ISSN:0003-2654    

DOI:10.1039/AN99116BP023

出版商:RSC    

年代:1991

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN99116FX025

出版商:RSC    

年代:1991

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN99116BX027

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991, VOL. 116 68 1 Occurrence, Handling and Chromatographic Determination of Pesticides in the Aquatic Environment A Review Damia Barcelo Environmental Chemistry Department, CID-CSlC, c/Jordi Girona 18-26, 08034 Barcelona, Spain Summary of Contents I nt rod uct i on Pesticides in Soil and Groundwaters Pesticides in Aquatic Organisms Sample Handling Water Samples Off-line methods On-line methods Extraction Clean-up Sediment and Biota Samples Chromatographic Analysis Common detectors Post-column reactors Mass spectrometric detection Supercritical fluid chromatography Miscellaneous Gas Chromatographic Met hods Liquid Chromatographic Met hods Conclusions References Keywords: Pesticide residue analysis; occurrence, chromatography; mass spectrometry; re view sample extraction and clean-up; gas and liquid Introduction Several hundred pesticides of different chemical nature are currently widely used for agricultural purposes throughout Europe and the USA.Some are substitutes for the organo- chlorines which were banned after evidence of their toxicity, persistence and bioaccumulation in the environment became available. According to a recent report published by the US Environmental Protection Agency (EPA), an over-all 500 000 tonnes of pesticides were used in 1985.1 In Europe pesticide usage is difficult to estimate; however, the UK reportedly used 14000 tonnes per year during the period 1980-1983.2 As far as details for specific pesticides are concerned, world-wide consumption of malathion and atrazine in 1980 amounted to 24000 and 90000 tonnes, respectively.3.4 In the Mediter- ranean countries 2100 tonnes of malathion (active ingredient) were sprayed during the same period (1980), 40% in Spain alone, versus 9700 tonnes in Asia.The EC countries also expend malathion at a rate of nearly 2000 tonnes per year.3 Finally, France consumed an estimated 5000 tonnes of atrazine in 1985,' whereas in 1986 in Switzerland only 120 tonnes of the same pesticide were used.4 Because of their widespread use, pesticides are currently detected by determination of their residues in various environ- mental matrices, such as soil, water and air. Hence, atrazine, one of the herbicides most widely used in the USA and European countries over the last 30 years, and employed for both pre- and post-emergence weed control among crops of corn, wheat, barley and sorghum, also on railways and roadside verges,s has, not surprisingly, been detected in surface and groundwaters throughout the world.It has been detected in some US groundwaters at concentrations in the range 0.1-3 pg 1-1,1-6J and also in surface waters in Canada,* Italy,9 Spain,lO the UK2 and France,s and in lake waters in Switzerland .4 Environmental studies should allow the transport and fate of pollutants, such as pesticides, to be predicted and the observed levels related to patterns in the usage and physico- chemical properties. Some workers have already pointed out that analytical chemistry should play a major role in environ- mental research." This review is intended to assist workers involved in monitoring pesticides in environmental matrices by: (i) addressing the behaviour of pesticides in soils, waters and biota; (ii) considering sample handling aspects including a comparison of off-line and on-line clean-up methods; and (iii) discussing some analytical aspects, particularly the choice of gas or liquid chromatographic techniques in combination with mass spectrometry (MS) for the unequivocal characterization of pesticides and their breakdown products in environmental matrices. This last point is of particular significance on account of the variety of chemical structures found in the different pesticide groups ( e .g . , phosphate, acidic and amphoteric). Pesticides in Soil and Groundwaters Data on the environmental fate of pesticides are required in order to determine the potential of a pesticide to reach groundwater including information on its hydrolysis, photol- ysis, aquatic metabolism, leaching and field dissipation.According to Aharonsons a pesticide can reach groundwater if its water solubility is greater than about 30 mg 1-I, its adsorptivity ( Koc) ( Koc = partition coefficient between soil organic carbon and water) is less than 300-500 cm3 g- 1, its soil half-life is longer than about 2-3 weeks, its hydrolysis half-life682 ANALYST, JULY 1991, VOL. 116 is longer than approximately 6 months and its photolysis half-life is longer than 3 d. The potential of a pesticide to contaminate groundwaters can also be assessed from the field conditions denoted by the acronym DRASTIC, an environ- mental index that includes seven key hydrogeological factors, namely, Depth to the water table, Recharge of groundwater by precipitation and irrigation, Aquifer medium, Soil medium, Topography (slope), Impact of the vadose zone, and Conductivity (hydraulic) of the aquifer.DRASTIC scores of greater than 150, recharge values larger than 25 cm per year, and thin soils with high nitrate levels (a few ppm of natural background), are indicative that pollution of groundwater by pesticides is feasible.6 One of the key parameters in determining groundwater pollution via soils is the mobility (leachability) of the pesticide, which in turn is determined by such factors as Koc, and the soil and hydrolysis half-lives. According to Gustafson , I * atrazine, aldicarb, picloram and diuron are the pesticides most liable to leach. The types of prevailing soil and water affect pesticide degradation, particularly non-biotic degradation processes (e.g., hydrolysis and photolysis).The photolysis of chlorotriazine, organophosphorus and phenylurea pesticides was studied in different aquatic media under laboratory conditions.13-1s The experiments revealed differences in the degradation rate between natural (whether surface or sea- water) and distilled waters arising from the presence of particulate matter and dissolved substances in natural waters, which influence photolysis by fostering radical reactions and hence increase the degradation rates.16 Photosensitizing and quenching effects have been observed for a variety of pesticides (e.g., atrazine, diuron) in photodegradation experi- ments involving distilled or sea-water.14-17 One other aspect to be considered in the interaction between pesticides and soils is biological transformation. This can be either non-enzymic, ( e . g . , involvement in photochem- ical reactions) or enzymic, with co-metabolism, detoxification or metabolism by widely available enzymes. Pesticides can also serve as a source of energy for some micro-organisms.6 Pesticides in Aquatic Organisms Some physico-chemical properties such as the water and fat solubility, the sorptivity in soil and sediments, and lipids and proteins, the relative molecular mass, hydrogen-bond forma- tion, dissociation constant, vapour pressure and melting-point are all of direct relevance to the bioaccumulation of pesti- cides.18 As a rule, however, the bioaccumulation of pesticides in aquatic organisms is related to the octanol-water partition coefficient (Flow); bioaccumulation factors are reportedly linearly related to log Kow, but only up to log KOw values of 5-6.18 Alternative correlations including quadratic terms accounting for steric factors and metabolism terms have also been reported. Such additional factors are considered to be essential19 for compounds such as polychlorinated biphenyls, which have log Kow values greater than 5-6. Modern pesticides have different log Kow values depending on the structural family. Thus, the values of log Kow of the triazines vary between 1.95 and 3.38,s suggesting a low bioaccumulation potential, whereas the organophosphates and carbamates have values of log Kow that cover the wide range from 0.7 to 5.9.20 The highest values (>5) correspond to the organophosphorus pesticides, such as bromophos-ethyl , chlorpyrifos, ethion, iodofenphos, leptophos, temephos and trichloronat, whereas the lowest (<2) are usually those of the carbamates.Sample Handling Pesticide sample preparation is usually achieved by dissolu- tion followed by liquid-liquid extraction (LLE) or by enrich- ment of trace compounds of interest with suitable sorbents, i.e., by solid-phase extraction (SPE). Although most of the official methods for pesticide analysis in water use LLE because of its simplicity and because it is a fully developed technique,’ SPE techniques have, however, gained in popu- larity and some have already been validated by different official institutions, viz., the US EPA and Ames labora- tories .* 1.22 Liquid-liquid extraction is an effective method for extract- ing pesticides from water samples; it is based on the partitioning of the different substrates between the aqueous sample and an immiscible organic solvent.The selectivity of LLE is dependent on the solvent used and the nature of the water matrix. Other parameters such as pH, ionic strength, water : solvent ratio, number of extractions and type and concentration of analyte must also be taken into account. These considerations were discussed in two reviews on the isolation of organic compounds from water samples .23.24 However, LLE methods do have some disadvantages: they are laborious, time-consuming and expensive and are subject to problems arising from the formation of emulsions, the evaporation of large solvent volumes and the disposal of toxic or inflammable solvents.One potential alternative to LLE methods is the increas- ingly used SPE, which uses a column containing a suitable sorbent to trap the analyte. The sample and solvent flow through the column by gravity or by positive (syringe) or negative (vacuum manifold) pressure. For on-line sample handling and trace enrichment, samples and flush solvents are applied via a second, inexpensive pump. Elution can be achieved by simply switching the valve to the ‘inject’ position, either in the forward-flush mode (i.e., in the same direction as sample loading) or in the back-flush mode (in the opposite direction). A typical SPE sequence involves the following steps: activation of the sorbent (wetting); removal of the excess of activating solvent (conditioning); sample applica- tion; removal of interferences (clean-up); elution of concen- trated analytes; and regeneration of the column.25 Solid-phase extraction is currently gaining acceptance and will probably be increasingly used as a wider variety of solid-phase supports become available.Sorbents such as carbon, alumina, silica, porous polymers, CX,CIX, aminopropyl-silica and Florisil are either laboratory prepared by packing the adsorbent into glass columns, or commerically prepared in disposable cartridges under trade names such as Supelclean, Quick-Sep, Sep-Pak and Bond-Elut. Cartridges for SPE are useful for field sampling and reduce sample manipulation, solvent consump- tion and labour costs by allowing batches of 12-24 samples to be prepared simultaneously, with no risk of sample contami- nation.Disadvantages of SPE include the need for extensive flushing of the solvents through the column when performing trace enrichment on sorbents with low selectivity, which results in concentration of many sample constituents and difficulties in the re-use of sorbents.21.22.25 Both LLE and SPE are commonly used off-line, with clearly separate handling steps (extraction and clean-up) and chro- matographic determination. Off-line SPE can be converted to on-line SPE by direct connection and elution of the adsorbed compounds onto the chromatographic system; the SPE columns are then usually referred to as ‘concentrator col- umns’.Such systems often involve the microprocessor control of the stages for switching of samples and flushing of solvents and eluents through the concentrator and chromatographic columns. This automated approach allows large-scale screen- ing and monitoring of pesticides in water. Although the chemistry and principles behind both off-line and on-line methods are essentially identical, the eluents used in off-line methods might require replacement in order to ensure compatibility with the subsequent chromatographic separation if an on-line method is to be used. In relation to the on-line methods, the use of miniaturized concentrator columns, membrane extraction discs and enzyme-linkedANALYST, JULY 1991, VOL.116 683 immunosorbent assays provides feasible alternatives to the on-line preconcentration and determination of pesticides in different types of water.'S-3() Water Samples Off-line method3 The identification and quantification of pesticides in matrices such as drinking and surface water is required for measuring environmental waste levels. Various preconcentration methods based on different physico-chemical principles are used for this purpose. Of these, LLE, dynamic and static headspace analysis, SPE and membrane processes are com- monly used and were recently described in a review arti~le.3~ A variety of extraction solvents are currently used for LLE. Thus, concentration of organophosphorus pesticides from water samples can be accomplished by using organic solvents such as hexane,32 dichloromethane, both with33 and without34 prior acidification (to avoid hydrolysis of the pesticides) ,33 and chloroform.3~ The use of these and other extraction solvents such as ethyl acetate and acetonitrile was discussed in a monograph on the determination of pesticides in water.36 Other pesticides, such as triazine herbicides, have been extracted into dichloromethane,2-37.3* ethyl acetate36 and mixtures of dichloromethane plus ethyl acetate or ammonium formate.39 Sometimes the extraction has to be increased by using alternative solvents or mixtures in order to analyse compounds that cannot be readily extracted by conventional methods (e.g., some breakdown products of triazine pesti- cides such as de-alkylated and hydroxylated triazine metabol- ites formed under environmental conditions) .39 However, pesticide-grade organic solvents should always be used for trace organic analyses.In SPE, the water sample is passed through a short bed of packing material, which can contain functional groups of different polarity, such as C8- or Clx-bonded phase, graphi- tized carbon black or Amberlite XAD resins. C8- and Clx-bonded phase cartridges-"),41 have been used for the analysis of various organophosphorus pesticides in surface and sea-water. Sep-Pak C8- and ClX-bonded phase silica cartridges have been employed for screening different pesticides such as triazines, molinate, trifluralin, alachlor and organophos- phorus compounds in Italian drinking and groundwaters at the sub-ppb Ie~el.~2 Amberlite XAD resins (mainly XAD-2) have been widely used in off-line analyses (e.g., for the extraction of organophosphorus pesticides from water samples at the ppb level using XAD-4)43 and for the isolation of atrazine'2.44 and chlorinated pesticides"34J6 from river and sea-water.Two excellent reviews of the use of XAD resins for the concentra- tion of organics in water were recently published.31-47 The greatest disadvantage of XAD resins is the generation of artefacts that are subsequently laborious to eliminate. One other group of selective SPE cartridges contain graphitized carbon black, and have been used to isolate triazines48 and phenoxyacid herbicides.49 The use of different sorbents in SPE involves a whole sequence of stages in the method, including sorbent activation, conditioning, sample application, interference removal, water removal and elution of the pesticides.Typically, methanol is used to wash the cartridges prior to use, and desorption of the pesticides is accomplished with either methanol-water or methanol-di- chloromethane mixtures when using graphitized carbon, and ethyl acetate, hexane or diethyl ether when using Cx- and C 8-bonded phases. Ace toni trile , acetone-water ,46 diethyl ether,45 acetone44 and acetonitrile-hexane47 have all been used for desorption of pesticides from XAD resins. On-line methods Solid-phase extraction procedures applied prior to the separa- tion and detection of pesticides in water by on-line liquid chromatographic techniques are areas of increasing interest. The sorbents, i.e., silica, alkylsilane-modified silica, alumina, porous polymers (with or without ion-exchange groups) or carbon modified materials, are usually packed into short stainless-steel or glass columns, or the 'concentrator columns' (in on-line procedures involving a chromatographic column) and can be used under high pressures.Solid-phase extraction columns are normally employed for trace enrichment, clean- up, sample storage, protection of the main column and, when needed, for derivatization. Trace components in the water sample are trapped by an appropriate sorbent packed into a concentrator column that is coupled to an analytical column via switching valves; the pesticides adsorbed are eluted from the concentrator column direct to the analytical column by a suitable mobile phase.Selectivity towards specific compounds can be achieved by coupling, in series, a variety of pre- columns containing different selective adsorbents such as C18, ion exchangers or metal-loaded phases. An overview and detailed description of the use of on-line concentrator columns in liquid chromatography for the determination of pesticides is provided elsewhere .253)-52 The host of pesticide determinations carried out on water samples over the last few years testifies to the increasing interest in this approach. Triazines at low-ppb levels have been concentrated from drinking and surface waters by using two pre-columns, whether isolated or serially connected, packed with C18 and PRP-1 (a styrene-divinylbenzene copolymer), respectively, followed by liquid chromatographic analysis using different detection systems, viz., ultraviolet (UV), fluorescence and electrochemical,sl.~3 or with a mem- brane disc containing 500 mg of C18 material coupled on-line to a liquid chromatograph with UV detection.27 Phenylurea herbicides and their corresponding anilines have been preconcentrated from river water using a platinum phase packed in a short concentrator column acting as an 'aniline filter'.Coupling of this pre-column to a C18 concentra- tor column allowed preconcentration of phenylureas at the low-ppb level for subsequent UV detection.54 A similar approach using electrochemical detection (ECD) was applied to the same herbicides for comparison. As expected, the sensitivities achieved with ECD were in the sub-ppb region and allowed detection of levels of about 30 ppt (parts per trillion) in surface waters.55 On-line preconcentration of diuron and its corresponding aniline from polluted well-water has also been reported.56 Phenylurea herbicides have been preconcentrated using a Clx column prior to analysis by liquid chromatography (LC) and detection by MS.The system used afforded an enrichment factor of from 100 to 1000-fold and characterization by LC-MS in the positive-ion mode.s7 Chlorinated phenoxy acids are another group of pesticides of interest for analysis by on-line preconcentration in LC. These compounds, which require derivatization for analysis by gas chromatography (GC), have been enriched from sub-ppb levels in drinking and surface water with further analysis by LC-UV.58 Also, various stationary phases such as octadecyl-silica, nitrile silica and macroporous polystyrene- divinylbenzene copolymers were used for preconcentration of the same group of phenoxy acid herbicides.*"sY The on-line liquid chromatographic system used was also applied to the determination of the fungicide iprodione in surface water and of carbamates, namely, carbaryl , chlorpro- pham, propoxur and carbofuran, in different types of water with detection limits in the sub-ppb range using UV detec- On-line LC-GC is another interesting approach that has gained in popularity over the last few years and is particularly suitable for on-line analysis of pesticides in water samples.61 In this respect, this technique combines the advantages of LC (viz., the possibility of using liquid chromatographic columns and pumps for preconcentration of pesticides) and the major assets of the selective detectors used in GC.The selectivity of ti on .3O. 60684 ANALYST, JULY 1991, VOL. 116 both the mobile and the stationary phase can be varied in LC, which can be exploited for preliminary clean-up of real samples. Liquid chromatographic techniques also allow mix- tures of components to be split into groups on the basis of chemical classes (group-type separations), and enrichment of very dilute samples. The LC-GC coupled techniques gener- ally use uncoated and de-activated capillary pre-columns, also known as retention gaps, which accommodate the liquid chromatographic effluent while it vaporizes, thereby provid- ing solute preconcentration.By using this approach, atrazine was analysed in water at the ppt level with the aid of a reversed-phase LC column and a solvent-exchange pro- cedure .61 The two main drawbacks of the technique are as follows: ( i ) the need for the derivatization of the preconcentrated polar herbicides and other metabolites that cannot be analysed directly by GC; and ( i i ) the high costs involved, which make it unaffordable by most of the laboratories routinely involved in pesticide analysis. On-line concentrator column technology has a promising future as a whole for the analysis of pesticides in water. Current technology allows on-line clean up and preconcentra- tion of pesticides with detection at the sub-ppb level. In addition, systems can be fully automated, a great aid for monitoring purposes. Recently, carbamate pesticides were determined completely automatically in 30 polluted water samples over the range 0.02-1 pg 1-1 by using disposable SPE cartridges.@ The water volumes concentrated on the concen- trator columns are typically between about 50 and 500 ml, i.e . , similar to those of off-line SPE methods. The ‘use of miniaturized concentrator columns in conjunction with nar- row-bore liquid chromatographic systems reduces water volumes to between 1 and 5 ml, i.e., much smaller volumes than those of off-line SPE methods. The breakthrough volume, one of the most critical parameters for preconcentra- tion in solid adsorbents, can usually be measured more accurately by on-line techniques, which thereby allow better optimization of the over-all preconcentration method.Finally, more accurate results in trace analysis should be expected from on-line than from off-line preconcentration as the former involves no sample manipulation between preconcentration and analysis (the whole of the concentrated organic material is transferred and analysed) . Sediment and Biota Samples Extraction Freeze-drying is a common method for conditioning sediment and biota samples for the extraction of pesticides.23.62 Freeze-drying of biological material destroys many cell membranes and hence increases the contact between the sample matrix and the extraction solvent. Despite the tedium, Soxhlet extraction allows efficient removal of pesticides from sediment and biota samples. Efficiency depends on, amongst other things, a suitable choice of solvent for each group of pesticides.The matrix is also of importance in this respect. Some pesticides (i.e., triazines) in sediment samples are involved in complex physical and chemical adsorption mechanisms that hinder extraction of the residues.62.63 Other serious problems encountered in the analysis of organic compounds in sediments are caused by elemental sulphur, often present in such materials, and by the various amounts of lipids occurring in biota samples. The following are all used for the extraction of organophos- phorus pesticides from soil matrices:62 acetone; acetone- water; dichloromethane; ethyl acetate;36 acetone-hexane (4 + 1 and 6 + 4)32,64365 (also suitable for pyrethroids); acetonedichloromethane (1 + 1)66 (also used for carba- mates); and methanol-water (9 + 1).Chlorinated phenoxy acid herbicides are extracted from soil matrices under different conditions. As these compounds are free organic acids, the presence of water will weaken the ionic bonds and facilitate separation. Consequently, water-miscible polar sol- vents such as methanol, acetone and acetonitrile are usually used in preference to sparingly miscible solvents such as diethyl ether and ethyl acetate or immiscible solvents such as dichloromethane. Most workers facilitate the breakage of the ionic bonds by replacing the charged soil particles with more electropositive cations (K+ , Na+ and NH4+). This is why alkali extractions in wate1-36.~~ or mixtures of acetonitrile or acetone-water in basic solutions are being increasingly used for this purpose.36 These extraction methods are also useful for other acid herbicides such as glyphosate, which is extracted from soil samples with KOH.68 Triazine herbicides are usually extracted with met hanol-water ,36,62,69 ace tone ,63 ace toni trile-water, ace- tone-ethyl acetate,36 methanol and diethyl ether.70 The solvents used to extract pesticide residues from biota are similar to those employed in analysing sediment samples.Extraction of pesticides from the biota matrix is one of the key steps of the analytical procedure involved. Extraction of pesticide residues from lipids is facilitated by strong acids and alkalis; however, some of the pesticides (e.g., organophos- phorus compounds and triazine) are decomposed as a result.The part of the matrix extracted with the pesticides must be removed by using an appropriate clean-up method. Ethyl acetate is commonly used for the extraction of organophos- phorus pesticides from biota samples,34~36~71 although aceto- nitrile,72 hexane32 acetone and methanol36 are also used for this purpose. As with soil samples, chlorinated phenoxy acids are extracted with the aid of alkali solutions.36 One problem still remains, however, how to achieve high pesticide extraction recoveries from soil and biota matrices without affecting the analytes of interest; much still needs to be done in this respect. Gentle but efficient extraction techniques such as supercritical fluid extraction (SFE) have proved to be of help in the extraction of triazines from soil samples with adsorptive properties.73 Clean-up Analysing pesticide residues in soil and biota samples entails isolation of the residues by a clean-up of the extract.Because most of the pesticides might be destroyed by strong acid and alkali, non-destructive methods involving partitioning sol- vents and column chromatography with silica gel, alumina or Florisil, gel-permeation chromatography (GPC) or normal- phase LC are to be preferred. Partitioning solvents such as acetonitrile have been employed in conjunction with Florisil, silica gel and basic alumina for the isolation of organophos- phorus and carbamate pesticides.74 Clean-up methods are only applied to water samples when strictly necessary. The clean-up efficiency of column chromatography has been improved by using silica gel with 1% acetic acid prior to the addition of samples containing organophosphorus pesti- cides.75 The US EPA clean-up methods are based on the use of Florisil column chromatography for organophosphates76 and triazines.77 Four different fractions of 6, 15,50 and 100% diethyl ether in hexane are eluted in both instances.A similar protocol was employed for organophosphorus and triazine residues in soil samples, from which they were extracted with hexane-diethyl ether (50 + 50).62 Gel-permeation chromatography has a molecular size discrimination capacity if applied with low-pressure chromato- graphic columns, and can be used after extraction in most instances. Thus, large biogenic compounds such as lipids are excluded from the pores of the polymeric material and eluted before smaller analytes, which are retained in the pores. Separation mechanisms other than size exclusion, vzz., adsorption and partition, must be considered.The prevalence of one type of mechanism over other types is largely determined by the mobile phase and the pore size of the packing chosen. With large pore GPC packings [1000-2000ANALYST, JULY 1991, VOL. 116 685 relative molecular mass ( M , ) exclusion], e.g., Bio-Beads SX-3, SX-4 and SX-8, both size exclusion and adsorption occur in the presence of poorly solvating mobile phases. This is true with Bio-Beads SX-3 (2000 M , exclusion limit), the GPC packing most frequently used to analyse different pesticides in conjunction with a variety of eluting solvents, such as ethyl acetate-cyclohexane ,34362,78379 acetone- cyclohexane, acetone-light petroleum (b.p.40-60 “C)80 and cyclohexane-dichloromethane.*1?82 In this way a wide range of pesticides, including organophosphorus compounds, phen- oxy acids and nitrogen-containing pesticides, can be success- fully isolated from their corresponding matrices. A less common approach for the clean-up of samples from environmental matrices that are to be analysed for pesticides involves the use of normal-phase LC8334 with scaled-up silica columns and dichloromethane-hexane or dichloromethane- pentane as eluents. Matrix interferences, usually lipids in biota, appear at the end of the chromatogram as the pesticides are eluted beforehand. This approach has been successfully applied to the isolation of organochlorine and organophos- phorus pesticides from both biota and other lipid-containing samples.However, a major drawback is that the more polar organophosphorus pesticides co-elute with the lipid peak and are thus poorly resolved from the lipid matrix.84 Chromatographic Analysis Gas Chromatographic Methods Capillary GC, in conjunction with selective detectors [mainly nitrogen-phosphorus (NPD), electron capture (ECD) and flame photometric (FPD)] or MS, is still one of the most common techniques for the determination of environmental pesticide residues. The low detection limits, high selectivity and affordability of gas chromatographic instrumentation are appealing to most laboratories involved in pesticide residue analysis. Several reviews on the use of GC-NPD and GC-ECD85.86 and GC-MS in various operational modes such as electron impact (EI) and positive (PCI) and negative (NCI) chemical ionization have been published.87,88 The GC-NPD analyses can be performed by the phosphorus and the nitrogen mode.The phosphorus mode, which is about one order of magnitude more sensitive than the nitrogen mode, allows organophosphorus pesticides to be routinely determined at concentrations of a few ng g-1 in different environmental mat rice s .34,62,64,66,74,76,79.8135 The nitrogen mode permits triazines to be determined at the ng g-1 level .62.69.7735 Electron-capture detectors, the most com- monly used to analyse classical chlorinated pesticides such as dichlorodiphenyl trichloroethane (DDT) and endrin, are resorted to when the molecule contains chlorinated groups (e.g., chlorpyrifos, a chlorine-containing organophosphorus pesticide35,6”7*) or when carbamate derivatives (trichloro- acetyl), chlorinated phenoxy acids (pentafluorobenzyl) or phenylurea pesticides (heptafluorobutyric) are to be ana- lysed.85.89 Other detectors used for organophosphorus pesti- cide analysis include GC-FPD71-81 and alkali flame-ionization detectors (FID).7* Also, photo-ionization detectors are used to analyse sulphur-containing and chlorinated pesticides, generally connected in parallel with other detectors (e.g., FID).”) Gas chromatography-mass spectrometry is widely used by environmental laboratories involved in pesticide residue analysis.The most common practice in this context is to perform GC-MS in the EI mode with a library search for the unequivocal identification of the pesticide or with a second injection to check for co-elution with an authentic standard of the pesticide of interest. Thus, GC-MS with EI is commonly used to confirm the presence of organophosphorus pesti- cides,42.*7.88 triazines,42.62.91 pyrethroidsg’ and paraquat and diquat after dehydrogenation.93 The use of GC-MS with NCI is a selective approach, particularly suitable for pesticides containing electron withdrawing groups (e.g., chlorine and nitro groups), which can stabilize negative charges.The main advantage of NCI is its high selectivity and sensitivity to organochlorine and organophosphorus pesticides in environ- mental samples. Gas chromatography-mass spectrometry in the NCI mode has been used to confirm the presence of organophosphorus ~ompounds34.*7~91~94,95 and tria~ines.~6 The use of GC-MS in the PCI mode has also been applied to organophosphorus pesticides95 and triazines.91.97 However, GC-MS with PCT and NCI with selected ion monitoring of 2-3 characteristic fragment ions of each analyte allows the unequivocal identification of organophosphorus pesticide residues in different environmental matrices.95 The need to use GC-MS in environmental pesticide analyses for confirma- tion purposes arises from the enormous variety of pesticides currently used.Selectivity and sensitivity can be further enhanced by using various ionization modes. However, only a few libraries of standard mass spectra are available for GC-MS with PCI and NCI, so each laboratory must collect their own.This is one of the main drawbacks of this technique and arises because the instrumental parameters such as the source temperature and reagent gas pressure have a critical influence on the relative ion intensities of the mass spectra obtained under chemical ionization conditions. When there are no GC-MS facilities available, other identification methods must be used. One inexpensive alternative in this respect is the use of linked retention data, parallel FPD and ECD, and linear temperature retention indices. This approach has been used for the identification of a wide range of organophosphorus, sulphur-containing and organochlorine pesticides.98 Liquid Chromatographic Methods Liquid chromatography systems used for environmental pesticide analysis were extensively reviewed in a recent paper.52 The increasing availability of liquid chromatographic methods for pesticide analysis is mainly the result of their suitability for thermally labile and polar herbicides, which require derivatization prior to gas chromatographic analysis.The liquid chromatographic methods of analysis also have a major advantage over GC, in that on-line pre-column and post-column systems are compatible with LC. In addition, the most powerful detectors for environmental organic analysis, namely mass spectrometers, have also been successfully coupled to LC. Some examples of the use of on-line pre-column systems in LC were mentioned above in relation to the analysis of pesticides in different water matrices. Common detectors Ultraviolet detectors are the most common choice for LC, set at different wavelengths according to the pesticides to be analysed (e.g., 241 nm for phenylurea pesticides,51.53,54 229 nm for iprodione,25.60 230 or 280 nm for phenoxy acids27,4”.”8,59 and 220 nm for carbamates52).Phenylureas have also been analysed by using ECD,51354 fluorescence detectors” and MS57 coupled on-line with pre-column liquid chromatographic systems. The use of the UV detector in LC in conjunction with off-line sample preparation is still the most common choice in environmental pesticide analysis. Information on specific wavelengths, eluents and columns for LC of over 200 pesticides is available in the literature. Ultraviolet detectors are the most commonly available in laboratories and also traditionally the most frequently used in LC.They have been used in the analysis of chlorotriazines at 220 nm ,62.63.70.99-101 organophosphorus compounds at wavelengths such as 205, 210 and 280 nrn,62,*02 phenoxy acids at 230 and 280 1311167 and carbamates at 295 and 247 nrn.103.104 The most serious drawbacks of UV detectors arise from matrix interferences and from working below 205 nm. According to the literature,686 ANALYST, JULY 1991, VOL. 116 the selectivity towards the pesticide asulam105 in a food matrix can be enhanced by choosing the most appropriate wavelength (280 nm) to avoid matrix interferences. Post-column reactors Although, as mentioned above, UV detection has found widespread application, it can only be used to detect mol- ecules that possess a chromophore.This has led to the development of chemical reaction methods to enhance detec- tion in LC. Two excellent, recently published reviews showed the great potential of post-column liquid chromatographic systems. 106,107 However, they require reagent solutions to be added and hence for additional pumps to be used; in addition, extra-column peak broadening can result from the reactor. Post-column systems are used in two different ways: the first involves employing normal-phase separators to extract ana- lytes into an immiscible organic solvent. Thus, aqueous and organic plugs are created, which are passed through a phase separator in order to isolate the latter. The pesticides in the organic phase are driven to the detector while the aqueous phase is sent to waste.The second alternative, more com- monly in use with post-column systems, involves post-column reactors based either on changes in physico-chemical proper- ties (viz., electrochemical, redox, hydrolytic) or photochem- ical and derivatization reactions (e.g., ion-pair formation and ligand-exchange) . Post-column extraction systems have scarcely been applied to pesticide analysis. Such systems have been used for polar pesticides or metabolites that are difficult to assay under typical conditions used in LC. Hydroxyatrazine, a polar metabolite of atrazine, was analysed by using ion-pair post-column extraction with 9,10-dimethoxyanthracene-2-sul- phonate as the counter ion and fluorescence detection.108 Buffers and ion-pairing agents present in the LC mobile phases that are likely to interfere with detection can be removed by using a post-column extraction system to transfer the organic phase to the detector while the inorganic ions remain in the aqueous layer.This procedure has been used for the ion suppressed extraction of chlorinated phenoxy acid pesticides with on-line MS detection.109 A typical application of solid-phase reactors is the analysis of N-methyl carbamates with on-line hydrolysis and addition of o-phthalaldehyde (OPA). The carbamates yield methyl- amine on hydrolysis which is converted into a fluorescent compound by the addition of OPA in a second reactor. This approach was recently applied to the analysis of carbamate residues at the picogram level in environmental mat- rices.11",111 The same group of pesticides was also assayed by post-column photolysis, reaction with OPA and fluorescence detection. 112 Other representative examples are the use of oxidation and derivatization reactions with OPA for the determination of glyphosate, one of the most polar herbicides that can only be extracted in water, with detection by fluorogenic labelling,113 and the determination of diquat and paraquat using a post-column reaction with sodium hydro- sulphite and diode-array detection. 114 Mass spectrometric detection The on-line combination LC-MS occupies a prominent place in environmental organic analyses and surpasses GC-MS in the analyses of polar pesticides and herbicides in some aspects. One book115 and two review articles52,87 have been published and provide excellent overviews of LC-MS applications to environmental pesticide analysis.An outstanding review of the different types of interfaces used in this context including transport systems, direct liquid introduction, thermospray, atmospheric-pressure ionization, particle beam, open tubular LC and continuous flow fast atom bombardment has also been published recently.116 Of the different LC-MS methodol- ogies, the thermospray (TSP) interfacing system is probably the most widely used and typically involves reversed-phase columns and volatile buffers, with or without a filament or discharge. Thermospray LC-MS has been applied to the analysis of a variety of pesticides including carbamates,117-'21 organophosphorus compounds,121-125 pyrethroids, 124 phenyl- triazine~.~~~~O~.~15."27.'28 Filament-off, filament-on and positive- and negative-ion modes are common choices in this context.The filament-off mode is associated with TSP ionization, whereas the process involved in filament-on is closer to chemical ionization and discharge ionization, which are applied to liquid chromatographic eluents with high water contents, in order to provide further structural information with additional fragments. These two procedures are used routinely; however, the filament-on alternative is more common owing to its higher sensitivity. The choice between the positive- or negative-ion mode depends on the compound concerned; the positive-ion mode is more frequently used. It normally yields [M + H]+ and/or [M + NH4]+ base peaks. The negative-ion mode has been shown to be more sensitive than the positive-ion mode to electronegative compounds such as chlorinated phenoxy acids; the negative ion mode yields [M + acetatel- or [M + formatel- base peaks if ammonium acetate or ammonium formate is used as the ionizing additive.126 Most sophisticated equipment, such as on-line LC-MS-MS (LC-tandem MS) or continuous flow fast atom bombardment instruments, also permits pesticide characterization. Thus, LC-MS-MS was applied to the analysis of organophos- phorus,1*5 carbamate118.120 and triazine115 pesticides, and was found to provide more fragmentation than typical TSP urea,115,117.121.123.126.127 chlorinated phenoxy aCids115.128 and LC-MS. Supercritical fluid chromatography Supercritical fluid chromatography (SFC) has been used to analyse various pesticides. 129-136 As supercritical fluids are gas-like in some aspects and liquid in others (they are typically 10-100 times less viscous than liquids), they can be used as mobile phases thus forming complementary aids to LC and GC.The main advantages of this technique are the shorter retention times involved in the analysis of moderately polar and thermally labile pesticides with large M, values, and the compatibility with most LC and GC detectors [UV, FID, NPD, MS and Fourier transform infrared spectrometry (FTIR)]. Reported applications of SFC include the analysis of carbamate, 129-133 organophosp horus134 and phenylurea l319l35 pesticides. In most of these, carbon dioxide is used as the supercritical mobile phase and propan-2-01 as the polarity modifier.134 The combination SFC-MS131.133 permits the characterization of different pesticides with EI or chemical ionization, which makes the technique potentially useful for environmental analysis. In addition, this technique allows improved extraction of pesticides from environmental mat- rices. Thus, SFE coupled to SFC was successfully used to isolate and analyse carbofuran and its 3-keto and 3-hydroxy metabolites. 136 Other applications include the extraction of pesticides ( e . g . , 2,4-D) from soil samples; the pesticides are then determined (without prior clean-up) by SFC using ion-mobility detection. 137 Miscellaneous Analysis for the most polar pesticides and others such as the ubiquitous 'quats' is rather tedious, as a great many analytical steps are involved.Polar and amphoteric pesticides, such as glyphosate115 and paraquat ,138 respectively, have been identi- fied by fast atom bombardment. This technique allows identification, but not quantitative analysis, thus indicating that not all pesticides can be analysed by the methods described in this work. Other chromatographic methods for the analysis of pesti- cides include thin-layer chromatography (TLC), which hasANALYST, JULY 1991, VOL. 116 687 proved to be effective for a variety of pesticides.24.139.140 The separation efficiency in TLC can be increased by working under overpressure, i. e . , under constant-flow conditions. Thin-layer chromatography features a number of advantages such as short analysis times, easy manipulation, and quan- tification and data manipulation capabilities similar to those available with GC and LC, with the added asset of the possibility of using MS detectors.140 Gas chromatography-FTIR and the more recent GC- FTIR-MS coupled technique provide useful information for screening complex environmental matrices; to date they have been applied to the analysis of 2,4-D salts in soil mat- Organometallic compounds used as pesticides have been assayed by electrothermal atomic absorption spectrometry (ETAAS) coupled to GC and LC.105,*39,142-144 Determination of organotin compounds in environmental samples was accomplished by using GC-ETAAS or LC-ETAAS with or without hydride generation,142 by direct flame ionization from a normal-phase eluent143 or an ion-exchange column .I44 The improved detection limits achieved usually range between 5 and 50 ng and can be further improved to a few picograms by using hydride generation.rices. 139,141 Conclusions Pesticide samples from the aquatic environment can obviously be handled and analysed in many ways. Solid-phase extraction (SPE) is gradually replacing LLE for the extraction of pesticides from water in many analytical methods because of the wide availability of selective adsorbent materials, and also to avoid the necessity of disposing of organic solvents. In addition, the ease of automation of the on-line concentration techniques coupled to LC has fostered their use for trace enrichment and clean-up of pesticides in environmental samples. The sample preparation step, whether used with LLE or SPE (with further clean-up), cannot be regarded as a separate entity.Sample handling should always be viewed in terms of the total analytical procedure and consequently be matched to the separation system and the detection mode used. Clearly, the choice of a particular sample handling and analytical technique should be dictated by the particular problem addressed. Determining pesticides by GC is relatively simple provided an adequately selective detector is used. This will be the methodology of choice for most laboratories involved in pesticide residue analysis on account of the lower detection limits (approximately two orders of magnitude) compared with LC. On the other hand, LC allows polar and thermally labile pesticides to be assayed without derivatization and the direct injection of ‘dirty’ extracts into water-miscible solvents.The inability of LC to analyse pesticides containing no chromophores or fluorophores is being partly overcome by the increasingly common LC-MS applications, as most pesticides are responsive to MS and can, therefore, be readily identified. Complementary use of GC and LC should obviously allow monitoring of most pesticides and their degradation products in the aquatic environment. In addition, there exists a variety of commercially available equipment for GC and LC with mass spectrometric detection that can solve problems asso- ciated with the confirmation of residues. The choice of analytical techniques has increased enor- mously over the last few years. Thus, both SFC and on-line LC-GC systems incorporating a variety of detectors (FID, UV, NPD and ECD) are already commercially available and have a promising future in this field.One constraint on the widespread use of this type of instrumentation is the high cost, typically more than twice that of a complete GC or LC system. Other coupled techniques include the use of the multi-element atomic emission detector for GC and LC or SFC coupled to fast atom bombardment MS, nuclear magnetic resonance spectroscopy o r inductively coupled plasma MS. Future studies of analytical methodology should concen- trate on the still scarcely explored aspects, such as: (i) the development of SFE methods for the extraction of pesticides from soil and biota and the coupling of these methods to other separation techniques such as TLC, LC, GC and SFC; (ii) one or more clean-up procedures for use with multiresidue methods to determine pesticides of diverse polarities including the more polar ones (e.g., by using novel selective adsorbents such as the new types of graphitized carbon, by collection of several fractions in normal-phase LC and/or high-resolution GPC); (iii) the use of mass spectrometric techniques to confirm pesticides that can be determined by GC ( e .g . , by using NCI, high-resolution MS and MS-MS, which, although well established and available in many laboratories, are used reluctantly); (iv) the determination of polar and amphoteric pesticides by mass spectrometric techniques (e.g., analysis for glyphosate or the ‘quats’ by micro-LC coupled to continuous flow fast atom bombardment MS or capillary zone electro- phoresis-MS); and (v) increasing automation in pesticide analysis ( e .g . , by using on-line disposable SPE cartridges as concentrator columns in LC, on-line LC-GC or on-line SFE-SFC coupled to a variety of selective detectors such as ECD, NPD and MS). One of the referees is thanked for his suggestions for improving the quality of the manuscript. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 References US Environmental Protection Agency, Agricultural Chemicals in Groundwater: Proposed Pesticide Strategy, US EPA, Wash- ington, DC, 1987, pp. 1-150. Watts, C. D., Clark, L., Hennings, S., Moore, K., and Parker, C., in Pesticides: Analytical Requirements for Compliance With EC Directives, Water Pollution Research Report 11, eds.Crathorne, B., and Augeletti, G., Commission of the European Communities, Brussels, Belgium, 1989. pp. 16-34. Premazzi, G., Evaluation of the Impact of Malathion on the Aquatic Environment, Commission of the European Communi- ties, Joint Research Center, Ispra, Italy, 1983, pp. 1-67. Buser, H.-R., Environ. Sci. Technol., 1990, 24, 1049. Montiel, A., Welte, B., Franchet, C., and Legrand, S., Water Supply, 1989, 7, 213. Aharonson, N., Pure Appl. Chem., 1987, 59, 1419. Klaine, S. J., Hinman, M. L., Winkelmann, D. A., Sauser, K. R., Martin, J. R., and Moore, L. W., Environ. Toxicol. Chem., 1988, 7, 609. Frank, R., and Logan, Ll., Arch. Environ. Contam. Toxicol., 1988, 17, 741. Galassi, S., Battaglia, C., and Vigano, L., Chemosphere, 1988, 17, 783.Rivera, J., Ventura, F., Caixach, J., de Torres, M., and Figueras, A., Int. J. Environ. Anal. Chem., 1987, 29, 15. d’Elia, C. E., Sanders, J. G., and Capone, D. G., Environ. Sci. Technol., 1989,23, 768. Gustafson, D. I., Environ. Toxicol. Chem.. 1989, 4, 339. Esser. H. O., Dupuis, G., Ebert, E., Vogel, C., and Marco, G. J., in Herbicides: Chemistry, Degradation and Mode of Action, eds. Kearney, P. C., and Kauffmann, D. D., Marcel Dekker, New York, 1975, vol. I, pp. 129-208. Durand, G., Barcelo, D., AlbaigCs, J., and Mansour, M., Chromatographia, 1990,29, 120. Kotzias, D., Klein, W., and Korte, F., Chemosphere, 1974, 4, 161. Hwang, H. M., Hodson, R. E., and Lee, R. F., in Photo- chemistry of Environmental Aquatic Systems, eds. Zika, R. G., and Cooper, W.J . , American Chemical Society Symposium Series, No. 327, 1987, pp. 2743. Cotham, W. E., and Bidleman, T. F., J. Agric. Food Chem., 1989, 37, 824. Esser, H. O., Pestic. Sci., 1986, 17, 265. Barcelo, D., Porte, C., and AlbaigCs, J., Environ. Sci. Technol., 1989, 23, 617. Bowman, B. T., and Sans, W. W., J. Environ. Sci. Health, Part B , 1983, 18, 667.688 ANALYST, JULY 1991, VOL. 116 21 Lopez-Avila, V., Milanes, J., Dodhiwala, N. S., and Beckert, W. F., J. Chromatogr. Sci., 1989,27, 209. 22 Junk, G. A., and Richard, J. J., J. Res. Natl. Bur. Stand. (US), 1988, 93, 274. 23 Onuska, F. I., HRC CC, J . High Resolut. Chromatogr. Chromatogr. Commun., 1989, 12,4. 24 Liska, I., Krupcik, J., and Leclercq, P. A., HRC CC, J. High Resolut. Chromatogr. Chromatogr.Commun., 1989, 12, 577. 25 Nielen, M. W. F., in Selective On-line Pre-column Sample Handling and Trace Enrichment in Liquid Chromatography, eds. Frei, R. W., and Zech, K., Elsevier, Amsterdam, J. Chromatogr. Libr., 1988, vol. 39A, ch. 1. 26 Barrat, D. J., in Pesticides: Analytical Requirements for Compliance With EC Directives, Water Pollution Research Report 11, eds. Crathorne, B., and Augeletti, G., Commission of the European Communities, Brussels, Belgium, 1989, pp. 27 Brouwer, E. R., Lingeman, H., and Brinkman, U. A. Th., Chromatographia, 1990,29,415. 28 Steinwandter, H., Fresenius J. Anal. Chem., 1990, 336, 8. 29 Geerdink, R., van Balkom, C. A., and Brouwer, H.-J., J. Chromatogr., 1989, 481, 275. 30 Marvin, C. H., Brindle, I. D., Hall, C. D., and Chiba, M., J.Chromatogr., 1990,503, 167. 31 Namiesnik, J., Gorecki, T., Biziuk, M., and Torres, L., Anal. Chim. Acta, 1990, 237, 1. 32 Organophosphorus Pesticides in River and Drinking Water 1980. Tentative Method, HM Stationery Office, London, 1983, pp. 33 Wang, T. C., Lenahan, R. A., and Tucker, J. W., Jr., Bull. Environ. Contam. Toxicol., 1987, 38,226. 1 34 Barcelo, D., Porte, C., Cid, J., and AlbaigCs, J., Znt. J. Environ. Anal. Chem., 1990,38, 199. 35 Neicheva, A., Kovacheva, E., and Marudov, G., J. Chro- matogr., 1988,437, 249. 36 Chau, A. S. Y., and Afghan, B. K., Analysis of Pesticides in Water, CRC Press, Boca Raton, FL, 1982, vols. I, I1 and 111. 37 Chlorophenoxy Acid Herbicides, Trichlorobenzoic Acid, Chlorophenols, Triazines and G lyphosate in Water, 1985, HM Stationery Office, London, 1986, pp.1-150. 38 Muir, D. C. G., J. Agric. Food Chem., 1980, 28, 714. 39 Durand, G., and Barcel6, D., Toxicol. Environ. Chem., 1989, 25, 1. 40 Hinckley, D. A., and Bidleman, T. F., Environ. Sci. Technol., 1989, 23, 995. 41 Maiies Vinuesa, J., Molt6 CortCs, J. C., Igualada Caiias, C., and Font Ptrez, G., J. Chromatogr., 1989,472,365. 42 Bagnati, R., Benfenati, E., Davoli, E., and Fanelli, R., Chemosphere, 1988, 17, 59. 43 Verweij, A., van Liempt, M. A., and Boter, H. L., Znt. J. Environ. Anal. Chem., 1985, 21, 63. 44 Galassi, S., Battaglia, C., and Viganb, L., Chemosphere, 1988, 17,783. 45 Ishibashi, M., and Suzuki, M., J. Chromatogr., 1988,456, 382. 46 Gomez-Belinchon, J. I., Grimalt, J. O., and AlbaigCs, J.. Environ. Sci. Technol., 1988,22, 677.47 Daignault, S. A., Noot, D. K., Williams, D. T., and Huck, P. M., Water Res., 1988, 22, 803. 48 Di Corcia, A., Marchetti, M., and Samperi, R., J. Chromatogr., 1987,405, 357. 49 Di Corcia, A., Marchetti, M., and Samperi, R., Anal. Chem.. 1989, 61, 1363. 50 Frei, R. W., Nielen, M. W. F., and Brinkman. U. A. Th., Znt. J. Environ. Anal. Chem., 1986, 25, 3. 51 Subra, P., Hennion, M. C., Rosset, R., and Frei, R. W., Znt. J. Environ. Anal. Chem., 1989, 37, 45. 52 Barcel6, D., Chromatographia, 1988, 25, 928. 53 Subra, P., Hennion, M. C., Rosset, R., and Frei, R. W., J . Chromatogr., 1988,456, 121. 54 Goewie, C. E., Kawkman, P., Frei, R. W., Brinkman, U. A. Th., Maasfeld, W., Seshadri, R., and Kettrup, A., J. Chromatogr., 1984, 284, 73. 55 Nielen, M.W. F., Koomen, G., Frei, R. W., and Brinkman, U. A. Th., J . Liq. Chromatogr., 1985, 8, 315. 56 Goewie, C. E., and Hogendoorn, E. A., J. Chromatogr., 1987, 410, 211. 57 Maris, F. A., Geerdink, R. B., Frei, R. W., and Brinkman, U. A. Th., J. Chromatogr., 1985,323, 113. 58 Hamann, R., and Kettrup, A., Chemosphere, 1987, 16, 527. 89-99. 1-17. 59 Hamann, R., Meier, M., and Kettrup, A., Fresenius 2. Anal. Chem., 1989,334,231. 60 Marvin, C. H., Brindle, I. D., Hall, C. D., and Chiba, M., Anal. Chem., 1990,62, 1495. 61 Davies, I . L., Markides, K. E., Lee, M. L., Raynor, M. W., and Bartle, K. D., HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 1989, 12, 193. 62 Durand, G., Forteza, R., and Barcelo, D., Chromatographia, 1989,28, 597. 63 Battista, M., Di Corcia, A., and Marchetti, M., J.Chromatogr., 1988,454,233. 64 Kjolholt, J., Chemosphere, 1985, 14, 1763. 65 Elhag, F. A., Yule, W. N., and Marshall, W. D., Bull. Environ. Contam. Toxicol., 1989,42, 172. 66 Belisle, A. A., and Swineford, D. M., Environ. Toxicol. Chem., 1988, 7, 779. 67 Meier, M., Hamann, R., and Kettrup, A., Fresenius 2. Anal. Chem., 1989,334,235. 68 Miles, C. J., and Moye, H. A., J. Agric. Food Chem., 1988,36, 486. 69 Helling, Ch. S., Zhuang, W., Gish, T. J., Coffman, C. B., Isensee, A. R., Kearney, Ph. C., Hoagland, D. R., and Woodward, M. D., Chemosphere, 1988, 17, 175. 70 Cabras, P., Spanedda, L., Pellecchia, M., and Gennari. M., J. Chromatogr., 1989, 472, 411. 71 McLeese, D. W., Zitko, V., and Sergeant. D. B.. Bull. Environ. Contam. Toxicol., 1979, 22,800.72 Aquatic Biology Group, Chemosphere, 1984, 13, 19. 73 Janda, V., Steenbeke, G., and Sandra, P., J. Chrornatogr., 1989,479,200. 74 Sharp, G. J . , Brayan, J. G., Dilli, S., Haddad, P. R., and Desmarchelier, J. M., Analyst, 1988, 113, 1493. 75 Lores, E. M., Moore, J. C., and Moody, P., Chemosphere, 1987, 16, 1065. 76 Pressley, T. A., and Longbottom, J. E., The Determination of Organophosphorus Pesticides in Industrial and Wastewaters, Method 614, US EPA, Cincinatti, OH, 1982, pp. 1-24. 77 Pressley, T. A., and Longbottom, J. E., The Determination of Triazine Pesticides in Industrial and Wastewaters, Method 619, US EPA, Cincinatti, OH, 1982. pp. 1-23. 78 Thier, H. P., and Zeumer, H., Manual of Pesticide Residue Analysis. Clean-up Method6, VCH, Weinheim, 1987, vol.I, pp. 79 Roos, A. H., van Munsteren, A. J., Nab, F. M., and Tuinstra, L. G. M., Anal. Chim. Acta, 1987, 196,95. 80 Steinwandter, H., Fresenius 2. Anal. Chem., 1988, 331,499. 81 Lawrence, J. F., Znt. J. Environ. Anal. Chem., 1987, 29, 289. 82 Venant, A., Borrel, S., Mallet, J., and van Neste, E., Analusis, 1989, 17, 64. 83 Petrick, G., Schulz, D. E., and Duinker, J. C.. J. Chromatogr.. 1988,435,241. 84 Gillespie, A. M., and Walters, S. M., J. Liq. Chromatogr., 1986,9,2111. 85 Onuska, F. I . , HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 1984, 7,660. 86 Mansour, M., Hustert, K., and Muller, R., Znt. J. Environ. Anal. Chem., 1989,37,83. 87 Levsen, K.. Org. Mass Spectrom., 1988. 23, 406. 88 Stan, H. J., J. Chromatogr., 1989. 467, 85. 89 de Kok, A., van Opstal, M., de Jong, T., Hoogcarspel, B., Geerdink, R.B.. Frei, R. W., and Brinkman. U. A. Th., Int. J . Environ. Anal. Chem.. 1984. 18, 101. 90 Dressler, M., J. Chromatogr. Libr., 1986, 36, ch. 6. 91 Durand, G., and Barcelo, D., Anal. Chim. Acta. 1991,243,259. 92 Lidgard, R. 0.. Duffield. A. M., and Wells, R. J., Biomed. Environ. Mass Spectrom.. 1986, 13, 677. 93 Hajslova, J., Cuhra, P., Davidek. T.. and Davidek, J.. J. Chromatogr., 1989,479,243. 94 Stan, H. J., and Kellner. G., Biomed. Mass Spectrom.. 1982.9, 483. 95 Stan, H. J., and Kellner, G., Biomed. Environ. MassSpectrom., 1989, 18,645. 96 Huang, L. Q., and Mattina. M. J. L.. Biomed. Environ. Mass Spectrom., 1989, 18. 828. 97 Rostad, C. E . , Pereira, W. E.. and Leiker, T. J.. Biomed. Environ.Mass Spectrom., 1989, 18, 820. 98 Stan, H. J., and Mrowetz, D., J. Chromatogr.. 1983,279, 173. 99 Pacakova, V., Stulik, K., and Prihoda, M., J. Chromatogr.. 1988,442. 147. 75-78.ANALYST, JULY 1991, VOL. 116 689 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 Sanchez-Rasero. F., and Dios, G. C., J. Chromatogr., 1988, 447, 426. Durand. G., and Barcelo, D., J. Chromatogr., 1990, 502,275. Greve, P. A., and Goewie, C. E., Int. J. Environ. Anal. Chem., 1985,20,29. Cabras, P.. Spanedda, L., Tuberoso, C., and Gennari, M., J. Chromatogr., 1989, 478, 250. Spierenburg, Th. J., van Dijk, M. B. H. K., and Zoun, D. E. F., J. Chromatogr., 1987,393, 137. Lawrence, J. F., Chromatographia, 1987, 24,45. Brinkman, U. A. Th..Chromatographia, 1987,24, 190. Brinkman, U. A. Th., Frei, R. W., and Lingeman, H., J. Chromatogr., 1989,492,251. Lawrence, J. F., Brinkman, U. A. Th., and Frei. R. W., J. Chromatogr., 1979, 185,473. Barcelo, D., Durand, G., Vreeken, R. J., de Jong, G. J., and Brinkman, U. A. Th., Anal. Chem., 1990,62, 1696. de Kok, A., Hiemstra, M., and Vreeker, C. P., Chromato- graphia, 1987, 24, 469. Goewie, C. E., and Hogendoorn, E. A., J. Chromatogr., 1987, 404.352. Miles, C. J., and Anson Moye, H., Chromatographia, 1987,24, 628. Tuinstra, L. G. M. Th., and Kienhuis, P. G. M., Chroma- tographia, 1987, 24, 696. Simon, V. A., and Taylor, A., J. Chromatogr., 1989,479, 153. Rosen. J. D., Applications of New Mass Spectrometry Tech- niques in Pesticide Chemistry, Wiley, New York, 1987, chs. 11-13, 16 and 17. Tomer, K. B., and Parker, C. E., J. Chromatogr., 1989, 492, 189. Voyksner, R. D., Bursey, J. T., and Pellizzari, E. D., Anal. Chem., 1984,56, 1507. Chiu, K. S., van Langenhove, A. V., and Tanaka, C., Biomed. Environ. Mass Spectrom., 1989, 18, 200. Cairns, T.. Siegmund, E. G., and Stamp, J., J. Rapid Commun. Mass Spectrom., 1987, 1,89. Rudewicz, P. J., Finnigan Mat Application Report Number 211, San Jose, CA, 1988, pp. 1-7. Bellar. T. A.. and Budde, W. L., Anal. Chem., 1988,60,2076. Barcelo. D., Biomed. Environ. Mass Spectrom., 1988, 17, 363. Barcelo, D., and AlbaigCs, J., J. Chromatogr.. 1989, 474, 163. Barcelo, D., LC-GC. 1988, 6, 324. 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 Betowski, L. D., and Jones, T. L., Environ. Sci. Technol., 1988, 22, 1430. Barcelo, D., Org. Mass Spectrom., 1989, 24, 219. Hammond, I., Moore, K., James, H., and Watts, C., J. Chromatogr., 1989,474, 175. Barcelo, D., Org. Mass Spectrom., 1989, 24, 898. Kalinoski, H. T., Wright, B. W., and Smith, R. D., Biomed. Environ. Mass Spectrom., 1986, 13, 33. Wright, B. W., and Smith, R. D., HRC CC, J. High. Resolut. Chromatogr. Chromatogr. Commun., 1985, 8, 8. Kalinoski, H. T., Udseth, H. R., Wright, B. W., and Smith, R.D., J. Chromatogr., 1987,400,307. Berry, A. J., Games, D. E., Mylchreest, I. C., Perkins, J. A., and Pleasance, S., Biomed. Environ. Mass Spectrom., 1988,15, 105. France, J. E., and Voorhees, K. J., HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 1988. 11, 692. Kalinoski, H. T., and Smith, R. D., Anal. Chem., 1988,60,529. Shah, S., and Taylor, L. T., HRC CC, J. High Resolut. Chromatogr. Chromatogr. Commun., 1989, 12, 599. Davies, I. L., Xu, B., Markides, K. E., Bartle, K. D., and Lee, M. L., J. Microcolumn Sep., 1989, 1, 71. Morrissey, M. A., and Hill, H. H., Jr., J. Chromatogr. Sci., 1988, 27, 529. Tondeur, Y., Sovocool, G. W., Mitchum, R. K., Niederhut, W. J., and Donnelly, J. R., Org. Mass Spectrom., 1987,14,733. Nubbe, M. E., Adams. V. D., Watts, R. J., and Robinet-Clark, Y. S., J. Water Pollut. Control Fed., 1988, 60, 773. Brinkman, U. A. Th., J. Planar Chromatogr. Mod. TLC, 1988, 1, 150. Gurka, D. F., Betowski, L. D., Jones, T. L., Pyle, S. M., Titus, R., Ballard, J. M., Tondeur, Y., and Niederhut, W., J. Chromatogr. Sci., 1988,26,301. Han, J. S., and Weber, J. H., Anal. Chem., 1988, 60, 316. Burns, D. T., Glockling, F., and Harriott, M., Analyst, 1981, 106, 921. Jewett, K. L., and Brinckman, F. E., J. Chromatogr. Sci., 1981, 19, 583. Paper Ol00787K Received February 20th, 1990 Accepted March 8th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600681

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST. JULY 1991. VOL. 116 69 1 Comparison of the Measurement of Serum Cotinine Levels by Gas Chromatography and Radioimmunoassay Ian G. M. Anderson and Christopher J. Proctor BAT Fundamental Research Centre, Regent's Park Road, Southampton SO9 I PE, UK Lars Husager Medi-Lab AS, 7 Adelgade, DK- 1304, Copenhagen, Denmark Standard continine solutions, controls and human serum samples containing cotinine have been measured by both radioimmunoassay (RIA) and gas chromatographic (GC) techniques. Cross-checks on standards and controls showed good agreement. However, for samples containing > 50 ng ml-1 of cotinine, RIA gave results on average 60% higher than GC. Determinations by using RIA and GC on samples containing <7 ng ml-1 of cotinine gave no significant correlation. The importance of the age of the serum sample has been investigated, and it is suggested that the age may affect the determination when dealing with low levels of cot i n i n e.Keywords : Radioimm unoassa y; gas chromatography cross-c h ecks; cotin ine; human serum; en vironmen ta I tobacco smoke Cotinine, a major metabolite of nicotine, has been used as a biomarker in many studies of both smokers and non- smokers. 1 4 Principally two methods of analysis have been used i. e., gas chromatography (GC) and radioimmunoassay (RIA). Analysis for cotinine using GC is generally more precise and sensitive than RIA, but has the disadvantage of having a much lower sample throughput.' Typical levels of cotinine in the serum of smokers are about 300 ng ml-1 and in non-smokers exposed to environmental tobacco smoke (ETS) of the order of 1-2 ng ml-1.6 Radioim- munoassay has commonly been used in studies of non- smokers, usually because of the large number of samples to be analysed.3.7 The accuracy of such analysis is important as some workers have attempted to carry out risk analyses and made claims about the extent of the population's exposure to ETS based on such data.8.Y The aim of the investigation presented here was to compare data from GC and RIA on split samples of serum taken from a lifestyle study of Danish men.Standard solutions and control samples were also cross-checked. Experimental Procedure The radioimmunoassay was performed at Medi-Lab, Copen- hagen, using the method developed by Knight et al. 1 0 but using a different batch of rabbit anti-cotinine antibody.This QSI-based RIA improves on the work of Langone and Van Vunakis" by employing an absorbed antiserum in order to remove anti-bridge antibodies and hence improve the sensitivity of the assay. The method has an estimated lower limit of detection of 5 ng ml-1.12 Capillary gas chromatographic analysis was performed using the method detailed by Feyerabend and Russell.13 This method uses 0.1 ml of sample which is made alkaline and extracted with dichloroethane using 5-methylcotinine as an internal standard. Afer vortex mixing, a high-speed mini- centrifuge is used for phase separation. The dichloroethane extract is transferred into a GC sample vial, and 2 p1 are injected onto a 0.32 mm i.d. X 10 m FFAP (free fatty acid phase) capillary column (Hewlett- Packard), the compounds being detected by nitrogen-phosphorous specific detection.The lower limit of detection was found to be 0.1 ng ml-1. Cross-checks of Standards and Samples Cross-checks were made between the two techniques on both standard solutions and on samples. Radioimmunoassay stan- dards were prepared by G. Knight, and consisted of fresh horse serum spiked with cotinine at five levels ranging from 12.5-800 ng ml-1. Three further standards (75, 150 and 300 ng ml-1) were used as controls throughout the RIA analysis. Cross-check samples (36) were taken randomly from a set of 3382 serum samples collected and frozen at -20" C in 1985 as part of a lifestyle study of Danish men. Aliquots of both these samples and RIA standards were transferred to the UK for GC analysis, packed in crushed dry ice in order to maintain their frozen state.Results The control standards at concentrations of 75, 150 and 300 ng ml-1 of cotinine were re-analysed on a number of occasions throughout the RIA analysis of the serum samples. The data from these analyses are presented in Table 1 where a comparison is made with GC cross-check results from duplicate determinations. The data suggest a consistent difference of about 5% between the results obtained by RIA and GC, with the RIA determination giving the higher values. A GC cross-check was also made against the five standard solutions used for RIA quantification using an independent set of standards made by spiking known concentrations of cotinine into bovine serum.The results are given in Table 2. As before, the data show a small variance, ranging from 4 to 6.9%, with the results from the RIA analysis being slightly higher. Table 3 presents GC and RIA cross-check data for the 36 randomly selected human serum samples. The results, ordered according to increasing levels of cotinine, as deter- mined by GC, are segregated into two groups, those contain- ing <7 ng ml-1 and those with >50 ng ml-1 of cotinine, no results being obtained for intermediate levels. Table 1 Mean cotinine levels (ng ml-1) found in controls as mcasured by RIA and GC RIA determination No. of GC deter- Mean value minations Mcan ( n = 2) 75 129 73 70 1 50 134 148 141 300 120 294 280 Control692 ANALYST, JULY 1991, VOL. 116 Table 2 GC cross-checks of RIA standards Table 4 Comparative RIA and GC results on cotinine levels (ng ml-1) from fresh human serum Cotinine in RIA standard/ ng ml-1 12.5 50 200 400 800 Cotinine measured by GC/ ngml-1 Variance (%) 11.7 6.4 48 4 189 5.5 377 5.8 745 6.9 Table 3 Comparative RIA and GC data on cotinine levels (ng ml-I) in human serum Levels of continine <7 ng ml-1 Levels of continine >50 ng ml-1 GC 0.1 0.2 0.2 0.3 0.4 0.5 0.6 0.8 1.7 2.1 3.0 3.3 3.6 3.8 5.1 6.5 RIA 18.5 <5 12.5 <5 <5 49.5 21.5 <5 7.5 <5 8.0 7.0 29.0 22.0 15.0 51.5 RIA : GC 185 <25 62.5 4 6 .7 <12.5 99 35.8 <6.3 4.4 <2.4 2.7 2.1 8.1 5.8 2.9 7.9 GC 53.3 145.4 149.2 159.2 164.0 165.1 221.2 229.5 246.0 273.3 279.7 281.8 288.1 304.6 330.7 334.6 348.9 352.6 359.7 385.2 RIA 86 220 275 285 285 275 420 405 355 355 535 450 365 580 495 525 430 580 520 535 RIA : GC 1.61 1.51 1.84 1.79 1.74 1.67 1.90 1.76 1.44 1.29 1.91 1.60 1.27 1.90 1.50 1.57 1.23 1.64 1.45 1.39 For the group of samples giving levels of cotinine >50 ng ml-1 by GC, the mean of the ratio of RIA : GC values is 1.60 (range 1.23-1.91).Hence, the RIA data suggest cotinine levels on average 60% higher than those obtained by GC. Analysis by GC resulted in a mean of 254 ng ml-1 of cotinine, whilst RIA gave a value of 399 ng ml-1. However, the GC and RIA data are highly correlated, with an analysis of variance giving an r2 value of 0.83 ( F = 85.2) which is significant at the 99% confidence level. For the samples with values of <7 ng ml-1 as determined by GC, the mean ratio of R1A:GC values is 29.9 (range 2.1-185).Analysis of variance between the RIA and GC data gives an r2 value of 0.16 ( F = 2.7), which is not significant at the 90% confidence level. Mean GC values from this set of samples is 2.0 ng ml-1, whilst the corresponding RIA value is 16.7 ng ml-1. At GC values of >50 ng ml-1 of cotinine (presumably corresponding to smokers), the data suggests a reasonably consistent bias of 60% between RIA and GC but little correlation is shown between the data sets for values of <7 ng ml-1 of cotinine (ie., from non-smoking subjects exposed to ETS). The serum samples tested in this study had been stored frozen for more than 4 years. In order to investigate whether the age of the samples was influencing the cross-check, ten fresh serum samples (five from smokers and five from non-smokers known be to be exposed to ETS) were acquired from Medi-Lab staff in Copenhagen and analysed both by RIA and GC.The results are given in Table 4. For smokers, GC analysis gave a mean cotinine level of 170.5 ng ml-1, whilst RIA gave a mean of 265 ng ml-1. The mean difference between the RIA data and GC data was 55.4% (similar to that found in aged serum samples). Analysis Non-smokers Smokers RIA R1A:GC GC RIA RIA:GC GC 1.4 10.5 7.5 42.9 65 1.52 1.4 11.0 7.9 78.7 127 1.61 1.9 11.0 5.8 186.2 311 1.67 2.7 13.5 5 .O 221.7 378 1.71 5.4 19.5 3.6 323.0 442 1.39 of variance gave an r2 value of 0.96 ( F = 73.0), significant at the 99% confidence level, and both of the analysis techniques ranked the samples in the same order.For serum from non-smokers, the mean level of cotinine was 2.6 ng ml-l by GC and 13.1 ng ml-1 by RIA (with a mean ratio of RIA : GC of 5.04). Analysis of variance gave an r2 value of 0.99 ( F = 261) which is significant at the 99% confidence level. Both the RIA and GC analyses ranked the subjects in the same order. Discussion This study suggests that there may be considerable differences in the quantification of cotinine in human serum dependent upon determination by GC or RIA. For levels found in smokers (generally >50 ng ml-1) RIA data gave on average values that were 60% higher than those obtained by GC. This difference is greater than that found in previous interlabora- tory cross-checks on serum, although similar differences have been found in urine sample~.~J4 As standards and controls gave similar results by GC and RIA, it is presumed that the differences derive from the nature of the interaction between the sample and the form of analysis.It is possible that the age of the initial samples had allowed some degradation resulting in an interference to the RIA, but this is unlikely as the difference between methods was similar for aged and fresh serum samples. It has been suggested that current RIA assays result in values for cotinine that are 30-50% higher than comparable GC data owing to RIA cross-reactivity of cotinine antibodies with other nicotine metabolites, primarily trans-3-hydroxy- cotinine .15 trans-3-Hydroxycotinine has been shown by Neurath and Pein16 to be the major nicotine metabolite in urine and a major metabolite in plasma.Another possible explanation for the difference between RIA and GC is that it has been shown that cotinine forms glucuronic acid conjugates. 17,18919 If the conjugates remain intact it is likely that they will be lost in the GC analysis during sample work-up, but may be still detected by RIA. This possibility requires further investigation. The difference in RIA and GC data for low levels of cotinine is somewhat disturbing. Data obtained using RIA was found to give much higher values than GC. Moreover, there was no significant correlation between the two sets of data obtained for the set of aged serum samples. For fresh serum, both the difference between RIA and GC values was reduced and the correlation between the two sets of data became highly significant.This suggests that the aged samples may have deteriorated in a manner that affects low level analysis, and this clearly has significant implications for the use of samples stored for several years, albeit frozen. Both the RIA and the GC analysis of serum cotinine are likely to be able to distinguish between regular smokers and non-smoking subjects. However, from the data produced in this study, the true levels of cotinine in serum remain uncertain. The study also provides a caution for research where data obtained from the determination of cotinine at low levels is related to exposure to ETS.ANALYST, JULY 1991. VOL. 116 693 The authors thank Dr. F. Gyntelberg and Dr. H. 0. Hein of the Rigshospitalet, Copenhagen, for the supply of serum samples used in this study.References Thompson, S. G., Stone, R., Nanchaahal, K.. and Wald, N. J., Thorax, 1990, 45, 356. Haley, N. J., Colosimo, S. G., Axelrad, C. M., Harris, R., and Sepkovic, D. W., Environ. Res., 1989, 49, 127. Sepkovic, D. W., and Haley, N. J., A m . J. Public Health, 1985, 75, 663. Benowitz, N. L., Hall, S. M., Herning, R. I., Jacob, P., Jones, R. T., and Osman, A-L., N. Engl. J. Med., 1983, 309, 139. Watts, R. R., Langone, J. L., Knight, G. J., and Lewtas, J., Environ. Health Perspect., 1990, 84, 173. Jarvis, M. J., Tunstall-Pedoe, H., Feyerabend, C., Vesey, C., and Saloojee, Y., A m . J. Public Health, 1987, 77, 1435. Greenberg, R. A., Haley, N. J.. Etzel, R. A., and Loda. F. A., N. Engl. J. Med., 1984,310, 1075. National Research Council, Environmental Tobacco Smoke: Measuring Exposure and Assessing Health Affects, National Academy Press, Washington DC, 1986.Cummings. K. M., Markello, S. J., Mahoney. M., Bhargava, A. K., McElroy, P. D., and Marshall, J. R., Arch. Environ. Health, 1990, 45, 74. 10 11 12 13 14 15 16 17 18 19 Knight, G. J., Wylie, P., Holman, M. S., and Haddow, J. E., Clin. Chem. (Winston Salem N.C.), 1985, 31, 118. Langone, J. J., and Van Vunakis, H. V., Methods Enzymol., 1982, 84,628. Knight, G. J., Foundation for Blood Research, Scarborough, ME, personal communication. Feyerabend, C., and Russell, M. A. H., J. Pharm. Pharmacol., 1990,42, 450. Biber, A., Scherer, G., Hoepfner, I., Adlkofer, F., Heller, W., Haddow, J. E., and Knight, G. J., Toxicol. Lett., 1987, 35, 45. Schepers, G., and Walk, R.-A., Arch. Toxicol, 1988, 62, 395. Neurath, R. B., and Pein, F. G., J. Chromatogr., 1987, 415, 400. Curvall, M., Kazemi Vala, E., and Englund, G., paper presented at the Euchem Conference Nic 1990, Visby, Sweden, June lOth-l4th, 1990. Curvall, M. Kazemi Vala, E., and Englund, G., paper presented at the Symposium of XIth International Congress of Pharmacology, Hamburg, June 28th-30th, 1990. Byrd, G. D., Chang, K-M., Greene, J. M., and de Bethizy, J. D., paper presented at the 44th Tobacco Chemists’ Research Conference, Winston-Salem NC, USA, October lst-3rd, 1990. Paper 010551 1 E Received December 7th, 1990 Accepted March Sth, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600691

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991, VOL. 116 695 Ion-exchange Chromatographic Determination of Anions by Indirect Photometric Detection: Comparison of Eluent Ions With Respect to Sensitivity Enhancement Shoji Motomizu, Mitsuko Oshima and Takashi Hironaka Department of Chemistry, Faculty of Science, Oka yama University, 3- I - I Tsushimanaka, Oka yama 700, Japan Aromatic sulphonate and carboxylate eluent ions were examined for use in the sensitive determination of inorganic anions by indirect photometric ion chromatography. The naphthalene-I ,3,6-trisulphonate ion was found to be the most sensitive for use as the eluent ion, the detection limit being as low as 1 x 10-8 mol dm-3. The naphthalene-I ,5-disulphonate ion is recommended for the analysis of water samples containing anions at concentrations of between 1 x 10-6 and 1 x 10-5 mol dm-3.These two eluent ions have several advantages over other choices: (i) detection is carried out at longer wavelengths (near 300 nm); (ii) the eluent ions are easily soluble in water and subsequently stable; (iii) their elution strength is not influenced by pH change; (iv) the eluent ions do not form any metal complexes; and (v) the reagents are inexpensive and commercially available. Keywords: Ion-exchange chromatography; inorganic anion; photometric detection; naphthalene- 1,5-disuI- phonate; naphthalene- 1,3,6-trisulphonate Ion-exchange chromatography (IEC) has been widely used in analytical chemistry for the separation of ionic species, which is often followed by the detection of the physical or chemical properties. A major breakthrough in analytical methods using IEC was the development by Small et a1.1 of ion chromato- graphy, which consisted of a separator and a suppressor column with a conductimetric detector.Later, Small and Miller2 reported the concept of indirect photometric chromat- ography (IPC). Since then, many workers have reported studies using IPC, and have proved its analytical advantages. Although the theoretical aspects of the IPC technique have been thoroughly evaluated and practical applications of IPC to the determination of anions and cations have been reported, little has been published on the characteristics of the eluent ions. In IPC, the eluent ions play an important role in enhancing the analytical sensitivity and sample throughput. In IPC of anions, Small and Miller2 examined iodide (mono- valent), sulphobenzoate and phthalate (divalent), and trimes- ate (trivalent) ions.Some other workers have examined such eluent ions as isophthalate,3 the copper complex of ethylene- diaminetetraacetic acid (EDTA),4 benzene-l,2,3-tricarboxyl- ate, benzene-l,2,4-tricarboxylate and benzene-l,2,4-5-tetra- carboxylate,5>6 Tiron' and 7-hydroxynaphthalene-1,3-disul- phonate ions.* Small has also recommended the eluent ions nitrate, iodide, benzoate, phthalate, sulphobenzoate and Cu(EDTA) for anionic analytes in a recent publication.9 In this work, various aromatic carboxylate and sulphonate ions are evaluated for their suitability for the enhancement of sensitivity in anion determination. In indirect photometric detection, the total ionic values for the anion and the cation in an effluent are always constant, according to the electroneutrality (charge balance relation- ship).Thus, the following equation can be written for anions: rnc = rn[Em-] + n[S"-] (1) where [Em-] and [S"-] are the concentrations of an rn valent eluent ion and an n valent sample ion in the effluent, respectively, and c is the concentration of the eluent ion. The absorbance ( A ) of the effluent is written as follows: where E~ and ES are the molar absorptivities of the eluent and the sample ion, respectively, and 1 is the pathlength of the flow cell; here 1 cm is used. Hereafter, the sample ions are considered ultraviolet (UV) transparent. The peak height (PH, in cm), which is recorded as a negative peak from a high background absorbance, is given by eqn.(3): PH = (-UAU) ( E E / ~ ) ( n [ S n - ] ) (3) where AU is the absorbance corresponding to a 1 cm PH on the recorder. From eqn. (3), to obtain a higher peak, an eluent ion with a large value of EE/m can be used and a detector condition with a small AU can be adopted. When a 1 x 10-4 mol dm-3 eluent ion solution is used to determine analyte ions at 1 x 10-6 mol dm-3 in the sample, the molar absorptivity of the eluent ion necessary to obtain a 1 cm PH can be calculated from eqn. (3), provided that the analyte ions are monovalent and are hardly diluted when they pass through the UV detector, and AU for the detector is adjusted to 0.001. The ~ ~ / m value is calculated as 1 X 103. Noise, which is mainly caused by the UV detector (and is dependent on the quality) and pulsation of the pump, is assumed to be proportional to the absorbance of the background ( E ~ c ) .Noise kEEC (4) or Noise/cm = ~EEc/AU ( 4 4 where k is a proportionality constant, which is determined by the quality of the detector. Thus, the sensitivity corresponding to the signal-to-noise (SIN) ratio is given by eqn. (5): Sensitivity = S/N = n[S"-]/krnc (5) Eqn. ( 5 ) shows that higher sensitivity will be obtained at lower concentrations of a lesser charged eluent ion. Another problem occurs when a low concentration of an eluent ion is used: the lower the concentration, the longer the retention times of the analyte ions and the shorter the peaks. Thus, a more concentrated eluent ion is required. Real samples often contain other substances that absorb UV light at 200-250 nm.Such substances sometimes give a serious error. Thus, an eluent ion that absorbs UV light at longer wavelengths, i . e . , near 300 nm, is recommended. In conclusion, the following characteristics are desirable in an eluent ion: (i) UV light is absorbed at longer wavelengths, (ii) a large molar absorptivity, (iii) an adequate elution ability at low concentrations and (iv) a low valency. Furthermore, when in practical use, the eluent has to be stable for a long period of time and over a wide pH range.696 ANALYST, JULY 1991, VOL 116 Experimental Apparatus The chromatographic equipment consisted of a dual-head reciprocating pump (Tosoh HLC-803D), a sample injection valve (Rheodyne 7125) with a loop (20 or 100 pl), anion- exchange columns, and a UVNIS spectrophotometric detec- tor (Tosoh UV-8000).Two different types of columns (both 50 X 4.6 mm id.) were used: one was packed with TSK gel IC-Anion-PW (a methacrylate-based anion exchanger; par- ticle size 10 pm; capacity 0.03 k 0.005 mequiv ml-I), the other was packed with TSK gel IC-Anion-SW (a silica-based anion exchanger; particle size 5 pm; capacity 0.4 k 0.1 mequiv g- 1). The columns were kept in a water-bath (30 "C) while in use. The eluent was propelled at a flow rate of 1.0 ml min-1. A Shimadzu UV-300 recording spectrophotometer was used to measure the UV spectra of the eluent ions. Reagents Eluent substances, carboxylic acids and sulphonic acids of benzene and naphthalene, except for naphthalene-1 ,&dicarb- oxylic acid, were available commercially (Tokyo Kasei Kogyo Co.), and were used after recrystallization from water or a mixture of water and ethanol.Naphthalene-1 ,8-dicarboxylic acid was prepared by the hydrolysis of its anhydride in a sodium hydroxide solution at 80 "C and acidification of the product. The product was recrystallized from ethanol. The eluent substances of acidic type were dissolved in an equi- molar sodium hydroxide solution, and those of the sodium salt type were dissolved in distilled water. Stock solutions of CI- , NO3- and S042- were prepared by dissolving sodium chloride, potassium nitrate and sodium sulphate, respectively, after drying the salts at 60 "C under reduced pressure (5 mmHg). Working solutions were prepared daily by accurate dilution.Results and Discussion Detection Wavelength Some of the absorption spectra of the eluent solutions are shown in Fig. 1. The absorption maxima of the naphthalene derivatives are at longer wavelengths than those of the benzene derivatives. The molar absorptivities of the benzene derivatives, near 280-300 nm, are about 1 x 103 0.6 8 e 2 (0 9 0.4 0.2 I I 260 280 300 Wavelengthlnm Fig. 1 Absorption spectra of eluent ions: 1, 1,5-NDS; 2, 2,6-NDC; 3, 2,6-NDS; 4, 1,3,6-NTS; 5 , 1,2,4,5-BTeC; 6, 1,3,5-BTC; and 7, picrylsulphonate ion. Concentration, 8 x 10-5 mol dm-3 Table 1 Comparison of eluent ions (1 x 10-3 mol dm-3) Eluent anion (charge) Benzene derivatives- -c02- (-1) 1,2-COXII- (-2) 1,3-CO2- (-2) 1,4-CO2- (-2) 3-NO2,I ,2402- (-2) 4-C1,1,2,-CO2- (-2) 4-SO3-, 1,2-CO2- (-3) 1,2,3-CO2- (-3) 1,2,4-CO2- (-3) 1,3,5-CO2- (-3) 1,2,4,5-CO2- (-4) 1,3-s03- (-2) 1-SO3- (-1) 2-so3- (-1) 1-0- ,4-S03- (-2) 1,5-SO3- (-2) 2,6- S O3 - (-2) 2,7-S03- (-2) 2-0- ,3,6-S03- (-3) 1,3,6-S03- (-3) 1 ,8-C02- (-2) 2,3-CO2- (-2) 2,6-C02- (-2) Naphthalene derivatives- 2,3-0- ,OH ;6-S03- (- 2) * IC-Anion-PW column.t IC-Anion-SW column. Abbreviation 1-BC 1,2-BDC 1,3-BDC 1,4-BDC 1,2-BDC-N02 1,2-BDC-CI 1,2-BDC-S03- 1,2,3-BTC 1,2,4-BTC 1,3,5-BTC 1,2,4,5-BTeC 1,3-BDS 1-NS 2-NS 4-NS-0- 6-NS-0- ,OH 1,5-NDS 2,6-NDS 2,7-NDS 3,6-NDS-0- 1,3,6-NTS$ l,8-NDC 2,3-NDC 2.6-NDC $ Eluent ion, 2.5 X mot dm-3. Detection Log k'* Log k' t wavelength/ Y nm 250 255 254 280 310 265 265 256 278 257 289 240 300 290 329 320 310 290 290 280 305 320 297 305 c1- 1.16 0.63 0.41 0.50 0.55 0.54 0.17 0.33 0.21 0.06 -0.01 0.36 0.36 0.48 0.07 0.15 0.13 0.13 0.14 -0.14 -0.13 0.54 0.37 0.23 N03- 1.52 1.04 0.84 0.86 0.95 0.87 0.59 0.77 0.64 0.47 0.47 0.67 0.64 0.71 0.32 0.39 0.43 0.44 0.44 0.12 0.20 0.94 0.74 0.55 so42- Very large 1.20 0.92 0.98 1.07 1.04 0.26 0.52 0.31 0.10 - -0.07 0.77 0.94 1.15 0.56 0.58 0.42 0.41 0.41 -0.12 -0.05 0.99 0.73 0.55 c1- - 0.69 0.61 - - - - - 0.09 -0.07 - 0.53 - - - - 0.33 0.32 - - - - 0.49 0.35 NO3- - 0.87 0.77 - - - - - 0.29 0.11 0.69 - - - - - 0.47 0.46 - - - - 0.67 0.50 S042- - 1.54 1.33 - - - - - 0.48 0.24 1.23 - - - - - 0.86 0.86 - - - - 1.14 0.91ANALYST, JULY 1991, VOL.116 697 dm3 mol-1 cm-*, whereas those of the naphthalene deriva- tives are several times larger. In respect of the enhancement of the sensitivity and the reduction of the interference from UV absorbing substances, naphthalene derivatives are recom- mended for use as the eluent ions.Comparison of the Elution Strength of Carboxylate and Sulphonate Ions The logarithmic values of the capacity factors ( k ' ) for chloride, nitrate and sulphate ions with various eluent ions are shown in Table 1. The parameters contributing to eluent ability for analyte anions include the charge and size of the eluent ion and the anionic group and its site, in addition to the concentration of the eluent ions. From Table 1, the following general aspects for elution ability are seen: (i) the larger the charge, the stronger the elution ability; (ii) sulphonate ions are stronger than carboxylate ions; (iii) naphthalene derivatives are stronger than benzene derivatives; (iv) groups such as NOz and CI strengthen the elution ability, whereas OH groups weaken it; and (v) a pair of adjacent charged groups contribute to the elution ability less than a separated pair of charged groups.The general aspects [(ii)-(v)] are in agree- ment with those seen commonly in ion-pair extraction. In IPC, phthalate ions (benzene-l,2-dicarboxylate ions; 1,2-BDC) have often been used. However, 1,2-BDC is not a powerful eluent ion and does not absorb UV light at longer wavelengths. Of the benzenedicarboxylate ions, the nitro derivative of 1,2-BDC is the most useful for sensitive determinations. In respect of the elution ability and the detection wavelength, naphthalenecarboxylate and naphthalene- sulphonate are favourable.However, the elution ability of carboxylate is severely affected by pH change. As a result, the most powerful and sensitive eluent ion of those examined is the naphthalene-l,3,6-trisulphonate ion (1,3,6-NTS). Naph- thalenedisulphonate ions are recommended as eluent ions: they absorb UV light at longer wavelengths; possess mod- erately strong elution ability; and are not affected by pH change. In practical use, the correct concentration of an eluent ion must be selected on the basis of the separation, the retention time and the concentration of analyte ions, although lower concentrations of an eluent ion are advantageous for sensitivity enhancement. Examples of chromatograms obtained with 1,3,6-NTS are shown in Fig.2. A sample containing small amounts of analytes could be separated effectively and analysed sensitively. In a sample that contains large amounts of analytes, however, the peaks are very broad because of overloading. Thus, 1,3,6-NTS is useful for samples containing very small amounts of anions. Time- Fig. 2 Chromatogram obtained with 1,3,6-NTS as an eluent ion: concentration, 5 x mol dm-3; pH, 5.7; and A, 280 nm. Concentration of C1-, 2 x mol dm-3: and NO2-, NO3- and S012-, 1 X mol dm-3. Hydrogen carbonate ion is contaminated from air Relationship Between the Capacity Factor of an Anion and the Concentration of an Eluent Ion In IEC, the following equation generally holds: Log k' = -n/m X log [Em-] + l/m x (log KE + nlog[Em-I,) (6) K E = [Sn-],m [Em-]"/[Sn-Im [Em-]," where KE is an equilibrium constant or a selectivity coeffi- cient, and the brackets [ ] and [ Ir indicate the ion concentration in mmol ml-1 for the solution phase and in mmol g-1 for the resin phase, respectively.Eqn. (6) shows that a plot of log k' against log[Em-] is a straight line with a slope of -n/m under the conditions where [S"-l], is negligible compared with EM-]^, and the selectivity coefficient is constant. Typical plots of log k' against log[eluent] are shown in Fig. 3. The slopes for several eluent ions, which are in 1.6 1.2 L 3 0.8 0 0.4 0 \ 0' '\ 1 '\ '\ 1 I I I I -3.4 -3.2 -3.0 -2.8 -2.6 Log ([eluentl/mol dm-3) Fig. 3 1,2-BDC and 2, 1S-NDS: V , Graph of log k' against log ([eluent]/mol dm-3) for 1, a, NO,-; and 0, C1- ~~ Table 2 Slopes of the plots of log k' versus log[eluent ion] Slope Eluent anion IC-Anion-PW column- 1,ZBDC 1,3-BDC 1,4-BDC 1,2,3-BTC 1,2,4-BTC 1,3,5-BTC 1,2,4,5-BTeC 1,8-NDC 2,3-NDC 2,6-NDC 1,5-NDS 1,3,6-NTS IC-Anion-SW column- 1,2-BDC 1,2,4-BTC 1,3,5-BTC 1,3-BDS 2,3-NDC 2,6-NDC 1,5-NDS 2,6-NDS C1- -0.60 -0.57 -0.57 -0.34 -0.37 -0.40 -0.31 -0.64 -0.62 -0.59 -0.56 -0.50 -0.52 -0.39 -0.46 -0.62 -0.58 -0.53 -0.61 -0.66 -0.59 -0.56 -0.55 -0.35 -0.36 -0.39 -0.27 -0.62 -0.63 -0.57 -0.48 -0.42 -0.50 -0.32 -0.42 -0.61 -0.54 -0.52 -0.59 -0.64 S042- -1.13 -1.07 - 1.07 -0.68 -0.70 -0.73 -0.57 -1.14 - 1.20 -1.09 - 1.01 -0.85 - 1.03 -0.68 -0.75 -1.14 - 1.06 -1.02 -1.10 -1.14698 ANALYST, JULY 1991, VOL.116 Table 3 Comparison of capacity factors for anions Log k' value of eluent ion Ionic A Anion radius*/A value? AGhS 1S-NDSQ 2,6-NDS 2,3-NDC 2,6-NDC 1,3-BDS 1,2-BDC 1.4-BDC 1,3,5-BTC 1 S-NDSfl 1 0 ~ - F- Br03- Cl- NOz- c103 - Br- N03- I- SCN- c104- 1.22 1.26 1.54 1.72 1.92 1.71 1.88 1.79 2.10 2.13 2.40 - -11.06 -8.08 -8.11 -6.89 -7.05 -5.32 -5.17 -5.03 - - - 0.185 -434 - - 0.298 -317 0.330 - 0.388 - 0.405 -303 0.454 -270 0.465 -257 0.710 - 0.850 -190 0.863 - 0.324 0.382 0.443 0.462 0.693 0.834 0.849 - - 0.493 0.565 0.646 0.669 0.946 1.107 1.102 - - 0.387 - 0.462 0.354 0.530 0.418 0.608 - 0.618 0.491 0.673 0.495 0.694 0.741 0.957 0.884 1.116 0.902 1.118 - - * Thermochemical ionic radii given by Jenkins and Thakur.lo t A parameter of the extractability of anions obtained in solvent extraction of ion pairs (log unit)." $ Free energy of hydration, kJ g-1 ion-1.12.*3 9 1 x 10-3 mol dm-3 eluent ion, IC-Anion-SW column.7 4 x mol dm-3 eluent ion, IC-Anion-PW column. - - -0.259 - - - - - -0.141 0.636 0.474 -0.066 0.700 0.543 0.019 - 0.023 0.760 0.600 0.084 0.796 0.605 0.110 1.043 0.834 0.382 1.189 0.957 0.551 1.207 0.964 0.565 - 0.179 0.162 0.348 0.377 0.510 0.648 0.603 0.704 1.134 1.493 1.691 agreement with those expected from eqn. ( 6 ) , are shown in Table 2. The slopes for sulphate are twice those for mono- valent anions, although experimentally obtained slopes (absolute values) are slightly larger than those expected from eqn. (6). Comparison of Capacity Factors of Monovalent Anions The values of log k' for several anions are shown in Table 3. In practical analyses, it is important to identify each peak among the many present.The values of log k' for anions increase with an increase in the extractability of the anions (A value in Table 311) into an organic phase as ion associates; that is, the bulkier the ion, the more it is retained at the ion exchanger, and is extracted into an organic solvent as an ion associate. In the solvent extraction of ion associates, the extractability of anions shows a linear relationship to the reciprocal of the ionic radius (Y). 11 Reichenberg14 showed that in ion exchange selectivity the net change in free energy (AGM/N) on removing an ion, M, from solution and exchanging M for an ion, N, on the functional group of the resin (A) can be expressed as follows: -AGMIN = [e2/(rA + rN) where AGh(M) and AGh(N) are the standard free energies of hydration of the ions M and N, respectively, and e is the electron charge.It was concluded that the electrostatic/ hydration theory on the basis of eqn. (7) explained the order of both cation and anion selectivities. The term in square brackets on the right-hand side can be said to be the electrostatic interaction between the functional group and the counter ions and the term in parentheses to be the difference in the hydration or hydrophobicity of the ions. The relationship between log k' and the reciprocal values of the thermochemical ionic radii of anions is shown in Fig. 4. In anions that have radii larger than the radius of the chloride ion, a linear relationship was obtained, except for SCN-, N02-, NO3-, and C103-, although the slopes are different from each other owing to variation of the type of ion-exchange packing and the charge number of the eluent ions.In the region where the linear relationship is valid, the term in square brackets will be small, and the term in parentheses will be the predominant one. Thus eqn. (7) can be written as - e2/(rA + YM)] - (AGh(N) - AGh(M)) (7) Log k' cc -AGM,N AGh(N) -k AG,(M) = Constant -t AGh,,) ( 7 4 1.6 1.4 1.2 1 .o 0.8 0.6 0.4 0.2 0 -0.2 -1 200 300 400 500 C .- & d I 0.4 0.5 0.6 0.7 0.8 0.9 r-IIA-1 Fig. 4 Graph of log k' against the reciprocal value of the thermo- chemical ionic radius ( r ) . and hydration energy (AGh) against r-l. Eluent ion: 0, 1,5-NDS (IC-Anion-PW column): a, 1,3-BDS (IC-Anion-SW column); and 0, 1,3,5-BTC (IC-Anion-SW column). A , C104-; B. SCN-; C, I-; D, NO2-; E, Br-; F, N03-; G, CI03-; H, Br03-; I , F-; and J.1 0 3 - where AGh,,) is the hydration energy of the sample ion, in this instance, the anion. The plot of AGh against Y-1 is almost linear (as shown in Fig. 4), and hence, the plot of log k' against Y-1 shows a linear relationship, according to eqn. (7a). One of the reasons why SCN-, NO2-, NO3- and C103- ions, and probably two other ions, Br03- and 103-, deviate from linearity is that these ions are not spherical but rod-like or planar. Contrary to the expectation from the ionic radius and eqn. (7a), the fluoride ion, in addition to IO3-, exhibits a large value of log k ' . This is probably because in these ions the electrostatic interaction between the functional group and the counter anions becomes predominant; that is, the term in square brackets in eqn.(7) is more important than the term in parentheses.ANALYST, JULY 1991, VOL. 116 699 For different eluent ions E l m l - and E2m*-, the capacity factors, log kl’ and log k2’, are written from eqn. (6) in the following equations, eqns. (8) and (9). LOgk1’ = (-n/m1) X log [ E l m l - ] + (l/ml) X (IogKE, + nlog [Elm1-Ir) + (l/m?) X (IogKE, + nlog [Ezmz-],) (8) (9) Lo&*’ = (-n/m?) X log [E*”2-] From eqns. (8) and (9), the following equation can be derived: Log (kl ’lk2’) = -nlog( [E 1m I--]l’m I/[ E2m2-]l’”Q) + log(KE, I’mI/KEzl’m*) + n l 0 g ( [ E l m l - ] ~ ” m l / [ E 2 m 2 - ] ~ ” ~ 2 ) ( 10) If the concentrations of the eluent ions in the aqueous solution are assumed to be identical, for certain analyte ions, in addition to those in the ion-exchange resin, and the charge numbers of certain analyte ions are identical and are n , the first and third terms of the right-hand side are constant.Thus, the following equation is derived: Log k l ’ = constant + log(KE11/ml/KE21/m2) + log k2’ (11) The plots of log k‘ for several eluent ions against log k’ for 1,5-NDS are shown in Fig. 5. The plots are all straight lines with a slope of unity. These results show that the value of log( KElllml/KEzllm?,) is constant, regardless of the type of analyte ions. This relationship is very useful for the prediction of the capacity factors of analyte ions. 1.2 1 .o 0.8 4 0.6 u) -I 0.4 0.2 0 -0.2 Recommended Eluent Ions in Practical Analysis For the most sensitive determination of anions, 1,3,6-NTS is the recommended ion of the eluent ions examined in this work.For good separation, however, the total concentration of anions in a sample must be less than that of the eluent ion. The detection limit, corresponding to an S/N ratio of 2, for monovalent anions is as low as 1 x 10-8 rnol dm-3. The most versatile eluent ions are naphthalenedisulphonate ions such as the 1,5-, 2,6- and 2,7-derivatives. They all exhibit similar elution strengths; however, the 1,5-derivative strongly absorbs UV light near 300 nm (see Fig. 1). Other advantages of this eluent ion are as follows: (i) its elution ability is moderate and the reagent is easily soluble in water; (ii) it is a chemically stable reagent; (iii) its elution ability does not change over a wide pH range; (iv) the reagent is less expensive and easily available; and ( v ) the reagent does not form any complexes with metal ions.A typical chromatogram for several anions is shown in Fig. 6 and those for water samples in Fig. 7: the peaks for C1-, N03- and S042- are completely resolved within 5 min. In order to improve the background, the river water samples were treated with a cartridge type pre-treatment column packed with cation-exchange resin (Toyopak IC-SPM, Na+ type) and Ca2+ and Mgz+ were eliminated simultaneously (Fig. 7(c)]. I I I I I I 1 I I I I I I I I I 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 Retention time/min Fig. 7 Chromatograms of ( a ) rain water; ( b ) river water, without pre-treatment; and (c) river water with pre-treatment. Eluent.4 X rnol dm--’ 1.5-NDS; A, 305 nm; and sample volume: ( a ) 100 pI and (b) 10 pl. 1. [Cl- = 6.7 x 10-6 rnol dm-3; 2, [NO3- = 3.5 x rnol dm--’; 3, [SO4’-\ = 11.7 x rnol dm--’; 4, [Cl-/= 2.5 x 1 0 F rnol dm--’; 5 , [NO3-] = 4.0 x rnol dm--’; and 6, (SO4*-] = 1.0 x 10-4 rnol dm-’ Fig. 5 Graph of log k ‘ of eluent ion against log k‘ of 1.5-NDS: 0. 1.2-BDC; 0. 1.3-BDS; 0.2.6-NDC; V. 2.6-NDS; and W, 1,3,5-BTC. Column; IC-Anicrl-SW Table 4 Detection limits and linear ranges of calibration graphs with 1.5-NDS ion as the eluent ,2 min, CI - Time - Fig. 6 Typical chromatogram. Eluent. 2 x lo--‘ rnol dm--’ 1S-NDS; A, 290 nm; flow rate. 1.0 ml min-1; sample volume, 10 PI; and concentration of anions. 1 x 10-4 rnol dm--‘ Sample volume 20 pl* loo PI* 20 yl? 100 pl-i Lower limit$/l0-6 rnol dm--’ c1- 0.57 0.12 1.7 0.13 NOJ- 1.8 0.36 3.6 0.46 sop 2.1 0.22 1.8 0.20 CI - 1.6 0.30 5.6 0.50 NO-’- 2.4 0.60 7.2 0.70 SO,’- 1.6 0.30 2.8 0.30 Upper limit9/10-4 rnol dm--‘ * [1.5-NDS] = 2 x 10-1 rnol dm--’ (A = 300 nm).-1 [1.5-NDS] = 4 x rnol dm--’ ( h = 305 nm). $ Detection limit corresponding t o an S/N ratio of 3. B Below these limits, calibration graphs are linear.700 ANALYST, JULY 1991, VOL. 116 2 r 4 Time- Fig. 8 Chromatograms of rain water sampled on ( a ) February 14th, 1986 and (b) February 18th, 1986: eluent, 6 X mol dm-3 1,2,4,5-BTeC; h, 250nm; flow rate, 1.1 ml min-1; and sample volume, 20 p1. 1, [Cl- = 1.7 x 10-5 mol dm-3; 2, [S042-] = 5.8 x 10-5 rnol dm-3 ; 3, /NO3-] = 3.71 x 10-5 mol dm-3; 4, [Cl-] = 3.85 x 10-5 mol dm-3; 5, [S042- = 2.17 x rnol dm-3; and 6, [NO3-] = 1.05 x 10-5 mol dm- 1, The detection limits and linear ranges of the calibration graphs are summarized in Table 4.Lowering the concentra- tion of the eluent ion does not necessarily improve the detection limits, because the retention times are longer with a decrease in the concentration of the eluent ion. An increase in sample volume improves the detection limit. Anions at concentrations of from 1 x 10-7 to 1 X 10-4 rnol dm-3 can be determined by varying the eluent concentration and the sample volume. The other eluent ion for the sensitive determination of anions is 1,2,4,5-BTeC. The retention times of the analyte ions, however, are subject to pH and any metal ions present, because the acid dissociation constants of 1,2,4,5-BTec are 1.92,2.82,4.49 and 5.64 and the eluent ion can form a chelate through its carboxyl groups. The chromatogram obtained for a rain water sample with this reagent is shown in Fig. 8. In this instance, S042- is between C1- and NO3-, and NOz-, if present in the water sample, is shielded by S042-. The authors are grateful to the Steel Industry Foundation for the Advancement of Environmental Protection Technology (SEPT, Japan) for financial support. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Small, H., Stevens, T. S., and Bauman, W. C., Anal. Chem., 1975,47, 1801. Small, H., and Miller, T. E., Anal. Chem., 1982, 54, 462. Naish, P. J., Analyst, 1984, 109, 809. Hayakawa, K., Sawada, T., Shimbo, K., and Miyazaki, M.. Anal. Chem.. 1987,59, 2241. Motomizu, S . , Sawatani, I., Hironaka, T., Oshima, M., and Tdei, K., Bunseki Kagaku, 1987, 36, 77. Hironaka, T., Oshima, M., and Motomizu, S., Bunseki Kagaku, 1987,36, 503. Sato, H., Anal. Chim. Acta, 1988,206, 281. Yokoyama, Y., and Sato, H., J. Chromarogr. Sci., 1988, 26, 561. Small, H., Zon Chromatography, Plenum Press, London, 1990, p. 199. Jenkins, H. D. B., andThakur, K. P., J. Chem. Educ., 1979,56. 576. Motomizu, S.. Bunseki Kagaku, 1989,38, 147. Rosseinsky, D. R . , Chem. Rev., 1965, 65, 467. Vasil’ev, V. P., Zolotasev, E. K., Kapustinskii, A. F., Mishchenko, K. P., Podgornaya, E. A., and Yatsimirskii. K. B., Zh. Fiz. Khim., 1960, 34, 1763. Reichenberg, D., in Ion Exchange, ed. Marinsky, J. A., Marcel Dekker, New York, 1966, vol. 1. Paper I l00024A Received January 2nd, 1991 Accepted February 26th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600695

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991, VOL. 116 701 Spectrophotometric Flow Injection Procedure for the On-line Monitoring of Sulphite in High Ionic Strength Brine Paul MacLaurin and Paul J. Worsfold* Department of Environmental Sciences, Polytechnic South West, Drake Circus, Plymouth, Devon P148AA, UK Alan Townshend School of Chemistry, University of Hull, Hull HU6 7RX, UK Neil W. Barnettt I.C.I. Chemicals and Polymers Ltd., The Heath, Runcorn, Cheshire WA7 4QD, UK Michael Crane I. C. I. Engineering Department, Winnington, North wich, Cheshire CW8 4DJ, UK A spectrophotometric flow injection procedure for the determination of sulphite in potassium chloride brine is described. The reagent used was the organic disulphide 2,2’-dinitro-5-5’-dithiodibenzoic acid and the effect of the reaction pH on the kinetics is discussed.A laboratory method for the determination of sulphite in the range 0.1-20 mg 1-1 is presented for three matrices; water, standard potassium chloride solutions and potassium chloride brine. The robustness of the method is discussed in terms of sample pH, sample temperature and potassium chloride concentration. A modified procedure for on-line analysis with a linear range of 3-100 mg 1-1 is described and the results are compared with those obtained off-line by iodimetric titration. The relative standard deviation for 168 results (triplicate determinations of a 62.3 mg I-’ sulphite standard) over a 7 d cycle was 2.1 %. Results are also presented for a 21 d on-line trial. Keywords: Sulphite; flow injection; spectrophotometry; process analysis; brine Flow injection (FI) is being increasingly applied to the on-line chemical monitoring of process streams.1-7 Flow injection has many attractions for process monitoring,*-11 particularly its short response time which provides real time pseudo-continu- ous data. In addition, on-line sample treatment improves both sensitivity and selectivity, FI instrumentation can be simple, inexpensive and robust, additionally, consumption of reagent is potentially very low. As a detection system, spectropho- tometry suffers little detector fouling and offers a large range of established laboratory methods. These factors, coupled with microcomputer control and autocalibration, render long term unattended operation and reliable process control possible. During the large scale production of potassium hydroxide and chlorine by the electrolysis of potassium chloride brine, potassium sulphite is added to the electrolyte stream as a chlorine scavenger.For plant control and economic reasons the continuous monitoring of sulphite in 18% m/m potassium chloride brine at 70-80 “C and pH 11-12 is required on a near real time basis. In a previous study6 a spectrophotometric FI procedure for the determination of sulphite in aqueous media was described, and a method for the on-line determination of sulphite in high ionic strength potassium chloride brine proposed. The methodology was based on the reaction of sulphite with the organic disulphide 2,2’-dinitro-5-5’-dithio- dibenzoic acid (DNTB). 11.12 Rigorous study of the suitability of this procedure for process analysis revealed a sensitivity to sample pH in the pH range 11-12.This paper describes a study of the effect of pH on the reaction of DNTB with sulphite and presents a more robust FI manifold for on-line application. The design and operation of a fully automated spectrophotometric FI analyser3.4.7 is outlined and the results from a 21 d on-line trial are presented. The performance of the monitor was validated by comparison of the results with those from a standard off-line iodimetric procedure. * To whom correspondence should be addressed. t Present address: Department of Chemical and Analytical Sciences. Deakin Univcrsity, Geelong, Victoria 3217. Australia. Experimental Reagents All solutions were prepared in distilled, de-ionized water and all reagents were of AnalaR grade (BDH) unless otherwise indicated.A 1000 mg 1-1 sulphite (as S032-) stock solution was prepared by dissolving 1.5743 g of sodium sulphite (dried for 2 h at 105OC) in 1 1 of 1 x 10-3 mol dm-3 ethylenediam- inetetraacetic acid (EDTA) (0.3722 g of EDTA disodium salt dissolved in 1 1 of water). Appropriate dilutions of this stock solution were made in the three matrices for the respective calibrations. Sulphite is rapidly oxidized during the prepara- tion of aqueous solutions and hence the concentration of stock solutions was determined iodimetrically. A solution contain- ing an excess of iodine was acidified with hydrochloric acid, then sulphite solution was added carefully with stirring. The remaining iodine was titrated with sodium thiosulphate.13 Solutions of DNTB (Aldrich) were prepared by dissolving an appropriate amount in ethanol ( 5 ml 1 - 1 of solution) and diluting with pH 6.9 buffer.The pH 6.9 buffer was prepared by dissolving 3.55 g of disodium hydrogen phosphate and 3.41 g of sodium dihydrogen phosphate in 1 1 of water. The pH 9.9 buffer was prepared by dissolving 19.07 g of disodium tetraborate decahydrate (borax) and 2.0 g of sodium hydrox- ide in 1 1 of water. The pH 11.7 buffer was prepared by dissolving 3.80 g of trisodium phosphate (BDH; general purpose reagent) in 11 of water. All pH adjustments were made using hydrochloric acid or sodium hydroxide of various concentrations. Instrumentation and Procedures Batch experiments Measurement of the rate of change of absorbance at 412 nm was carried out using a diode-array spectrophotometer (Hewlett-Packard 8451A) fitted with a silica cuvette with a pathlength of 1 cm.In all experiments the response corre- sponds to the addition of a 3.0 ml aliquot of sample solution (0 or 16.7 mg 1 - 1 of SO3’-) to 0.3 ml of DNTB solution (4.0 g I - ’ ; 24-fold excess).ANALYST. JULY 1991, VOL. 116 702 Automated laboratory analysis A schematic diagram of the automated monitor is shown in Fig. 1. The FI manifolds were made up of 0.8 mm i.d. polytetrafluoroethylene (PTFE) tubing (Anachem) with PTFE T pieces for stream merging. The absorbance was monitored by a spectrophotometer (LKB Ultrospec 11) fitted with an 18 pl silica flow cell with a pathlength of 1 cm (Hellma), and the analogue output relayed to a strip-chart recorder (Chessel BD 40 04).Injections (20 pl) were made via a 12 V solenoid-activated injection valve (Chemlab Instruments) and standardhample selection was controlled by a two-way 12 V solenoid valve (Lee). Sample, reagent and carrier streams were propelled by two peristaltic pumps (Ismatec Mini $820) with poly(viny1 chloride) (PVC) pump tubing (Labsystems). A single board microcomputer (Control Universal) was used for control as shown in Fig. 1. The analogue output from the spectrophotometer was digitized (13 bit analogue to digital converter, Control Universal), and after data processing the result was sent to a miniature printer and a liquid crystal display. Software for control and data handling was written in BBC BASIC in a form that facilitated operator interaction, such that modifications to the sampling frequency, data acquisition mode and data treatment could be implemented.The three-line manifold (Fig. 2) is a modification of that previously reported6 and incorporates a pH 9.9 disodium tetraborate decahydrate-sodium hydroxide buffer. The sensi- tivity to sample pH was investigated by analysing standard solutions of 10 mg 1 - 1 of sulphite in water and in 20% m/v potassium chloride. The effect of sample temperature was studied by analysing a standard solution of 10 mg 1 - 1 of sulphite in 20% m/v potassium chloride at pH 11 .O maintained at various temperatures in a heated water-bath (Grant W14). The effect of potassium chloride concentration in the sample matrix was determined by analysing a standard solution of 10 mg 1-1 of sulphite prepared in potassium chloride solutions covering the range 0-20% m/v at pH 11.0.Reagent Pump 1 injection I valve , Data I acquisition1 I I ' T h i L g I - - - - I -- - - - - - - - Monitor co ntro I I I Sample process control corn puter output Fig. 1 Schematic diagram of the automated monitor 20 pl DNTB (2.5 x 10-3 mol dm-3) i Water 1.6 41 2 nm U ml min-' Fig. 2 high ionic strength potassium chloride brine Flow injection manifold for the determination of sulphite in On-line analysis The instrumentation used for on-line analysis was as described above and the FI manifold is shown in Fig. 3. Brine, at a temperature of 70-80 "C, was drawn from the process stream via a polypropylene pipe to a 1 1 constant head device from which the sample stream was abstracted.Control software was developed for unattended operation and incorporated a triplicate standard analysis every 60 min and a triplicate sample analysis every 15 min. Each sulphite concentration was calculated by ratioing each mean sample response to the preceding mean standard response. A number of routines were included in the software in order to minimize the gathering of spurious data resulting from entrainment of air in the manifold. The time, sulphite concentration and standard deviation were down-loaded to the local printer and display unit after every 15 min analysis cycle. Results and Discussion Batch Experiments The effect of the reaction pH on the kinetics of the DNTB-sulphite reaction14 indicates the need for an effective buffering system for quantitative analysis.For the analysis of aqueous sulphite standards a pH 6.9 buffer is suitable6 but for the analysis of highly alkaline process streams a buffer of higher pH is required. Experiments conducted using pH 11.7 buffer, however, showed that precise pH control in alkaline media is much more important than in neutral solutions. As can be seen in Fig. 4, the reaction profile is significantly affected by increasing the pH of the reaction medium from pH 6.9 to pH 11.7. The ultraviolet-visible spectrum (200-800 nm) of DNTB in pH 11.7 buffer is very similar to that of the thiolate anion (the monitorand of the DNTB-sulphite reac- 20 PI DNTB (2.5 x 10-3 mol dm-3) n 1 Water 1.6 9.9 buffer 1.6 Water 1.6 0.5 ml min-' Brine I I / I I Fig.3 tion of sulphite in high ionic strength potassium chloride brine Modified flow injection manifold for the on-line determina- 2.5 r I 2.0 Q, 1.5 s. 2 1.0 4 0.5 D B A I I I I 1 0 10 20 30 40 50 60 Time/s Fig. 4 Response profiles for the reaction of DNTB in buffer and buffer containingsulphite: A, pH 6.9 buffer; B, pH 11.7 buffer; C, pH 6.9 buffer containing 10 mg 1-1 sulphite; and D, pH 11.7 buffer containing 10 mg 1 - 1 sulphiteANALYST, JULY 1991, VOL. 116 703 tion), suggesting cleavage of the DNTB sulphur-sulphur bond at high pH in the absence of sulphite. It is known that aromatic disulphides, particularly nitro-substituted aromatic disulphides, are susceptible to cleavage in alkaline conditions,15,16 yielding the corresponding thiolate anion and sulphinic acid.The molar absorption coefficient calculated in terms of the thiolate anion was found to be in agreement (E = 13500 1 mol-1 cm-1) with that previously reported,12.13 thus confirming cleavage of the DNTB sulphur-sulphur bond to yield the thiolate anion in a 1 : 1 ratio. The effect of pH on cleavage of the DNTB sulphur-sulphur bond in the absence of sulphite was further investigated over the pH range 10.9-12.8. Fig. 5 shows that at pH 10.9 there is no significant increase in absorbance with time but at a higher pH the rate of cleavage is significantly affected by very small changes in pH. For on-line analysis it is, therefore, necessary to buffer the sample stream to pH 10.9 or below in order to minimize this effect. Automated Laboratory Analysis Calibration Calibration data are presented in Table 1 for sulphite in water, 20% m/v potassium chloride standard solution and potassium 3.0 2.5 8 2.0 B C 1.5 a 1.0 0.5 n "0 20 40 60 80 100 120 Time/s Fig.5 Response profiles for the alkaline cleavage of the DNTB sulphur-sulpher bond at different pH values in the absence of sulphite: A, gH 10.9; B, pH 11.5; C, pH 11.8; D, pH 12.1; E, pH 12.5; and F, pH 1 .8 Table 1 Calibration data for sulphite in aqueous media; RSD (%) values quoted in parentheses (n = 10) Mean signal/A Concentration/ mg 1-1 Water KCI (20% d v ) KCI brine 0 0.050 5 0.228 10 0.409 15 0.555 20 0.716 ( 1 .o> (0.4) (0.6) (0.5) (0.3) 0.059 0.251 0.442 0.616 0.777 (3.3) (0.8) (0.9) (0.3) (0.4) 0.056 0.237 0.417 0.584 0.743 (4.2) (0.7) (0.6) (0.7) (0.4) Linear regression and correlation coefficients ( n = 5 ) : (1) Response (absorbance) = 0.033 x [S032-] (mg 1-1) + 0.060; r = (2) Response (absorbance) = 0.036 x [SO3*-] (mg I - l ) + 0.069; r = (3) Response (absorbance) = 0.034 x [S032-] (mg 1-1) + 0.063; r = 0.9992.0.9993. 0.9996. chloride brine (obtained from ICI Chemicals and Polymers). A sample injection frequency of 60 samples h-1 was used throughout and ten replicate analyses of each solution were made. The results show good correlation in the range 0.1-20 mg 1-1 of sulphite. pH stability Owing to the sensitivity of the reaction to pH the effect of sample pH on the response of the system was investigated for standards in water and in 20% m/v potassium chloride. The results given in Table 2 show that below pH 12.0 there is no significant variability in response and that above pH 12.0 only a slight increase in signal is observed.The manifold is therefore suitable for process applications because the pH of the sample stream rarely exceeds 12.0. Temperature stability The temperature of the KCl brine on the plant is maintained in the range 7&8OoC, and therefore the temperature of the abstracted sample varies considerably owing to external environmental influences. This could have a significant effect on the rate of reaction but the results given in Table 3 show that such an effect is eliminated by sample dilution in the FI manifold. Effect of potassium chloride concentration Comparison of the calibration data for water, 20% m/v potassium chloride standard solution and potassium chloride brine (Table l ) , shows that the concentration of potassium chloride has an effect on sensitivity.The results presented in Fig. 6 show the increased response for a 10 mg 1-1 standard sulphite solution with increasing potassium chloride con- centration, which is due to an increased rate of reaction at higher ionic strength. However the process stream contains 18% m/m potassium chloride, which corresponds to the region that exhibits the least variation in response, and small changes in process ionic strength will therefore have minimal effect. Table 2 Effect of sample pH on system response for 10 mg I- standard sulphite solutions prepared in water and 20% m/v KCI (n = 5) Matrix Sample pH Mean signal/A RSD (YO) Water 6.6 0.297 0.6 9.5 0.289 0.4 10.4 0.293 1 .o 11.2 0.298 0.3 11.9 0.299 0.7 12.2 0.311 0.4 KCI (20% d v ) 5.6 0.423 0.5 10.4 0.422 0.4 10.8 0.423 0.5 11.4 0.427 0.9 12.1 0.435 0.9 12.5 0.436 0.5 Table 3 Effect of sample temperature on system response for a 10 mg 1-1 standard sulphite solution in 20% m/v KCl at pH 11.0 (n = 5).Mean Sample temperaturePC signaVA RSD (Yo) 25 35 45 55 70 90 95 0.463 0.464 0.463 0.466 0.470 0.462 0.468 0.4 0.3 0.6 0.1 0.2 0.5 0.7704 ANALYST, JULY 1991, VOL. 116 On-line Analysis Calibration A 16 h continuous trial of the manifold and software using a 10 mg 1-1 sulphite standard solution produced a mean response of 0.452 A with a relative standard deviation (RSD) of 0.46% (n = 16). Analysis of a simulated sample solution gave a mean concentration of 15.85 mg 1-1 with an RSD of 0.51% (n = 64) and subsequent analysis by iodimetric titration gave the same result.Preliminary on-line trials showed the levels of sulphite to be outside the linear range of the proposed method (0.1-20 mg 1-1). In order to extend the linear range, the absorbance of the thiolate anion was measured at increasing wavelengths from 412 nm (the wavelength of maximum absorbance). Measurement at 500 nm extended the linear range (0.540 mg 1-1) with a reduced sensitivity (0.139 A for 40 mg 1-1 of sulphite). The linear range could not be further extended in this way owing to the poor signal-to-noise ratio at higher wavelengths. Furthermore, it was not possible to increase the concentration of DNTB because of its limited solubility. The sample stream was therefore further diluted on-line (Fig.3) and calibration data covering the range 3-200 mg I-* are presented in Table 4. Sulphite can be determined over the range 3-200 mg I - I ; the response is linear for the concentration range 3-100 mg 1-1 (n = 5, r = 0.9999). Validation of on-line method Results from the on-line method were validated against the standard iodimetric procedure over an 8 h period (Fig. 7). A 0.4 a C +? a . 3 0.3 0.2 ' 1 I I I 0 5 10 15 20 [KCII (% m/v) Fig. 6 Effect of potassium chloride concentration in the sample matrix on the response Table 4 Calibration data for the on-line determination of sulphite in 20% d v KCI at pH 11.0 (n = 10) Concentration/ Mean mg 1-1 signaYA RSD (%) 0 25 50 75 100 150 200 0.015 0.038 0.063 0.088 0.112 0.150 0.177 2.9 2.6 0.0 0.9 0.8 0.6 0.7 Linear regression and correlation coefficient (n = 5): Range 3-100 mg 1-l SO3*-: Response (absorbance) = (9.7 X lop4) X [S03*-] (mg 1-1) + 0.014; Range 3-200 mg 1-1 S032-: r = 0.9941. r = 0.9999.portion of the brine was taken from the constant head device every 15 min to coincide with the analysis cycle of the monitor. Three hours into the trial the addition of sulphite to the process stream was increased in a step-wise manner over a period of 2.5 h. After the trial had been in progress for 5.5 h, the rate of sulphite addition was returned to its original level and then rapidly increased and reduced again over a 1 h period. It can be seen from Fig. 7 that the corresponding changes in sulphite concentration are closely followed by the monitor and by iodimetric analysis.Fig. 8 shows the regression of the monitor response on the results of the standard iodimetric procedure (Y = 0.9964) and reveals no observable systematic error. Extended on-line trial Fresh reagents were prepared weekly and details of reagent consumption are given in Table 5. The response to a standard (62.3 mg 1-1) over a 1 week operating period (168 triplicate determinations) was 0.084 A with an RSD of 2.1%. Peristaltic pump tubing was replaced after 14 d, and in 21 d of continuous on-line use only one failure was reported (owing to the 130 r I - \ F 2 0 .o 110 E c 4- 6 100 0 90 ," 80 4- .- r Fig. 7 Method comparison study: solid line, monitor response; and 0, off-line iodimetric results 130 I 70 80 90 100 110 120 130 lodimetric analysis/mg I-' Fig.8 Regression of on-line flow injection results on the off-line iodimetric results; r = 0.9964 Table 5 Reagent consumption over a 7 d period (672 analyses) Reagent Consumption/l De-ionized water pH 9.9 buffer DNTB reagent Sulphite standard 10.8 5.4 1.4 0.9ANALYST, JULY 1991, VOL. 116 L I I I I I 0 100 200 300 400 500 Tirne/h Fig. 9 Results of a 21 d on-line trial ~ ~~ ~~~ ~ ~~ Table 6 Performance characteristics and specifications Parameter Over-all accuracy Precision Response time Dynamic range Maintenance Running costs System costs Plant specification Proposed monitor (10% <3% <5% <1% ( n = 64) 15 min <5 min 0.1-100 mg 1- 30 min per week - <fl .OO per day - f 6 500 1-20 mg 1-1 <1 h per week blockage of the injection valve).A plot of monitor output over the 21 d on-line trial is presented in Fig. 9. The sharp increase in concentration at approximately 400 h corresponds to a temporary plant shut-down, during which time sulphite addition was continued. This trial demonstrates the reliability of the instrumentation in a process environment and the good correlation of the on-line and manual results. A feedback loop for process control (as shown in Fig. 1) could therefore be confidently and easily incorporated. The reaction chemistry is sufficiently stable in terms of pH, temperature and potassium chloride concentration for on-line analysis. The performance characteristics of the proposed 705 monitor also meet the original plant specifications as shown in Table 6.The monitor was reliable over a 21 d on-line trial and the results are in good agreement with those obtained from a standard off-line iodimetric procedure. The authors acknowledge the assistance of A. A. Davies and A. Afnan during the on-line trial, and I.C.I. Chemicals and Polymers Ltd. for financial support. P. M. thanks the SERC for a research studentship. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 References Recktenwald, A., Kroner, K. H., and Kula, M. L.. Enzyme Microb. Technol., 1985,7,607. Gisin, M., andThommen, C., Anal. Chim. Acta, 1986,190,165. Clinch, J. R., Worsfold, P. J . , and Casey, H., Anal. Chirn. Acta, 1987,200, 523. Benson, R. L., Worsfold, P. J., and Sweeting, F. W., Anal. Proc., 1989, 26, 385. Gisin, M., and Thommen, C., TrAC, Trends Anal. Chem. (Pers. Ed.), 1989, 8, 62. MacLaurin, P., Parker, K. S., Townshend, A., Worsfold, P. J., Barnett, N. W., and Crane, M., Anal. Chim. Acta, 1990, 238, 171. Benson, R. L., Worsfold, P. .I., and Sweeting, F. W., Anal. Chim. Acta, 1990, 238, 177. Ranger, C. B . , in Automated Stream Analysis for Process Control, ed. Manka, D. P., Academic Press, New York, 1982. Van der Linden, W. E., Anal. Chim. Acta, 1986, 179,91. RhiiEka, J . , Anal. Chim. Acta, 1986, 190, 155. Lazaro, F., Luque de Castro. M. D., and Valcarcel, M., J. Pharm. Biomed. Anal., 1988, 190, 585. Humphrey, R. E., Ward, M. H., and Hinze, W., Anal. Chem., 1970, 42, 698. Vogel, A. I . . Quantitative Inorganic Analysis, Longman, London, 1961. Brown, D. S . , and Jenke, D. R., Analyst, 1987, 112, 899. Jocelyn, P. C., Biochemistry of the SH Group, Academic Press, London, 1972. Parker, A. J . , and Kharasch, N., Chem. Rev., 1959, 59, 583. Paper I I001 941 Received January 15th, 1991 Accepted March 15th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600701

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991. VOL. 116 707 Shipboard Flow Injection Method for the Determination of Manganese in Sea-water Using In-valve Preconcentration and Catalytic Spectrophotometric Detection 1. Ya. Kolotyrkina, L. K. Shpigun and Yury A. Zolotov Kurnakov Institute of General and Inorganic Chemistry, Akadem y of Sciences, Leninsky Prospect 37, Moscow I 7 7907, USSR G. 1. Tsysin Department of Chemistry, Moscow University, Moscow 179899, USSR A flow injection method for the determination of manganese at trace level concentrations in sea-water is proposed. The in-valve ion-exchange microcolumn for preconcentration was directly coupled to the spectrophotometer. The spectrophotometric detection was based on the catalytic effect of Mn” on the oxidation of N,N-diethylaniline by potassium periodate in a neutral medium.Variation of the preconcentration time from 10 s to 10 min allowed the determination of manganese in the concentration range 20 pg 1-1-10 ng I-’. A sampling frequency of up to 15 h-1 and a relative standard deviation of 5-8% were achieved. The proposed method was successfully applied to the direct shipboard measurement of the manganese content in deep sea-water samples for the purpose of indicating the presence of active hydrothermal vents on the sea floor. Keywords: Manganese determination; sea-water; flow injection; catalytic reaction; in-valve precon- cen tra ti0 n One of the most interesting problems in analytical chemistry, applied to the investigation of the oceans, is the chemical indication of the presence of hydrothermal springs on the sea floor.’.’ In addition to the fact that metal-rich hot springs are associated with mineral deposits in ocean basins and finding their locations could eventually lead to mining, studies of hydrothermal activity can help to explain the origin of many major mineral deposits on land.Buoyant emanations of hot, hydrothermal vent fluids form ‘plumes’ which are charac- terized by a chemical composition different from the sur- rounding sea-water. Large manganese and 3He anomalies in sea-water samples provide evidence of intense present-day activity due to hydrothermal springs on the sea floor. Manganese has been established as a very sensitive tracer of hydrothermal and, therefore, it is necessary to have a simple and fast analytical method for the direct determination of manganese aboard a research vessel.The routine batch methods usually used for this purpose tend to be laborious and time consuming, requiring large sample vol- umes and complicated procedures. In most of these pro- cedures, the required sensitivity is achieved in combination with preconcentration and separation techniques such as solvent extraction.s.6 Some procedures are based on the catalytic effect of Mn” on the oxidation of an organic compound with the formation of a coloured product.7 Nevertheless, catalytic methods are complicated and trouble- some because of their time-dependent characteristics. The combination of a catalysed chemical reaction with a flow injection (FI) system has been reported to be ideal in overcoming many previous difficulties and greatly simplifying the analytical procedure .X-1° An FI method for the determination of manganese in sea-water which is based on the sensitive catalytic oxidation of N, N-diethylaniline (DEA) by potassium periodate in neutral media has been proposed.’] Although this method is relatively sensitive and interference free, it requires a separation- preconcentration step in order to remove the sea-water matrix effect and to increase the precision of the determination of manganese at the ultra-trace level.This paper describes a new FI procedure, based on the catalytic detection principle, associated with an in-valve preconcentration of manganese using a microcolumn filled with fibrous diethylenetriamine- tetraacetic acid (DETATA) as sorbent. Experimental Reagents All the solutions were prepared from analytical-reagent grade chemicals with re-distilled water.A stock manganese solution (0.1 pg ml-1 of Mn”) was prepared by suitable dilution of an aqueous standard solution (1000 yg ml-I of Mn”) (BDH). Working solutions of Mn” were prepared daily by diluting the stock solution with sea-water depleted of manganese, collec- ted from a depth of 1000-1500 m and pumped through a column containing Chelex-100. N ,N-Diethylaniline. Purified by distillation. 17- Potussium periodate. Recrystallized twice. The phosphate buffer solutions (pH 7.3-7.4) of 0.2 rnol dm-3 Na2HPO4-10H2O-0.05 rnol dm-3 KH2P04 or 0.25 rnol dm-3 KH2P044.2 rnol dm-3 NaOH were prepared from salts recrystallized twice. The working solution RI consisted of the buffer saturated with the potassium periodate solution (1 X 10-2 rnol dm-3).Solution R2 was prepared by dissolving 45-50 yl of DEA solution in 4 ml of 0.1 rnol dm-3 HCI; re-distilled water was added to give a total volume of 200 ml. Hydrochloric acid (0.5 rnol dm-3) or HN03 (0.5 rnol drn-”, used as the eluent solution, was prepared from quartz distilled, 2 rnol dm-3 acid. Microcolumn Preparation Newly synthesized cellulose fibrous DETATA sorbentl3 was used. The concentration of the DETATA functional groups was 0.22 mmol per gram of cellulose fibre. The microcolumn for the preconcentration and separation of manganese from the sea-water matrix was a glass tube (10 x 2.5 mm i.d.) packed with fibrous DETATA sorbent in the H+ form. Glass wool plugs were placed at each end of the sorbent zone (3-5 mm) in order to retain the fibre and were in turn held in place by short pieces of polytetrafluoroethylene (PTFE) manifold-tubing (0.5 mm i.d.).All new columns were conditioned in the FI manifold by pumping strong acid solution (2 rnol dm-3 HCl or HN03) through the column for about 10 min at a flow rate of 0.4 ml min-1, thereby eluting all metal ions from the column. Re-distilled water at a flow rate of 2 ml min-1 flushed the remaining acid solution out of the sorbent.708 ANALYST, JULY 1991, VOL. 116 w U w Fig. 1 Schematic diagram of the FI manifold for the determination of manganese in sea-water: R1 and R2, reagents; E, eluent; S, sample; v l , v2, v3 and v4, flow rates; P1 and P2, pumps; I, injection valve; ll and 12, reaction coils; D, detector; and W, waste. The solid line indicates the load position and the broken line the elution position of the injection valve (see text for details) Apparatus and FI Manifold Configuration A FIAStar 5020 flow injection analyser (Tecator), consisting of a Model 5020 main unit, a Model 5024 filter photometer (D, Fig.l ) , equipped with a flow-through cell (light path, 10 mm; volume, 18 pl), with a Model 5032 controller were used. All tubing was PTFE (0.5 or 0.7 mm i.d.). The Chemifold I1 was used in order to provide the chosen FI manifold configuration shown in Fig. 1. The ion-exchange microcolumn was incor- porated instead of the injection loop in the six-channel rotary valve (I) of the FI device. The eluent line (E) merged directly with the reagents (R1 and R2) at T connections.The increase of absorbance at 470 nm was registered by the detector (D). Collection and Filtration of Sea-water Samples Sea-water samples were collected by a Rosette water sampler. Samples were removed from Niskin bottles into acid-cleaned polyethylene bottles, filtered through Nuclepore membrane filters (pore size, 0.4 pm) and analysed immediately. Results and Discussion Ion-exchange Preconcentration and Elution Conditions Ion-exchange columns of Chelex 100 are usually used in FI systems for sorption of heavy metals.14 However, there are several drawbacks with Chelex lOO,15 and it is not sufficiently effective for the preconcentration of manganese from sea- water.16 For that reason the newly synthesized aminocarboxyl DETATA sorbent was used.13 The most suitable dynamic conditions for the separation of manganese from sea-water by DETATA sorbent were chosen. It was shown that complete separation can be achieved in the pH range 3-8 and at a sorbent flow rate (vl) of 5 ml min-1.The enhancement coefficient for manganese was completely independent of the salinity of the solution. The DETATA sorbent was found to be very suitable as a concentration-separation tool for sea-water at its natural pH (7.5-8.5); the sample does not need to be mixed with a buffer before pumping through the column. It was shown that column size can be minimized to 2 x 3 mm at a sample flow rate of 2-3 ml min-'.'Increasing the sample flow rate could cause compaction of the sorbent fibres in the column making it impossible to use for an extended period of time.The manganese elution conditions were also studied. It was possible to elute manganese by using weak acid solutions (0.1-0.5 mol dm-3 HCI or HN03), which can be explained by the ion-exchange mechanism of the reaction between man- ganese and the functional groups of the sorbent. Therefore, the microcolumn could be used in a cycle of two steps, preconcentration-elution, more than 150 times. The elution of the adsorbed manganese by 0.5 rnol dm-3 HCI was complete when the acid was pumped through the column for 90 s at a flow rate of 0.4 ml min-l. As an increase in the elution time broadens the peak, 0.5 rnol dm-3 HCl or HN03 was chosen as the most suitable .eluent. Flow Injection Variables in a Catalytic Detection System In the system described under Experimental, spectrophoto- metric detection, based on the catalytic effect of manganese on the oxidation of DEA by potassium periodate, is followed by ion-exchange preconcentration.The optimum conditions for the catalytic spectrophotometric determination were chosen using an injection coil of 30 p1 instead of an ion-exchange column. The standard solutions, containing 10-100 pg 1-1 of manganese in 0.5 mol dm-3 HCI, were injected into the system. In a previous paper," the effects of the concentrations of DEA and periodate, and the pH, were studied in detail. The same concentrations of the reagents were therefore used here. The buffer concentration of the reagent mixture, R1, and flow rates of R1 and R2 were chosen in order to obtain a value of pH = 7 in the reaction coil, 12.[The concentration of the buffer solution (0.2 mol dm-3) was limited by the solubility of some of the salts ( e . g . , Na2H- PO4. 10H20) .] The low flow rates of the reagents (v3 = 1.2 and v4 = 0.8 ml min-1) were chosen to give a maximum catalysed reaction time before the spectrophotometric detection. The length of the reaction coils 11 (0.9 m, 0.7 mm i.d.) and 12 (1.2 m, 0.7 mm i.d.) were selected so that a large sharp manganese peak was obtained. The peak height was found to be virtually independent of the temperature in the range 20-30 "C and, therefore, all analyses were carried out at room temperature. Procedure for Analysis of Sea-water Samples The same system was used for the analysis of sea-water samples.The sea-water sample (S) was pumped through the microcolumn containing DETATA sorbent at v1 = 2.8 ml min-1 for a time (tc) of 10 s-10 min (tc depended on the concentration of manganese in the analyte). During the preconcentration step only the pump, P1, was on and the injection valve (I) was in the load position. At the same time the acid eluent solution was driven to the reaction manifold by the same pump (by-passing the injection loop) in order to wash the coloured oxidation product from the tubing of the system. During the next step the re-distilled water was pumped through the column for 1 min in order to flush the remaining sea-water out of the sorbent. At this step the pump, P2, was switched on. After stabilizing the baseline, the rotor of the injection valve was automatically switched to the elution position, and the eluent was directed through the column, eluting the adsorbed manganese and carrying it into the reaction manifold.The flow rate v2 was 0.4 ml min-1, corresponding to the elution conditions discussed above. This eluent stream merged consecutively with the streams of R1 and R2. The mixture was driven into the reaction coil l2 and through a flow cell. The increase in absorbance at 470 nm, due to the colour of the spectrophotometric catalysed reaction which took place in the reaction coil after mixing all the reagents, was registered and recorded. The injection loop was turned to the load position again, P2 was switched off and the re-distilled water was pumped through the column for 1 min at a flow rate of 2.8 ml min-1 in order to flush out the remaining acid solution.The first step of preconcentration was repeated with the other samples. The elution and detection time was 90 s but the preconcen- tration time was variable.ANALYST, JULY 1991, VOL. 116 709 The calibration graph was obtained by measuring the peak height (in terms of absorbance) for standard solutions prepared with sea-water depleted of manganese. The intercept of the standard additions curve was used to determine the manganese concentration in the depleted sea-water, assuming that sea-water free from manganese would not give a signal. The calibration graphs for different preconcentration times are shown in Fig. 2. Variation of the preconcentration time between 10 s and 10 min allows the determination of a wide concentration range (10 ng 1-1-15 pg 1-1) of manganese giving 10- to 800-fold enhancement factors (Table 1).A linear dependence of the manganese enhancement factor on preconcentration time was observed. The detection limit of this method, determined as the average signal from three blank solutions, is about 2 ng 1-1 and depends on the purity of the re-distilled water. It could be lowered by using a preconcentration time of more than 10 min. A sampling rate of up to 15 samples h-1 can be achieved. 0 0.4 0.8 1.2 1.6 2.0 [Mn]/pg 1-1 Fig. 2 s; 2, 1 min; 3, 2 min; and 4, 5 min Calibration graphs for different preconcentration times: 1.30 Table 1 Preconcentration conditions for various concentrations of manganese in sea-water with corresponding sampling rates Pre- Mean concentration Range of enhancement Sampling time/s cM,lng I- factor ratelh- 600 10-100 800 4 300 50-200 400 6 120-180 100-500 150-250 8-10 60 300-1 500 80 12 30 500-2000 40 15 10 2 X 103-1.5 X 10' 10 15 The accuracy of the proposed method was tested by performing recovery experiments on manganese standards prepared with sea-water depleted of manganese.Results for the determination of manganese in a series of these synthetic mixtures are given in Table 2. It can be seen that the proposed method gives satisfactory repeatability. Shipboard Analytical Results The proposed method was tested at sea during the 21st cruise of RV Akademik M . Keldysh during April-June 1990 to the western part of the Pacific Ocean. The purpose of the expedition was to investigate hydrothermal processes on the sea floor.About 400 sea-water samples were collected from the surface and from 32 depth stations (the vertical profiles ranging from several metres to several hundred metres above the sea floor) and analysed for dissolved manganese. Table 3 shows the results of the routine analyses of several different sea-water samples. The relative standard deviation is less than 8% for the concentration range 40-600 ng 1-1 of manganese. The accuracy of the method was verified during the cruise by comparing our results for sea-water with the results obtained by a modified Klinkhammer method.6 The compari- son between the manganese results obtained by these two Table 2 Recovery of manganese added to sea-water depleted of manganese: P = 0.95 Manganese@ 1-1 Recovery Added Found RSD* n (Yo 1 50 59 0.08 4 118 150 146 0.05 4 97 300 327 0.06 5 109 500 5 10 0.06 6 102 1000 1038 0.05 5 103 1500 1493 0.05 4 100 1 x 10' 1.02 x 1o-r 0.02 12 102 * RSD = relative standard deviation.Table 3 Determination of dissolved manganese in sea-water samples: P = 0.95; and n = 6 Manganese foundhg 1 - 1 RSD* 44 t 4 0.08 102 * 9 0.08 357 t 24 0.06 632 k 34 0.05 * RSD = relative standard deviation. Cali bration Samples from the plume (station 2254) Standards h f A \h Mn II concentration/ 540 1040 I 1 nn 1-1 I Depthh 2141 2100 2050 795 2340 L - 2000 838 529 154 h 692 498 328 144 1975 1955 1900 ~ ~~ Fig. 3 Output signals for a series of manganese standards and for sea-water samples collected from the hydrothermal plumeANALYST, JULY 1991, VOL.116 710 0 500 1000 1500 I E 2000 . 0 400 800 1200 1600 P 5 500 Iooo 1500 1 t 3000 2500 L 1000 50T 2500 0 200 400 600 800 1000 1000 50:: 3000 I I I I 0 1000 2000 3000 0 500 1000 1500 2000 Concentration of Mnlng I-’ Fig. 4 Vertical distribution of dissolved manganese at the stations of the Manus basin. ( a ) Station 2243; (6) station 2247; (c) station 2254; and (d) station 2260 distinctly different techniques is fairly good. A least-squares fit of the data gives the equation [Mn] (Klinkhammer) = 5.6 + 0.93 X [Mn] (FI) with P = 0.95; n = 29; correlation coefficient r = 0.974; and concentration in ng 1-1. The concentration of manganese determined ranged from 20-100 ng 1-1 for sea-water to anomalously high concentra- tions of up to 100 times the sea-water value in hydrothermal plumes.The output signals for a series of samples from the hydrothermal plume at one of the stations is shown in Fig. 3. Some vertical profiles of manganese concentration corres- ponding to the position of known plumes are shown in Fig. 4. In the Manus basin the existence of active hydrothermal springs was confirmed by dives of the submersible Mir. The manganese concentration in a sample collected by Mir from the area near the hydrothermal fluid outlet was 16 pg 1-1. Conclusion The results clearly demonstrate that the FI technique based on in-valve ion-exchange separation with spectrophotometric detection is very promising for the trace analysis of sea-water. The proposed FI method for shipboard determination of manganese in sea-water has the advantages of good accuracy, a fast sampling rate and automation.The essentially closed environment of an FI system enables the determination of manganese at nanogram levels to be carried out in a standard ship’s laboratory without significant contamination. The only necessary requirement for obtaining precise repeatable results is a sufficient purity of all the reagents used and re-distilled water. We thank A. P. Lisitsin for inviting us to participate in the cruise, V. V. Gordeev and his group for providing the manganese data used in the inter-calibration, and L. V. Dmitriev for his interest in our work and helpful discussions on the geochemistry of hydrothermal vents. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Edmond, J. M., and Von Damn, K. L., Sci. Am., 1983,248,78. Rona, P., Sci. Am., 1986, 254.84. Klinkhammer, G.. Rona, P., Greaves, M., and Elderfield, H., Nature (London), 1985, 314, 727. Klinkhammer, G., Bender, M., and Weiss, R. F., Nature (London), 1977,269,319. Klinkhammer, G. P., Anal. Chem., 1980, 52, 117. Atnashev, V. B., Muzgin, V. N., and Atnashev, Yu. B., Zh. Anal. Khim., 1982, 37. 1590. Hadjiioannou, T. P., and Kephales, T. A., Mikrochim. Acta, 1969, 1215. Yamane, T., Anal. Chim. Acta, 1981, 130, 65. Yamane, T., Anal. Sci., 1986, 2, 191. Maspoch. S., and Blanco, M., Analyst, 1986, 111, 69. Kolotyrkina. I. Ya., Shpigun, L. K., and Zolotov, Yu. A., Zh. Anal. Khim., 1988, 43. 284. Organicum. Obschiy Praktikum po Org. Khim.. ed. Kost, A. N., Mir, Moscow, 1965, p. 315. Tsysin, G. I., Formanovsky, A. A., Mikhura, I. V., Evtikova, G. A., Sokolov, D. P., and Marov, I. N., Zh. Neorg. Khim., 1990, 34, 960. Kingston, H. M., Barnes, I. L.. Brady. T. J., Rains, T. C., and Champ, M. A., Anal. Chem., 1978. 50, 2064. Sakamoto-Arnold, C. M., and Johnson, K. S.. Anal. Chem., 1987, 59, 1789. Nevoral, V.. Chem. Lisfy, 1987, 81, 323. Paper 0105465H Received December 4th, 1990 Accepted March 11 th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600707

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991, VOL. 116 Flow Injection Spectrophotometric Determination D-Sorbitol Using Methyl Orange as an Indicator Kazuko Nose Environmental Resources Research Centre, Oka yama University of Science, 700, Japan Michio Zenki* 71 1 of Boron With I - 1, Ridai-cho, Oka yama Department of Chemistry, Faculty of Science, Oka yama University of Science, I - I , Ridai-cho, Oka yama 700, Japan A spectrophotometric method for the determination of boron by flow injection, based on the complexation reaction between D-sorbitol and boric acid followed by the acid-base reaction of Methyl Orange, is described. Under the optimum conditions, the calibration graph was essentially linear up to 1.2 mg 1-1 of boron (120 pl injections) and the detection limit was 0.02 mg 1-1 (signal to noise ratio = 3).Twenty samples can be analysed per hour with a relative standard deviation of 0.28454%. The method was applied to the determination of boron in commercial eye lotions. By using a 0.6% m/v sodium chloride solution as the carrier stream, boron in sea-water samples was also determined. Keywords: Boron determination; flow injection; Methyl Orange; 0-sorbitol; spectrophotometry Boron is one of the elements that is difficult to determine quantitatively. Various analytical methods have been pro- posed for the determination of boron. Spectrophotometry is widely used in the field of trace analysis, because other methods such as atomic absorption spectrometry122 or flame emission spectrometry334 are less sensitive. Several specific and sensitive reagents, for example, curcumin,5.6 carmine,536 Methylene Blue6-7 and azomethine H,s have been reported for the spectrophotometric determination of boron.All of these methods, however, have some disadvantages, requiring distil- lation of methyl borate, evaporation to dryness, extraction with dichloroethane, the use of concentrated sulphuric acid and/or long standing times for colour development. Hence, automated procedures for the determination of boron could not be developed. Boric acid is too weak an acid (pK, = 9.2) to be titrated directly with a standard alkali solution. It is known that some polyhydroxy compounds react with boric acid to form complexes which are sufficiently acidic to be titrated directly. For milligram amounts of boric acid, titration with NaOH- mannitol is one of the most important methods.6 If a polyhydroxy compound-boron complex is dissociated to a large extent in a stream containing a pH indicator, a colour change reaction occurs which is proportional to the concentra- tion of the complex; hence boric acid should be detectable with high sensitivity by using spectrophotometry. In the present paper an automated procedure for the determination of boron by means of a flow injection (FI) method is described.The method is based on the spectrophotometric detection of the increased acidity of a sorbitol solution by the addition of boric acid. Methyl Orange (MO) was used as a pH indicator. The method was applied to the determination of boron in eye lotions and sea-water samples. Experimental Reagents All chemicals were of analytical-reagent grade and were used without further purification.Distilled, de-ionized water was used throughout to prepare the solutions. * To whom correspondence should be addressed. A stock solution of boron (lo00 mg 1-1) was prepared by dissolving 5.715 g of boric acid (Wako Pure Chemical Industries, Osaka, Japan) in 1 1 of water. Working standard solutions were prepared from the stock solution by appro- priate dilution with water. A stock solution of the pH indicator was prepared by dissolving MO (Colour Index 13 025, Nakarai Chemical, Kyoto, Japan) in water to give a concentration of 2.0 X 10-3 mol dm-3. The reagent solution for FI was prepared by dissolving 57 g of D-sorbitol (Nakarai Chemical) in about 200 ml of water, adding 25 ml of MO solution (2.0 X 10-3 mol dm-3) and diluting to 250 ml with water.Apparatus A two-channel manifold [one channel is the carrier stream (distilled water or 0.6% m/v NaCl solution) and the other is the reagent stream containing the polyhydroxy compound (sorbitol) and MO] was designed in order to obtain reasonable sensitivity and reproducibility. A schematic diagram of the FI system is shown in Fig. 1. A reciprocating pump of the double plunger type (DM2U, Sanuki Kogyo, Tokyo, Japan) was used to propel both solutions at an equal volume ratio (1.0 ml min-1 each). Sample injections into the carrier stream were made with a six-way rotary valve (SVM, Sanuki Kogyo) with a sample loop having a volume of 120 pl. Spectrophotometric measurements were carried out by using a visible spectropho- tometer (S-3250, Soma Kogaku, Tokyo, Japan) equipped with an 8 PI flow cell with a pathlength of 10 mm.Absorbances were measured at 520 nm and the output was fed to a strip-chart recorder (U-228, Nippon Denshi Kagaku, Tokyo, Japan). All connections were made with 0.5 mm i.d. poly- tetrafluoroethylene (PTFE) tubing, except for the back-pres- sure coil which had an i.d. of 0.25 mm (length 1 m). Procedure Commercially available eye lotions were diluted 2500-fold with water. Sea-water samples were filtered through a 0.45 pm membrane filter and diluted 5-fold with water. Boron samples and standards were injected into the water stream and merged with the sorbitol-MO solution pumped at the same flow rate. For the analysis of sea-water, a 0.6% m/v NaCl solution was used as the carrier stream instead of distilled water.The boric acid-sorbitol complex was formed in the coiled reaction tube712 ANALYST, JULY 1991, VOL. 116 CS.+* RC BC I I W 11.0 I RS I 1 U mi min-1 Fig. 1 Schematic dia ram of the FI system. CS, Carrier stream (distilled water or 0.62 m/v sodium chloride solution); RS, reagent stream (2.0 X mol dm-3 MO in 1.25 mol dm-3 D-sorbitol solution); P, pump; S, sample injection (120 PI); M, mixing joint; RC, reaction coil (4 m X 0.5 mm i.d.); D, detector (520 nm); R, recorder; BC, back-pressure coil (1 m x 0.25 mm i.d.); and W, waste and the colour change of the MO indicator occurred at the same time. The absorbance of the resulting solution was measured at 520 nm and recorded on the strip-chart recorder.Results and Discussion Selection of Polyhydroxy Compounds The reactivity of polyhydroxy compounds towards boric acid depends on the configuration of the hydroxyl groups in the molecules.9 The concentration of the polyhydroxy compound in the reagent stream was set at 0.8 rnol dm-3 (mannitol had the lowest solubility in water and a more concentrated solution could not be prepared), and the complexation with boric acid was investigated (MO concentration: 2.0 X rnol dm-3). The results obtained for five compounds, i . e . , D-sorbitol, D-mannitol, D-fructose, xylitol and D-ribose, are shown in Fig. 2. Other compounds, e . g . , galactose, maltose, D-arabi- nose, L-rhamnose, lactose, D-xylose, erythritol and glycerol, were not effective. Of the compounds tested, sorbitol was found to be the most appropriate with respect to detection sensitivity.Although mannitol is recommended as a complex- ing agent in the titration methods,6 other workers10711 have shown that sorbitol is slightly more effective. In addition, the solubility of sorbitol in water is superior to that of mannitol and the concentrated sorbitol solution could be easily pre- pared. Therefore, sorbitol was used as the complexing agent for boric acid in this work. The sorbitol concentration in the reagent stream was varied from 0.5 to 1.75 rnol dm-3 and the peak height (sensitivity) was measured for various boron concentrations. Fig. 3 shows that the peak height increases with an increase in the concentration of sorbitol. Although the sensitivity increases significantly at low sorbitol concentra- tions, the increments are small at a concentration above 1 rnol dm-3.By assuming the upper detection limit of the boron concentration to be 1.2 mg ml-1, and considering the cost factor, the concentration of sorbitol was fixed at 1.25 rnol dm-3. Selection of an Indicator The choice of a pH indicator is also important. Methyl Yellow, MO, Methyl Red, Bromophenol Blue, Bromocresol Green, Bromocresol Purple, Bromothymol Blue, Neutral Red and Phenol Red were tested. An ordinary sorbitol solution exhibits a pH of about 4.8-5.0; hence MO (pK, = 3.75) was found to be the most favourable indicator because the colour reaction occurred without the need for adjustment of the pH and the sensitivity was excellent. Methyl Orange has an absorption maximum at 520 nm (acidic form) and at 465 nm (basic form).The colour change reaction is monitored at 520 nm, the wavelength corresponding to the maximum absorp- tion of the protonated form of MO, so that the slope of the calibration graph is positive. Increasing the concentration of MO provides larger peaks and increases the amount of boric 0 1 2 Boron concentration/mg 1-1 Fig. 2 Comparison of different polyhydroxy compounds for the determination of boron using the FI system. Reaction coil, 2 m; other conditions as in Fig. 1. 1, D-Ribose; 2, xylitol; 3, D-fructose; 4, D-mannitol; and 5 , D-sorbitol 1.0 1.5 0.5 D-Sorbitol concentration/mol dm-3 Fig. 3 Effect of the D-sorbitol concentration in the reagent stream. Boron concentration: 1, 0.4; 2, 0.8; 3, 1.2; 4, 1.6; and 5 , 2.0 mg 1-1 acid that can be determined; however, this will also cause large reagent blanks.Therefore, a 2.0 x 10-4 rnol dm-3 solution of MO was used for the determination of 0-1.2 mg 1-1 of boron. Optimum Conditions for FI On the basis of these observations, appropriate conditions for FI were examined. The flow rates of the carrier and reagent streams were set at 1.0 ml min-1 each. In this system, the response increased linearly with the sample volume injected up to 120 PI; however, the injection of larger sample volumes produced only a marginal effect on the response. Therefore, an injected volume of 120 PI was found to be the optimum. The effect of the reaction coil length on the peak height was examined for various coil lengths of between 0.5 and 10 m.The peak height decreased with increasing coil length owing to dispersion; however, the reproducibility was poor with a coil length of less than 3 m, hence a 4 m coil was used. Interferences As the proposed method is based o n the acid-base reaction of MO, common cations and anions were found not to affect the determination of boron. Only fluoride and molybdenum interfered seriously when present at the same concentration as boron. This is due to the formation of a fluoroborate between fluoride and boric acid, whereas molybdic acid forms a complex with sorbitol similar to those of germanic and phosphoric acid.9.10ANALYST, JULY 1991. VOL. 116 I 713 ' Ill 1 O z i n [o.o 2 1 -c.. Time - Fig. 4 Typical FI signals for the determination of boron. Concentra- tion of boron: 1,0; 2,0.2; 3,0.4; 4,0.6; 5.0.8; 6,l.O; and 7,1.2 mg I-' Table 1 Determination of boron in commercial eye lotions Boron contentlg I-' Spectro- Proposed photometric HPLC Sample method* methodt method$ A9 2.12f0.03 2.03 2.08 B§ 1.05 f 0.01 0.95 1.03 Cy 2.05 f0.03 2.03 2.05 * Average of three determinations f standard deviation.t Reference 8. $ Reference 11. 9 For relieving eye strain. 7 For cleaning contact lenses. ~~ Table 2 Recovery test for the determination of boron Boron in Boron Boron sample/ added/ found/ Recovery Sample* mgl-1 mgl-1 mgl-l (Yo) Eye lotion- A 0.85 0.20 1.03 98 0.40 1.26 101 B 0.42 0.20 0.62 100 0.40 0.82 100 C 0.82 0.20 1.04 102 0.40 1.22 1 00 Hagi 0.72 0.20 0.88 96 0.40 1.12 100 Hinase 0.56 0.20 0.79 1 04 0.40 0.94 98 Ushimado 0.55 0.20 0.74 99 0.40 0.96 101 * Eye lotion and sea-water samples were diluted 2500- and 5-fold, Sea- water- respectively.On the other hand, control of the pH of the sample solution injected is important in this method. The sample solutions were prepared at a pH of 5.0, which is similar to that of the sorbitol solution, in order to avoid colour changes in the indicator due to the acid-base reaction. Calibration Graph The calibration graph has a tendency to deviate from linearity as shown in Figs. 2 and 3. However, the calibration graph is 10 min - 1 3 II 3 3 1 Time - Fig. 5 mol dm-3 MO; 2, 1% m/v NaCl; and 3, 3% m/v NaCl. Wavelength: ( a ) 400; (6) 500; (c) 520; and (d) 600 nm. Reagent and carrier solutions, distilled water. Other conditions as in Fig. 1 Effect of sodium chloride on peak height.1, 2.0 X Table 3 Determination of boron in sea-water samples Boron contentlmg I-' Sample* Proposed Spectrophotometric methodt method$ Artificial9 4.39 f 0.04 4.52 Japan Sea7 3.61 ? 0.02 3.80 Set0 Inland Seal1 2.80 k 0.03 2.49 Set0 Inland Sea** 2.75 k 0.02 2.72 * Sea-water samples were collected on July 21, 1990. t Average of three determinations k standard deviation. $ Reference 8. 9 Theoretical content: 4.50 mg 1-1 of boron. fi At Hagi, Yamaguchi Prefecture. 11 At Hinase, Okayama Prefecture. ** At Ushimado, Okayama Prefecture. essentially linear up to 1.2 mg 1-1 of boron under the optimum conditions (Fig. 4). The regression coefficient was 0.9992 (six data points), and the detection limit for a signal-to-noise ratio of 3 was 0.02 mg 1 - 1 of boron.The precision of the proposed method was determined by injecting ten samples and record- ing the peak absorbances. Relative standard deviations were 0.28 and 0.54% for 0.5 and 1.2 mg 1 - 1 of boron, respectively. The sampling rate was about 20 samples h-1. Applications The boron content in a number of eye lotions was determined by the proposed FI method. Owing to the high content of boric acid, the eye lotion samples were diluted 2500-fold with water and 120 p1 portions were injected directly without adjustment of the pH. The results obtained and details of the recovery tests are summarized in Tables 1 and 2. An attempt was also made to apply the method to the determination of boron in sea-water. The direct injection of sea-water samples into the FI system caused serious positive errors.In order to investigate the effect of the matrix (NaCI), distilled water was used as both the reagent and carrier stream, and a 2.0 x 10-5 mol dm-3 MO solution (reference), and 1 and 3% m/v NaCl solutions were injected over the entire visible wavelength range. The flow signals obtained at 400, 500, 520 and 600 nm are shown in Fig. 5 The signals given by the MO solution were simple and obvious, and no signal was observed at 600 nm because of the lack of absorption. However, the signals given by the NaCl solutions were714 ANALYST, JULY 1991, VOL. 116 complex (double peaks) and did not change over the wavelength range studied. This is mainly due to the difference in the refractive index of NaCl solutions and distilled water.In order to eliminate this matrix effect, a 0.6% m/v NaCl solution was used as the carrier stream instead of distilled water. The slope of the calibration graph obtained was 12% lower than the original graph. The pH of the sea-water samples was adjusted to 5.0 with dilute HC1 and the samples were diluted 5-fold with water. The results are summarized in Tables 2 and 3. Results for the determination of boron in eye lotions and sea-water samples using the proposed method were compared with those obtained by batch spectrophotometry using azo- methine H* and high-performance liquid chromatography (HPLC) with conductimetric detection.11 Owing to the broad matrix peak, the boron content in sea-water could not be determined by the HPLC method. The results given by the proposed method were in good agreement with those obtained with the spectrophotometric and HPLC methods (Tables 1 and 3). The authors thank Professor Kyoji T8ei for helpful discussions. References 1 Spielholtz, G. I., Toralballa, G. C., and Willsen, J. J., Mikrochim. Acta, 1974,4, 649. 2 Horta, A. M. T. C., and Curtius, A. J., Anal. Chim. Acta, 1978, 96, 207. 3 Agazzi, E. J., Anal. Chem., 1967,39, 233. 4 Castillo, J. R., Mir, J. M., Martinez, C., and Bendicho, C., Analyst, 1985, 110, 1435. 5 Rand, M. C., Standard Methods for the Examination of Water and Wastewater, American Public Health Association, Wash- ington, DC, 1975, pp. 287-291. 6 Williams, W. J., Handbook of Anion Determination, Butter- worth, London, 1979, pp. 23-39. 7 Ducret, L., Anal. Chim. Ada, 1957, 17, 213. 8 Capelle, R., Anal. Chim. Acta, 1961, 24, 555. 9 Boeseken, J., Adv. Carbohydr. Chem., 1949,4, 189. 10 Okada, T., and Kuwamoto, T., Fresenius 2. Anal. Chem., 1986, 325, 683. 11 Zenki, M., Ohtani, J., Ikeda, T., and Tdei, K., Chem. Lett., 1990, 885. Paper 0l05563H Received December 1 I th, 1990 Accepted February 26th, 1991

ISSN:0003-2654    

DOI:10.1039/AN9911600711

出版商:RSC    

年代:1991

数据来源: RSC

摘要:

ANALYST, JULY 1991, VOL. 116 715 Bidentate Organophosphorus Compounds as lonophores for Calcium- and Uranyl-selective Electrodes Oleg M. Petrukhin, Elga N. Avdeeva, Alex F. Zhukov, lrena B. Polosuchina, Svetlana A. Krylova and Svetlana L. Rogatinskaya D. I. Mendeleev Institute of Chemical Technology, Miusskaja sq. 9, Moscow 125820, USSR Georgiy V. Bodrin, Nina P. Nesterova, Yury M. Polikarpov and Martin 1. Kabachnik A. N. Nesmeyanov Institute of Organoelement Compounds, Vavilova str. 28, Moscow 117813, USSR Neutral bidentate organophosphorus compounds have been investigated as electroactive compounds for Ca2+- and U022+-selective electrodes. The results obtained enabled Ca2+ and U022+ electrodes to be developed and the optimum conditions for the operation of these electrodes to be determined.Some general conclusions on the nature of the selectivity of the ligands investigated are also presented. Keywords: Calcium-selective electrode; uran yl-selective electrode; neutral organophosphorus compound; selectivity coefficient; membrane composition The development of highly selective membrane electrodes based on neutral carriers is one of the most promising trends in ionometry. (The term ionometry is proposed as the name for the branch of electroanalytical chemistry based on the use of ion-selective electrodes, by analogy with the term poten- tiometry.) The successful development of these electrodes is, in many respects, determined by the availability of a reliable theory explaining the selective behaviour of membranes with neutral carriers and allowing one to formulate the principles governing the design of ionophore structures with pre- assigned properties.To date, a number of ionophore-based ion-selective electrodes (ISEs) have been developed and investigated.’-3 Neutral bidentate organophosphorus com- pounds capable of selective complexation with metal cations have been extensively applied in extraction studies, owing to the specific structure of the mutual arrangement of the phosphoryl groups (P=O).4 The possibilities of varying the bidentate reagent structures make it possible to carry out a search for the optimum reagent for a specific purpose. In the present work, the bidentate organophosphorus compounds chosen for investigation for Ca’+ and U022+ electrodes were of the type for which theoretical and experimental data were available, indicating their high efficiency and selectivity in complexation and extraction proces~es.~.s Based on the analysis of the experimental data obtained, those compounds that proved to be the most efficient as Ca’+ and U022+ ionophores were selected.The optimum composition of the membrane and the conditions for the operation of these electrodes have been determined. Experimental Reagents and Reactants All the reagents used were of analytical-reagent grade. Poly(viny1 chloride), brand S-70 (Chlorcaustik, Volgograd, USSR), was used. Plasticized membranes were prepared by using tetrahydrofuran purified by the procedure described in reference 6, sodium tetraphenylborate (NaTPB) (VEB Jena- pharm, Jena, Germany), ortho-nitrophenyl octyl ether (0- NPOE) and potassium tetrakisb-ch1orophenyl)borate (KTCPB) (Fluka, Buchs, Switzerland), and also sodium tetrakis (p-chloropheny1)borate (NaTCPB), which was synthesized by the procedure of Cassaretto et a1.7 The neutral organophosphorus compounds studied were synthesized at the A.N. Nesmeyanov Institute of Organoele- ment Compounds. These compounds can be sub-divided into three groups. R R / \ R R R R \ / / \ R R Fig. 1 investigated Structures of the bidentate organophosphorus compounds (1) Ligands containing a phosphoryl and a carbamoyl group (Fig. 1, ligands 1-8). Oxides of dialkyl (ary1)-dialkylcarbam- oylmethylphosphines were synthesized by the reaction between the ethers of trivalent phosphorous acids and amides of chloroacetic acid according to the following reaction: R2P(OC2H5) + CICH2C(O)NR’2 -.c RZP-CH2-CNR‘2 I I 1 0 where R = alkyl, alkoxy, phenyl or tolyl and R’ = alkyl.716 ANALYST, JULY 1991, VOL. 116 (2) Ligands containing two phosphoryl groups with an alkylene or vinylene bridge (Fig.1, ligands 9-14). The reactants of this group were obtained by the action of a phosphine oxide, containing a chloromethyl group, on a secondary phosphinous acid ether. As a result, a tetrasubsti- tuted methylenediphosphine dioxide is formed by the reaction Compounds 12-14 were obtained in a similar way (Fig. 1). (3) Ligands containing two phosphoryl groups with a xylylene bridge (Fig. 1, ligands 15-20). These ligands were synthesized by starting from ortho- and rneta-xylene, which were first converted into the corresponding bromo derivatives by direct bromination of the side chains.The bromo com- pounds thus obtained were then subjected to the Arbuzov reaction with the ethyl ether of diphenylphosphinous acid: A 0.1 rnol dm-3 calcium chloride solution was prepared by dissolving a weighed amount of the salt in water. The initial solution was standardized complexometrically.8 Calibration solutions, from 1 x 10-2 to 1 x 10-6 rnol dm-3, were prepared by successive dilution of the 0.1 rnol dm-3 calcium chloride solution. In order to prepare solutions with a constant ionic strength, a 0.5 rnol dm-3 potassium chloride solution was used. The pH of the calcium chloride calibration solutions was adjusted by adding hydrochloric acid or sodium hydroxide solution.A 0.1 rnol dm-3 solution of uranyl nitrate was prepared by dissolving a weighed amount of the salt in 0.6 rnol dm-3 potassium nitrate or chloride solution. The initial solution was standardized by using a gravimetric method.9 Calibration solutions, from 1 x 10-2 to 1 x 10-5 rnol dm-3, were prepared by successive dilution of the 0.1 rnol dm-3 uranyl nitrate solution with the solution of the background salt (0.6 mol dm-3 KN03). The pH of the calibration solutions was adjusted to 2.70 k 0.05 by adding nitric or hydrochloric acid. Membrane Preparation and Electrode Fabrication Plasticized membranes were prepared by a standard pro- cedure.10Two types of electrode were investigated: (i) with an internal reference solution and (ii) with a solid contact. In this work a graphite bar (pure for analysis) 5 mm in diameter was used.E. m. f. Measurements The electrode characteristics were investigated by using the following electrochemical cells: Ag, AgCl I Internal solution I Plastic membrane I and Graphite I Redox system I Plastic membrane 1 Solution to be investigated 11 Saturated KC1 I AgCl, Ag Solution to be investigated 11 Saturated KC1 I AgCl, Ag Two redox systems were used: (1) Fe3+-Fe2+ [equal volumes of 0.1 rnol dm-3 solutions of Fe2(S04)3 and FeS04] and (2) quinone-hydroquinone (up to 30% of quinhydrone was introduced into the internal layer of a two-layer mem- brane). The potentials were measured with an 1-120 M digital ionometer and a pH-121 pH meter (SIP, Gomel, USSR). The solution used for the electrodes with an internal reference solution had the following composition: 0.1 rnol dm-3 UO2(NO3)2 + 0.1 rnol dm-3 KCI, pH = 2.70 f 0.05, for U022+ electrodes; and 0.1 rnol dm-3 CaCI2 + 0.1 rnol dm-3 KCI, pH = 6, for Ca2+ electrodes. Electrodes with an internal reference solution and with a solid contact have similar characteristics.An EVL-1M AgCl electrode was used as the external reference electrode. The ionic strength (p) and/or the acidity of the solutions investigated were kept constant in each series of measure- ments, which made it possible to construct calibration graphs of potential (EImV) versus the negative logarithm of the potential-determining ion concentration (-log c). The selectivity coefficients, ki,j, were determined by the mixed solution method.10 As shown by the investigations performed, the selective properties of the membrane depend to a considerable extent on the properties of the membrane solvent [specifically, on its relative permittivity (E)].The following were investigated as membrane solvents: dibutyl phthalate (E = 6.1), diisooctyl sebacate (E = 4.1) and o-NPOE (E = 24). The highly polar o-NPOE proved to be the most effective for Ca2+ and U022+ membranes. Results and Discussion Preliminary investigations showed that the Ca2+ ISEs based on dialkylcarbamoyl compounds (Fig. 1, ligands 1-6) have linear electrode functions in the concentration range 1 x 10-3-1 x 10-1 rnol dm-3. The electrode function slopes are less than the theoretical slopes. Replacement of substituents on the nitrogen atom of the carbamoyl group does not significantly improve the selectivity of the electrode for Ca2+. However, the introduction of a phenyl group at the phos- phorus atom substantially improves the characteristics of the electrode (Fig.2). All the U022+ electrodes based on dialkylcarbamoyl compounds (Fig. 1, ligands 2 4 , 7 and 8) have non-linear electrode functions. Therefore, only the diphosphine dioxides were subsequently investigated. The structure of the bridging group in these compounds plays an important role in complexation. By studying the dioxides of tetraaryl(alkyl)alkylenediphosphines (Fig. 1, ligands 9-14) and tetraaryl(alkyl)xylylenediphosphines (Fig. 1, ligands 15-20), it was found that, for complexation with cations, the most favourable situation is the presence of a xylylene bridge in the ligand molecule, which imparts a conformational rigidity greater than that for a methylene or vinylene bridge.Compounds with a xylylene group are therefore of particular interest as potential electroactive compounds for Ca2+ and U022+ membranes. By using tetratolyl-o-xylylenediphosphine dioxide (o-TXDO, Fig. 1, ligand 18) and tetratolyl-rn-xylylenediphosphine dioxide (m- TXDO, Fig. 1, ligand 20), Ca2+ electrodes were fabricated. The use of rn-TXDO as the electroactive compound leads to an improvement in the electrochemical characteristics of the electrode: detection limit, 2.1 x 10-6 rnol dm-3, electrode function slope, 24.0 mV (for the electrode based on o-TXDO the corresponding values are 2.2 x 10-5 rnol dm-3 and 19.7 mV), and a considerable increase in the selectivity of the electrode for Ca compared with Mg.According to Morf's model, the selectivity of neutral ligands is determined by the constant of the formation and the distribution of metal complexes. 11 If the complexes formed are assumed to have the same composition, then the ex- pression for the selectivity coefficient, kj,j, will take the following form: where i is the determined ion, j the interfering ion, S the neutral ligand, pjs and pis are the stability constants in theANALYST, JULY 1991, VOL. 116 0 1 --= -1 2; 5 cn 1 -2 717 - Na+, K+ Li + - - Mg2+ - Sr2+ - 0a2+ K+ Na+ - - Sr2+ - Sr2+ - - Li+, Sr2+ - Li+ - Mg2+ - Ba2+ - Sr*. - Mg2+, Ba2+ - - Mg2+, Ba2+ - Mg2+ - Mg2+ - Ba2+ - Sr2+' Li + - - Li+ - Lit 0a2+ - - Compound number K' I - K+ - K' - I - Na+ - K+ - Na+ Na+ - - Na+ I -3 ' I Fig.2 Selectivity coefficients of Ca electrodes based on compounds 1-6 (see Fig. 1) Table 1 Selectivity cocfficicnts of Ca2+ electrodes (pH = 6, 1.1 = 0.5 mol dm-3, concentration of interfering ion = 0.1 mol dm-3 Interfering ion (j) Electroactive compound Li + K+ Mg'+ Sr2+ Ba2+ O-TXDO - 5.2 x 10-3 6.3 x 10-3 1.9 x 10-2 3.2 x 10-2 WZ-TXDO 3.2 x 10-3 1.0 x 10-3 1.7 x 10-5 4.2 x 10-1 4.0 X 10-2 aqueous phase and Pis and Pis the partition constants of the complexes. This equation defines the dependence of the selectivity coefficient on the complexation properties of the cations and ligands and on the properties of the membrane solvent that influence the distribution of the complexes. The expression for the selectivity coefficient, in the form of equation (l), can be applied to the systems considered here, which are designed for determining Ca2+ and UOZ2+ ions.The selectivity of the complexation reaction is determined both by stereochemical and electronic effects. The influence of steric effects is evidenced by a comparison of the selectivity coefficients, ki,+ for Ca2+ electrodes with o-TXDO and rn-TXDO (Table 1). The electrode based on rn-TXDO has a higher selectivity for Ca compared with Mg than that based on o-TXDO. It is of interest to note that for Sr and Ba, cations that are larger in size than Ca, the difference in the selectivity coefficient of the Ca*+ electrode with o-TXDO and rn-TXDO as ionophores decreases. Before choosing the electroactive compounds for the UO2'+-selective electrode, a large number of compounds were tested (Fig.1, ligands 2,4, 7 , s and 15-20). The ligands 17 [tetraphenyl-o-xylylenediphosphine dioxide (o-PXDO)] and 18 (o-TXDO) proved to be the most promising. The electrode function slopes of 28 k 1.3 mV per pU022+ unit (o-PXDO) and 30 k 2.2 mV per pUOZ2+ unit (o-TXDO) show that the potential-determining reaction, viz., the inter- phase transfer of uranium(vi), can be described in its general form by the equation: UO~*+ + nS + W n 2 + (2) where the overbar indicates the compounds in the membrane phase, i.e., the potential-determining ion in a given situation is the U02*+ ion. As for the stoichiometric coefficient, n , this can vary from 1 to 3, in accordance with the coordinating possibilities of the UO22+ ion.The linear portion of the electrode function follows the theoretical slope over a comparatively narrow range of U02*+ ion concentration: 1 x 10-5-1 x 10-3 mol dm-3 (o-TXDO) and 1 x 10-4-1 x 10-2 mol dm-3 (o-PXDO). The narrow linear region of the electrode function is explained by the fact that, in addition to the main reaction (2), a number of other reactions between uranium(v1) and the membrane com- ponents are taking place: ~ 0 2 2 + + X- + n S - UO~XS,+ M + A- -+ MA [A = TXPB-(TPB-), (X = C1-, OH- and NO3-) M+ = background salt cation] (X = C1-, NO3-, SO4*- and OH-) (3) (4) ( 5 ) A shift in the linear region of the electrode function, which depends on the type of ionophore, is explained by the different stabilities of the U02Sr,*+ complexes. The influence of the electronic effects of the substituents in the organo- phosphorus compounds on the complexation process is clearly seen when one compares the results for two groups of ligands acting as electroactive compounds: the first group are ligands with tolyl (o-TXDO) and phenyl (o-PXDO) substituents on the phosphorus atom of the phenyl group; the second group are ligands with butyl (Fig, 1, ligand 15) and butoxy substituents (Fig.1, ligand 16). In both instances, in going to less electronegative substituents, from phenyl to tolyl and from butoxy to butyl, an increase in the electron density on the donor oxygen atoms occurs, as a result of which the stability of the U02Sn2+ complexes is increased. The greater stability of the complex of uranium(v1) with the ligand o-TXDO, com- pared with o-PXDO, leads to a lower detection limit (Fig.3). The increasing stability of the uranium(v1)-ionophore complex increases the concentration of the positively charged U02Sn*+ complex in the membrane phase and, correspond- ingly, the concentration of mobile inorganic counter ions. Hence, the greater the stability of a complex the lower is the detection limit; however, the uranium(v1) concentration at which the effect of the anion manifests itself is also lower, i.e., the linear range of the electrode function becomes narrower. A change in the electrode function slope with an increase in the uranium(v1) concentration, more noticeable for hydro- chloric acid solutions, is probably associated with the forma- tion and transport of singly charged U02XS,,+ complexes [equation (3)] (X = NO3- and Cl-).An increase in the U02*+ + nXz- + U02X, (2-zn)718 60 > E 40’- a 20 ANALYST, JULY 1991. VOL. 116 - - 100 I 1 80 1 O t -20 I ’ I I I I 6 5 4 3 2 1 -Log (cuoz2+lrnol ~ I r n - ~ ) Fig. 3 Electrode function of UOZ2+ electrodes based on: 1, o-PXDO and 2, o-TXDO loo i ‘k 80 . > E 60 - a 40 - I .\ I I I I I 0 0.2 0.4 0.6 0.8 1.0 cKNO3/rno1 dm-3 Fig. 4 Influence of the background electrolyte concentration on the potential of U022+ electrodes based on o-PXDO (pH = 2.7 f 0.05). Uranium(vi) concentration ( C ~ V I ) : 1, 1 X 10-3; 2, 1 X 10-4; and 3, 1 x 10-5 mol dm-3 electrode function slope with an increase in the pH of the solution, and also with increasing uranium(v1) concentrations, can be explained by the fact that, together with the U02Sn2+ and U02XS,+ complex formation reactions, an increasing contribution to the exchange current is being made by the U020HS,+ complex.Hence, together with reaction (2), a substantial contribution to the potential can be made by the distribution of the U022+ ion according to reaction (3). In order to eliminate the effect of anions, KTPB o r KTCPB was introduced into the membrane. Both these substances, however, are electroactive with respect to the alkali o r alkaline earth metal ions [equation (4)]. Therefore, an increase in the lipophilic anion concentration in the membrane above a certain limit also leads to a narrowing of the linear range of the electrode function. The effectiveness of the use of an ISE is also governed by the reactions of the ion to be determined with the background electrolyte anions in the aqueous phase [equation ( 5 ) ] .A study of the influence of the concentration of the background electrolyte on the U02*+ ion-selective electrode potential showed that an increase in the concentration of all the salts Table 2 Stability constants of complexes formed between U02’+ ions and acid ligands.12 All values were measured at 25 “C unless stated otherwise Equilibrium Log B P [ UO?_SO,] s [ UO?_’+].[SO4’-] 1.81 2.7 [UO2(SO,)2*-] G [U02*+].[S04’-]2 2.5 k 0.2 2.7 [UO,(S0,),4-] e [U02*+]*[SO~2-]3 3.7* 1 .o* [UO*Cl+] e [U022+].[Cl-] -0.06 2.0 [U02NO3+] S [U02*+].[N03-] -0.6 2.0 [U020Hi] $ [U022+].[OH-] 8.0 * 0.0 0.5 * 20 “C. Table 3 Selectivity coefficients, kuo22+, ,, for the U022+ electrode j kUo22+.x 104 j kUO2z+. x 10“ Na+ 0.4 co2+ 0.8 K+ 0.2 Ni’+ 1.3 Ca2+ 0.6 cu*+ 0.7 Mg2 + 0.6 AF+ 0.2 Ba2+ 0.6 Fe3 + 7.4 investigated (KN03, KCI and Na2S04) initially resulted in a decrease in the potential, which then remained constant (Fig. 4). A decrease in the uranium(v1) ion activity determined is proportional to the stability constant of the complex formed between the U022+ ion and the background salt anion (Table 2). In those solutions with a high background salt concentra- tion the electrode potential is therefore actually determined by the equilibrium concentration of the free uncomplexed U022+ ion, and not by its over-all concentration. The U022+ ion concentration is inversely proportional to the stability constant of the corresponding complex of uranium(v1) with the acid ligand.Hence, the most stable complex with the sulphate ion (see Table 2) provides the maximum value for the detection limit. The pH is an important factor in the determination of uranium(v1). As the stability constant of U 0 2 0 H + is high (see Table 2), it is necessary to ensure that the pH is sufficiently low to prevent hydrolysis of the U022+ ion which starts to occur at a pH of approximately 3. However, at low pH values the electrode potential is influenced by the H+ function of KTPB and the plasticizer. These constraints make it necessary to conduct the measurements of the solutions investigated over a very narrow pH range (2.70 k 0.05). As a result of the complexity of the system studied, caused by the fact that, in addition to the main reaction, viz., the interphase transfer of uranium(v1) [reaction (2)], a number of other electrochemical and chemical reactions are taking place, there appears to be a comparatively narrow range for the optimum ionometric conditions for a given U022+ ion-selective electrode.The U022+ electrode based on o-TXDO has a high selectivity with respect to alkali and akaline earth metal ions. The selectivity coefficients, kUO22+ j , for the U022+ electrode based on o-TXDO at pH 2.70 k 0.05 with a 0.1 mol dm-3 concentration of the interfering ion against the background solution (0.6 mol dm-3 KN03) are given in Table 3. Neutral organophosphorus compounds therefore constitute an interesting class of ionophores with great potential; these compounds will be investigated further.References 1 Morf, W. E., and Simon, W., Hung. Sci. Instrum., 1977,41, 1. 2 Buchi, R., Pretsch. E.. and Simon, W., Tetrahedron Lett., 1976, 20, 1709. 3 Oggenfuss, P., Morf. W. E., Oesch, U., Ammann, D., Pretsch, E . , and Simon, W., Anal. Chim. Acta, 1986, 180. 299. 4 Chmutova, M. K.. in Theory and Practice of Extraction Methods, eds. Alimarin, I . P., and Bagreev, V. V.. Nauka. Moscow, 1985, p. 120.ANALYST, JULY 1991, VOL. 116 719 5 Zhukov, A. F.. Koldajev, A. B., Nesterova, N. P., Medved, T. J., and Kabatchnik, M. I., Neutral Bidentate Organophos- phorus Reagents as Electrode Active Compounds in Ion- selective Electrodes, 4th Symposium on Ion-selective Elec- trodes, Matrafured, 1984, pp. 711-721. 6 Gordon, A. J., and Ford, R. A., The Chemist’s Companion, Wiley, New York, London, Sydney and Toronto, 1972. 7 Cassaretto, F. P., McLafferty, J. J., and Moore, C. E., Anal. Chim. Acta, 1965, 32, 376. 8 Schwarzenback, G., and Flaschka, H . , Die Komplexometrische Titration, Ferdinand Enke Verlag, Stuttgart, 1965. 9 Markov, V. K., Verniy, E. A,, Vinigradov, A. V., Elinson, C. V., Kligin, A. E., and Moiseev, I. V., Uranium. Methods for its Determination, Atomizdat , Moscow, 1964. 10 Cammann, K., Das Arbeiten mit Ionen-selektiven Electroden, Springer-Verlag, Berlin, Heidelberg and New York, 1977. 11 Morf, W. E., Principles of Ion-selective Electrodes and of Membrane Transport, AkadCmiai Kiad6, Budapest, 1981. 12 Smith, R. M., and Martell, A. E., Critical Stability Constants, Volume 4. Inorganic Complexes, Plenum Press, New York, 1976. Paper 9/04789A Received November 8th, 1989 Accepted October 30th, 1990

ISSN:0003-2654    

DOI:10.1039/AN9911600715

出版商:RSC    

年代:1991

数据来源: RSC