burner flame until the carbon had been expelled completely.After cooling, 5 ml of 0.1 m hydrochloric acid were added. In cases where a residue was observed, 1 g of sodium carbonate and 0.2 g of boric acid were added to the crucible. The sample was placed in a muffle furnace and fused at 650 °C for 15 min. After cooling, 5 ml of hydrochloric acid were added to the crucible, and the sample was diluted to 100 ml with water. Results and Discussion Microwave Digestion of Sodium and Potassium With Water Sodium and potassium were extracted from photoresists with water using a convection electric oven in previous work.12 The organic samples studied here were similarly analysed with a conventional microwave oven replacing the convection oven.The sodium and potassium concentrations obtained by the convection and microwave oven digestion methods were as shown in Table 3. There was little deviation in the values for the two methods, suggesting that the microwave digestion method is a consistent technique for the determination of ultratrace concentrations of sodium and potassium in the organic materials studied.Virtually complete recovery was also obtained using the standard additions method. Digestion by microwave energy conspicuously shortened the digestion period by a factor of 20; furthermore, the complications of the acid digestion process as described above were obviated. Microwave Digestion of Magnesium and Calcium With Dilute Acid The two digestion methods using water and dilute acid were also applied to the determination of magnesium and calcium.For the procedure using dilute acid, the digests became opaque, suggesting that alkaline earth elements were not extracted completely. The values of the magnesium and calcium concentrations obtained by the aqueous and dilute acid digestion methods were as shown in Table 3. Comparison of the results for the two methods shows that both methods were adequate for the digestion of magnesium and calcium at relatively low concentration levels. In addition, the values of the deviations in their concentrations were also small. However, for comparatively high concentration levels, it can be seen that there was little correlation of the values for the actual concentration levels obtained by the two methods, and that the deviations in the actual concentration values were conspicuously large.Comparison of Microwave Digestion Method With Other Methods Ultratrace level metallic concentration reference standards for electronic devices are not yet available.Therefore, the ashing and acid digestion methods were used to confirm the results of the microwave digestion method. Metallic concentrations in the Table 1 Furnace conditions and instrumental parameters for Zeeman-effect background corrected atomic absorption spectrometry Argon gas Tempera- flow rate/ Step ture/°C Ramp/s Hold/s ml min21 Drying 1 120 1 10 250 Drying 2 130 5 10 250 Ashing 700* 5 10 250 800† 900‡ Atomizing 1600* 0 5 1900† 2400‡ Cleaning 2500 1 2 250 Lamp current/ Slit-width/ Element Wavelength/nm mA nm Na 589.6 8 0.2 K 766.5 12 0.2 Mg 285.2 30 0.4 Ca 422.7 20 0.4 * Na and K.† Mg. ‡ Ca. Table 2 Ion chromatograph parameters Column IonPac AS-4A Effluent 1.8 mm Na2CO3 1.7 mm NaHCO3 Effluent flow rate 1.2 ml min21 Suppressor ASRS Detector Electric conductivity Detector sensitivity 30 ms Injection loop 50 ml 130 Analyst, February 1997, Vol. 122organic materials obtained by the two methods were also as listed in Table 3. The results of the ashing and acid digestion methods for comparatively high metallic concentration values were in good agreement. The high contamination levels obtained using the platinum crucible are thought to have resulted from the platinum crucible itself. Moreover, low level metallic concentration values for the acid digestion method were in good agreement with those obtained by the microwave digestion method.Comparison with the total impurity values shown in Table 3 for the analyses performed using the acid digestion and ashing methods shows that nearly 100% digestion was obtained after digestion for 3 min with the microwave oven. It was also found that the EP-2 and EP-3 samples in Table 3 remained as white precipitates after ashing in the platinum crucible. The precipitate of the EP-3 sample was dissolved easily by 1 ml of concentrated hydrochloric acid, whereas that of the EP-2 sample could not be dissolved by any acid.The alkali fusion method was necessary to decompose the EP-2 sample completely. After decomposition of the EP-2 sample, a high barium concentration (12.8 ± 0.6%) was observed, and is thought to have resulted from the presence of BaSO4. (BaSO4 is typically used as a plasticizer in the manufacture of white plastics.) Consequently, it is considered that the high calcium concentration observed in the EP-2 sample resulted from the high barium plasticizer concentration.A high calcium concentration (4.9 ± 0.1%) was also observed in the EP-3 sample, probably as a consequence of the use of a calcium plasticizer in the manufacture of this sample. It can be concluded that nearly complete microwave digestion was obtained when the deviation in the measured concentration value was comparatively small. Conversely, microwave digestion can be considered to have been incomplete when the measured deviation was large, thus indicating that regardless of their complexities, either of the acid digestion or ashing methods should be used to obtain adequate sample digestion.Digestion of Chloride Ion The digestion method was also applied to the determination of chloride ion impurities in the organic samples. The chloride ion Table 3 Elemental concentrations (mg g21) of impurities in organic materials by several analytical methods* Na Sample Conventional† Microwave‡ Acid digestion Ashing PR-1§ 0.42 ± 0.03 0.46 ± 0.02 0.45 ± 0.01 < 0.5¶ PR-2§ 0.086 ± 0.002 0.088 ± 0.003 0.086 ± 0.004 < 0.5¶ LQ-1· 0.041 ± 0.004 0.038 ± 0.004 0.042 ± 0.002 < 0.5¶ LQ-2· 0.39 ± 0.05 0.38 ± 0.03 0.37 ± 0.03 < 0.5¶ EP-1** 0.12 ± 0.03 0.15 ± 0.03 0.15 ± 0.01 < 0.5¶ EP-2** 2.1 ± 0.3 1.8 ± 0.4 1.8 ± 0.3 NM†† EP-3** 3.1 ± 0.2 2.9 ± 0.2 2.8 ± 0.5 2.5 ± 0.3 K Sample Conventional† Microwave‡ Acid digestion Ashing PR-1§ 0.21 ± 0.02 0.23 ± 0.02 0.23 ± 0.01 < 0.5¶ PR-2§ 0.055 ± 0.003 0.059 ± 0.001 0.059 ± 0.005 < 0.5¶ LQ-1· 0.053 ± 0.002 0.055 ± 0.003 0.056 ± 0.003 < 0.5¶ LQ-2· 0.24 ± 0.03 0.23 ± 0.02 0.23 ± 0.02 < 0.5¶ EP-1** 0.018 ± 0.003 0.017 ± 0.001 0.018 ± 0.002 < 0.5¶ EP-2** 1.0 ± 0.3 1.2 ± 0.2 1.1 ± 0.2 NM†† EP-3** 2.1 ± 0.2 2.3 ± 0.3 2.4 ± 0.1 2.2 ± 0.3 Mg Sample Water‡‡ Dilute acid§Ø Acid digestion Ashing PR-1§ 0.12 ± 0.03 0.15 ± 0.04 0.12 ± 0.02 < 0.3¶ PR-2§ 0.078 ± 0.004 0.077 ± 0.004 0.083 ± 0.002 < 0.3¶ LQ-1· 0.032 ± 0.003 0.039 ± 0.002 0.036 ± 0.003 < 0.3¶ LQ-2· 0.19 ± 0.02 0.25 ± 0.02 0.23 ± 0.01 < 0.3¶ EP-1** 0.082 ± 0.003 0.079 ± 0.002 0.087 ± 0.002 < 0.3¶ EP-2** 0.05 ± 0.04 (%) 0.12 ± 0.07 (%) 0.13 ± 0.03 (%) 2.1 ± 0.1 (%) EP-3** 5.8 ± 4.5 9.5 ± 3.7 15 ± 1 15 ± 2 Ca Sample Water‡‡ Dilute acid§Ø Acid digestion Ashing PR-1§ 0.20 ± 0.03 0.19 ± 0.01 0.18 ± 0.02 < 0.3¶ PR-2§ 0.052 ± 0.004 0.059 ± 0.005 0.053 ± 0.004 < 0.3¶ LQ-1· 0.022 ± 0.002 0.023 ± 0.003 0.022 ± 0.003 < 0.3¶ LQ-2· 0.18 ± 0.03 0.23 ± 0.01 0.18 ± 0.02 < 0.3¶ EP-1** 0.13 ± 0.03 0.12 ± 0.02 0.12 ± 0.01 < 0.3¶ EP-2** 0.03 ± 0.02 (%) 0.04 ± 0.04 (%) 0.05 ± 0.03 (%) 0.10 ± 0.01 (%) EP-3** 0.16 ± 0.12 (%) 1.5 ± 1.3 (%) 4.9 ± 0.1 (%) 4.9 ± 0.1 (%) * Number of samples = 3; ± value is for a single standard deviation; all values listed are blank-corrected.† Convection oven digestion method (60 min); solution is 5 ml of water. ‡ Microwave oven digestion method (3 min); solution is 5 ml of water.§ Photoresist samples were supplied by two manufacturers. ¶ < 0.3 and < 0.5 denote values below detection limits. · Liquid crystal samples were supplied by two manufacturers. ** Epoxy resin samples were supplied by three manufacturers. †† Because the alkali fusion method was employed, these values could not be measured. ‡‡ Microwave oven digestion method (3 min); solution is 5 ml of water. §§ Microwave oven digestion method (3 min); solution is 5 ml of 0.1 m hydrochloric acid.Analyst, February 1997, Vol. 122 131concentrations obtained by the aqueous digestion method were as shown in Table 4, and were compared with those obtained by the alkali fusion method.13 It can be seen that the values for the two methods were in close agreement. In addition, the values of the deviations in chloride ion concentration showed little variation, indicating that microwave digestion is a consistent technique for the determination of chloride ion concentration as well as those of sodium and potassium.It was found that complete chloride ion digestion was obtained after 3 min. A notable advantage of the proposed digestion method is that the chloride ion concentration can be measured using the same sample used for the measurement of alkali and alkaline earth element concentrations. Blank Levels The blank levels obtained for the three procedures described above were as listed in Table 5.Except for those of the ashing method, minimal values of less than 1 ng per analysis were achieved. The blank levels obtained by the microwave digestion method developed here are the lowest of the three methods. Conclusions The simple microwave digestion method reported here is suitable for the routine determination of the relevant trace impurities of sodium, potassium, calcium and magnesium in organic materials for microelectronics devices at the ng g21 level and chloride ion at the 10 ng g21 level.The time required for analysis was only one-twentieth of that required for digestion with a convection oven. Additionally, the probability of contamination was reduced compared with that inherent in the acid digestion and ashing methods. The proposed method also permits chloride ion determination using the same sample solution as used for the above-mentioned cations. References 1 Neuhaus, H. J., Day, D. R., and Senturia, S. D., J. Electron. Mater., 1985, 14, 379. 2 Gercken, B., Pavel, J., and Reus, G., presented at the 12th International Symposium on Microchemical Techniques, Cordoba, Spain, 1992. 3 Sacher, E., IEEE Trans. Electr. Insul., 1983, EI-18, 369. 4 May, T. C., and Woods, M. H., IEEE Trans. Electr. Devices, 1979, ED-26, 2. 5 Basson, W. D., and B�ohmer, R. G., Analyst, 1972, 97, 482. 6 Koirtyohann, S. R., and Hopkins, C. A., Analyst, 1976, 101, 870. 7 Boer, J. L., and Maessen, F. J., Spectrochim. Acta, Part B, 1983, 38, 739. 8 Friel, J. K., Skinner, C. S., Jackson, S. E., and Longerich, H. P., Analyst, 1990, 115, 269. 9 Pratt, K. W., Kingston, H. M., MacCrehan, W. A., and Koch, W. F., Anal. Chem., 1988, 60, 2024. 10 Bettinelli, M., Baroni, U., and Pastorelli, N., Anal. Chim. Acta, 1989, 225, 159. 11 Fernando, L. A., Heavner, W. D., and Gabrielli, C. C., Anal. Chem., 1986, 58, 511. 12 Takenaka, M., Kozuka, S., and Hashimoto, Y., Bunseki Kagaku, 1993, 42, 71. 13 Hashimoto, Y., Nikkei Microdevices, 1992, 1, 82. Paper 6/04802A Received July 9, 1996 Accepted November 7, 1996 Table 4 Chloride ion concentration in organic materials (mg g21)* Cl Sample Conventional† Microwave‡ Alkali fusion PR-1§ 0.25 ± 0.03 0.27 ± 0.02 0.28 ± 0.05 PR-2§ 4.2 ± 0.2 4.5 ± 0.3 4.3 ± 0.2 LQ-1¶ < 0.02 < 0.02 < 0.1 LQ-2¶ < 0.02 < 0.02 < 0.1 EP-1· 38 ± 5 35 ± 6 40 ± 2 EP-2· 51 ± 3 52 ± 5 55 ± 10 EP-3· 110 ± 20 120 ± 10 120 ± 5 * Number of samples = 3; ± value is for a single standard deviation. † Convection oven digestion method (60 min). ‡ Microwave oven digestion method (3 min); extract was 5 ml of water. § Photoresist samples supplied by two manufacturers. ¶ Liquid crystal samples supplied by two manufacturers. · Epoxy resin samples supplied by three manufacrers. Table 5 Blank values found for each procedure (ng per 5 ml per analysis)* Element Digestion† Acid digestion Dry ashing Na 0.05 ± 0.01 0.12 ± 0.03 530 ± 30 K 0.02 ± 0.01 0.05 ± 0.02 510 ± 20 Ca 0.03 ± 0.01 0.13 ± 0.04 320 ± 10 Mg 0.02 ± 0.01 0.17 ± 0.02 280 ± 10 Cl 20 ± 3 NM‡ 110 ± 10§ * Number of samples = 3; ± value is for a single standard deviation. † Aqueous digestion method using microwave oven (3 min); extract was 5 ml of water. ‡ NM = Not measured. § This value was obtained by the alkali fusion method. 132 Analyst, February 1997, Vol. 122

ISSN:0003-2654    

DOI:10.1039/a604802a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

Optimization of a Microwave-assisted Extraction Method for Phenol and Methylphenol Isomers in Soil Samples Using a Central Composite Design Mar�ýa P. Llomparta, Rosa A. Lorenzoa, Rafael Cela*a and J. R. Jocelyn Par�eb a Departamento de Qu�ýmica Anal�ýtica, Nutrici�on y Bromatolog�ýa, Facultad de Qu�ýmica, Universidad de Santiago de Compostela, 15706 Santiago de Compostela, Spain. E-mail: qnrctd@usc.es b Environment Canada, Environmental Technology Centre, Ottawa, Ontario, Canada, K1A 0H3 A simple and rapid microwave-assisted extraction (MAE) procedure was developed and optimized for phenol, o-cresol, m-cresol and p-cresol in soil samples.The spiked soil used was prepared 25 d before treatment to simulate weathering processes and to allow for the formation of analyte-matrix interactions. The samples, immersed in hexane–acetone, were irradiated with microwaves in a closed-vessel system. Optimization of the method was achieved by using a factorial design approach on parameters such as the volume of solvent, the percentage of acetone and the extraction temperature.The MAE procedure yielded extracts that could be analyzed directly using a GC–flame ionization detection system without any preliminary clean-up or concentration steps. Keywords: Microwave extraction; phenol; cresols; factorial designs; soil samples The first applications of microwave-assisted extraction (MAE) to be reported dealt with the extraction of flavors from plant material,1–3 although it can be applied to the extraction of a variety of chemicals from a wide range of matrices such as soils, plant and animal tissues and a variety of manufactured products.1–9 In general, the compounds can be extracted more selectively and more quickly with similar or better recoveries in comparison with conventional extraction processes.Also, MAE uses less solvent and energy than conventional techniques.3,4 In the last 2 years the number of environmental applications of this technique has increased rapidly.Onuska and Terry extracted organochlorinated pesticides from sediments10 and PCBs from water.11 Fish and Revesz12 extracted organochlorinated pesticides from soil. Hasty and Revesz13 applied this technique to the extraction of petroleum hydrocarbons from soil. Lopez-Avila and co-workers14–16 used microwave energy to extract several groups of pollutants such as PAHs, PCBs, pesticides, phenolics and base/neutral compounds in soils and sediments.In all these studies, MAE proved to be similar to or more efficient than Soxhlet and sonication methods. Phenol and cresols are constituents of crude oil and coal tar, also being used in the chemical industry. It is well known that these compounds exhibit properties that are hazardous to human health,17,18 thus making it necessary to identify the occurrence and levels of contamination in the environment, especially in soils. In this study, we used MAE for the extraction of phenol and the three methylphenols (cresols) from soils in a closedvessel system.The solvent employed was hexane–acetone. The spiked soil used in the optimization process was prepared 25 d before analysis. Parameters studied included volume of solvent, percentage of acetone and extraction temperature. The soil samples were treated by this procedure using the optimum conditions obtained by applying a central composite design. Experimental Microwave-assisted Extraction MAE experiments were performed with a 950-W MES-1000 microwave solvent extraction system (CEM, Matthews, NC, USA).This extractor has provision for 12 simultaneous extractions. A 1–5 g aliquot of soil was accurately weighed into a Teflon-lined extraction vessel and 10–50 ml of acetone– hexane were added to each sample in the proportions dictated by the experimental design. The extraction vessels were closed after ensuring that a new rupture membrane was used for each extraction.For this study, 1–6 simultaneous extractions were performed using full power in less than 15 min at an optimized temperature of 130 °C. When the irradiation period was completed, the sample carousel was removed from the microwave cavity and cooled in a water-bath. The control vessel was returned to the microwave system to check that the extract was at room temperature before opening. Solvent losses were checked in randomly selected experiments and were found to be < 1%.Using a glass pipette, 1 ml of the clear supernatant was transferred into an injection vial and the raw extract was analyzed by GC–flame ionization detection (FID) without any preliminary clean-up or concentration procedure. Fig. 1 shows the chromatogram for a spiked soil sample. Ultrasonic Bath Extraction Extractions using an ultrasonic bath (Selecta, Barcelona, Spain) were performed using 2 g portions of soil. The soil samples were sonicated for 60 min with continuous power under identical chemical conditions, namely 8 ml of acetone and 2 ml of hexane.The clear supernatant extracts were analyzed without any clean up or concentration steps. Fig. 1 Chromatogram of a spiked soil sample showing the resolution obtained using the operating conditions given in Table 1. Analyst, February 1997, Vol. 122 (133–137) 133Reagents and Chemicals Phenol standards were supplied by Aldrich Chemie (Steinheim, Germany). Methanol, hexane and acetone were purchased from Romil Chemicals (Cambridge, UK).Phenol stock standard solutions (1 g l21) were prepared by weighing an appropriate amount of the standard and dissolved in 10 ml of hexane. Working standard solutions were prepared by appropriate dilution of the stock standard solutions. All the solutions were stored at 5 °C in the dark when not in use. For quantitative GC determinations, calibration was carried out at four concentration levels spanning the range 1–10 mg ml21.Optimization experiments were performed using a spiked garden soil sample obtained from the campus of Santiago de Compostela University (Galicia, Spain), the carbon content of which was 2%. The soil was dried in an oven at 104 °C for 48 h, ground and sifted to a particle size below 300 mm. A 300 g aliquot was slurried with 250 ml of methanol solution of phenols. The sample was then allowed to air-dry with occasional mixing at ambient temperature, protected from draughts, for 5 d.The soil was then bottled and stored in a dry, dark location for 20 d before the first extractions. The concentrations in the soil, on the basis of added amounts, were 10.84, 14.23, 12.17 and 12.10 mg g21 for phenol, o-cresol, mcresol and p-cresol, respectively. It was also assumed that the contaminants were uniformly distributed in the sample and that, because the soil still retained residual moisture throughout the storage period, any analyte–matrix interactions would have occurred, over the weathering period, to a similar extent to those in real contaminated soil with similar properties. Analysis Extracts were analyzed on a Hewlett-Packard (Avondale, PA, USA) HP5890 Series II gas chromatograph equipped with a flame ionization detector and a Hewlett-Packard Model 7673A autosampler.A 60 m 3 0.56 mm id, 0.2 mm phase thickness, fused-silica chromatographic column coated with diisodecyl phthalate (DIIDP) (Restek, Bellefonte, PA, USA) was used, which allows good resolution of the three cresols.Chromatographic data were acquired and processed with a Hewlett- Packard Model 3365A data station. Table 1 summarizes the chromatographic conditions used. Safety Considerations Microwave-assisted processes are simple and can be readily understood in terms of the operating steps to be performed. However, the application of microwave energy to flammable organic solvents can pose serious hazards in inexperienced hands.Thus extreme safety precautions and great attention to details when planning and performing experiments must be used by all personnel dealing with microwaves. The authors urge all readers to ensure that they seek proper information from knowledgeable sources and that they do not attempt to implement these techniques unless proper guidance is provided. Only approved eqientifically sound procedures should be used at all times. Results and Discussion Optimization of the Procedure for Phenol and Methylphenols by Gas Chromatography The chromatographic conditions were optimized for the resolution of the four phenols considered, as summarized in Table 1.By using a DIIDP chromatographic column an adequate resolution of the methyl phenol isomers is obtained (Fig. 1). Other stationary phases do not provide resolution between m-cresol and p-cresol. As noted under Experimental, calibration curves were constructed at four concentration levels using appropriately diluted standards.Each concentration level was injected in triplicate. Chromatographic peak areas were fitted by linear regression and the results are given in Table 2. The repeatability of the chromatographic procedure was assessed by performing eight consecutive injections of a standard solution containing the four analytes. The results (between-injection repeatability data) are also given in Table 2 along with detection and quantification limits for direct injections of standards at signal-to-noise ratios of three and ten, respectively.Evaluation of the Homogeneity of Laboratory-spiked Soil Samples In order to test the homogeneity of the spiked sample to be used in optimizing the extraction process, a set of six 5 g extractions were performed under the conditions given in the second column of Table 3. The average recoveries obtained were 77.2, 67.7, 70.3 and 59.9% for phenol, o-, m- and p-cresol, respectively. The variability (RSD 7.3–9.6%) was very close to that obtained from the injection of calibration standards alone (Table 2).In conclusion, the material was considered to be homogeneous as far as the target analytes are concerned, taking portions of at least 5 g. Some blank samples were extracted under the same conditions (Table 3, second column). The results showed the absence of an analytical signal at the retention times of the compounds studied. Factorial Design. Evaluation of the Response Surfaces The variables considered in the MAE optimization process were temperature, volume of solvent and percentage of acetone.A Table 1 GC operating conditions Injection port temperature 125 °C Injection mode Splitless Injection volume 2 ml Splitless time 60 s Column 60 m30.56 mm id 0.2 mm film thickness, DIIDP (Restek) Carrier gas Nitrogen (99.9995%) Carrier gas flow rate 5.8 ml min21 Carrier gas pressure at column head 50 kPa Oven temperature 100 °C FID temperature 150 °C Table 2 Calibration and statistical validation parameters Parameter Phenol o-cresol m-cresol p-cresol Calibration range/ mg ml21 0.98–7.82 1.28–10.27 1.10–8.78 1.09–8.74 Correlation coefficient 0.9999 0.9999 0.9999 0.9999 Detection limit/ ng ml21 (S/N = 3) 20.2 18.3 23.7 24.0 Quantification limit/ng ml21 (S/N = 10) 69.2 61.0 79.1 79.9 Between-injection RSD (%) 3.8 5.5 6.2 6.0 134 Analyst, February 1997, Vol. 122three-level central 23 + star orthogonal composite design involving 14 runs and three central points was chosen.This model allows the direct evaluation of the variables considered, and also the first- and second-order interaction terms. The low and high levels assigned to the variables regarding the value assigned to the fixed factors are listed in the third column of Table 3. The data analysis was performed using the statistical package Statgraphics Plus V.6.0.19 Table 4 gives the design matrix for this experiment and the recoveries obtained in each run.An analysis of the results given in Table 4 produced the Pareto chart shown in Fig. 2, which is the result of combining the individual Pareto charts for each species. As can be seen, temperature was statistically significant in all cases, being affected by a positive sign. Also, the interaction between percentage of acetone and volume was significant and negative for m-cresol. For the other compounds this interaction was close to the significance level, also being negative.The volume and the percentage of acetone were not significant but the former was affected by a negative sign and the latter by a positive sign. Fig. 3(a) shows the response surface function developed by the model considering temperature and percentage of acetone in the case of o-cresol. As can be seen, the extraction efficiency was directly proportional to the temperature and to the percentage of acetone. Response surfaces modeled for phenol and p-cresol led to identical conclusions.Fig. 3(b) shows the same function for m-cresol. In this case, the response presents a maximum between 120 and 130 °C when the percentage of acetone is at its highest levels. Fig. 4(a) and (b) show the response surface function developed by the model considering temperature and volume for phenol and m-cresol, respectively (we do not show the response surface function for the other two compounds because, as can be seen in Fig. 2, their influence in the system is very far from the significance boundary and close to zero).In both cases, the response reaches the maximum value when the volume is at its lowest levels. Also, temperature presents a maximum for mcresol between 110 and 120 °C. Fig. 5 shows the response surface function developed by the model considering percentage of acetone and volume for mcresol. The response obtained is maximum when one of the two factors is at its highest level and the other is at its lowest level.When the two factors have the lowest value simultaneously, or the highest value simultaneously, the response obtained is minimum. A possible explanation of this behavior could be presented in terms of differential microwave energy absorption by the solid sample and/or the solvent. When the total solvent volume is high compared with the solid sample volume and contains high proportion of a microwave non-transparent cosolvent (e.g., acetone), energy should be mostly absorbed by the supernatant solvent rather by the solid sample material.Therefore, heating of the sample is produced not only by direct interaction with microwaves but mainly by convective heating from the top hot solvent mixture layer, thus showing slower extraction kinetics. On the other hand, for solvent mixtures containing low proportions of acetone, convective heating is negligible and microwave energy has to be absorbed by the solid material because the extraction solvent appears to be mostly transparent to microwaves.When low volumes of solvent are used, most of the liquid is in direct contact with the solid material. Also in this case an increase in the proportion of acetone improves the extraction efficiencies and kinetics. In any case, given these findings, it was decided to work with 10 ml of acetone–hexane (80 + 20) solvent at 130 °C. Number of Simultaneous Extractions and Size of Sample Using the same soil and the optimum conditions developed above (Table 3, last column), single and multiple (six samples) extraction experiments were performed.The results obtained were identical, within experimental error. Another set of Table 3 MAE and ultrasonic bath extraction parameters employed and optimum MAE values Central Homogeneity composite Sonic bath Optimum Parameter study design extraction value Temperature/°C 90 70–130 — 130 Proportion of acetone (%) 50 20–80 80 80 Volume of solvent/ml 20 15–50 10 10 Time/min 10 10 60 10 No.of samples extracted simultaneously 6 1 5 1–6 Sample size/g 5 5 2 1–5 Table 4 Design matrix and response values in the central composite design Recovery (%) Run Temperature/ Acetone Volume/ No. °C (%) ml Phenol o-Cresol p-Cresol m-Cresol 1 100 50 32.50 57.2 63.9 45.6 63.2 2 100 50 56.2 59.6 54.6 39.2 62.1 3 130 80 50 85.5 82.8 82.8 73.3 4 70 80 50 49.4 54.0 35.7 23.3 5 70 80 15 85.2 69.3 59.7 82.0 6 70 20 50 76.1 71.0 65.4 82.1 7 130 20 15 80.8 68.6 53.9 61.9 8 100 9.4 32.50 72.6 61.3 47.5 70.4 9 100 80 32.50 69.5 80.4 49.5 63.2 10 140.6 50 32.50 103.7 86.6 79.4 81.6 11 130 20 50 78.1 73.0 58.6 69.4 12 70 20 15 45.5 43.5 37.0 37.1 13 130 80 15 94.3 80.2 63.9 88.0 14 100 80 8.8 86.3 77.6 70.1 82.6 15 54.4 80 32.50 39.5 32.2 20.2 25.0 16 100 90.6 32.50 83.9 68.3 66.2 83.1 17 100 50 32.50 75.5 65.1 51.7 73.2 Analyst, February 1997, Vol. 122 135experiments was also performed working with a 1 g sample size. The recovery results were again identical for all these extractions.Fig. 6 summarizes the results obtained. In conclusion single or multiple extractions can be performed with a sample size between 1 and 5 g using an extraction time as short as 10 min after reaching 130 °C. The total sample preparation time is around 15 min and depends on the sample size and the characteristics of the soil. Comparison of Sonication and Microwave-assisted Extraction Methods and Final Recoveries Obtained For comparison purposes, we carried out five 2 g extractions on the same soil using a sonication bath (extraction time 60 min).The sonication extraction conditions and the MAE conditions are given in Table 3 (fourth and fifth columns, respectively). The sonication recoveries were about 50% of the real amount added to the soil whereas the MAE recoveries were 100% for all Fig. 2 Combined Pareto chart for the standardized effects in the central composite design including two-factor interactions. Dotted vertical lines indicates the statistical significance bounds for the effects.Fig. 3 Response surfaces estimated from the factorial design by plotting temperature versus percentage of acetone in the solvent for (a) o-cresol and (b) m-cresol extraction. Fig. 4 Response surfaces estimated from the factorial design by plotting temperature versus extractant volume for (a) phenol and (b) m-cresol extraction. Fig. 5 Response surface estimated from the factorial design by plotting percentage of acetone in the extraction solvent versus solvent volume for mcresol extraction. 136 Analyst, February 1997, Vol. 122the species studied. The results are given in Table 5. The RSD is around 7% in both the microwave and sonication methods. The authors acknowledge the support of the Spanish Interministerial Commission for Science and Technology and to the Xunta de Galicia, within the framework of projects PB92-0372 and XUGA20906B95. One of the authors (M.P.L) is indebted to the Spanish Education Ministry for a doctoral grant.References 1 Ganzler, K., Bati, J., and Valko, K., in Chromatography, the State of the Art, Akad�emiai Kiad�o, Budapest, 1985. 2 Ganzler, K., Salgo, A., and Valko, K., J. Chromatogr., 1986, 371, 299. 3 Zlotorzynski, A., Crit. Rev. Anal. Chem., 1995, 25, 43. 4 Par�e, J. R. J., US Pat., 5 002 784, 1991. 5 Par�e, J. R. J., US Pat., 5 338 557, 1994; 5 377 426, 1995; 5 458 557, 1995; 1996; 5 519 947. 6 Par�e, J.R. J., Belanger, J. M. R., and Stafford, S. S., Trends. Anal. Chem., 1994, 13, 176. 7 Collin, C. J., Lord, D., Allarire, J., and Gagnon, D., Parfums Cosm�et. Aromes, 1991, 97, 107. 8 Par�e, J. R. J., and B�elanger, J. M. R., in Instrumental Methods in Food Analysis, ed. Par�e, J. R. J., and B�elanger, J. M. R., Elsevier, Amsterdam, 1994, ch. 10. 9 Croteau, L. G., Akhtar, M. H., B�elanger, J. M. R., and Par�e, J. R. J., J. Liquid Chromatogr., 1994, 17, 2971. 10 Onuska, F. E., and Terry, K.A., Chromatographia, 1993, 36, 191. 11 Onuska, F. E., and Terry, K. A., J. High Resolut. Chromatogr., 1995, 18, 417. 12 Fish, J. R., and Revesz, R., presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, New Orleans, LA, March 1995. 13 Hasty, E., and Revesz, R., Am. Lab., 1995, 2, 66. 14 Lopez- � Avila, V., Young, R., Benedicto, J., Ho, P., Kim, R., and Beckert, W. F., Anal. Chem., 1995, 67, 2096. 15 Lopez- � Avila, V., Young, R., and Beckert, W.F., Anal. Chem., 1994, 66, 1097. 16 Lopez- � Avila, V., Benedicto, J., Charan, C., Young, R., and Beckert, W. F., Environ. Sci. Technol., 1995, 29, 2709. 17 Environmental Health Criteria 161: Phenol, US Environmental Protection Agency, Cincinnati, OH, 1992. 18 Environmental Health Criteria 168: Cresols, US Environmental Protection Agency, Cincinnati, OH, 1995. 19 Statgraphics Plus V.6., Reference Manual, Manugistics, Rockville, MD, 1992. Paper 6/05447A Received August 5, 1996 Accepted October 10, 1996 Fig. 6 Percentage recoveries for the studied species using different MAE conditions. Series 1, single extraction (n = 6), sample size 5 g; series 2, multiple extraction (n = 6), sample size 5 g; and series 3, multiple extraction (n = 6), sample size 1 g. Table 5 MAE versus sonication, with operating conditions as given in the fourth and fifth columns in Table 3 (n = 5 in both cases) MAE Sonic bath Compound Recovery (%) RSD (%) Recovery (%) RSD (%) Phenol 104.4 6.6 55.6 7.9 o-Cresol 94.5 6.1 52.7 3.7 m-Cresol 98.4 8.9 58.7 7.4 p-Cresol 89.1 9.9 45.3 7.9 Analyst, February 1997, Vol. 122 137 Optimization of a Microwave-assisted Extraction Method for Phenol and Methylphenol Isomers in Soil Samples Using a Central Composite Design Mar�ýa P. Llomparta, Rosa A. Lorenzoa, Rafael Cela*a and J. R. Jocelyn Par�eb a Departamento de Qu�ýmica Anal�ýtica, Nutrici�on y Bromatolog�ýa, Facultad de Qu�ýmica, Universidad de Santiago de Compostela, 15706 Santiago de Compostela, Spain.E-mail: qnrctd@usc.es b Environment Canada, Environmental Technology Centre, Ottawa, Ontario, Canada, K1A 0H3 A simple and rapid microwave-assisted extraction (MAE) procedure was developed and optimized for phenol, o-cresol, m-cresol and p-cresol in soil samples. The spiked soil used was prepared 25 d before treatment to simulate weathering processes and to allow for the formation of analyte-matrix interactions. The samples, immersed in hexane–acetone, were irradiated with microwaves in a closed-vessel system.Optimization of the method was achieved by using a factorial design approach on parameters such as the volume of solvent, the percentage of acetone and the extraction temperature. The MAE procedure yielded extracts that could be analyzed directly using a GC–flame ionization detection system without any preliminary clean-up or concentration steps. Keywords: Microwave extraction; phenol; cresols; factorial designs; soil samples The first applications of microwave-assisted extraction (MAE) to be reported dealt with the extraction of flavors from plant material,1–3 although it can be applied to the extraction of a variety of chemicals from a wide range of matrices such as soils, plant and animal tissues and a variety of manufactured products.1–9 In general, the compounds can be extracted more selectively and more quickly with similar or better recoveries in comparison with conventional extraction processes.Also, MAE uses less solvent and energy than conventional techniques.3,4 In the last 2 years the number of environmental applications of this technique has increased rapidly. Onuska and Terry extracted organochlorinated pesticides from sediments10 and PCBs from water.11 Fish and Revesz12 extracted organochlorinated pesticides from soil. Hasty and Revesz13 applied this technique to the extraction of petroleum hydrocarbons from soil.Lopez-Avila and co-workers14–16 used microwave energy to extract several groups of pollutants such as PAHs, PCBs, pesticides, phenolics and base/neutral compounds in soils and sediments. In all these studies, MAE proved to be similar to or more efficient than Soxhlet and sonication methods. Phenol and cresols are constituents of crude oil and coal tar, also being used in the chemical industry. It is well known that these compounds exhibit properties that are hazardous to human health,17,18 thus making it necessary to identify the occurrence and levels of contamination in the environment, especially in soils.In this study, we used MAE for the extraction of phenol and the three methylphenols (cresols) from soils in a closedvessel system. The solvent employed was hexane–acetone. The spiked soil used in the optimization process was prepared 25 d before analysis. Parameters studied included volume of solvent, percentage of acetone and extraction temperature.The soil samples were treated by this procedure using the optimum conditions obtained by applying a central composite design. Experimental Microwave-assisted Extraction MAE experiments were performed with a 950-W MES-100Matthews, NC, USA). This extractor has provision for 12 simultaneous extractions. A 1–5 g aliquot of soil was accurately weighed into a Teflon-lined extraction vessel and 10–50 ml of acetone– hexane were added to each sample in the proportions dictated by the experimental design.The extraction vessels were closed after ensuring that a new rupture membrane was used for each extraction. For this study, 1–6 simultaneous extractions were performed using full power in less than 15 min at an optimized temperature of 130 °C. When the irradiation period was completed, the sample carousel was removed from the microwave cavity and cooled in a water-bath. The control vessel was returned to the microwave system to check that the extract was at room temperature before opening. Solvent losses were checked in randomly selected experiments and were found to be < 1%.Using a glass pipette, 1 ml of the clear supernatant was transferred into an injection vial and the raw extract was analyzed by GC–flame ionization detection (FID) without any preliminary clean-up or concentration procedure. Fig. 1 shows the chromatogram for a spiked soil sample. Ultrasonic Bath Extraction Extractions using an ultrasonic bath (Selecta, Barcelona, Spain) were performed using 2 g portions of soil.The soil samples were sonicated for 60 min with continuous power under identical chemical conditions, namely 8 ml of acetone and 2 ml of hexane. The clear supernatant extracts were analyzed without any clean up or concentration steps. Fig. 1 Chromatogram of a spiked soil sample showing the resolution obtained using the operating conditions given in Table 1. Analyst, February 1997, Vol. 122 (133–137) 133Reagents and Chemicals Phenol standards were supplied by Aldrich Chemie (Steinheim, Germany). Methanol, hexane and acetone were purchased from Romil Chemicals (Cambridge, UK). Phenol stock standard solutions (1 g l21) were prepared by weighing an appropriate amount of the standard and dissolved in 10 ml of hexane. Working standard solutions were prepared by appropriate dilution of the stock standard solutions. All the solutions were stored at 5 °C in the dark when not in use.For quantitative GC determinations, calibration was carried out at four concentration levels spanning the range 1–10 mg ml21. Optimization experiments were performed using a spiked garden soil sample obtained from the campus of Santiago de Compostela University (Galicia, Spain), the carbon content of which was 2%. The soil was dried in an oven at 104 °C for 48 h, ground and sifted to a particle size below 300 mm. A 300 g aliquot was slurried with 250 ml of methanol solution of phenols.The sample was then allowed to air-dry with occasional mixing at ambient temperature, protected from draughts, for 5 d. The soil was then bottled and stored in a dry, dark location for 20 d before the first extractions. The concentrations in the soil, on the basis of added amounts, were 10.84, 14.23, 12.17 and 12.10 mg g21 for phenol, o-cresol, mcresol and p-cresol, respectively. It was also assumed that the contaminants were uniformly distributed in the sample and that, because the soil still retained residual moisture throughout the storage period, any analyte–matrix interactions would have occurred, over the weathering period, to a similar extent to those in real contaminated soil with similar properties.Analysis Extracts were analyzed on a Hewlett-Packard (Avondale, PA, USA) HP5890 Series II gas chromatograph equipped with a flame ionization detector and a Hewlett-Packard Model 7673A autosampler. A 60 m 3 0.56 mm id, 0.2 mm phase thickness, fused-silica chromatographic column coated with diisodecyl phthalate (DIIDP) (Restek, Bellefonte, PA, USA) was used, which allows good resolution of the three cresols.Chromatographic data were acquired and processed with a Hewlett- Packard Model 3365A data station. Table 1 summarizes the chromatographic conditions used. Safety Considerations Microwave-assisted processes are simple and can be readily understood in terms of the operating steps to be performed.However, the application of microwave energy to flammable organic solvents can pose serious hazards in inexperienced hands. Thus extreme safety precautions and great attention to details when planning and performing experiments must be used by all personnel dealing with microwaves. The authors urge all readers to ensure that they seek proper information from knowledgeable sources and that they do not attempt to implement these techniques unless proper guidance is provided.Only approved equipment and scientifically sound procedures should be used at all times. Results and Discussion Optimization of the Procedure for Phenol and Methylphenols by Gas Chromatography The chromatographic conditions were optimized for the resolution of the four phenols considered, as summarized in Table 1. By using a DIIDP chromatographic column an adequate resolution of the methyl phenol isomers is obtained (Fig. 1). Other stationary phases do not provide resolution between m-cresol and p-cresol.As noted under Experimental, calibration curves were constructed at four concentration levels using appropriately diluted standards. Each concentration level was injected in triplicate. Chromatographic peak areas were fitted by linear regression and the results are given in Table 2. The repeatability of the chromatographic procedure was assessed by performing eight consecutive injections of a standard solution containing the four analytes.The results (between-injection repeatability data) are also given in Table 2 along with detection and quantification limits for direct injections of standards at signal-to-noise ratios of three and ten, respectively. Evaluation of the Homogeneity of Laboratory-spiked Soil Samples In order to test the homogeneity of the spiked sample to be used in optimizing the extraction process, a set of six 5 g extractions were performed under the conditions given in the second column of Table 3.The average recoveries obtained were 77.2, 67.7, 70.3 and 59.9% for phenol, o-, m- and p-cresol, respectively. The variability (RSD 7.3–9.6%) was very close to that obtained from the injection of calibration standards alone (Table 2). In conclusion, the material was considered to be homogeneous as far as the target analytes are concerned, taking portions of at least 5 g. Some blank samples were extracted under the same conditions (Table 3, second column). The results showed the absence of an analytical signal at the retention times of the compounds studied.Factorial Design. Evaluation of the Response Surfaces The variables considered in the MAE optimization process were temperature, volume of solvent and percentage of acetone. A Table 1 GC operating conditions Injection port temperature 125 °C Injection mode Splitless Injection volume 2 ml Splitless time 60 s Column 60 m30.56 mm id 0.2 mm film thickness, DIIDP (Restek) Carrier gas Nitrogen (99.9995%) Carrier gas flow rate 5.8 ml min21 Carrier gas pressure at column head 50 kPa Oven temperature 100 °C FID temperature 150 °C Table 2 Calibration and statistical validation parameters Parameter Phenol o-cresol m-cresol p-cresol Calibration range/ mg ml21 0.98–7.82 1.28–10.27 1.10–8.78 1.09–8.74 Correlation coefficient 0.9999 0.9999 0.9999 0.9999 Detection limit/ ng ml21 (S/N = 3) 20.2 18.3 23.7 24.0 Quantification limit/ng ml21 (S/N = 10) 69.2 61.0 79.1 79.9 Between-injection RSD (%) 3.8 5.5 6.2 6.0 134 Analyst, February 1997, Vol. 122three-level central 23 + star orthogonal composite design involving 14 runs and three central points was chosen. This model allows the direct evaluation of the variables considered, and also the first- and second-order interaction terms. The low and high levels assigned to the variables regarding the value assigned to the fixed factors are listed in the third column of Table 3. The data analysis was performed using the statistical package Statgraphics Plus V.6.0.19 Table 4 gives the design matrix for this experiment and the recoveries obtained in each run.An analysis of the results given in Table 4 produced the Pareto chart shown in Fig. 2, which is the result of combining the individual Pareto charts for each species. As can be seen, temperature was statistically significant in all cases, being affected by a positive sign. Also, the interaction between percentage of acetone and volume was significant and negative for m-cresol.For the other compounds this interaction was close to the significance level, also being negative. The volume and the percentage of acetone were not significant but the former was affected by a negative sign and the latter by a positive sign. Fig. 3(a) shows the response surface function developed by the model considering temperature and percentage of acetone in the case of o-cresol. As can be seen, the extraction efficiency was directly proportional to the temperature and to the percentage of acetone.Response surfaces modeled for phenol and p-cresol led to identical conclusions. Fig. 3(b) shows the same function for m-cresol. In this case, the response presents a maximum between 120 and 130 °C when the percentage of acetone is at its highest levels. Fig. 4(a) and (b) show the response surface function developed by the model considering temperature and volume for phenol and m-cresol, respectively (we do not show the response surface function for the other two compounds because, as can be seen in Fig. 2, their influence in the system is very far from the significance boundary and close to zero). In both cases, the response reaches the maximum value when the volume is at its lowest levels. Also, temperature presents a maximum for mcresol between 110 and 120 °C. Fig. 5 shows the response surface function developed by the model considering percentage of acetone and volume for mcresol.The response obtained is maximum when one of the two factors is at its highest level and the other is at its lowest level. When the two factors have the lowest value simultaneously, or the highest value simultaneously, the response obtained is minimum. A possible explanation of this behavior could be presented in terms of differential microwave energy absorption by the solid sample and/or the solvent. When the total solvent volume is high compared with the solid sample volume and contains high proportion of a microwave non-transparent cosolvent (e.g., acetone), energy should be mostly absorbed by the supernatant solvent rather by the solid sample material.Therefore, heating of the sample is produced not only by direct interaction with microwaves but mainly by convective heating from the top hot solvent mixture layer, thus showing slower extraction kinetics. On the other hand, for solvent mixtures containing low proportions of acetone, convective heating is negligible and microwave energy has to be absorbed by the solid material because the extraction solvent appears to be mostly transparent to microwaves. When low volumes of solvent are used, most of the liquid is in direct contact with the solid material.Also in this case an increase in the proportion of acetone improves the extraction efficiencies and kinetics. In any case, given these findings, it was decided to work with 10 ml of acetone–hexane (80 + 20) solvent at 130 °C.Number of Simultaneous Extractions and Size of Sample Using the same soil and the optimum conditions developed above (Table 3, last column), single and multiple (six samples) extraction experiments were performed. The results obtained were identical, within experimental error. Another set of Table 3 MAE and ultrasonic bath extraction parameters employed and optimum MAE values Central Homogeneity composite Sonic bath Optimum Parameter study design extraction value Temperature/°C 90 70–130 — 130 Proportion of acetone (%) 50 20–80 80 80 Volume of solvent/ml 20 15–50 10 10 Time/min 10 10 60 10 No.of samples extracted simultaneously 6 1 5 1–6 Sample size/g 5 5 2 1–5 Table 4 Design matrix and response values in the central composite design Recovery (%) Run Temperature/ Acetone Volume/ No. °C (%) ml Phenol o-Cresol p-Cresol m-Cresol 1 100 50 32.50 57.2 63.9 45.6 63.2 2 100 50 56.2 59.6 54.6 39.2 62.1 3 130 80 50 85.5 82.8 82.8 73.3 4 70 80 50 49.4 54.0 35.7 23.3 5 70 80 15 85.2 69.3 59.7 82.0 6 70 20 50 76.1 71.0 65.4 82.1 7 130 20 15 80.8 68.6 53.9 61.9 8 100 9.4 32.50 72.6 61.3 47.5 70.4 9 100 80 32.50 69.5 80.4 49.5 63.2 10 140.6 50 32.50 103.7 86.6 79.4 81.6 11 130 20 50 78.1 73.0 58.6 69.4 12 70 20 15 45.5 43.5 37.0 37.1 13 130 80 15 94.3 80.2 63.9 88.0 14 100 80 8.8 86.3 77.6 70.1 82.6 15 54.4 80 32.50 39.5 32.2 20.2 25.0 16 100 90.6 32.50 83.9 68.3 66.2 83.1 17 100 50 32.50 75.5 65.1 51.7 73.2 Analyst, February 1997, Vol. 122 135experiments was also performed working with a 1 g sample size. The recovery results were again identical for all these extractions. Fig. 6 summarizes the results obtained. In conclusion single or multiple extractions can be performed with a sample size between 1 and 5 g using an extraction time as short as 10 min after reaching 130 °C. The total sample preparation time is around 15 min and depends on the sample size and the characteristics of the soil.Comparison of Sonication and Microwave-assisted Extraction Methods and Final Recoveries Obtained For comparison purposes, we carried out five 2 g extractions on the same soil using a sonication bath (extraction time 60 min). The sonication extraction conditions and the MAE conditions are given in Table 3 (fourth and fifth columns, respectively). The sonication recoveries were about 50% of the real amount added to the soil whereas the MAE recoveries were 100% for all Fig. 2 Combined Pareto chart for the standardized effects in the central composite design including two-factor interactions. Dotted vertical lines indicates the statistical significance bounds for the effects. Fig. 3 Response surfaces estimated from the factorial design by plotting temperature versus percentage of acetone in the solvent for (a) o-cresol and (b) m-cresol extraction. Fig. 4 Response surfaces estimated from the factorial design by plotting temperature versus extractant volume for (a) phenol and (b) m-cresol extraction.Fig. 5 Response surface estimated from the factorial design by plotting percentage of acetone in the extraction solvent versus solvent volume for mcresol extraction. 136 Analyst, February 1997, Vol. 122the species studied. The results are given in Table 5. The RSD is around 7% in both the microwave and sonication methods. The authors acknowledge the support of the Spanish Interministerial Commission for Science and Technology and to the Xunta de Galicia, within the framework of projects PB92-0372 and XUGA20906B95.One of the authors (M.P.L) is indebted to the Spanish Education Ministry for a doctoral grant. References 1 Ganzler, K., Bati, J., and Valko, K., in Chromatography, the State of the Art, Akad�emiai Kiad�o, Budapest, 1985. 2 Ganzler, K., Salgo, A., and Valko, K., J. Chromatogr., 1986, 371, 299. 3 Zlotorzynski, A., Crit. Rev. Anal. Chem., 1995, 25, 43. 4 Par�e, J. R. J., US Pat., 5 002 784, 1991. 5 Par�e, J. R. J., US Pat., 5 338 557, 1994; 5 377 426, 1995; 5 458 557, 1995; 1996; 5 519 947. 6 Par�e, J. R. J., Belanger, J. M. R., and Stafford, S. S., Trends. Anal. Chem., 1994, 13, 176. 7 Collin, C. J., Lord, D., Allarire, J., and Gagnon, D., Parfums Cosm�et. Aromes, 1991, 97, 107. 8 Par�e, J. R. J., and B�elanger, J. M. R., in Instrumental Methods in Food Analysis, ed. Par�e, J. R. J., and B�elanger, J. M. R., Elsevier, Amsterdam, 1994, ch. 10. 9 Croteau, L. G., Akhtar, M. H., B�elanger, J. M. R., and Par�e, J. R. J., J. Liquid Chromatogr., 1994, 17, 2971. 10 Onuska, F. E., and Terry, K. A., Chromatographia, 1993, 36, 191. 11 Onuska, F. E., and Terry, K. A., J. High Resolut. Chromatogr., 1995, 18, 417. 12 Fish, J. R., and Revesz, R., presented at the Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, New Orleans, LA, March 1995. 13 Hasty, E., and Revesz, R., Am. Lab., 1995, 2, 66. 14 Lopez- � Avila, V., Young, R., Benedicto, J., Ho, P., Kim, R., and Beckert, W. F., Anal. Chem., 1995, 67, 2096. 15 Lopez- � Avila, V., ., Anal. Chem., 1994, 66, 1097. 16 Lopez- � Avila, V., Benedicto, J., Charan, C., Young, R., and Beckert, W. F., Environ. Sci. Technol., 1995, 29, 2709. 17 Environmental Health Criteria 161: Phenol, US Environmental Protection Agency, Cincinnati, OH, 1992. 18 Environmental Health Criteria 168: Cresols, US Environmental Protection Agency, Cincinnati, OH, 1995. 19 Statgraphics Plus V.6., Reference Manual, Manugistics, Rockville, MD, 1992. Paper 6/05447A Received August 5, 1996 Accepted October 10, 1996 Fig. 6 Percentage recoveries for the studied species using different MAE conditions. Series 1, single extraction (n = 6), sample size 5 g; series 2, multiple extraction (n = 6), sample size 5 g; and series 3, multiple extraction (n = 6), sample size 1 g. Table 5 MAE versus sonication, with operating conditions as given in the fourth and fifth columns in Table 3 (n = 5 in both cases) MAE Sonic bath Compound Recovery (%) RSD (%) Recovery (%) RSD (%) Phenol 104.4 6.6 55.6 7.9 o-Cresol 94.5 6.1 52.7 3.7 m-Cresol 98.4 8.9 58.7 7.4 p-Cresol 89.1 9.9 45.3 7.9 Analyst, February 1997, Vol. 122 1

ISSN:0003-2654    

DOI:10.1039/a605447a

出版商:RSC    

年代:1997

数据来源: RSC

摘要:

O O CH3 SO3Na O O O O CH3 MSB OH– + HSO3 – + Na+ O OH CH3 Menadione ( a) Zn H+ ( b) CH3 Spectrofluorimetric Determination of Vitamin K3 by a Solid-phase Zinc Reactor Immobilized in a Flow Injection Assembly I. Gil Torr�oa, J. V. Garc�ýa Mateoa and J. Mart�ýnez Calatayud*b a Departamento de Qu�ýmica, Colegio Universitario CEU, Moncada (Valencia), Spain b Departamento de Qu�ýmica Anal�ýtica, Universidad de Valencia, 46100 Burjasot (Valencia), Spain The spectrofluorimetric determination of vitamin K3 (menadione) using a flow injection (FI) assembly provided with a solid-phase reactor with immobilized zinc is described. The naphthohydroquinone was produced by means of two coupled steps in the FI system: hydrolysis of the sodium sulfite derivative of the menadione in a basic medium and reduction of the generated menadione in the zinc reactor in an acidic medium.The fluorescent product was monitored spectrofluorimetrically (lex 325 nm; lem 425 nm). The calibration graph was linear over the range 0.1–18 mg ml21 with a reproducibility of 1.6%; the limit of detection was 0.005 mg ml21 and the sample throughput was 70 h21.The influence of foreign compounds was studied and the procedure was applied to the determination of vitamin K3 in three different pharmaceutical formulations. Keywords: Vitamin K3; menadione sodium hydrogensulfite; fluorimetry; unsegmented continuous flow Vitamin K is a generic term for a homologous group of fat soluble vitamins consisting of 2-methyl-1,4-naphthoquinone derivatives.They have in common their physiological role as a required cofactor in the synthesis of blood clotting and their highly lipophilic nature. Synthetic derivatives with similar activity to natural vitamins include menadione or vitamin K3. Water-soluble derivatives of menadione such as menadione sodium hydrogensulfite (MSB) have been prepared. These compounds yield menadione after decomposition in the organism.Several methods have been proposed for the determination of MSB. In complex matrices extraction using an aqueous alcoholic solvent and subsequent back-extraction of menadione into a organic solvent after hydrolysis of the vitamin in a basic medium is usual. Chromatographic techniques have provided suitable means of determination, e.g., GC with flame ionization detection,1–3 HPLC with UV detection4 and spectrofluorimetric detection with post-column derivatization based on the reduction of the menadione to a fluorescent derivative (2-methyl-1,4-dihydronaphthoquinone). Post-column reduction with sodium tetrahydroborate (NaBH4) has also been reported.5 Recently, packed columns with powdered zinc and post-column derivatization procedures in chromatography have been used.6 Both electrochemical7,8 and photochemical reduction9 have been exploited for fluorimetric detection. Other methods include spectrophotometric detection,10,11 HPLC12 and polarographic determination of menadione in anhydrous acetonitrile with lithium perchlorate as carrier electrolyte against a calomel reference electrode.13 Finally, the British Pharmacopoeia14 describes a cerimetric titration of menadione with a previous reduction to the hydroquinone with zinc in hydrochloric acid medium with light excluded.This paper describes a flow injection (FI) procedure for the determination of vitamin K3. The method includes the on-line hydrolysis of MSB in a basic medium and subsequent chemical reduction of the menadione by a solid-phase zinc reactor (Fig. 1). The product formed was monitored spectrofluorimetrically (lex 325 nm, lem 425 nm). Experimental Reagents Analytical-reagent grade chemicals were used unless indicated otherwise. Menadione sodium hydrogensulfite was obtained from Guinama (Valencia, Spain), acetonitrile from Scharlau (Barcelona, Spain) and NaOH and HCl from Panreac (Barcelona, Spain). The zinc bed reactor was prepared with zinc ‘gross’ (Merck, Darmstadt, Germany) confined in PTFE tubing of 1.5 mm id.All solutions were prepared with deionized water. Zinc powder was immobilized in a polyester resin bed as described earlier.15,16 Apparatus The apparatus consisted of a Model 5041 rotary injection valve (Rheodyne,Cotati, CA, USA), a Minipuls 2 peristaltic pump from Gilson (Worthington, OH, USA) an F-4500 spectrofluorimeter (Hitachi, Barcelona, Spain) and flow cells from Hellma (M�ullheim, Germany) of 8 ml inner volume and 8 3 1 mm aperture and 25 ml inner volume and 11 31.5 mm aperture.The flow system was made of PTFE tubing of 0.8 mm id and a PTFE Y confluence from Omnifit (Cambridge, UK). Flow Injection Assembly The flow injection manifold is depicted in Fig. 2. A solution of MSB in acetonitrile–water (9 + 1 v/v) merged with 0.035 m Fig. 1 (a) Hydrolysis of the MSB and (b) chemical reduction of menadione. Analyst, February 1997, Vol. 122 (139–142) 139NaOH solution.The water-soluble vitamin was hydrolysed in the basic medium in loop L1 and then merged with 0.08 m HCl solution in acetonitrile–water (4 + 6). The menadione formed by means of this treatment was injected via the sample loop into the carrier solution [acetonitrile–water (9 + 1)] and forced through the zinc reactor (R) (35 cm 31.5 mm id) and into loop L3. After the reduction step, the fluorescence of the product (2-methyl- 1,4-dihydronaphthalene) was monitored (lex 325 nm, lem 425 nm).Results and Discussion Preliminary Tests Preliminary experiments allowed us to establish the chemical conditions for the spectrofluorimetric determination of MSB. A single-line assembly including a 10 cm 3 1.5 mm id reactor containing zinc immobilized on polyester resin (particle size 150–200 mm) was used to pass 40 mg ml21 solutions of MSB in methanol–water (3 + 2). Because each reaction must be conducted at a different pH, the solutions were prepared in the presence and absence of 2% sodium carbonate, followed in some cases by addition of 0.1 m HCl to pH 2.15.The same experiment was repeated in the absence of the zinc reactor. The assembly excluding the reactor provided no fluorescence signal; in the alkaline medium, a yellow colour was observed as a result of the hydrolysis of MSB to menadione; the former decomposed to a considerable extent under the action of light via a variety of complex photodegradation pathways. In the presence of the zinc reactor, only vitamin K3 exhibited fluorescence (lex = 325 nm, lem = 425 nm), following hydrolysis in the alkaline medium.Both the hydrolysis and the reduction of the vitamin took place virtually instantaneously. Selection of the Reductant Column The reductant power of the zinc reactor, its reactivity against acids and the particle size used suggested that the reducing metal could be gradually exhausted. For this reason, an experiment was performed in order to assess the influence of the reaction medium on the column lifetime.To this end, the assembly in Fig. 3 was used to monitor the fluorescence intensity of 2-methyl-1,4-dihydronaphthalene for 1 h. Subsequently, the column was flushed by passing a washing solution (deionized water) for 10 min. Finally, the reactor was again subjected to the above-mentioned exhaustion conditions for a further 1 h. The experiment was carried out with various reductants, viz., zinc and cadmium (both of which were assayed separately as powder immobilized on polyester resin in a metal to resin mass ratio of 1 : 1 and a particle size of 150–200 mm), copperized cadmium and zinc grains. The immobilized zinc and cadmium powder beds proved ineffective as they gradually lost their reducing power.Copperized cadmium exhibited a higher resistance to exhaustion; however, the fluorescence signals obtained were scarcely reproducible and considerably smaller (about 25% lower than those provided by zinc and cadmium powder under identical conditions).A 12 cm 35 mm id column packed with zinc particles of size between 0.5 and 1.2 mm was finally chosen. The readings obtained under these conditions remained constant throughout the working period (3 h). This column exhibited an increased reducing power against cadmium as a result of the increased contact surfanc particles, which was adopted for subsequent experiments. Study of the Chemical System The determination of the vitamin involves the on-line alkaline hydrolysis of the sodium hydrogensulfite in menadione to give a product that is insoluble in water but soluble in most organic solvents, followed by reduction in an acidic medium.These two steps (hydrolysis and reduction) were examined in greater detail by using an FI assembly similar to that in Fig. 2. The hydrolysis step was studied using 1% sodium carbonate and NaOH at concentrations from 0.001 to 0.07 m.At equal concentrations, sodium carbonate had a slight buffering effect and released some CO2 on acidification that interfered with detection. The highest transient analytical signals (FI peaks) were obtained with 0.035 m NaOH; lower concentrations resulted in inadequate hydrolysis whereas higher concentrations increased the pH and hindered the subsequent reduction. The reduction of menadione was studied using hydrochloric, acetic, sulfuric, phosphoric and perchloric acid, all at 0.1 m concentration.The fluorescence intensity obtained was similar for all except HCl, which provided higher values. In a subsequent experiment, the HCl concentration was varied in order to study the influence of the acidity of the medium on the response of the reductant column. An acid concentration over 0.1 m caused the production of small hydrogen bubbles that interfered with detection and shortened the lifetime of the reductant bed at higher acid concentrations. An HCl concentration of 0.1 m was therefore chosen for subsequent experiments.Study of the Fluorescence of 2-Methyl-1,4-dihydronaphthalene The characteristics of the detection system used prompted us to perform several experiments in order to investigate the potential influence of different variables on the fluorescence of 2-methyl- 1,4-dihydronaphthalene, viz., the irradiation time at the excitation wavelength (325 nm), solvent, ionic strength, dissolved oxygen and temperature.The experiments were carried out in the assembly depicted in Fig. 3. The influence of the irradiation time was revealed by stopping the flow in the FI system and recording the fluorescence intensity of naphthohydroquinone in the spectrofluorimeter cuvette in a continuous manner. The emission intensity was found to decrease with time, particularly within the first few seconds. Fig. 2 Flow injection assembly for determination of MSB. Flow rates, Q: a, Qa = 3.5 ml min21, MSB in acetonitrile–water (9 + 1 v/v); b, Qb = 0.3 ml min21, 0.035 m NaOH; c, Qc = 0.3 ml min21, 0.08 m HCl in acetonitrile–water (2 + 3 v/v); d, Qd = 3.0 ml min21, acetonitrile–water (9 + 1 v/v).L1 = 330 cm; L2 = 10 cm; L3 = 25 cm. L = sample loop; V = injection valve; R = solid-phase zinc reactor (35 cm 3 1.5 mm id). Fig. 3 Flow injection assembly for preliminary studies. a, 40 mg l21 of MSB in 60% methanol, Qa = 1 ml min21; b, 1% m/v Na2CO3 solution, Qb = 0.5 ml min21; c, 0.1 m HCl, Qc = 0.5 ml min21.L1 = 49 cm; L2 = 94 cm. R = Immobilized zinc on a polyester resin (zinc to resin ratio = 1 : 1 m/m); particle size = 150–250 mm; D = spectrofluorimeter. 140 Analyst, February 1997, Vol. 122Because the polarity of the medium is known to affect excitation and de-excitation mechanisms for fluorescent species, we studied various water-miscible organic solvents including methanol, ethanol, acetonitrile, dioxane and acetone. The solutions used contained 50 mg ml21 MSB in 60% of the solvent.The blank signal was zero in every case except for dioxane and acetone, which provided no analytical signal. Acetonitrile gave the best results as the likely consequence of hydrogen bonding being less favoured than in the aqueous alcoholic mixtures, so it was adopted for further experiments. The influence of the ionic strength was determined by comparing the emission intensities recorded by replacing the 0.1 m HCl solution in the flow system with a solution also containing 4% NaCl.The analytical signals obtained in the presence of this salt were 15% smaller. The presence of oxygen in solution can have an adverse effect on the fluorescence of naphthohydroquinones as it alters the competitive reaction mechanisms that favour the formation of non-fluorescent products by a quenching effect. We therefore chose to degas the different solutions by immersion in an ultrasonic bath for 30 min. The readings obtained did not differ significantly from those recorded for the non-degassed solutions.These results are consistent with chromatographic practice regarding the use of columns packed with zinc particles that efficiently cancel the effect of dissolved oxygen. The effects of temperature on the hydrolysis and reduction reactions were studied separately. The former were examined by immersing a reactor of length L1 in a bath thermostated at 20, 35 or 56 °C (see Fig. 2). The latter were investigated by immersing the carrier solution (deionized water) in a bath at 35 °C.Increasing the temperature decreased the fluorescence emission and reproducibility. Room temperature was therefore chosen for the subsequent optimization step. Optimization of the Flow Injection Manifold The FI manifold (Fig. 2) was optimized by using the univariate method. Chemical variables (reactant concentrations) were optimized first, then the dimensions of the reductant column and finally the FIA variables.Chemical variables The NaOH and HCl concentrations were optimized jointly since, although the hydrolysis and reduction of the vitamin can be considered as two independent processes, the fact that the two channels were merged determined the pH (medium) for the second reaction (reduction): the acid must be concentrated enough to neutralize the NaOH and provide a suitable medium for reducing menadione; however, excess acid can release hydrogen and hinder detection.A 1 : 3 base to acid ratio proved appropriate. At this ratio, HCl concentrations between 0.1 and 1.5 m were tested. Concentrations near the upper end of this range released hydrogen and led to rapid exhaustion of the reactor; 0.035 m NaOH and 0.1 m HCl provided the best compromise between peak height and lifetime of the zinc column. The acetonitrile concentration in the MSB, NaOH and HCl solutions was optimized. The signal was found to rise markedly with increase in the solvent content in the MSB solution.The opposite effect, although not so marked, however, was observed in the NaOH and HCl solutions (Fig. 4). The hydrolysis and neutralization reactions took place to a greater extent in the aqueous medium, but the naphthohydroquinone exhibited more intense fluorescence in acetonitrile. A solvent content of 90% was used for the sample solution despite the increased intensity owing to the solubility limit of MSB in acetonitrile. Under these conditions, the HCl concentration must be lowered to 0.08 m owing to the formation of small hydrogen bubbles at higher acid levels.Reductant column The reductant column was optimized by altering the reactor length and diameter. The size of the zinc particles was imposed by the commercially available form of the reagent. We studied three different inner diameters for the column, viz., 5, 3.5 and 1.5 mm. The results obtained are shown in Fig. 5. A 35.5 cm 3 1.5 mm id PTFE column was selected for subsequent experiments.Larger diameters resulted in considerable sample dispersion, whereas smaller diameters and longer lengths led to increased reduction yields. However, the fluorescence signal decreased above 35.5 cm owing to increased dispersion of the sample, with no change in the reduction yield. FI variables The effects of FI variables were investigated by performing five injections of 2.5 mg ml21 MSB per value of the variable studied. The values adopted as optimum were those resulting in the best possible compromise between peak height (fluorescence intensity), reproducibility and throughput.The effects of the reactant flow rates in the sequence and of the ranges of carrier 0.47–3.6 ml min21, MSB 0.47–4.3 ml min21 and NaOH and HCl jointly 0.11–0.67 ml min21 (for both) were studied. The sample flow rate was found to be critical; the NaOH and HCl flow rates also proved influential to some extent. The sample flow rate was directly related to the amount of analyte present in the system: the greater this value the higher was the menadione concentration in the sample loop and hence the higher the fluorescence intensity.A flow rate of 3.5 ml min–1 was selected as optimum. The NaOH and HCl flow rates were altered jointly at a constant acid to base mole ratio of 3 : 1 and with minimal dilution of MSB. A flow rate of 0.3 ml min–1 for both reagents met this criterion. Increasing the carrier flow rate slightly decreased the analytical signal but increased the throughput.A flow rate of 3 ml min21 offset the dispersion produced by low flow rates without decreasing the reduction yield. The distances between merging points, between the reactor and detector, and the sample volume were optimized. Reactor– Fig. 4 Influence of the concentration of acetonitrile on the sample and reagent solutions. A, Sample; B, HCl; C, carrier; D, NaOH. Fig. 5 Influence of length and inner diameter of the zinc-bed reactor on the reduction of menadione.A, id = 5 mm, B, id = 3.5 mm; C, id = 1.5 mm. Analyst, February 1997, Vol. 122 141detector lengths greater than 25 cm decreased the signal through increased instability of the fluorescent product. The distance between the merging point for HCl and the valve was uninfluential because the acid–base reaction was instantaneous. On the other hand, MSB was hydrolysed in a fairly short time, so 330 cm was chosen as the optimum length.Finally, we studied the influence of the injected sample volume on the analytical signal. The range examined was 160–1015 ml. The fluorescence intensity was found to increase markedly with increase in the sample volume up to 850 ml, which was adopted as optimal. Once all the system variables had been optimized, those which proved critical in the optimization step (viz., the sample, NaOH and HCl flow rates) were re-optimized over narrower ranges. The results thus obtained coincided with those found in the optimization step.Analytical Applications The calibration graph was linear over the range 0.1–18 mg ml21 and fitted the equation I = 47.45 + 442.63c (n = 13), r = 0.9993, where I is the fluorescence intensity (in arbitrary units) and c the MSB concentration (in mg ml21). The day-today reproducibility was studied from five calibration curves constructed on different days. The average of the slopes obtained was 456.78 and the RSD was 2.1%.The reproducibility of the determination was calculated from replicate injections containing 10 mg ml21 of MSB (flow cell of 8 ml inner volume and 8 3 1 mm aperture). The RSD for 32 replicates was 1.6%, the sample throughput was 70 h21 and the limit of detection (three times the background noise) was 0.005 mg ml21. The sensitivity and limit of detection of the proposed method were substantially modified by increasing the aperture of the flow cell. In fact, replacing the flow cell with another one of 11 3 1.5 mm aperture and 25 ml inner volume resulted in the linear calibration graph I = 129.51 + 1451.50c (r = 0.9990).The RSD for 22 replicates and 3 mg ml21 was 1.2%, the sample throughput was 52 h21 and the limit of detection was 0.001 mg ml21. The potential effects of compounds (excipients) frequently accompanying MSB in its formulations were studied by using solutions containing the drug (10 mg ml21) and various concentrations of each interferent.The interferent concentrations tested and the relative errors they produced were as follows: ascorbic acid, 100 mg ml21, 24.1%; lactose, 500 mg ml21, 23.2%; saccharose, 400 mg ml21, 4.1%; saccharine, 100 mg ml21, 23.4%; and nicotinamide, 1000 mg ml21, 4.6%. Pectin, sodium glutamate and sulfadiazine, owing to their low solubility in water, were tested from saturated solutions and yielded errors of 0%. Finally, the proposed method was applied to the determination of MSB in three pharmaceutical formulations, Quercetol K (from Ferrer International, Barcelona, Spain), Citroflavona solution (from Funk, Barcelona, Spain) and vitamin K3 (from F.A.S., Valladolid, Spain).The results obtained were compared with those obtained with a reference method. Quercetol and Citroflavona solution were reduced to hydroquinone by zinc in hydrochloric acid with light excluded. After the zinc had been filtered off, the hydroquinone formed was determined by cerimetric titration, which re-oxidized it to quinone (according to the British Pharmacopoeia14).The determination of vitamin K3 (from F.A.S.) was carried out spectrometrically at 308 nm owing to the absence of interferences accompanying the drug. The experimental results are summarized in Table 1. Conclusions FI methodology allows the use of solid-phase zinc reactors and avoids the shortcomings associated with solid reagents: control of reactivity caused by excess of reagents and parallel reactions.The reactivity of this metal is suitable under controlled acid conditions (0.08 m HCl). The life span of the column is at least 1 month and no pretreatment of the bed reactor is necessary to obtain reproducible transient FI peaks after a few injections. The chemical reduction coupled with fluorimetric detection provide limits of detection at ppb levels and the method allows the automation of the on-line reduction of menadione as the previous step described in the British Pharmacopoeia14 as a standard method of analysis.References 1 Libby, D. A., and Sheppar, A. J., J. Assoc. Off. Anal. Chem., 1965, 48, 973. 2 Winkler, V. W., and Yoder, J. M., J. Assoc. Off. Anal. Chem., 1972, 55, 1219. 3 Winkler, V. W., J. Assoc. Off. Anal. Chem., 1973, 56, 1227. 4 Manz, U., and Mauer, R., Int. J. Vitam. Nutr. Res., 1982, 52, 248. 5 Speek, A. J., Schrijver, J., and Scheurs, H. P., J. Chromatogr., 1984, 301, 441. 6 Billedeau, S. M., J. Chromatogr., 1989, 472, 371. 7 Haroon, Y., Bacon, D. S., and Sadowski, A. J., Biomed. Chromatogr., 1987, 2, 4. 8 Wang, L. Z., Ma, C. S., Zhang, X. L., and Xu, L., Microchem. J., 1994, 50, 101. 9 Lefevere, M. F., Frei, R. W., Scholten, H. M. T., and Brinkman, U. A. Th., Chromatographia, 1982, 15, 459. 10 Ruan, Y., Duan, F., Ning, Y., and Fan, X., Fenxi Huaxue, 1988, 16, 746. 11 Iskander, M. L., Medien, H. A. A., and Khalil, L. H., Anal. Lett., 1995, 28, 1513. 12 Hu, O. Y. P., Wu, C. Y., Chan, W. K., and Wu, F.H. H., J. Chromatogr. B, 1995, 666, 299. 13 Pharmaceutical Chemistry. Vol. 2: Drug Analysis, ed. Roth, H. J., Ellis Horwood, Chichester, 1988, pp. 111–117. 14 British Pharmacopoeia, 1993, HM Stationery Office, London, 1993, p. 410. 15 Mart�ýnez Calatayud, J., and Garc�ýa Mateo, J. V., Anal. Chim. Acta, 1993, 274, 275. 16 Mart�ýnez Calatayud, J., Garc�ýa Mateo, J. V., and Lahuerta Zamora, L., Anal. Chim. Acta., 1992, 265, 81. Paper 6/06881B Received October 8, 1996 Accepted December 3, 1996 Table 1 Determination of menadione in pharmaceutical formulations Pharmaceutical Reference Error† preparation Declared value method (RSD)* Proposed method* (%) Quercetol K 5 mg per tablet 4.83 mg per tablet (0.6%) 4.77 mg per tablet 21.2 Citroflavona 0.66 mg in 5 ml 0.656 mg in 5 ml (1.2%) 0.654 mg in 5 ml 20.3 Vitamin K3 10 mg per ampoule 7.51 mg per ampoule (0.2%) 7.56 mg per ampoule 0.7 * Each value is the average of five determinations. † Error against the reference method. 142 Analyst, February 1997, Vol. 122 O O CH3 SO3Na O O O O CH3 MSB OH– + HSO3 – + Na+ O OH CH3 Menadione ( a) Zn H+ ( b) CH3 Spectrofluorimetric Determination of Vitamin K3 by a Solid-phase Zinc Reactor Immobilized in a Flow Injection Assembly I. Gil Torr�oa, J. V. Garc�ýa Mateoa and J. Mart�ýnez Calatayud*b a Departamento de Qu�ýmica, Colegio Universitario CEU, Moncada (Valencia), Spain b Departamento de Qu�ýmica Anal�ýtica, Universidad de Valencia, 46100 Burjasot (Valencia), Spain The spectrofluorimetric determination of vitamin K3 (menadione) using a flow injection (FI) assembly provided with a solid-phase reactor with immobilized zinc is described.The naphthohydroquinone was produced by means of two coupledtem: hydrolysis of the sodium sulfite derivative of the menadione in a basic medium and reduction of the generated menadione in the zinc reactor in an acidic medium. The fluorescent product was monitored spectrofluorimetrically (lex 325 nm; lem 425 nm).The calibration graph was linear over the range 0.1–18 mg ml21 with a reproducibility of 1.6%; the limit of detection was 0.005 mg ml21 and the sample throughput was 70 h21. The influence of foreign compounds was studied and the procedure was applied to the determination of vitamin K3 in three different pharmaceutical formulations. Keywords: Vitamin K3; menadione sodium hydrogensulfite; fluorimetry; unsegmented continuous flow Vitamin K is a generic term for a homologous group of fat soluble vitamins consisting of 2-methyl-1,4-naphthoquinone derivatives.They have in common their physiological role as a required cofactor in the synthesis of blood clotting and their highly lipophilic nature. Synthetic derivatives with similar activity to natural vitamins include menadione or vitamin K3. Water-soluble derivatives of menadione such as menadione sodium hydrogensulfite (MSB) have been prepared. These compounds yield menadione after decomposition in the organism.Several methods have been proposed for the determination of MSB. In complex matrices extraction using an aqueous alcoholic solvent and subsequent back-extraction of menadione into a organic solvent after hydrolysis of the vitamin in a basic medium is usual. Chromatographic techniques have provided suitable means of determination, e.g., GC with flame ionization detection,1–3 HPLC with UV detection4 and spectrofluorimetric detection with post-column derivatization based on the reduction of the menadione to a fluorescent derivative (2-methyl-1,4-dihydronaphthoquinone). Post-column reduction with sodium tetrahydroborate (NaBH4) has also been reported.5 Recently, packed columns with powdered zinc and post-column derivatization procedures in chromatography have been used.6 Both electrochemical7,8 and photochemical reduction9 have been exploited for fluorimetric detection.Other methods include spectrophotometric detection,10,11 HPLC12 and polarographic determination of menadione in anhydrous acetonitrile with lithium perchlorate as carrier electrolyte against a calomel reference electrode.13 Finally, the British Pharmacopoeia14 describes a cerimetric titration of menadione with a previous reduction to the hydroquinone with zinc in hydrochloric acid medium with light excluded.This paper describes a flow injection (FI) procedure for the determination of vitamin K3.The method includes the on-line hydrolysis of MSB in a basic medium and subsequent chemical reduction of the menadione by a solid-phase zinc reactor (Fig. 1). The product formed was monitored spectrofluorimetrically (lex 325 nm, lem 425 nm). Experimental Reagents Analytical-reagent grade chemicals were used unless indicated otherwise. Menadione sodium hydrogensulfite was obtained from Guinama (Valencia, Spain), acetonitrile from Scharlau (Barcelona, Spain) and NaOH and HCl from Panreac (Barcelona, Spain).The zinc bed reactor was prepared with zinc ‘gross’ (Merck, Darmstadt, Germany) confined in PTFE tubing of 1.5 mm id. All solutions were prepared with deionized water. Zinc powder was immobilized in a polyester resin bed as described earlier.15,16 Apparatus The apparatus consisted of a Model 5041 rotary injection valve (Rheodyne,Cotati, CA, USA), a Minipuls 2 peristaltic pump from Gilson (Worthington, OH, USA) an F-4500 spectrofluorimeter (Hitachi, Barcelona, Spain) and flow cells from Hellma (M�ullheim, Germany) of 8 ml inner volume and 8 3 1 mm aperture and 25 ml inner volume and 11 31.5 mm aperture. The flow system was made of PTFE tubing of 0.8 mm id and a PTFE Y confluence from Omnifit (Cambridge, UK).Flow Injection Assembly The flow injection manifold is depicted in Fig. 2. A solution of MSB in acetonitrile–water (9 + 1 v/v) merged with 0.035 m Fig. 1 (a) Hydrolysis of the MSB and (b) chemical reduction of menadione.Analyst, February 1997, Vol. 122 (139–142) 139NaOH solution. The water-soluble vitamin was hydrolysed in the basic medium in loop L1 and then merged with 0.08 m HCl solution in acetonitrile–water (4 + 6). The menadione formed by means of this treatment was injected via the sample loop into the carrier solution [acetonitrile–water (9 + 1)] and forced through the zinc reactor (R) (35 cm 31.5 mm id) and into loop L3. After the reduction step, the fluorescence of the product (2-methyl- 1,4-dihydronaphthalene) was monitored (lex 325 nm, lem 425 nm).Results and Discussion Preliminary Tests Preliminary experiments allowed us to establish the chemical conditions for the spectrofluorimetric determination of MSB. A single-line assembly including a 10 cm 3 1.5 mm id reactor containing zinc immobilized on polyester resin (particle size 150–200 mm) was used to pass 40 mg ml21 solutions of MSB in methanol–water (3 + 2). Because each reaction must be conducted at a different pH, the solutions were prepared in the presence and absence of 2% sodium carbonate, followed in some cases by addition of 0.1 m HCl to pH 2.15.The same experiment was repeated in the absence of the zinc reactor. The assembly excluding the reactor provided no fluorescence signal; in the alkaline medium, a yellow colour was observed as a result of the hydrolysis of MSB to menadione; the former decomposed to a considerable extent under the action of light via a variety of complex photodegradation pathways.In the presence of the zinc reactor, only vitamin K3 exhibited fluorescence (lex = 325 nm, lem = 425 nm), following hydrolysis in the alkaline medium. Both the hydrolysis and the reduction of the vitamin took place virtually instantaneously. Selection of the Reductant Column The reductant power of the zinc reactor, its reactivity against acids and the particle size used suggested that the reducing metal could be gradually exhausted.For this reason, an experiment was performed in order to assess the influence of the reaction medium on the column lifetime. To this end, the assembly in Fig. 3 was used to monitor the fluorescence intensity of 2-methyl-1,4-dihydronaphthalene for 1 h. Subsequently, the column was flushed by passing a washing solution (deionized water) for 10 min. Finally, the reactor was again subjected to the above-mentioned exhaustion conditions for a further 1 h. The experiment was carried out with various reductants, viz., zinc and cadmium (both of which were assayed separately as powder immobilized on polyester resin in a metal to resin mass ratio of 1 : 1 and a particle size of 150–200 mm), copperized cadmium and zinc grains.The immobilized zinc and cadmium powder beds proved ineffective as they gradually lost their reducing power. Copperized cadmium exhibited a higher resistance to exhaustion; however, the fluorescence signals obtained were scarcely reproducible and considerably smaller (about 25% lower than those provided by zinc and cadmium powder under identical conditions).A 12 cm 35 mm id column packed with zinc particles of size between 0.5 and 1.2 mm was finally chosen. The readings obtained under these conditions remained constant throughout the working period (3 h). This column exhibited an increased reducing power against cadmium as a result of the increased contact surface with the zinc particles, which was adopted for subsequent experiments.Study of the Chemical System The determination of the vitamin involves the on-line alkaline hydrolysis of the sodium hydrogensulfite in menadione to give a product that is insoluble in water but soluble in most organic solvents, followed by reduction in an acidic medium. These two steps (hydrolysis and reduction) were examined in greater detail by using an FI assembly similar to that in Fig. 2. The hydrolysis step was studied using 1% sodium carbonate and NaOH at concentrations from 0.001 to 0.07 m.At equal concentrations, sodium carbonate had a slight buffering effect and released some CO2 on acidification that interfered with detection. The highest transient analytical signals (FIeaks) were obtained with 0.035 m NaOH; lower concentrations resulted in inadequate hydrolysis whereas higher concentrations increased the pH and hindered the subsequent reduction. The reduction of menadione was studied using hydrochloric, acetic, sulfuric, phosphoric and perchloric acid, all at 0.1 m concentration.The fluorescence intensity obtained was similar for all except HCl, which provided higher values. In a subsequent experiment, the HCl concentration was varied in order to study the influence of the acidity of the medium on the response of the reductant column. An acid concentration over 0.1 m caused the production of small hydrogen bubbles that interfered with detection and shortened the lifetime of the reductant bed at higher acid concentrations.An HCl concentration of 0.1 m was therefore chosen for subsequent experiments. Study of the Fluorescence of 2-Methyl-1,4-dihydronaphthalene The characteristics of the detection system used prompted us to perform several experiments in order to investigate the potential influence of different variables on the fluorescence of 2-methyl- 1,4-dihydronaphthalene, viz., the irradiation time at the excitation wavelength (325 nm), solvent, ionic strength, dissolved oxygen and temperature.The experiments were carried out in the assembly depicted in Fig. 3. The influence of the irradiation time was revealed by stopping the flow in the FI system and recording the fluorescence intensity of naphthohydroquinone in the spectrofluorimeter cuvette in a continuous manner. The emission intensity was found to decrease with time, particularly within the first few seconds. Fig. 2 Flow injection assembly for determination of MSB.Flow rates, Q: a, Qa = 3.5 ml min21, MSB in acetonitrile–water (9 + 1 v/v); b, Qb = 0.3 ml min21, 0.035 m NaOH; c, Qc = 0.3 ml min21, 0.08 m HCl in acetonitrile–water (2 + 3 v/v); d, Qd = 3.0 ml min21, acetonitrile–water (9 + 1 v/v). L1 = 330 cm; L2 = 10 cm; L3 = 25 cm. L = sample loop; V = injection valve; R = solid-phase zinc reactor (35 cm 3 1.5 mm id). Fig. 3 Flow injection assembly for preliminary studies. a, 40 mg l21 of MSB in 60% methanol, Qa = 1 ml min21; b, 1% m/v Na2CO3 solution, Qb = 0.5 ml min21; c, 0.1 m HCl, Qc = 0.5 ml min21.L1 = 49 cm; L2 = 94 cm. R = Immobilized zinc on a polyester resin (zinc to resin ratio = 1 : 1 m/m); particle size = 150–250 mm; D = spectrofluorimeter. 140 Analyst, February 1997, Vol. 122Because the polarity of the medium is known to affect excitation and de-excitation mechanisms for fluorescent species, we studied various water-miscible organic solvents including methanol, ethanol, acetonitrile, dioxane and acetone.The solutions used contained 50 mg ml21 MSB in 60% of the solvent. The blank signal was zero in every case except for dioxane and acetone, which provided no analytical signal. Acetonitrile gave the best results as the likely consequence of hydrogen bonding being less favoured than in the aqueous alcoholic mixtures, so it was adopted for further experiments. The influence of the ionic strength was determined by comparing the emission intensities recorded by replacing the 0.1 m HCl solution in the flow system with a solution also containing 4% NaCl. The analytical signals obtained in the presence of this salt were 15% smaller. The presence of oxygen in solution can have an adverse effect on the fluorescence of naphthohydroquinones as it alters the competitive reaction mechanisms that favour the formation of non-fluorescent products by a quenching effect.We therefore chose to degas the different solutions by immersion in an ultrasonic bath for 30 min.The readings obtained did not differ significantly from those recorded for the non-degassed solutions. These results are consistent with chromatographic practice regarding the use of columns packed with zinc particles that efficiently cancel the effect of dissolved oxygen. The effects of temperature on the hydrolysis and reduction reactions were studied separately. The former were examined by immersing a reactor of length L1 in a bath thermostated at 20, 35 or 56 °C (see Fig. 2). The latter were investigated by immersing the carrier solution (deionized water) in a bath at 35 °C. Increasing the temperature decreased the fluorescence emission and reproducibility. Room temperature was therefore chosen for the subsequent optimization step. Optimization of the Flow Injection Manifold The FI manifold (Fig. 2) was optimized by using the univariate method. Chemical variables (reactant concentrations) were optimized first, then the dimensions of the reductant column and finally the FIA variables. Chemical variables The NaOH and HCl concentrations were optimized jointly since, although the hydrolysis and reduction of the vitamin can be considered as two independent processes, the fact that the two channels were merged determined the pH (medium) for the second reaction (reduction): the acid must be concentrated enough to neutralize the NaOH and provide a suitable medium for reducing menadione; however, excess acid can release hydrogen and hinder detection. A 1 : 3 base to acid ratio proved appropriate.At this ratio, HCl concentrations between 0.1 and 1.5 m were tested. Concentrations near the upper end of this range released hydrogen and led to rapid exhaustion of the reactor; 0.035 m NaOH and 0.1 m HCl provided the best compromise between peak height and lifetime of the zinc column. The acetonitrile concentration in the MSB, NaOH and HCl solutions was optimized.The signal was found to rise markedly with increase in the solvent content in the MSB solution. The opposite effect, although not so marked, however, was observed in the NaOH and HCl solutions (Fig. 4). The hydrolysis and neutralization reactions took place to a greater extent in the aqueous medium, but the naphthohydroquinone exhibited more intense fluorescence in acetonitrile. A solvent content of 90% was used for the sample solution despite the increased intensity owing to the solubility limit of MSB in acetonitrile.Under these conditions, the HCl concentration must be lowered to 0.08 m owing to the formation of small hydrogen bubbles at higher acid levels. Reductant column The reductant column was optimized by altering the reactor length and diameter. The size of the zinc particles was imposed by the commercially available form of the reagent. We studied three different inner diameters for the column, viz., 5, 3.5 and 1.5 mm. The results obtained are shown in Fig. 5.A 35.5 cm 3 1.5 mm id PTFE column was selected for subsequent experiments. Larger diameters resulted in considerable sample dispersion, whereas smaller diameters and longer lengths led to increased reduction yields. However, the fluorescence signal decreased above 35.5 cm owing to increased dispersion of the sample, with no change in the reduction yield. FI variables The effects of FI variables were investigated by performing five injections of 2.5 mg ml21 MSB per value of the variable studied.The values adopted as optimum were those resulting in the best possible compromise between peak height (fluorescence intensity), reproducibility and throughput. The effects of the reactant flow rates in the sequence and of the ranges of carrier 0.47–3.6 ml min21, MSB 0.47–4.3 ml min21 and NaOH and HCl jointly 0.11–0.67 ml min21 (for both) were studied. The sample flow rate was found to be critical; the NaOH and HCl flow rates also proved influential to some extent.The sample flow rate was directly related to the amount of analyte present in the system: the greater this value the higher was the menadione concentration in the sample loop and hence the higher the fluorescence intensity. A flow rate of 3.5 ml min–1 was selected as optimum. The NaOH and HCl flow rates were altered jointly at a constant acid to base mole ratio of 3 : 1 and with minimal dilution of MSB. A flow rate of 0.3 ml min–1 for both reagents met this criterion.Increasing the carrier flow rate slightly decreased the analytical signal but increased the throughput. A flow rate of 3 ml min21 offset the dispersion produced by low flow rates without decreasing the reduction yield. The distances between merging points, between the reactor and detector, and the sample volume were optimized. Reactor– Fig. 4 Influence of the concentration of acetonitrile on the sample and reagent solutions. A, Sample; B, HCl; C, carrier; D, NaOH.Fig. 5 Influence of length and inner diameter of the zinc-bed reactor on the reduction of menadione. A, id = 5 mm, B, id = 3.5 mm; C, id = 1.5 mm. Analyst, February 1997, Vol. 122 141detector lengths greater than 25 cm decreased the signal through increased instability of the fluorescent product. The distance between the merging point for HCl and the valve was uninfluential because the acid–base reaction was instantaneous. On the other hand, MSB was hydrolysed in a fairly short time, so 330 cm was chosen as the optimum length.Finally, we studied the influence of the injected sample volume on the analytical signal. The range examined was 160–1015 ml. The fluorescence intensity was found to increase markedly with increase in the sample volume up to 850 ml, which was adopted as optimal. Once all the system variables had been optimized, those which proved critical in the optimization step (viz., the sample, NaOH and HCl flow rates) were re-optimized over narrower ranges.The results thus obtained coincided with those found in the optimization step. Analytical Applications The calibration graph was linear over the range 0.1–18 mg ml21 and fitted the equation I = 47.45 + 442.63c (n = 13), r = 0.9993, where I is the fluorescence intensity (in arbitrary units) and c the MSB concentration (in mg ml21). The day-today reproducibility was studied from five calibration curves constructed on different days.The average of the slopes obtained was 456.78 and the RSD was 2.1%. The reproducibility of the determination was calculated from replicate injections containing 10 mg ml21 of MSB (flow cell of 8 ml inner volume and 8 3 1 mm aperture). The RSD for 32 replicates was 1.6%, the sample throughput was 70 h21 and the limit of detection (three times the background noise) was 0.005 mg ml21. The sensitivity and limit of detection of the proposed method were substantially modified by increasing the aperture of the flow cell.In fact, replacing the flow cell with another one of 11 3 1.5 mm aperture and 25 ml inner volume resulted in the linear calibration graph I = 129.51 + 1451.50c (r = 0.9990). The RSD for 22 replicates and 3 mg ml21 was 1.2%, the sample throughput was 52 h21 and the limit of detection was 0.001 mg ml21. The potential effects of compounds (excipients) frequently accompanying MSB in its formulations were studied by using solutions containing the drug (10 mg ml21) and various concentrations of each interferent. The interferent concentrations tested and the relative errors they produced were as follows: ascorbic acid, 100 mg ml21, 24.1%; lactose, 500 mg ml21, 23.2%; saccharose, 400 mg ml21, 4.1%; saccharine, 100 mg ml21, 23.4%; and nicotinamide, 1000 mg ml21, 4.6%.Pectin, sodium glutamate and sulfadiazine, owing to their low solubility in water, were tested from saturated solutions and yielded errors of 0%. Finally, the proposed method was applied to the determination of MSB in three pharmaceutical formulations, Quercetol K (from Ferrer International, Barcelona, Spain), Citroflavona solution (from Funk, Barcelona, Spain) and vitamin K3 (from F.A.S., Valladolid, Spain).The results obtained were compared with those obtained with a reference method. Quercetol and Citroflavona solution were reduced to hydroquinone by zinc in hydrochloric acid with light excluded. After the zinc had been filtered off, the hydroquinone formed was determined by cerimetric titration, which re-oxidized it to quinone (according to the British Pharmacopoeia14).The determination of vitamin K3 (from F.A.S.) was carried out spectrometrically at 308 nm owing to the absence of interferences accompanying the drug. The experimental results are summarized in Table 1. Conclusions FI methodology allows the use of solid-phase zinc reactors and avoids the shortcomings associated with solid reagents: control of reactivity caused by excess of reagents and parallel reactions. The reactivity of this metal is suitable under controlled acid conditions (0.08 m HCl). The life span of the column is at least 1 month and no pretreatment of the bed reactor is necessary to obtain reproducible transient FI peaks after a few injections. The chemical reduction coupled with fluorimetric detection provide limits of detection at ppb levels and the method allows the automation of the on-line reduction of menadione as the previous step described in the British Pharmacopoeia14 as a standard method of analysis. References 1 Libby, D. A., and Sheppar, A. J., J. Assoc. Off. Anal. Chem., 1965, 48, 973. 2 Winkler, V. W., and Yoder, J. M., J. Assoc. Off. Anal. Chem., 1972, 55, 1219. 3 Winkler, V. W., J. Assoc. Off. Anal. Chem., 1973, 56, 1227. 4 Manz, U., and Mauer, R., Int. J. Vitam. Nutr. Res., 1982, 52, 248. 5 Speek, A. J., Schrijver, J., and Scheurs, H. P., J. Chromatogr., 1984, 301, 441. 6 Billedeau, S. M., J. Chromatogr., 1989, 472, 371. 7 Haroon, Y., Bacon, D. S., and Sadowski, A. J., Biomed. Chromatogr., 1987, 2, 4. 8 Wang, L. Z., Ma, C. S., Zhang, X. L., and Xu, L., Microchem. J., 1994, 50, 101. 9 Lefevere, M. F., Frei, R. W., Scholten, H. M. T., and Brinkman, U. A. Th., Chromatographia, 1982, 15, 459. 10 Ruan, Y., Duan, F., Ning, Y., and Fan, X., Fenxi Huaxue, 1988, 16, 746. 11 Iskander, M. L., Medien, H. A. A., and Khalil, L. H., Anal. Lett., 1995, 28, 1513. 12 Hu, O. Y. P., Wu, C. Y., Chan, W. K., and Wu, F. H. H., J. Chromatogr. B, 1995, 666, 299. 13 Pharmaceutical Chemistry. Vol. 2: Drug Analysis, ed. Roth, H. J., Ellis Horwood, Chichester, 1988, pp. 111–117. 14 British Pharmacopoeia, 1993, HM Stationery Office, London, 1993, p. 410. 15 Mart�ýnez Calatayud, J., and Garc�ýa Mateo, J. V., Anal. Chim. Acta, 1993, 274, 275. 16 Mart�ýnez Calatayud, J., Garc�ýa Mateo, J. V., and Lahuerta Zamora, L., Anal. Chim. Acta., 1992, 265, 81. Paper 6/06881B Received October 8, 1996 Accepted December 3, 1996 Table 1 Determination of menadione in pharmaceutical formulations Pharmaceutical Reference Error† preparation Declared value method (RSD)* Proposed method* (%) Quercetol K 5 mg per tablet 4.83 mg per tablet (0.6%) 4.77 mg per tablet 21.2 Citroflavona 0.66 mg in 5 ml 0.656 mg in 5 ml (1.2%) 0.654 mg in 5 ml 20.3 Vitamin K3 10 mg per ampoule 7.51 mg per ampoule (0.2%) 7.56 mg per ampoule 0.7 * Each value is the average of five determinations. † Error against the reference method. 142 Analyst, February 1997,

ISSN:0003-2654    

DOI:10.1039/a606881b

出版商:RSC    

年代:1997

数据来源: RSC