ISSN:0003-2654    

DOI:10.1039/AN98409FX017

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BX019

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BP033

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 533 Use of Enzymes in lmmunoassay Techniques A Review Christopher Blake and Barry J. Gould Department of Biochemistry University of Surrey Guildford Surrey GU2 5XH UK Summary of Contents Principles of im m u noassays Radioimmunoassay (RIA) Alternatives to RIA Enzyme i m m u n oassay Heterogeneous enzyme i m m u no assay Homogeneous enzyme immunoassay Enzyme-based i m mu noassay Simultaneous assay of two haptens Use of combined techniques Choice of enzyme label Chemical reactions in EIA Protein - enzyme conjugates Glutaraldehyde method Periodate method Dimaleimide method Maleimide ester method Disulphide interchange method Purification of enzyme - protein conjugates Hapten - enzyme conjugates Mixed anhydride method Ca rbodi i m ide met hod Periodate method m-Maleimidobenzoic acid N-hydroxysuccinimide ester (MBSE) and related methods Methods using bifunctional imidates Method of controlled coupling of amino compounds to enzymes Coating of solid surfaces with protein and hapten Concluding rem a r ks References Keywords Review; enzyme immunoassay; ELISA and EMIT; simultaneous assay of two haptens; hapten -enzyme conjugate formation Principles of Immunoassays Immunoassay techniques have over the past 25 years allowed biochemists to understand many physiological pathological and pharmacological processes.The initial studies by Yalow and Berson1.2 involved the use of radioactive iodine as a label for the peptide hormone insulin and radioactive atoms or groups are still the most commonly employed labels in current methodology.The basic principle of competitive immunoassays is shown in equation (1). Ab + Ag + Ag* =AbAg + AbAg* . .(1) The free antigen (Ag) and labelled antigen (Ag") compete for a fixed and limited number of specific binding sites on the antibody (Ab) molecules. After an incubation period the free and antibody-bound antigen are separated from each other and the amount of labelled antigen in one of the fractions is determined. At higher concentrations of unlabelled antigen, fewer labelled antigen molecules will be bound by antibody. Therefore a calibration graph can be produced and from this the concentration of antigen in biological samples can be determined. The number of immunoassays performed in 1983 probably exceeded 70 million and the numbers are likely to continue to increase to between 500 and 1000 million annually.3 This growth of immunoassay as a bioanalytical technique is due to a number of advantageous properties? .. it is applicable in principle to the assay of any compound; only a specific antibody is needed and these are either available commercially or can be raised in animals; it is capable of high sensitivity amounts as low as lo-'* mol have been reported as being measurable; the degree of sensitivity is dependent on the avidity of the antibody for the antigen and on the detection systems available to measure the label; little or no pre-treatment of sample is necessary; many manufacturers produce equipment for the (semi)-automation of immunoassays.Radioimmunoassay (RIA) The current predominance of the use of radioactive labels for immunoassays is partly historical. Since the initial develop-ment of RIAs many laboratories have already made the capital investment required. However radioactive labels also have advantages for immunoassays. Radioactivity can be detected with great sensitivity by simple methods that are virtually unaffected by the reaction milieu and independent of other environmental factors. Radiolabelling procedures are relatively simple as many labelled compounds are cornmer-cially available and because radiolabels are comparatively small they do not normally affect reaction kinetics.5 However radioactive labels also have drawbacks. Prepara-tion of the labelled antigen involves real risks which are cumulative.The labelled antigen shows batch-to-batch varia 534 ANALYST MAY 1984 VOL. 109 tion and generally has a shelf-life limited to 2 months, although it may be shorter if the radioactive decay causes destruction to the molecular structure. Separation of reacted from unreacted labelled compounds is essential and hence a heterogeneous system is required. The separation process usually involves centrifugation after the addition of a reagent that will selectively precipitate one of the two phases. The supernatant is then removed by aspiration or decantation. Such a separation process is difficult to automate and has limited the development of radioimmunoassay . In recent years various "non-centrifugation" separation methods have been developed for use in heterogeneous assays,f-*" but none of these systems has been widely applied.Alternatives to RIA The disadvantages of radiolabels encouraged the search for alternatives,ll particularly with the aim of developing a homogeneous assay system that does not require a separation step. Several possible labels have been investigated. These include the use of bacteriophages ~ 1 3 chemiluminescent groups and bioluminescent groups l$-19 fluorescent groups (see Soini and Hemmila20 and Visor and Schulman21 for reviews of this subject) metal atoms,22 and stable free radicals .23 Each of these labels has some positive features that may be of advantage in a particular analytical situation but all are in one way or another limited in the number of potential applications in which they can be used.11 Enzyme Immunoassay (EIA) Enzymes or molecules that interact with enzymes when employed in immunoassay methods possess more potential than any of the other labels quoted above because their use not only overcomes the problems involved in the use of radioisotopes but allows the development of a diverse range of assay protocols so enabling enzyme immunoassay to offer significant analytical advantages in a wide range of situations.The use of enzyme immunoassay has been frequently reviewed,2&37 but most of these reviews have been concen-trated on one particular area of enzyme immunoassay and other techniques were excluded from consideration. In this review we consider the various ways in which enzymes and enzyme-active substances such as cofactors prosthetic groups, substrates and inhibitors can be employed in immunoassay techniques to enable a broad understanding of the strengths, weaknesses and future potential of the various branches of the techniques to be obtained.However before discussing the different applications of enzymes in immunoassay the advantages and disadvantages of enzyme labels compared with radiolabels can be considered in general terms and are summarised in Table 1. Most of the advantages listed are self-explanatory. The reference to multiple simultaneous assays is amplified in a separate section. In the situations where EIAs are used for qualitative tests to detect the presence or absence of an infecting agent the formation of a coloured product by the enzyme reaction allows simple visual assessment of the results that is unlikely to be achieved by the use of any other type of label.Enzymes may be affected by constituents of biological samples. These may increase (activators) or more likely decrease (inhibitors) enzyme activity. This will generally only be a problem with homogeneous assays. The activity of enzymes is also affected by substrate and coenzyme concentra-tion pH temperature and ionic strength. All of these require accurate control if serious errors are to be avoided. Many of these factors are more easily controlled by dedicated auto-mated equipment which adds to the capital cost of EIAs but allows a high throughput of these assays. The fact that the assay of enzyme activity involves an extra incubation stage in comparison with RIA does not seem to have increased the time required for heterogeneous EIAs.The other disadvan-tages of enzyme labels referred to in Table 1 are mentioned where appropriate. The various assay possibilities in which enzymes are involved will now be discussed. Heterogeneous Enzyme Immunoassay Heterogeneous EIA as an analytical technique originated more than 10 years ago in the work of Van Weeman and Schuurs38 and Engvall and Perlmann.39 The principles behind the various assay protocols are identical with those in which other labels are employed and EIAs have been developed using each of the four systems illustrated in Figs. 1-4. Competitive EIA for antigen (Fig. 1) is analogous with the classical radioimmunoassay and involves competition between labelled and unlabelled antigen for a limited amount of specific antibody ( e .g . Engvall et ~ 1 . ~ 9 . As with the radioim-munoassay system a sequential additional variation of this method is possible in which the additional of the labelled antigen is delayed until the binding between the antibody and unlabelled antigen is complete (e.g. Belanger et al.41). Table 1. Comparison of enzyme with radiolabels in immunoassay ( a ) Advantages of enzyme labels (i) No radiation hazards occur during labelling or disposal of (ii) Enzyme-labelled products can have a long-shelf life e.g. 1 (iii) Equipment for enzyme assay can be inexpensive and is (iv) Homogeneous assays can be completed in a few minutes and (v) Heterogeneous assays are ideal for visual qualitative tests.(vi) Multiple simultaneous assays are possible. waste. year or more. generally available. are readily automated. ( b ) Disadvantages of enzyme labels (i) Plasma constituents may affect enzyme activity. (ii) Assay of enzyme activity can be more complex than measurement of some types of radioisotope. (iii) Less control of enyzme labelling reactions. (iv) At present homogeneous EIAs have limited sensitivity. n Antigen from +n Labelled antigen Antibody bound to labelled antigen Fig. 1. Classical competitive EIA for antigen. All components are mixed with a limited amount of antibody specific for the antigen. Separation of bound enzyme label is necessary before measurement of enzyme activit ANALYST MAY 1984 VOL.109 535 Fig. 2 shows the principle of the immunoenzymometric assay in which the enzyme-labelled antibody is reacted with antigen from the sample and then added to excess of solid-phase antigen. The enzyme activity bound to the solid phase after washing is inversely proportional to the concentra-tion of free antigen (e.g. Maiolini et 4 . 4 2 In the sandwich assay for antigen solid-phase antibody in excess is incubated with the antigen to be measured. The solid phase is then washed and enzyme-labelled antibody is added. The enzyme activity attached to the solid phase is proportional to the concentration of antigen (e.g. Clark and Adam+). The principle of this method is shown in Fig. 3. The sandwich assay for antibody detection (Fig.4) has also been employed with enzyme labels.44 In this technique the amount of enzyme activity bound to the solid phase is proportional to the amount of an antigen-specific antibody present in the sample. In addition to the four techniques detailed above a heterogeneous EIA for antigen has been demonstrated using an enzyme-labelled second antibody. In this method free antigen prevents antigen-specific antibody from binding to solid-phase antigen. Thus the amount of enzyme-labelled second antibody bound to the solid phase is inversely proportional to the amount of free antigen in the ~ample.~S This method has the advantage that one label can be used to measure any antigen for which a suitable first antiserum is available. Two other general labelling systems have been used in ELISA.One is based on the very high affinity between avidin and biotin.46 The other system utilises conjugates to Protein A which has a high affinity for the F(c) portion of immunoglobulin molecules from several animal species .47 Both systems offer advantages in sensitivity and the possibility of using common reagents that are now available commer-cially. The main disadvantage is that an additional step is required in the assay. The choice of separation method in heterogeneous EIA is limited by the large size of the enzyme label. This precludes the use of some methods used in RIA such as charcoal separation which are based on differences in relative mole-cular mass. Separation methods that have been employed include the use of second antibodies (e.g.Miyai et a1.49, Antigen from sample Labelled antibody bound to antigen Labelled anti body Solid-phase antigen bound to labelled antibody Fig. 2. Competitive ELISA for antigen. Enzyme-labelled antibody reacts specifically with antigen in the sample and is then added to excess of solid-phase antigen. After washing the enzyme label still attached to the solid phase is measured covalent chrornatography,49 and magnetic particles (polyacrylamide - agarose beads impregnated with iron oxide) linked to either antigen or antibody.50 A variety of solid phases have also been used. Some such as cellulose, agarose and polyacrylamide particles require centrifugation and as such have little advantage over the use of second antibodies.It is however in the use of solid supports in which no centrifugation is needed that the heterogeneous EIA methods have made most impact. The use of such solid supports in both competitive and non-competitive techniques was given the name Enzyme (excess) Antigen from 11 sample Antigen bound to Labelled solid- phase antibody antibody Labelled antibody attached to solid phase via antigen from sample h Fig. 3. Sandwich assay for antigen. Antigen in the sample is mixed with excess of solid-phase antibody. After washing of the solid phase enzyme-labelled antibody which is specific for another site on the antigen is added. The enzyme label which remains bound after washing is measured n antigen + II + La belled second antibody solid-phase antigen Labelled second antibody attached to solid phase via antibody from sample h Fig.4. Sandwich assay for antibody. Antibody in the sample is mixed with excess of solid-phase antigen. After washing of the solid phase enzyme-labelled second antibody is added. Bound enzyme activity is measured after washin 536 ANALYST MAY 1984 VOL. 109 Linked ImmunoSorbent Assay (ELISA) by Engvall and Perlmann.39 ELISA has been used with enthusiasm particu-larly by those working on infectious diseases where the need is often for the detection of antibodies. The indirect sandwich assay illustrated in Fig. 4 has proved particularly applicable. The solid support most frequently used in the large-scale screening work required in the study of infectious diseases has been the microtitration plate.51 These plates are cheap and require only small volumes of reagents.Commercially avail-able equipment now allows incubation washing and direct “through-the-plate” reading of absorbance or fluorescence. An inexpensive portable battery-operated photometer for reading ELISA tests in microtitration plates has been devel-oped by Rook and Cameron.52 Antigens and antibodies have also been adsorbed on to the walls of spectrophotometric cuvettes and used in single-tube ELISA methods in which the enzyme activity was determined directly using reaction rate analysis.53,54 Microtitration plates and multiple cuvette systems may suffer from “edge effects,” which are caused by inconsistent antigen-binding characteristics.55 An alternative, commercially available system using coated beads should not suffer from this disadvantage.Other systems using a dip-sticks6 or a syringe57 allow for easy washing between incubation stages and may be more suitable for small numbers of qualitative assays. One novel approach to heterogeneous EIA recently des-cribed is the use of tagged enzyme - ligand conjugates.58 In this assay the insoluble phase is provided by a receptor to the tag which has been linked to the enzyme - ligand conjugates. The receptor is insolubilised by being joined to Sepharose. Competition for antibody binding occurs between free ligand from the sample under test and the tagged enzyme - ligand conjugates. Antibody binding serves to mask the tag on the conjugates so that it can no longer bind to the insolubilised receptor.Accordingly the amount of enzyme activity asso-ciated with the insoluble fraction is proportional to the amount of free ligand in the sample under test. The model system used biotin as the tag and immobilised avidin as the receptor. The major advantage of this system is that the insolubilised receptor is a “universal” separation reagent and can be used for any ligand or enzyme provided the enzyme is linked to the same tag. The use of a solid-phase receptor would eliminate the centrifugation process and be an additional advantage of the technique. A novel separation method for use in a heterogeneous EIA was reported by Monji and Castro.59 These workers used 0-galactosidase as a label and an insoluble pseudosubstrate of the enzyme agarose - aminocaproyl-P-galactosylamine to separate the phases in assays for human choriomammotro-pin.Monji and Castro referred to their assay as Steric Hindrance Enzyme Immunoassay (SHEIA). When the incu-bation- of enzyme-labelled hormone sample containing the free hormone and antibody is complete the insolubilised pseudosubstrate is added. The hormone - enzyme conjugate, which is antibody bound is prevented from binding to the pseudosubstrate by steric hindrance whereas the conjugate, which is not antibody bound can reversibly attach to the pseudosubstrate. Either the bound or free enzyme activity can be measured and used for the quantitation of the hormone being studied. One other variation of the solid-phase technique that deserves mention is the Thermometric Enzyme Linked Immunosorbent Assay (TELISA).In this method the enzyme reaction is monitored by measuring the heat produced in a continuous-flow system.60 Too many variables are involved to allow general state-ments as to the sensitivity of heterogeneous EIA in compari-son with other immunoassay techniques.26 The importance of some such variables has been investigated in the EIA of steroids by Joyce et al.61 These workers found that the sensitivity was influenced both by the separation method used and the molar ratio of steroid to enzyme in the conjugate employed. Statements have been made that radioimmunoassay is always more sensitive than EIA (e.g. Watson62 and Ekins63), but much of the experimental data reviewed by Schuurs and Van Weeman26 in 1977 showed that such a generalisation was not justified.Indeed with the use of fluorescent detection methods that have been made easier by the recent marketing of an automatic microplate fluorescence reader the potential sensitivity of EIA is high. Using fluorescent assay of CJ-galacto-sidase and affinity-purified antibodies and enzyme-labelled conjugates one group of workers have been able to measure 0.2 attomoles (amol 10-18 mol) of ferritin.64 Another group used a fluorogenic substrate with alkaline phosphatase and were able to detect attograms of antigen.65 Of course many EIAs described in the literature are less sensitive than the corresponding radioimmunoassays but it does not seem that this difference is caused by factors inherent in the two types of assay.27 Certainly for small molecules where the attachment of a label of high relative molecular mass such as an enzyme, would theoretically influence the reactivity of the molecule more than the presence of the smaller radioisotope the observed sensitivity is not always less than that seen for radioimmunoassay .One final point should be made in reference to hetero-geneous EIA which is that although serum factors are removed prior to the measurement of enzyme activity and cannot therefore interfere in that reaction interference in the immunological process of an heterogeneous EIA system has been noted.66 The inclusion of digested gelatin in the reaction buffer was found to remove this interference,66 but the two publications by Kato et ~1.66367 serve as a useful reminder that even in heterogeneous systems the presence of serum factors may produce undesirable non-specific effects.Homogeneous Enzyme Immunoassay The Enzyme Multiplied Immunoassay Technique (EMIT; Syva Maidenhead UK) is the most widely used homogen-eous enzyme immunoassay system. The assay principle is shown in general terms in Fig. 5. The technique was introduced by Rubenstein et al.68 and depends on a change in the specific enzyme activity when antibody binds to enzyme-labelled antigen. The activity of the unseparated assay mixture corresponds to the proportion of enzyme-labelled antigen to which antibody is bound. Three distinct variants of the EMIT principle have been devised and applied commercially. The earliest EMIT systems68 employed lysozyme as the enzyme label.The natural substrate for this enzyme is the peptidoglycan of the cell wall of certain bacteria. When lysozyme acts on a suspension of Micrococcus luteus the turbidity of the suspen-sion is reduced thus providing a convenient measure of enzyme activity. Lysozyme - hapten conjugates are active with the bacterial suspension but when hapten antibodies bind to an enzyme - hapten conjugate a steric barrier is presented to the bulky substrate. Antibody binding in this way can inhibit up to 98% of the enzyme activity.@ This effect is employed in the EMIT assays for drugs of abuse.”) Enzymes with small substrates have also been successfully used in EMIT assays. Hapten conjugates of glucose-6-phosphate dehydrogenase and malate dehydrogenase were found to be inhibited up to 80% by excess of hapten antibodies.The mechanism with malate dehydrogenase was investigated by Rowley et al.,71 who concluded that the inhibition was due to conformational effects. It is most likely that an inhibitory conformational change is induced by the antibody binding to haptens linked to certain active groups on the enzyme surface although an alternative explanation would be that a conformational change required for activity is prevented as a result of the antibody binding69 The dehy ANALYST MAY 1984 VOL. 109 537 FJ Haptenfrom Fig. 5. lJ Antibody bound to hapten sample ’ Limited amount ofantibody \ Substrate Product Antibody bound to hapten Hapten-la belled inactivates enzyme active enzyme Principle of the EMIT system of homogeneous enzyme immunoassay.conjugation of hapten to enzyme does not destroy enzyme activity but combination with hapten-specific antibody causes marked inhibition of enzyme activity. The measured enzyme activity is dependent on the relative amounts of free hapten and hapten-labelled enzyme drogenase enzymes are used in this way in the EMIT assays for therapeutic drugs (e.g. Pippenger et al.72). In the lysozyme- and dehydrogenase-based assays described above antibody binding results in an inhibition of the activity of enzyme - hapten conjugates. There is one EMIT assay, however in which antibody binding stimulates the activity of the conjugate namely the assay for thyroxine. In this method, the conjugation of thyroxine to malate dehydrogenase results in a substantial inhibition of enzyme activity.The binding of the thyroxine antibodies to the conjugate can in certain circumstances partly reverse the inhibition. These complex phenomena have been studied in detail by Ullman et al. ,73 and have been used as the basis for the assay for thyroxine.74 Until 1979 the EMIT system had been applied only to small molecules but Gibbons et al.75 investigated the use of the technique to measure proteins. They realised that the confor-mational effects of the binding of the antibody to the enzyme-linked antigen as utilised in the dehydrogenase-based systems were likely to be much smaller when the antigen was a protein rather than a small molecule. Steric considerations restricted the number of protein molecules that could be attached to an enzyme molecule.Using a conforma-tional change-based system the antibody binding to the protein would need to exert effects across the bulk of the antigen. Preliminary experiments with dehydrogenase protein conjugates confirmed the expectation that antibody binding would cause only weak inhibition.6YIt seemed more likely that the steric exclusion of a macromolecular substrate as utilised in the lysozyme-based systems would be of use in the measurement of proteins provided that the presence of the antigen proteins attached to the enzyme did not prevent the substrate reaching the active site. The enzyme P-galactosidase was selected for use because of the absence of serum activity at the pH optimum of the bacterial enzyme used for conjugation, and because the enzyme has been shown to retain its activity even after conjugation to protein.76 The substrate o-nitro-phenyl galactoside was converted into a macromolecular form by attachment to a dextran carrier via a hydrophobic spacer.Inhibition of enzyme activity was observed up to a maximum of about 80% as a result of the addition of excess of antibody, and this provided the basis for the homogeneous assay of free antigen. The expansion of the EMIT system to enable large molecules to be measured is a significant extension to its usefulness. Another novel assay recently produced by the same group of workers is “enzyme enhancement imrnunoa~say.”~~ This is an assay specifically for polyvalent ligands and antibodies and avoids the need to label the antigen.The technique involves antibodies labelled with P-galactosidase succinylated antibo-dies and a macromolecular form of the o-nitrophenyl galacto-side substrate. The enzyme-labelled antibody and the succiny-lated antibody form an immune complex in the presence of sample antigen An enzyme within this negatively charged microenvironment forms a product that gives a second light-scattering phase whereas the product formed by free enzyme remains soluble. Thus the amount of antigen in the sample under test modulates the rate of increase of light-scattering. Two other variations of the EMIT system are an “end-point” approach and the use of “enzyme channelling immu-noassay.” The end-point approach to the EMIT assays78 involves the use of antibodies against the enzyme label to inhibit enzyme activity completely.In an enzyme-channelling immunoassay the immune reaction places two enzymes which catalyse consecutive reactions in close proximity. In the method of Litman et al. ,79 glucose-4-phosphate dehydroge-nase is co-immobilised with the antigen on agarose beads. Hexokinase which produces the glucose-6-phosphate that acts as the substrate for the dehydrogenase is linked to the antibody. Binding of hexokinase-labelled antibody to the bead-bound antigen results in an accelerated conversion of glucose ATP and NAD+ into 4-phosphogluconolactone, ADP and NADH. The reaction is monitored via the increase in NADH concentration. This approach could be used to increase the specificity of assays for macromolecules if the two enzymes were attached to monoclonal antibodies raised to adjacent antigenic sites on the macromolecule.80 This would eliminate interference by other molecules containing one but not both of the antigenic sites.The major advantage of the EMIT system is that it can be automated with relative ease and then produces rapid results. The commercially available “kit” method has been adapted for use with centrifugal analysers,sl continuous-flow systems8* and automated kinetic reaction equipment.83 As with other homogeneous immunoassays the EMIT method is prone to interference from non-specific factors present in the biological fluid under test. Interference has been shown to occur due to the presence of hyperlipidaemia,84 myeloma immunoglobulin^^ and excess of sodium chloride in urine samples.8h787 This last effect is due to an increase in ionic strength and produces false negative results.Salt has been added to urine samples by drug addicts wishing to avoid detection by the EMIT drugs of abuse assays.86 Enzyme-based Immunoassay The various ways in which enzymes have been employed directly as immunoassay labels have been discussed above. However molecules acting with or upon enzymes have also been used as immunoassay labels. Such assays depend on the steric hindrance caused by antibody attaching to the hapten-labelled conjugate modulating the enzyme activity by altering the natural function of the label with the enzyme. Enzyme cofactors prosthetic groups substrates and inhibitors have all been shown to be suitable for use as labels in the development of homogeneous enzyme-based immunoassay systems (Figs.6-8). These methods have two potential advantages as the enzyme is not labelled as it is with EMIT. This allows a wider range of reaction conditions to be used as the cofactors, prosthetic groups substrates and inhibitors are more stabl 538 ANALYST MAY 1984 VOL. 109 sample + to hapten Hapten from Limited amount ofantibody \ a + P Hapten-labelled prosthetic group Apoenzyme (inactive) Antibody bound to labelled 4 ( 1 hapten such that prosthetic group is not available to n a Doenzvme Product k U Fig. 6. Principle of “Prosthetic Group-Labelled Immunoassay” (PGLIA). The apo-enzyme only becomes active when it has been converted into the holoenzyme by binding to the prosthetic group, which is part of the active site of the enzyme than enzymes.The products of these reactions should also be more specific than reactions involving enzymes which gener-ally have several comparable functional groups many of which are likely to react. The use of enzyme cofactors was introduced in two consecutive publications.g”s9 The technique takes advantage of the amplifying effects of enzymic cycling. The method requires that the ligand - cofactor conjugate be active in enzyme-catalysed reactions and that the activity be modified when the ligand is bound to a specific protein. In the published methods ligands were conjugated to nicotinamide 6-(2-aminoethy1amino)purine dinucleotide (AENAD) a deriva-tive of nicotinamide adenine dinucleotide (NAD).The conjugate was active with dehydrogenases with a reduced form of the cofactor conjugate being produced (i.e. ligand -AENADH). The ligand - AENADH was then recycled its oxidation being responsible for the formation of either reduced thiazolyl blue ,88 or bioluminescence via a luciferase reaction.89 The cofactor cycling system employed in the initial studies gave about 200 cycles h-1 with the conjugates.88 The use of cycling systems with higher cycling rates would improve the sensitivity of the assay and alcohol dehydrogenase which has been used to cycle NAD at 30000 cycles h-1 was used in the later studies.89 The assay is homogeneous rapid (using bioluminescence incubations of only 1 min were required) and has the potential for the production of sensitive assays.The use of a prosthetic group as a label in immunoassay could also lead to the development of sensitive systems as illustrated in Fig. 6. Such a method for the detection of haptens was reported by Morris et ~ 1 . 9 0 In their method the prosthetic group was joined covalently to the ligand in such a way that the conjugate was able to combine with inactive apoenzyme to form the active holoenzyme. The assay can be Hapten from sample + Hapten labelled with enzyme substrate I non-f luorog enic) r n /Antibody bound to hapten G i’ Antibody bound to hapten such that substrate is not available to enzyme b’D Products formed one of which fluoresces Fig.7. Principle of “Substrate-Labelled Fluorescent Immunoassay” (SLFIA). Only a limited amount of the substrate-labelled hapten can be used and therefore only a small amount of substrate is available for reaction which necessitates the formation of a fluorescent product to obtain adequate sensitivity. Either the rate of formation of fluorescent product or the total amount formed can be measured carried out without the need for separation because the ability of the prosthetic group - ligand conjugate to regenerate active holoenzyme is substantially inhibited when the labelled ligand is complexed with its antibody. Lin and Pardue91 have studied the kinetics of this relatively complex assay system and have shown that reliable results can be obtained within 3 min of mixing.The model system used flavin adenine dinucleotide (FAD) as the prosthetic group and glucose oxidase as the enzyme in a colorimetric immunoassay for theophylline. All of the reagents required were well defined and readily prepared. The assay was rapid and required only a 1-yl sample of serum for assays in the therapeutic range of theophylline. The system has also been applied to the measurement of protein mol-ecules.9” The prosthetic group label system has also been adapted to a reagent strip.92 Enzyme substrates have also been employed as immuno-assay labels particularly in a method that combines features of both enzyme-based and fluorescence immunoassay tech-niques. The method has been given the name “substrate-labelled fluorescent immunoassay” (SLFIA) and is illustrated in Fig.7. It was developed by Burd and co-~orkers93,93 and has been used in a variety of assay systems (e.g. Wong et al.95 and Ngo et a1.96). In the method neither the labelled ligand nor the antibody-bound ligand is fluorescent. However the labelled ligand can be cleaved by enzymatic hydrolysis to release part of the labelled molecule which is then able to fluoresce. Cleavage of the labelled ligand is sterically preven-ted when it is bound to the specific ligand antibody. Competition for antibody binding sites by ligand from the sample under test makes more of the labelled ligand available to act as an enzyme substrate. The extent of the fluorescence is proportional to the amount of free ligand in the test sample. I ANALYST. MAY 1984 VOL.109 539 Hapten from sample Hapten labelled with enzyme inhibitor R Limited amount ofantibody \ + II +El Antibody bound to hapten Enzyme (active) U Antibody bound to hapten Ik such that inhibitor cannot also bind to enzyme Inhibited enzyme (in active) Fig. 8. Principle of an inhibitor-based homogeneous enzyme immuno-assay. The inhibitor-labelled hapten combines with enzymes that lose activity. In principle the reaction between enzyme and inhibitor could be reversible (as shown) or irreversible the assay for gentamicin,g4 galactosidase was used as the enzyme and the synthetic fluorogenic substrate was a @-galactosylumbelliferone - gentamicin conjugate. The sensi-tivity of the assay which was comparable to that of radioim-munoassay relies on fluorescence rather than any amplifica-tion of the enzyme.The assay method is homogeneous rapid and simple to perform. Only 1 pl of sample is required and linear calibration graphs are obtained. The SLFIA system has also been adapted to a dry-strip format97 in which the results obtained compare well with those found using established methods. Enzyme inhibitors are another group of enzyme-asso-ciated substances of potential use in the development of homogeneous enzyme-based immunoassays. If a homogen-eous system were to be developed then the hapten - inhibitor conjugate would need to exert its inhibitory action while present free in the reaction solution but be sterically hindered from interacting with the enzyme when the antigenic portion of the conjugate was antibody bound (Fig.8). The possibility of developing such an inhibitor-based system was first noted by Wisdom25 in 1976 but it was not until 1980 that the first publication concerning practical investigations was produced. A consideration of the reactions involved indicates that the inhibitor used must be a potent one having an affinity for the enzyme of 10-8 M or stronger if sensitive assays are to be developed. Ngo and Lenhoff98 highlighted this fact when they outlined the principle of the inhibitor-based system and noted that such affinities could be achieved by the use of either anti-enzyme antibodies or enzyme receptors as well as more conventional enzyme inhibitors. In fact the experimental work they reported was confined to the use of anti-enzyme antibodies in that they detailed the development of an assay for dinitrophenyllysine in which the acitivity of a peroxidase enzyme could be inhibited by antiperoxidase antibodies linked to dinitrophenol residues, provided that the dinitrophenol portion of the conjugate was not bound to its specific antibodies.Ngo and Lenhoffqs pointed out the advantage of the use of inhibitors over the use of substrates in that the inhibitor approach provides an intrinsic amplifying power because the uninhibited enzyme can continuously generate products. In their closing comments Ngo and Lenhoff noted that there are numerous theoretical ways in which potent modulators of enzyme activity could be employed in immunoassay systems. They listed as examples the use of the avidin - biotin interaction (affinity constant 10-14 M) the use of transition-state analogues such as coformycin (which inhibits adenosine deaminase with an inhibition constant of 10-10 M) and the use of enzyme activators which can produce a dramatic increase in the activity of some mutant enzymes.Later in 1980 the first practical use of a conventional enzyme inhibitor in a homogeneous enzyme immunoassay was published by Finley et al.99 These workers evaluated an enzyme inhibition immunoassay for serum thyroxine. The method is based on the inhibition of hydrolysis of the substrate acetyl-P-(methy1thio)choline iodide by acetylcholine esterase (acetylcholine hydrolase). Thyroxine covalently linked to a cholinesterase inhibitor (a phosphonate) irreversibly inhibits the enzyme when free but is not inhibitory when bound to thyroxine antibodies.N-Methylorphenandrine sulphate is added to the reaction mixture to block any interfering pseudocholinesterase present in the serum. This method has been used to produce a commercial "kit" assay for thyroxine but it cannot be used with haemolysed samples because of the possibility of interference by ery-throcyte membrane-bound acetylcholinesterase. Owing to the fact that depending on genotype the enzyme may be tightly or loosely bound to the cell membrane the results of tests on haemolysed samples vary unpredictably and Finley et al.99 sensibly suggest that samples that are even slightly haemo-lysed should not be assayed. Despite the limitation of this method it is possible that enzyme inhibitors may yet be important in the future development of immunoassays.Simultaneous Assay of Two Haptens One of the theoretical advantages given in Table 1 resulting from the use of enzymes as immunoassay labels is the possibility of measuring more than one hapten simul-taneously. This possibility has been noted by at least two previous reviewers (Wisdom25; O'Sullivan et al.31) but the first working model system has only recently been devel-oped.1"" The advantages of simultaneous multiple-hapten systems compared with single-hapten methods include savings on reagent costs and sample volumes and a reduction in the over-all assay time. One or two other labels have been employed in the development of dual assay systems but at present each would seem to have significant limitations.The use of the two iodine isotopes 1251 and 1311 has been the most commonly used approach (e.g. Mitsuma et al.")l and Haynes and Goldie102), but their usefulness is hindered by the short half-life of I311 and the less than ideal distinction between the gamma emissions of the two isotopes. The use of enzyme labels in dual measurements offers both stability and ease of distinction between the signals produced. In the method of Blake et a1.,1()0 alkaline phosphatase was conjugated to thyroxine and detected at 540 nm by the formation of phenolphthalein from phenolphthalein mono-phosphate while @-galactosidase was linked to triiodothyro-nine and detected at 420 nm by the formation of o-nitrophenol from o-nitrophenyl P-galactoside.The assay was carried out in a heterogeneous manner using double antibody precipitation. The choice of the two enzymes enabled each to be measured in the presence of the other without cross-detection and allowe 540 ANALYST MAY 1984 VOL. 109 the determination of both thyroid hormones in a single tube. The assay compared well with existing methodology and confirmed that the theoretical advantages in time and cost savings could be achieved in practice. Dean et a1.103 have described an assay for the simultaneous determination of phenytoin and phenobarbital using substrate-labelled fluorescent immunoassay. The assays, although of adequate sensitivity required staggered reading because both fluorophore products have maximum fluores-cence emissions at the same wavelength.Use of Combined Techniques In recent years there has been increasing interest in the use of combinations of two techniques as the basis for immunoassay protocols. These methods do not easily fit into any classifica-tion of the methods available but as such techniques may play an important role in the future they are discussed here. Three examples of combination techniques in which enzymes are involved will be given. One such combination technique is the viroenzymoim-munoassay for penicillin that was developed by Mamas and Dray. 104 The classical viroimmunoassay (e.g. Haimovich and Sela12) involves the counting of plaques produced by bacterial lysis. In viroenzymoimmunoassay the bacterial lysis is moni-tored not by plaque formation but by the measurement of the release from the bacteria of the endocellular enzyme P-galac-tosidase.The combination method is similar to viroimmuno-assay in all other respects in that competition for antibody binding occurs between bacteriophage-labelled hapten and free hapten from the sample under test. The performance of the assay is better than pure EIA methods and comparable to classical viroimmunoassay techniques. However the viroenzymoimmunoassay is much quicker than the classical plaque-counting method and is more suitable for automation. Enzyme-enhanced luminescence immunoassay is a combi-nation method that has only recently been developed.105 The model system involved transferrin labelled with pyruvate kinase. After the antibody - antigen reaction and subsequent bound - free separation the enzyme was used to generate adenosine triphosphate (ATP) from adenosine diphosphate (ADP) and phosphoenolpyruvate.The ATP was then moni-tored using a firefly luciferin - luciferase reagent. The assay for transferrin compared well with an established radioim-munoassay procedure. The authors point out that for hapten assays enzymes with a lower relative molecular mass than that of pyruvate kinase would appear to be more suitable. The method is not limited to the ATP - ADP system as other enzymes with high turnover rates could be linked to a bacterial NADP - NADPH-dependent luciferin - luciferase. The third and in some ways most promising combination technique discussed in this section is radioenzyme immuno-assay.The method involves the use of tritium and hence does not possess all the advantages of a non-radioactive system but it does allow the development of assays of very high sensitivity without the use of the potentially more hazardous radioactive iodine. The principle was initially discussed in a short publication by Van Der Waart and Schuurs.106 These workers prepared a conjugate between human chorionic gonadotropin (hCG) and acetylcholinesterase. Competition between free and labelled antigen for binding to anti-hCG antibodies that had been adsorbed to microtitre wells was allowed to take place. After washing of the solid phase the bound enzyme activity was measured using tritiated acetylcholine as substrate. The contents of the wells were transferred into glass tubes and extracted with toluene - isoamyl alcohol and the radioactivity present in the non-aqueous phase was determined.The authors stated that their assay was not optimised but claimed that it was more sensitive than either radioimmunoassay or enzyme immunoassay for hCG. The first detailed use of this system was described by Harris and co-workers,1"7.1"8 who refer to the technique as "ultra-sensitive enzymatic radioimmunoassay" (USERIA). These workers used both competitive and non-competitive assay protocols but in all systems the final stage was the addition of an alkaline phosphatase - goat anti-IgG conjugate. After washing of the solid phase the bound alkaline phosphatase activity was measured by the conversion of tritiated adenosine monophosphate into tritiated adenosine.Product and sub-strate were separated using small DEAE-Sephadex columns and the enzyme activity observed was used to calculate the antigen concentration in the sample under test. The USERIA for cholera toxin was claimed to be 100-1 000-fold more sensitive than either of its two parent assays. The detection limit was 10-16 g of cholera toxin which is equal to approximately 600 molecules. When applied to the assay of acetylaminofluorine - DNA adducts108 the detection limit was 2 fmol which was a 60-fold increase in sensitivity over available radioimmunoassay methods. The USERIA system is recognised as having potential for the future ( e . g . Ekinslog) particularly if simpler methods of substrate - product separation can be developed.Choice of Enzyme Label Enzymes function as useful labels because they are very efficient catalysts. A single molecule of enzyme typically converts 103-104 molecules of substrate into product per minute but for some enzymes this figure can be as high as 106-107. Ideally the enzyme label should have the properties listed in Table 2 although it is the properties of the conjugated form that are most critical. The velocity (v) of an enzyme-catalysed reaction [equation E + S s ES E + P ( 2 ) enzyme substrate enzyme - substrate enzyme product intermediate (211 is usually described by the Michaelis - Menten equation [equation (3)] . . . . . * (3) V[Sl v=-K + [S] * . and is dependent on the substrate concentration S and two constants the maximum velocity (V) and the Michaelis constant (Km).It follows that enzymes with high V and low K will be the most sensitive labels. The requirement for a simple spectrophotometric assay is particularly relevant to qualitative assays and to those laboratories without more sophisticated equipment. Reactions are normally followed by the formation of product which should have a high molar Table 2. Properties of an ideal enzyme label (i) High enzyme activity (V) at low substrate concentrations (low (ii) Enzyme stable at the pH required for good antibody - antigen binding (generally neutral pH) and active at this pH particularly for homogeneous EIA. (iii) Cheap accurate and sensitive assay method preferably with a spectrophotometric end-point. (iv) Presence of reactive groups through which enzymes can be covalently linked to antibody.antigen or hapten with minimum loss of enzyme or immune activities. (v) Enzyme-labelled conjugates which are stable under routine storage and assay conditions. (vi) Availability of soluble purified enzyme at low cost. (vii) Absence of health hazards attributable to enzyme substrates and cofactors. (viii) Absence of enzyme activity and factors affecting enzyme activity from the test fluid (particularly for homogeneous EIA). K m ) ANALYST MAY 1984 VOL. 109 54 1 absorptivity. The enzymes that are most commonly used as labels together with some of their properties and the end-points used are listed in Table 3. Several of the enzymes are from bacterial or vegetable sources.The enzyme labels most commonly used for heterogeneous EIAs are peroxidase alkaline phosphatase and 8-galactosidase; all three can be detected at femtomole levels by spectrophotometry. Peroxidase is the cheapest of these enzymes and has a carbohydrate content of 1&15% which is utilised in certain conjugation reactions. Several chromogens are available that produce dark colours and are suitable for visual determinations. However most of these chromogens, including benzidine o-dianisidine o-phenylenediamine (OPD) o-toluidine and 2,2'-azinodi(3-ethylbenzothiazoline-6-sulphonate) (ABTS) have been shown to be carcinogenic or mutagenic in test systems.363111 Despite this two of these three are amongst the most commonly used chromogens which are OPD ABTS and 5-aminosalicylic acid (5-AS).These were compared by Al-Kaissi and Mostratos112 and found to be of comparable sensitivity when measured at their optimum wavelengths i.e. OPD 475 nm ABTS 414 nm and 5-AS 500 nm. Poor solubility is a problem with 5-AS and its product. If 5-AS is recrystallised in the presence of sodium hydrogen sulphite this improves the sensitivity and reduces the background.113 It is known that peroxidase is inactivated by contact with polystyrene surfaces but this inactivation is prevented if Tween is added to the reaction mixture or used to pre-treat the surface.114 The search for sensitive safe chro-mogens for peroxidase continues and further examples are available.ll5.116 Glucose oxidase can be measured with the same chromogens as for peroxidase but as it has a lower specific activity the assays tend to be less sensitive.However, glucose oxidase does have the advantage that it is absent from human plasma and thus the enzyme could have applications in homogeneous EIA. Alkaline phosphatase and its conjugates are very stable, and the colorimetric and fluorimetric substrates that give sensitive assays are safe chemicals. The main disadvantage is that the purified calf intestine enzyme which is preferred because of its relatively high specific activity compared with other sources is expensive as the supply of source material is limited.36 P-Galactosidase which is absent from plasma is an enzyme that is used for both heterogeneous and homogeneous EIAs. The reactions with one of the colorimetric substrates 0- or p-nitrophenyl-P-D-galactoside or the fluorogenic substrate, 4-methylumbelliferyl-(3-~-galactoside are easily followed.The latter substrate allows 1 amol of enzyme to be detected after incubation for 60 min.117 Several other enzymes listed in Table 3 have been used for heterogeneous assays. Acetylcholinesterase has a high specific activity and a low K value. It has been measured with a glass pH electrode118 or by a spectrophotometric method.99 Adeno-sine deaminase has a low I( value. It has been measured with an ammonia gas-sensing electrode at levels of 60 pmol.1'9 Catalase has the highest specific activity quoted for any enzyme but its product absorbs in the ultraviolet range which is generally unsatisfactory when assaying biological materials.Another detection method is based on thermometry but its sensitivity is limited to 100 pmol at present.120 Urease has been detected in two ways in EIAs. The spectrophotometric assay in which the ammonia produced reacts with bromocresol purple has been used for a simple sensitive and rapid qualitative assay.57 However stability problems of conju-gates probably owing to free thiol groups have been reported.121 A rapid sensitive assay using an ammonia sensitive electrode has also been published. 122 The three enzymes most frequently used for homogeneous EIAs were discussed earlier. Lysozyme is really of historical interest as it was the first enzyme used for an homogeneous Table 3. Enzymes commonly used as labels for EIA. Data compiled for Barman110 and M a g g i ~ ~ ~ EIA Heterogeneous Homogeneous Enzyme Source Acetylcholinesterase Electrophorus Adenosine deaminase Calf intestine electricus Alkaline phosphatase Calf intestine Catalase Calf liver P-Galactosidase Escherichia coli Glucose oxidase Aspergillus niger Peroxidase Horseradish Urease Jack beans Acetylcholinesterase Electrophorus 6-Galactosidase Escherichia coli electricus Glucose 6-phosphate Leuconostoc dehydrogenase mesenteroides Lysozyme Chicken egg white Malate dehydrogenase Pig heart pH Specific activity 7-8 1400 90 VM optimum at 37 "C/units mg-* K m 7.5-9 200 60 WM 8-10 1000 0.2 mM for PNPP 6-8 40 OOO * 6-8 600 1 mM 4-7 200 C ' " = 3 3 mM; K:2=0.2 mM 5-7 4 500 * 6.5-7.5 10 OOO 10 mM 7-8 1400 90 VM 6-8 600 1 mM 7.8 400 g h P = O .l mM K:AD= 0.15 mM * 4.5-5.5 -8.5-9.5 1 000 K'"'=0.3mM K P D = O . 1 rnM Relative molecular mass 54 000 100 OOO 250 000 540 OOO 186 000 40 000 483 000 54 000 540 000 104 000 14 5 0 70 000 End-point pH electrode or spectrophotometric Ammonia gas-sensing electrode Spectrophotometric or fluorimetric UV absorbtion or thermometry Spectrophotometric or fluorimetric H202 combines with chromogen H202 combines with chromogen Ammonia reacts with chromogen or gas-sensing electrode pH electrode or spectrophotometric Spectrophotometric or fluorimetric Formation of NADH by UV absorption or fluorimetric Formation of cell wall fragments (AA at 450 nm) Formation of NADH * K depends on substrate 542 ANALYST MAY 1984.VOL. 109 EIA.68 Glucose 6-phosphate dehydrogenase from L. mesen-teroides is useful as it has high activity with NAD+. This overcomes the possibility of interference in haemolysed samples as the corresponding human red cell enzyme is specific for NADP+. Malate dehydrogenase is probably the dehydrogenase with the highest specific activity but its use in homogeneous assays is limited to urine analyses. The two dehydrogenases are assayed at the absorption peak of NADH, which is at 340 nm. Chemical Reactions in EIA Certain chemical reactions are essential for successful EIA. These include protein - protein and protein - hapten conjuga-tion reactions. Protein or hapten coating of solid surfaces for use in ELISA may be by chemical reaction or more usually by physical adsorption.The major reactive groups of proteins that are employed are carboxyl amino sulphydryl and phenolic groups which are generally most reactive in their non-protonated form. Protein - Enzyme Conjugates Glutaraldehyde method Proteins are readily cross-linked by reaction with the bifunc-tional reagent glutaraldehyde mainly through the E-amino groups of lysyl residues in proteins. The exact chemical reaction is uncertain. The cross-linkages are more stable than are found with Schiff‘s base. 123 A proposal that an unsaturated condensation polymer of glutaraldehyde is involved124 is not consistent with the monomeric nature of glutaraldehyde, which may react with amino groups to form quaternary pyridinium compounds.125 When glutaraldehyde was added to a mixture of an enzyme and an immunoglobin the conjugates produced were hetero-geneous and of high relative molecular mass.126 This is mainly due to polymerisation of IgG with consequent loss of antibody activity which is found particularly when using peroxidase or alkaline phosphatase.40.127 Some of these problems are reduced by the two-step glutaraldehyde method. The enzyme normally used is peroxidase as it has few free amino groups. It is treated with glutaraldehyde the excess of glutaraldehyde is removed and then the activated enzyme is allowed to react with the antibody protein. 128 Enzyme polymers are formed, but in the method of Boorsma and Streefkerkl29 a conjugate preparation containing an enzyme to antibody molar ratio of approximately 1:l is more easily obtained.The build-up of large enzyme - protein complexes can also be prevented by carrying out the second stage of the reaction during molecular sieve chromatography.130 Periodate method Commercially available peroxidase has only one or two lysine residues available for reaction. It also has eight carbohydrate chains on its surface which can be readily oxidised by NaI04 to yield aldehyde groups. These undergo a Schiff‘s base reaction with E-amino groups of antibody protein. The product is stabilised by reduction with sodium tetrahydro-borate(II1). In the orginal procedure the oxidation was carried out after reaction at pH 8 with an amino-blocking reagent such as l-fluoro-2,4-dinitrobenzene to minimise self-conjugation of the peroxidase.131 However 35% of self-coupled peroxidase was found despite these precautions. A more efficient procedure was described by Wilson and Nakane,l32 in which the NaI04 oxidation was carried out at pH 4-5. The pH was raised to 9.5 for the conjugation reaction with IgG or Fab’. In this process only 5% of self-coupled peroxidase molecules were formed. It also avoids the use of reagents to block amino groups which may be essential if enzyme activity is not to be lost during the conjugation reaction. The periodate method has now been used with glucose oxidasel33 and alkaline phosphatase,134 and advantages were claimed compared with peroxidase for use in ELISA. Dimaleimide method Under mild conditions maleimide groups react fairly rapidly with thiol groups and very slowly with other functional groups of proteins.The most commonly used of these reagents is N,N’-o-phenylenedimaleimide which has been used for the preparation of several conjugates with P-galactosidase as this enzyme contains free thiol groups. This homobifunctional reagent was first used in a two-stage process to produce an enzyme-labelled insulin. 135 Initially thiol groups were intro-duced into insulin by reaction with S-acetylmercaptosuccinic anhydride followed by reaction with hydroxylamine. This activated insulin was reacted with the dimaleimide the excess of reagent being removed prior to reaction with 6-galactosi-dase. No enzymic activity was lost during the conjugation reaction.The dimaleimide method is potentially applicable to all proteins that contain thiol groups. If necessary these can be introduced as mentioned above or by reaction with 4-methyl-mercaptobutyrimidate45 or by limited reduction of a disul-phide bond using 2-mercaptoethylamine. 136 Another dimalei-mide N N’-o-oxydimethylenedimaleimide has been used to conjugate IgG with peroxidase glucose oxidase P-galactosi-dase and penicillinasel37 with good retention of enzyme activity. The conjugates also gave low non-specific binding. Maleimide ester method The m-maleimidobenzoyl N-hydroxysuccinimide ester (MBS) was first used to couple insulin to P-galactosidase in 1976.138 Initially insulin was reacted with MBS to form MBS-acylated insulin by reaction of amino groups with the active ester.The excess of MBS was removed and in the second stage thioether bonds were formed between the thiol groups of P-galactosidase and the double bond of the maleimide moiety. Self-polymerisation was prevented as antibodies lack thiol groups and a heterobifunctional reagent was used. The whole process of conjugate formation was performed at pH 7 without loss of enzyme activity or immune reactivity of the individual components. The N-hydroxysuccinimide ester of N-(4-carboxycyclo-hexylmethy1)maleimide is more stable than MBS at neutral pH. It was used to form maleimide peroxidase which was then reacted with thiol groups of Fab’.139 The activites of enzyme and antibody were well preserved. When used in enzyme immunoassay the non-specific binding of Fab‘ - peroxidase conjugates was low probably owing to minimum poly-merisation during the conjugation reaction.Disulphide interchange method The N-hydroxysuccinimide ester of 3-(2-pyridyl-dithio)propionic acid was used to introduce 2-pyridyl disul-phide groups into both peroxidase and IgG by reaction with amino groups.140 The 2-pyridyl disulphide groups on one protein were then reduced by dithiothreitol to form thiol groups and then the two proteins were conjugated by a disulphide interchange reaction. This process avoids intra-molecular cross-linking. Similar reaction schemes have been used by other groups.141.142 Methyl 3-(4’-dithiopyridy1)propionimidate introduced a 4-dithiopyridyl group into one protein which could be reacted with a second protein that had been thiolated.Antibody activity was well retained and the specific activity of the peroxidase conjugate was 74% of that of the original enzyme. Thiol groups were introduced into a potentially useful enzyme A5J-ketosteroid isomerase by reaction with 5-acetylmercaptosuccinic anhydride. 143 After removal of th ANALYST. MAY 1984 VOL. 109 543 excess of reagents the modified enzyme was reacted with 5,5'-dithiobisnitrobenzoate and then allowed to react with thiol groups on the antigen with the release of thionitroben-zoate. The thiol groups on the antigen had been introduced by reaction with 4-methylmercaptobutyrimidate. Only 20% of enzyme activity was lost during the entire preparation. The conjugate prepared was used for a solid-phase immunoassay of human placental lactogen.Enzyme activity was measured directly when attached to the immunoadsorbent or more efficiently after the enzyme had been released from the solid phase by incubation with dithiothreitol. Purification of enzyme - protein conjugates All conjugation methods result in a final reaction mixture that should contain the desired product i.e. enzyme coupled to protein. In general the enzyme loses some activity and antibodies lose some immunoreactivity. The final activity is affected by the number of enzyme molecules bound to each antibody molecule and immunoreactivity is significantly reduced in highly polymerised conjugates. However in addition there are usually monomeric and polymeric forms of the enzyme and protein. These should be removed because otherwise they increase the background and decrease sensi-tivity.The species of lower relative molecular mass can be removed by the use of molecular sieve chromatography. Agarose or polyacrylamide gels are most usefu1,144 but pure enzyme - antibody conjugates are not obtained. Affinity chromatography is a more rapid procedure that gives purer reagents. Peroxidase conjugates because they contain car-bohydrate have been isolated using columns of concanavalin-A - Sepharose.145.146 Conjugates with A"3-ketosteroid isomerase were also purified by affinity chromatography. Boorsma and Streefkerk'zq used a Protein A - Sepharose column which interacts with IgG from several species, followed by use of a concanavalin-A - Sepharose column, which isolated the conjugates containing peroxidase.If affinity chromatography materials are not available the continuous removal of complexes from unbound molecules, by allowing the coupling procedure to take place during molecular sieve chromatography can be used to prepare enzyme - antibody conjugates free from large complexes as well as free enzyme.130 Hapten - Enzyme Conjugates Numerous cross-linking procedures have been used to form hapten - enzyme conjugates and most of them are similar to those used for the preparation of hapten - protein immu-nogens. If possible a different site and cross-linking method should be used for these two processess as this avoids interaction between the antibody and the bridge between hapten and enzyme; such an interaction can decrease assay sensitivity by 100- to 200-fold.148 It is also generally useful to incorporate a spacer group of 4-6 atoms between the hapten and enzyme which reduces steric effects on the interaction between hapten and antibody.149 The site of attachment between enzyme and hapten should be carefully selected as it influences the specificity of the assay.148 Purification of the enzyme - hapten conjugate affects the sensitivity of EIA. The presence of free enzyme may cause high background signals while unconjugated hapten tends to dilute out the enzyme-labelled form. Standard techniques such as dialysis or gel filtration readily remove free low relative molecular mass haptens. The removal of unconju-gated enzyme is more difficult but can be achieved by affinity chromatography.150,151 The important features of an enzyme - hapten conjugate are its enzymic activity and immunoreactivity. These are affected by the ratio of hapten to enzyme and in homogeneous EIA by the location of hapten relative to the active site. In heterogeneous EIA insufficient work has been carried out to clarify the importance of the hapten to enzyme ratio. Thus a sensitive assay for oestradiol-17P was developed using oestradiol-P-galactosidase with 100% immunoreactivity and a steroid to enzyme ratio of only 1:1.7,151 while Comoglio and Celedal52 found that 10 cortisol molecules per molecule of P-galactosidase were required to give 100% immunore-activity . Mixed anhydride method The mixed anhydrides of acids e.g. alkylchloroformates are formed at low temperatures and in inert organic solvents and then slowly added to cooled enzyme solutions.Under these conditions only protein lysyl and tyrosyl residues react. Hydroxylamine can be used to remove substituents from tyrosine side-chains. ' 5 3 This method has been used with many haptens including oestradiol oestrio1,tjj cortisol 152 mor-phine155 and methotrexate.156 Normally a 10- to 20-fold molar excess of mixed anhydride is used and the yields are generally 20-30%. Carbodiimide method The water-soluble carbodiimides l-ethyI-3-(3-dimethyl-aminopiopyl) carbodiimide and 1-cyclohexyl-3-( 2-morpholino-4-ethyl) carbodiimide methyl-p-toluenesulphonate have been used to couple haptens includ-ing testosterone,l57 cortisol158 and progesterone,lj9 to a variety of enzymes.The coupling reaction is generally carried out at pH 5.5-6 and amino groups on the protein are involved. One problem with the carbodiimide method is that signifi-cant intra- and intermolecular cross-linking of enzymes occur. A two-step conjugation has been used for conjugation of hippuric acid to IgG that should be generally applicable to hapten - enzyme conjugate formation where the hapten lacks amino groups.160 The acid was activated for 2 min with carbodiimide at room temperature and then added to protein in strong phosphate buffer at pH 8 to minimise the activation of the carboxyl groups of the protein. This reduced protein cross-linking from 7&8O% to less than 10%. Both the mixed-anhydride and carbodiimide methods produce peptide bonds.The carboxylic acid group can be introduced into the hapten if necessary. Oxygen and nitrogen substituents can be alkylated with halo esters followed by ester hydrolysis. 161 Reactions of hydroxy groups with succinic anhydride and ketone or aldehydes with carboxymethyl-oxime'62 have also been widely used. Periodate method Carbohydrate residues are cleaved with periodate to generate dialdehydes that can be coupled to amines. This approach was used to prepare a digoxin immunogen163 and an adenosine -P-galactosidase conjugate €or EIA. 164 In theory any amino-containing hapten can be conjugated with a glycoprotein after periodate cleavage. However the result of coupling T4 with peroxidase was not efficient. 165 m-Maleimidobenzoic acid N-hydroxysuccinimide ester (MBSE) and related methods The heterobifunctional reagent MBSE has been used in a two-stage reaction to couple haptens such as viomycin which contain an amino group to P-galactosidase which contains about 10 free thiol residues.166 The same enzyme has been conjugated with viomycin gentamicin and penicillin using N-(3-maleimidopropionylglycyloxy)succinimide.167 The use of these two reagents one to produce the immunogen and the other the enzyme label incorporates different bridging groups and improves sensitivity in the resulting EIA. In the prep-aration of the immunogen the protein used bovine serum albumin lacks thiol groups. These have been introduced b 544 ANALYST MAY 1984 VOL. 109 the reduction of disulphide bridges with dithiothreitoll67 or sodium tetrahydroborate(II1) .I68 Another cross-linker has also been used N-(y-maleimidobutyryloxyjsuccinimide this time for a very sensitive EIA of blasticidin S.Maleimidoben-zoate derivatives have been used to make thyroxine,169 cortisol170 and digoxigeninl71 derivatives of P-galactosidase and they have all been used in EIA procedures. An important advantage of the conjugates made using the methods referred to in this section is that the products retain very high enzyme activity and immunoreactivity . Methods using bifunctional imidates Dimethyl adipimidate reacts with amino groups under alkal-ine conditions. Using this reagent conjugates have been formed between P-galactosidase and desmethylnortripty-line172 and triiodothyronine.173 The reaction was carried out in two stages.Initially the hapten and dimethyladipimidate were reacted under anhydrous conditions before being allowed to react with enzyme in aqueous conditions. The desmethylnor-triptyline conjugate retained about 80% enzyme activity and immunoreactivity . Method of controlled coupling of amino compounds to enzymes The direct conjugation of amino groups on haptens with enzymes can lead to extensive loss of enzyme activity and low yields of enzyme conjugate These difficulties were overcome by Singh et al. ,174 who added a thiol group to an amino group of gentamicin. The enzyme employed in their system glucose 6-phosphate dehydrogenase was modified with bromo-acetylglycine. The two compounds were then used to form the enzyme - hapten conjugate which was successfully used in an homogeneous enzyme immunoassay although 70% of the enzyme activity was lost in the initial modification reaction.Coating of Solid Surfaces with Protein and Hapten Heterogeneous EIA requires a separation stage. In ELISA techniques this is facilitated by the use of surfaces that are coated with antigen antibody or hapten and are disposable after use. The surfaces most commonly used are microtitre plates (generally made from polymethacrylate or poly-styrene) large (3.2 or 6.4 mm diameter) specular beads (made of nylon or polystyrene) one per determination or polystyrene tubes. Generally the required protein or hapten conjugated to protein is allowed to adsorb on to the plastic surface which is usually negatively charged.The rate and extent of protein binding are dependent on the type of protein and its concentration and the time pH and temperature of incu-bation. 175 Cantarero et al. 175- recommended adding about 500 ng of protein per 6.5 cm2 of polystyrene which appears to give a monolayer of protein. Up to this amount of protein adsorption to polystyrene occurs probably by hydrophobic interactions. However if more protein is used protein -protein interactions occur which are wasteful and do not increase the working range of assays. Although protein adsorption is simple and convenient it has been claimed that adsorbed antibodies may undergo denaturation or loss of adsorbed protein either on washing176 or when other proteins plus non-ionic detergents are added to prevent non-specific adsorption of immunoreagents.177 These effects can limit both the sensitivity and the reproducibility of the ELISA technique. The use of proteins covalently attached to the surface should overcome these difficulties. Nylon balls have been activated by partial hydrolysis with 3.5 M HC1 for 24 h followed by reaction with glutaraldehyde or a carbodi-imide.177 This treatment increased the capacity of nylon about 20-fold in comparison with simple adsorption of IgG. The treated nylon balls had about a 10-fold increased sensitivity and gave a lower background when used in EIA than poly-styrene balls to which IgG had been adsorbed. Hendry and Herrmannl77 also showed that the covalently immobilised antibodies allowed re-use of the nylon balls.Similar methods for covalent linking of proteins to polystyrene surfaces are not available although some success has been claimed after pre-treatment with glutaraldehyde. 178,179 Another approach is to adsorb fetuin a protein containing some carbohydrate to polystyrene. This can then be activated with periodate. It was found that when IgG was covalently attached to activated beads the resulting EIA had improved sensitivity and de-creased background. 18” Concluding Remarks In general the use of enzyme labels and enzyme-associated labels in immunoassay systems has provided assays that are specific potentially sensitive and relatively inexpensive to set up and run. The reagents involved are stable and for that reason are well suited to assay systems that are only used infrequently.In most instances the enzyme systems are much safer than the techniques in which radioisotopes are used. The presently available homogeneous systems have limited sensitivity but this disadvantage is balanced and frequently outweighed by their convenience and ease of automation. The rapidity of these homogeneous assays has made them particularly suitable for the analysis of drugs in emergencies. The heterogeneous methods can be very sensitive and are relatively unaffected by interfering factors. For qualitative analyses the heterogeneous system can be unrivalled in its simplicity. Despite the plethora of publications on the subject it must be remembered that EIA is still a relatively new technique in comparison with RIA and much of the methodology involved has not yet been fully evaluated.One particular area where research would be valuable is in the direct comparison of different EIA techniques. Some workers have directly com-pared different substrates (e.g. Groomel81; Porstmann er al. 182; Labrousse et al.183; Yolken and Leisterl*4) or different enzymes (e.g. Johnson etal.1”) but the greater availability of such information would be a significant aid to enzyme immunoassay development. Areas of methodology suitable for further improvement include the examination of new enzyme labels the production of more sensitive methods of enzyme assay the development of more specific cross-linking reagents and improved tech-niques for coating solid surfaces with proteins and haptens.This review has briefly described the various ways in which enzymes can be harnessed in the development of immunoas-say methods. The enzymes are employed in a variety of ways, most but not all of which utilise the unique amplifying ability of enzymes to produce suitably sensitive assays. An evaluation of the usefulness of the various assay protocols cannot easily be made in the context of a review such as this as this is influenced to a great extent by the analytical situation in which the assay is employed. The EMIT system is for example most suited when rapid determinations are required whereas ELISA methods are to be preferred in large-scale screening procedures in which large numbers of samples are involved. The enzyme-based systems that have been developed only in recent years may combine sensitivity with the advantages of homogeneous techniques.The continued development of a wider range of assay protocols will result in an even greater matching of assay capabilities to analytical requirements. This further exploitation of the potential of enzymes as immuno-assay labels will hopefully provide yet further improvements in laboratory efficiency ANALYST MAY 1984 VOL. 109 545 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30, 31. 32. 33. 34. 35. 36. 37. 38. 39. 30. 41. References Yalow R . S . and Berson S . A Nature (London) 1959,184, 1648. Yalow K. S . and Berson S.A . J . Clin. Invest. 1960 39, 1157. Landon J. Antibiot. Chemother. 1979 26 118. Weinryb I. Schroff. J. R . and Olson D . R . Drug Metab. Rev. 1979 10 271. Ekins R . “Radioimmunoassay and Related Procedures in Medicine 1977,” Volume I International Atomic Energy Agency Vienna 1978. p. 241. Mattiasson B. and Ling T. G. I . . J . Immunol. Methods 1980, 38 217. Cais M. and Shimoni. M. Ann. Clin. Biochem. 1981 18, 317. Cais M. and Shimoni M. Ann. Clin. Biochem. 1981 18, 324. Thorell J. 1 Clin. Chem. 1981 27 1969. Smith K. O. and Gehle W. D. Methods Enzymol. 1980,70, 288. Schall R . F. and Tenoso H. J . Clin. Chem. 1981.27 1157. Haimovich J. and Sela M. Science 1969. 164 1279. Andrieu J.-M. Mamas S . and Dray F. in Cameron, E. H. D. Hillier S.G . and Griffiths K. Editors “Steroid Immunoassay,” Alpha Omega Publishing. Cardiff 1975 p. 189. Pratt J. J . Woldring M. G. and Villerius L. J. Immunol. Methods 1978 21 179. Velan B. and Halrnann M. Imrnunochemistry 1978,15,331. Pazzagli M. Kim J. B. Messeri G . Kohen F. Bolelli, G. F. Tommasi A . Salerno R. Moneti G . and Serio M. J . Steroid Biochem. 1981 14. 1005. Pazzagli M. Kim J. B. Messeri G. Kohen F. Bolelli, G. F. Tommasi A Salerno R. and Serio M. J. Steroid Biochem. 1981 14 1181. Schroeder H. R. darrico R. J . Boguslaski. R. C and Christner J. E. Anal. Biochem. 1976 72 283. Wannlund J Azari J . Levine L. and DeLuca M., Biochem. Biophys. Res. Commun. 1980 96 440. Soini E. and Hemmila I. Clin. Chem. 1979 25 353. Visor G. C and Schulman. S.G . J . Pharm. Sci. 1981 70. 469. Cais M. Dani S . Eden Y. Gandolfi O. Horn M. Isaacs, E . E. Josephy. Y. Saar Y . Slovin E. and Snarsky L., Nature (London) 1977 270 534. Leute R. K. Ullman E . F. and Goldstein. A . J . Am. Med. Assoc. 1972 221 1231. Scharpk. S. L. Cooreman W. M Blomme W. J . and Laekeman G. M. Clin. Chem. 1976 22 733. Wisdom G . B Clin. Chem. 1976 22 1243. Schuurs A. H. W. M. and Van Weeman B. K. Clin. Chim. Acta 1977 81 1. Schuurs A. H . W. M. and Van Weeman B. K. J . Immunoassay 1980 1 229. Belanger. L. Scand. J . Immunol. 1978 8 (Suppl. 7) 33. Voller A . Bartlett A . and Bidwell D. E . J. Clin. Parhol., 1978 31 507. O’Beirne A. J. and Cooper H . R. J . Histochem. Cytochem., 1979 27 1148. O’Sullivan M. J . Bridges. J . W. and Marks V.Ann. Clin. Biochem. 1979. 16 221. Van Weeman B. K Bosch A. M. G Dawson E. C. and Schuurs A. H. W. M. J . Steroid Biochem. 1979 11 147. Oellerich M. J . Clin. Chem. Clin. Biochem. 1980. 18 197. Maggio E . T. Editor “Enzyme Immunoassay,” CRC Press, Cleveland OH 1980. Carlier Y. Bout D and Capron. A Bull. Inst. Pasteur, Paris 1981 79 313. Yolken R. H Rev. Infect. Diseases 1982 44 35. Ngo T. T. and Lenhoff H. M. Mol. Cell. Biochem. 1982,44, 3. Van Weeman B. K. and Schuurs A H. W. M. FEBS Lett,, 1971. 15 232. Engvall E. and Perlmann P. Irnmunochernistry 1971.8,871. Engvall E. Jonsson K and Perlmann P. Biochem. Biophy. Acta 1971 251 327. Belanger. L. Hamel D. Dufour. D . and Pouliot M. Clin. Chem. 1976 22 198. 42. 43. 44. 45. 46. 47.48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70, 71. 72. 73. 74. 75. 76. 77. 78. Maiolini B. Ferrua B. and Masseyeff R. J. Immunol. Methods 1975 6 355. Clark M. F. and Adams A. N. J . Gen. Urol. 1977,34,475. Engvall E . and Perlmann P. J. Immunol. 1972 109 129. Gnernmi E. O’Sullivan M. J. Chieregatti G. Simmons M., Simmonds A . Bridges J. W. and Marks V. in Pal S . B Editor “Enzyme-linked Immunoassay of Hormones and Drugs,” Walter de Gruyter Berlin and New York 1978 p. 29. Guesdon J.-L. Ternynck T. and Avremeas S . J . Histo-chem. Cytochem. 1979 27 1131. Tate D. Y. Carlton G. T. Nesbit M. E. White J. G., Krivit W. and Sorenson R. L.,Am. J . Hematol. 1980,9,349.Miyai K. Isbibashi K. and Kumahara Y. Clin. Chim. Acta, 1976 67 263. Yamamoto R. Umeda Y. Kosaka A . and Kato K J . Biochem. 1981 89 223. Guesdon J.-L. and Avremeas S . Methods Enzymol. 1981, 73 471. Voller A. Bidwell D. Huldt G . and Engvall E. Bull. W.H.O. 1974,51 209. Rook G. A . W. and Cameron C. H. J. Immunol. Methods, 1981 40 109. Chooi M. K. ICRS Med. Sci. Biochemistry 1980 8 394. Tsang V. C. W. Wilson B. C. and Maddison S. E. Clin. Chem. 1980 26 1255. Goodburn R . Williams D. L. and Marks V. J . Automat. Chem. 1982,4 129. Felgner P. Zentralbl. Bakteriol. Parasitenkd. Znfektionskr. Hyg. Abt. 1 Orig. Reihe A 1978 240 112. Chandler H. M. and Hurrell J. G. R. Clin. Chim. Acta. 1982 121. 225. Ngo T. T. and Lenhoff H. M. Biochem. Biophys.Res. Commun. 1981 99 496. Monji N. and Castro A . Res. Commun. Chem. Pathol. Pharmacol. 1979 26 187. Borrebaeck C . Borjeson J. and Mattiasson B. Clin. Chim. Acta 1978 86 267. Joyce B. G . Turkes A. Ozoran A. Read G. F. and Fahmy R . B. in Pal S . P. Editor “Enzyme-linked Immuno-assay of Hormones and Drugs,” Walter de Gruyter Berlin and New York 1978 p. 247. Watson D . Lancet 1976 2 570. Ekins R. Lancet 1976 2 569. Ishikawa E. Imagawa M. Yoshitake S . Niitsu Y. Urushi-zaki I . Inada M. Imura H. Kanazawa. R . Tachibana S . , Nakazawa N. and Ogawa H. Ann. Clin. Biochem. 1982,19, 379. Shalev A . Greenberg A. H. and McAlpine P. J. J. Immunol. Methods 1980,38 125. Kato K. Umeda Y. Suzuki F. and Kosaka A. J. Appl. Biol. 1979 1 479. Kato K. Umeda Y.Suzuki F. and Kosaka A. Clin. Chim. Acta 1980 102 261. Rubenstein IS. E . Schneider R. S . and Ullman E. F Biochem. Biophys. Res. Comrnun. 1972 47 846. Crowl C. P. Gibbons I. and Schneider R. S . in Nakamura, R. M. Dito W. R . and Tucker E. S . Editors “Immuno-assay Clinical Laboratory Techniques for the 1980s,” A. K. Liss New York 1980 p. 89. Schneider R. S . Lindquist P. Wong E. T Rubenstein, K. E. and Ullman E. F. Clin. Chern. 1973. 19 821. Rowley G. L. Rubenstein K. E . Huisjen J and Ullman, E. F J . Biol. Chem. 1975 250 3759. Pippenger C. E. Bastiani R. J. and Schneider. R. S. in Schneider H. Janz D. Gardener-Thorpe. C. Meinard H., Sherwin A. L. Editors “Clinical Pharmacology of Antiepilep-tic Drugs,” Springer-Verlag New York 1975. p. 331. Ullman E.F. Yoshida R . A Blakemore J. I Maggio E., and Leute. R. Biochim. Biophys. Acta 1979 567 66. Jaklitsch A. P, Schneider R. S . Johannes R. J . Levine, J. E . and Rosenberg G. L. Clin. Chem. 1976. 22 1185. Gibbons I. Skold. C. Rowley G . L and Ullman E. F., Clin. Chern 1979 25 1078. Kitagawa. T. and Aikawa T. J . Biochem. 1976 79 233. Gibbons I. Hanlon T. M. Skold C. N. Russell M. E and Ullman E. F . Clin. Chem. 1981 27 1602. Yoshida R. A . Johns-Stephens. L and Greenwood H. M., Clin. Chern. 1980 26 982 546 ANALYST MAY 1984. VOL. 109 79. Litman D. J. Hanlon T. M. and Ullman E. F. Anal. Biochem. 1980 106 223. 80. Thompson R. J. Trends Biochem. Sci. 1982 7 419. 81. Lasky F. D. Ahuja K. K. and Karmen. A. Clin. Chem., 1977 23 1444. 82. Nolan J.P Di Benedetto. G. and Tarsa N. J. Clin. Chem., 1981 27 738. 83. Galen R. S . and Forman D. Clin. Chem. 1977 23 119. 84. Ying C. Miyada D. Gwinup D. and Simpkins H. Clin. Chem. 1981 27 344. 85. Alexander N. M . Gattra R. and Nishimoto. M Clin. Chim. Acta 1980 100 301. 86. Kim H. J. and Cerceo E. Clin. Chem. 1976 22 1935. 87. Andersen O. and Eriksen P. B. Clin. Chem. 1977,23,751. 88. Carrico R. J. Christner J . E. Boguslaski R. C. andYeung, K. K. Anal. Biochem. 1976 72 271. 89. Schroeder H. R. Carrico R. J. Boguslaski R. C. and Christner J. E. Anal. Biochem. 1976 72 283. 90. Morris D. L. Ellis P. B Carrico R. J. Yeager F. M Schroeder H. R. Albarella J. P. Boguslaski R. C. Hornby, W. E. and Rawson D. Anal. Chem. 1981 53 658. Lin J. D. and Pardue H. L.Clin. Chem. 1982 28 2081. Tyhach R. J. Rupchock P. A. Pendergrass J. H. Skjold, A. C Smith P. J. Johnson R. D Albarella J. P and Profitt J. A. Clin. Chem. 1981 27 1499. Burd J. F. Carrico R. J. Fetter M. C. Buckler R. T., Johnson R. D. Boguslaski R. C. and Christner J. E. Anal. Biochem. 1977 77 56. Burd J. F. Wong R. C. Feeney J. E . Carrico R. J. and Boguslaski R. C. Clin. Chem. 1977 23. 1402. Wong R. C. Burd J . F. Carrico R. J. Buckler R. T . , Thoma J. and Boguslaski R. C. Clin. Chem. 1979,25,686. Ngo T. T. Carrico R. J. Boguslaski R. C. and Burd J. F., J . Immunol. Methods 1981 42 93. Greenquist A. C. Walter B. and Li T. M. Clin. Chem., 1981 27 1614. Ngo. T. T. and Lenhoff H. M. FEBS Lett. 1980 116 285. Finley P. R Williams R. J. and Lichti D. A.Clin. Chem., 1980 26 1723. 100. Blake C. Al-Bassam M. N Gould B. J. Marks V Bridges J . W. and Riley C. Clin. Chem. 1982 28 1469. 101. Mitsuma T. Colucci J . Shenkman L. and Hollander C. S . , Biochem. Biophys. Res. Commun. 1972 46 2107. 102. Haynes. S. P. and Goldie D. J. Ann. Clin. Biochem. 1977, 14 12. 103. Dean K. J . . Thompson S. G. Burd J. F and Buckler R. T., Clin. Chem. 1983 29 1051. 104. Mamas S . and Dray F . Ann. Zmmunol. 1979 103C 407. 105. Fricke H. Strasburger C. J. and Wood. W. G. J . Clin. Chem. Clin. Biochem. 1982. 20 91. 106. Van Der Waart M. and Schuurs. A. H. W. M. 2. Anal. Chem. 1976 279 142. 107. Harris C. C. Yolken R. H. Krokan H. and Hsu I. C., Proc. Natl. Acad. Sci. USA 1979 76 5336. 108. Hsu I. C. Poirier M. C. Yuspa S.H. Yolken R. H. and Harris C. C. Carcinogenesis 1980 1 455. 109. Ekins. R Nature (London) 1980 284 14. 110. Barman T. E. “Enzyme Handbook,” Volumes 1 and 2, Springer-Verlag Berlin Heidelberg New York 1969. 111. Voogd C. E . Van der Stel J. J . and Jacobs J. J. J. A. A. J . Immunol. Methods. 1980 36 55. 112. Al-Kaissi E. and Mostratos A. J . Immunol. Methods. 1983, 58 127. 113. Ellens D. J. and Gielkens A. L. J. J . Immunol. Methods, 1980 37 325. 114. Berkowitz D. B. and Webert D. W. J . Immunol. Methods, 1981 47 121. 115. Ngo T. T. and Lenhoff H. M. Anal. Biochem. 1980 105, 389. 116. Carr P. W. and Bowers L. D. “Immobilized Enzymes in Analytical and Clinical Chemistry,” Wiley New York Chi-Chester Brisbane Toronto 1980. Ishikawa E. and Kato K. Scand.J. Immunol. 1978 8 43. Mascini. M. Zolesi F. and Palleschi G. Anal. Lett. 1982, 15 101. Gebauer C. R . and Rechnitz G. A. Anal. Lett. 1981,14,97. Mattiasson B . Borrebaeck C. Sanfridson B. and Mosback, K. Biochem. Biophys. Res. Commun. 1977,483 221. 91. 92. 93. 94. 95. 96. 97. 98. 99. 117. 118. 119. 120. 121. 122. 123. 124. 125 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. Samake H. Rajkowski K. M. and Cittanova. N Clzn. Chim. Acta 1983 130 129. Meyerhoff M. E. and Rechnitz G. A. Anal. Biochem 1979. 95 483. Hardy P. M. Nicholls A. C. and Rydon H. N .J . Chem. SOC. Perkin Trans. I 1976 9 958. Richards F. M. and Knowles. J . R. J. Mol. Biol. 1968 37. 231. Hardy P. M. Hughes G. J . and Rydon. H. N. J . Chem. SOC. Chem. Commun. 1976 5 157. Avrameas S. Immunochemistry 1969 6 43. Boorsma D. M. and Kalsbeek G. L. J. Histochem. Cytochem. 1975 23 200. Engvall E. Scand. J . Immunol. 1978 8 (Suppl. 7) 25. Boorsma D. M. and Streefkerk J. G. in Knapp W., Holuber K. and Wick. G. Editors “Immunofluorescence and Related Staining Techniques,” ElsevieriNorth-Holland Biomedical Press Amsterdam and New York. 1978 p. 225. LannCr M. and Bergquist R. in Knapp W. Holubcr. K., and Wick G. Editors ‘‘Immunofluorescence and Related Staining Techniques,” Elsevier/North-Holland Biomedical Press Amsterdam New York 1978 p. 237.Nakane P. K. and Kawaoi A. J. Histochem. Cytochem., 1974 22 1084. Wilson M. B. and Nakane P. K. in Knapp W. Holuber K., and Wick G. Editors “Immunofluorescence and Related Staining Techniques,” Elsevier/North-Holland Biomedical Press Amsterdam and New York 1978 p. 215. Johnson R. B. and Nakamura R. M. in Nakamura R. M., Dito W. R. and Tucker E. S . Editors “Immunoassays: Clinical Laboratory Techniques for the 1980s,” A. R. Liss. New York 1980 p. 144. Hazlett D. T. G. and Garner P. Afr. J. Clin. Exp. Immunol. 1981 2 325. Kato K. Hamaguchi Y. Fukui H. and Ishikawa E J. Biochem. 1975 78 235. Kato K. Hamaguchi Y. Fukui H. and Ishikawa E. FEBS Lett. 1975 56 370. Weston P. D. Devries J. A. and Wrigglesworth R., Biochim. Biophys. Acta 1980 612 40. Kitagawa T.Aikawa T. J . Biochem. 1976. 79 233. Yoshitake S . Imagawa M. and Ishikawa E. Anal. Lett., 1982 15 147. Carlsson J. Drevin H. and AxCn R. Biochem. J . 1978,173, 723. King T. P. Li Y and Kochoumian L. Biochemistry 1978, 17 1499. King T. P. and Kochoumian L. J. Immunol. Methods. 1979. 28 201. Terouanne B. Nicolas J. C. Descomps B. and Crastes de Paulet A. J . Immunol. Methods 1980 35 267. Boorsma D. M . and Streefkerk J. G. J . Histochem. Cytochem. 1976 24 481. Arends J. J . Immunol. Methods 1979 25 171. Poxstmann T. Porstmann B. Schmecta H. Nugel. E., Seifert R. and Grunow R. Acta Biol. Med. Ger. 1981 40, 849. Terouanne B. Nicolas J. C. Descomps B. and Crastes de Paulet A. J . Immunol. Methods 1980 35 277. Van Weeman B. K. and Schuurs A.H. W. M. Immuno-chemistry 1975 12 667. Mould G. P. Aherne G. W. Morris B. A. Teale J. D and Marks V. Eur. J . Drug Metab. Pharm 1977 2 171. Exley D. and Abuknesha R FEBS Lett 1977. 79. 301. Exley D. and Abuknesha R. FEBS Lett. 1978. 91 162. Comoglio S . and Celeda F. J. Immunol. Methods 1976 10, 161. Riordan J. F. and Vallee B. L. Biochemistry 1963,2 1460. Van Weeman B. K. and Schuurs A. H. W. M. FEBS Lett., 1972 24 77. Schneider R. S . Lindquist P . Wong E. T. Rubenstein, K. E. and Ullman E. F. Clin. Chem. 1973 19 821. Marks V. O’Sullivan M. J. Al-Bassam M. N. and Bridges, J. W. in Pal S. B. Editor “Enzyme-linked Immunoassay of Hormones and Drugs,” Walter de Gruyter Berlin and New York 1978 p. 419. Tateishi K. Yamamoto H. Ogihara T. and Hayashi C., Steroids 1977 30 25 ANALYST MAY 1984 VOL.109 547 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. Ogihara T. Miyai K. Nishi K. Ishibash K. and Kumahara, Y. J. Clin. Endocrinol. Metab. 1977 44 91. Dray F. Andrieu J.-M. and Renaud F. Biochim. Biophys. Acta 1975 403 131. Davis M.-T. B. and Preston J. F. Anal. Biochem. 1981,116, 402. Spector S . and Parker C. W . Science 1970 168 1347. Erlanger B. F . Borek G . Beiser S. M. and Lieberman S . , J. Biol. Chem. 1959 234 1090. Butler V. P. and Chen J . P. Proc. Natl. Acad. Sci. USA, 1967 57 71. Lauer R. C. and Erlanger B. F. Zmmunochemistry 1974,44, 533. Schall R . F. Fraser A. S . Hansen H. W. Kern C. W. and Tenoso H. J. Clin. Chem. 1978 24 1801.Kitagawa T. Fujitake T. Taniyama H. and Aikawa T. J. Antibiot. 1976 29 1343. Kitagawa T. Kanamaru T. Kato H. Yano S . and Asanuma Y. in Pal. S. B . . Editor “Enzyme-linked Immuno-assay of Hormones and Drugs,” Walter de Gruyter Berlin and New York 1978 p. 59. Kitagawa T. Kawasaki T . and Munechika H . J . Biochem., 1982 92 585. Monji N. Malkus H. and Castro A. Biochem. Biophys. Res. Commun. 1978 85. 671. Monji N. and Castro A. A. Experientia 1980 36 347. Monji N. Ali H. and Castro A. Experientia 1980,36 1141. Al-Bassam M. N. O’Sullivan. M. J. Gnemmi E . Bridges, J. W. and Marks V. Clin. Chem. 1978 24 1590. O’Sullivan M. J. Gnemmi E. Morris D. Al-Bassam M. N., Simmons M. N. Bridges J. W. and Marks V. in Pal S. B., 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. Editor “Enzyme-linked Immunoassay of Hormones and Drugs,” Walter de Gruyter Berlin and New York. 1978, p. 301. Singh P. Leung D. K . Rowley G. L. Cragne C. and Ullman E. F. Anal. Biochem. 1980 104 51. Cantarero L. A. Butler J. E. and Osborne J . W. Anal. Biochem. 1980 105 375. Pesce. A. J. Ford D. J and Gaizutis M. A . Scand. J. Immunol. 1978 8 (Suppl. 7) 1. Hendry R . M. and Herrmann J. E. J. Immunol. Methods, 1980 35 285. Place J. D . and Shroeder H. R . J. Immunol. Methods 1982, 48. 251. Klasen E . A. Rigutti A. Bos A. and Bernini L. F J. Immunol. Methods 1982 54 241. Ziola B. and Tuokko H. Acta Pathol. Microbiol. Scand., Sect. C 1980 88 127. Groome N. P. J. Clin. Chem. Clin. Biochem. 1980. 18,345. Porstmann B. Porstmann T. and Nugel E . J. Clin. Chem. Clin. Biochem. 1981 19 435. Labrousse H. Guesdon J.-L. Ragimbeau J. and Avremeas S . J. Immunol. Methods 1982 48 133. Yolken R . H. and Leister F. J. J. Clin. Microbiol. 1982 15, 757. Johnson R . B. Libby R. M. and Nakamura R . M. J. Immunoassay 1980 1 27. Paper A31349 Received October loth I983 Accepted December 23rd I98

ISSN:0003-2654    

DOI:10.1039/AN9840900533

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 549 The Berthelot or lndophenol Reaction and Its Use in the Analytical Chemistry of Nitrogen A Review Philip L. Searle New Zealand Soil Bureau Private Bag Lower Hutt New Zealand Summary of Contents Introduction Historical Nature of the Reaction Mechanism Reagents Phenols Hypohalite sources Cata I ysts of monochloramine formation of phenol concentration of indophenol dissociation of absorption maxima of rate of colour development of colour stability of gas volatilisation and absorption Reagent concentrations Phenols Hypo ha1 ite sources Nitroprusside Development time and temperature Order of addition of reagents Colour stability Absorption maximum Sensitivity Interferences Nitrogen compounds Metals Non-metals Salt effect Light Reagents Reaction conditions Interferences Analysis rates Radiochemical methods Applications Agriculture Clinical Food Waters Meta I lu rgy Pet ro I eu m Gases Pharmaceutical products Conclusions Phenols Hypoc h lo rite source Catalysts PH Other variables Interferences Summary Reaction conditions pH dependence Solvent extraction met hods Automated methods Keywords Review; Berthelot or indophenol reaction; nitrogen determination; ammonium determinatio 550 ANALYST MAY 1984 VOL.109 Introduction The Berthelot reaction (sometimes called the indophenol reaction) is the name given to the reaction of ammonium ions and a phenol which under suitable oxidising conditions, results in the formation of an indophenol dye.Indophenol dyes are highly conjugated and absorb strongly between 630 and 720 nm. Methods based on this reaction are sensitive and relatively specific for the ammonium ion. There are numerous methods in which ammonium-nitrogen is determined by the reaction and this also applies to total methods in which nitrogen is converted into the ammonium form by suitable pre-treatment such as Kjeldahl digestion. The reaction mechanism is complex and not completely understood and this led to problems in early manual methods. The proliferation of interest in the reaction in recent years has stemmed almost entirely from the development of automated analysis parti-cularly continuous flow analysis (CFA) in which the reaction variables are controllable and reproducible.The widespread use of the reaction coupled with recent research on the reaction mechanism make it an appropriate time to review the history nature and applications of the reaction. Historical Berthelot’ noted that a blue coloration was obtained when ammonia phenol and hypochlorite were mixed and Lex2 also found that phenol and ammonia reacted under oxidising conditions to form a blue colour. Cotton3 independently observed the reaction and also noted that hypobromite could be substituted for hypochlorite. Engel4 reported that some aliphatic and aromatic amines reacted with hypochlorite and phenol to give green and blue colorations. Tarugi and L e n ~ i ~ in studying the action of hypochlorites on compounds contain-ing amino groups found that blue colorations were obtained for a range of nitrogen compounds when p-benzoquinone was present.These workers also described a method for the determination of ammonia in water that was more sensitive than Nessler’s method.6 Thomas7 suggested that the reaction was suitable for the determination of ammonia but disagreed with the findings of Tarugi and Len$ regarding the specificity of the reaction. Thomas found that amongst the amino acids only aminoacetic acid gave an intense colour under the reaction conditions. Colorations were also obtained with some amines but Thomas7 qualified his statement by saying that these positive reactions with amines could be due to ammonia impurities. In a following paper Thomas8 utilised the reaction to determine ammonia in cerebrospinal fluid and showed that the reaction was not affected by the presence of proteins or sugars.Foxwell9 used the reaction for the determination of ammonia in coal waste liquors and found it highly sensitive, although the reaction failed when free acid was present in the sample. Foxwell also found that heating was essential to the success of the method which when applied to specimen samples gave comparable results to the distillation and titration of ammonia. Ayers et a l l 0 and Hucker and Wall” used the reaction as described by Thomas8 to measure ammonia produced by bacteria. Orr12 utilised Thomas’s8 method to determine ammonia in urine and found that the reaction could be performed without heating thereby over-coming interference from the breakdown of urea caused by heat under alkaline conditions.Orr12 also found that under his reaction conditions only a-alanine and glycine produced the reaction and only to a minor extent. Contrary to these findings Butcher13 claimed that a number of organic substances interfered in Thomas’s8 method for determining ammonia as several aliphatic amines produced blue colorations identical with that given by ammonia. Murray14 utilised Orr’sl2 spectrophotometric method to determine urea in blood and urine by measuring ammonia produced after enzymatic conversion with urease and found that the results compared closely to those given by an aeration method. Hansenls used the reaction to detect ammonia in agar cultures but found thymol to be more sensitive than phenol and used hypobromite instead of hypochlorite, because it was easier to prepare a solution of known strength.Hansen found his method to be less susceptible to inter-ferences from methylamine and a number of other amino acids although greater interference from glycine was encoun-tered. Furth et al. 16 investigated the reaction of various amino acids with phenol and hypochlorite and showed that some amines reacted in a similar manner to ammonia. This was attributed to the splitting off of the amine group under the reaction conditions. Furth et al. isolated the blue dye produced by the reaction and concluded that it belonged to the indigo-phenol type. Van Slyke and Hiller17 used calcium hypochlorite and phenol for the determination of ammonia in blood and found the method to be more sensitive than the Nessler reaction.Harrow et af.,18 in the investigation of 20 different phenols, reached the conclusion that phenol itself was the best for the reaction the colour being more pronounced and stable than that produced by thymol. Lapin and Hein” noted the reaction of phenols sodium hypobromite and ammonia and found that the coloured material produced with thymol could be extracted with organic solvents. These workers also noted that salicylic acid and its salts reacted but many polyhydric phenols and a- and P-naphthol did not. BorsookZO modified the method of Van Slyke and Hiller17 to achieve greater sensitivity and colour stability and used the reaction to determine ammonia in blood urine and tissue fluids. Simi-larly Polonovski and Boulanger21 used the reaction for measuring ammonia in biological fluids and Wittermans22 measured urea by determining ammonia after urease treat-ment.Furth and Gotz123 studied the colour rection of ammonia and glycine with hypobromite and phenols. They extracted and purified the dye formed when thymol was used and concluded that the nitrogen atom was essential for the formation of the dye which they concluded consisted of a mixture of different high relative molecular mass substances with various phenolic compounds grouped around an oxidised halogenated quinone phenolimide. Crismer24 used chloramine-T (N-chloro-p-toluene-sulphonamide sodium salt) and phenol for the determination of ammonia in urine and body fluids but felt that the method was not sensitive enough for the determination of ammonia in blood.Hinsberg and Mucke25 employed the reaction using phenol and hypochlorite for the determination of ammonia in Kjeldahl digests and showed the importance of hypochlorite concentration and pH in determining the optimum conditions for maximum colour development. Veinberg26327 used thymol and sodium hypobromite followed by solvent extraction of the indothymol to determine nitrogen in steel after an acid digestion. Hansen and Nielson28 modified the spectropho-tometric technique of Crismer24 and found sodium hypobro-mite more convenient to use than chloramine-T and thymol better than phenol for producing a stable cobur. Alten and Haupt29 used the reaction for measuring ammonia in the presence of organic matter because the high sensitivity reduced interferences from colloidal matter.Russell30 studied the reaction between ammonia phenol and calcium hypoch-lorite in some detail and showed that the phenol concentration and pH were important factors in improving sensitivity. Russell also found that iron chromium and manganese(I1) ions catalysed the reaction and used a standard addition of manganese(I1) ions to increase the sensitivity. Copper(1) ions were found to inhibit the reaction. Vazhenin31 used a modification of Foxwell’s9 method for determining adsorbed ammonium in soils. Casas and Domin-g ~ e z ~ ~ reported the reaction between ammonia sodium hypochlorite and phenol apparently unaware of the previou ANALYST. MAY 1984 VOL. 109 55 1 work and Kakita33 used the phenol - hypochlorite reaction to measure nitrogen in steel after acid digestion and removal of the metals from solution by alkaline precipitation.Riley34 used a modified version of Russell’~3~) method to measure ammonia in sea water and found that heating the reaction mixture to 70 “C produced a slightly less intense but more stable colour. Riley34 also found that the method could not be used directly on sea water owing to interference by magnesium and calcium ions. Lapin35 examined the basis of the reaction between thymol and hypochlorite and postulated a three-step reaction mechanism. Lubochinsky and Zalta36 discovered that the presence of sodium nitroprusside greatly increased the sensitivity of the reaction more than the use of the man-ganese(I1) ion.These workers also realised the importance of pH and employed a phosphate buffer at pH 12 to stabilise the reaction conditions. Noble37 utilised the phenol - hypochlorite reaction to determine low levels of nitrogen in petroleum after Kjeldahl digestion and developed a method that used less complex reagents and a shorter development time than those described by Riley.34 Scheurer and Smith38 used the manganese(I1)-catalysed reaction to determine nitrogen in simple organic compounds but used hypochlorous acid instead of hypo-chlorite as this was found to produce a more stable colour and no heating was necessary. Exley39 determined small amounts of nitrogen in biological material after Kjeldahl digestion and used Russell’s3() method with a minor modification to the amount of manganese(I1) ion catalyst used.Crowther and Large40 found that the sensitivity of Riley’s34 method could be increased in the presence of organic solvents so that heating and the use of a manganese(I1) ion catalyst were unnecessary if acetone was present. Takei and Kato41 noted that Riley’s34 method could be used without heating but found that cyanide and thiocyanate interfered with the colour reaction. These workers also altered the reagent concentrations to measure nitrogen in steels after digestion and distillation. Brown used sodium nitroprusside to increase the sensitivity of the reaction as suggested by Lubochinsky and Zalta,36 and used this method to determine ammonia in rat tissues and human plasma. Riley and Sinhaseni43 measured ammonia in sea water using the acetone modification of Crowther and Large40 to increase the rate of colour development.Lapin et al.44 measured ammonia in soil using the thymol - hypobromite reaction followed by solvent extraction of the indothymol prior to colorimetric measurement. Fleury and Eberhard45 utilised the nitroprusside-catalysed reaction to determine urea after enzy-matic conversion into ammonia with urease. Bohnstedt46 evaluated Riley’s34 method and proposed modifications for determining ammonia contents of steel digestates and Datsko and Kaplin47 measured ammonia in natural waters but used hypobromite instead of hypochlorite because a more sensi-tive stable colour was produced. Since 1960 the nitroprusside-catalysed reaction has become increasingly popular as a means of measuring ammonium ions in solution and this version has been most widely reported in the literature.Studies of the reaction mechanism and alterna-tive reagents have also been made and are summarised in the following sections of this review. Nature of the Reaction Mechanism Tarugi and Lenci5 studied the action of hypochlorites on substances containing the groups NH2 and NH and found that p-benzoquinone when heated with ammonia hypochlorite and zinc chloride produced N-chloro-p-hydroxybenzo-quinone monoimine (I) (Fig. l) which when treated with alkaline phenol produced a blue coloration. 2,6-Dibromobenzoquinone (11) treated in the same way produced 0 NCI I Fig. 1. Formation of I according to Tarugi and Lencis OH 1 I I B* Br Fig.2. Formation of I11 according to Tarugi and Len& a red crystalline substance that was identified as N-p-hydroxyphenyl-2,6-dibromo-p-benzoquinone monoimine (111) (Fig. 2) which is blue in alkaline solution. These workers found that in general amino acids and amino aldehydes react with phenol and hypochlorite to give intense blue colorations and that the same substances react with quinone and hypochlorite to give I indicating that this compound may be an intermediate step in the formation of the indophenol. Gibbs48 described a number of tests for identifying phenols and noted that amines and halogenated phenols react under oxidising conditions to produce indophenols of type 111. Gibbs also noted that I and phenol react to form the blue indophenol derivative of type I11 (Fig.3) and that quinone chloroimides do not react with phenols with substituted para-positions. This fact was confirmed in a later study of 65 phenols by Soloway and Santoro.49 Some phenols with unsubstituted para-positions but with substituted ortho- and rneta-positions do not react probably because of steric hindrance by the adjacent groups on the ring.48 Furth et ~ 1 . 1 6 isolated the product of the reaction between ammonium carbonate, phenol and sodium hypochlorite and concluded that the dye was of type 111. It is now established that the coloured end-product of the reaction is an indophenol of type 111. Lapin35 investigated the theoretical basis of the reaction and proposed a three-stage reaction mechanism (Fig.4); the first most specific and characteristic phase being the formation of a p-aminophenol the second the conversion of the p-amino-phenol into a halogen derivative of quinone imine and the third the synthesis of indophenol. This mechanism is consistent with the statement by Rodd’o that indophenols can be formed by condensation of a phenol and a quinone chloroimine although Rodd indicated that indophenols can also be formed by oxidation of a mixture of a p-aminophenol and a phenol. Bolleter et a1.51 proposed a mechanism (Fig. 5 ) in which the first stage of the reaction sequence was the reaction between ammonia and hypo-chlorite to form monochloroamine which then reacts with a phenol to form quinone chloroimine. This then reacts with another molecule of phenol to form the yellow associated indophenol which dissociates in alkali to give the blue indophenol 552 ANALYST MAY 1984 VOL.109 OH 1 (1) NH3 + OCI - . NH2Cl OH 0 NCl OH 0 Ill Q (j 0 OH 4 NCI Fig. 3. Formation of indophenol according to G i b b ~ ~ ~ H-OH-- - + OH OH NH2 Yellow Blue OH ++ NCI ( 2 ) Q + NH2 Fig. 5. Reaction sequence according to Bolleter et a/. 5 1 OCI- -(1) N H ~ + HOCI +NH& + ti2o OH Q 0 0 ( 3 4 + NCI NHp 5-Aminosalicylate 0 OH Fig. 4. Reaction sequence according to Lapin35 0 - 6 NH coo- @coo-These workers supported their theory by showing that p-aminophenol reacts with phenol in alkaline solution to form an indophenol with the same visible spectrum as the indo-phenol produced directly from ammonia.This over-all reac-tion sequence has been reiterated by Dambacher et al.52 and used as the basis for further research of the reaction mechanism.~3~~4 Bottcher et al.55 suggested that nitrous acid produced by the oxidation of ammonia by hypochlorite formed p-isonitrosophenol (mesomeric with quinone monox-ime) which then reacted with the phenate ion. In their limited conclusions these workers stated that the hypochlorite ion was essential to the reaction. Newel156 showed spectroscopic evidence for the presence of three chlorinated species (NC13, NHZC1 and NHC12) when chloramine-T and ammonium ions were present in distilled water systems at pH values between 4.1 and 5.3. The respective amounts of the species were pH dependent and Newel156 concluded that a transient compound of ammonium and chloramine-T was present under the reaction conditions.The formation of monochloramine (NH2C1) as the first stage of the reaction appears to be generally accepted.s1,52,57 Patton and Crouch53 investigated the kinetics and mechanism of monochloramine formation and concluded that there was a relatively narrow pH range (10.5-11.5) suitable for forming stable monochloramine under the Berthelot reaction conditions. These workers concluded that overall their results were consistent with the mechanism 0-dYco0- + (4) 0 Fig. 6. Reaction sequence according to Krorn5' proposed by Bolleter et al.51 Krom 54 investigated the reaction between sodium salicylate and sodium dichloroisocyanurate (sodium dichloro-S-triazine-2,4,6-trione) as the phenol and source of hypochlorite respectively and proposed a reaction sequence (Fig.6) with the formation of 5-aminosalicylate as an additional intermediate. Krom54 supported the inclusion of 5-aminosalicylate as a reaction intermediate by showing that the indophenol forms rapidly when 5-aminosalicylate is present. This contrasted markedly with the slow formation of the indophenol when the solution contained only sodium salicylate ammonia and the chlorinating agent ANALYST MAY 1984 VOL. 109 553 Under the reaction conditions there are a number of possible side reactions between reagents and reaction inter-mediates especially the chlorination of phenol.58 Lange et al.59 found that the chlorination of phenol proceeded via the sequence 2-chlorophenol 2,4-dichlorophenol 2,4,6-trichlorophenol independent of the reagent concentrations and reaction pH.These workers also found that the availabili-ties of phenol and chloramine-T were limited by phenol chlorination. The latest research has helped to clarify the probable reaction mechanisms although more data are needed on the equilibria involved. Reagents Phenols Phenols that undergo the Berthelot reaction normally have an unsubstituted para-position,48~4~ although some phenols with vacant para-positions may not react if there is steric hindrance from adjacent gr0ups.~8 Some para-substituted halogenated phenols undergo the reaction apparently because the para-halogen atom can migrate to other ring postions under suitable oxidising conditions leaving the para-position free.49 Phenol has been the most widely used phenolic compound, although early workers recognised that other phenols could be used in the reaction.15.18 These include thymol,” salicylic acid and its salts,l9 a-naphthol,@J guaiacol,61 o-phenylphenol,62 o-chlorophenol,63 2-methyl-5-hydroxyquinoline64 and rn-cres-With the exception of thymol and sodium salicylate the use of other phenolic compounds does not appear to be a regular occurrence.The use of thymol appears to be limited mainly to methods where the indothymol is concentrated by solvent extraction to increase sensitivity (see solvent extraction methods). Sodium salicylate is popular because it is a readily soluble solid and is less toxic and more convenient to prepare than sodium phenate.The use of phenol can result in the formation of poisonous o-chlorophenol under the reaction conditions.66 Glebko et al.67 concluded that the use of higher relative molecular mass phenols does not increase the sensitivity of the reaction significantly enough to justify their use and Krom54 stated that the sensitivities of optimised analytical systems using different phenolic compounds is similar. It appears that there are no major analytical advan-tages in using higher relative molecular mass phenols although Yamaguchi and Machidab2 reported a very high molar absorptivity when o-phenylphenol was used in the reaction. This finding however has not resulted in the regular use of this compound. 01.65 Hypohalite sources The formation of monochloramine as the first stage in the reaction mechanism is usually achieved in the presence of hypochlorite although the reaction proceeds when hypobro-mite is substituted probably through a similar mechanism.Hansenls found that solutions of hypobromite are easier to prepare at known strength and there have been a number of published methods using hypobromite especially when thymol is used as the phenolic compound. Datsko and Kaplin47 found that more indophenol was produced when hypobromite was used and that it was more stable than that produced with hypochlorite. However for optimised analy-tical procedures there appears to be no evidence that hypobromite is superior and its use has not been reported widely in recent literature. Solutions of known hypochlorite concentration can be readily prepared from solid organochlorine compounds that quantitatively hydrolyse to hypochlorite and Crismer24 used chloramine-T for this purpose.Hansen and Nielson28 criti-cised the use of this compound because heat was necessary to hydrolyse the compound effectively. However it has been used in subsequently published methods by Newe1156, Haussler and Hadju68 and Stegemann and Loescheke .69 Reardon et al. 70 used sodium dichloroisocyanurate (sodium dichloro-S-triazine-2,4,6-trione hereafter referred to as NaDTT) as the source of hypochlorite and Krom54 has described the mechanism of hydrolysis of this compound. NaDlT is convenient to use and more stable in solution than sodium hydrochlorite.54 There is an increasing usage of NaDTT reported in the literature although it may not be suitable for some applications.For instance Seely et al.71 found that NaDTT reacts with protein and amines so that if these are present in the sample solution a negative interfer-ence occurs because of a reduction in the hypochlorite concentration. Catalysts Russell30 noted that the sensitivity of the Berthelot reaction increased in the presence of chromium iron and man-ganese(I1) ions and described a method in which extra sensitivity was achieved by the standard addition of a small amount of manganese(I1) ions although Takei and Kato41 found that it did not affect the sensitivity of their method. This technique was used by subsequent workers until Crowther and Large40 proposed a method incorporating the addition of acetone which overcame the need for the manganese(I1) ion.Lubochinsky and Zalta36 discovered that sodium nitroprus-side [sodium nitrosopentacyanoferrate(III)] increased the sensitivity of the reaction considerably more than the man-ganese(I1) ion. Fenton72 described a method that used both acetone and nitroprusside but subsequent studies by Horn and Squire73 and by Harwood and Huyser74 concluded that this was unnecessary and that nitroprusside was superior to acetone as it provided more rapid colour development and a stable colour . Most methods now use the nitroprusside catalysed reaction, although the actual mechanism is not completely understood. Horn and Squire73 proposed that the actual catalyst was an iron(I1) nitritopentacyano complex which is the product of the reaction between sodium nitroprusside and sodium hydroxide.This conclusion was supported by Weichselbaum et aE.,75 who suggested that this complex promoted the formation of monochloramine from ammonia and hypoch-lorite. Patton and Crouch53 noted that sodium nitratopenta-cyanoferrate hydrolyses to sodium aquopentacyanoferrate. On the basis of their data they proposed a reaction mechanism involving the formation of a complex between aquopentacyanoferrate and monochloramine which reacts directly with the phenol to form the amine-substituted phenol. In support of this they quoted the work of Liddicoat et ~ 2 1 . ~ ~ and Hampson,77 who used potassium hexacyanoferrate(I1) in the presence of ultraviolet light to catalyse the reaction these conditions being suitable for the formation of aquopenta-cyanoferrate.78 Patton and Crouch53 decided to call the aquopentacyanoferrate a coupling reagent because it appeared to be destroyed in the course of the reaction and was therefore not a catalyst in the strictest sense.[Fe(CN),I4-Hexacyanoferrate(l1) Ultraviolet light ‘Iz b [Fe(CN)5H2012- [F~(CN)SNOI~- [Fe(CN15H20]3-A pentacyanoferrate(ll1) complex I I / H ~ O ~ ~ N H z C I [Fe(CNI50N0l4- b [Fe(CN)5NHzCl13-NHzCl Intermediate Fig. 7. Berthelot reaction outlined by Kroms4 Reactions involving complex cyanides as catalysts in th 554 ANALYST MAY 1984 VOL. 109 Krom54 investigated the role of complex cyanides in the reaction and represented the reactions diagrammaticaly (Fig.7). Kromj4 concluded that these complex cyanides could stabilise monochloramine and he suggested [Fe(CN)SNH2C1]3- as a possible complex. In addition, Kroms4 noted that both aquopentacyanoferrate(I1) and nitro-prusside in the presence of chlorine can form a pentacyano-ferrate(II1) complex which is used in organic chemistry as an oxidising agent in the oxidation of phenates to benzoquinones and in the oxidative coupling of phenol with p-aminophenol. On the basis of this latest research it appears possible that cyanide complexes can increase the rate of indophenol production by influencing more than one step of the reaction mechanism. Reaction Conditions pH Dependence The pH-dependent nature of most of the steps of the reaction mechanism means that the control of pH is important and this is recognised by early workers.9,30 The effects of pH are summarised below.p H dependence of monochloramine formation The formation of monochloramine is considered to be the first reaction step and Patton and Crouch53 suggested that the optimum pH values for maximum monochloramine produc-tion are 8.5-11:O. Alternatively Kroms4 suggested that the maximum concentrations may occur between pH 10.5 and 11.5. A number of workers have found that with the nitroprusside-catalysed reaction maximum indophenol forma-tion occurs above pH 11 S.36Jj9.74379-85 Kromj4 has suggested that the formation of indophenol can take place at higher pH values than would otherwise be possible because of the stabilisation of monochloramine by the catalysts derived from nitroprusside.This theory is supported by the fact that lower pH optima are reported in methods where nitroprusside is not present.8689 In contrast SagPo and Solorzano91 reported maximum indophenol formation at pH values between 9.8 and 10.4 in the presence of nitroprusside but these anomalies could be explained by the relationship between optimum pH and phenol concentration (as under p H dependence of phenol concentration). Generally it appears that the nitroprusside-catalysed reaction is less susceptible to minor pH changes near the optima than the uncatalysed reaction.73 p H dependence of phenol concentration RusselPJ and Bottcher et a1.jS found that with the manganese(I1)-catalysed reaction the phenol concentration affected the absorbance versus pH relationship.These work-ers found that the pH for maximum colour development varied with phenol concentration and the absorption maxima shifted to higher pH values with increasing phenol concentra-tion. Harwood and Huyser74 performed similar experiments with both acetone and nitroprusside as catalysts and found that with acetone there were two pH maxima (pH 11.3 and 12.3) and the pH of maximum absorbance decreased with increasing phenol concentration. Harwood and Huyser74 also found that for the nitroprusside-catalysed reaction the phenol concentration was less important because the absorbance versus pH graphs are essentially the same once a threshold concentration of phenol is reached. This is confirmed in the literature as the concentrations used show that there is no apparent relationship between final phenol concentration and optimum pH.pH dependence of indophenol dissociation The yellow associated indophenol dissociates in alkaline medium to give the blue form and the colour change is pH dependent and reversible51 (see under Mechanism). The pK, for phenol is 9.9 and the maximum dissociation of the respective indophenol is expected at a value a unit or two greater than this although the relative pH optimum will change with the phenol used because ring substitution will affect the pK of the phenolic hydrogen. Accordingly the pK, of the indophenol formed from salicylate is higher because the pK of phenolic hydrogen in this instance is 13.4. This is confirmed in the literature as the optimum pH for colour development is higher when salicylate is used.54 Pym and Milham92 investigated the salicylate - nitroprusside system and concluded that the reaction pH affected only the dissociation of the indophenol rather than affecting its rate of formation.In a similar study however Krom54 found that the main effect of pH was on the other stages of the reaction sequence rather than on the dissociation of the indophenol. This finding makes sense because a stable absorption maximum should occur once the pH of maximum dissociation of the indophenol is reached rather than an optimum “peak” reported in most studies. Pym and Milham92 noted an inflection in the absorbance -pH relationship at about pH 11.5 in addition to a maximum at about pH 13 and they suggested that this could be due to the existence of at least two chromogenic species which differ considerably in absorption maxima.p H dependence of wavelength of maximum absorbance (hmaX,) Pym and Milham92 observed that the h,,, changed with pH in the nitroprusside-catalysed salicylate system. Errors from this source however are probably negligible because of the broad absorbance band of indophenol and also because the rates of change of A,,, and absorbance with pH are effectively zero near the optimum pH. p H dependence of rate of colour development The effects of pH on the rate of colour development for both the phenol - acetone and phenol - nitroprusside reactions have been studied by Harwood and H ~ y s e r . 7 ~ Their work showed that with both systems the initial rate of colour development increased with increasing pH.Nitroprusside provided a more rapid rate than acetone if the pH was above 11. However at lower pH values (10.5) the reaction rate was faster with acetone. When acetone was used there was no induction period in the rate characteristics but with nitroprusside a definite induction period was observed although this disap-peared above pH 12. This behaviour was also observed by Weichselbaum et al.75 Under conditions of high pH nitroprus-side provided considerably faster colour development and reached a stable maximum after about 20 min whereas acetone required twice as long and this is why nitroprusside is favoured in most current methods. p H dependence of colour stability Harwood and Huyser74 found that the nitroprusside-catalysed reaction yields an essentially stable colour at pH values between 11.4 and 12.4 but that instability occurred at pH values above 12.7.Jones80 found that at low pH values there was a tendency for the colour to increase slowly with time. The colour produced when acetone is used as a catalyst decreases at pH values above 11.574 and it appears that for this and the nitroprusside-catalysed reaction stability decreases at pH values above the pH for maximum colour development. Lange et al.93 found that indophenol hydrolysed to 1,4-benzoquinone and 4-aminophenol in the presence of acids, alkalis and hypochlorite ions although the decomposition rate passes through a minimum between pH 11.5 and 13.Generally the nitroprusside-catalysed reaction offers advantages in terms of colour stability as it is less susceptible to small changes in reaction pH.73>g ANALYST MAY 1984 VOL. 109 555 p H dependence of gas volatilisation and absorption Loss of nitrogen as ammonia can occur from strongly alkaline solutions and Searcy et al. 94 assumed that this was the cause of low results obtained if the phenol reagent was not added directly after the alkaline hypochlorite. Weatherburn95 found that low results occurred only in the presence of nitroprusside and this suggests that a mechanism other than volatilisation is the cause of low results. Pym and Milham92 found that the colour intensity decreased with time and they attributed this to the absorption of C02 with the subsequent decrease in sensitivity being the result of decreasing pH.Problems with volatilisation and absorption have not been widely reported and it appears that errors from these sources are not a major problem. Reagent Concentrations Kroms4 emphasised the complex inter-relationships existing among the reactants and reaction intermediates so that the choice of optimum reagent concentrations and other reaction conditions is difficult. This is further emphasised by the large range of reagent concentrations used in the literature and it is evident that for any one combination of reagents there are probably several combinations of experimental conditions giving similar sensitivities and precision. For example similar sensitivities are quoted for several methods which differ widely in reagent concentrations and reaction conditions as given in Table 1.The broad aspects of reagent concentrations are discussed below. Phenols Sodium salicylate and phenol are the two main phenolic compounds used and it is common for higher concentrations of salicylate to be used because it is less reactive than phenol.54 The literature shows that there is wide variation in final concentrations used and it appears that if other factors are constant any increase in sensitivity with increasing phenol concentration effectively plateaus once a threshold concentra-tion is reached,74 although if very large excesses of phenol are used then sensitivity may decrease. Jf salicylate is used instead of phenol a higher optimum pH may result in order to produce an equivalent concentration of ionised compound because it has a higher p&.54 Hypohalite sources Sodium hypochlorite is a common source of the hypochlorite ion and solutions of organochlorine compounds (NaDTT, chloramine-T) are readily prepared although the rate of hydrolysis of these compounds to hypochlorite may depend on the reaction conditions.Harwood and Huyser74 reported a sharp optimum hypochlorite concentration if acetone was used as the catalyst. If nitroprusside is used the sensitivity of the reaction levels off once a threshold hypochlorite concen-tration is reached although increasing the hypochlorite well past the optimum decreases sensitivity. When chloramine-T is used Lange et al.101 found that maximum sensitivity and colour stability are reached if the chloramine-T concentration is limited to between 29 and 44% of the phenol concentration.However with other hypochlorite sources the concentrations involved are often an order of magnitude less than the phenol (Table 1). This means that the hypochlorite concentration is not particularly critical once the optimum is reached when nitroprusside is used as the catalyst. Nitrop russide It has been well established that nitroprusside is the source of the best catalyst for the reaction although the formation of the effective species may depend on reaction conditions. Harwood and Huyser74 showed that if other reagent concen-trations were kept constant increasing the nitroprusside concentration above the optimum did not affect the sensitivity of the reaction.Development Time and Temperature In some of the first uncatalysed methods using phenol and hypochlorite reagents a development temperature of 100 "C was often used and maximum colour development occurred relatively rapidly although the intensity of the colour faded rapidly with time.20 With the nitroprusside-catalysed reaction Weatherburn95 found that there was a complicated inter-relationship between reaction temperature development time and nitroprusside concentration and this makes comparisons between methods difficult as a considerable range of nitro-prusside concentrations are used. Colour development is apparently faster at 37°C than at room temperature (15-25°C) and a development temperature of 37 "C has been widely used.Weatherburn95 observed that increasing the reaction temperature above 37 "C generally resulted in an increased maximum absorbance although this was accompan-ied by an increase in the time taken to obtain a stable maximum. Weatherburn95 also found that colour develop-ment at 100 "C was rapid although the colour intensity decreases rapidly with time and this may explain why Muftic102 stated that heat retarded the reaction. It appears that the use of high temperatures is of little benefit in influencing the sensitivity and development time for the nitroprusside-catalysed reaction especially as the possibility of interferences caused by the hydrolysis of organic nitrogen compounds increases with temperature. If organochlorine Table 1. Reagents final concentrations and sensitivities of some methods based on the Berthelot reaction Sodium nitroprusside Phenolic concentration/ Hypochlorite concentration/ concentration/ Final Final final Reference compound g I-' compound g I-' mg 1- l Scheiner".. . . . . . Phenol 12 JaenickeS4 . . . . . . Phenol 8.3 KorolefPS . . . . . . . Phenol 0.9 HarwoodandKuhnyy . . . . Phenol 18.8 Verdouw et a1.66 . . . . . . Na salicylate 40 Pym and Milham92 . . . . Na salicylate 10.2 Krom54 . . . . . . . . Na salicylate 40 FraserandRussell100 . . . . Nasalicylate 17 Patton and Crouch53 . . . .o-Chlorophenol 0.45 Kempers" . . . . . . Phenol 5.6 * Sodium dichloro-S-triazine-2,4,6-trione. NaOCl NaOCl NaOCl NaOCl NaOCl NaOCl NaDDT* NaDDT NaDDT NaOCl 1 .0 0.167 0.22 0,004 1 .0 0.745 0.3 0.2 5.0 0.045 40 27 83 1 1 40 1400 72 1200 120 2 Optimum PH 11.3-11.7 11.411.8 11.5-1 3.0 10.4 12.0 12.0 12.8-13.1 12-1 2.4 -10.5-1 1.5 E x 1041 lmol lcm-1 1.23 2.04 0.87 2.01 2.04 2.04 1.87 1.83 1.23 2.0 556 ANALYST MAY 1984 VOL.109 compounds are used as the source of hypochlorite higher reaction temperatures may be necessary to obtain maximum sensitivity because the rate of hydrolysis of these compounds to hypochlorite may be temperature dependent. Order of Addition of Reagents In most methods the phenol is added prior to the hypochlorite and at high concentrations of hypochlorite little or no indophenol is produced if the hypochlorite is added first.lo3,104 At lower concentrations the hypochlorite can be added first with no loss of sensitivity38379 and this order of reagent addition fits the proposed reaction sequence .53 Some workers have reported that sensitivity decreases with increasing interval between the addition of hypochlorite and phen01,~~JO~ and the timing between the addition of hypochlorite and phenol becomes increasingly critical as reaction temperature increases.86 An explanation for this phenomenon is that the initial reaction product monochloramine decays with time, and this is supported by the fact that there is no decrease in sensitivity when all reagents are added in quick succession, irrespective of the order of addition.92 The decrease in sensitivity could also be explained if ammonia was volatilised after the addition of alkaline hypochlorite,Y4 but Weather-burn95 found that this only occurred when nitroprusside was present indicating that a factor other than alkalinity is responsible.The addition of phenol first has advantages in some instances as it stabilises the sample solution against bacterial changes. 106 However the most appropriate order of reagent addition probably depends on the application of the method, as the order of reagent addition may affect the susceptibility of the reaction to interferences caused by the hydrolysis of organic nitrogen compounds by the alkaline hypochlorite. lo7 Colour Stability Indophenols produced by the Berthelot reaction appear to be stable. The duration of reported stability ranges from 4 to 6 hwJ08 and up to 24 h or longer if the solutions are protected from carbon dioxide absorption92 or direct sunlight .9* Long-term stability has also been reported for the indophenols produced from sodium salicylate92 and o-chlorophen~l,~~ so that for practical purposes colour stability is independent of the constituent phenol used.The stability of indophenols also appears to be largely independent of the method of formation, provided that high development temperatures are not used .20 Absorption Maxima The indophenols have broad absorption bands and the position of the absorption maximum is dependent on the constituent phenol although minor changes may occur depending on reagent concentrations,105 pHY2 and the presence of an organic solvent.Table 2 lists the reported absorption maxima for various indophenols. Table 2. Absorption maxima (Amax.) for some indophenols Constituent phenol Lilax./nm Phenol . . . . . . . . . . 630 Sodiumsalicylate . . . . . . 667 o-Chlorophenol . . . . . . 650* Thymol . . . . . . . . . . 6604 Guaiacol . . . . . . . . . . 680 o-Phenylphenol . . . . . . 690 m-Cresol . . . . . . . . . . 665 a-Naphthol . . . . . . . . 720 2-Methyl-5-hydroxyquinoline . . 710$ * In acetone. t In acetone - ethanol. $ In ethanol. Reference 105 100 53 109 62 62 65 60 64 Sensitivity The sensitivities of optimised analytical systems using differ-ent phenolic compounds have been reported as essentially similar,54 and molar absorptivities (E 1 mol-1 cm-1) of the order of 2 x 104 have been reported for phenol,97 sodium salicylate ,66 o-chlorophenol53 and guaiacol.6l Higher molar absorptivities have been reported for o-chlorophenol (E = 3.14 X 104 1 mol-1 cm-1),63 rn-cresol (E = 2.48 x 104 1 mol-1 cm-1)a and o-phenylphenol (E = 6.74 x lo4 1 mol-1 cm-1),6* although the application of the higher sensitivities of these methods has not been reported in the literature.In practical terms an analytical system giving a molar absorptivity of 2 x 104 1 mol-1 cm-1 will give an absorbance of 0.14 in a 10-mm cuvette with 0.1 pg ml-1(7.1 x M) of nitrogen in the final solution. Interferences Nitrogen Compounds Positive interferences can arise when nitrogen groups such as amino amide or amine are hydrolysed to ammonia under the however that the apparent positive interference by a number of organic nitrogen compounds was a result of ammonia contamination.Urea breaks down to ammonia when heated in alkali but Orr12 found that ammonia could be measured in the presence of urea if the reaction was carried out in an excess of phenol and without heating. Domnasll2 found that if calcium hypochlorite was added to the sample at pH < 5 before the addition of alkaline phenol then the indophenol blue colour was obtained with a large number of nitrogen compounds and WearnelO7 noted that the reaction is more specific for ammonia if the reaction pH is greater than 7. Searcy et ~ l . 9 ~ showed that the specificity of the reaction was affected by the order of addition of reagents. Amino acids treated first with alkaline hypochlorite exhibited greater chromogeneity than those reacted initially with phenol and the amount of colour obtained did not appear to be related to the number of nitrogen groups present in the amino acid.These workers found that there was no chromogenic response on the compounds tested by Wearne107 and they attributed this to the fact that alkaline rather than acid hypochlorite was used. Inhibition of the reaction has been reported in methods using the catalysed reaction where lower hypochlorite concen-trations are used. Reardon et al.70 found that a range of amino acids gave no chromogenic response when NaDTr was used as the hypochlorite source. However Seely et al.71 found that amino acids and protein inhibit the reaction because these compounds react with the NaDTT and thereby effectively reduce the hypochlorite concentration.Organic nitrogen compounds also react directly with hypochlorite113 and the inhibition of the reaction by nitrogen compounds has been Ngo et a1.120 found that indophenol formation was strongly depressed by groups with primary secondary and tertiary amino groups and by aromatic amines. These workers postulated that nucleophilic addition of alkylamine to quinone chloroamine retarded indophenol formation. Scheinergh noted that nitrate ions can complex with ammonia and this could be a possible source of interference. Nitrogen com-pounds that are possible interferents are summarised in Table 3 together with the relevant references; however the nature and extent of interferences from these sources will depend on individual reaction conditions.reaction conditions. 7.10.13~5,34~1,90,110,111 Fenton72 found, widely reported.41.81,96,114,116-119 Metals Some metals interfere with the reaction if present in sufficient amounts. Enhancement and inhibition of the colour reaction have been reported for the same metal. Table 4 surnrnarise ANALYST MAY 1984 VOL. 109 557 Table 3. Interferences nitrogen compounds Reference numbers Positive Negative Compound interference interference Allantoin - . . . . . . . . . . 112 Aniline - . . . . . . . . . . 110,121 Amides* - . . . . . . . . . . 107,112 Amines* . . . . . . . . . . 15,51,110 71,77,81,118, 121-123 120 Amino acids* . . . . . . . . 7,10,12,34 71,81,95,114. 90,94,107 116,117,119, 112.124-126 120 Aminophenol .. . . . . . . 19 -Alloxan - . . . . . . . . . . 11 2 Creatine - . . . . . . . . . . 17,112 Creatinine . . . . . . . . Cyanide - . . . . . . . . . . i 4,117 Ethylcarbamate . . . . . . 112 -Hydrazinesulphate . . . . . . -Hydroxylammonium sulphate . . -113 -118 118 115,117,118, 120 71 Protein -Thiocyanate . . . . . . . . 114,117 41 Thiourea . . . . . . . . . . - 118,120 Urea . . . . . . . . . . 12.107,112 -Uricacid . . . . . . . . . . 107,111.112 113 Hydroxylamine . . . . . . . . 34,55 127 -Nitrite . . . . . . . . . . - 77,96,114. . . . . . . . . . . 123 * All compounds in these groups are possible interferents. Table 4. Interferences metals Reference EIement" A l . . . . . . B a .. . . . . C a . . . . . . c o . . . . . . Cr . . . . . . C u . . . . . . F e . . . . . . H E . . . . . . L i . . . . . . Mg . . . . Mn . . . . N a . . . . . . Ni . . . . . . P b . . . . . . Z n . . . . . . Positive interference - . . - . . - . . . . 114,117,137,138 . 30. 137 - . . . . 30 . . 57,139 - . . - . . . . 30,34,51.90 . . 92,137 . . 114,117 . . 51,137 - . . Negative interference 97,114,117 34 34,97,130 102,120 30,34.55,68,89.90, 98,104,118,130-133 96 114,117,118, 132,134-136 120,137 97,120,130,136 118 120 136 136 34 --* Check individual references for oxidation states. Table 5. Interferences non-metals Reference Species or compound Br- . . . . . . F- . . .. . . . . I- . . . . . . . . S2- . . . . . . sop . . . . . . Sz03'- . . . . . . Thiols . . . . . . Dimethyl sulphoxide Ascorbicacid . . . . Se . . . . . . . . Positive Negative interference interference - 34.51,87.115 114 117 -114,117 123 98,118,120,126 120 41 - 114,117,118 115 -120 120 120 ---- 83,132 the reported metal interferences and the apparent contradic-tion in some instances indicates that the nature and extent of any interference depends on the reaction conditions used. There is considerable evidence that copper and mercury can cause inhibition of the reaction and as these elements are widely used as catalycts in digestion methods for determining total nitrogen (e.g. Kjeldahl digestion) possible inter-ferences from these sources should be investigated for any particular method.Interferences by most other metals appear to occur only when the metal is present in more than trace amounts or when they are present in higher concentrations than nitrogen. Interferences from some elements can be minimised by the use of chelating agents such as EDTA citrate or tartrate,'*g although inhibition of indophenol formation by such agents has also been reported.57.129 The mechanism of the inter-ference by elements such as copper and mercury is not understood although the fact that such elements form complexes with ammonia may be significant. There appears to be little understanding of the way in which manganese(I1) ions catalyse the reaction although interferences from this element are reduced if the nitroprusside-catalysed reaction is used.Non-metals Sulphur in a range of oxidation states selenium and the halogens are reported as interferences among the non-metallic elements (Table 5 ) . Bolleter et a1.51 have suggested that the bromine ion interferes by precipitating phenol as tribromo-phenol. However the mechanism of these interferences is unclear and for sulphur compounds may result as a modifi-cation of the oxidising power of the hypochlorite source or interference by a sulphur nucleophile12() reacting with reaction intermediates. Salt Effect Samples with high salinity such as sea water require special treatment.34 However the problems associated with these samples may be a function of reaction pH,14" as high salinity samples tend to lower the reaction pH possibly as a result of precipitation of insoluble hydroxides.Froelich and P i l ~ o n l ~ ~ noted that in continuous flow analysis (CFA) systems, differences in salt content of sample solutions can cause errors because of the nature of the flow cell (see Sea water under Applications). Light Cri~mer2~ observed a decrease in indophenol colour intensity caused by decomposition of the indophenol compound on exposure to light and Fawcett and Scott'os noted that the depression in absorbance by direct sunlight may be of the order of 10%. Fawcett and Scottl"' found that filament and strip lighting did not decompose indophenol and recommen-ded that colour development should proceed in darkness or where natural lighting is minimal.Kempersy7 found that if phenol and nitroprusside are used as a mixed reagent then the blank value increases on exposure to sunlight. Gravitz and Gleye142 found that exposure to sunlight of the mixed phenol -nitroprusside reagent results in the formation of a compound the same as that formed in the presence of ammonia. The visible spectra of blanks that had been exposed to sunlight were similar to the spectra obtained in the presence of ammonia and Gravitz and Gleye142 suggested that the nitrogen in the indophenol formed in the blank originated from the nitrosyl group of the nitroprusside. Liddicoat et al.'6 and Hampson77 used potassium hexacyanoferrate( 11) and ultraviolet light for colour development and obtained lower blanks compared with the use of nitroprusside as the catalyst.Verdouw et a1.66 found that with phenol hexacyanoferrate(I1 558 ANALYST MAY 1984 VOL. 109 was preferred to nitroprusside as it produced better colour stability in both the light and the dark although laboratory light did not interfere when salicylate was used as the phenolic compound. Bower and Holm-Hansenl43 found that the indophenol produced when salicylate was used was light sensitive and the light sensitivity of methods may be dependent on the phenol used. Under normal laboratory conditions the effects of light are likely to be of concern only if very low levels of nitrogen are being determined where it is desirable to keep the blank to the lowest possible level. Under these circumstances it appears to be desirable to use the methods that use the catalytic action of the aquopentacyanoferrate(II1) complex for colour development and to avoid the use of mixed phenol -nitroprusside reagents.Solvent Extraction Methods Indophenols produced by the Berthelot reaction are soluble in a variety of organic solvents and solvent extraction has been used in uncatalysed methods as a means of concentrating the indophenol to improve sensitivity. The indophenol produced when thymol is used as the phenolic compound appears to be particularly suitable for this p r o ~ e d u r e l ~ . 2 ” 2 8 ~ 3 5 . ~ ~ ~ 1 ~ ~ ~ ~ ~ and methods in which phenol has been used have been des-cribed,56,86J31J50 as has a-naphthol. 60 The increase in sensi-tivity resulting from the use of the nitroprusside-catalysed reaction appears to be sufficient for most applications and solvent extraction as a means of increasing sensitivity has not been reported in recent literature.Automated Methods Automated versions of the Berthelot reaction are widely used where there are large numbers of samples for nitrogen determination. The automated reaction has also proved popular because problems caused by the complex reaction equilibria are effectively eliminated in systems where reaction conditions are closely controlled. The majority of published automated methods are based on the use of CFA although methods using automated discrete analysis have been des-cribed,151 as have flow injection analysis (FIA) methods.152 Automated methods are adaptations of manual methods and as such are subject to the same reaction variables described in the previous section so that only the main aspects of CFA methods are outlined below.Reagents The nitroprusside catalysed reaction is the most commonly used automated version of the Berthelot reaction because the reaction is more sensitive and reaches equilibrium faster than the uncatalysed reaction. These attributes are desirable in flow systems where there are advantages in minimising flow time to achieve better sample resolution. Greater sensitivity also means that the sample input can be reduced and the resulting dilution minimises interferences. The uncatalysed reaction may be suitable however when large concentrations of ammonium-nitrogen are measured. 153 All the published automated methods are based on aqueous chemistry and the phenolic compounds used are either phenol or salicylic acid because of the solubilities of the respective sodium salts.Sodium hypochlorite solutions are the most widely used source of hypochlorite ion but there appears to be increasing use of organochlorine salts such as NaDTT. Reaction Conditions The reaction conditions closely parallel those used in manual methods and reaction temperatures of betwen 30 and 50 “C are used for the nitroprusside-catalysed reaction,’”,1jj whilst temperatures of the order of 95°C are used in methods without catalysts.l56,1’7 The addition of phenol to the sample at the first stage of the reaction has been traditional in manual methods and the literature on automated methods has reflected this trend although addition of hypochlorite first has been reported in recent studies.’s8.159 Factors such as the order of addition of reagents are likely to cause less problems in automated methods and SearleI6” found that the order of reagent addition did not affect the sensitivity as long as the reagents were added separately and in quick succession.Interferences Automated methods are subject to the same interferences as manual methods although they are more susceptible to interferences that affect the rate of reaction. Measurement often occurs before the reaction reaches equilibrium and this may explain some apparent anomalies in the nature of some interferences. For example it is well established that mercury has an inhibiting effect on the reaction in manual methods (Table 4) but Gehrke et al.5’ and Davidson et al. l35~ report enhancement of the reaction in automated methods when mercury is present. This anomaly could occur if less indo-phenol is formed in the presence of mercury despite an increased rate of formation. There are essentially three ways of minimising interferences in CFA. Flow dialysis has been used by many workers156J61.162 and the use of automated distillation of ammonia prior to the colorimetric reaction has also been described. 157.163 Chemical chelating agents are adequate for many applications128.12y and have the advantage that the simplicity of the flow system is retained so that there is no loss in performance. Analysis Rates Apart from the increase in precision that automated analysis normally offers over manual methods the other main advan-tage is the potential for greatly increased sample analysis rates.Rates of 60 h-1 are readily obtainable although miniaturisation of the flow system allows higher analysis ratesl6t) and even higher rates are obtainable by FIA. The inclusion of dialysers distillation modules or dilution loops normally lowers the analysis rate because the flow path of the system is increased. The balance between analysis rate and precision will vary depending on the application of the method. Radiochemicai Methods Awl64 increased the sensitivity of the nitroprusside-catalysed reaction by radioiodinating the phenol with iodine- 125. The radioactive indophenol produced by the reaction is extracted into aqueous solution and counted with a gamma counter.The method is highly sensitive although subject to some interfer-ences from amino acids and other substances that interfere in the indophenol reaction. Applications A quick guide to references for particular applications is given in Table 6. A more complete discussion of applications is given below. Agriculture Fertilisers Ammonium-nitrogen CFA methods have been described by Docherty,l65 Wrightman,166 King and Scobie,’67 Nebe1,168 Ridleyl6Y and Viswanathan et a1.17O None of these methods utilise the catalysed reaction probably because of the large amounts of ammonium present in fertiliser extracts. Viswana ANALYST MAY 1984 VOL. 109 559 ~-Table 6. Quick reference guide to applications Reference Reference Manual Application method Fertilisers NH4-N .. . . -TotalN . . . . -CFA* method Manual method CFA* method Application Meat NH4-N . . . . TotalN . . . . Milk NH4-N . . . . TotalN . . . . Sugar NH4-N . . . . TotalN . . . . TotalN . . . . TotalN . . . . Brewing materials Unspecified food Sea water NH4-N . . . . 165- 170 57,71,162,172-174,352 285 --286-289 Feeds Soils TotalN . . . . -NHA-N . . . . 31,44,85,97,100, 179-182 TotalN . . . . 85,97. 181,378 290 --29 1-295 175-1 78 128,160,183-187 296 297 296 297 123,133,153,157, 160,161,163,171, 185,188-194 195,196 298-300 Ureaseactivity . -364 Plant material Total N . . . . 29,83,197-203, 353.371 129,139,151,153, 163,188,189,191, 206216,353,354, 373,374,386 34,43,66,76,77,87, 91,98,106,114,117, 118,127,130.134, 140,142,143,148, 150,301-303, 365368,370 123,126,140,154, 304306,369,375 Blood NH4-N .. . . 17,42,72,73,75, 111,122,124,125, 135,182,217-234, 250,355357,362 Urea . . . . . . 14,20,22,45,105, 217,239-246,360-362,372 Urine NH4-N . . . 12,18,20,24,113, 250,251 Urea . . . . . . 14,105,256-258 Other biological extracts NH4-N . . . . 10,11,15,21,38, 112,119,124,252, 260-268,362 80,81,84,88 103, 132,135,269,270-277 Urea . . . . . . 377 Total N . . . . 35,39,46,52,55,69, Fish NHj-N . . . . 147,149,283 156,235-238,358, 359 TotalN . . . . Non-saline water NH4-N . . . . 159 312-320.379 47,96,99,104,127, 370,376 130,146,307-311, -247-249 TotalN .. . . 136,138,158,321-325 25-3-2551 248,255,259,363 Metals Total N . . . . 26,27,33,89,115, 131,144,326-336, 381-385 Petroleum Gases Pharmaceutical products TotalN . . . . 37,340,341 NH4-N . . . . 343-345 Paracetamol . . 347,348 Phenacetin . . . . 347 Sulphanilamide . . 349-351 * Continuous flow analysis. 137.337-339 268 342 278-282,380 346 346 284 than et al.170 found the method suitable for measuring ammonium in the presence of urea indicating that hydrolysis of urea was not encountered under the reaction conditions used. Total nitrogen methods have been described that include a digestion step to convert other nitrogen forms into ammonia prior to colour development by CFA. Gehrke et aLS7 used a manual digestion procedure to reduce nitrate and convert organic nitrogen prior to using the manganese(I1)-catalysed phenol - hypochlorite reaction.Mercury used to catalyse the digestion inhibited the sensitivity of the colour reaction but errors from this source were minimised by adding mercury to the standards. Moore and Kelly171 used a manual nitrate reduction step prior to automated Kjeldahl digestion before measurement of ammonia with the manganese(I1)-catalysed reaction. Tartrate was used to prevent precipitation of hydroxides in the flow stream. A similar procedure was described by Gehrke et ~1.172 Gehrke et a1.173 described a method that included nitrate reduction with chromium and titanium as part of the automatic continuous digestion procedure; however the procedure was replaced by a manual conversion with a block digester174 because the automated digestion was complex and did not completely convert all organic nitrogen into ammonium.This method also used the nitroprusside-catalysed salicylate - hypochlorite reaction for increased sensitivity. In a further development Wall and Gehrke'62 described a method for separately measuring ammonium nitrate and urea-nitrogen. Ammonium is measured by the nitroprusside-catalysed salicylate - hypo-chlorite reaction on three sub-samples one untreated one subjected to nitrate reduction and the other to enzymatic urea hydrolysis. The combination of these three chemistries means that individual contributions of any one fraction to the total nitrogen can be determined by substituting distilled water for the reagents of individual chemistries.Feeds CFA methods for the measurement of total nitrogen in feeds has been described by Law et al. ,175 Gehrke et al. ,176 Wall and Gehrkel77 and Basson. 178 These methods involve a manual Kj eldahl digestion because automated continuous digestion is not satisfactory.17* The amount of nitrogen in feeds is usually high and the uncatalysed reaction provides adequate sensitiv-ity.175,'78 Wall and Gehrke,177 however used flow dialysis as a means of dilution prior to measurement using the nitroprusside-catalysed reaction. Soils and soil extracts In the analysis of salt extracts of soils interferences from extracted cations are not a major problem although the use o 560 ANALYST MAY 1984 VOL.109 complexing agents is recommended In some soils extracted organic matter may cause errors by increasing the background absorbance,lbl but this is not a major problem if the nitroprusside-catalysed reaction is used. In Kjeldahl digests interferences from metals solubilised by the digestion need to be minimised by the use of complexing agents in conjunction with either high dilution dialysis or distillation of the sample digest. In such procedures it is also necessary for the acid digest to be accurately neutralised so that the same reaction pH is obtained for all samples and standards. Manual methods. Many methods for the measurement of exchangeable or adsorbed ammonium in salt extracts of soils have been described.31.44.85.179-182 Fraser and RusselllOO used the salicylate - nitroprusside - NaDTT reaction to determine the cation-exchange capacity of clays by measuring the ammonium present directly in the ammonium-saturated clay after acid digestion.Insoluble hydroxides which formed under the reaction conditions were separated by centrifuging prior to spectrophotometric measurement. Kudeyarovlsl used the reaction to determine total ammonium- a-amino-and amide-nitrogen after an oxidation procedure and Kem-per97 measured Kjeldahl nitrogen with the nitroprusside-catalysed reaction after digestion and distillation. Nitrate and nitrite were also measured if a reduction step was included prior to distillation. Hinds and Lowe85 measured ammonium in Kjeldahl digests and used EDTA as a complexing agent to remove metal interferences.CFA methods. Ammonium can be determined directly in salt extracts of ~oils128J60~18~18~ and on Kjeldahl digests. 12*~53,16~~8-191 In methods that are applied directly to sample digests the possibility of interferences can be mini-mised by high sample dilution and the use of a complexing agent. Under these conditions the nitroprusside-catalysed reaction is nearly always used because the increased sensitivity offsets the dilution factor.153.160 Other methods of over-coming interferences are the use of flow dialysisl337161J92 and automated distillation. 1 5 7 ~ 6 3 ~ 8 5 . 1 9 3 ~ 9 4 Distillation is the most effective means of removing possible interferents. The use of automatic distillation or dialysis however does increase the complexity of the flow system and this usually decreases the rate of sample analysis.These factors have to be balanced against the advantages of such systems when considered for any particular application. Methods for the determination of urease activity by measuring the ammonia produced have been described by Hofman and Teicher195 and Searle and Speir.196 Plant material Total nitrogen in plant material is determined after a digestion procedure (usually Kjeldahl digestion) and is subject to fewer interference problems compared with materials such as soils. This occurs because of the relatively high nitrogen and low mineral content of plant material with the consequent low concentrations of possible interfering ions in the digestate. In addition the free acid content of plant digests is relatively constant and errors arising from differing reaction pH values is less likely.Manual methods. A number of methods based on the nitroprusside-catalysed reaction use pheno183-197-199 and sali-cylate206203 as the active phenols. Mitchell83 used EDTA to complex digestion catalysts such as copper or mercury and neutralised the digests to pH 7 before colour development. Phosphate buffers to stabilise reaction pH have been u~ed,197~201,203 although Burrows198 found that the method applied to Kjeldahl digests of leaves was unaffected by normal fluctuations in digestion acidity residual catalyst or mineral contents. Colour development times have ranged from 90 min at room temperature197 to 30 min at 25 OC203 and 20 min at 37 OC.201 CFA methods.The reaction based on phenol - hypochlorite reagents has been used for the analysis of Kjeldahl digests"1.18*~2""2*7; procedures that combine automatic con-tinuous digestion have also been used. 139.208~09 In addition, flow systems capable of measuring other elements such as phosphorus simultaneously on the plant digest have been described. 1*9,*91,21(~15 The nitroprusside-catalysed salicylate reaction has been used by Crooke and Simpson,129 who used high dilution and EDTA to overcome interferences from the mercury catalyst used in the digestion. Methods using these reagents have also been described by Bietz216 and Holz and Kremers. 153 Keay and Menage163 used automated distillation prior to the phenol - hypochlorite reaction to minimise interferences and Cadahia1g9 used flow dialysis for the same reason.Only a few of the described procedures have incorporated a complexing agent to prevent interferences from catalysts used in the digestionl29-*39 or have reported the need to control reaction pH with a buffer solution.191 Clinical Blood ammonia In modern clinical laboratories ammonia in blood is usually determined by an enzymatic method in which glutamate dehydrogenase catalyses the condensation of ammonia with a-ketoglutenate with simultaneous oxidation of reduced nicotinamide adenine dinucleotide phosphate (NADPH). The decrease in absorbance at 340 nm caused by the disappear-ance of NADPH is directly proportional to the ammonia concentration and precipitation of protein or separation of ammonia from plasma constituents is unnecessary.217 This method is commonly used in laboratories that use multi-channel CFA systems (J.Newton Wellington Hospital Laboratory personal communication). Over the years, however a number of methods utilising the Berthelot reaction have been described although there is little reference to such methods in recent literature indicating that the enzymatic method is generally favoured in modern laboratories. Manual methods. With the Berthelot reaction a pre-treatment is necessary to isolate the ammonia from the nitrogenous components of blood (e.g. protein) and from the red blood cells which contain enzymes that can rapidly increase the ammonia content with time.217 Amongst the methods of separation used have been diffusion"J18-2l9 and aeration17J2" of ammonia from alkaline blood to an acid collecting solution suitable for the spectrophotome tric proce-dure.The disadvantage of these methods is that hydrolysis of nitrogen compounds can occur under alkaline conditions, resulting in an over-estimation of ammonia. Gangolli and Nicholson,220 however overcame this problem by precipitat-ing protein and amino acids with mercury(I1) chloride - lead acetate prior to aeration. A large number of methods using ion-exchange resins to isolate ammonia have also been decribed72,73,**.122.125.221-227 and these involve passing blood through an ion-exchange resin so that ammonium ions are retained for subsequent elution and spectrophotometric deter-mination. Sources of error can occur in these methods as plasma protein can adsorb on to the surface of resin granules72 and some basic amino acids are also adsorbed.81 Both of these can cause errors in the colorimetric step although they can be minimised if diluted blood is used.82 Methods involving de-proteinisation of blood by precipitation prior to spectro-photometric determination overcome the problem of falsely high ammonia values caused by enzymatic ammonia produc-tion in stored samples.1 1 1 . 1 2 ~ . 2 2 0 . 2 2 ~ 3 ~ Some amino acids, particularly proline can still inhibit the colorimetric reaction if present in sufficient amounts.111.230 Weischelbaum et al. 75 described a kinetic method based on the measurement of the rate of the nitroprusside-catalysed reaction but this required special instrumentation.CFA methods. Flow dialysis to separate the ammonia from possible interferents has been used in some labora ANALYST MAY 1984 VOL. 109 561 tories.156,235-"* This appears to overcome the problems of hydrolysis of non-ammonia nitrogen or inhibition of the colour reaction that can be experienced in manual methods if a separation step is not included. Blood urea Manual methods. The Berthelot reaction has often been used for the indirect measurement of urea by determinations of the ammonia produced by enzymatic urease treat-ment.14,20,22,45,105.217.23~246 Interferences from amino acids and non-protein nitrogen are not normally a problem if the correct reaction conditions are ~ s e d . 2 ' ~ Richterich and Kuef-fer2.17 used a discrete analyser to measure blood urea and investigated a number of possible interferences including lipemia bilirubin amino acids and drugs and found that interferences only occurred in the presence of marked haemolysis.CFA methods. Methods using automated urease treatment, followed by flow dialysis to separate the ammonia produced for spectrophotometric determination have been de-scribed.246.248.249 Wilson248 noted that a small amount of urea was broken down to ammonia under the reaction conditions but this can be avoided if low concentrations of hypochlorite are used. 196 Most clinical laboratories using automated equipment use the diacetyl monoxime reaction (J. Newton, Wellington Hospital Laboratory personal communication), which has the advantage that a conversion step is not required, although it is not as sensitive as the Berthelot reaction.Urine ammonia Manual methods. The reaction has been applied directly to urine12,250,251 although the reaction conditions must be such that hydrolysis of compounds such as urea does not occur. This can be achieved if an excess of phenol is added first and heat is not used for colour development .12,252 Urine may contain many nitrogenous compounds that interfere with the reaction and separation of possible interferences by micro-diffusionlsJ0 or distillation24 before colour development has been used. Gipps et al. 113 overcame the need for a separation step by employing large dilutions (up to 1000-fold) before using the nitroprusside-catalysed reaction. CFA methods.Methods based on dialysis of the sample before colour development have been described .253-255 These methods are similar to those described for blood analysis and dialysis appears to effectively remove interfering compounds. Urine urea Manual methods. Methods previously described for use with blo0d1~,~~)5,*56258 are also applicable to urine and are all based on the conversion of urea into ammonia by urease. CFA methods. As for manual methods some methods previously described for bl00d248~259 are also applicable to urine as is a two-channel system for measurement of ammonia and urea in urine described by Burnette et al.255 These methods are also based on dialysis of the urease treated sample prior to colour development. Other biological extracts Ammonia.The Berthelot reaction has been used to determine ammonia in a wide range of biological extracts. These include the measurement of ammonia produced by bacteria,10.11-1~ ammonia in cerebrospinal fluid,8 in urea,252 in tissue extracts,I24 in faeces,260 in kidney slices,119 in feeding solution ,Zhl in glucose reaction products262 and in unspecified biological iiquid."JhL'-'h' The reaction has also been used for determining (x-NH2 groups in amino acids,26h for detecting L-asparagine and L-arginine in enzyme reactors,267 for deter-mining the relative molecular masses of compounds contain-ing nitrogen which is easily converted into ammonia,38 and for detecting amides and related compounds. 112 In addition, Mor et a1.268 described a CFA method for measuring ammonia in yeast.Total nitrogen. (i) Manual methods. Most digestion pro-cedures effectively destroy all potentially interfering nitrogen compounds and many methods based on Kjeldahl digestion have been published. These include the use of the uncatalysed reaction,269 the manganese(I1)-catalysed reaction23.46355 and the acetone-catalysed reaction.88 In addition reactions based on alternative reagents have been used such as the thymol - hypobromite reaction,270 the a-naphthol - hypo-chlorite reaction271 and the salicylate - hypochlorite reac-tion.135 The most recently published methods are based on the nitroprusside-catalysed reaction.52.69,80,103,132,13S,272-274 The reaction is equally adaptable to procedures using other digestion reagents such as sulphuric acid - hydrogen perox-ides1 or perchloric a~id.80~84~275-277 Perchloric acid digestion is rapid although quantitative conversion of all nitrogen in some biological materials may not be achieved.gO (ii) CFA methods.Automated colour development has been used on sulphuric acid - perchloric acid digests,278 Kjeldahl digests279J80 and sulphuric acid - phosphoric acid -hydrogen peroxide digestion mixtures.281 A method using automated continuous digestion before colour development has also been described? Food Fish Burnett283 showed a good correlation between the concentra-tion of ammonia in crabmeat and its degree of decomposition, and used the thymol - hypobromite reaction with solvent extraction of the indothymol to measure the ammonia concentration.This method was further refined to improve sensitivity by Fernandez-Flores and Salwin147 and recommen-ded as an index of decomposition of crabmeat by Stein-brecher.149 Ruiter and Weseman284 described a CFA method for measuring ammonia in fish and shrimp and used dialysis to prevent interferences from amines. Meat Wasilewski et a1.285 used the nitroprusside-catalysed salicylate - hypochlorite reaction in a manual method to measure microgram amounts of ammonia in meat and found that normal concentrations of amino acids peptides and urea did not interfere in the determination. CFA methods for the measurement of total nitrogen after acid digestion have also been described.28"289 Milk products Behringer and Klostermeyer29" measured ammonia in cheese with the phenol - hypochlorite reaction after distillation from cheese slurry.CFA methods for measuring total nitrogen in milk after acid digestion have also been d e ~ c r i b e d . ~ ~ ' - ~ ~ ~ Sugar De Villiers et al.296 described manual and CFA methods for measuring ammonia in sugar. Kubadinow and R o e s r ~ e r ~ ~ ~ compared Kjeldahl digestion and titration with a CFA method for measuring total nitrogen in sugar and found good agreement between the methods. Brewing materials CFA methods for measuring total nitrogen in malt and barley have been described by Kremkow298 and Mitcheson and Stowe11.299 Moll et al.300 used a similar method to determine total nitrogen in beer and wort. Waters The main problems encountered in the application of the reaction to waters arise from the low levels of ammoni 562 ANALYST.MAY 1984 VOL. 109 present and the wide range of inorganic and organic com-pounds that if present in sufficient amounts can interfere with the reaction. Salinity differences affect reaction sensi-tivity by causing changes in solution pH and this is likely to be the main source of error in waters with high salt contents such as sea water. Sea water Manual methods. Krasnova301 used Thomas's method7 for measuring ammonia in sea water although it is not clear if the ammonia was separated before measurement. R i l e ~ 3 ~ and Riley and Sinhaseni43 used distillation and diffusion respec-tively to isolate and concentrate the ammonia before spectro-photometric determination. For the same reasons Newel1 and Dal Pont,lso Newel156 and Matsunaga and Nishimura148 applied the reaction directly to sea water in acid conditions prior to solvent extraction before rendering the solvent alkaline to measure the extracted indophenol.These methods are slow and complicated and the most recently published methods use the nitroprusside-catalysed reaction directly on a sea water sample with a buffering or chelating agent to stop precipitation of hydroxides and to minimise the effects of variations in salinity. Among the complexing agents used are 1,2-diaminocyclohexane-N,N,N' ,"-tetraacetic acid (CDTA),127 EDTA114 and citrate.134 S a g F noted that differ-ent calibration graphs were obtained with standards in sea water and distilled water and produced standards from nitrogen-free sea water to overcome the salinity error a method used also by Liddicoat et al.76 and Dal Pont et al.134 SagiYO also noted that salts of calcium and magnesium appeared to be the cause of salt error and that sodium chloride was not involved.Emmet87 observed the depressive effect of increasing salinity and he attributed this to bromide ion interference which decreases the effective hypochlorite con-centration. Solorzanoy] found that with colour development at 22-27 "C for 1 h identical calibrations were obtained with standards in distilled water and sea water. This was confirmed by Zadorojny et ~ 1 . 1 ~ ~ and the method has been used subsequently by McCarthy and Kamyko~ski3~2 and Degob-bisl06 and for the measurement of urea in sea water after urease pre-treatment .303 Solorzanoy1 also found interference from sulphide and recommended removal of this interferent by acidification and aeration prior to the addition of reagents.KorolefP8 used a similar method and recommended colour development for 6 h at room temperature in the dark. Koroleffge found that the reaction was complete in 2 h in distilled water but took at least 6 h in sea water. The "salt effect" was directly related to differences in reaction pH and precipitates formed with sea water as no complexing agents were used. No interferences from a range of organic nitrogen compounds at their normal level of occurrence were encoun-tered but hydrogen sulphide interfered. Nimura118 applied the reaction after adding a large excess of solid potassium carbonate to render the salt content of all samples nearly equal the indophenol being measured after precipitation of hydroxides.Matsunaga and Nishimura148 used the thymol - chloramine-T reaction and measured low ammonia concentrations by extracting the indothymol into hexanol. Liddicoat et ~ 1 . ~ 6 found that the use of the nitroprusside-catalysed reaction resulted in high and erratic blanks which they overcame by using hexacyanoferrate( 11) in the presence of ultraviolet light. Gravitz and Gleye142 found that a photochemical side reaction involving the nitroprusside anion interfered in S o l o r z a n o ' ~ ~ ~ method and recommended that the reagents be protected from light exposure. Hamp-son77 found that nitrite caused suppression of colour develop-ment particularly if present in higher concentrations than ammonia.Amines and related organic nitrogen compounds also interfered but the levels at which interference occurred was not reported. Verdouw et al. 66 used salicylate hexacyanoferrate(I1) and ultraviolet light to measure the ammonia in sea water and listed a table of inorganic interferences. These workers concluded that salicylate was a good alternative to phenol because it is non-toxic and the method is reproducible. The lower sensitivity of the reaction in sea water was attributed to a pH effect that was overcome by increasing the final sodium hydroxide concentration. Bower and Holm-Hansen1Q fol-lowed up this work but found the method to be susceptible to variation in salinity and that the reaction when salicylate was used as the phenol decreased in sensitivity after prolonged exposure to sunlight.CFA methods. Head304 modified Solorzano'syl method and found that there was minimum salinity effect and negligible responses from urea and amino acids added to synthetic sea water indicating that these compounds did not hydrolyse under the reaction conditions. Grasshoff and Johannsen 126 concluded that salinity errors were inevitable even with the use of buffer solutions. These workers found interferences from hydrogen sulphide if present in concentrations above 50 pg 1-I and from amino acids at the heating temperature of 70 "C used to accelerate colour development although errors from these sources were not considered critical at the normal levels of these substances in sea water.Benesch and Mangels-dorf305 described a method that they claimed was applicable over a wide salinity range and free from interference from most non-ammonium nitrogen compounds occurring naturally in sea water. This method was subsequently investigated by Grasshoff and Johannsen14" who showed that the method was not free from salinity interferences because the sensitivity decreased with increasing magnesium chloride concentration. The disagreement about the effect of salt concentrations arose because Benesch and Mangelsdorf305 only tested their method over a range of sodium chloride concentrations and colour formation is largely unaffected by this salt.90 Grasshoff and J ~ h a n n s e n ' ~ ~ suggested that the strong inhibition of the reaction by magnesium could be caused by the blocking of the phenol used (salicylate) by magnesium and by the shift in reaction pH caused by the formation of MgOH+ and Mg( OH)2.Slawyk and MacTsaac306 used a CFA version of Koroleff's98 method for shipboard ammonia analysis and concluded that it was reliable for natural sea waters. Reusch Berg and Abd~llah15~ used a reaction temperature of 40 "C and reported no hydrolysis of amino acids at this temperature. The method was reported to show little or no salt effect although the composition of the salinity standards was not stated. Adamski1s9 used the automated method of Grasshoff and Johannsenl26 to measure total nitrogen in sea water after a manual persulphate digestion step and obtained good recoveries of nitrogen from spiked marine samples.Helder and De Vriesl23 adapted Solorzano's91 method to CFA and claimed that the method was applicable over a wide salinity range. These workers also noted that in CFA salinity differences can cause differences in the refractive indices of samples which can lead to erratic absorbance measurements in the flow cell of the spectrophotometer. This phenomenon is attributed to the curvature of the ends of the flow cell and had been previously noted by Froelich and Pils0n.1~~ Helder and De Vries'23 found that small interferences from amino acids and urea could be tolerated but in contrast to the findings of Grasshoff and Johannsen,126 they found that the presence of sulphide even in low concentrations leads to an over-estimation of ammonia.Nitrite caused an over-estimation of ammonia but this was not considered a problem when analysing natural sea water. Non -salin e water Manual methods. Datsko and Kaplan47 used an uncatalysed phenol - hypobromite reaction for the analysis of waters and found maximum colour development was obtained in 50 min at room temperature. The use of hypobromite increased th ANALYST MAY 1984 VOL. 109 563 sensitivity and stability of the reaction compared with hypo-chlorite. Kaplin-307 modified the method to increase the detection limit by extracting the indophenol into chloroform prior to measurement. Rossum and Villarru~3~)~ used the manganese(I1)-catalysed reaction in preference to Nessler'sh method which was unsatisfactory in moderately hard waters.Tetlow and Wilsonl"3 used the acetone-catalysed reaction for the analysis of boiler feed water and found no appreciable interferences from the impurities normally present. Cock-ing3Og used Riley's33 diffusion method to measure ammonia excreted by fresh water fish. Sutcliffe and Jones"() compared the Nesslerh and the uncatalysed phenol - hypochlorite methods for sewage after steam distillation and showed that the indophenol method had higher sensitivity. Maunter,311 in a comparison of methods recommended the salicylate -hypochlorite method for the analysis of drinking waters because this was more sensitive than the phenol - hypochlorite method. Harwood and Kuhngg described a nitroprusside-catalysed phenol - hypochlorite method that was not sensitive to the timing of reagent addition and produced a stable colour after 20 min.These workers compared their manual method with a CFA adaption312 and obtained close agreement between the two methods although no mention was made of interference studies. Fujinuma et al. 146 improved the sensi-tivity of Roskam and de Langen'sl27 method for water analysis by extracting and concentrating the indophenol into isobu-tanol. ScheineP6 described a method in which a phosphate -citrate - EDTA buffer was used to minimise differences in sample pH and listed a table of interferences. This method produced maximum colour after 45 min although low values were obtained if the hypochlorite reagent was not added promptly after the phenol - nitroprusside reagent. Boo and Ma130 described a simple method designed for water laboratories without sophisticated equipment.The uncatalysed phenol - hypochlorite reaction was used and these reagents were added to the sample in boric acid before adding sodium hydroxide to render the solution alkaline. Under these conditions maximum colour was obtained in 6 min and was stable for 12 h. A distinct blue colour. easily detected by the naked eye is obtained at 2 pg ml-1 of ammonium nitrogen, which is considered to be the threshold concentration for many purposes. Interferences from a number of ions were noted (Fe2+ Fe3+ Ca2+ Cu2+ Mg2+) and a diffusion technique was recommended if these were present in large amounts. CFA methods. Britt313 adopted the method of Scheurer and Smith38 for the analysis of reactor waters and used a flow dialyser to separate possible interferents.Similar dialysis methods using the uncatalysed reaction have been reported for waters effluents and sewage that have suspended sol-ids.31*-317 Ellerker and Collinson318 found that dialysis was not necessary for fully treated sewage and used EDTA to prevent precipitation of hydroxides. Crowther and Evans319 described a nitroprusside-catalysed method for waters in which natural turbidity or colour is compensated for by a separate reference channel. The same w0rkers3~~ also des-cribed a method that involved automated distillation prior to spectrophotometric analysis as a means of eliminating errors from sample colour turbidity and sample matrix. A compari-son of the two methods gave a very good correlation, indicating that the automatic blank correction was not susceptible to most common interferences although the capacity of the reference stream to compensate for humates or tannates was limited.The automated distillation method was free from inorganic interferences and there was minimum hydrolysis of organic nitrogen compounds because of the short reaction time between the sample and hot alkali. A number of methods have been described for measuring Kjeldahl nitrogen in waters using automated continuous digestion. 136,138.158.321-324 The methods using the uncatalysed reaction are not sensitive enough to measure low nitrogen levels and Harwood and Huyser"2 described a system that could be used for high and low levels of ammonia depending on whether nitroprusside or acetone were used as the catalyst.A method using batch Kjeldahl digestion prior to CFA has been described325 and this method offers the advantage of more complete digestion than that achieved in continuous digestion systems. Metallurgy Metals such as steel contain very low amounts of nitrogen (usually <0.02%) and further dilution occurs during dissolu-tion of the sample so that adequate sensitivity of the spectrophotometric measurement can be a problem. Further problems can occur because the Berthelot reaction is suscept-ible to interference from a wide range of metallic ions such as manganese and chromium which are often present in metallurgical samples. Manual methods. Many methods have been described that apply the reaction directly to sample solutions following acid dissolution.These include the use of solvent extraction to increase sensitivity which has been used for the analysis of steels ,2633131,326 titanium carbide144 and aluminium. 13' Simi-larly methods that do not use catalysts or solvent extraction have been described for the analysis of stee1,33-115 beryllium,*9 uranium ,327,328 zirconium ,327 vanadium and titanium .328 To overcome the problems of interfering metal ions and also to concentrate the ammonia prior to spectrophotometric analysis steam distillation of the sample extract has been used for steels,329-332 molybdenum,333 vanadium,334 zirconium335 and tantalum alloys.336 None of these procedures are particu-larly suitable for the routine analysis of large numbers of samples and in only two examples3303 331 has the more sensitive nitroprusside-catalysed reaction been used.CFA methods. Manual steam distillation prior to CFA has been used for steelsl37.337,33* and uranium,339 and there is no indication that more fully automated systems such as the use of flow dialysis or auto-distillation have been applied to the analysis of metals. Petroleum The uncatalysed phenol - hypochlorite reaction has been used to measure total nitrogen in petroleum after Kjeldahl diges-tion and di~tillation.37~3"~341 Heistand342 described a method using automated continuous extraction and digestion coupled with CFA with the nitroprusside-catalysed reaction to measure nitrogen in petroleum. Good recovery of nitrogen from petroleum-based nitrogen compounds was obtained, although the analysis rate of 40 d-1 is relatively slow for an automated system.Gases Methods for the measurement of ammonia in air have been described by Kurczatova343 and Kifune et a1.344 These methods involve collection of ammonia in acid solution or a filter prior to elution and spectrophotometric measurement. Harrell et al.345 measured ammonia in tobacco smoke after collection of the gases in boric acid. These workers found that it was necessary to charcoal treat the acid extract to remove organic compounds derived from the smoke in order to prevent interference in the colorimetric reaction. Pharmaceutical Products The Berthelot reaction can be used for determining various drugs usually after hydrolysis of the compound to an intermediate form in the reaction sequence.M ~ r f i n 3 ~ h J ~ ~ described automated and manual methods for the determina-tion of phenacetin and paracetamol. The reaction of either compound with acidic hypochlorite formed p-quinone chlor 564 ANALYST MAY 1984 VOL. 109 imide which reacts with alkaline phenol to produce indo-phenol. Davis et af.348 modified the manual procedure to include an additional hydrolysis step to obtain a more complete reaction. Ellcock and Fogg349J50 used the nitro-prusside-catalysed reaction of sulphanilamide with alkaline hypochlorite and phenol to measure sulphanilamide and found that other sulphonamides did not interfere. Frings and Salloom351 measured acetaminophen in serum by hydrolysis to p-aminophenol and the subsequent reaction with ammonia and phenol to form indophenol.Conclusions It is difficult to make definitive statements about optimum reaction conditions of the Berthelot reaction because there are several combinations of conditions that are equally suitable for a particular application. It is possible however to make some generalisations which are listed below. Phenols The use of phenol as the phenolic compound is traditional and is probably used in many instances for this reason alone. In practice however sodium salicylate offers advantages over phenol because it is easier to prepare more stable in solution, less toxic and safer to handle. It is often claimed that the reaction is more sensitive when phenol is used but the fact that the sodium salicylate reaction requires different solution concentrations and higher pH for optimum colour develop-ment is often overlooked.Because of the practical advantages sodium salicylate is recommended although it may not be suitable for all applica-tions. For instance it has been reported that the reaction involving sodium salicylate is particularly susceptible to interferences from magnesium140 and light. 143 Hypochlorite Sources Organochlorine compounds particularly NaDTI' form more stable solutions of the hypochlorite ion than sodium hypo-chlorite and because this salt is readily available it offers a practical alternative. NaDTT may not be suitable for some applications as it reacts with amino acids and protein and these compounds interfere with the reaction by lowering the effective hypochlorite concentration.71 Catalysts The reaction using catalysts derived from nitroprusside is now almost exclusively used because of the increased reaction rate, sensitivity and stability of the indophenol produced.The exact nature of the catalysing cyanide complex can vary depending on the reaction conditions and either the nitroprusside ion or an aquopentacyanoferrate( 111) complex may be the active species. Nitroprusside may however give high and variable blanks and it is desirable to avoid these when very low levels of nitrogen are being determined. High blanks are caused by a light-induced reaction involving nitroprusside breakdown products that form an indophenol dye with a similar visible spectrum to that formed with ammonia.54.141 This problem can be avoided if aquopentacyanoferxate(II1) is used and this complex if formed under reaction conditions favourable to the hydrolysis of nitroprusside53 or by the action of ultraviolet light on potassium hexacyanferrate(I1) .77 PH The optimum reaction pH depends on the reagents catalysts and concentrations used and many optimum values have been suggested.However all steps in the reaction are pH depen-dent and control of reaction pH is essential. In practice this is achieved by the use of buffer solutions neutralised sample extracts or the use of small volumes so that differing sample acidities do not affect the reaction pH. The pH dependence of the reaction varies with the catalysing complex and the aquopentacyanoferrate(II1)-catalysed reaction is more sensi-tive to small changes in pH than the nitroprusside-catalysed reaction .54 Other Variables Reaction temperature time and the order of addition of reagents need to be carefully controlled and this is why the reaction has been most successfully used in automated chemistry systems such as CFA.High reaction temperatures (>50 "C) are unnecessary with the catalysed reaction and the use of higher temperatures may cause hydrolysis of organic nitrogen compounds. The order of addition of reagents does not appear to be particularly important for reaction sensitivity provided the reagents are added in quick succession. However the magnitude of interferences may depend on the order of reagent addition. 107 Indophenol dyes once formed, are stable for many hours provided they are not exposed to direct sunlight,98 or formed at high temperatures.*" Interferences Interferences from a large number of sources can occur, although most can be successfully eliminated by modifying the reaction conditions.Interference problems are usually only severe when there is a large ratio of interferent to nitrogen and most applications papers report techniques to deal with possible interferences. Summary The Berthelot reaction has found wide analytical use for the measurement of nitrogen and the literature relating to the theory and application of the reaction is extensive. The reaction mechanism is complex and reaction conditions such as pH temperature reagents reagent concentrations and order of addition all affect the sensitivity of the reaction.In addition there are a number of elements compounds and conditions that can interfere with the reaction. This review records all these factors along with current applications so that they can be taken into account by users and potential users of this sensitive and versatile reaction. The author is grateful to the library staff Soil Bureau DSIR, for their assistance with the literature search and to Dr. M. D. Krom Department of Geology and Geophysics Yale Uni-versity USA for permission to reproduce Fig. 7. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Berthelot M. P. E Rep. Chim. Appl. 1859 1. 284. Lex R. Ber. Chem. Ges. 1870 3 457. Cotton S .Bull. SOC. Chim. Fr 1874 21. 8. Engel M. R. Bull. SOC. Chim. Fr. 1875. 23 435. Tarugi N . and Lenci F. Boll. Chim. Farm 191 I 50 907 Nessler J. Chem. Center. 1856 27 529. Thomas P . Bull. SOC. Chim. Fr. 1912 11 796. Thomas P. Bull. SOC. Chim. Fr. 1913 13 398. Foxwell G. E. Gas World 1916 33 10. Ayers S. H. Rupp P. and Mudge C. S . . J. Infect. Dis. 192 29 235. Hucker G. J. and Wall. W. A . J. Bucteriol. 1922 7 515. Orr A. P. Biochem. J 1924 18 806. Butcher B. H J. Absfr. Bacferiol. 1924 8 32. Murray M. M. Biochem. J . . 1925 19 294. Hansen P. A . J. Bacteriol 1930 19 223. Furth O. Friedrich A. and Scholl. R. Biochem. Z 193 240 50 ANALYST MAY 1984 VOL. 109 565 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.32. 33. 34. 35. 36. 37. 38. 39. 40. 41, 42. 43. 44. 45 * 46. 47. 48. 49. SO. 51. 52. 53 * 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. Van Slyke. P. D. and Hiller A J . Biol. Chem. 1933 102, 499. Harrow B. Chamelin I. M and Wagreich H. Science 1933, 78 54. Lapin L. N. and Hein W. Z. Anal. Chem 1934 98 236. Borsook H. J . Biol. Chem. 1935 110 481. Polonovski M. and Boulanger P. Bull. SOC. Chim. Biol., 1935 17 944. Wittermans A. W.? Ned. Tijdschr. Gerteeskd. 1935. 79 311. Furth O. and Gotzl F. Biochem. Z . 1936 283 358. Crismer M. R. Bull. SOC. Chim. Biol. 1937 19 1000. Hinsberg K and Mucke K. Biochem. Z. 1938 297 332. Veinberg G . J . Zavod. Lab. 1938 7 1251. Veinberg.G . J. Zavod. Lab. 1940. 9 1073. Hansen P. A . and Nietsen. V. J . Biol. Chem. 1939,131,309. Alten F. and Haupt W. Bodenkd. Pflanzenernaehr 1941, 25 140. Russell J. A . J . Biol. Chem. 1944. 156 457. Vazhenin 1. G . Pochvovedenie. 1949 359. Casas J. R. and Dominguez. R. An. R. SOC. ESP. Fis. Quim. Ser. B . 1949 45 1607. Kakita Y. Nippon Kinzoku Gakkaishi 1949 13 39. Riley J. P. Anal. Chim. Acta 1953 9. 575. Lapin L. N. Tr. Kom. Anal. Khim. Akad. Nauk SSSR. 1954, 5 77. Lubochinsky B. and Zalta J . P. Bull. SOC. Chim. Biol 1954. 36 1363. Noble E . D . Anal. Chem. 1955 27 1413. Scheurer P. G . and Smith F. Anal. Chem 1955 27 1616. Exley D. Biochem. J . 1956 63. 496. Crowther A. B. and Large R. S. Analyyt lb.56 81 64. Takei S. and Kato T. Technol.Rep. Tokohu Imp. Univ., 1955 20 71. Brown R. H. Duda G. D. Korkes. S . and Handler P Arch. Biochem. Biophys. 1957 66 301. Riley J. P . and Sinhaseni P . J . Mar. Riol. Assoc. U. K . 1957, 36 161. Lapin L. N . Zamanov K. Kh and Makarova V. P., Pochvovedenie 1957 95. Fleury M. P. and Eberhard R . Ann. Pharm. Fr. 1958,743, 465, Bohnstedt U. Fresenius Z . Anal. Chem. 1958. 163 415. Datsko V. G . and Kaplin V. T. Gidrokhim. Muter. 1959, 29. 230. Gibbs H . D . J . Biol. Chem 1927 72. 649. Soloway S . . and Santoro A . Anal. Chem 1955. 27 798. Rodd E. H. “Chemistry of Carbon Compounds,” Volume 111. Elsevier Amsterdam 1956 p. 721. Bolleter W. T. Bushman C. J . . and Tidwell P. W. Anal. Chem. 1961 33. 592. Dambacher M. Gubler A . and Haas. H. G . Clin.Chem., 1968 14 615. Patton C. J. and Crouch S. R Anal. Chem. 1977 49 464. Krom. M. D . Analyst 1980 105 305. Bottcher C. J. F. Van Gent. C. M. and Pries C Recl. Trav. Chim. Pays-Bas 1961 80. 1157. Newell B. S . J . Mar. Biol. Assoc. U. K . 1967 47 271. Gehrke C . W. Kaiser F. E. and Ussary. J . P. J . Assoc. Off. Agric. Chem. 1968 51 200. Karayannis M. I . . Chim. Chron 1975 4 15. Lange K. Schliemann. J Friebe. R and Linow. F., Nuhrung. 1980 24 457. Morita. Y. and Kogure. Y Nippon Kagiiku Kaishi. 1963 84, 816. Yamaguchi R. and Machida. T. Yakugaku Zasshi 1969 89, 804. Yamaguchi. R and Machida T. Yiikugaku Zasshi. 1968,88, 1383. Yamaguchi R . Suzuki. M. and Hiroshima H Chem. Pharm. Bull 1970 18 1866. Morita. Y and Kogurc. Y. Nippon Kagaku Kaishi 1970 91, 653.Yamaguchi R . Machida T and Ueki M. Yakagaku Zusshi, 1969 89. 1534. Verdouw H. Van Echteld. C . J. .4. and Dekkers E. M. J., J . Water Res 1978. 12 399. Glebko. L. I. Ulkina. Zh. I and Lognenko E. M., Mikrochim. Acfa. 1975 11. 641. Haussler A and Hajdu. P. Klin. Wochenschr. 1960 17 874. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96 97 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. Stegemann. H. arid Loescheke V. Hoppe-Seyler‘s Z. Physiol. Chem. 1962. 329 241. Reardon J. E . Foreman. J .R. and Searcy. R. L. Clin. Chim. Acta 1966 14 403. Seeiy J. H. Petitclerc J . C. and Benoiton. L. Clin. Chim. Acta 1967. 18. 85. Fenton J . C. B. Clin. Chim. Acta 1962 7 163. Horn. D. B. and Squire C . R . Clin. Chim. Acta 1967,17,99. Harwood J. E and Huyser D . J . Wuter Res 1970 4 501. Weichselbaum T. E Hagerty J . C. and Mark €-I. B Anal. Chem. 1969 41 848. Liddicoat M. I Tibbitts S . and Butler E . I . . Limnol. Oceanogr. 1975 20 131. Hampson B. L. Water Res. ~ 1977 11 305. Aspenger S . Murati. I. Pavlovic D. J J . Chem. Soc. 1960, 730. Horn D. B. and Squire. C. R. Clin. Chim. Acfa. 1966. 14, 185. Jones G . R. N Lab. Pract. 1967 16 1486. Beecher G. R. and Whitten G . K. Anal. Biochem 1970,36, 243. Kurahasi K . Ishihara A. and Uehara H. Clin.Chim. Acfa, 1972 42 141. Mitchell H. L. J . Assoc. Of5 Anal. Chem. 1972 55 1. Jaenicke L. Anal. Biochem. 1974 61 623. Hinds A. A. and Lowe L. E. Commun. Soil Sci. Plant Anal., 1980 11 469. Rommers P. J. and Visser J. Analyst 1969. 94 653. Emmet R . T. Anal. Chem. 1969 41 1648. Bohley P. Hoppe-Seyler’s 2. Physiol. Chem 1967,348. 100. Bundy J. K. and Goode G. C. Anal. Chim. Acta 1967 37, 394. Sagi T. Oceanogr. Mug 1966 18 45. Solorzano L. Limnol. Oceanogr. 1969 14 799. Pym R. V. E. and Milham. P. J.,Anal. Chem 1976,48,1413. Lange R. Friebe R . and Linow F. Nahrung. 1980,24,907. Searcy R. L. Simms N. M. Foreman J. A . and Bergquist, L. M Clin. Chim. Acta 1965 12 170. Weatherburn M. W. Anal. Chem. 1967 39 971. Scheiner D . Water Res 1967 10 31. Kempers A.J. Geoderma 1974 12 201. Koroleff F. Int. Counc. Explor. Sea C. M. Hydrography Committee 1969 C:9. Harwood J. E. and Kuhn A. L. Water Res 1970 4. 805. Fraser A . R. and Russell J. D. Clay Miner. 1969 8 229. Lange R. Friebe R. and Linow. F. Nahrung 1980.25.447, Muftic M. K. Nature (London) 1964 201. 622. Mann L. T. Anal. Chem 1963. 35 2179. Tetlow J . A. and Wilson A. L. Analyst 1964 89. 453. Fawcett. J . K. and Scott J. E. J . Clin. Pathol. 1960. 13. 156. Degobbis D . Limnol. Oceanogr. 1973 18 146. Wearne J. T. Anal. Chem. 1963. 35. 327. Liddicoat M. I . Tibbitts. S . and Butler E. I Water Res., 1976 10 567. Hashitani H. and Yoshida H. Bunseki Kagaku 1968 17, 1136. Schach H. G. Limnologica. 1966 4 431. Okuda H Fujii S . and Kawashima. Y Tokushima J .Exp. Med. 1965 12 11. Domnas A . J . Biochem. 1961 50 46. Gipps. C. H . Reitsema A . and Wibbens-Albirts M. Clin. Chim. Acta 1970 29 501. Manabe T. Nippon Suisan Gakkaishi 1969. 35 897. Matsunaga K. Sugino K and Nishimura M. Runseki Kagaku. 1970 19 1255. O’Donovan D. J. Clin. Chim. Acta. 1971 32. 59. Zadorojny C. Saxton. S. and Finger R J . Water Pollut. Control Fed 1973. 45 905. Nimura. Y. Nippon Suisan Gukkaishi. 1973 39 1315. Forgan-Smith. J . . Slaughter B. D. and Cross R. B Clin. Chim. Acta 1976 69 139. Ngo. T. T. Phan A. P. H. Yam. C. F. and Lenhoff H. M Anal. Chem. 1982. 54 46. Whitaker J. F. Stephens A . . and Naftalin A . Clin. Chirn. A m 1965 12 466. Forman D. T Clin. C’him. Acfa 1964. 10. 497. Helder W De Vries. R. T. P Neth.J . Sea Res 1970. 13, 1.54. Konitzer K. and Voight. S . Clin. Chim. Acta. 1963. 8. 5 566 ANALYST MAY 1984. VOL. 109 125. Rahiala E. L. and Kekomaki M. P. Clin. Chim. Acta 1970, 30 761. 126. Grasshoff K. and Johannsen H. J. Cons. Int. Explor. Mer., 1972 34. 516. 127. Roskam R. Th. and de Langen P. Anal. Chim. Acta 1964, 30 56. 128. Dabin D. Cah. ORSTOM Ser. Pedol. 1965 3 338. 129. Crooke W. M. and Simpson W. E. J. Sci. FoodAgric. 1971, 22,9. 130. Boo H. and Ma T. S . Mikrochim. Acta 1976,II 515. 131. Namiki M. Kakita Y. and Goto H. Talanta 1964,11,813. 132. Shahinian A. H. and Reinhold J. G. Clin. Chem. 1971 17, 1077. 133. Henzell E. F. Vallis I. and Lindquist J. E. Trans. Znt. Congr. Soil Sci. 9th 1968. 3 513. 134. Dal Pont G. Hogan M. and Newell B.CSIRO Division of Fisheries and Oceanography Melbourne 1974 Rep. No. 55. 135. Lauber K. Clin. Chim. Acta 1976 67 107. 136. Kramme D. G. Griffin R. H. Hartford C. G. andcorrado, J. A. Anal. Chem. 1973 45 405. 137. Frazer J. W. Jones G. D. Lim R. Waggoner M. C. and Rodgers L. B. Anal. Chem. 1969 41 1485. 138. Harwood J. E. and Huyser D. J. Water Res. 1970 4 539. 139. Davidson J. Mathieson J. and Boyne A. W. Analyst 1970, 95 181. 140. Grasshoff K . and Johannsen H. J. Cons. Int. Explor. Mer., 1974 36 90. 141. Froelich P. N. and Pilson M. E. Q. Water Res. 1978 12, 599. 142. Gravitz N. and Gleye L. Limnol. Oceanogr. 1975,20,1015. 143. Bower C. E. and Holm-Hansen T. Can. J. Fish Aquat. Sci. 1980 37 794. 144. Klibus A. Kh. and Nazarchuk Z. Zh. Anal.Khim. 1961,16, 79. 145. Kulenok M. I. Izv. Vyssh. Uchebn. Zaved. Khim. Khim. Technol. 1961 4 558. 146. Fujinurna H. Shimada Y . and Hirano S . Bunseki Kagaku, 1971 20 131. 147. Fernandez-Flores E. and Salwin H. J. Assoc. Off. Agric. Chem. 1968 51 1109. 148. Matsunaga K. and Nishimura M. Bunseki Kagaku 1971,20, 993. 149. Steinbrecher K. J . Assoc. Off. Anal. Chem. 1973 56 598. 150. Newell B. S. and Dal Pont G. Nature (London} 1964,201, 36. 151. Haslemore R. M. and Roughan P. G. J. Sci. Food Agric., 1976 27 1171. 152. Betteridge D. Anal. Chem. 1978. 50 832A. 153. Holz F. and Kremers H. Lundwirtsch. Forsch. Sonderh., 1971 26 177. 154. Reusch Berg B. and Abdullah M. I. Water Res. 1977 11, 637. 155. Connetta A. Buccafuri A and Jansen J. Am. Lab.1976, 8 103. 156. Logsdon E. E. Ann. N.Y. Acad. Sci. 1960,87 801. 157. Skjernstad J. O. and Reeve R. Commun. Soil Sci. Plant Anal. 1976 7 229. 158. Elkei O. Anal. Chim. Actu 1976 86 63. 159. Adamski J. M. Anal. Chem. 1976 48 1194. 160. Searle P. L. N.Z. J. Agric. Res. 1975 18 183. 161. Selmei-Olsen A. R. Analyst 1971 96 565. 162. Wall L. L. and Gehrke C. W. J. Assoc. Off. Anal. Chem., 1979 62 503. 163. Keay J. and Menage P. M. A. Analyst 1969 94 895. 164. Aw S . Analyst 1981 106 114. 165. Docherty A. C. Lab. Pract 1969. 18 47. 166. Wrightman K. Adv. Autom. Anal. Technicon Int. Congr., 1967 2 61. 167. King G. H. and Scobie A. G. Chem. Process Eng. 1970,51, 62. 168. Nebel S . “Technicon Italy Symposium,” April 1971 p. 173. 169. Ridley. A. G. Lab.Equip. Dig. 1976 14 79. 170. Viswanathan T. S. Nagarajan M. and Haribabu B. J. Assoc. Off. Anal. Chem. 1980 63 1248. 171. Moore H. P. and Kelly R. E. Adv. Autom. Anal. lechnicon Int. Congr. 1969 2 47. 172. Gehrke C. W. Wall L. L. Killingley J. S . and Inada K. J. Assoc. Off. Agric Chem 1971 54 651. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. Gehrke. C. W. Killingley J. S . and Wall L. L. J. Assoc. Off. Agric. Chem. 1972 55. 467. Wall L. L. and Gehrke C. W J . Assoc. Off. Anal.Chem 1977 60 881. Law A. R. Nicolson N. J. and Norton R. L. J . Assoc. Ofl. Anal. Chem. 1971 54 764. Gehrke C. W. Wall L. L. and Absheer J . S. J. Assoc. Off. Agric. Chem 1973 56. 1096. Wall L. L. and Gehrke C. W. J. Assoc. Off. Anal. Chem 1975 58 1221. Basson W. D Lab. Pract. 1976 25 763. Shkonde E. I. and Koroleva I. E. Pochvovedenie 1964,54. Kudeyarov V. N. Agrokhimiya 1970 9 127. Kudeyarov V. N. Khim. Sel’sk. Khoz. 1970 8 708. Samokhvalov S. G. Prizhukova V. G. Privalenkov S. Ya., and Yurlova 0. V. Khim. Sel’sk. Khoz. 1974. 12 622. Burdin S . and Fortier M. Agron. Trop. (Paris) 1971 25, 392. Burdin S . and Egoumenides C. Agron. Trop. (Paris) 1973, 28 1193. Robertson P. D. Annu. Rep. Agric. Res. Council Letcombe Lab. 1974 73. McLeod S . and Zarcinas B.Commun. Soil Sci. Plant Anal., 1976 7 743. Samokhvolov S. G. and Prizhukova V. G. Khim. Sel’sk. Khoz. 1975 13 753. Schuman G. E. Stanley M. A. and Knudsen D. Soil Sci. Am. Proc. 1973 37 480. Cadahia C. An. Edafol. Agrobiol. 1973 32 479. Saas A. Nucl. Sci. Abstr. 1975 31 19140. Petts K. W. and Belcher M Water Pollut. Control 1980,79, 399. Brown M. W. J. Sci. FoodAgric. 1973 24 1119. Keay J. and Menage P. M. A. Analyst 1970,95 379. Hannawalt R. B. and Steckel J. E. Autom. Anal. Chem., Technicon Symp. 1966 1 133. Hofmann G. and Teicher K. Z . Pflanzenerniihr. Ding. Bodenkd. 1961 95 55. Searle P. L. and Speir T. W. Commun. Soil Sci. Plant Anal., 1976 7 365. Novozamsky I. van Eck R. van Schouwenburg J. Ch. and Walinga I. Neth. J. Agric.Sci. 1974 22 3. Burrows J. R. Proc. Annu. Congr. S. A f r . Sugar Technol. Assoc. 1977 51 68. Smith V. R. Commun. Soil Sci. Plant Anal. 1980 11 709. Samokhvalov S. G. Prizhukova V. G. and Kondrat’er, M. N. Khim. Sel’sk. Khoz. 1976 14 71. Felker P. Anal. Chem. 1977 49 1080. Havilah E. J. Wallis D. M. Morris R. and Woolnough J. A. Lab. Pract. 1977 26 545. Nkonge C. and Ballance G. M. J. Agric. Food. Chem., 1982,30 416. Warner M. H. and Benton Jones J. Commun. Soil Sci. Plant Anal. 1970 1 109. Uhl D. E. Lancaster E. B. and Voinovich C.,Anal. Chem., 1971 43 990. Ayers C. W. Proc. SOC. Anal. Chem. 1971 8 17. Eastin E. F. Commun. Soil Sci. Plant Anal. 1978 9 107. Ferrari A . Ann. N.Y. Acad. Sci. 1960 87 792. Siriwardene J. A. de S . Thomas A. J. Evans.R. A and Axford R. F. E. J. Sci. Food. Agric. 1966 17 456. Varley J. A. Analyst 1966. 91 119. Williams C. H. and Twine J. R. CSIRO Division of Plant Industry Sydney 1967 Technical Paper No. 24. O’Neill J. V. and Webb R. A. J. Sci. Food Agric. 1970 21, 217. Benintende M. “Technicon Italy Symposium,” April 1971, p. 205. Faithfull N. T. Lab. Pract. 1971 20. 41. Kirneu B. S . and Aduayi E. A. East Afr. Agric. For. J . , 1972 38 207. Bietz J. A, Anal. Chem. 1974 46 1617. Kaplan A. Methods Biochem. Anal. 1969 17 311. Ternberg J. L. and Hershey F. BJ. Lab. Clin. Med 1960, 5 766. Weller H. Roentgen. Laboratoriumsprax. 1962 15 L142. Gangolli S. and Nicholson T. F. Clin. Chim. Acta 1966 14, 585 ANALYST MAY 1984 VOL. 109 221. 222. 223. 224. 225.226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 235. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. Miller. G. E and Rice J . D. Am. J . Clin. Pathol. 1963 39, 97 Van Nunen. E. M. M. A . Chem. Weekbl. 1966 62 64. Van Oudheusden A. P. M Phurni. Weekbl. Ned. 1966. 101, 385. Kingsley G. R . and Tager H. S. Stand. Methods. Clin. Chenz. 1970 6 115. Wenzel. E. Oppolzer R . and Schuster D . Mikrochim. Acta, 1971 680. Hegedus-Wain I. and Szam. I. Med. Welt. 1973. 24 1205. Vujovic A . Rass. tnt. Clin. Ter. 1974 54 940. Muller-Beissenhirtz W and Keller H.Klin. Wochenschr., 1965. 43 43. Keller H . Muller-Beissenhirtz W. and Neumann E. Klin. Wochenschr. 1967 45. 314. McCullough H Clin. Chim. Acta 1967 17 297. Vigneron C. Roos F. Dailloux M. and Regnie M.-A., Pathol. Biol. 1970 18 217. Pre J . and Burnichon J Pathol. Biol. 1971 19 103. Leffler H. H . Am. J . Clin. Pathol. 1967 48 233. Hofmann M. 2. Med. Labortech. 1969 10 86. Dropsy G and Boy J. Ann. Biol. Clin. 1961 M. 313. Manston R. Biochem. Med. 1970 4. 486. Boy J. Manceaux J. C. and Choisy H. Am. Biol. Clin., 1970 28 441. Imler. M. Frick A. Stahl A. Peter B. and Stahl J. Clin. Chim. Acta 1972 37 245. Rashkovan B. A . Sb. Nauchn. Rab. Vitebsk. Med. tnst. 1957, 8 108. Searcy R. L. Gough G . S . Korotzer J. L. and Bergquist, L. M. Am. J. Med.Technol. 1961 27 225. Searcy R . L. and Gough G. S . J . Med. Lab. Technol. 1962, 19 272. Watson D . and Jonnason I. Med. J . Aust. 1963 2 227. Macfarlane T. W. Proc. Assoc. Clin. Biochem. 1964 3 21. Cirje M. and Sandru D. Viata Med. 1965 12 1617. Naftalin L. Whitaker. J . F. and Stephens A. Clin. Chim. Acta 1966 14. 771. Searcy R . L. Foreman J . A . Ketz A. and Reardon J . E., Am. J . Clin. Pathol. 1967 47 677. Richterich R . and Kueffer H. 2. Klin. Chem. Klin. Biochem. 1973 11 553. Wilson. B. W. Clin. Chem. 1966 12 360. Wilcox A. A. Carroll. W. E. Sterling R. E Davis H. A , and Ware A. G. Clin. Chem. 1966 12 151. Nikochev G . Stoichiev St and Tikholova S. Vutr. BoleJ., 1970. 9 73. Cunarro J . A. and Weiner M. W. Kidney Int. 1974,5 303. Jacobs P.Pharm. Weekbl. 1963 98. 505. Dobbs C. R . Castleberry H. B. and Shaw E. G. Clin. Biochem. 1969 2 381. Dieu. J. P. Clin. Chim. Acta 1970 28 209. Burnette M. A . Markley R. A . and Babcock M. J. Nutr. Rep. Irit.- 1976 14 125. Berry H. K. Univ. Texus Puhl. 1951 No. 5109 88. Mathies. J. C Clin. Chem. 1964 10 346. Liemann F. 2. Gesamte Exp. Med. 1964 138 191. Van der Heiden D . A. van Triet. A. J . and Faber A . , Pharm. Weekhl. Ned. 1968. 103 793. Bailenger J . . Dubourg-Casamayou. L. and Merle J . Feuill. Biol 1974. IS. 57. Ayoub W . Castel. J . Jean R. and Huguet. R Trav. SOC. Pharm. Montpellier 1976. 36. 301 . Komoto. M. Nippmi Nogei Kagaku Karshi 1962. 36. 310. Schweiger H. G. Rapoport S. M. Schozel E. and Richer. K. Vopr. Med. Khim. 1957 3 65.Walker R. I. and Shipman W. H. J. Chromatogr. 1970. 50, 157. Setsuro F . and Okuda H Tokushima J. Exp. Med. 1975. 22 1. Konitzer. K. Korner I and Voigt S . Fresenius Z. Anal. Chem. 1966. 233 426. Ngo T. T. Int. J . Biochem. 1975 6 663. Mor J. R. Zimnierli A . and Fiechter A. Anal. Biochem., 1973 52 614. Cuny L. J. Pharm. Chim. 1926 3. 150. Glebko L. I . . Ulkina. J. I. and Vas’kovsku. V. E . Anal. Biochem. 1967 20 16. Nahagawa Y Acta Med. Univ. Kugoshimu 1970. 12. 23. 272. 273. 274. 275, 276. 277. 278. 279. 280, 281. 282. 283. 284, 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. 313.314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 567 Weller. H. Roentgen Laboratoriumsprax. 1962. 15 L77. Khramov V. A . Lab. Delo 1965 7. 396. Sylvester-Bradley R. and Howitt S. G . J. Sci. Food Agric., 1977 28. 312. Sloane-Stanley G. H . and Jones G. R . N. Riochem. J. 1963, 86. 16. Sloane-Stanley G. H. Biochem. J . 1967. 104. 293. Meijers C. A. M. and Rutten J. C. J . M. Clin. Chim. Acta, 1969 24 308. Hofstader R . A . Microchem. I. 1966 10 444. Peaston M. J. T. Wilson S. and Pemberton J . Clin. Chem., 1968 14 871. Lange R. Friebe R. Linow F. and Schaffner B. Nahrung, 1979 23 549. Tingvall P. Analyst 1978 103 406. Jacobs S . C. J . Clin. Pathol. 1968 21 218. Burnett J. L. J . Assoc. Off. Agric. Chem. 1965,48 624. Ruiter A .and Weseman J. M. J . Food Technol. 1976. 11, 59. Wasilewski S . Slowik M. and Mroczek M. Przem. Spozyw. 1968 22 272. Schmidhofer T. Ejli H . K. and Weber R. Technicon, England 1973 Paper No. TI. Weber R. Technicon International Division Geneva 1973, Technical Report No. 6. Gantenbein W. M. Schermerhorn J. L. and George E. J. Assoc. Off. Agric. Chem. 1974 57 838. Gantenbein W. M. J . Assoc. Off. Anal. Chem. 1973,56,31. Behringer G . and Klostermeyer H. Z. Lebensm. Unters. Forsch. 1981 72 362. Schafer H. W. and Olson. N. F. Anal. Chem. 1975,47,505. Brisson G. J. Autom. Anal. Chem. Technicon Symp., Technicon New York 1965 p. 54. Heeschen W. “Technicon England Meeting,” Nov. 1970. Hindle E. J. and Wheelock J. V. Proc. Nutr. SOC. 1970,29, 89. Kuwata T .“Technicon Japan Symposium,” August 1971 De Villiers P. Detavernier R. and Groult M. Sucr. Fr., 1977 118 66. Kubadinow N. and Roesner G. Sucr. Belge 1977 96 9. Kremkow C . Monatsschr. Brau. 1970 23 36. Mitcheson R. C. and Stowell I<. C. J . Inst. Brew. 1970,76, 335. Moll M. Flayeux R. and Le huede. J. M. tnd. Aliment. Agric. 1975 92 631. Krasnova V. S . J. Appl. Chem. USSR 1936 9 2068. McCarthy J. J. and Kamykowski P. Fish Bull. 1972 70, 1261. McCarthy J. J . Limnol. Oceanogr. 1970 15 309. Head P. C. Deep Sea Res. 1971 18 531. Benesch R . and Mangelsdorf P. Helgol. Wiss. Meeresunt-ers. 1972 23 365. Slawyk G. and MacIsaac J. J. Deep Sea Res. 1972 19,521. Kaplin V. T. Gidrokhim. Muter. 1961 32 148. Rossum J. R. and Villarruz P. A . J .Am. Water Works ASSOC. 1963 55 657. Cocking A . W. J . Fish. Res. Board Can. 1967 24 1419. Sutcliffe R . A . and Jones G. A . Water Pollut. Control 1968, 67 209. Maunter G. Khig. Zdraveopaz. 1970 13 319. Harwood J. E. and Huyser D. J. Water Res. 1970 4 695. Britt R. D . Anal. Chem. 1962 34 1728. Chapman. B. Cooke G . H. and Whitehead R . Water Pollut. Control 1967 66 280. Millar A . S. Effluent Water Treat. J . 1967 7 468. Williamson T. and Millar A. S . Water Treat. Exam. 1969. 18,150. O’Brien J. E. and Fiore J. Wastes Eng. 1962 33 352. Ellerker R. and Collinson. B. Water Pollut. Control 1972, 71 540. Crowther J . and Evans J . Analvst 1980 105 849. Crowther. J . and Evans. J. Analyst 1980 105 841. Tenny. A . M Autom. Anal. Chem Technicon Symp. 1966, 1 580.McDaniel W. H. Hemphill R. N . and Donaldson W. T . , Autom. Anal. C‘hem. Technicon Symp 1967. 1. 363. Kammerer. P. A. Rodel M. G . Hughes R. A. and Lee. G . F. Eniiiron. Sri. Technol. 1967. 1. 340. I p. 375 568 ANALYST. MAY 1984 VOL. 109 324. 325. 326. 327. 328. 329. 330. 331. 332. 333. 334. 335. 336. 337 338. 339. 340. 341. 342. 343. 344. 345. 346. 347. 348. 349. 350. 351. 352. 353. 354. 355. 356. Canelli E. Water Air Soil Pollut. 1976 5. 339. Crowther J . Wright B. and Wright W. Anal. Chim. Acta, 1980 119 313. Goto H. Kakita Y. and Atsuya. I. Bunseki Kagaku 1963, 12 727. Hashitani H. and Yoshida H. Bunseki Kagaku 1970 19, 1564. Hashitani H. Yoshida H. and Adachi T. Anal.Chim. Acta, 1975 76 85. Calmettes J. and Drain J. Rev. Metall. 1956 LIII 682. Bach B. B. Willis R. R. and Reid R. B. Metallurgia 1966, 74 191. Jenkins R. H. Br. Steel Corp. Open Rep. 1972 MGiCCI520I 72. Petrov P. Tr. Nauchno-Issled. Inst. Cherna Met. 1972,6,295. UKAEA Industrial Group Chem. Services Dept. U. K.At. Energy Auth. Rep. 1958 IGO-AMIS-43. UKAEA Industrial Group Chem. Services Dept. U.K.At. Energy Auth. Rep. 1958. IGO-AMIS-9. UKAEA Production Group. Springfields. Lancs C’. K.At. Energy Auth. Rep. 1961 P.G. 15(S). Davis W. F. Graab J . W. andMerkle E. K . . Talanta 1971, 18 263. Willis R. R. Metallurgia 1969 79 129. Clapham C. A. Hall G. D. and Scholes P. H. Br. Steel Corp. Open Rep. 1970 MGlDl647170. Prall J . R. Rep. At. Energy Comm.U . S . 1970 NCLO-1061. Milner 0. I. Zatner R. J. Hepner L. S . andcowell W. H., Anal. Chem. 1958 30 528. Smith A. J. Cooper F. F. Rice J. O. and Shaner W. C., Anal. Chim. Acta 1968. 40 341. Heistand R. N. Anal. Chem. 1970 42 903. Kurchatova. G. Roczn. Panst. Zakl. Hig. 1971 22 289. Kifune I. Iwatani M. and Kawata K. Bunseki Kagaku, 1979 28 633. Harrell T. G. Rush K. L. and Sensabaugh A. J . Tob. Sci., 1975 19 154. Murfin. J . W. Analyst 1972. 97 663. Murfin J . W. and Wragg. J . S . Analyst 1972 97 670. Davis D. R. Fogg A. G. Thorburn Burns D and Wragg, J . S Analyst 1974 99 12. Ellcock C. T. H. and Fogg A. G. Lab. Pract 1974.23.555. Ellcock C. T. H. and Fogg A. G. Analyst 1975 100 16. Frings C. S and Saloom J. M. CIin. Toxicol. 1979 15 67. Rexroad P.R. and Krause G. F. J . Assoc. Off. Agric. Chem. 1970 53 450. Kaul A. K. and Sharma T. R. Freseniits Z . Anal. Chem., 1976 280 113. Collins P . F. Lawrence W. W and Williams J . F Beitr. Tabakforsch. Int. 1972. 6 167. Dienst S. G. and Morrk. B J . Lab. Clin. Med. 1964. 64. 495. Englhardt A Haertwig. B . and Franken F Z. Klin. Chem. Klin. Biochem. 1971. 9 229. 357. 35P. 359. 360. 361. 362. 363. 364. 365. 366. 367. 368. 369. 370. 371. 372. 373. 374. 375. 376. 377. 378. 379 * 380. 381, 382. 383. 384. 385. 386. Secchi A. R. Rezzonico A. and Cocchi A, Clin. Chim. Acta 1965 12 235. Ijpma S . T. Blizenberg B. G. and Lei,jnse. B. Clin. Chem 1978 24 489. Stahl A Journees de L’Internat en Pharmucie de Grenoble, “Les Aminoacidopathies” Volume 11. 1972 p. 269. Searcy R. L. Reardon J . E. and Foreman J . A. Am. 1. Med. Technol 1967 33 I. Searcy R. L and Cox F. M. Clin. Chim. Acta 1963 8 810. Chaney A. L. and Marbach E. P. Clin. Chim. 1962,8 130. Bennett D. J. Am. J . Clin. Pathol 1974 61 616. Hambraeus. L. Forsum E. Abrahamsson L. and Loenner-dal B. Anal. Biochem. 1976 72 78. Sapozhnikov V. V. Medvedeva. T. W. Sychkova G . I., Khim. Morei Okeanov Mnfer. Vses. Kmf. Khim. Morw 6th, 1972. Anderson A. J. and Rey F. Int. Counc. Explor. Sea C . N . , Hydrography Committee. 1972 C 18. Racanelli E . Legittimo. P. Cellini A,. and Pantani F. Rass. Chim. 1978 30 309. Mackereth F. J. H. Sci. Publ. Freshwater Biol. Assoc. No. 21 1963. Petts K. W. Tech. Rep. Water Res. Cent. (Medmenham, Engl.) 1979 TR119. Hagihara K. Kogai Bunseki Shishin 1972 5 25. Rowland A. P. Commun. Soil Sci. Plant Anal. 1983 14 49. La Rosa C. F. Rev. Assoc. Bioquim. Argent. 1958. 23. 228. Isaac R. A. and Johnson W. C. J . Assoc. Off. Anal. Chem., 1976 59 98. Engels W. and Orbe G. Landwirtsch. Forsch. 1970.22 116. Jenkins D. J . Water Pollut. Control Fed. 1967 39 159. Dahl. I. NORDFORSK Miljoeverdssekr. Publ. 1973 1. 22. Gasanov S. G. Lab. Delo 1962 8 3. Zhong S. Fenxi Huaxue 1982. 10 191. Wilhelms A. and Bernandt H. Wasser 1969. 36 353. Maddix C. Norton R. L. and Nicholson N. .I. Anctlysf, 1970 95 738. Malikova. E. D. Kunin L. L. and Uritskaya T. P Zh. Anal. Khim. 1977 32 88. Kakita Y. Nippon Kinzoku Gakkaishi 1949 13 39. Kallmann S . Hobart E. W. Oberthin H. K. and Brienza. W. C. Anal. Chem. 1968 40 332. UKAEA Industrial Group Chem. Services Dept. U. K.At. Energy Auth. Rep. 1958 IGO-AMIS-121. Hashitani H. and Yoshida. H. Bunseki Kagaku 1970 19, 1081. Gales M. E. and Booth R. L. U.S. Environ. Protect. Agency Rep. 1974 EPA-67014-74-002. -Prcper A31358 Received October 13th 1983 Accepted December 22nd 198

ISSN:0003-2654    

DOI:10.1039/AN9840900549

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 5 69 Mechanisms of Transition Metal Interferences in Hydride Generation Atomic-a bsorption Spectrometry Part 1. Influence of Cobalt Copper Iron and Nickel on Selenium Determination Bernhard Welz and Marianne Melcher Department of Applied Research Bodenseewerk Perkin-Elmer 8t Co. GmbH 0-7770 Uberlingen FRG Transition metal interferences in the determination of selenium using the hydride generation AAS technique were investigated in a system in which the selenium hydride is generated in pure acid solution and comes into contact with the metal ions only in a second flask. All investigated elements interfere in the ionic form with selenium when they are present in sufficiently high concentrations. Whenever a precipitation of the metallic species occurs however capture and decomposition of the selenium hydride by the finely dispersed metal appear to be the predominant mechanism of interference.In addition this solid -gas reaction occurs typically at considerably lower interferent concentrations than the liquid - gas reaction of the ionic species. A substantial increase in the range of interference-free determination of selenium can be achieved by increasing the acid concentration of the solution for measurement because of the increased solubility of the reduced metal in the strong acid. Keywords H ydride generation atomic-absorption spectrometry; transition metal interferences; mechanisms of interferences; selenium determination Owing to the separation of the analyte element from the matrix constituents by volatilisation in the form of the gaseous hydride the hydride generation AAS technique is free from spectral interferences caused by background absorption when suitable equipment is used.Because of the very limited number of elements that can be volatilised by this technique, vapour-phase interferences are also very unlikely except for mutual interferences of the hydride-forming elements. 132 It is well known however that a number of transition metals, mainly those of Groups VIII and IB can cause severe signal depressions in the hydride generation technique and sel-enium is one of the elements that is most affected by these interferences. In a review in 1979 Robbins and Caruso3 found that the studies published only consider whether or not and to what extent a substance interferes.Further work elucidating the mechanisms by which a particular substance interferes with hydride formation was not then available. This situation has not changed much in the meantime and only a few publica-tions include some information on the possible mechanisms involved in hydride generation and the observed transition metal interferences. Smith’ published a general study involving 48 elements and observed that many of the elements that interfered formed precipitates after the addition of sodium tetrahydroborate-(111). He proposed that preferential reduction of the metal ion interferent in solution to a different valency state or to the free metal can cause precipitation of that species which can then either co-precipitate the analyte metal or adsorb the volatile hydride formed and catalytically decompose it.Smith assumed that some of the tetrahydroborate is used up this way and less is available for reduction of the analyte element to the hydride. Pierce and Brown4 found that the order of tetrahydroborate and hydrochloric acid addition in their automated system caused significant differences in the analytical data. If the tetrahydroborate was added first a dark precipitate was observed which was most marked for Cd” Cox+ Fe3+ and Pb‘f. This precipitation in all instances except for cadmium resulted in total suppression of the analytical signal for selenium. They proposed that these cations compete with the selenium for reduction and consume the reducing agent. No precipitation and no cationic interferences (up to 33 mg 1-1 of interferent) were noted when the hydrochloric acid was added prior to the tetrahydroborate solution.Kirkbright and Taddia5 also noticed that in the presence of elements such as nickel palladium or platinum after the addition of the reductant a very finely dispersed black precipitate was formed. They also reported for the determina-tion of arsenic that on the addition of nickel powder virtually complete suppression of the signal was observed. The test was performed in the presence of a masking agent to complex the trace amounts of metal eventually dissolved. They pointed out that nickel and other Group VIII elements are hydrogenation catalysts and can also adsorb hydrogen in large amounts. Hence capture and decomposition of the hydride by the finely dispersed metal can occur.Mayer et a1.,6 in a careful study of the determination of selenium did not mention an interferent precipitation. They proposed that the selenium hydride after its generation, forms insoluble selenides or stable complexes with the free ions of the interfering metals in a secondary reaction when it is transported through the sample solution by the carrier gas. Such a gas - liquid reaction would depend only on the speed of diffusion of the selenium hydride to the gas - liquid interface and of the concentration of the interfering ion in the solution for measurement. Meyer et a1.6 also found that the transition metal inter-ferences observed for selenium depend very strongly on the acid concentration in the solution for measurement.Whereas in 0.3 M HCl Cu in masses above 2 pg started to depress the selenium signals more than 200 pg of Cu could be tolerated in 5 M HCl. An additional influence comes from the volume of acid used for the measurement because the transition metal interferences in hydride generation do not depend on the analyte to interferent ratio but on the interferent concentra-tion in the final solution applied for measurement. Working with 5 M HCl we found no interferences from up to 40 pg of Cu and 60 pg of Ni in 10 ml of acid whereas up to 500 pg of Cu and 1000 pg of Ni could be tolerated in 50 ml of acid.’ The discrepancies in the degree of signal depression reported in the literature for the same interferent can frequently be explained by the different acid concentration and volume applied as well as other instrumental details.Vijan and Woods found 80% depression of the selenium signal by 2 mg 1-1 of Cu and 20% depression by 2 mg 1-l of Ni or 20 mg 1-1 of Fe in an automatic flow system using 6 M HC1. Pierce and Brown,j using a similar system and 5 M HCI. found no interferences from up to 31 mg I-* of Co Cu or Fe. McDaniel et al.9 using a batch system with SO ml of 4 M HCl 570 ANALYST. MAY 1984 VOL. 109 reported less than 5% signal depression from 100 mg 1-1 of Co Cu or Ni and 200 mg 1-1 of Fe. In the analysis of low-alloy steels we have shown that addition of HN03 to 0.5 M HC1 has a similar effect in the determination of Se.10 Kirkbright and Taddias explained the decrease in inter-ferences with increasing acidity of the solution by the greater solubility of the reduced interfering metal in the stronger acid.To distinguish between the proposed mechanisms we carried out interference studies in a system where the hydride was generated in pure dilute HCl first and then conducted into a second flask containing the interferent in acid solution. With this arrangement a reduction of the interfering metal ion and its precipitation can be excluded as it does not come into contact with the tetrahydroborate solution. The results were compared with those obtained when the interfering metal was present directly in the reaction flask so that a reduction and precipitation were possible. The influence of the acid concen-tration in both arrangements was also investigated.Experimental Apparatus A Perkin-Elmer Model 4000 atomic-absorption spectro-photometer equipped with an electrodeless discharge lamp for selenium operated at 6 W from an external power supply was used for all determinations. A spectral band pass of 2 nm was selected to isolate the 196.0-nm line and the signals were recorded on a Perkin-Elmer Model 56 recorder. The Perkin-Elmer Model MHS-20 hydride system used has been des-cribed in detail elsewhere.2 The system was modified in such a way that the selenium hydride formed was conducted through a gas wash-bottle with a volume of approximately 100 ml before it reached the heated quartz cell (Fig. 1). The time settings on the MHS-20 controller were 35 s for Purge I (this is the time to drive all air out of the system) 8 s for Reaction [during this time sodium tetrahydroborate(II1) solution is continuously fed into the reaction vessel] and 35 s for Purge I1 (during this time the reaction settles and the hydride and hydrogen are driven out of the system).The electrically heated quartz cell atomiser was held at 900 "C. The solution volumes used throughout these experiments were 10 ml in the reaction flask and 50 ml in the gas wash-bottle. The solution in the gas wash-bottle was 0.5 M HCl during all "blank" runs. Reagents Sodium tetrahydroborate(III) solution 3% m/V. Prepared by dissolving sodium tetrahydroborate( 111) powder (Riedel-de-Haen) in de-ionised distilled water and stabilising with 1% m/V sodium hydroxide solution. The solution was filtered before use and could be stored for a few days only.Heated quartz cell 900 "C T NaBH, Gas Fig. 1. Experimental arrangement for separating the generation of selenium hydride in the reaction vessel from the interfering ions in the gas wash-bottle Selenium(ZV) stock standard solution 1000 mg 1-1. Pre-pared by diluting a Titrisol solution (Merck) containing 1 .000 g of selenium to 1 1 with de-ionised distilled water. Aliquots were diluted with 0.5 M HCl to obtain appropriate working reference solutions. Transition metal solutions. The transition metal solutions for the interference tests were prepared from cobalt chloride (CoC12.6H20) copper sulphate (CuS04.5H203 nickel nitrate [Ni(N03)2.6H20] iron(II1) nitrate [Fe(N03)3.9H20] and iron(I1) sulphate (FeS04.7H20).All these salts (Merck) were of analytical-reagent grade or higher purity. Hydrochloric acid. High-purity HCI was prepared from analytical-reagent grade acid (Merck) by sub-boiling distilla-tion. Nitric acid. High purity prepared as for HCI. Results and Discussion In a first set of experiments a pure selenium standard (100 ng of Se in 10 ml of 0.5 M HCl) was in the reaction flask of the hydride system and the interfering metal ions Co(II) Cu(II), Fe(II) Fe(II1) and Ni(II) were in the gas wash-bottle in 50 ml of 0.5 M HCI. The results for the determination of selenium with this arrangement are compared with those when the selenium and the interfering metal ion are both in the reaction flask and pure 0.5 M HCI in the gas wash-bottle (Fig.2). From the interference curves it becomes apparent that all investi-gated elements cause some signal depression when they are present in the ionic form in the gas wash-bottle. The effect, however is very moderate and essentially identical for the bivalent ions of iron cobalt and nickel. Only copper has a pronounced influence on the selenium signal. When present in the reaction flask however all bivalent ions investigated show a much more dramatic signal depres-sion that starts at more than two orders of magnitude lower interferent concentrations. In all these instances a dark precipitate is formed immediately after the addition of the sodium tetrahydroborate(II1) solution. A direct qualitative correlation could be found between the visible appearance of a precipitate and the observation of a signal depression.The only exception is trivalent iron where the interferences are B 50 I 0 10-4 10-3 10-2 10-1 1 10 lnterferent concentrationig 1-1 Fig. 2. Influence of cobalt copper. iron and nickel on the determination of selenium in 0.5 M hydrochloric acid. A Interferent in the reaction flask; and B. interferent in the gas wash-bottl ANALYST MAY 1984 VOL. 109 57 1 virtually identical for the interferent present in the gas wash-bottle and in the reaction flask. In this instance a dark precipitate is also formed in the reaction flask but not immediately after the addition of the tetrahydroborate sol-ution but with a substantial delay and only after the atomisation signal for selenium has already been recorded.These observations allow the conclusion that all investi-gated metals interfere in the ionic form according to the mechanism proposed by Meyer et a1.6 by forming insoluble selenides or stable complexes in a gas - liquid reaction. When present in the reaction flask however which is the normal situation the first reaction at least for the bivalent ions investigated seems to be a reduction of the metal ion to the metal and its precipitation. A competition of the interfering compounds with selenium compounds for reduction as proposed by Pierce and co-workers,jJl is not very likely however. One of the reasons is that the tetrahydroborate is always present in a large excess over the interfering ion so that a depletion of the reductant is very unlikely.In addition these interferences are typically not affected by the tetrahydroborate concentration. Finally a competition of the interferent with the selenium for reduction should only slow down or delay but not inhibit the production and evolution of the selenium hydride when a large enough excess of tetrahydroborate is added. Integration over the absorbance signal should therefore eliminate or substan-tially reduce these interferences which however is not the case. The effect appears to be even more pronounced in most instances when signal integration is used for peak evaluation. We therefore believe that the interference is caused by capture and decomposition of the evolved gaseous hydride by the finely dispersed metal precipitate as already discussed by Smith' and proposed later by Kirkbright and Taddia.5 This mechanism is in agreement with the finding that a change in the tetrahydroborate concentration does not affect the inter-ference pattern because a higher reductant concentration will speed up both hydride formation and the precipitation of the interfering metal.This mechanism is also confirmed by the visible coincidence of the formation of a precipitate and the observation of a signal depression seen by us and confirmed by most other workers. From all of the ions investigated in this study the only exception from the general interference pattern is trivalent iron where the signal depressions are essentially identical for the ion in the gas wash-bottle and in the reaction flask. As mentioned earlier in this instance the precipitation in the reaction flask is visibly delayed until after the selenium signal is recorded.We assume that in this instance the reduction is a two-step process with trivalent iron being reduced to bivalent iron first followed by the reduction to the metal and its precipitation only after all trivalent iron has been reduced to its bivalent stage. This explains the delayed precipitation and the even smaller signal depression when the interferent is present in the reaction flask because ionic bivalent iron has a smaller effect than ionic trivalent iron on the selenium signal. The actual effect of the metal ions however is typically obscured by the influence of reduced species formed by reaction with tetrahydroborate. The strong depressing effect of bivalent iron reported by several workers is in fact due to precipitated metallic iron and not to the ion.Selenium hydride is formed independent of, and in parallel with these reduction reactions because there is always a large enough excess of tetrahydroborate to leave the formation of the hydride unaffected from concomitants in the solution. The depression of the selenium signal by trivalent iron is in our opinion an ionic interference both in the reaction flask and in the gas wash-bottle. When the same experiments are carried out in 5 M instead of 0.5 M HCI some changes in the interference pattern become apparent (Fig. 3). When the interferent is present in the reaction flask the range of interference-free determination is extended considerably for all bivalent metal ions.The only exception is again trivalent iron for which an increase in the acid concentration does not bring about a major effect. The same is true for all metal ions investigated when they are present in the gas wash-bottle in 5 M HC1. These observations are expressed more quantitatively in Table 1 where the effect of the acid concentration on the I l I ' I I 1 I I I 10-4 10-3 10-2 10-1 0 10 50 L 0 lnterferent concentrationig I-Fig. 3. Influence of cobalt copper iron and nickel on the determination of selenium in S M hydrochloric acid. A Interferent in the reaction flask; and B. interferent in the gas wash-bottle Table 1. Comparison of interferences from transition metal ions on the determination of 100 ng of selenium in 0.5 and 5.0 M hydrochloric acid with the interfering element in the reaction flask or in the gas wash-bottle.The start of the interference is defined as a 10% signal depression caused by the interferent. Interferent concentrations in g 1 - I Start of transition metal interference ~~ ~ In the reaction flask In the gas wash-bottle Interfering Interferent concentration metal 5 M HCli ion 0.5 M HCl 5.0 M HCI 0.5 M HCI Cu(1I) . . . . . . . . 0.000 07 0.003 40 Co(I1) . . . . . . . . 0.00s 0.2 40 Fe(I1) . . . . . . . . 0.003 0.04 13 Ni(I1) . . . . . . . . 0.0004 0.004 10 Fe(1II) . . . . . . . . 0.7 0.6 0.85 Interferent concentration 5 .O M HCli 0.5 M HC1 5.0 M HCl 0.5 M HC1 0.02 0.007 0.35 1 3 3 1 4 4 1 5 5 0.3 0.3 572 range of interference-free determination in the presence of the different transition metal ions is shown.“Start of interfer-ence” is defined as a 10% signal depression by the interferent. From the comparison in Table 1 it becomes apparent that the range of interference-free determination is increased by a factor of 1040 by a 10-fold increase in the acid concentration when the bivalent ions are present in the reaction flask. For trivalent iron in the reaction flask and for all metal ions investigated when they are present in the gas wash-bottle the effect of the higher acid concentration is a factor between 0.35 and 5 which is considerably smaller than the corresponding reaction flask results. Our explanation for this observation is that the increased acid concentration has a substantial effect on the solubility of the precipitated metal or in other words precipitation occurs only at higher metal ion concentrations.The same mechanism has already been proposed by Kirkbright and Taddias for arsenic and the interference of Group VIII elements. The theory is further supported by the fact that the addition of increasing concentrations of HN03 to the 0.5 M HC1 has a comparable releasing effect on the determination of selenium in the presence of Group VIII and IB elements.7,10312 It is well known that these elements show excellent solubility in HN03 and HCl mixtures. Increasing the acid concentration has only a minor effect when the interfering metal ions are in the gas wash-bottle and the influence of the acid can be positive or negative.This is in agreement with our theory because only a minor influence can be expected from the HC1 concentration on the concentration of interfering metal ions. We know that several transition metal ions form complex compounds in more concentrated HCI and these complexes are most probably responsible for the extension of the range of interference-free determination for Co2+ Fe2+ and Ni*+ when present in 5 M HC1 in the wash-bottle. The concentration of free ions however is still high enough so that they can react with selenium hydride to form insoluble selenides or complexes. Trivalent iron is almost unaffected by the changes in the acid concentration or composition when present in both the reaction flask and the gas wash-bottle. As expected for this ion the first reaction step is again a reduction of the trivalent to the bivalent state.This also explains the flatter slope of the interference curve at higher iron concentrations because the bivalent iron ion has a less pronounced influence on selenium than the trivalent ion. In essence using a system where the selenium hydride is separated from the interferent it could be shown that Co Cu, Fe and Ni cause a signal depression in the determination of selenium when present in sufficiently high concentrations. A gas - liquid reaction of the hydride formed with the interfering ions in solution is the most likely mechanism. Whenever precipitation of the metallic species occurs after the addition of tetrahydroborate solution which is the case for all bivalent ions investigated capture and decomposition of ANALYST MAY 1984 VOL.109 the hydride by the finely dispersed metal appear to be the predominant mechanism of interference. This interference occurs typically at considerably lower interferent concentra-tions but can be substantially reduced by increasing the acid concentration. Trivalent iron when present in the reaction flask is reduced to its bivalent state in a first-step reaction on addition of tetrahydroborate solution. The precipitation of metallic iron occurs in a second step typically after the selenium signal has already been recorded. As expected this reaction cannot be influenced to any significant extent by the acid concentration of the solution and is virtually independent of the conditions chosen.Conclusion It has been shown that the investigated transition metals Co, Cu Fe and Ni start to affect the determination of selenium only at high interferent concentrations (above 10-100 mg 1-1) when present as the ions. The severe interferences reported in the literature are caused by the reaction of gaseous hydrogen selenide with the metallic species that are precipitated in a finely dispersed form by the tetrahydroborate. The most obvious means of increasing the range of interference-free determination of selenium is therefore to prevent reduction and precipitation of the transition metal ions. One way that has been found effective is to increase the acid concentration of the solution for measurement. A concentration of HCl of 5 M appears to be a favourable environment for the deter-mination of selenium in the presence of various interferents. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Smith A. E . Analyst 1975 100 300. Welz B. and Melcher M Anal. Chirn. Acta 1981 131 17. Robbins W. B. and Caruso J. A . Anal. Chern. 1979 51, 889A. Pierce F. D and Brown H. R. Anal. Chern. 1976.48.693. Kirkbright G . F and Taddia M. Anal. Chirn. Acta 1978, 100 145. Meyer A. Hofer Ch. Tolg G Raptis S . and Knapp G., Fresenius Z. Anal. Chern. 1979 296 337. Welz B . and Melcher M. Wasser 1984 62 in the press. Vijan P. N. and Wood G. R. Talanta 1979 23 89. McDaniel M. Shendrikar A. D. Reiszner. K. D. and West, P. W. Anal. Chern. 1976 48 2240. Welz B. and Melcher M. Spectrochirn. Acta Part I? 1981, 36 439. Pierce F. D . Lamoreaux T. C. Brown H. R and Fraser, R. S . Appl. Spectrosc. 1976 30 38. Welz B. Grobenski Z . . and Melcher M in Koch K. H., and Massmann H. Editors “13. Spektrometertagung,” Walter de Gruyter Berlin New York 1981 p. 337. Paper A31313 Received September 12th 1983 Accepted December 13th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900569

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST. MAY 1984 VOL. 109 573 Mechanisms of Transition Metal Interferences in Hydride Generation Atomic-a bsorption Spectrometry Part 2.* Influence of the Valency State of Arsenic on the Degree of Signal Depression Caused by Copper Iron and Nickel Bernhard Welz and Marianne Melcher Department of Applied Research Bodenseewerk Perkin-Elmer & Co. GrnbH 0-7770 herlingen FRG A significant influence of the valency state of arsenic was found on the degree of the signal depression owing to transition metals in hydride generation AAS. This phenomenon can be explained by the slower evolution of arsine from As(V) in comparison with As(lll) and the more complete precipitation of the interfering metals by the time the hydride is formed. This observation supports the theory that the interferences are due to a gas -solid reaction.Keywords Hydride generation atomic-absorption spectrometry; transition metal interferences; valency state of arsenic; arsenic determination; mechanism of interferences The difference in sensitivity between trivalent and pentavalent arsenic in the hydride generation AAS technique is well documented. There are however. some inconsistencies in the reported magnitude of the effect. Vijan et a1.I found a sensitivity difference of 25-30°/i. Thompson and Thoresby2 reported that the ratio of the signals for the two oxidation states varies with the major anion present. Solutions in H2S04 were found to give more consistent results with only a slight reduction in sensitivity. In 1.5 M H2S04 the absorbance signal for As(V) was 54% of that for As(II1).Peat$ reported that the signal for As(V) is about 30% lower in peak height but only 15% lower in peak area. Siemer et aL4 found different peak heights but identical peak areas in 7 M HCl. They admitted however that in general this is not the case. They found that the reduction of As(V) is often incomplete during the time in which tetrahydroborate actually exists in the highly acidic solutions to which it is added. Hinners’ also reported that As(II1) and As(V) respond similarly in HCl when sufficient sodium tetrahydroborate(II1) reagent is used. Thompson and Thomerson6 found that As(V) gives only 37% of the signal of As(I1I) with 1% mlV sodium tetrahydroborate(II1) solution but 90% when a 4% mlV tetrahydroborate(IT1) solution is used for reduction.Siemer and Koteel7 reported that the often observed difference in response seen for As(II1) and As(V) disappears when the arsine is collected in a cold trap before it is introduced into the atomiser cell provided that enough tetrahydroborate is used. This would indicate that the effect is due to reduction kinetics of the arsenic species. Mainly because of this difference in sensitivity most workers prefer to reduce arsenic to the trivalent state before its determination by the hydride technique. Goulden and Brooksbanks used a potassium iodide - tin(I1) chloride mixture for this purpose and found that the reduction is complete in approximately 20 s. Peat9 used a 10% mlV solution of potassium iodide in 1% mlV ascorbic acid for reduction whereas Sinemus et al.9 found that 10% mlV potassium iodide solution alone is sufficient if the reduction is carried out in 5 M HCl.The reaction is complete after about 15 min at ambient temperature. Some workers however prefer to determine arsenic directly from its pentavalent state when the element is obtained in that valency state after sample decomposition because the sensitivity difference is only small.6 Vijan et al. determined arsenic in soil and vegetation after digestion in nitric and perchloric acids and found that the “normalising * For Part 1 of this series see p. 569. effect of the decomposition treatment of both samples and references practically eliminates the difference” between the two valency states. Evans et al. 10 determined arsenic and other volatile hydride-forming elements in foodstuffs after a dry-ashing or wet-oxidation procedure.Using this method arsenic is obtained in the highest valency state and converted into the respective hydride without pre-reduction. For the analysis of low-alloy steels” we decomposed the samples in aqua regia and obtained arsenic and antimony in their pentavalent states. Both elements were determined against acid references of the same valency state for simplicity reasons. There are numerous reports in the literature on inter-ferences from transition metals on the determination of arsenic in the hydride AAS technique. To our knowledge, however there is only a single short note that the valency state of a hydride-forming element has an influence on these interferences.Fleming and Idel2 reported for the determina-tion of antimony in steels that they initially experienced serious interferences and that most of these disappeared when iodide was added to reduce Sb(V) to Sb(II1). We have found that the extent of signal depression is considerably more pronounced for pentavalent arsenic so that a reduction to the trivalent state would not only improve sensitivity but also reduce interferences. Experimental Apparatus A Perkin-Elmer Model 4000 atomic-absorption spectropho-tometer equipped with an electrodeless discharge lamp for arsenic operated at 8 W from an external power supply was used for all determinations. A spectral band pass of 0.7 nm was selected to isolate the 193.7-nm line and the signals were recorded on a Perkin-Elmer Model 56 recorder set at the 10-mV range.The Perkin-Elmer Model MHS-20 hydride system used has been described in detail elsewhere.13 The electrically heated quartz cell atomiser was held at 900 “C and the time settings at the controller were 35 s for Purge I 10 s for Reaction [during this time sodium tetrahydroborate(II1) solution is transported to the reaction flask] and 40 s for Purge 11. Reagents All acids used were of analytical-reagent grade. HCl and HN03 were further purified by sub-boiling distillation. The metal salts used for interference studies were of analytical-reagent grade 574 ANALYST MAY 1983 VOL. 109 Sodium tetrahydroborute(III) solution 3% mlV. Prepared by dissolving sodium tetrahydroborate( 111) powder (Riedel-de-Haen) in de-ionised distilled water and stabilising with 1% mlV sodium hydroxide solution.The solution was filtered before use and could be stored for a few days only. Arsenic( V) stock standard solution 1 000 mg 1-l. Prepared by diluting a Titrisol solution (Merck) containing 1.000 g of arsenic (as HiAs04) to 1 1 with de-ionised distilled water. Aliquots were diluted with 0.5 M HC1 to obtain appropriate working reference solutions. Arsenic(III) stock standard solution 1 000 mg 1-1. Prepared by dissolving 1.320 g of arsenic(II1) oxide (As203) in 25 ml of 3.5 M H2S04 adjusting to a phenolphthalein end-point and diluting to 1 1 with 0.2 M H2S04. Aliquots were diluted with 0.5 M HCl to obtain appropriate working reference solutions. Sample Preparation The reduction of As(V) to As(II1) is best carried out in 5 M HC1 medium.A 1-ml volume of 10% mlV potassium iodide solution is added to 10 ml of sample solution and left at ambient temperature for at least 15 min. After this time the reduction is complete and the solution is transferred into the hydride system for the determination of arsenic. Results and Discussion We have shown earlier” that the range for the interference-free determination of arsenic and the other volatile hydride-forming elements can be extended by one to several orders of magnitude when the reaction is carried out in higher acid concentrations or mixtures of acids such as hydrochloric and nitric acids. The optimum acid environments determined earlier for arsenic to give the greatest freedom from inter-ferences in the presence of a single metal ion are summarised in Table 1.The presence of high nickel concentrations requires an increase in the acid concentration and the addition of iron as a releasing agent14 in order to obtain the maximum freedom from interferences. Except for the iron and nickel pair nothing was known on the kind of influence and the optimum acid concentration when more than one interfering element is present. This, however is of particular interest in the analysis of alloys, high-alloy steels and industrial effluents. We therefore investi-0.3 0.2 0.1 0 Q) 0 0.4 $ a CX 0.3 0.2 0.1 0 (a’) ( 2% ) gated the releasing effect of the different acid concentrations and mixtures given in Table 1 for the determination of arsenic in the presence of all three interfering metals.The concentra-tions chosen were 50 mg 1-1 for copper and iron and 10 mg 1-1 for nickel. The signals obtained for 50 ng arsenic under the various conditions are shown in Fig. 1. The peak-height sensitivity for pure As(V) standards was found to be between 60 and 70% of that for As(II1) standards, fairly independent of the acid concentration and mixture used. This difference disappeared almost completely when peak-area integration was used. The influence Gf the investigated metals however is not only different for the different acid mixtures and concentrations but also for the two valency states of arsenic. In a 0.5 M HC1 environment the presence of the three metals causes essentially complete signal suppression.In 0.5 M HC1 - 0.3 M HN03 (a mixture that helped to overcome the interference of up to 2000 mg 1-1 of iron alone) the simultaneous presence of copper iron and nickel again causes almost complete (96%) suppression for As(V) whereas for As(I1I) a clearly lower depression of only 85% is observed. In 0.5 M HC1 - 1.75 M HN03 with the addition of 2 mg of iron to the 10 ml of original sample solution (a mixture that allowed an interference-free determination of arsenic in the presence of up to 100 mg 1-1 of nickel) a 65% signal depression is found for As(V) in the presence of all three metals but only a 15% signal depression for As(II1). Finally in 5 M HCl medium (which allows up to 1000 mg 1-1 of copper alone to be Table 1. Range of interference-free determination of arsenic(II1) in the presence of copper iron or nickel in 0.5 M HCl and in an optimised acid environment.The “maximum interferent concentration” is defined as giving less than 5% signal depression compared with the matrix-free solution Maximum Maximum interferent interferent concentration concentration in 10 ml of in 10 ml of Interfering 0.5 M Optimised acid optimised element HCl/mg environment acidimg c u . . . . 1 5 M HCI 10 Fe . . . _ 1 0.5 M HC1- 0.3 M HN03 20 Ni . . . . 0.001 0 . 5 ~ H C l - 1 . 7 5 ~ H N 0 1 +2 mg Fe(II1) d) Fig. 1. Influence of the valency state of arsenic on the signal depression caused by 50 mg 1- of Cu 50 m 1-1 of Fe and 10 mg I - ’ of Ni in different acid concentrations and mixtures.(u)-(d) 50 ng of As(V); (u’)-(d’) 50 ng of As(II1). dixtures ( a ) (a’). 0.5 M HCl; ( b ) (b’) 0.5 M HCl - 0.3 M HN03; ( c ) ( c ’ ) 0.5 M HCl - 1.75 M HN03 + 2 mg Fe(III) per 10 ml; and (d) (d’). 5 M HC1. In each box the pair of peaks on the left represent the metal-free standard and those on the right with interfering metals added; the percentages in parentheses are the residual signals in comparison with the metal-free standards in the same acid mediu ANALYST. MAY 1984 VOL. 109 575 0.5 0 10 20 30 40 50 Ti m eis Fig. 2. Superimposed recorder tracings for 50 ng of As(II1) and 50 ng of As(V) in 10 ml of 0.5 M HCI. Chart speed 240 mm min-l. Higher acid concentrations shift both signals slightly to the left tolerated) the three metals together cause a 35% signal depression for As(V) but only less than 10% depression for As( I1 I).This observation leads to the recommendation that 5 M HCl is the best and most universally applicable medium for the determination of arsenic in the presence of mixed transition metals such as copper iron and nickel. It has been found, however that iron(II1) has to be present in excess to control the influence of nickel on the determination of arsenic completely.15 A 0.2-mg amount of iron was therefore added to all solutions in 5 M HCl similar to earlier procedures,12.14 independent of the iron content of the samples to make sure that a minimum amount of iron is present in every instance. In addition and even more importantly arsenic should be present in the final solution for measurement in its trivalent and not its pentavalent state.According to our results this should provide a considerably greater freedom from inter-ferences from transition metals. As 5 M HCl is an ideal medium for the reduction of arsenic with potassium iodide,9 sample preparation becomes a very simple procedure. The difference in the degree of signal depression for As(II1) and As(V) by transition metals found in this work has not been reported previously but may be a much more general problem than realised or anticipated in the past. The explanation for this phenomenon must be found in the reaction mechanisms involved. A reaction of the unreduced ions can be excluded because the signal depression is independent of the contact time of analyte and interferent prior to the addition of the reducing agent.Smith16 and also Kirkbright and Taddial7 proposed a preferential reduction of the interfering metal ion to a different valency state or to the free metal which is precipitated. This can then adsorb the volatile hydride formed when it passes through the solution in a bubble of carrier gas, catalytically decompose it and form insoluble compounds. A finely dispersed dark precipitate is typically observed after the addition of the sodium tetrahydroborate(II1) solution when higher concentrations of transitions metals are present in the solution for measurement. We carried out experiments for selenium in which the hydride was generated from pure aqueous solution in one flask and the interfering ion was in a second flask through which the gaseous hydride had to be passed.18 For most elements the interferences were several orders of magnitude smaller in this arrangement compared with the interferent present in the reaction flask where it can be reduced by the tetrahydro-borate.This also demonstrates that the reduction of the inter-ferent is an important step and that capture and decomposi-tion of the evolved hydride by the finely dispersed precipi-tated interfering metal is the most likely mechanism. The only major difference in the behaviour of the two valency states of arsenic in the hydride technique is the slower evolution of arsine from As(V) than from As(III) as shown in Fig. 2. This means that the precipitation of the interfering metal is more complete at the time when the arsine is evolved from As(V) leading to a more pronounced interference.In addition the evolution of the hydride is distributed over a longer period of time for the higher valency state. This means that the same mass of arsine is in contact with the interferent on average for a longer period of time. Such a gas - solid reaction essentially depends only on the speed of diffusion of the hydride to the surface of the gas bubble in which it is embedded. According to the laws of diffusion the mass of hydride transported to the surface of the gas bubble is largely independent of the concentration of the analyte element in the sample solution. This is in agreement with the finding that the signal depressions do not depend on the analyte concentration and the analyte to interferent ratio but only on the concentra-tion of the interfering metal in the final solution for measurement.Conclusion All the effects of copper iron and nickel observed in the determination of arsenic can be explained by the proposed mechanism of preferential reduction of the interferent to the metal and capture and decomposition of the evolved hydride by the finely dispersed precipitated interfering metal. Vice versa the theory of the preferential reduction and the gas -solid reaction is supported by the finding that the signal depression caused by transition metals is more pronounced for As(V) than for As(II1). The recommended conditions for the determination of arsenic using hydride generation include reduction to As( 111) with potassium iodide and determination in 5 M HCI as the final solution for measurement.In the presence of nickel the addition of iron(II1) further increases the range of interference-free determination. 1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Vijan P. N. Rayner A . C. Sturgis D. and Wood G . R Anal. Chim. Acta 1976 82 329. Thompson A . J and Thoresby P. A . Analysr 1977 102 9. Peats S . At. Absorpt. Newsl. 1979 18 118. Siemer D. D. Koteel P . and Jariwala V. Anal. Chem., 1976 48 836. Hinners. T. A . Analyst 1980 105 751. Thompson K. C and Thomerson D. R. Analyst 1974 99, 595. Siemer D. D . and Koteel P. Anal. Chem. 1979 49 1096. Goulden P. D . and Brooksbank P . Anal. Chem. 1974.46, 143 1. Sinemus H. W Melcher M. and Welz. B. A t . Spectrosc 1981 2 81. Evans W. H. Jackson F. J . and Dellar D Analyst 1979, 104 16. Welz B. and Melcher M. Spectrochim. Acta Part B 1981, 36. 439. Fleming H . D and Ide R. G. Anal. Chim. Actu 1976. 83, 67. Welz. B. and Melcher M. Anal. Chim. Actu 1981 131 17. Welz B. Grobenski Z. and Melcher M. in Koch K. I-I., and Massmann H. Editors “13 Spektrometertagung.” Walter de Gruyter Berlin New York 1981. p. 337. Welz B . and Melcher M. Wasser 1982 59 407. Smith A. E Analyst 1975 100 300. Kirkbright G . F . and Taddia M Anal. Chim. Acta 1978, 100 145. Welz B. and Melcher M. Analyst 1984 109 569. Note-Reference 18 is to Part 1 of this series. Paper A3131 4 Received September 12th 1983 Accepted December 13th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900573

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 577 Mechanisms of Transition Metal Interferences in Hydride Generation Atomic-absorption Spectrometry Part 3.* Releasing Effect of Iron(ll1) on Nickel Interference on Arsenic and Selenium A . Bernhard Welz and Marianne Melcher Department of Applied Research Bodenseewerk Perkin-Elmer & Co. GmbH 0-7770 Uberlingen FRG Iron(lll) has a releasing effect on the interference due t o nickel on the determination of arsenic and selenium. Together with an increase i n the acid concentration of the solution for measurement this effect can extend the range of interference-free determination for these elements by three orders of magnitude. A preferential reduction of Fe(lll) t o Fe(ll) which inhibits the precipitation of the interfering nickel as the metal is the proposed mechanism of reaction.Keywords Hydride generation atomic-absorption spectrometry; nickel interference; iron releasing effect; arsenic determination; selenium determination Nickel has a strongly depressing effect on the signal of arsenic and the other hydride-forming elements when present in relatively low concentrations. 1-5 Various proposals have been made to overcome these interferences and the applica-tion of masking agents is one of them. Guimont et a1.6 suggested thiocyanate Aggett and Aspell7 employed acetate and citrate at pH 4 and Belcher et a1.8 applied EDTA for this purpose. RubeSka and Hlavinkova,9 however found that these masking agents seem to work only if after mixing the sample and the tetrahydroborate solutions the final pH is alkaline or neutral or if pellets of sodium tetrahydro-borate( 111) are used.Kirkbright and Taddial0 proposed thiosemicarbazide and/or 1,lO-phenanthroline to minimise the strong suppressive effect of nickel and Dornemann and Kleistll found pyridine-Zaldoxime most effective for this purpose. Thompson et al. 12 applied a coprecipitation on lanthanum hydroxide previously proposed for the separation of hydride-forming elements from a copper matrix to overcome the interference from nickel and also other transition metals. For the determination of selenium in njckel-based materials we proposed removing the nickel from the solution by precipita-tion with sodium hydroxide. 13 Selenium remains quantita-tively in solution and can be readily determined after acidification.Fleming and Ide4 found serious interferences from nickel and some other transition metals in the determination of all hydride-forming elements in the absence of iron. In the presence of iron however most of these interferences disappeared so that a wide variety of steel samples could be analysed without problems. They kept the amount of iron in the hydride generation cell constant and an appropriate amount of iron was added when a dilution was used or when a highly alloyed steel was being analysed. We found no interferences from up to 2% of nickel in low-alloy steels when the hydride-forming elements were determined in a sufficiently high acid concentration.15 The determination could be carried out directly against acid standards after appropriate dilution.A more careful investiga-tion of the nickel influence16 showed that an increase in the acid concentration alone cannot provide sufficient freedom from interferences and that iron obviously plays an important role as a releasing agent. This effect of iron became apparent again in the determination of arsenic17 and selenium18 in * For Part 2 of this series see p. 573. industrial effluents where complete freedom from the nickel interference could be obtained only in the presence of a sufficiently high iron concentration. We believe that the releasing effect of iron(TI1) is due to a preferential reduction of this species prior to the reduction of nickel ions to the metal which then adsorbs and decomposes the hydride formed. Experimental Apparatus A Perkin-Elmer Model 4000 atomic-absorption spectro-photometer equiped with electrodeless discharge lamps, operated at 8 W for arsenic and at 6 W for selenium from an external power supply was used for all determinations.A spectral band pass of 0.7 nm was selected to isolate the 193.7-nm arsenic line and 2 nm was used for the determina-tion of selenium at the 196.0-nm line. The signals were recorded on a Perkin-Elmer Model 56 recorder set at the IO-mV range. The Perkin-Elmer Model MHS-20 hydride system used has been described in detail elsewhere19; the instrumental settings used for the determination of arsenic and selenium are summarised in Table 1. Reagents All reagents used except for the sodium tetrahydro-borate( 111) powder were of analytical-reagent grade or higher purity.HC1 and HN03 were further purified by sub-boiling distillation. Sodium tetrahydroborate( III) solution 3% mlV. Prepared by dissolving sodium tetrahydroborate(II1) powder (Riedel-de-Haen) in de-ionised distilled water and stabilising with 1 YO mlV sodium hydroxide solution. The solution was filtered before use and could be stored for only a few days. Arsenic and selenium stock standard solutions 1 000 mg 1 - I . Prepared by diluting Titrisol solutions (Merck) containing Table 1. Operating parameters for the MHS-20 hydride system Cell Element S S S "C As . . . . 35 10 40 900 Se . . . . 35 8 35 900 Purge I/ Reaction/ Purge III temperature 578 ANALYST. MAY 1984 VOL. 109 100 t-U 20 \ \ 'I b '\ I I I I I I 1 I I I I 1 1 000 Ni concentrationirng I-' 0 ' 0.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50 100 Fig.1. Influence of nickel on the determination of SO ng of As(II1) in hydrochloric acid and hydrochloric acid - nitric acid mixtures with and without the addition of trivalent iron as a buffer. A 0.5 M HCl; B 0.5 M HCl - 0.3 M HN03; C 0.5 M HC1 - 1 M HNO,; D. 0.5 M HCl + 2 mg Fe(II1); E 0.5 M HC1 - 0.3 M HN03 + 2 mg Fe(II1); and F 0.5 M HCl - 2 M HN03 + 2 mg Fe(II1) I I I 0 0.1 1 .o 10 100 1 000 Ni concentration/rng I-' Fig. 2. Influence of nickel on the determination of 50 ng of Se(1V) in hydrochloric acid hydrochloric acid - nitric acid mixtures and with the addition of trivalent iron as a buffer. A 0.5 M HCl; B 0.5 M HCI - 0.2 M HNO,; and C 0.5 M HCI - 2 M HNO + 2 mg Fe(II1) 1.000 g of the element to 1 1 with de-ionised distilled water.Aliquots were diluted with 0.5 M HCl to obtain appropriate working reference solutions. Results and Discussion It has been shown previously that the range of interference-free determination for the volatile hydride-forming elements can frequently be extended by several orders of magnitude when the concentration of the acid in the reaction flask is increased or when acid mixtures are used.15-'8,2() When nickel is the interfering metal however this procedure is not as successful as it is for several other interferents. With arsenic, the addition of 0.3 M HN03 acid to the 0.5 M HC1 extends the range of interference-free determination only by a factor of 2, and the addition of 1 M HN03 gives a factor of 20 (Fig.1). The situation is even worse for selenium (Fig. 2) where the addition of 0.2 M HN03 to the 0.5 M HC1 does not extend the range of interference-free determination at all and only reduces the degree of signal depression. The use of 5 M instead of 0.5 M HC1 increases the range of interference-free determi-nation for arsenic by a factor of 50 and for selenium by a factor of 20. All these improvements are very useful for the determina-tion of volatile hydride-forming elements in samples that do not contain too much nickel such as natural water biological materials and rock and soil samples. The freedom from interferences is not sufficient however for the analysis of some industrial effluents ores high-alloy steels and nickel-base alloys etc.For these samples the addition of iron has been recommended in order to reduce the nickel interference further. i & * X It can be seen from Figs. 1 and 2 that the addition of 2 mg of iron to the solution for measurement in the reaction flask typically extends the range of interference-free determination for arsenic and selenium by a further factor of about 50 over the nickel concentration tolerable in the same acid in the absence of iron. With this increased freedom from inter-ferences arsenic and selenium can be determined in almost all types of samples directly against standards that contain only the same acid and iron concentrations. Nothing has been mentioned up to now about the possible mechanism by which iron acts as a releasing agent.The effect of iron is surprising at least at a first glance because iron itself causes signal depressions in the determination of all hydride-forming elements when present in higher concentrations. l5 In addition some workers suspect that possible synergistic effects occur when two or more interferents are present in a sample.12 Raptis et a1.21 found systematic errors in the determination of selenium in biological matrices. They assumed that the combined effect of several interfering elements is responsible for the signal depressions because no single interferent exceeded its tolerance limit in the solution used for the determination. No details on possible mechan-isms were given however. It is important to note that only trivalent iron can be applied as a releasing agent for the nickel interference whereas bivalent iron has no effect.From previous experiments we know that the interference of nickel in the determination of selenium probably involves capture and decomposition of the selenium hydride formed by finely dispersed metallic nickel. 22 We also know that the first step in this reaction is the preferential reduction of the nickel ions by the tetrahydrobor-ate and the formation of a nickel precipitate. Finally we could demonstrate that an increase in the acid concentration diminishes the influence of nickel probably because it increases the solubility (decreases the precipitation) of nickel. The interferences are more pronounced however when the evolution of the hydride is delayed and thus the precipitation of the metallic interferent is more complete.23 In an investigation of transition metal interferences in the determination of selenium22 we have already found that trivalent iron behaves differently from the bivalent ions of Groups VIII and IB.We have concluded from these results that trivalent iron interferes as the ion and that the first step in the reaction is the reduction of trivalent to bivalent iron which has only a very moderate influence in the ionic form. The hydrides of arsenic and selenium are formed in parallel to these reduction reactions and there is always a large enough excess of tetrahydroborate to leave this formation of the hydrides unaffected by the transition metals in solution. The electrochemical potentials24 of the possible reduction reactions involved when tetrahydroborate is added to a solution of trivalent iron and bivalent nickel are as follows: Fe3+ + e- Fe2+ +0.77 V -0.23 V Ni2f + 2e- Ni Fez+ + 2e- f Fe -0.41 v The high positive potential of the reduction of trivalent t ANALYST.MAY 1984 VOL. 109 579 bivalent iron suggests a preferential reduction of this species, which in addition is present in a large excess. Nickel will be reduced to the metal and precipitated only after all the iron has been reduced to the bivalent state. Iron ions in both valency states as well as nickel ions however have a considerably less pronounced influence on the arsenic signal compared with metallic nickel. This explains the releasing effect of trivalent iron in the presence of nickel; the reduction of bivalent iron to the metal has an even more negative potential than the reduction of nickel ions to the metal so that no effect can be expected for this species as can be confirmed by experiment.We believe that transition metal ions at no time during the reaction are depleting the available tetrahydroborate reagent. A “competition” is therefore very unlikely despite the comments of several workers and the volatile hydrides are formed in parallel to the reduction unaffected by the presence of transition metal ions in the solution. We have not seen a delay in the signal of the volatile hydrides by transition metals but only a reduction in signal height and/or signal area. The interference apparently starts only after the hydrides have been formed and is moderate when the interferent is present in the ionic form.Severe transition metal interferences are observed only when the interferent is precipitated from the solution as the metal and the mechanism probably involves capture and decomposition of the gaseous hydride at the freshly precipitated metal. All measures that prevent the precipitation of the reduced metal or delay it until after the volatile hydride has been evolved from the solution are suitable for extending the range of interference-free determination of the hydride-forming element. Working with higher acid concentrations or in mixed acids has proved very successful for many interferents in the hydride technique. The addition of a buffer that. owing to its electrochemical potential is reduced preferentially prior to the interferent appears to be another powerful analytical aid.It is essential however that the reduced species of the buffer does not interfere with the hydride evolution from the solution. Iron(II) which is formed in the reduction of the iron(II1) buffer has only a very moderate influence at high concentrations on arsenic and selenium.2‘ The pronounced interference of bivalent iron reported by several workers is in fact the interference from the precipitated metallic iron that is formed after the addition of the tetrahydroborate solution. This second reduction step however takes place only after all trivalent iron has been reduced to the bivalent form which is after the volatile hydrides have been evolved from the solution.Trivalent iron is therefore very useful as a buffer for controlling transition metal interferences specifically that of nickel in the determination of arsenic and selenium. Conclusion Severe transition metal interferences in the determination of arsenic and selenium using hydride generation are found only if the interferent is reduced and precipitated as the metal prior to or during the evolution of the hydride. All measures that prevent or delay the precipitation of the interferent can therefore be applied to decrease these interferences. Higher acid concentrations have been found to be effective in earlier work. The addition of up to 2 mg of iron(II1) to 10 ml of the solution for measurement has been found to increase further the range of interference-free determination of arsenic and selenium in the presence of nickel because it delays the precipitation of metallic nickel.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. References Smith A. E Analyst 1975 100 300. Pierce. F. D. and Brown H. R. Anal. Chern. 1976,48,693. Vijan P. N. and Wood G. R. Talanta 1976 23 89. Pierce F. D. and Brown H. R. Anal. Chem. 1977,49,1417. Meyer A. Hofer Ch. Tolg G. Raptis S. and Knapp. G., Fresenius Z . Anal. Chem. 296 337. Guimont J. Pichette M. and Rhkaume N . At. Absorpt. Newsl. 1977 16 53. Aggett. J and Aspell A. C. Analyst 1976. 101 341. Belcher R. Bogdanski S. L. Henden E. and Townshend, A . Analyst 1975 100 522. RubeSka I .. and Hlavinkova V. At. Absorpt. Newsl. 1979, 18 5. Kirkbright G. F. and Taddia. M. Anal. Chim. Acta 1978, 100. 145. Dornemann. A. and Kleist H. Fresenius 2. Anal. Chem., 1981. 305 379. Thompson M. Pahlavanpour B. Walton S. J . and Kirk-bright. G. F. Analyst 1978 103 705. Welz B. and Melcher M. Anal. Chim. Acta 1983 153 297. Fleming H. D. and Ide R. G. Anal. Chim. Acta 1976 83. 67. Welz B. and Melcher M. Spectrochim. Acta Part B 1981. 36 439. Welz B. Grobenski Z . and Melcher M in Koch. K. H and Massmann H. Editors “13. Spektrometertagung.” Walter de Gruyter Berlin New York 1981 p. 337. Welz B. and Melcher M. Wasser 1982 59 407. Welz B. and Melcher M. Wusser 1984 62 in the press. Welz B. and Melcher M. Anal. Chim. Acta 1981 131 17. Berndt. H. Willmer P. G. and Jackwerth E. Fresenius 2. Anal. Chem. 1979 296 377. Raptis S. Knap G. Meyer A. and Tolg G Fresenius Z. Anal. Chem. 1980 300 18. Welz B. and Melcher M. Analyst 1984 109 569. Welz B. and Melcher M. Analvst 1984 109 573. West R. C. and Astle M. J . Editors “Handbook of Chemistry and Physics,” Sixty-first Edition CRC Press. Cleveland. OH 1981 p. D155. Note-References 22 and 23 are to Parts 1 and 2 of this series, respectively. Paper A31315 Received September J2th 1983 Accepted December 13th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900577

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 581 Rapid Atomic-absorption Spectroscopic Analysis of Molybdenum in Plant Tissue with a Modified Carbon Rod Atomiser John W. Steiner and Kevin M. Ryan Department of Primary Industries Animal Research Institute 665 Fairfield Road Yeerongpilly Brisbane, 4 105 Australia ~~~~ A modified carbon rod atomiser was used for the determination of molybdenum in plant material. The sequential introduction of hydrogen methane nitrogen and oxygen into the atomiser eliminated the tedious and time-consuming chelation and solvent extraction procedure commonly used for the pre-treatment of samples. Methane also effected continuous re-coating of the atomiser with pyrolytic graphite. An optical device increased the precision of the injection of the samples. The improvement in precision accuracy, sensitivity rapidity and repeatability gained through the use of this modification was demonstrated by the determination of molybdenum in certified samples.Keywords Molybdenum determination; in situ gaseous pre-treatment of plant samples; matrix modification; addition of gases to atomiser; atomic-absorption spectroscopy with electrothermal a tom isa ti0 n Molybdenum is an essential element in plant and animal nutrition. The molybdenum requirement of plants varies with species so that deficiency is exhibited in some when its concentration is as high as 0.28 mg kg-1 and in others when its concentration is as low as 0.03 mg kg-1.1 Animals also require low levels of molybdenum in their diets for normal growth and development .2 Consequently molybdenum is frequently determined in plant tissue and requires a rapid reliable and sensitive method.Although there are many analytical techniques available for the determination of molybdenum in biological materials, recent interest has focused on the use of AAS with elec-trothermal atomisation. There is however a major problem in the use of this technique. Determination of molybdenum in biological material is extremely susceptible to matrix interfer-ences with loss of sensitivity.3 Several attempts have been made to overcome this problem. These include chelation and solvent extraction,4.5 wet ashing with complete removal of residual oxidising acids6 and correlation of the rate of loss of sensitivity to the number of firings.7 All of these methods are tedious time consuming and/or susceptible to error.In this paper we describe a new method for eliminating matrix interference and preventing loss in sensitivity. This method uses different gases for the in situ pre-treatment of digested plant samples directly in the atomiser. The effective-ness of this method is illustrated by the analysis of certified orchard leaves. Experimental Apparatus A Varian Techtron AA-175 spectrophotometer equipped with a modified carbon rod atomiser (CRA-90) was used.8 Reagents Gases The gases (high-purity nitrogen; industrial dry hydrogen; and medical grade oxygen) were obtained from Commonwealth Industrial Gases (CIG) Ltd. Brisbane. Methane (99.9% pure) was obtained from Matheson. Standard solutions The concentrations of the working standards ranged from 0.02 to 0.06 mg 1-1 of molybdenum.These were prepared daily from concentrated stock solutions (BDH Chemicals Ltd.) by dilution with 1 + 4 HCl. Standard reference material Freeze-dried orchard leaves (SRM 1571) were obtained from the US National Bureau of Standards. The certified concen-tration for molybdenum was 0.3 k 0.1 mg kg-1 (95% confidence limit). Procedure Glassware cleaning New glassware digestion tubes and caps were initially decontaminated by soaking in 10% nitric acid (Univar grade, Ajax Chemicals) (re-distilled before use). They were subse-quently boiled for 1 h in 3% non-ionic detergent (Decon-90, Decon Laboratories) and then rinsed in glass-distilled water and then with a final rinse of doubly glass-distilled water.Preparation of samples Plant samples were finely ground through a 1-mm screen in a stainless-steel mill (Glen Creston Laboratory) and were then stored in clean glass jars. Samples were dried in a forced air oven at 100 "C for 12 h prior to digestion. Digestion of SRM 1571 and test samples The SRM 1571 and test samples were prepared by accurately weighing suitable amounts (0.5-0.7 g) into digestion tubes (150 x 25 mm screw cap culture tubes Pyrex No. 9826) and adding 5 ml of the digestion acid (1 + 4 V/VHC104 + HN03). The tubes were allowed to stand for 4 h to complete frothing before transferring into a cool sand-bath which was heated slowly to a temperature of 90 "C and then kept at this temperature for 90 min. The digestion was completed by heating the tubes with a Bunsen burner until dense white fumes of perchloric acid were liberated and less than 0.5 ml of liquid remained in the tubes.(If the digest began to char during digestion 1 ml of re-distilled nitric acid was added.) After cooling the digests were diluted to 15 ml with 1 + 4 HC1. Instrument operation and optimisation The optical device with the transverse gas jets was lowered into position and the selected gases allowed to flow through the interior of the atomiser tube. The atomiser was pre-heated to 93 "C before manually injecting the sample. The injection rate was carefully monitored with the optical device and an equilibrium between the injection and evaporation rates was maintained. At the transition from the drying to the ashing stage the tim 582 ANALYST MAY 1984 VOL.109 elapsed was recorded with a stop watch. After 16 s the optical device and transverse gas jets were rapidly slid out of the atomisation compartment and the instrument was changed from total- to corrected-absorption mode. The absorption signal was recorded with the digital readout meter using peak mode with continuous background correction. Exact instrumental conditions are listed in Table 1. The optimum flow-rates of the gases used for the in situ gaseous pre-treatment of the sample and the continuous re-coating of the atomiser are listed in Table 2. Results and Discussion Initial Observations The effects of using different gases during the injection, drying ashing and atomisation stages for the determination of molybdenum were investigated using SRM 1571 and a digested plant sample.For each gas or combination of gases an experiment consisting of 65 firings of the one tube was conducted to monitor changes in accuracy precision and repeatability. The experimental procedure was as follows. 1. Firings 0-16 initial calibration of instrument. 2. Firings 17-25 determination of SRM 1571. 3. Firings 26-35 determination of digested plant sample. 4. Firings 36-51 re-calibration of the instrument. Table 1. Instrumental conditions Parameter Value Volume of injection/$ . . . . . . . . 5 Drying temperature/'C . . . . . . . . 93 Ashing temperature/'C . . . . . . . . 640 Drying time/s . . . . . . . . . . 60 Ashing time/s . . . . . . . . . . 40 Hold time/s .. . . . . . . . . . . 1.0 Ramp rate/'C s-1 . . . . . . . . . . 500 Residence time of transverse gas jets/s . . 16 PM voltage/mV . . . . . . . . . . 406 Atomising temperature/"C . . . . . . 2480 Table 2. Optimum flow-rates (mean i s.d.) of gases for the determination of molybdenum Gas introduction systems Flow-ratelm1 min-l Sheath . . . . . . . . Nitrogen 8823 +81 Verticalgasjet . . . . . . Methane 5.8 L 0.01 Transverse gas jets . . . . Hydrogen 240 k 1.3 Hydrogen 340 k 8.2 Methane 20.4 f 0.12 Oxygen 157 k 1.2 5. Firings 52-60 re-determination of SRM 1571. 6. Firings 61-65 re-determination of the digested plant sample. The concentration of molybdenum in the SRM 1571 determined during the two trial periods (viz. firings 17-25 and 52-60) are shown in Table 3 and expressed as 95% confidence ranges.This allowed a comparison with the measured and certified concentration of molybdenum in SRM 1571. Data presented in Table 3 show that the appropriate addition and manipulation of gases in the atomiser when synchronised with the instrumental process virtually elimi-nates loss of sensitivity with the consequent improvement of accuracy. When the optimum conditions i.e. line 8 in Table 3 were used no significant differences were noted in accuracy, precision and sensitivity. It was found that the HC104 in the digested material did eventually destroy the pyrolytic coating of the carbon tube between firings 80 and 90 and a rapid loss in sensitivity resulted. Evaluation of New Procedure Table 4 shows the accuracy and precision obtained with the modified procedure when SRM 1571 was analysed in quadrup-licate on each of 4 d.Further assessments of accuracy and precision were repeated regularly approximately every 100 determinations during the course of over 2 000 determina-tions. The recoveries of the analyte added to the plant samples before digestion are shown in Table 5. The new procedure was used each time and no variations were observed. Table 6 shows a comparison of analytical parameters determined with and without using the modified procedure. The absolute error and the relative standard deviation (r.s.d.) were significantly lower using the modified procedure. There was also an improvement in the correlation coefficient of the calibration graph and the sensitivity when the modified procedure was used.The advantages of the modified procedure depend on using the correct mixing ratios of the gases and introducing them sequentially into the atomiser. Each gas plays a distinct role in the operation of the instrument. Methane and hydrogen function as reducing agents during atomisation of the sample while methane also provides carbon for pyrolytic re-coating of the atomiser. Flow-rates of both gases must be controlled accurately to those shown in Table 2. Hydrogen flow via the sheath must be set in the range 33CL350 ml min-1 while methane flow via the sheath must be set in the range 19.9-20.7 ml min-1 and through the vertical gas jet in the range 5.6-6.4 ml min-1. Flow-rates of hydrogen outside the above range resulted in an unacceptable deterioration of precision.When methane flow-rates were higher than the optimum range only Table 3. Experimental conditions and results of the repetitive determination of molybdenum in SRM 1571 containing 0.3 L 0.1 mg kg-1 of molybdenum Experimental conditions Gas introduction systems Vertical Transverse Sheath jet jets - - NZ . . . . . . . . . . . . N2 + H . . . . . . . . . . - -- - N + Hz + CH4 . . . . . . . . N + H2 + CH4 N2 + HI + CH4 CH4 H? NZ + H + CHJ CH4 0 2 - . . . . . . . . CH4 . . . . . . . . . . . . . . . . Nz + HZ + CH4 . . . . . . . . CH4 H ? + O ? N + HZ + CH CH4 Hz+02 . . . . . . . . * Mean + 95% confidence limit. t Firing interval 17-25 inclusive (based on n = 9). $ Firing interval 52-60 inclusive (based on n = 9).Estimated concentration/ mgkg 1 * Optical device Period I t Period 2 t - 0.326 k 0.094 0.075 2 0.017 - 0.251 k 0.024 0.114 i 0.012 - 0.148 i 0.030 0.155 i 0.055 - 0.081 k 0.023 0.081 k 0.018 - 0.086 L 0.007 0.147 i 0.015 - 0.228 & 0.023 0.155 2 0.055 - 0.256 ? 0.009 0.248 i 0.010 $ 0.280 i 0.009 0.282 L 0.00 ANALYST MAY 1984 VOL. 109 5 83 ~~ ~~ ~ ~ ~~ Table 4. Analysis of molybdenum in orchard leaves (SRM 1571) Certified No. of No. of analyses Mo contenttl Mo found+/ 4 4 0.3 f. 0.1 0.281 f. 0.007 sub-samples per sub-sample* mg kg-1 mg kg-* Performed on separate days. t Mean 95% confidence limit. Table 5. Recovery of molybdenum added to plant samples No. of No. of analyses Amount of Mo Recovery, 4 4 0.01 104.5 k 5.5 4 4 0.02 95.8 k 2.9 4 4 0.03 100.9 f.2.0 sub-samples per sub-sample* added/pg Yo ?. s.d. * Performed on separate days. Table 6. Comparison of performance using the modified and unmodified procedure Modification Analytical parameter With Without Absoluteerror*/mgkg-1 . . . . 0.03 -0.14 R.s.d.,% . . . . . . . . . . 3.1 12.7 Slope (b,) . . . . . . . . . . 1.930 2.225 Correlationcoefficient(r) . . . . 0.9987 0.989 3 measured and the certified value of the SRM 1571. * The absolute error is defined as the difference between the The presence of these gases in the atomiser facilitated a uniform injection - drying stage and improved the evaporation - combustion rate of various components during ashing of the sample. These improvements appeared to be associated with the presence or absence of a white layer of material in the lower section of the atomiser immediately after ashing commenced.The removal of this material in the absence of hydrogen and oxygen was impossible even at an ashing temperature of 1 050 “C. However it was completely removed within 6 s by the introduction of hydrogen and oxygen at a temperature of 640 “C. It is reasonable to assume that the identity of this material might hold the key for selecting a suitable chemical modifier to be used instead of the gases. Conclusion The method for the determination of molybdenum in plant tissue proposed in this paper is rapid sensitive free from the loss of sensitivity and matrix interferences usually associated with the determination of molybdenum by atomic-absorption spectroscopy with electrothermal atomisation.Chemical pre-treatment directly in the atomiser using various gases, synchronised with the over-all operation of the instrument, has resulted in this improvement and made the search for a chemical modifier unnecessary. 1. 2. slight deterioration in accuracy was observed. However when the methane flow-rate was below optimum effects on accuracy were more pronounced. For example when the methane flow-rate was 4.6 ml min-1 (optimum 5.8 ml min-1) molybdenum results were reduced by 12% for the first period of firings (Table 3) and 21% for the second period. Slight variations in the methane flow-rate in the vertical jet caused more profound changes in accuracy than when the flow-rate of methane in the sheath gas was varied. Further the mixing ratios of hydrogen and oxygen added to the atomiser via the transverse gas jets were also critical. 3. 4. 5. 6. 7. 8. References Jarrel W. M. Page A. L. and Elgeewi A. A. Residue Rev., 1980 74 1. Underwood E. J. “Trace Elements in Human and Animal Nutrition,” Fourth Edition Academic Press New York 1977. Studnicki M. Anal. Chem. 1979 51 1336. Khan S. U. Cloutien R. O. and Hidiroglou M . 1. Assoc. Off. Anal. Chem. 1979 62 1062. Horak O. Bodenkultur 1977 28 405. Newman D. R. and Munshower F. F. Anal. Chim. Acta, 1981 123 325. Wilson D. O. Comrnun. Soil Sci. Plant Anal. 1979,10,1319. Steiner J . W. and Kramer H. L. Analyst 1983 108 1051. Paper A311 99 Received July 5th 1983 Accepted October 17th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900581

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST MAY 1984 VOL. 109 585 Use of Aliquat-336 for the Extraction of Cadmium from Aqueous Solutions Kate Grudpan" and Colin G. Taylor Department of Chemistry and Biochemistry The Polytechnic Byrom Street Liverpool 13 3AF UK A procedure is described for the concentration of cadmium from a large volume of aqueous solution. Cadmium is extracted from a solution containing chloride ion into a solution of Aliquat-336 in carbon tetrachloride and stripped from the extract into aqueous perchloric acid. Conditions for quantitative extraction and stripping have been established by use of a radioactive tracer. Recoveries of cadmium standards added to fresh estuarial or sea water have been measured by flame atomic-absorption spectrometry. Keywords Cadmium extraction; Aliquat-336; liquid - liquid extraction; radioactive tracer; atomic-absorption spectrometry Both solvent-extraction and ion-exchange techniques have been used for the concentration of trace amounts of cadmium from aqueous solutions and for the separation of cadmium from other metals prior to its determination usually by spectrophotometry.Cadmium in sea water has been determined by solvent extraction as the dithizonatel or by enrichment using a basic anion-exchange resin .2 Other methods have been developed that employ solvent extraction or ion exchange for separation and/or pre-concentration and atomic-absorption spectrometry (AAS) (flame or graphite furnace) for the determination of cadmium in bl00d,3 organic matter ,4,5 foodstuffs6 or natural waters.738 Disadvantages inherent in these methods include lengthy procedures,' an excessively long sorption time2 and reagent in~tability.6.~ Cadmium in sea water has also been determined by carbamate extraction (tetramethylenedithiocarbamate and diethyldithiocarbamate) into Freon TF (1,1,2-trichloro-l,2,2-trifluoroethane)lO or into chloroform," followed by stripping into nitric acid.A liquid ion-exchange resin (Amberlite LA-2) has been recommended for use in the determination of sub-p.p.m. levels of cadmium in organic matter following wet digestion of the sample,4 and a number of other ion-exchange extraction steps in cadmium determinations has been reported. Extractants include trioctylamine in cyclohexane,12 tri-benzylamine in dichloroethanel3 and methyltrioctylam-monium chloride (Capriquat) in chloroform14; however none of these has been applied to the determination of cadmium in natural samples.The quaternary ammonium salt tricaprylmethylammonium chloride (Aliquat-336) has been shown to be a versatile extractant for metals that are capable of existing as complex anionic species. Aliquat-336 has been used in the detection of gallium and indium,15 in the determination of lead-210 in bone ash16 and in the determination of chromium,l7J8 molybdenum,19 gold20 and cobalt.21 The salt has also been used for the separation of lead-212 from natural thorium22 and for the separation of cadmium and zinc,23-25 cadmium and mercury26 and actinide - lanthanide element~27-~9 from each other. The availability of Aliquat-336 and the simple rapid procedures described for its use led us to examine the reagent as a means for concentrating cadmium from aqueous solutions containing p.p.b.levels of the metal to levels of 0.1 p.p.m. or higher suitable for determination by flame AAS. * Present address Department of Chemistry Faculty of Science, The University Chiang Mai Thailand. Extraction and stripping procedures have been followed radiometrically. Experimental and Results Reagents All reagents were of analytical-reagent grade except where otherwise stated. Aliquat-336. This is a mixture of high relative molecular mass quaternary ammonium salts containing Cs-Clo straight carbon chains with C8 predominating; the average relative molecular mass is 442 (General Mills Chemicals Minneapolis, MN USA).Aliquat-336 solution 0.1 M in carbon tetrachloride. The solution was purified by shaking for successive 10-min periods with aqueous solutions of 20% sodium hydroxide 6 M hydrochloric acid and 20% sodium chloride. The volume of the aqueous phase was always one fifth of the volume of the Aliquat solution.21 Other Aliquat-336 solutions were pre-pared by appropriate dilutions of this stock solution with carbon tetrachloride, Standard cadmium solution 1000 p.p.m. A stock solution was prepared by disso1vir.g 2.282 g of cadmium sulphate (3CdS04.8H20) in 90 ml of water. Sulphuric acid (2 M) was added drspwise until a clear solution was obtained which was then diluted with water to 1 1. The stock solution was standardised by titration with ethylenediaminetetraacetic acid in the usual way.30 Other standard solutions of cadmium were prepared as required by making appropriate dilutions of the stock solution.Cadmium-109 as chloride in 0.1 M hydrochloric acid 1.4 pg ml-1 16 MBq ml-1. Radiochemical Centre Amersham. Radioactively labelled cadmium solutions of appropriate specific activity and concentration were prepared by mixing portions of the radionuclide with portions of the correspond-ing standard solution and diluting to the required concentra-tion. Ethylenediamine 0.2 and 0.1 M in water. Perchloric acid 0.4 and 0.2 M in water. Ethylenediamine perchlorate 0.1 M. Equal volumes of ethylenediamine (0.2 M) and perchloric acid (0.2 M) were mixed. Ethylenediamine - ethylenediamine perchlorate buffers. Ethylenediamine (0.1 M) was added to ethylenediamine perchlorate (0.1 M) until the required pH was reached.Synthetic sea water.31 The following were dissolved in order in distilled water and the solution was diluted to 5 1: NaCl(l17.39 g); MgC12.6H20 (53.18 g); SrC12.6H20 (0.20 g); H3B03 (0.13 g); Na2S04 (19.59 g); CaC12 (5.51 g); KC1 (3.32 g); KBr (0.48 8); and NaHC03 (0.96 g) 586 ANALYST MAY 1984 VOL. 109 Apparatus Radiometric measurements were performed using either an Ortec Spectrometer Model 3030 or an Ekco Type N529C gamma scintillation counter both fitted with well-type thallium-activated sodium iodide crystals. AAS measurements were performed using an air - acetylene flame with either a Perkin-Elmer 290B or an IL 151 spectrometer. Concentration of Cadmium by Extraction and Stripping Preliminary radiometric experiments showed that 1 pg of cadmium in 1 1 of synthetic sea water could be extracted into 10 ml of 0.1 M Aliquat-336 in carbon tetrachloride and stripped from the extract into 5 ml of 5% aqueous ethylenediamine, with an over-all recovery of 90%.The effects of pH concentration of chloride ion and concentration of Aliquat-336 on the extraction of cadmium were examined by the following procedure. Solutions containing 0.1 p.p.m. of radioactively labelled cadmium ion and various concentrations of sodium chloride were prepared. During preparation the pH of each solution was adjusted to a measured value by the addition of hydrochloric acid or sodium hydroxide. A portion (10 ml) of each solution was extracted with 10 ml of a solution of Aliquat-336 in carbon tetrachloride.Extractions were carried out in stoppered centrifuge tubes by shaking for 3 min and centrifuging for 5 min. The extent of extraction of cadmium was obtained by measuring the activities of known portions of the aqueous phase before and after the extraction. The results in Table 1 indicate that cadmium is quantita-tively extracted from aqueous solutions at pH 2-9 containing a 0.5 M concentration of chloride ion into a 0.1 M solution of Aliquat-336 in carbon tetrachloride. The results further indicate that at pH 4 the extraction into 0.1 M Aliquat-336 is quantitative over a range of chloride ion concentrations from 0.1 to 0.7 M and also that at pH 4 and a chloride ion concentration of 0.5 M the extraction is quantitative over a range of Aliquat-336 concentrations from 0.01 to 0.1 M.For subsequent work a pH of 4 a chloride concentration of 0.5 M and an Aliquat-336 concentration of 0.1 M were usually selected. In order to study the stripping of cadmium from the organic phase a stock solution of cadmium and Aliquat-336 in carbon tetrachloride was prepared as follows. An aqueous solution (100 ml) containing 0.1 p.p.m. of labelled cadmium ion and 0.5 M in sodium chloride was equilibrated for 5 min with 100 ml of 0.1 M Aliquat-336 in carbon tetrachloride. The solvent phase was separated and before storage was centrifuged for 5 min to remove trace amounts of water. Preliminary experiments indicated that cadmium could be stripped from this stock solution by aqueous ethylenediamine of any concentration between 0.003 and 0.375 M with a recovery of 99% or better.A series of experiments were then carried out with strips containing ethylenediamine - ethylenediamine perchlorate buffer or perchloric acid. The stock solution and the strip (10 ml of each) were equilibrated using the procedure already described. The extent of removal of cadmium was obtained by measuring activities of portions of the stock solution before and after stripping. The result quoted above and those displayed in Table 2 indicate that solutions of ethylenediamine or perchloric acid, at certain concentrations separately or in admixture can strip cadmium quantitatively from a carbon tetrachloride phase containing Aliquat-336.The work so far described demonstrates the feasibility of concentrating microgram amounts of cadmium quantitatively from a large volume of aqueous solution by simple extraction and stripping. In order to verify this a series of aqueous Table 1. Extraction of cadmium from aqueous solution effect of pH, chloride concentration and Aliquat-336 concentration. Equal volumes of aqueous solution and extractant taken PH Chloride/M Aliquat-336/~ Recovery of Cd % 2 0.5 0.1 99.9 4 99.9 6 99.9 7 99.9 9 99.8 4 0.1 0.1 99.7 0.3 99.9 0.7 99.8 4 0.5 0.01 99.3 0.05 99.7 Table 2. Stripping of cadmium from a carbon tetrachloride extract: effect of strip composition. Extractant 0.1 M Aliquat-336 in CC1,; equal volumes of extractant and strip ConcentratiodM PH 9.5 10.5 11.0 2.0 1.1 1 .o 0.1 Buffer Perchloric acid Recovery of Cd % 0.1 - 99.9 0.01 - 43.2 0.1 - 99.9 0.01 - 99.8 0.1 - 99.7 0.01 - 99.9 - 0.01 3.1 - 0.08 31.5 - 0.1 99.7 - 0.8 99.9 solutions (100 ml) containing 3 pg of cadmium with a sodium chloride concentration of 0.5 M and/or a hydrochloric acid concentration of 0.01-0.6 M were equilibrated successively with 0.1 M Aliquat-336 in carbon tetrachloride (5 ml) and two portions (5 mi each) of carbon tetrachloride.The extract was combined with the washings and the cadmium was stripped by equilibrating with 5 ml of 0.4 M perchloric acid. After centrifugation the aqueous phase was separated and diluted to 10 ml with water washings. Cadmium was determined in this diluted aqueous phase by flame AAS using standard cadmium solutions in 0.2 M perchloric acid for calibration.Some of the original aqueous solutions were acidified with hydrochloric acid in order to study the possibility of loss of cadmium by sorption. Perchloric acid was chosen for stripping in preference to the equally effective ethylenediamine firstly because perchloric acid is more readily available than ethyl-enediamine in a high state of purity and secondly because some evidence was found of slow sorption of cadmium from solutions of ethylenediamine on to glass surfaces. The results in Table 3 show reproducible quantitative recoveries of cadmium under all the conditions examined. Recommended Procedure for the Concentration of Cadmium from Natural Water Samples The following procedure has been developed based on the results so far reported for the concentration of cadmium present in contaminated natural waters from the 1 p.p.b.to the 0.1 p.p.m. level in preparation for determination by flame AAS. All operations should be carried out in a clean area in order to avoid any contaminatiolr from trace amounts of cadmium in the environment. All solutions should be pre-pared with doubly distilled water and all glassware soaked in 0.1 M nitric acid and rinsed with doubly distilled water before use ANALYST MAY 1984 VOL. 109 5 87 Table 3. Recovery of 3 pg of cadmium from 100 mt of aqueous solution following extraction and stripping. Extractant .5 ml of 0.1 M Aliquat-336 in CCI,; strip 5 ml of 0.4 M HC104 Aqueous solutionh Cadmium Sodium chloride Hydrochloric acid recovered/pg 0.5 0.01 3.0 0.5 0.12 2.9,3.0,3.0 - 0.36 2.9 - 0.60 2.9,2.9,2.9 Table 4.Determination of cadmium recovery from spiked samples of natural waters following the recommended concentration procedure. Volume of sample = 1 1; volume of strip = 10 ml. Consistent blanks of 0.01 p.p.m. in the strip were obtained Cadmium found* Cadmium added In strip In sample. Recovery, Sample p.p.b p.p.m. p.p.b. % Atlantic sea water . . - n.d. n.d. 11.5 - n.d. 1 0.11 1.1s 1 0.12 - 0.04 1 0.14 1.35 1 0.13 - 0.01 1 0.13 1.25 1 0.12 River Mersey water - 0.04 0.40 9.5 Liverpool tap water - 0.02 0.1.5 110 * Corrected for blank; n.d. = not detected. The concentration of chloride ion in the water sample should be about 0.5 M.The salinity of the sample if not known should be determined titrimetricaliy using silver nitrate and an appropriate amount of sodium chloride (analytical-reagent grade) added to the sample if necessary. Into a separating funnel fitted with a Teflon tap place 10 ml of concentrated hydrochloric acid. Wet the walls of the funnel with the acid and add 1 1 of the water sample followed by the calculated amount of solid sodium chloride. Add 10 ml of carbon tetrachloride and equilibrate the mixture by shaking the funnel vigorously for 5 min. Allow the phases to separate and discard the carbon tetrachloride phase together with any solid that has appeared at the interface. Add to the aqueous phase 5 ml of 0.1 M Aliquat-336 in carbon tetrachloride.Equilibrate the mixture as before allow the phases to separate and run the carbon tetrachloride phase into 5 ml of 0.4 M perchloric acid contained in a centrifuge tube that can be fitted with a ground-glass stopper. Repeat the extraction with a further 5 ml portion of Aliquat - carbon tetrachloride and combine the second extract with the first. Wash the aqueous phase with two successive portions ( 5 ml) of carbon tetrachloride shaking for 2 min each time and combine the washings with the extracts in the tube. Shake the contents of the tube vigorously for 5 min to strip the cadmium into the perchloric acid. Centrifuge the tube for 5 min and transfer the aqueous phase to a 10-ml calibrated flask. Wash the surface of the carbon tetrachloride phase with successive 1-ml portions of water and use the washings to make up the perchloric acid strip to a volume of 10 ml.Carry out a blank by repeating the whole procedure using 1 1 of doubly distilled water containing sodium chloride at the same concentration as that in the sample. If no sodium chloride has been added to the water sample (e.g. estuarial river water) carry out the blank by equilibrating 10 ml of Aliquat - carbon tetrachloride with 5 ml of perchloric acid, following the stripping procedure described above. Application The recommended procedure has been applied to the deter-mination of cadmium by flame AAS in samples of sea water, estuarine river water and tap water spiked with standards. The results in Table 4 indicate that the procedure is capable of yielding consistent results with quantitative recoveries.Reproducible blanks were obtained throughout the deter-mination the blank value (0.1 p.p.m. of cadmium in the strip) being close to the limit of detection for cadmium by flame AAS under the instrumental conditions employed. Conclusion The moderate stability of the tetrachlorocadmiate ion [CdC14]2- (log p = 3) allows cadmium when present in saline aqueous solution to exist in this anionic form and to be quantitatively extracted into carbon tetrachloride as an ion-association compound with the cation of Aliquat-336. Cadmium when extracted in this way can be quantitatively stripped into an aqueous solution containing ethylenediamine , perchloric acid or an ethylenediamine - perchloric acid buffer.Both of these reagents release cadmium from the carbon tetrachloride phase by converting the metal into a cationic species the former by complexing cadmium as the stable bisethylenediaminocadmium ion [Cd(NH2C2H4NH&l2+, (log = 10) and the latter by the displacement reaction in equation (1) followed by equation (2). {(Aliquat)2[CdC14]}cc14 + (2C104-)aq = (2 AliquatC104)cc14 + ([CdCI4]2-),, (1) These reactions form the basis of the procedures that have been examined in this work for the concentration of cadmium from a large to a small volume of aqueous solution prior to its determination by flame AAS. Perchloric acid rather than ethylenediamine or ethylenediamine - ethylenediamine per-chlorate is recommended as the preferred strip solute because of the observed tendency of cadmium to be sorbed on to glass surfaces under the alkaline conditions created by ethylene-diamine.The primary purpose of this work was to demonstrate the use of Aliquat-336 to separate and concentrate cadmium from large volumes of aqueous solution. It is recognised that limits of detection that are lower than those exhibited in this work would need to be achieved if the method were to be applied the determination of cadmium in most environmental water samples including sea water where the cadmium concentra-tion can range from 0.02 to 0.12 p.p.b.11 Such levels could be reached by the use of electrothermal atomisation but substantial improvements in the blanks would also be neces-sary. Present results indicate that reductions in reagent blanks could be brought about without significant loss of extraction efficiency by reducing the concentration of sodium chloride added to non-saline samples from 0.5 to 0.1 M and by reducing the concentration of Aliquat-336 in the carbon tetrachloride extractant from 0.1 to 0.01 M (Table 1).In conclusion the recommended procedure illustrates the value of liquid ion exchangers such as Aliquat-336 as reagents for the extraction concentration and separation of metals. [CdC14]2- + 6H20 = [Cd(H20)h12+ + 4C1- (2) The authors are grateful to the Liverpool Education Authority for supporting this work to Henkel Corporation (formerly General Mills Chemicals Inc.) Minneapolis MN USA for a gift of Aliquat-336 and to the Department of Oceanography, Liverpool University for sea water samples 588 ANALYST MAY 1984 VOL.109 Petrow G. H. and Cover A. Anal. Chem. 1965. 37 1659. Kinhikar G. M. and Dara S. S. Talanta 1974 21. 1208. Adam J. and Pribil R. Talanta 1971 18. 91. Adam J . and Pribil. R. Talanta 1971 18 349. Adam J . and Pribil R. Talanta 1971 18 405. Adam J. and Pribil R. Talanta 1971 18 733. Bogen D. C. and Kleinman M. T. Analyst 1967 92 611. Moore F. L. Environ. Lett. 1975 10 37. Moore F. L. Sep. Sci. 1975 10 489. McDonald C. W. and Rhodes T. Anal. Chem. 1974. 46, 300. McDonald C. W. and Moore F. L. Anal. Chem. 1973.45, 983. Moore F. L. Anal. Chem. 1964 36 2158. Moore. F. L. Anal. Chem. 1965 37 1235. Moore F. L. Anal. Chem. 1966 38 510. Vogel A. I. “A Text book of Quantitative Inorganic Analysis,” Third Edition Longmans London 1961 p.444. Williams W. J. and Gillam A. H. Analyst 1978 103 1239. Paper A31327 Received September 18th) 1983 Accepted December 7th) 1983 1. 2. 3. 4. 5 . 6. 8. 9. 10. 11. 12. 13. 14. 15. References Mullin J. B. and Riley J. P. J. Mar. Res. 1959 15 103. Brooks R. R. Analyst 1960 85 745. Delves H. T. Shepherd G. and Vintner P. Analyst 1971, 96 260. Analytical Methods Committee Analyst 1969 94 1153. Analytical Methods Committee Analyst 1975 100 761. Evans W. H. Read J . I. and Lucas B. E. Analyst 1978, 103 580. Kuwata K. Kisatomi K. and Kasegawa K. At. Absorpt. Newsl. 1971 10 111. Riley J. P. and Taylor D. Anal. Chim. Acta 1968 40 479. Dornemann A. and Kleist H. Analyst 1979 104 1030. Danielsson L. G. Magnusson B. and Westerlund S . Anal. Chim. Acta 1978 98 47. Bruland K. W. Franks R. P. Knauer G. A. and Martin, J. H. Anal. Chim. Acta 1979 105. 233. Ghersini G. and Marriottoni S. Talunta 1971 18 442. Green H. Talanta 1973 20 139. Hideo A. Hiroshi K. and Toshikazu T Bunseki Kaguku, 1978 27 449. Pribil R. and Veseley V. Talanla 1970 17 801. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31

ISSN:0003-2654    

DOI:10.1039/AN9840900585

出版商:RSC    

年代:1984

数据来源: RSC