ISSN:0003-2654    

DOI:10.1039/AN99621FX001

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

ANALAO 121(1) 1R-IOR, 1-94, 1N-16N (1996) JANUARY 1996IGUEST EDITORIAL 1NEDITORIAL 3NREVIEWS1RIllSAMPLE HANDLING1713ATOMIC SPECTROSCOPY/SPECTROMETRY1931MOLECULARSPECTROSCOPY/SPECTROMETRY374349SEPARATION SCIENCE5561B I 0 AN A LY T I C A L6771ELECT R 0 AN A LYTl C A L7579~ '"An a I y stThe analytical journal of The Royal Society of ChemistryCONTENTSJ. N. MillerHarp MinhasMethods for the Determination of fi-Agonists in Biological Matrices-A Review-Damien Boyd, MichaelO'Keeffe, Malcolm R. SmythContinuous-flow Method for the Determination of Phenols at Low Levels in Water and Soil LeachatesUsing Solid-phase Extraction for Simultaneous Preconcentration and Separation-Zheng-hang Zhi, AngelRios, Miguel ValcarcelDetermination of Trace Amounts of Cadmium in a Hydrometallurgical Zinc Refining Process Stream by aFlow-injection Method With On-line Preconcentration and Spectrophotornetric Detection-YutakaHayashibe, Yasumasa SayamaSpeciation of Nitrogen in Wastewater by Flow Injection-A.Cerda, M. T. Oms, R. Forteza, V. CerdaDetermination of Methylmercury in Sediments Using Supercritical Fluid Extraction and GasChromatography Coupled With Microwave-induced Plasma Atomic Emission Spectrometry-HAkanEmteborg, Erland Bjorklund, Fredrik Odman, Lars Karlsson, Lennart Mathiasson, Wolfgang Frech,Douglas C. BaxterDetermination of Ultra-trace Amounts of Selenium(iv) by Flow Injection Hydride Generation AtomicAbsorption Spectrometry with On-line Preconcentration by Co-precipitation with Lanthanum Hydroxide.Part II.On-line Addition of Co-precipitating Agent-Steffen Nielsen, Jens J. Sloth, Elo H. HansenStudy of the Chemiluminescent Characteristics of Ninhydrin and its Application-Guo Nan Chen, Xue QinXu, Fan ZhangExistence of Two Basic Sites in Triazolo-l,4-diazepines: Determination of Two pK, Values for a ModelCompound in Water-Beatrice Legouin, Jean-Louis BurgotSensitive Peroxyoxalate Chemiluminescence Determination of Psychotropic lndole DerivativesduanaCepas, Manuel Silva, Dolores Perez-BenditoMicrobore Liquid Chromatography-Electrospray Mass Spectrometry of Selected Synthetic PyrethroidInsecticides-Ian A. Fleet, John J. Monaghan, Derek B. Gordon, Gwyn A. LordInvestigation of On-line Reversed-phase Liquid Chromatography-Gas Chromatography-MassSpectrometry as a Tool for the Identification of Impurities in Drug Substances-Elise C.Goosens, Karel H.Stegman, Dirk de Jong, Gerhardus J. de Jong, Udo A.Th BrinkmanEstablishing the Cut-off Concentration for the Detection of Etrophine in Horse Urine-Robert F. Smith,Laurence S. Jackson, Andrew MooreAmperometric Biosensor for the Determination of the Artificial Sweetener Aspartame With an lnirnobilizedBienzyme System-Shu-Fen ChouDetermination of Low Concentrations of Nickel and Aluminium in Membrane Electrolyser Liquors-MichaelCullen, Susan LancashireElectrochemical Trace Analysis of Gold in Ore-Jyotsna Shukla, K. S. PitreTHE ROYALC H EM I STRYInformationServices Cambridge, EnglandTypeset and printed by Black Bear Press Limited,Continued on inside back cover-0003-2654C199611:1-OTHER METHODS838993iVvi ixixivCertified Reference Materials (CRMs 479 and 480) for the Quality Control of Nitrate Determination inFreshwater-Ph. Quevauviller, M.Valcarcel, M. D. Luque de Castro, J. Cosano, I?. MoselloSensitive Determination of Nitrite Using its Catalytic Effect on the Bromate Oxidation ofProchlorperazine-Ashraf A. Mohamed, Mohamed F. El-shahat, Tsutomu Fukasawa, Masaaki lwatsukiCUMULATIVE AUTHOR INDEXINSTRUCTIONS TO AUTHORSGUIDELINES FOR SUBMISSION ON DISKIUPAC PUBLICATIONS ON NOMENCLATURE AND SYMBOLISMREFEREEING PROCEDURE AND POLICY (1996)COPYRIGHT LICENCENEWSANDVIEWS 5N7N11N12N13N15NBook ReviewsConference DiaryCoursesPapers in Future IssuesTechnical Abbreviations and AcronymsConference Report-Peter HarrowingCover picture: Determination of @-agonists in food samples using solid-phase extraction (see p. 1 R).Image kindly supplied by Michael O’Keeffe, The National Food Centre, Teagasc, Dublin, Ireland

ISSN:0003-2654    

DOI:10.1039/AN99621BX003

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. 121 5N Book Reviews Science and the Detective By Brian H. Kaye. Pp. xviii + 38. VCH. 1995. Price DM68.00. ISBN 3-527-29252-7. This is an interesting compilation of selected areas of the forensic sciences and is to be highly recommended to those with little scientific knowledge. The author clearly uses the public appeal of forensic science as a vehicle for spreading the scientific gospel to a wide audience and he demonstrates an easy reading style which most should find enjoyable. Professor Kaye has a clear interest in the derivation of words from the Greek and Latin and while the reviewer found this appealing some readers may see this as an irritation. Science students about to enter university, lawyers and perhaps even forensic scientists will find this volume partic- ularly enjoyable and informative, but anyone with an enquiring mind will find it enthralling.This does not mean that there are no faults; a little better proof-reading would not have gone amiss when correcting the names in some of the cases cited (e.g., Scoose, Skuse) as well as the many serious errors in chemical structures; the structures relating to explosives (p. 107) and odours (p. 166) are particularly poor. ‘Science students about to enter university, lawyers and perhaps even forensic scientists will find this volume particularly enjoyable and informative, but anyone with an enquiring mind will find it enthralling. ’ The application of modem technology and the fundamental principles underlying these technologies are well covered and lucidly explained.What is perhaps missing is an appraisal of the relative merits of individual techniques and an explanation of why a particular method was used in preference to others in specific cases. This is a book of 12 chapters taking the reader from the crime scene through the laboratory and into the courts. It is well referenced for further general reading and is a well presented volume with good layout and illustration. This is a book which will and should be read by many. It’s only major fault is its price. At &30 for a paperback this could be a serious disincentive to buy! Professor B. Caddy 51900401 University of Strathclyde, Glasgow Plants and the Chemical Elements-Biochemistry, Up- take, Tolerance and Toxicity. Edited by M. E. Farago. Pp.292. VCH. 1994. Price f79.50. ISBN 3-527-28269-6. The text consists of 10 chapters written by experts in their topic. The scope of the work is considerable and complex and covers many multidisciplinary areas. Chapter 1 by Brian E. Davies introduces the book with an overview of soil chemistry and bioavailability with special reference to trace elements. His account provides a concise description of rock and soil geochemistry, soil processes and bioavailability. In Chapter 2 Aradhana Mehra and Margaret Farago focus on metal ions and plant nutrition providing a thorough treatment of the role of macro- and micro-nutrients and a brief review of uptake processes. Mark Macnair and Alan Baker present a fascinating account of the evolutionary aspects of plant populations which can tolerate high concentrations of metal ions in chapter 3.The review, which concentrates on Cu, Zn and Cd, also provides an insight into various experimental approaches. Following on, Robert Brooks explores plants that hyperaccumulate the heavy metals Co, Cu, Ni (and Zn). Included are some useful and extensive tables of plant species with location and concentra- tion. Chapter 5 , written by Robert Hay from Massey University in New Zealand, contains detailed information on the chemistry and structure of enzymes and proteins which by its nature is the most difficult information to digest. ‘the text is a valuable forum for exchange of current thinking across a broad range of disciplines. ’ In the book reviewed, figure 5.22 on page 141 is blank! The main discussion in the following chapter is on the toxic effects of metals at the cellular level written by Jaco Vangronsveld and Herman Clijsters from Belgium.Inhibition and induction of enzymes are the main topics. The authors use conclusions for each subject to provide helpful summaries to guide the reader through the chapter. Chapter 7 provides a good general summary of plants and radionuclides. G. Shaw and J. N. B. Bell review aspects of environmental pollution and contamination from a mechanistic approach including discussion on atmos- pheric deposition. Margaret Farago considers plants as in- dicators of mineralization and pollution in chapter 7, briefly reviewing mineral exploration and plants as indicators of pollution. In my view, this chapter logically follows on from chapter 4! The extensive topic of the analytical approach for plant analysis is briefly reviewed by Margaret Farago and Aradhana Mehra.Chapter 8 could provide botanical students with a useful introduction to analytical techniques. Finally Margaret Farago describes methods for the study of inorganic species in plant tissue using phytochemistry. The editor has managed to produce a book with appeal to a wide audience. I can summarize my feelings no better than the cover which concludes that the text is a valuable forum for exchange of current thinking across a broad range of disciplines. A. P. Rowland 41901 59B Institute of Terrestrial Ecology, Grange-over-Sands, Cumbria Thermal Plasmas. Fundamentals and Applications. Volume 1 By Maher I. Boulos, Pierre Fauchais and Emil Pfender.Pp. xiii + 452. Plenum. 1994. Price $85.00. ISBN 0-306-44607-3. This book is the first of two volumes concerned with plasma technology; a subject which covers a wide and diverse range of disciplines from chemical analysis through plasma physics to materials science. Volume 1 is concerned with the fundamental concepts, gaseous kinetics, thermodynamics and transport properties of plasmas. The second volume will be concerned more with scientific and engineering applications of plasmas. The authors state in the preface that both are aimed at scientists seeking a broad view of the subject and for graduate students entering the subject of thermal plasma technology. It is with this criterion that I have attempted to review Volume 1. In addition I have reviewed the book through the eyes of an analytical chemist to judge its worth to this group of the scientific community. Volume 1 can be broadly described as a book which provides the reader with a complete description of the plasma state containing all of the necessary theoretical concepts in an accessible presentation.The book divides more or less equally6N Analyst, January 1996, Vol. I21 into descriptions of the microscopic properties and macroscopic properties of the plasma state. Chapter one is very much an introductory chapter which sets the scene, describing high intensity arcs, free burning arcs, RF discharges and microwave plasmas, followed by a discussion of relevant properties of plasmas giving the reader a useful synopsis of the subject.The rest of chapter one then switches to descriptions of several industrial applications of plasmas, e.g., plasma deposition and plasma metallurgy. Clearly chapter one forms an introduction to both volumes. Chapters two to four are concerned with the microscopic properties of plasmas covering basic atomic and molecular theory, elementary particle kinetics and fundamental concepts in gaseous electronics. These chapters provide an important basic framework for understanding the plasma state but possibly the most useful for analytical scientists is chapter four which builds upon earlier chapters to provide the reader with a clear understanding of thermal equilibrium and local thermal equilibrium (LTE) in the plasma state. An understanding of these concepts is important in understanding the emission characteristics of analytical plasmas (e.g., ICP) as well as the fundamental influence of temperature in such plasmas.As expected the treatment is based upon Maxwell-Bolzmann statistics and the Boltzmann distribution, concepts which rely upon the existence of LTE. The conditions under which LTE prevails are clearly discussed and examples of deviations from LTE are given. This approach leads to a rigorous definition of the plasma state and a discussion of Debye length, Quasi- neutrality and charge carrier separation. LFor graduate students wishing to enter the research field of thermal plasmas this book provides an excellent grounding in the subject. ’ Macroscopic properties of thermal plasmas are discussed in chapters five, six, seven and eight.Again, the importance of temperature in determining the properties of these plasmas is stressed. In Chapter five the plasma equations for current, mass and heat flow are derived. Chapter six is concerned with the thermodynamic properties of the plasma state and considers partition functions and the species number density temperature dependence of mixed gas plasmas. Chapter seven is a discussion of transport properties in plasmas under conditions of electric field and temperature gradients. Thus, self diffusion, thermal and electrical conductivity and viscosity of the plasma medium are elucidated. The theme of mixed gas plasmas is picked up again in this chapter showing the effects of gas composition on thermal conductivity. Finally in chapter eight a fairly routine but brief presentation of radiation transport and emission is given which is of some importance to analytical spectroscopists using a plasma for chemical analysis.Given the range of material covered in Volume 1 the authors have made a very good job of achieving their aims. Not many books exist which will take the reader from basic atomic theory right through to the thermodynamics of the plasma state. Even fewer do so in the lucid fashion which these authors have done. I suspect however that the book is not pitched at graduate level analytical chemists, being rather more concerned with the physics of plasmas. This is reflected in the relatively high level of applied mathematics knowledge which it is assumed that the reader possesses. Nevertheless, for analytical chemists with a real interest in plasma diagnostics this book is really a very useful bringing together all of the underlying concepts of the plasma state.For graduate students wishing to enter the research field of thermal plasmas this book provides an excellent grounding in the subject. Professor R. D. Snook 4/90] 26F UMIST, Manchester Split and Splitless Injection in Capillary GC. 3rd Edition By Konrad Grob. Chromatographic Methods. Pp xxiv + 548. Huthig. 1993. Price DM148.00. ISBN 3-7785-2151-9. Many of us feel after a few hours with a gas chromatograph, that there is nothing much to it beyond choosing the right column and a suitable temperature programme. Konrad Grob believes otherwise, and has been at great pains in his writing to show that if one really wants to identify all the volatile components present, and still more, to measure them quantitatively, there is a great deal more to be learned about the technique.This substantial book is on injection techniques alone, which he demonstrates can have many pitfalls. If after reading it, one despairs of success, he does say at one point; ‘of course many analyses are free from such difficulties . . . assume the reader is sufficiently experienced to maintain a balanced picture of the technique ’ . ‘This substantial book is on injection techniques alone, which demonstrates can have many pitfalls.’ He divides the technique first into ‘classical’ vaporizing injection (which forms the greater part of the book, and is subdivided into split, splitless and direct injection) and programmed temperature vapourization (PTV) injection, known to some as cold injection, which forms a much shorter part and which is further subdivided into split, splitless, solvent split and direct injection.Grob is well known for his wamings about the dangers of split injection. To be able to write 212 pages on this subject alone, in a concise and unrepetitive way, clearly indicates there is a great deal to be considered. Even subjects like measuring the flow rate through the column, or the correct length of a syringe needle are considered and shown to be far from simple or unbiased. Silanizing inserts and glass wool for plugs has received a lot of study and he comes down clearly on the need for high temperature silanizing. It is nice to see ones own experiences (or prejudices?) supported, e.g., that a glass wool plug in a straight glass tube is the best kind of injector liner, and that retention gaps are well worth using, and they do not need to be more than about 50 cm long, though he does advocate glass press-fit connectors for joining gap and column.Personally, they seem to give a great deal of trouble and glass-lined metal fittings seem much more reliable, and cheaper in the long run. A section of 60 pages is devoted to the vexed question of evaporation from the needle and just how much sample is injected when ‘1 pl’ is delivered from a syringe. There is no simple answer. It is not an easy book in which to browse, but once one is familiar with some of it, it is easy reading, because the text is split into many sections with a marginal heading or note to every paragraph (three or four to the page), and with important statements always in bold type. There is a good subject index, a one-page summary of how to select the correct injection technique and five-page glossary of terms. Sometimes one is confused by apparently conflicting advice under different headings, although he does his best to conclude at the ends of chapters and a two-page summary of parameter selection at the end. The subject has advanced considerably, even since the second edition in 1990, so this book necessarily supersedes the earlier ones. It is not a book to force on to undergraduates, and its price is high, but no one should be allowed to direct a laboratory doing quantitative analysis by GC without first being thoroughly familiar with this book, or its contents. Professor E . D. Morgan 3190309E Keele University, StafSordshire

ISSN:0003-2654    

DOI:10.1039/AN996210005N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. 121 (7-11) 7 Determination of Trace Amounts of Cadmium in a Hydrometallurgical Zinc Refining Process Stream by a Flow-injection Method With On-line Preconcentration and Spectrophotometric Detection Yutaka Hayashibe and Yasumasa Sayama Materials Characterization Center, Central Research Institute, Mitsubishi Materials Co., 1-297, Kitabukuro-cho, Omiya, Saitama 330, Japan A flow-injection (FI) method was developed for the spectrophotometric determination of cadmium in a hydrometallurgical zinc refining process stream using l-(4-nitrophenyl)-3-(4-phenylazophenyl)triazene (Cadion) as the chromogenic reagent. The sample solution was injected into a carrier containing potassium iodide. The sample was then passed through an anion-exchange mini-column on which the analyte was concentrated as a cadmium-iodo complex.In order to extend the detectable range of cadmium, a multiple sample injection method, in which the sample solution was repeatedly injected into the carrier at regular intervals of 30 s, was applied. Cadmium on the column was eluted with 1 moll-1 nitric acid and merged with a stream of a mixture of masking agent (trisodium citrate-potassium sodium tartrate-potassium hydroxide) and Cadion. Finally, the absorbance of the cadmium-Cadion complex was measured at 480 nm. The proposed FI was fully controlled by a personal computer. The proposed system permitted throughputs of 6 samples h-l for single injection, and 2 samples h-l for single injection, and 2 samples h-1 for multiple sample injection (50 injections). The reproducibility was satisfactory with a relative standard deviation of less than 5.0% (0.14 pg ml-1 Cd level, n = 5) for the single injection method and 10% (2.0 ng ml-1 Cd level, n = 5) for the multiple sample injection method (50 injections).The detection limits were 0.028 pg ml-1 of cadmium for the single injection method and 0.83 ng ml-1 of cadmium for the multiple sample injection method (50 injections). The absolute amount of cadmium detectable, defined as the analytical signal equal to twice the uncertainty in the background, was 10 ng. Keywords: Flow injection; cadmium determination; on-line preconcerztration; anion exchange; spectrophotometric detection Introduction For the preparation of the zinc electrolyte in hydrometallurgical zinc refining, it is necessary to eliminate impurities which originate from zinc concentrate (sphalerite, calamine and smithsonite).In the preparation of the electrolyte, zinc powder is commonly utilized to eliminate impurities by cementation. The amount of zinc powder added to the process stream is controlled on the basis of the cadmium content, which must be monitored continuously. Flow injection (FI) is a rapid and precise technique, and a number of automated FI systems are commercially available. Although several methods for the determination of cadmium with FI have been published,2-5 their applications to on-line process analysis have scarcely been reported. Sensitive spectrophotometric methods for the determination of cadmium have been reported and several sensitive chromo- genic reagents for the spectrophotometric detection of cadmium are commercially available. Among these, 1 -(4-nitrophenyl)- 3-(4-phenylazophenyl)triazene (Cadion) forms a water-soluble complex with cadmium in basic media containing Triton X-100.6 In order to improve the selectivity and sensitivity of the Cadion method, however, the separation of cadmium from matrix zinc and preconcentration of cadmium are necessary before spectrophotometric detection.C~precipitation,~ solvent extraction,*-lO ion exchange, etc., have usually been used for separation of cadmium from the zinc matrix. Because of the applicability to aqueous media, we selected an ion-exchange technique. Various ion-exchange procedures have been de- scribed for the separation of cadmium from other elements.Although cadmium can be separated from zinc by cation- exchange or anion-exchange methods,' '-I4 their selectivity for cadmium is poor and a large amount of resin is required for zinc to be adsorbed. The anion-exchange adsorption behaviour of many elements in hydriodic acid media has been reported by Marsh et al.15 We have also investigated the anion-exchange behaviour of 12 elements in potassium iodide media and found that cadmium is selectively adsorbed on the strongly basic anion-exchange resin Bio-Rad AG1 -X8 from dilute solutions of potassium iodide, and completely separated from matrix zinc.16 In this work, an FI system for the determination of trace amounts of cadmium in a hydrometallurgical zinc refining process stream was developed. In order to determine cadmium accurately and rapidly, the sample solutions were taken directly from a continuous-flow process stream.The sample solution was injected into a carrier (potassium iodide solution) and cadmium was concentrated on an anion-exchange mini-column as the iodo complex. After elution with nitric acid, cadmium was detected spectrophotometrically with Cadion as the chromogenic reagent. In general, the sensitivity of an FI system is defined and controlled easily by varying the volume of the injection loop. However, replacement of the injection loop for the adjustment of sensitivity to detect various amounts of analyte (from nanograms to micrograms) is tedious and time- consuming. Hence, a multiple sample injection method was studied in order to control the sensitivity of the proposed FI system.Experimental Reagents All reagents used were of analytical-reagent grade and all solutions were prepared with distilled water.8 Analyst, January 1996, Vol. 121 w2 A cadmium stock standard solution was prepared by dissolving 1.00 g of cadmium (99.999% purity) in 30 ml of nitric acid, expelling the nitrogen oxides and diluting to 1000 ml with 1 moll-1 nitric acid to yield a 1 mg 1-1 Cd-1.4 mol 1-1 nitric acid solution. Working standard solutions were prepared by appropriate dilution of the stock standard solution. A stock solution of Cadion was prepared by dissolving 200 mg of the commercial reagent (Aldrich, Milwaukee, WI, USA) in 1000 ml of 0.1 moll-1 potassium hydroxide-O.l% v/v Triton X-100 solution. A working solution of Cadion was prepared by diluting the stock solution appropriately with 0.1 mol 1-1 potassium hydroxide-O.l% v/v Triton X- 100.For the prepara- tion of 0.1 mol 1-l potassium iodide solution as the carrier solution, 16.8 g of potassium iodide were dissolved in the appropriate amount of water and diluted to 1000 ml with water. About 5.0 g of trisodium citrate dihydrate, 2.5 g of potassium sodium tartrate tetrahydrate and 112 g of potassium hydroxide were dissolved in 1000 ml of water to yield 1.7 X 10-2 moll-' trisodium citrate-8.8 X 10-3 mol 1-l potassium sodium tartrate-2.0 mol 1-1 potassium hydroxide mixture as the masking solution. The anion-exchange resin Bio-Rad AGl-X8 (100-200 mesh, C1- form, Bio-Rad Labs.) was used as received for preparing the anion-exchange mini-column.The capacity of the resin used was 1.2 mequiv. ml-1. The mini-column was prepared in the following manner: the resin was slurry-packed in a PTFE tube (100 X 1.0 mm id) and each end was plugged with the appropriate amount of cotton wool. - p2 6 - FI Manifold and General Procedure Fig. 1 shows a schematic diagram of the manifolds used. A Hitachi U- 1000 ratio-beam spectrophotometer equipped with a 60 pl flow cell (20 mm pathlength) was used as the detector. Sanuki-kogyo DMX-2400T double-plunger pumps were used for delivering the carrier, the eluent, the chromogenic reagent solution and the masking solution. A peristaltic pump, ATTO AC-2120, was used to pump sample solutions. Sanuki-kogyo SVA-6M2H automated six-way valves, made of ceramic material, were used for sample introduction and line switching for the ion-exchange column.PTFE tubing (1.0 mm id) was used to construct the analytical manifold. The dimensions are specified in the discussion. All of the pumps and valves and the spectrophotometer were controlled by an NEC-PC980 1 perso- nal computer, for which a control program (MS-DOSm88- BASIC ver.6.1) written in this laboratory was used. - C D P3 - Carrier solution was pumped into the analytical line at a flow rate of 1 .O ml min-l by the pump P1. The sample, taken directly from the zinc refining process with an on-line sampler, was injected into the carrier stream with the six-way valve V1 (350 pl), and passed through the mini-column connected to the six- way valve V2 to adsorb the cadmium-iodo complex.For the multiple sample injection method, 350 pl portions of sample solution were repeatedly injected into the carrier at regular intervals of 30 s. At 220 s after the sample injection, the eluent (1.0 moll-1 nitric acid) was introduced into the mini-column at a flow rate of 1.0 ml min-1 by switching V2. The effluent was merged with the chromogenic reagent solution and the masking solution at a flow rate of 0.75 ml min-1, and the absorbance of the cadmium-Cadion complex was monitored in the flow- through cell at 480 nm. Results and Discussion Anion-exchange Adsorption Behaviour of Cadmium and Zinc in Acidic Potassium Iodide Media Table 1 shows the compositions of typical zinc electrolytes used for hydrometallurgical zinc refining. Because of the high matrix concentration and interferences caused by the co-existing elements, it is difficult to determine cadmium directly with a spectrophotometric method.The adsorbabilities of many ele- ments on an anion-exchange resin in hydriodic acid media have already been published. 1 1915 However, details of the distribution coefficients in lower concentrations of hydriodic acid have not been reported. We have measured the anion-exchange distibu- tion coefficients of 12 elements in acidic potassium iodide media.I6 It is well known that cadmium forms stable anionic iodo complexes ([CdI3]- and [CdI,]*-). We believe that the form of the cadmium-iodide complex is [CdI#-- in 0.1 moll-' potassium iodide solution from an estimate of the stability constants of the two species.17 The results obtained suggest that cadmium is almost completely sorbed on the anion-exchange resin column at potassium iodide concentrations above 0.05 moll-1, whereas zinc is weakly adsorbed at such concentration levels, and can be washed out from the column with 0.05 moll-' potassium iodide.Although iodide is unstable in acidic media, potassium iodide in neutral solution can be stored for at least 7 d. Furthermore, because of the short mixing time of the iodide solution with the acidic solution, and hence little liberation of iodine, the anion-exchange method in iodide media can be employed for the on-line preconcentration of cadmium. The cadmium-iodo complex can be rapidly and completely eluted from the column with dilute nitric acid. Spectrophotometric Detection of Cadmium The conventional methods used for the determination of trace amounts of cadmium are AAS and extraction photometry with dithizone.The dithizone method, which requires extraction and back-extraction, is tedious and time-consuming. Other conven- ~ ~~~~ Table 1 Composition of typical zinc electrolyte used for hydrometallurgical zinc refining Species Ca2+ Co*+ Fez+ Mn2+ Ni*+ SW+ s o p * mg ml-1. Content/ pg ml-' 200 0.1 15 5* <0.1 <0.1 = 180* Species Cd*+ c u2+ K+ Mg2+ Pb2+ Bi'+ Zn'+ Content/ pg ml-* 0.2 0.1 5* 8* <0.1 <0.1 > 150*Analyst, January 1996, Vol. 121 9 ient methods, which are more selective and sensitive than the dithizone method, have scarcely been reported. Hence, the Cadion6 and 5,10,15,20-tetraphenyl-2 lH,23H-porphyrinetetra- sulfonic acid, disulfuric acid, tetrahydrate (TPPS)18 methods were considered for the spectrophotometric detection of cadmium.These reagents react with cadmium and form water- soluble complexes with high molar absorptivity (E > 105 m2 mol-I). Although Cadion is less sensitive than TPPS, 0.1 pg ml-1 levels of cadmium can be detected with the former in alkaline media in combination with Triton X- 100 as solubilizing agent. Taking into account the high reaction rate and low cost, Cadion was chosen as the chromogenic reagent for the detection of cadmium. Optimization of the FI System The single-line manifold shown in Fig. 1 was constructed to introduce the sample into the FI system. It would seem to be more appropriate to use a multiline arrangement and merge the iodide stream with the sample.However, the total time for sample introduction using a multiline manifold would probably be longer than that using a single-line manifold. Hence, the single-line manifold was used. The dispersion of the sample injected is sufficient to ensure mixing with the carrier and complex formation with iodide in the proposed system. The influence of the concentration of the potassium iodide solution (carrier) was examined in the range 0.01-0.5 rnol 1-1 and it was found that a potassium iodide concentration above 0.05 mol 1-l was necessary for cadmium to be adsorbed strongly on the anion-exchange mini-column. However, a concentration of potassium iodide of more than 0.2 mol 1-l cannot be employed because matrix zinc is also strongly adsorbed on the mini-column and cadmium cannot be com- pletely separated from matrix zinc.A potassium iodide concentration of 0.1 mol 1-1 was selected. The effect of the concentration of nitric acid as the eluent was investigated in the range 0.1-2.0 mol 1-1. The peak height increased steadily with increasing concentration of nitric acid, and hence the cadmium-iodo complex adsorbed on the column was rapidly eluted with increasing concentration of nitric acid. A concentration below 0.5 moll-1 is not recommended because the system peak caused by the zinc weakly adsorbed on the column appears in close proximity to the cadmium peak. Hence, 1 .O moll-1 nitric acid was employed to elute the cadmium-iodo complex from the mini-column. Cadmium can be separated from most of the co-existing elements in the zinc refining process stream by the anion- exchange mini-column.However, small amounts of bismuth, copper and lead accompany cadmium. These elements react with Cadion and interfere with the spectrophotometric detection of cadmium. Hence, 1.7 X 10-2 mol 1-1 trisodium citrate-8.8 X 10-3 moll-1 potassium sodium tartrate was utilized to mask these interferents.5 Cadion reacts with cadmium to form a stable complex in alkaline media; therefore, the effluent from the column must be made alkaline. Hence, a mixture of 1.7 X 10-2 moll-' trisodium citrate-8.8 X mol I-' potassium sodium tartrate-2 mol 1-1 potassium hydroxide was used as the masking stream. The effect of the Cadion concentration in the range 0.00005-0.02% m/v was tested, using 350 pl of sample.A maximum and constant response for 1.0 pg ml-1 of cadmium was obtained at a reagent concentration above 0.002% m/v. In order to dissolve Cadion in water, potassium hydroxide and Triton X- 100 were necessary as solubilizing agents.5 Hence, a mixture of 0.002% m/v Cadion solution-0.1% v/v Triton X-100-0.1 rnol 1-1 potassium hydroxide was used as the chromogenic reagent. The effect of the length of the reaction coils (Fig. 1) was examined at various flow rates of the carrier and reagent solutions. Coils of 0,0.5 and 1 .O m were tested for C1. The coil length was found to have no effect on the response of cadmium. This result suggested that cadmium reacts rapidly with iodide ion to form a stable [CdLJ2+ complex. The length of C2 was then varied from 1 to 5 m.A 3 m coil gave the most sensitive and precise results. Hence, the optimum lengths were 0.5 m for C1 and 3 m for C2. The effect of the flow rate of the carrier (Pl) on the peak height was studied in the range 0.5-2.0 ml min-l, by injecting 350 p1 of cadmium standard solution and a real sample solution which was spiked with cadmium standard solution to contain 0.4 pg ml-1 of cadmium. A constant response was obtained in the range tested. This result demonstrated that the anion- exchange reaction between the cadmium-iodo complex and iodide is rapid and that the complex is strongly adsorbed on the anion-exchange resin. A flow rate of PI of 1.0 ml min-l was selected. The influence of the flow rate of P2 for delivering the eluent was examined in the range 0.6-2.0 ml min-l.The peak height decreased steadily (Fig. 2) and the peak width broadened with increasing flow rate. This suggests that [CdIJ*- is strongly adsorbed on the ion-exchange resin and that the ion- exchange reaction between [Cd14]2- and Nos- is slow. The peak separation between the system peak caused by zinc adsorbed weakly on the mini-column and the cadmium peak was improved by decreasing the flow rate. The optimum flow rate of P2 was 1.0 ml min-1 to ensure sufficient sensitivity for the determination of 0.1 pg ml-1 levels of cadmium. The effect of the flow rate of P3 was investigated in the range 0.6-2.0 ml min-1. An optimum flow rate of 1.0 ml min-1 was employed to reduce the consumption of expensive chromogenic reagent and also to avoid dilution of the effluent.The influence of the sample volume on the absorbance was investigated by injecting various volumes ( I 65-870 pl) of cadmium standard and sample solutions into the carrier stream at the recommended flow rate and coil length. The peak height increased steadily with increasing injection volume. It was found that sufficient sensitivity to detect 0.1 pg ml-1 levels of cadmium was obtained by injecting volumes above 350 p1. Therefore, an injection volume of 350 pl was used subse- quently. Switching Sequence of Valves The influence of the adsorption time, which is the time from sample injection to the start of elution of cadmium by switching 0 0.5 1.0 1.5 2.0 Flow rate of P2 / ml min-1 Fig. 2 Flow rate of P2 versus maximum peak height absorbance. Filled circle, standard solution (0.2 yg ml-I Cd); open circle, sample solution (0.4 yg ml-I Cd-150 mg ml-1 Zn).The flow rate of pump P1 was kept at 1.0 ml min-1. The ratio of the flow rate of pump P2 to that of pump P3 was kept at 1.3.10 Analyst, January 1996, Vol. 121 V2, was investigated in the range 140-480 s by injecting 350 pl of cadmium standard solution or sample solution into the carrier stream at the recommended flow rate. The adsorption time had no effect on the response of cadmium and the maximum peak height absorbance was consistently obtained 80-90 s after switching V2. Furthermore, the system peak decreased with increasing adsorption time. This suggested that the cadmium- iodo complex is strongly adsorbed on the anion-exchange mini- column, whereas zinc is weakly adsorbed and washed out by the carrier.Hence, 220 s was chosen as the adsorption time. The elution time, which is the period from the start of the elution to the next sample injection, was also examined. An elution time of 200 s was found to be necessary for the quantitative recovery of cadmium, for which the column wash volume by the eluent was about 3.3 ml. The signal profiles obtained by using the proposed system are shown in Fig. 3. Concentration of Cadmium on the Anion-exchange Mini-column by the Multiple Sample Injection Method As mentioned above, cadmium is strongly adsorbed on the anion-exchange mini-column from potassium iodide media. The concentration of cadmium as the cadmium-iodo complex on the mini-column was attempted. In general, the sensitivity of an FI system is defined and controlled easily by varying the volume of the injection loop, However, replacement of the injection loop for the adjustment of sensitivity to detect various amounts of analyte (from nanograms to micrograms) is tedious and time-consuming.The multiple sample injection method was studied to vary the total injection volume of the sample by using the proposed FI system. Portions (350 pl) of the standard solution (20 ng ml-1 Cd) were repeatedly injected at regular intervals of 30 s. After the last sample injection, the anion- exchange mini-column was washed with the carrier for 220 s, and then the adsorbed cadmium was eluted with 1 moll-’ nitric acid. As can be seen in Fig. 4, the relationship between the number of sample injections and the peak height absorbance was linear from 1 to 50 injections, and the maximum peak height absorbance was consistently observed 80-90 s after switching V2.A ‘carryover effect’ can be observed in Fig. 4 when more than 50 multiple injections are performed. This suggests that a portion of the cadmium-iodo complex adsorbed on the resin was flowing out from the column. For multiple injections at regular intervals of 30 s using a 350 pl loop at a carrier flow rate of 1.0 ml min-1, the total time for sample introduction is 1500 s and the total sample volume injected is 4 min A - 17.5 ml. On the other hand, for a single injection using a 17.5 ml loop at the same carrier flow rate, the total time required for sample introduction is more than 1050 s, because of the dispersion of the injected sample zone in the injection loop.It is suggested that the multiple injection method would allow the preconcentration of a large volume as rapidly as a single injection of the same volume of sample, and the injection of the total volume of sample without changing the injection loop. The reproducibility of the multiple injection method for 50 injec- tions was poorer than that of the single injection method at the same absolute amount of cadmium loaded, because of pulsa- tions caused in the carrier flow. Furthermore, in the multiple injection method, the adsorption band of the cadmium-iodo complex is extended in the mini-column by the large volume of carrier delivered. Consequently, an equivalent amount of the complex is eluted in a larger volume, bringing the peak down closer to the detection limit.Hence, the number of injections affects the reproducibility. If necessary, levels of cadmium of several ng ml-l can be determined by the multiple sample injection method. It should be noted, however, that co-existing elements in the sample solution affect the adsorptivity of the cadmium-iodo complex and that it is necessary to optimize the number of injections for each sample composition. Calibration Graphs A calibration graph was obtained by the procedure described under Experimental. The calibration graph was linear over the range 0.05-2.0 pg ml-1 of cadmium, using the single injection method, and from 0.002 to 0.04 pg ml-1 of cadmium, using the multiple sample injection method (50 injections).Equations of the calibration graphs obtained by the least-squares method are: single sample injection: y = (0.1645 k 0.0082) x + (0.0079 k 0.0021) (n = 3); multiple-sample injection: y = (6.1 f 0.0005) x + (0.0086 f 0.0025) (50 injections, n = 3), where y is the maximum peak height absorbance and x is the concentration of cadmium in pg ml-1. The data regarding each calibration were evaluated: the multiple injection calibration had a sensitivity that was almost 37 times larger than that of the single injection calibration. The peak height of a 0.04 pg ml-1 solution injected 50 times using a 350 p1 injection loop was 15% lower than that of a 2.0 pg ml-1 solution injected in one lot using the same injection loop. It is suggested that the adsorption band of the cadmium-iodo complex is extended in the mini-column by the large volume of carrier delivered.The responses that were obtained by using solutions prepared by the addition of various increments of the standard cadmium solution to the real sample solution were equal to those obtained with the standard 0.4 pg Cd rnl-1 h 0.1 pg Cd rnl-’ t- Scan Fig. 3 Sample solution contained 150 g 1-1 Zn. Typical analytical signals obtained with the proposed FI system. The total volume injected / ul 0 3500 7000 10500 14000 17500 21000 0.2 ! I 5: / I The number of sample injections Fig. 4 Number of sample injections versus maximum peak height absorbance. Sample solution contained 20 ng ml-1 Cd. Aliquots of 350 pl were injected at constant intervals of 30 s.Analyst, January 1996, Vol.121 11 cadmium solution. Therefore, it was decided to quantify cadmium by a simple calibration method, and by using the same multiple injection method for the sample solution. Effect of Foreign Ions The influence of foreign metal ions was studied. In the determination of 0.5 pg ml-1 of cadmium by using the single injection method, the following ions when present in the amounts (pg ml-1) shown in parentheses do not interfere: Cu2+ (lo), Fe2+ (lOOO), Fe3+ (loo), Sn2+ (lOOO), Bi3+ (50), and Pb2+ (30). These elements, therefore, do not affect the determination of cadmium in a hydrometallurgical zinc refining process stream. Other elements such as alkali metals and alkaline-earth metals, and species such as nitrate and sulfate do not interfere at all.Analysis of Zinc Electrolyte and River Water The proposed FI method was applied to the determination of cadmium in several samples. The results obtained for the high- salt concentration solutions of zinc refining process streams by the single sample injection method and for the river water collected from Shibakawa river (Omiya, Saitama, Japan) are given in Table 2. The values obtained with modified JIS HlllOl9 and JIS KO10220 methods are also listed. In order to compare the validity of the results obtained in Table 2, the I t I values for the determination of cadmium in the sample solutions were calculated. For sample No. 6, It1 was higher than the critical value (1.96, degrees of freedom = 00 for a 0.05 significance level), which can be interpreted as the probable existence of a systematic error for this determination. For the other determinations the presence of a systematic error was not proven. The reproducibility was satisfactory with a relative standard deviation of less than 5.0% (0.14 pg ml-1 Cd level, n ~~ Table 2 Results of the determination of cadmium in hydrometallurgical zinc electrolyte and river water (pg ml-l) Proposed FI Sample No.method JIS-method* I t I Zinc electrolyte 1 0.12, 0.13 0.14, 0.14 7.36 2 0.24, 0.24 0.24, 0.25 -t 3 0.26, 0.26 0.24, 0.25 -t 4 0.09, 0.08 0.08, 0.10 1.41 5 0.22, 0.24 0.21, 0.22 2.12 6 0.12 f 0.012$ 0.12, 0.13 15.4 River water <0.01§ < 0.01 -t * Zinc electrolyte: JIS-H11 lO;*9 river water: JIS-K0102.20 t Not calculated. * Average of 152 determinations k standard deviation.Sample analysed 5 Sample injected 50 times at regular intervals of 30 s. at regular intervals of 30 min. = 5 ) for the single injection method and 10% (2.0 ng ml-1 Cd level, n = 5) for the multiple sample injection method (50 injections). The detection limits were 0.028 pg ml-1 of cadmium for the single injection method and 0.83 ng ml-1 of cadmium for the multiple sample injection method (50 injections). The absolute amount of cadmium detectable, defined as the analytical signal equal to twice the uncertainty in the background, was 10 ng. The FI system permits a throughput of 6 samples h-1 for the single injection method and 2 samples h-1 for the multiple sample injection method (50 injections). These results indicate that the proposed FI method is suitable for the on-line routine determination of trace amounts of cadmium in the hydrometallurgical zinc refining process stream.The authors gratefully acknowledge Professor K. Oguma of Chiba University for his useful suggestions. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Hughes, D. V., and Nyberg, J. R., in Process Control and Automation in Extractive Metallurgy, ed. Partelpoeg, E. H., and Himmesoete, D. C., The Minerals, Metals & Materials Society, 1988, p. 157. Gomes Neto, J. A., Bergamin Filho, H., Sartini, R. P., and Zagatto, E. A. G., Anal. Chim. Acta, 1995, 306, 343. Xu, S., Sperling, M., and Welz, B., Fresenius’ J. Anal. Chem., 1992, 344, 535. Purohit, R., and Devi, S., Analyst, 1991, 116, 825. Hirata, S., Honda, K., and Kumamaru, T., Anal. Chim. Acta, 1989, 221, 65. Chung-Gin, H., Chao-Sheng, H., and Ji-Hong, J., Talanta, 1980, 27, 676. Fischer, H., and Leopoldi, G., Mikrochim. Acta, 1937. 1, 37. Saltzman, B. E., Anal. Chem., 1953, 25, 493. Bode, H., and Wulff, K., 2. Anal. Chem., 1966, 32, 219. Escriche, J. M., Estelles, M. L., and Reig, F. B., Talanta, 1983, 30, 915. Kallman, S., Oberthin, H., and Lin, R., Anal. Chem., 1958, 30, 1846. Kallman, S., Oberthin, H., and Lin, R., Anal. Chem., 1960, 32, 58. Korkisch, J., and Klakl, E., Talanta, 1969, 16, 377. Strelow, F. W. E., Anal. Chim. Acta, 1978, 97, 87. Marsh, S. F., Alarid, J. E., Hammond, C. F., Meleod, M. J., Roensch, F. R., and Rein, J. E., Los Alamos Scientific Laboratory report LA- 7084, 1978, February. Sayama, Y., Tokuda, M., and Hayashibe, Y., Anal. Sci., 1995, 11, 849. Stability Constants of Metal-Ion Complexes, ed. Sillen, L. G., and Martell, A. E., Chemical Society Special Publication No. 17, Suppl. 1, The Chemical Society, London, 1971. Igarshi, S., Itoh, J., Yotsuyanagi, T., and Aomura, K., Nippon- Kagakukai-shi, 1978, 2, 212. JIS H1110, Method for Determination of Cadmium in Zinc Metal, Japanese Industrial Standard Committee, Tokyo, 1989. JIS K0102, Testing Methods for Industrial Wastewater, Japanese Industrial Standard Committee, Tokyo, 1989. Paper 5/03134F Received 16 May, 1995 Accepted I 8 September, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100007

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

11N Analyst, January 1996, Vol. 121 Courses Date Conference 1996 February 5-6 5-6 5-6 5-6 5-6 6-8 12-14 Location The Next Step in Capillary Gas Chromatography for Trace Analysis Belgium Bruges, Pre- and Postcolumn Techniques in HPLC for Bruges, Improved Analyte Isolation, Derivatization, Clean-up, Separation and Detection Isotopically Labelled Compounds in Hyphenated GC-techniques Analytical Tools for GC-MS (Advanced Modes of Ion-trap Mass Spectrometry) Biomedical Applications of GC-MS HPLC Beginners Training Course Package Testing of Pharmaceuticals March 26 Intermediate HPLC Training Course 27 Intermediate HPLC Training Course 28 Intermediate HPLC 'Raining Course April 23-24 HPLC Troubleshooting Courses May 19-22 1996 International Symposium, Exhibit & Workshops on Preparative Chromatography, Ion Exchange, and AdsorptionLlesorption Processes and Related Techniques 21-23 HPLC Beginners Training Course Belgium Bruges, Belgium Bruges, Belgium Bruges, Belgium Macclesfield, UK Cobham, UK Macclesfield, UK Macclesfield, UK Macclesfield, UK Macclesfield, UK Washington D.C., USA Macclesfield, UK Contact Congress Secretariat, Ordibo bvba, L.Hennincksraat 18, B-26 10 Wilrijk, Antwerpen, Belgium Tel: +32 38 28 89 61. Congress Secretariat, Ordibo bvba, L. Hennickstraat 18, B-2610 Wilrijk, Antwerpen, Belgium Tel: +32 38 28 89 61. Congress Secretariat, Ordibo bvba, L. Henninckstraat 18, B-2610 Wilrijk, Antwerpen, Belgium Tel: + 32 38 28 89 61. Congress Secretariat, Ordibo bvba, L. Henninckstraat 18, B-2610 Wilrijk, Antwerpen, Belgium Tel: +32 38 28 89 61.Congress Secretariat, Ordibo bvba, L. Henninckstraat 18, B-2610 Wilrijk, Antwerpen, Belgium Tel: + 32 38 28 89 61. Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK, SKll 6PJ Tel: 01625 613848. Fax: 01625 616916 Dr. J. A. Clements, Room 403, Royal Pharmaceutical Society of Great Britain, 1 Lambeth High Street, London SE1 7JN Tel: +44 (0)171 735 9145. Fax: +44 (0)171 735 7629 Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK SKll 6PJ Tel: 01625 613848. Fax: 01625 616916 Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK SKll 6PJ Tel: 01625 613848. Fax: 01625 616916 Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK SKll 6PJ Tel: 01625 613848. Fax: 01625 616916 Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK SKll 6PJ Tel: 01625 613848. Fax: 01625 616916 Janet Cunningham, Ban- Enterprises, P.O. Box 279, Walkersville, MD 21793 USA Tel: +1 301 898 3772. Fax: +1 301 898 5596 E-mail: Janetbarr@aol.com Nikki Rathbone, HPLC Technology Ltd, Macclesfield, Cheshire, UK SKI 1 6PJ Tel: 01625 613848. Fax: 01625 616916 Entries in the above listing are included at the discretion of the Editor and are free of charge. If you wish to publicize a forthcoming meeting please send full details to: The Analyst Editorial Office, Thomas Graham House, Science Park, Milton Road, Cambridge, UK CB4 4WF. Tel: +44 (0)1223 420066. Fax: +44 (0) 1223 420247. E-mail:Analyst@RSC.ORG.

ISSN:0003-2654    

DOI:10.1039/AN996210011N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

12N Analyst, January 1996, Vol. 121 I COPIES OF CITED ARTICLES Future Issues Will Include The Royal Society of Chemistry Library can usually supply copies of cited articles. For further details contact: The Library, 1 Royal Society of Chemistry, Burlington House, Piccadilly, London W1V OBN, UK. Tel: +44 (0)171-437 8656. Fax: +44 (0)17 1-287 9798. Telecom Gold 84: BUR2 10. Electronic Mailbox (Internet) LIBRARY@RSC.ORG. If the material is not available from the Society’s Library, the staff will be pleased to advise on its availability from other sources. Please note that copies are not available from the RSC at Thomas Graham House, Cambridge. Electrochemical Decomposition of Cyanides on Tin(rv) Oxide Electrodes in Alkaline Media-A. V. Benedetti, C. S. Fugivara, A. A. Cardoso, P.T. A. Sumodjo Poly(viny1 chloride), Polysulfone and Sulfonated pol yether-Ether Sulfone Composite Membranes for Glucose and Hydrogen Peroxide Perm-selectivity in Amperometric Biosensors-Ian Christie, Y. Benmakroha, M. Desai, Pankaj Vadgama Bis-N-Methylacridinium Nitrate Immobilized on Silica as a Solid-phase Chemiluminescence Reagent-0. A. Zaporozhets, V. V. Sukhan, N. A. Lipkovska Evaluation of Anasorb CMS and Comparison with Tenax TA for the Sampling of Volatile Organic Compounds in Indoor and Outdoor Air by Breakthrough Measurements-Maria P. Baya, Panayotis A. Siskos Dual-detector System for the Shipboard Analysis of Halocarbons in Sea-water and Air for Oceanographic Tracer Studies-Stephen M. Boswell, Denise Smythe-Wright Determination of Trace Levels of Niguldipine in Urine and Blood by Adsorptive Stripping Voltammetry at the Hanging Mercury Drop Electrode-Dagmar Obendorf, Gottfried S tubauer Probabilistic Approach to Confidence Intervals of Linear Calibration-Yuzuru Hayashi, Reiko Matsuda, Russell B.Poe Simultaneous Determination of Trace Amounts of Copper, Nickel and Vanadium in Sea-water by High-performance Liquid Chromatography after Extraction and Back-extraction- Yoshio Shijo, Hidetoshi Sato, Nobuo Uehara, Aratake Sachiko Application of Environmental Analytical Supercritical Fluid Extraction to Compounds of Agricultural Significance. A Review-Iain A. Stuart, John Maclachlan, Arthur McNaughtan Chemiluminescence Determination of Penicillamine via Flow Injection Analysis Applying a Quinine-Cerium(rv) System- 2.D. Zhang, W. R. G. Baeyens, X. R. Zhang, G. Van Der Weken Wavelength-resolved Fluorescence Detection in Capillary Electrophoresis- Jonathan V. Sweedler, Aaron R. Timperman Inverse Scattering Theory of Fourier Transform Infrared Photoacoustic Spectroscopy-J. F. Power Gas-phase Detection of Cocaine by Means of Immunoanalysis-Torsten Ziegler, Oliver Eikenberg, Ursula Bilitewski, Michael Grol Spectrofluorimetric Determination of Trace Amounts of Aluminium with Salicylaldehyde Salicyloylhydrazone- Chongqiu Jiang, Chen Wang, Bo Tang, Xiaogang Zhang Influence of Selected Natural Complexants on the Mobilization/ Purging of Copper From Aqueous Media Into Supercritical Carbon Dioxide-William D. Marshall, Jin Wang What Exactly is Fitness for Purpose in Analytical Measurement?-Michael Thompson, Tom Fearn Oriented Immobilization of Antibodies and Their Applications in Immunoassays and Immunosensors-Bin Lu, Malcolm R.Smyth, Richard O’Kennedy Paraformaldehyde as an End-point Indicator in Hydrolytic Thermometric Titration-Oswaldo E. S. Godinho, Julio Cesar B. Fernandes, Luiz M. Aleixo, Graciliano Oliveira Net0 Sense and Traceability-Michael Thompson Formulation Optimization of Novel Multicomponent Photoprotective Liposomes by Using Response Surface Methodology-Yannis L. Loukas Development and Evaluation of a Chemiluminescent Immunoassay for Chlortoluron Using a Camera Luminometer-D. Stevenson, M. F. Katmeh, W. Aherne Potassium Ion-selective Optodes Based on the Calix[6]arene Hexaester and Application in Human Serum Assay-Wing Hong Chan, Albert W. M. Lee, Daniel W. J. Kwong, Wing Leong Tam, Ke-Min Wang Strategies for Decreasing Ascorbate Interference at Glucose Oxidase Modified Poly(o-pheny1enediamine)-coated Electrodes-Robert D. O’Neill, Karl McAteer ExtractionAtomic Absorption Method for the Determination of the Platinum Group Elements and Gold in Copper-Nickel Ores Using Autoclave Sample Decomposition Technique-V. G. Torgov, M. G. Demidova, T. M. Korda, N. K. Kalish

ISSN:0003-2654    

DOI:10.1039/AN996210012N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. I21 (13-17) 13 Speciation of Nitrogen in Wastewater by Flow Injection A. Cerda, M. T. Oms, R. Forteza and V. Cerda* Department of Chemistry, Universitat de les Illes Balears, 07071 -Palma de Mallorca, Spain A rapid method for the sequential determination of nitrite, nitrate and total nitrogen is proposed. Nitrite was determined directly by using the Griess reaction, which was also used to quantify nitrate after reduction to nitrite with hydrazine. For total nitrogen determination, the nitrogen-containing compounds (organic substances and nitrite and ammonium ions) were oxidized photochemically using a UV lamp and converted into nitrate, which was then reduced to nitrite and determined spectrophotometrically. Under the optimized conditions, up to 220 pmol l-1 N-NO2- and 240 pmol l-1 N-NO3- can be determined, the detection limits being 2 pmol 1-1 N-N02- and 8 pmol l-1 N-NO3-.The relative standard deviation for nitrite and nitrate are 1.5 and 2.3%, respectively. The photo-oxidation method for total nitrogen determination has a linear range of 30-1000 pmol 1-1 N, with a relative standard deviation of 3%. The proposed method was applied to the determination of nitrate, nitrite and total nitrogen in wastewaters. Keywords: Photo-oxidation; flow injection; total nitrogen; nitrate; nitrite; speciation Introduction Many methods have been proposed for the determination of inorganic nitrogen compounds (nitrite, nitrate and ammonium ions) and some of them have been adapted to automated procedures for their application for routine analysis.Usually, the determination of nitrite is based on the Griess reaction: the diazotization of nitrite ion with sulfanilamide in acidic medium, followed by reaction with N-( 1 -naphthyl)- ethylenediamine, gives an azo dye with a maximum absorption at 540 nm. The dye formation can be easily monitored spectrophotometrically . Nitrate can be determined using the same reaction by prior reduction to nitrite. One of the most frequently employed flow injection (FI) methods for the reduction of nitrate to nitrite is that involving a copper-coated cadmium column. *-I2 However, the interference of phos- phatel”14 and the need for frequent recalibration make the method inappropriate for continuous analysis of samples with complex matrices (e.g., wastewaters).Furthermore, the method has to be rejected when using strong oxidants such as persulfate that may destroy the column. Reduction by hydrazine either in a batch1s>l6 or in a contin~ous~~-19 mode has been successfully applied as an alternative to the Cd-Cu method with the advantage that less calibration steps are needed. For the determination of total nitrogen, the organic nitrogen- containing substances are first converted into inorganic com- pounds (usually nitrate or ammonium ion). Different batch methods for the determination of total nitrogen have been developed to overcome the drawbacks of the traditional * To whom correspondence should be addressed. Kjeldahl method (tedious, time-consuming and subject to contamination). In 1969, Koroleff20 developed a method based on the oxidation to nitrate using alkaline persulfate.The high temperature and pressure required were achieved by using an autoclave. The method was applied to the analysis of natural waters and sea-water,21 effluents from sewage sludge plants22 and soil extracts.23 Although this method allowed better recoveries for compounds with N-N and N-0 linkages24 than the Kjeldahl digestion, it was still too slow for monitoring purposes. Replacing the autoclave with a microwave fur- nace25326 substantially shortens the digestion time. One other alternative involved irradiation of the sample with UV light in the presence of an oxidizing reagent (typically hydrogen peroxide or persulfate ion). This approach, proposed by Armstrong and co-workers27~28 for the determination of organic carbon, nitrogen and phosphorus in sea-waters, was later modified and applied by other workers29330 to fresh and natural waters.Segmented31732 and non-~egmented33-~5 flow systems with high sampling frequencies have been described. The nitrogen compounds are photo-oxidized and converted into nitrate on-line. The determination of the resulting nitrate is achieved either by direct detection in the UV region.35 or by reduction with a Cd-Cu column31,33,34 or with the Devarda In this paper an FI system for the sequential determination of nitrite, nitrate and total nitrogen in wastewaters is proposed. The method used for nitrite determination is based on the Griess- Ilosvay reaction and was chosen on the grounds of its high sensitivity and selectivity.The same reaction is used to determine nitrate, following reduction to nitrite ion. The hydrazine method for the reduction of nitrate was chosen because of the drawbacks of the Cd-Cu column reduction method for wastewater analysis, as mentioned above. For total nitrogen determination the photo-oxidation method was chosen. In the presence of persulfate and UV light, organic nitrogen-containing compounds, ammonium and nitrite are converted into nitrate, which is determined as described. The method was applied to the determination of nitrite, nitrate and total nitrogen in raw and treated urban wastewaters. ai10y.32 Experimental Apparatus The FI manifold (Fig. 1) consisted of a Rheodyne (Cotati, CA, USA) 5020 injection valve with a loop of 90 pl, and two Rheodyne 501 1 selector valves (SVA and SVB in Fig.1). By using two peristaltic pumps (Gilson (Worthington, OH, USA) Minipuls Models 2 and 3), the aspiration rate for the sample and oxidizing solution was adjusted separately from that for the other reactants. Once flow rates were optimized, a single pump with appropriate tubing could be used. The UV light source was a Heraeus 15 W mercury lamp, with maximum emission at 254 nm, surrounded by an aluminium reflector and cooled by means of an electric fan. Spectrophotometric measurements were made on a Hewlett-Packard (Avondale, PA, USA) HP 8452 diode-14 Analyst, January 1996, Vol. 121 array detector equipped with a flow-through cell of 18 pl inner volume and 1 cm pathlength. Both the reactors and the injection loop were constructed from PTFE tubing of 0.5 mm id.The dimensions of the reaction coils are shown in Fig. 1. The temperature was maintained constant by means of a thermostated bath incorporated into the FI manifold. The de-bubbler was a T-piece made of methacry- late in which the photo-oxidized sample, entering from the left, was aspirated with a flow rate of 0.36 ml min-1 through the bottom of a conical inner chamber while bubbles and spare flow left for waste through the upper side (Fig. 1). For nitrate and nitrite determination, the sample was passed through a column pached with XAD-7 non-ionic resin (Amber- lite, 480 m3 g-', 80 A, 20-50 mesh) before being injected into the FI manifold. The pre-column (30 X 1 mm id) was inserted into the sample aspiration channel in order to avoid potential interferences from organic matter.The resin was chosen on the basis of the results obtained by Freeman et al.,36 which showed that the non-ionic resins are the most effective in eliminating interference due to organic matter. The XAD-7, a polar poly(methacry1ate) resin, is similar in characteristics to that proposed by these workers. Other manifold designs were tested in order to overcome the problems arising from the high concentrations of persulfate necessary for the oxidation step. An alternative configuration with an additional line of hydrogensulfite, merging with the photo-oxidized sample as proposed by McKelvie et al.,33 was investigated but it did not improve the performance of the system shown in Fig. 1. Reagents All the reagents used were analytical-grade chemicals and included the following.Alkaline persulfate, R 1. This contained 15 g 1- 1 potassium persulfate and 3.5 g 1-1 sodium tetraborate for pH adjustment. The solution was prepared daily by appropriate dilution of a stock containing 40 g 1-1 K2S208 and 35 g 1-I Na2B4-07- 1 OH20. Working reductant, R2. This contained 3 g 1-1 hydrazine sulfate, 0.006 g 1-l CuS04, 1 g 1-1 ZnS04 and 20 g 1-1 NaOH. This reagent is unstable, so it must be prepared daily from a stock consisting of 10 g 1-1 N2H6S04, 1 g 1-1 CuS04, 80 g 1-' NaOH and 10 g 1-1 ZnS04. Chromogenic reagent, R3. This consisted of 20 g of sulfanilamide, 0.5 g of N-(l-naphthy1)ethylenediamine (NED) and 25 ml of concentrated HCl (37%, d = 1.19 g ml-1). Resin Sample 'I mi min-' UV-source Sample R1 Debubbler sVA RC3 0.36 W RC2 1 2 Fig. 1 The FI arrangement for determination of nitrite, nitrate and total nitrogen. R1, persulfate alkaline solution; R2.reducing solution; R3, chromogenic reagent; Resine, Amberlite XAD-7 (480 m3 g-l; 8 nm, 20-50 mesh); UV-source, ultraviolet lamp (15 W, 254 nm); injection volume, 90 yl; SVA, SVB, selection valves; RCl, reaction coil (2 m X 0.5 mm id); RC2, reaction coil (1 m X 0.5 mm id); RC3, photo-oxidation coil (3 m X 0.5 mm id); W, waste; D, detector (540-420 nm). Nitrate and nitrite stock solutions, 0.001 mol 1-I. Prepared from their sodium salts and used to prepare working standard solutions by appropriate dilution. Model nitrogen compounds. In order to evaluate the photo- oxidation efficiency, ammonium chloride, urea, aspartic acid, barbituric acid, nicotinic acid, glutamic acid, EDTA and glycine stock solutions with a nitrogen content of 1 g 1-l each were used.All these reagents were >98% pure. All working standard solutions were prepared by dilution from the stocks. Procedure The proposed FI assembly (Fig. I ) allows the sequential determination of nitrite, nitrate and total nitrogen in the same sample by actuating the switching valves (SVA and SVB) as appropriate. Joint determination of nitrate and nitrite The sample was aspirated (via the pre-column) through port 1 of SVA and mixed with the reductant solution (R2), which was aspirated via port 2 of SVB. The resulting stream was passed through reactor RCl (2 m X 0.5 mm id), which was heated by immersion in a thermostated bath at 40 O C in order to accelerate the reduction. The emerging solution was injected into a distilled water carrier that was then merged with the chromo- genic reagent (R3).The absorbance of the azo dye formed was monitored at 540 nm. Spectral oscillations caused by changes in the refractive index were corrected for by subtracting the absorbance at 420 nm, where the absorbance of the reaction product was virtually zero. Determination of nitrite The nitrite content in the original sample was determined directly by using the Griess reaction under non-reductive conditions. For this purpose, both switching valves were turned to position 1 , so that the hydrazine solution was replaced with distilled water, and the above-described procedure for the joint determination of nitrate and nitrite was repeated.Determination of total nitrogen The oxidation of nitrogen-containing compounds and the determination of total nitrogen were addressed by using SVA in position 2. The sample was mixed with alkaline persulfate (Rl) and propelled by the peristaltic pump to the photoreactor (RC3, a piece of PTFE tubing of 3 m X 0.5 mm id coiled around the UV lamp) where ammonium, nitrite and organic nitrogen- containing compounds were converted into nitrate ion. The de- bubbler at the photoreactor outlet facilitated the sweeping of bubbles, formed during the photo-oxidation process, to waste. Only part of the photo-oxidized sample was introduced into the FI system for the determination of nitrogen content by the above-described procedure for nitrate and nitrite, while the remainder left for waste together with the bubbles formed during photo-oxidation.Also, while SVA was turned to position 1 to determine nitrate and nitrate in the unmineralized sample, the photo-oxidized sample was sent to waste via the de-bubbler. Results and Discussion Photo-oxidation Conditions Preliminary experiments were carried out in order to examine the influence of the irradiation time, oxidant concentration and nitrogen content on the mineralization process by using various model nitrogen-containing compounds.Analyst, January 1996, Vol. 121 15 The models were chosen on the grounds of chemical structure. They included substances frequently occurring in wastewaters (ammonia and urea), straight-chain amino acids (glycine, glutamic acid and aspartic acid), cyclic amino acids (nicotinic acid) and other types of compound usually employed in mineralization studies (EDTA and barbituric acid).Initially, experiments were performed by using solutions containing I mg 1-I N and 4 g 1-1 persulfate, and an irradiation time of 40 s. Under these conditions, conversion was always below 100% (about 90% for glycine, ammonium chloride and urea, and less than 75% for EDTA and nicotinic, barbituric and aspartic acid). Increasing the irradiation time to 80 s resulted in no significant improvement in the photo-oxidation yield. However, if the persulfate concentration was simultaneously raised, per cent. conversions increased substantially. The effect of the two variables was therefore studied simultaneously. Hence the persulfate concentration was varied between 4 and 20 g 1-l at various irradiation times from 40 to 80 s (adjusted by varying the photoreactor feeding flow rate, viz., the summation of those for the sample and persulfate, between 0.4 and 0.9 ml min-1).A persulfate concentration of 6 g 1-1 resulted in 80-100% mineralization at all the irradiation times tested. Increasing irradiation time gave rise to increasing photo-oxidation yield for all the compounds studied except for ammonium chloride, which exhibited the opposite trend probably owing to some ammonium being volatilized (as ammonia) during the digestion process. Irradiation times longer than 50 s resulted in no significant improvement and prolonged the mineralization time unduly.A persulfate concentration of 6 g 1-1 and an irradiation time of 50 s were therefore used in subsequent experiments. The effect of the nitrogen content in the samples was also examined. Solutions of test substances with nitrogen concen- trations ranging between 1 and 14 mg 1-1 N were photo- oxidized and analysed. Recoveries were calculated in each instance and plotted against the nitrogen concentration (Fig. 2). Irradiation of solutions containing concentrations below 7 mg 1-1 N resulted in photo-oxidation yields of 80-100% for all the compounds tested. As the concentration was raised, the mineralization efficiency decreased and was incomplete above 10 mg 1-I. Using the optimized conditions ([K2S208] = 6 g 1-1, t = 50 s), several raw and treated urban wastewaters were analysed; the samples were diluted in order to have a final nitrogen concentration lower than 10 mg 1-1.The nitrogen content determined in the unoxidized sample under reductive conditions I 120.00 \.---.--t 40.00 , 4.00 8.00 12.00 16.00 mg I-' N Fig. 2 Influence of the nitrogen content in the photooxidation yields for several nitrogen-containing compounds. 0, Urea; A, EDTA; +, nicotinic acid; 0, glycine; B, barbituric acid; A , aspartic acid; 0, glutamic acid; and +, ammonium ion. was subtracted from that in the photo-oxidized sample and the result compared with the Kjeldahl nitrogen value. The photo- oxidation results were always smaller than the Kjeldahl nitrogen values, as noted by other workers.30.37 This may be due to the presence of high levels of organic matter which decreases the amount of oxygen available for oxidizing nitrogen-containing compounds, since part of the oxygen is used to form other compounds, e.g., C02.The problem was solved by raising the concentration of oxidizing reagent up to 15 g 1-1. Higher concentrations should be avoided because of the large number of bubbles formed, which interfere with the detection step, while concentrations lower than 12 g 1-1 are not sufficient for quantitative photo-oxidation of the nitrogen compounds occur- ring in wastewaters. The influence of the flow rate ratio of sample (9,) to oxidant (qR1) was finally examined using urea standards and real samples at qR1 : qs ratios from 0.3 to 2. The optimum flow rates were chosen in order to ensure excess of persulfate even when samples with a high nitrogen content had to be analysed.The final ratio was q R 1 : q, = 1.8, adjusted as shown in Fig. 1. In this way the irradiation time was 50 s. Analytical Performance Calibration graphs were obtained and wastewaters analysed under the conditions given in Fig. 1 using a persulfate concentration of 15 g 1-1. The determination of nitrate, nitrite and total nitrogen in the sample entailed constructing four different calibration graphs. The graph for the determination of nitrite [eqn.(A)] was constructed from sodium nitrite standards of various concen- trations under non-reductive conditions. The graph was linear up to 220 pmol I-' N and the resulting detection limit was 2 pmol 1 - 1 N. The calibration equation was: Absorbance (AU) = 2.00 [mmol 1-1 N] - 0.0024 (r2 = 0.9999) (A) The calibration graph for the joint determination of nitrate and nitrite [eqn. (B)] was obtained from sodium nitrate standards that were aspirated through the resin and quantified under reductive conditions.The graph was linear up to 240 pmol l-1 N and the detection limit was 8 pmol 1-1 N. The graph equation was: Absorbance (AU) = 1.23 [mmol 1-1 N] + 0.0018 (r2 = 0.9998) (B) The graph for the determination of nitrite under reductive conditions [eqn. (C)] was constructed from nitrite standards under reductive conditions and was linear over the range 6-217 pmol 1-1 N. The detection limit was 6 pmol 1-l N. The calibration equation was: Absorbance (AU) = 1.20 [mmol 1-1 N] -0.002 (r2 = 0.9998) (C) Finally, the graph for the determination of total nitrogen [(eqn.(D)] was obtained by aspirating sodium nitrate solutions of various concentrations, followed by irradiation and analysis. The graph was linear up to 1000 pmol l-1 N and the detection limit was 30 pmol 1-1 N. The calibration equation was: Absorbance (AU) = 0.43 [mmol 1-l N] + 0.02 (r2 = 0.9993) (D) From the ratio of the slopes of calibration graphs (A) and (C) it follows that the absorbance of a nitrite standard under reductive conditions was 60% of that under non-reductive conditions. From the ratio of calibration graphs (B) and (C) it follows that the percentage conversion of nitrate into nitrite is nearly 100%.16 Analyst, January 1996, Vol. 121 Speciation The nitrite content in the samples was determined by direct interpolation on calibration graph (A) of the signal obtained under non-reductive conditions.The results obtained under reductive conditions allowed the summation of nitrate and nitrite to be performed. If nitrate is to be determined in the original sample, its content can be calculated by subtracting that for nitrite. The analytical determination of nitrate in the sample following photo-oxidation (by means of calibration graph D), provides the total nitrogen content. This includes nitrate and nitrite in the original sample. In order to obtain total Kjeldahl nitrogen (TKN), the nitrogen content in the photo-oxidized sample is subtracted from that in the unoxidized sample analysed under reductive conditions. Reproducibility The reproducibility of the reduction-detection process was determined by using sodium nitrate and nitrite standards.For this purpose, ten successive injections of a 80 pmol l-1 N-NO3- solution were carried out under reductive conditions, and ten injections of 100 pmol 1-1 N-NO2- under non-reductive conditions. The relative standard deviations (s,) thus obtained were 2.3% for nitrate and 1.5% for nitrite. The reproducibility of the photo-oxidation process was studied by aspirating and mineralizing ammonium and urea standards containing 400 pmol 1-1 N. The s, values obtained from ten injections were 3 and 2.7%, respectively. Interferences The interference from foreign ions in the reduction of nitrate to nitrite was examined. No interference due to the following ions and concentrations was observed: Ca2+, 200 mg 1- l; Mg2+, 200 mg 1-1; Fe3+, 100 mg 1-1; Zn*+, 100 mg 1-1; Ni*+, 100 mg 1-1; NH4+, 100 mg 1-1; C1-, 500 mg 1-l; S042-, 500 mg 1-1; C032-, 500 mg 1-l; and Po43-, 500 mg I-'.Concentrations as low as 6 mg 1-1 of Cr2+ and Cr3+ at 10 mg l-l, adversely affected the reduction of nitrate to nitrite. Potential interferences in the Griess reaction are well established and were not re- examined. With respect to the photo-oxidation step, the interference effect on the digestion process from organic matter was also studied. For this purpose different concentrations of carbon in the form of glucose, ranging from 1 to 24 mmol I-' C, were added to solutions of 1 mmol 1-l N-urea. For a molar relationship N/C of 1/20 the signal decrease was approximately lo%, but no interferent effect was observed at lower carbon concentrations.Analysis of Real Samples Samples were analysed directly, with no pre-treatment. For the determination of nitrate and nitrite in the unoxidized sample, the sample was aspirated through a column packed with XAD-7 resin in order to remove organic matter. For the determination of total nitrogen (TN), all samples were mixed with persulfate, irradiated and processed. The high nitrogen content in the wastewater samples required a prior dilution in order to ensure a linear working range.The results obtained after the analysis of several samples collected at the inlet and outlet of different wastewater treatment plants are compared with those obtained by the Kjeldahl digestion method (Table 1).The TN values of 26 samples obtained using the UV- FI method were plotted against the TKN + N-NO2- + N-N03- values. The regression equation thus obtained was: TN (UV-n) = 1.05 TN (TKN + NOz- + Nos-) - 1.00 (n = 26) Several samples with nitrate and nitrite contents below the detection limits were spiked with various amounts of nitrate and/or nitrite. The results obtained are summarized in Table 2. An example of the FI signals obtained using the proposed FI configuration is given in Fig. 3. Conclusions The proposed FI system allows the speciation of nitrogen- containing compounds in wastewaters. It constitutes a rapid choice for the determination of nitrate, nitrite and total nitrogen in the same sample (all three parameters can be quantified in less than 15 min).The mineralization involved is achieved by irradiation with a UV lamp. The digestion time is fairly short (50 s under the conditions used in this work). The photo- Table 1 Results of the analysis of raw (in) and treated (out) wastewater samples (t = 50 s, [K2S208] = 15 g 1-I) Sample 1 in 1 out 2 in 2 out 3 in 3 out 4 in 4 out 5 in 5 out 6 in 6 out 7 in 7 out 8 in 8 out 9 in 9 out 10 in 10 out 11 in 11 out 12 in 12 out 13 in 13 out Total nitrogen TKN* + Photo- N02- + TKN*/ N-N02-/ N-N03-/ oxidation/ NO3-/ mg 1-1 N mg 1 - 1 N mg 1-1 N mg 1-1 N mg 1-1 N 43.4 39.2 50.4 40.6 50.0 42.1 72.6 48.8 58.6 11.6 80.0 4.9 133.0 7.3 80.9 9.8 84.2 11.6 76.9 6.7 89.1 14.4 76.9 5.1 82.4 15.9 ND+ ND ND ND ND ND ND ND ND 2.3 ND ND ND ND ND 50.1 ND 1 .o ND ND ND ND ND ND ND 0.2 * TKN = Total Kjeldahl nitrogen.+ ND = Not detected. ND ND ND ND ND ND ND 3.6 ND 12.8 ND ND ND 18.1 ND 0.3 ND 14.4 ND 14.1 ND 0.2 ND ND ND 22.1 46.1 43.0 46.4 40.5 47.8 41.2 72.2 55.9 65.8 26.8 98.1 2.5 135.6 29.1 82.6 58.2 90.5 29.8 76.3 19.9 101.6 14.5 78.4 75.6 38.1 ND 43.4 39.2 50.4 40.6 50.0 42.1 72.6 52.4 58.6 26.7 80.0 4.9 133.0 25.5 80.9 60.2 84.2 27.0 76.9 20.8 89.1 14.6 76.9 5.1 82.4 38.2 Table 2 Results obtained for wastewater samples spiked with various amounts of nitrate and nitrite Sample 1 2 3 4 5 6 7 8 9 NO2- added/mg 1-1 1.84 3.68 7.36 1 1.04 0 1.84 3.68 5.52 3.68 NO2- found/mg 1-I 1.86 3.61 7.22 10.54 0 I .80 3.71 5.61 3.74 NO3- added/mg I-- I 2.50 5 .OO 10.00 14.88 4.96 2.48 9.92 4.96 7.44 N03- found/mg 1-l 2.64 5.84 10.82 14.23 4.60 2.86 9.10 5.36 7.50Analyst, January 1996, Vol.121 17 0.40 S [ om (s) 0.00 - Q 0.00 I 10.00 ?.a0 30.00 Time/rnn Fig. 3 Analysis of two different wastewater samples. Triplicate injection of: (A) and (D), the unoxidized sample under non-reductive conditions; (B) and (E), the unoxidized sample under reductive conditions; (C) and (F), the photo-oxidized sample under reductive conditions. AU = absorbance units. oxidation yield ranges from 90 to 100% and is comparable to that provided by the Kjeldahl method used as reference. The proposed method can readily be automated; in addition, it is much faster than the classical Kjeldahl method, hence it is suitable for monitoring total nitrogen in wastewater. Nitrate in the photo-oxidized and original sample is quanti- fied after reduction to nitrite.Using hydrazine rather than a copper-coated cadmium column for this purpose allows high concentrations of persulfate to be employed with no appreciable adverse effects on the reduction reaction. In this way, the use of a column that degrades with time and must be replaced or continuously renewed is avoided. The authors acknowledge the financial support of CICyT (Interministerial Council for Research in Science and Tech- nology, Spain) in the framework of Project AMB94-1033, and of the European Community (EC) through Project MATl- CT93-0008. References 1 Wood, E. D., Armstrong, F. A. J., and Richards, F. A., J . Mar. Biol. Assoc. U.K., 1967, 47, 23. 2 Henriksen, A., and Selmer-Olsen, A. R., Analyst, 1970, 95, 514. 3 Lambert, R. S., and Dubois, R.J., Anal. Chem., 1971,43,955. 4 Nydahl, F., Talanta, 1976, 23, 349. 5 Anderson, L., Anal. Chim. Acta, 1979, 110,123. 6 GinC, M. F., Bergamin, H., Zagatto, E. A. G., and Reis, B. F., Anal. Chim. Acta, 1980, 114, 191. 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 Mageson, J. H., Suggs, J. C., and Midgett, M. R. Anal. Chem., 1980, 52, 1955. van Staden, J. F., Anal. Chim. Acta, 1982, 138,403. Koupparis, M. A., Walczak, K. M., and Malmstadt, H. V., Anal. Chim. Acta, 1982, 142, 119. Clinch, J. R., Worsfold, P. J., and Casey, H., Anal. Chim. Acta, 1987, 200, 523. Maim6, J., Cladera, A., Mas, F., Forteza, R., Estela, J. M., and Cerdh, V., Int. J. Environ. Anal. Chem., 1989, 35, 161. McCormack, T., David, A. R. J., Worsfold, P.J., and Howland, R., Anal. Proc., 1994, 31, 81. Vandenabeele, J., Verhaegen, K., Sudrajat, Avnimelech, Y., van Cleemput, O., and Verstraete, W., Environ. Technol., 1990, 11, 1137. Skicko, J. I., and Tawfik, A., Analyst, 1988, 113, 297. Mullin, J. B., and Riley, J. P., Anal. Chim. Acta, 1955, 12, 464. Henriksen, A., Analyst, 1965, 90, 83. Hale, D. R., Znt. Lab., 1980, 10, 79. Madsen, B. C., Anal. Chim. Acta, 1981, 124, 437. Cerdh, A., Oms, M. T., Forteza, R., and Cerda, V., submitted for publication. Koroleff, F., Determination of Total Nitrogen in Natural Waters by Means of Persulphate Oxidation, International Council for the Exploration of the Sea (ICES), Council Meeting 1969, Paper C: 8, revised version, 1970. d’Elia, C. F., Stendler, P. A., and Corwin, N., Limnol. Oceanogr., 1977, 22, 760. Nydahl, F., Water Res., 1978, 12, 1123. Gallardo, A., and Schlesinger, W. H., Soil Biol. Biochem., 1990, 22, 927. Vandenabeele, J., Verhaegen, K., Sudrajat, Avnimelech, Y ., van Cleemput, O., and Verstraete, W., Environ. Technol., 1990, 11, 859. Kingston, H. M., and Jassie, L. B., Anal. Chem., 1986, 58, 2534. Johnes, P. J., and Heathwaite, A. L., Water Res., 1992, 26, 1281. Armstrong, F. A. J., Williams, P. M., and Strickland, J. D. H., Nature (London), 1966,211,481. Armstrong, F. A. J., and Tibbits, S., J . Mar. Biol. Assoc. U.K., 1968, 48, 143. Henriksen, A., Analyst, 1970, 95, 601. Gustafsson, L., Talanta, 1984, 31, 979. Afghan, B. K., Goulden, P. D., and Ryan, J. F., Adv. Autom. Anal., 1970, 2, 29 1. Lowry, J. H., and Mancy, K. H., Water Res., 1978, 12,471. McKelvie, I. D., Mitri, M., Hart, B. T., Hamilton, I. C., and Stuart, A. D., Anal. Chim. Acta, 1994, 293, 155. Kroon, H., Anal. Chim. Acta, 1993, 276, 287. Hinkamp, S., and Schwedt, G., 2. Wasser Abwasser Forsch., 1991, 24, 60. Freeman, P. R., Hart, B. T., and McKelvie, I. D., Anal. Chim. Acta, 1993,282, 379. Cabrera, M. L., and Beare, M. H., Soil Sci. SOC. Am. J., 1993, 57, 1007. Paper 5f03220B Received May 22,1995 Accepted September 19, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100013

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. 121 15N Conference Report Analytical Applications of Biosensors: February 23,1995 The analytical applications of biosensors were examined at a symposium organized by the Joint Pharmaceutical Analysis Group in collaboration with the Royal Society of Chemistry Analytical Division (East Anglia and South East Regions) and the M25 Measurement and Sensors Club on February 23, 1995. Analytical measurements are being performed more and more in situ than in the laboratory. As a result, biosensors are being developed as a subset of chemical sensors using enzymes and antibodies. The new approaches towards electrochemically- based biosensors were introduced by Professor H. A. 0. Hill (University of Oxford). Professor Hill’s team began research with the electrochemistry of the cytochrome-c.Near the surface of the molecule is the heme group which is involved in electron transfer. The development of so-called ‘promoters’ of the electron transfer marked a singular discovery. The detection and measurement of such electron transfer reactions forms the basis of amperometric biosensors. This fundamental research was followed in 1982 by key work involving glucose oxidase, using a ferrocene as the basis of an oxygen-insensitive enzyme electrode. This was developed by Professor Hill’s team and Medisense (Oxford) to form a home-test biosensor. The development of the glucose test kit for clinical use was discussed by Dr. Gordon Sanghera (Medisense). He described a biosensor as being the marriage of a biological component, i.e.a redox protein, enzyme or antibody, with a transduction system using cyclic voltammetry, which allows measurement and (if required) amplification of the signal. The market for blood glucose monitoring is worth more than $800 million worldwide with an estimated 12 million diabetics in the US alone. It was recognized that in the self-monitoring of blood glucose the major source of error came from the user and therefore there was a clear need for simpler and less user-dependent systems. Such systems should achieve less than 10% variability at glucose concentrations of 3040 milligrams per decilitre. Dr. Sanghera suggested, that a biosensor which uses a direct reaction is a simpler device and should have significant advantages over products generally available which use a colorimetric secondary reaction using a dye system.Having developed a commercial test kit for blood glucose, research and development work is continuing to produce similar kits for cholesterol, alcohol, and paracetamol measurement. The subject of the presentation given by Dr. Elizabeth Hall (University of Cambridge) was the use of optics and optrodes for biosensing. Dr. Hall explained how fibre optic technology may be used for biosensors with measurement being made at the interface. The recognition layer, i.e. that layer which is specific for the substance of interest, may be applied directly to the surface of the sensor or alternatively indirectly onto a modulation layer, i.e. silane. Dr. Hall said a small change in the modulation layer may cause a large change in measurement which in turn leads to calibration difficulties.In recent years a polymer deposition process has been developed for the biorecognition molecules, and photolithography allows depo- sition of different layers. Where the biorecognition molecules are antibodies, Dr. Hall explained that the amount of antibody was important for sensitivity of the biosensor and if applied directly onto the surface of the sensor it was difficult to manipulate the concentration and density on the surface. Dr. Hall discussed other uses for optical sensors aside from pharmaceutical analysis and explained how they were suitable for measuring air quality and in particular for the determination of nitrogen oxides. The practical applications of biosensors in the pharmaceut- ical industry was discussed by Dr.Jennifer Hall (Cranfield University). Within the pharmaceutical industry, biosensors have a wide range of uses, which include research and development; process monitoring and control; quality control; and monitoring of the work environment. Within research and development two key areas were of particular importance; the analysis of biomolecular interactions and the monitoring of the physiological effects of a drug, hormone or other ligand. Of the bimolecular interactions which may be investigated, Dr. Hall highlighted antigen-antibody and hormone-receptor inter- actions and a very elegant technology for the monitoring of these, surface plasmon resonance (SPR). This technique detects the changes in the refractive index of the surface layer of a solution in contact with a sensor chip.This sensor chip being a very thin metallic strip over which there is a hydrogel material to which the biomolecules are attached. Changes in refractive index can be caused by a variation of the mass on the surface of the chip due to interactions of biomoIecules. A number of SPR instruments are available from which various types of informa- tion may be obtained including specificity of binding, affinity, kinetics, and relative binding pattern. Applications of the SPR system which have recently been published include the detection and measurement of cyclosporin and peptide ana- logues. The recent work undertaken at Cranfield in the field of DNA liquid crystals was discussed by Dr. Hall. She emphasized the work undertaken by herself and her co-workers, particularly Professor A.P. F. Turner (in collaboration with Professor Y. Yevdokiniov, Institute of Molecular Biology in Moscow). Deoxyribonucleic acid (DNA) liquid crystals are based on the formation of cholesteric liquid-crystalline dispersions of DNA which have anomalous optical activity. Different optical properties are associated with ordered and non-ordered packing and only linear double-stranded DNA undergo ordered packing. This optical activity may then be measured by circular dichroism or polarization microscopy. However, they were working towards a simpler instrument for monitoring the optical properties which could be used as an analytical tool for measuring compounds which may in quantifiable manner, disrupt, these DNA liquid-crystalline dispersions. This was exciting new work which Dr.Hall felt could lead to the development of new biosensors and instrumentation. Moving on to methodology for the monitoring of physiological interactions, Dr. Hall introduced the technology of the light addressable potentiometric sensor (LAPS) device. She ex- plained the LAPS system was based on an insulated semi- conductor device that responds to surface potentials at an electrolyte/solid interface. Based on this technology a com- mercially available microphysiometer had recently been used to measure the extracellular proton flux from isolated gastric gland; the in vitro analyses of ocular irritancy; and the activity of muscarinic receptors. The particular uses of biosensors in process monitoring and control has also been the subject of research at Cranfield.In situ monitoring and automated off-line monitoring, for example in fermentation processes, have been examined. In the area of pharmaceutical quality control the analysis of DNA contamination can cause significant problems.16N Analyst, January 1996, Vol. 121 The use of LAPS technology had been applied in this area and using this technique it had been reported that 2 picograms of DNA could be detected within a couple of hours. Concluding her presentation, Dr. Hall turned to gas- and vapour analysis in the control of the workplace environment. Over the past few years Dr. Hall and her colleagues have been developing biosensors for the direct measurement of gases and vapours such as phenol and sulfur dioxide, which they believe will lead to the development of a range of gas sensors including personal monitors.Professor Brian Birch (University of Luton) introduced the subject of disposable and permanent biosensors based on thick- film technology. About ten years ago thick-film technology was identified as a valuable means of producing biosensors using processes taken from the electronics industry based on precision screen printing and the furnace firing of inks containing metals. Using this technique, a variety of precise, repeatable sensors could be manufactured which could be used particularly for electrochemically based devices. The capillary fill device (CFD) which consists of two parallel plates separated by a small air gap has been used to produce a generic range of disposable sensors.Liquid sample is drawn into the device by capillary action where it reacts with solid materials deposited in the device during manufacture. This technique makes the CFD an ideal device for the rapid and accurate sampling of liquids stated Professor Birch, although he warned that high viscosity liquids and air bubbles can affect the electrode causing false results, unless a thin layer of surfactant is sprayed to both internal surfaces. Sensors are available for the measurement of various parameters using electrochemical techniques such as potentio- metry, controlled potential coulometry and controlled current coulometry. The potentiometric CFD for example, is designed to measure the redox potentials of a solution and may be used to determine the concentration of a species in the sample solution.This type of sensor consists of a working redox electrode usually made of carbon or gold and of a solid reference silver/ silver halide electrode. It is the manufacture of the reference electrode which is an important part of the CFD concept emphasised Professor Birch. They may be manufactured either by electrochemical means or by the chemical oxidation of silver in the presence of chloride ions. In either case electrodes of equally reliable results are produced. The coulometric CFD utilizes the measurement of the charge produced by an electrochemical reaction at the surface of an electrode, where the potential of the electrode is kept constant, or varied in controlled manner. In principle three electrodes are required for coulometric experiments, these being a reference electrode, a working electrode and a counter electrode which used to complete the electrical circuit.In this case it is important for the working electrode to be relatively inert and that it allows only electron transfer between the solution and the electrode. In all types of CFD three of the important kinetic rates in electro- chemical reactions are the rate of the heterogeneous electron- transfer process; the rate of diffusion of the electrode reactants and products; and the rate of any chemical reaction coupled to the electron transfer. However, in the simplest electrode processes and experiments only the diffusional event needs to be considered. Through time, sideways diffusion effects can occur as molecules from the outer solution penetrate further in from the edges of the electrode.To counter this effect (which can distort the results) a guard electrode has been developed which surrounds the working electrode. It is maintained at the same potential as the working electrode although on separate electrical circuit. Since the CFD had been developed it had been used for a variety of measurements including glucose in blood, amino acids as well as nitrate and trace heavy metals. Professor Birch concluded his presentation by considering biosensors in its widest interpretation and discussed the role of permanent sensors in the water industry. This has been undertaken as a joint project between Siemens and the University of South- ampton and has led to the development of an integrated solid- state device for the measurement of temperature, conductivity, pH and dissolved oxygen in the monitoring of drinking and wastewater.Thus far, he said, these sensors had proved to be robust, have remained free from major fouling and have a working life in excess of one month. The success of this project has now led to a second project being undertaken with the approval of the Department of Trade and Industry for the development of sensors to measure chlorine, trace metals and ammonia. Biosensors have proved an attractive area for basic research in recent years and it is only with the recognition of an imperative to generate a viable measurement capability has research shifted to addressing practical and pragmatic issues. This was the belief of Professor Pankaj Vadgama (University of Manchester) in his presentation on the optimization of bio- sensors in the hospital environment. It was important to remember that there existed not only a host of competing technologies, but also high expectations about the de-skilling of analysis. Biosensors clearly provided technological benefits in that they promise simplified construction, ready miniaturization and direct electrical read-out.However, the end-user in hospital requires ease of operation independent of any measurement principle, together with low cost and reliability. Professor Vadgama highlighted the near-unique ability of biosensors to measure optically-opaque samples and a capability for con- tinuous measurement, which can confer a competitive ad- vantage.In the hospital environment the measurement of samples at the bedside is an attractive proposition, which would overcome the problems of sample storage and stability as well as effecting a fast turn-around time. The closer monitoring with biosensors would considerably aid the understanding of bio- chemical variability during chronic disease states which could lead to improved medical treatment. The ability of biosensors to provide real-time continuous monitoring could be of immense value in the acutely ill, diabetic or those receiving intravenous nutrition. Professor Vadgama identified the considerable bene- fits which could arise from the implantation of miniaturized biosensors at selected locations in the body to provide organ specific biochemical information.He believed differences existed between tissues and blood for many metabolites and drugs, and knowledge of local concentrations in specific organs may provide especially valuable insights for tailoring therapy. Although the most successful approach in the development of biosensors had involved electrochemical devices, the major challenge remains one of translating some interesting or novel chemical/transduction concept into a practical system capable of operating in blood, tissue or urine. He suggested that it appeared membranes such as those of microporous poly- carbonate, isopropyl myristate incorporated porous supports, or diamond-like carbon, provided the requisite generic technology for interfacing biosensors. These confer improved biocompati- bility; extended linear range; selectivity against interferent species (e.g., platelets and fibrin); a non-toxic/non-immuno- genic interface; and mechanical stability. In vivo monitoring requires not only the special adaptation of biosensors in terms of miniaturization and biocompatibility, he states, but must also be fail-safe. Concluding his presentation Professor Vadgama stressed that biosensors were a potentially powerful tool for biochemical monitoring in the hospital environment, but the realization of their potential required a close appreciation of the materials required for interfacing needs. The institution of an integrated approach that combines engineering with chemistry in both academic and industrial effort is vital if devices are to be of practical value in hospital medicine. Dr. Peter Harrowing Bristol Royal Infirmary Bristol, UK

ISSN:0003-2654    

DOI:10.1039/AN996210015N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. 121 (19-29) 19 Determination of Methylmercury in Sediments Using Supercritical Fluid Extraction and Gas Chromatography Coupled With Microwave-induced Plasma Atomic Emission Spectrometry HBkan Emteborg", Erland Bjorklundb, Fredrik Odmanc, Lars Karlssonb, Lennart Mathiassonb, Wolfgang Frech" and Douglas C. Baxter" a Department of Analytical Chemistry, Umed University, S-901 87 Umed, Sweden b Department of Analytical Chemistry, University of Lund, S-221 00 Lund, Sweden c Department of Applied Geology, University of Luled, 5-971 87 Luled, Sweden A method employing supercritical fluid extraction (SFE) and GC coupled with microwave-induced plasma atomic emission spectrometry (MIP-AES) is presented for the determination of methylmercury in sediments. Butylmagnesium chloride was used to derivatize the target compound to butylmethylmercury which is amenable to GC.Using a commercially available reference sediment (PACS-1, National Research Council of Canada) as the model sample, a factorial design was utilized to investigate the effect of three variables; density, temperature and flow rate, on the extraction efficiency. An extraction efficiency of 49 k 0.5% could be obtained for a 37.5 min dynamic extraction, corresponding to 25 thimble volumes of supercritical COz, and using purified support sand. Studies on the efficacy of SFE for another sediment matrix as a function of time have also been undertaken. Repeated pressure reductions in combination with support sand were found to increase the extraction efficiency of methylmercury from PACS-1 but not from a sediment issued by the Community Bureau of Reference (BCR) as part of an interlaboratory comparison.For PACS-1 this resulted in an increase in the average extraction efficiency to 96 % for duplicate determinations following 50 thimble volume sweeps. Distillation was used as a reference method for isolation of methylmercury from sediments. Parallel extractions of the BCR sediment, using GC-MIP-AES for the final determination, gave results that were in good agreement and corresponded well with data submitted during the intercomparison exercise. The detection limit for the methylmercury in sediment using the described SFE GC-MIP-AES method is estimated to be 0.1 ng 8-1 based on a 20 p1 injection, 0.5 g of sample and three times the blank level.It is proposed that the co-extracted sulfur from the sediment mediates the transport of methylmercury and, to some extent, inorganic mercury from the sediments. This is supported by the strong correlation between the concentrations of butylmethylmercury and dibutylsulfide found in the toluene extract. Using a stable isotope tracer, 199Hg, and ICP-MS, evidence for the spurious formation of methylmercury during SFE under certain conditions is also presented. Keywords: Supercritical fluid extraction; sediments; methylmercury; sulfur; gas chromatography; microwave-induced plasma atomic emission spectrometry Introduction Even though the use of methylmercury as an agricultural seed dressing or fungicide is prohibited in most countries, this species is still of substantial environmental concern.' This is particularly true since freshwater fish such as plke (Esox lucius) may contain more than 0.3-1.0 mg kg-' of methylmercury on a wet-mass basis and thus exceed current quality objectives.In remote lakes, to which there are no point sources of either inorganic mercury or methylmercury, pikes can be found containing elevated concentrations of methylmercury. The reason for this may be found in long distance transport of various mercury species via the atmosphere273 as well as methylation of inorganic mercury to methylmercury predom- inantly taking place in bottom sediments. 1,- Mercury thereby fulfills one of the criteria for being considered a global pollutant . The input of methylmercury and inorganic mercury to remote aquatic ecosystems stems mainly from precipitation, dry deposition and methylation of inorganic mercury.It has been observed by Hultberg et al.7 that most of the methylmercury found in a defined ecosystem results from precipitation. The other major source of methylmercury is believed to be various methylation processes taking place in bottom sediments and the water column.1.6 In pristine environments, the major source of methylation is of biogenic origin and is effected by various bacteria. The process is assumed to be a carbanion transfer of the methyl group in methylcobalamin to inorganic mercury, thus forming a compound that is 10-100 times more toxic than its precursor.' There are also demethylation processes per- formed by bacteria8 as well as formation of volatile mercury species such as dimethylmercury and elemental merc~ry.~ This means that the amount of methylmercury determined in fresh sediments is a result of the equilibrium between formation and removal processes.However, in freeze-dried and sterilized sediments methylmercury concentrations are very stable. Abio- tic methylation may also occur as a result of transalkylation reactions between organometallic compounds and inorganic mercury in heavily polluted environments.8 Finally, fulvic acids have also been found to have methylating properties, at least at elevated temperatures, as found by Nagase et a1.10 Recently, Alli et al. 11 demonstrated the occurrence of ethylmercury in a natural sediment. However, no explanation was given for the origin of this rarely encountered mercury species.This compound has previously only been detected in a heavily polluted environment outside a factory for organ- omercurial production.12 Nevertheless, there is no doubt that20 Analyst, January 1996, Vol. 121 methylmercury is the predominant organomercurial present in natural sediments. Consequently, this study only considers methylmercury in sediments, although the methodology pre- sented here offers the possibility to determine ethylmercury provided that this species is not destroyed in the extraction step. Numerous papers describing various methods for the extrac- tion of methylmercury from sediments have been published. The most frequently used sample pre-treatment and work-up procedures will be summarized below.In one of these studies,I3 three different extraction procedures were investigated, these involving HC1-leaching, KOH-methanol digestion or distilla- tion of methylmercury chloride. The subsequent determination was achieved by GC with cold vapour-atomic fluorescence spectrometry (CV-AFS) following aqueous phase ethylation of methylmer~ury.~3 According to this study and later work by Horvat et al., l4 HC1-leaching is insufficient for quantitative release of methylmercury from the matrix. Furthermore, certain implications for spiking must be fulfilled since incipient methylmercury was found to be more stable than that added. The released amounts of methylmercury were positively correlated with the amount of total organic carbon (TOC) in the sediments. KOH-methanol digestion gave better results, but some problems were encountered since it was observed that ethylmethylmercury could be formed from inorganic mercury and impurities in the derivatizing agent.The product formed is identical to ethylated methylmercury and therefore may lead to positive bias in the determined concentrations unless suitable caution is exercised. l3 Additionally, severe matrix effects could result if elevated concentrations of sulfide ions were present in the aliquot from the alkaline digestate in a direct ethylation procedure. The distillation of methylmercury chloride follow- ing addition of 10 ml of solution containing KC1 and H2S04 gave the highest recoveries and extraction efficiencies regard- less of the type of sediment investigated.Distillation is time consuming (1.5 h) as the collection rate should not exceed 7-8 ml h-l. Additionally, at most 85-90% of the added reagent volume should be collected to avoid methylmercury decompo- sition. '3915 Considering the above-mentioned facts, the need for simple and easily automated analytical techniques for the isolation of methylmercury from sedimdents is clear. In this study we have investigated the use of supercritical fluid extraction (SFE) for isolation of methylmercury from sediments. In parallel, steam distillation was used as a reference method. The SFE methodology provides fast and reliable extractions and is parsimonious in terms of solvents, labour and chemicals compared with other currently used method.1618 Concerning organometallic compounds, SFE has been used for isolation of organolead19 as well as organotin20-23 from sediments.Methyl-, dimethyl- and mercuric mercury have been transferred from filter papers using various complexing agents together with SFE.24,25 Here, supercritical C02 is used to extract methylmercury from marine sediments, and experi- mental designs26 employed to study three important parameters: density; flow rate; and temperature. Following SFE or distilla- tion, the target analyte is butylated using a Grignard reagent and determined by GC coupled with microwave-induced plasma atomic emission spectrometry (GC-MIP-AES).27.28 Experimental Apparatus The operating conditions for the techniques involving GC have been summarized in Table 1. Additional information on the GC-MIP-AES system may be found elsewhere.27.28 A Hew- lett-Packard (HP) 7680T (Wilmington, DE, USA) unit was used for SFE.Carbon dioxide (4.8 grade, AGA Gas AB, Sundbyberg, Sweden) was used as the extraction medium. Methylmercury was collected on a trap, packed with octadecylsilica (ODs), at a temperature of 5 "C. For cryogenic cooling of different zones in the SFE unit, C02 of lower purity (food technology grade, AGA Gas AB) was employed. Two portions of 1.5 ml of toluene Table 1 Specifications of the gas chromatographs and operating conditions used Injection Gas technique and chromatograph volumes Varian 3300; 1090 Direct injection septum-equipped (on column); programmable 2-20 pl injector Varian 3300; 1075 Split injection, splitlsplitless ratio 1 : 50; injector insert equipped with deactivated glass frit; 0.8-1.6 ~1 Hewlett-Packard Splitless HP 5890 injection (1 min); 1 p1 Column 15 m x 0.53 mm id coated with 1.5 pm DB-1 (J & W Scientific, Rancho Cordova, CA, USA); 5 m X 0.53 mm id retention gap (HP) 30 m X 0.32 mm id coated with 0.25 pm DB-1 (J 8c w> 30 m x 0.25 mm id coated with 0.25 pm HP-SMS Carrier gas and flow rate He; 18 ml min-* N2; 2.2 ml min-1 He; 1 ml min-1 Temperature programme Injector temp.180 "C; initial column temp. 50 "C for 1 min, then ramped at 40 "C min-1 to 180 "C; final hold time 1 rnin Injector temp. 180 "C; initial column temp. 50 "C for 1 min, then ramped at 40 "C min-' to 180 "C; final hold time 2 min 180 "C; initial column temp. 100 "C for 2 min, then ramped at 20 "C min-1 to 300 "C; final hold time 1 min Injector temp.Relays and solvent delays Detection system Pneumatic four- MIP-AES; way valve Hg 253.652 nm (Valco) under GC control; -1 (to vent) 0 -2.25 min; +1 (to MIP) 2.26- 5.25 min and C 247.857 nm or Pb 405.847 nm Solvent delay 3 min Flame ionization detector (FID); air flow 300 ml min-I; N2 make-up flow 25 ml min-1; H2 flow 30 ml min-1 MS; electron impact ionization (30 eV)21 Analyst, January 1996, Vol. 121 (pro analisi quality, Merck, Darmstadt, Germany) were used for elution of the trap. The operating conditions for the SFE step were varied as follows: density, 0.5-0.8 g ml-1; temperature, 40-80°C; flow rate, 1.0-4.0 ml min-1. Since many different extraction times, densities, flow rates and temperatures were used throughout, SFE conditions will be given separately for each group of experiments.The ICP-MS instrument was a PlasmaQuad PQ2+ (VG Elemental, Winsford, UK) modified with a Turbo Interface and new, upgraded PQVision OS/2-based software. The operating and scanning conditions are displayed in Table 2. The peak width in the mass range studied was 0.75 u at 5% peak height. The standard nebulization system was replaced by a Microneb 2000 Direct Injection Nebulizer (DIN)(Cetac Technologies Inc, Omaha, NE, USA). Operating conditions of the DIN may be found in Table 2. The voltages required for the ion lenses were adjusted daily to maximize the count rate at m/z 209. Internal standardization was accomplished by adding an aqueous solution of 203Tl to all blank, standard and sample solutions.Ions were detected using an electron multiplier operated in pulse-counting mode. Samples were weighed using balances having a resolution of 10 I.18. Reagents and Materials The chemicals used were of analytical-reagent grade and in some instances purified further prior to use. A stock standard solution of methylmercury chloride (MeHgC1, 320.5 mg 1-1 Hg) (Merck) was prepared by dissolving the salt in Milli-Q water obtained from Milli-Q equipment (Millipore, Bedford, MA, USA). Inorganic mercury solutions were prepared by diluting a 1000 mg 1-1 certified standard (HgC12, Referens- material AB, Ulricehamn, Sweden) which was used to verify the stability of the MeHgCl stock solution. As complexing agent, sodium diethyldithiocarbamate (DDTC) (99+ ACS grade, Aldrich, Steinheim, Germany) was used.A 0.5 mol 1-l solution was prepared by dissolving the salt in Milli-Q water. A borate buffer of pH 9 (Merck) was used to achieve a pH suitable for extraction. Derivatization was achieved using a 2.0 rnol 1-1 solution of butylmagnesium chloride in tetrahydrofuran (Al- drich). For reduction of inorganic mercury, making the samples amenable for detection of methyl-199Hg using DIN-ICP-MS, a 0.5 mol 1-1 solution of tin(1r) chloride (Merck) in 0.3 mol I-' H2S04 (Merck) was used. For preservation of samples destined for determination of methyl-I99Hg by DIN-ICP-MS, a (1 + 1) solution of distilled HN03 and 0.03 moll-' K2Cr207 was used. This solution was purified over Chelite S (Serva, Heidelberg, Germany), a resin with a high affinity for mercury.An L-cysteine (Sigma, St. Louis, MO, USA) solution was used for back extraction of mercury species from toluene in some experiments. An amount of 1 g of L-cysteine, 0.772 g of sodium acetate (Riedel de Haen, Seelze, Germany) and 12.48 g of sodium sulfate (Merck) were dissolved in 100 ml of Milli-Q water to obtain this solution. For distillation, a 2.7 mol 1-1 potassium chloride (Merck) solution was prepared in Milli-Q water. Sulfuric acid (8 moll-1) was obtained by dilution of concentrated sulfuric acid (Merck) with Milli-Q water. Reagents used for studies of methylmercury formation using DIN-ICP-MS were as described below. Nitric acid was purified by sub-boiling distillation of analytical grade feedstock in a quartz still prior to use. The enriched stable mercury isotope (199Hg) was purchased from Oak Ridge National Laboratory (Oak Ridge, TN, USA).The isotope was received as solid mercury oxide with an isotope enrichment of 91.09%. A stock solution of approximately 100 mg 1-1 was prepared by dissolution of the material in nitric acid and dilution to a final acid content of 0.16 mol 1-l. The concentration of this stock solution was verified by reverse-spike isotope dilution ICP-MS. A commercial standard solution (1000 mg 1-1) of T1 (Spex Plasma Standards, Edison, NJ, USA) was used for the preparation of a 5 mg 1-1 solution of T1 in 0.80 mol 1-1 nitric acid used for spiking the samples for internal standardization. Table 2 ICP-MS operating and scanning conditions for the determination of the isotopic ratio 199HgPHg Operating conditions- Forward r.f.power/W 1250 Coolant gas flow rate/l min-1 13.5" Auxiliary gas flow rate/l min-1 0.85* Nebulizer gas flow rate/l min-1 0.4* Sampling depth Sampling cone Skimmer cone Expansion pressure/kPa 2.5 Range, mJz 196.6-204.4 Number of channels/a.m.u. 20 Time sweep-11s 0.5 Dwell time/ps 320 Total run time/s 200 Number of runs 3 Sample flow-rate/pl min-1 75* Reflected power/W <5 20 mm from load coil, Ni; orifice id 1.00 mm Ni; orifice id 0.75 mm on centre* Analyser pressure/MPa 3.3 Scanning conditions- DIN operating conditions- Nebulizer pressure/MPa 0.64* Gas displacement pump pressure/MPa 175* Flow injection valve Fused silica capillary id/pm 50 * Typical values cited. These parameters were adjusted daily to Carrier solution 2% HN03 Rheodyne No.90 10-068 Tefzel loop size/ml 1 .oo optimize the ion signal. Sediment Samples The CRM PACS-1 was obtained from the National Research Council of Canada (NRCC, Ottawa, Canada). This sediment was collected from the harbour of Esquimalt, BC, Canada. The Community Bureau of Reference (BCR)-sediment was issued by the Joint Research Centre at Ispra, Italy for the European Commissions' Measurements and Testing Programme (for- merly BCR).29 This was carried out for initial screening of analytical methods implemented in various laboratories in Europe for the determination of methyl- and total mercury in sediments. Optimization To optimize the experimental parameters for the SFE of methylmercury from the model sediment sample (PACS- l), response surface methodology26 was employed.For calcula- tions the computer program MODDE v. 1.2 (Umetri AB, Umeii, Sweden) was used, graphical presentation of the optimization data being provided by the computer package SIMCA-S for Windows v. 5.01 (Umetri AB). The models were calculated including linear and interaction terms for the experimental variables only, as quadratic terms were found not to contribute significantly to the quality of the models. Procedures Supercritical fluid extraction Sediment (0.5-1 g) was accurately weighed and placed in the extraction thimbles either directly or after half filling the22 Analyst, January 1996, Vol. 121 volume with support sand. Both a commercially available, acid- washed, thermally pre-treated sea sand obtained from Merck and a natural, unpurified sand available in our laboratory, were employed for this purpose. When using the support sand, the sediment and the sand were mixed using an acid-washed glass rod before the thimble cap was screwed in position.The thimble was shaken manually to effect thorough mixing. Next, the thimble was filled to the mark with sand and installed in the SFE-system. Initially, an unpurified sand was used which resulted in spurious methylmerciry formation and was conse- quently only used for studies of methylmercury formation as described below. In total, about 10 g of sand was required. In some experiments approximately 2 g of fine granular copper metal (Mallinckrodt Speciality Chemicals, Paris, France) was added covering the sediment-sand mixture before the thimble was filled with sand.Following SFE, methylmer- cury was rinsed from the ODS trap using two 1.5 ml portions of toluene collected in two separate septum-equipped vials. The trap was rinsed twice with toluene to eliminate memory effects in the trap. The first vial normally contained 95-100% of the methylmercury. However, in some instances, methylmercury was also detected in the second vial, especially at high levels of methylmercury in the sediments. Immediately after elution, the vials filled with toluene were recapped and put in a refrigerator. As the samples were to be transported for 1-2 d, the stability of a 4 ng ml-1 solution of methylmercury in toluene in tightly capped vials was studied at room temperature. It was found that losses of methylmercury from the vials were negligible after 2 d.Very small mass losses, typically 0.1-0.3%0, could be detected after 2 d, no difference being observed in this respect between vials standing upright or lying down. The vials were, therefore, weighed prior to and after shipment to detect any abnormal mass losses which could give erroneous results. Following transferral of the toluene to centrifuge tubes prior to derivatization (see below), the empty vials were weighed to determine the exact volume of toluene. Distillation For isolation of methylmercury by distillation, an experimental set-up similar to that described by Horvat et al.30 was employed. Sediment (0.5 g) was weighed into a 50 ml glass tube and 5 ml of Milli-Q water was added followed by additions of 0.2 ml of 20% KC1, 0.5 ml of 8 mol 1-1 H2S04 and 4.3 ml of Milli-Q water.The glass tube was placed in a heating block kept at 160 "C and then connected to a 15 ml collection vessel standing in an ice-water-bath. Nitrogen, passing a plug of gold-wool to remove any mercury contamination, was bubbled through the sediment-reagent mixture at a gas flow rate of 60 ml min-1. The distillate was collected at a rate of 6 ml h-1, distillation being terminated after 85-90% of the added reagent volume had been transferred. Following distillation, 1 ml of pH 9 borate buffer (Merck) and 1 ml of 0.5 moll-' of DDTC was added to the distillate. Next, 1 ml of toluene (Burdick and Jackson) was added, the vessel was corked and shaken automatically (Janke and Kunkel Labortechnik, Staufen, Germany) for 10 min.A volume of 0.8 ml of the toluene phase was withdrawn using a micropipette, transferred into a 10 ml screw-capped centrifuge tube and the extracted methylmercury derivatized as described below. Evaluation of Spurious Methylmercury Formation Spurious methylmercury formation was observed using com- binations of the unpurified sand and various sediments. The unpurified sand, the sediments and the commercial sand were separately spiked with 2.5 pg of I99Hg using 1 ml of an ethanol- water (1 + 1) solution. Various combinations were investigated. Either the sediment or the sample support sand was spiked. Following SFE, the toluene phase (1.5 ml) was automatically shaken with 1.5 ml of the L-cysteine solution for extraction of the mercury species. Then, the inorganic mercury was quantita- tively reduced to Hgo by adding 100 pl of the 0.5 moll-' tin(I1) chloride solution. The mercury vapour was purged by bubbling nitrogen through the solution for 5 min at a rate of 15 ml min-1.An aliquot of the resulting aqueous solution was placed in another centrifuge tube where 0.5 ml of pH 9 borate buffer, followed by 0.5 ml of DDTC and 0.5 ml of toluene, was added. The centrifuge tube was then shaken for 5 min. Next, 0.4 ml of the toluene phase was withdrawn and butylated in another centrifuge tube using 0.1 ml of Grignard reagent (for further details see Derivatization section below). This solution was then analysed for methyl- and inorganic mercury by GC-MIP-AES. Only methylmercury could be detected, demonstrating the efficiency of this method for completely separating inorganic mercury, in agreement with previous results3' (although note that some losses of methylmercury are incurred32).The final aqueous extracts containing only methylmercury were pre- served with 2 ml of a strongly oxidizing K2Cr207/HN03 solution contained in 10 ml screw-capped glass vials. Mercury was then determined by DIN-ICP-MS. Prior to analysis using DIN-ICP-MS, 2 ml of blanks, samples and standards were spiked with 20 pl of a 5 mg 1-1 thallium solution as an internal standard. All solutions were subsequently filtered through a 0.45 pm syringe filter (Acrodisc, Gelman Sciences, MI, USA). The sample loop in the injection valve of the DIN-system was manually filled and then switched to inject the sample into the acidic carrier (0.32 mol 1-l HN03), data being acquired by scanning (see Table 2 for scanning par- ameters).A 1 ml sample volume yielded a steady-state signal for at least 10 min. After acquisition of the data, the injection valve was re-routed to the 'load' position, filling the sample loop for the next injection. Peak areas were calculated by the software and all responses were corrected for the blank and related to the internal standard. The isotopes used in the following calculations were l99Hg and 202Hg, the contribution of 202Hg from the enriched isotope spike being negligible. Spiking For assessment of matrix effects inhibiting the SFE of methylmercury from the BCR-sediment, spiking at two levels of methylmercury (three replicates at +30 ng g-l and two replicates at +60 ng g-1) was performed.PACS-1 was spiked with methylmercury at +9 ng g-1. Approximately 0.5 g of sediment was placed in a scintillation vessel and the spike was added. Thereafter Milli-Q water was added so that the final volume was 1 ml. The resulting mixture was blended using a Vortex apparatus (AB Termo-Glas, Gothenburg, Sweden) and then left standing to dry and equilibrate for 2 d. The dry sediment was then placed in an extraction thimble as described above and swept with 10 thimble volumes of C02 using optimum extraction conditions. Sand and sediments spiked with *99Hg were prepared differently. The standard was diluted in a Milli-Q water-ethanol (1 + 1) solution such that 1 .O ml of this solution would give a spike of 5 pg g-1 of 199Hg in the solid.This solution evaporated faster and could be placed in the SFE system after approx- imately 6 h. The BCR-sediment was also spiked prior to distillation by adding appropriate volumes of a 1 mg 1-1 (as Hg) methylmer- cury aqueous standard solution to 0.5 g sample masses wetted with 5 ml Milli-Q water. After an equilibration period in the dark at room temperature with the vessel sealed to prevent evaporation losses, distillation proceeded as described above. For equilibration periods in the range 1-3 h, no significant trends in the recoveries obtained were observed.Analyst, January 1996, Vol. 121 23 Derivatization Toluene containing methylmercury was transferred to 10 ml screw-capped centrifuge tubes standing in an ice-water-bath. Subsequently, using a micropipette, 200 pl (or 100 p1 for the methylmercury formation experiments described above) of the Grignard reagent was added to the toluene phase in the centrifuge tubes and left standing for at least 10 min.To destroy the excess Grignard reagent, 300 pl of 1 mol 1-1 of HC1 was added. The centrifuge tubes were then centrifuged at 5000 rpm (3200g) for 5 min. The resulting organic phase was withdrawn using a Pasteur pipette and emptied into 2 ml screw-capped glass vials. A volume of 2-20 p1 of this solution was injected on the GC-MIP-AES system for determination of methylmercury. The samples were stable for several weeks when stored below - 18 "C. Results and Discussion In order to avoid any misunderstanding of the following discussion, the terminology used is clarified.All concentrations are expressed as the mass of mercury, in the form of methylmercury, per unit mass of sediment. The term extraction efficiency is used to denote the percentage of methylmercury originally present in the sample that is extracted. All extraction efficiencies reported for PACS-1 below are related to the average value of two independent sets of results (8.23 ng g-1) reported by Horvat et al.13 This value is considered to correspond to 100% extraction efficiency. For the BCR- sediment the extraction efficiencies were similarly calculated on the basis of the mean of 11 values (49.4 ng g-1) obtained in the interlaboratory comparison.29 Recovery is defined as the percentage of added methylmercury extracted from a spiked sample. Initial Studies For optimization of the SFE procedure, a factorial design including three parameters was evaluated.The investigated parameters were: density (0.5-0.8 g ml-l); temperature (40-80 "C); and flow rate (1.0-4.0 ml min-1). The sediment PACS-1 was used for all optimization experiments. This is an appropriate model sample since results for the methylmercury content were available from both the distillation procedure in combination with GC-MIP-AES and from the work of Horvat et a1.13 In initial optimization experiments, no sample support was added and the extraction fluid amount was limited to 10 thimble volumes. With these parameters and the extractor operated at 0.8 g ml-l, 80 "C and 4.0 ml min-I, an extraction efficiency of 28% could be achieved at best. It was also found that the magnitude of the flow rate was of little relevance, implying that the extraction was limited by desorption of methylmercury from the matrix.33 Hence, this parameter was set to 4.0 ml min-1 in all following experiments to reduce the extraction time.Increasing the temperature and density were found to have significant, positive effects on the extraction efficiency, the former being the most important, as indicated by the model regression coefficients (not shown). To improve the extraction efficiency, higher temperatures up to 125 "C were evaluated at a density of 0.65 g ml-l. This density was chosen since the instrumentation does not allow pressures in excess of 38 MPa, which would be required in order to obtain densities greater than 0.65 g ml-l at 125 "C.A maximum extraction efficiency of 40% was achieved at 125 "C using 10 sweeps; use of this temperature should be avoided since the polymeric sealing in the thimble caps is slowly degraded eventually leading to leakage. At 80,95 and 110 "C extraction efficiencies of 16,27 and 36% were achieved using a density of 0.65 g ml-1. At 110 "C, the extraction efficiency was not dramatically improved over that at 80 "C using a density of 0.8 g ml-1 (28%). Further experiments were therefore limited to a maximum temperature of 80°C which is preferable for the long term stability of the thimble caps. An attempt was made to use 5% methanol as modifier but it was found that methylmercury was not efficiently retained on the octadecylsilane (ODS) trap at the temperatures required to prevent modifier condensation.Another problem is that the Grignard reagent reacts violently with residual methanol in the toluene eluent. Therefore, the use of modifiers was abandoned in the following work. Organic solvents containing no acidic protons but still exhibiting slightly polar properties such as tetrahydrofuran and thiophene will be investigated as modifiers for SFE of methylmercury from sediments in future work. To check the performance of the trapping procedure, four different trap temperatures were tested (-25, -15, -5 and 5 "Cj. The amount of methylmercury collected was identical within the temperature range investigated. Consequently a trap temperature of 5 "C was chosen to minimize the usage of cryogenic C02. Results discussed under Optimization [also see Fig.4(a)] indicate that the ODS sorbent quantitatively trapped desorbed methylmercury. Spurious Methylmercury Formation During these initial experiments to probe the characteristics of SFE for methylmercury isolation it was observed that, after extraction and apparently independent of flow rate, the sediment had been compacted in the thimble with visible channels through the cake. This would contribute to the low extraction efficiencies obtained since the supercritical C02 will flow through these channels rendering most of the desorbed methylmercury inaccessible. In order to disperse the sediment to improve the contact between the supercritical fluid and the sample, PACS-1 was mixed with a sample support consisting of unpurified sand. This measure, besides the obvious dis- advantage of limiting the maximum sample intake, led to determined methylmercury contents of 25 ng g-l being obtained by SFE GC-MIP-AES, i.e.some three times higher than values reported by Horvat et a1.l3 If Fig. l(a) and (b) are compared it can be seen that the peak for methylmercury in the former is significantly larger. The only difference between these runs is the sample support sand. The chromatogram shown in Fig. l(a) was obtained from a sediment which had been mixed with the unpurified sand prior to SFE, whereas a commercial, acid-washed, thermally treated (puri- fied) sand was used for Fig. l(bj, a blank of the latter also being shown in Fig. l(c). Even though a small blank for methylmer- cury could be detected from the unpurified sand (0.5 ng g-I), this is insufficient to explain the difference in the concentrations determined from the two chromatograms.Consequently, it was hypothesized that methylmercury was actually formed during SFE of PACS-1 when using the unpurified sand as sample support. In order to assesss the hypothesis that inorganic mercury may be methylated in the extraction, a stable isotope spike (199Hg) was added to sand or sediment samples. Blanks were also run to ensure that the formation of methylmercury was not induced by high concentrations of inorganic mercury and the supercritical fluid. These consisted of '99Hg-spiked support sands and gave no detectable excess quantities of methylmercury. In Table 3, the investigated spiked samples are listed, together with the isotope ratios obtained by DIN-ICP-MS.Peak areas were calculated by the software and all responses were corrected for the blank and related to the internal standard. The isotopes used in the following calculations were 199Hg and 202Hg, the contribution of 2*2Hg from the enriched isotope spike being negligible. The natural isotopic ratio of 199Hg/202Hg is 0.5637, although the determined ratio on the day of the measurement was found to be 0.60 f 0.02. The discrepancy can24 Analyst, January 1996, Vol. 121 be explained by different bias factors for the two masses due to variations in ion-transmission and detecti0n.3~ From Fig. 2 it can also be seen that under certain conditions a surplus of methylmercury (l99Hg) compared to the natural 2 3 4 5 2 3 4 5 2 3 4 5 Time/min Fig.1 GC-MIP-AES chromatograms (Hg channel) for derivatized supercritical fluid extracts of (a) 0.5 g PACS-1 + 10 g unpurified sand (apparent methylmercury concentration 28.3 ng g - 1 as determined from the calibration curve); (h) 0.5 g PACS-1 + 10 g purified sand (5.4 ng g-I); and (c) 10 g purified sand blank. Injection volumes were 4 p1. Peaks at 2.7 and 4.1 min correspond to methylmercury and inorganic mercury, respectively. SFE parameters: flow rate, 4 ml min-1; temperature, 80 "C; density, 0.8 g ml-I; 10 thimble volumes swept. isotopic ratio (displayed by the black bars) results. Table 3 demonstrates that significant formation of methylmercury occurs primarily when the sample support sand is spiked in combination with the two sediments investigated.If the sediment itself was spiked no large shifts in the isotopic ratio could be detected as depicted in Fig. 2(b). The mechanism for the formation is not clear but it can be anticipated that large amounts of *99Hg released from the sand collide with methyl groups in humic- and fulvic acids in the sediments at elevated temperatures, leading to methylation in analogous fashion to that described by Nagase et al.1° No highly significant35 shift in the isotopic ratio could be detected when the sediment was spiked, which can be explained in terms of the 1g9Hg being more tenaciously bound to sulfur and humic substances following equilibration, than to the sand matrix. Large amounts of inorganic mercury were detected in the unpurified sand, which may then be methylated as explained above.Sediments from polluted areas may also contain various alkylated tin and lead species. Organotin compounds are extensively used as antifouling agents in paint used on ships, however, often in the form of butyltins. Transalkylation reactions are rapid between free trimethyllead and inorganic mercury.* In order to assess the amount of alkyllead co- extracted from the sediment and the beach sand, the 405.782 nm emission line for lead was monitored simultaneously to the mercury channel using the GC-MIP-AES system. No lead species were detected at the retention times for trimethyllead and dimethyllead. The observed methylation of mercury could therefore not be explained by a transalkylation reaction. For successful SFE of organolead from a sediment matrix, methanol should be added as a modifier as described by Johansson et al.19 This may have implications for the design of SFE-based procedures for isolation of methylmercury, alone or together with other organometallic species,37 from heavily polluted sediment where the above mentioned risk for transalkylation is obvious.Optimization Having established conditions avoiding spurious methylmer- cury formation, further studies of the SFE parameters were undertaken, both for optimization purposes and to elucidate the extraction mechanism. A second factorial design, comprised of only two parameters (density and temperature), is described in Table 4, and the responses displayed in Fig. 3. These experiments were designed to optimize the initial rate of extraction where most methylmer- cury is removed from the matrix (compare Fig.4). The sediment samples were now mixed with purified sand and an extraction fluid amount corresponding to 25 thimble volumes was Table 3 Isotope ratios (199Hg/202Hg)* determined by DIN-ICP-MS for methylmercury isolated by SFE from various combinations of sediment and support sand spiked with an enriched stable isotope solution to yield 5 pg 199Hg g-1 sediment. Note that no signals for methylmercury were detected in the blank, i.e. a combination of purified sand and 199Hg. Sample 1 2 3 4 5 6 7 Sediment PACS-1 PACS-1 + 199Hg PACS-1 + 199Hg BCR BCR + 199Hg BCR + 99Hg BCR + 199Hg support * 99HgP02Hg Beach sand + 199Hg 1.89 f 0.03 Beach sand 0.63 f 0.04 Purified sand 0.69 f 0.02 Beach sand + 199Hg 0.98 k 0.01 Beach sand 0.64 k 0.02 Purified sand 0.65 f 0.01 - 0.74 k 0.15 [-statistic+ 61.97*** 1.16 5.51** 29.44*** 2.45 3.87** 1.60 * Natural isotope ratio is 0.5637, but was determined experimentally on the day of measurement to be 0.60 f 0.02.See text for further discussion. +Calculated t-statistic to test whether the isotopic ratio is significantly ( P = 0.05;**) or highly significantly (P = 0.005;***) different from the experimentally determined 'natural' value.Analyst, January 1996, Vol. I 2 I 25 employed. The best extraction efficiency obtained from this design was 49%. As indicated by the regression coefficients reported in Table 4, increases in both density and temperature improve the extraction efficiency. There is also a significant interaction term between these variables.These data show that release of methylmercury from a sediment matrix is improved at higher temperatures. A thermal desorption mechanism has previously been proposed for the SFE of alkyllead species19 and organic pollutants from certain environmental samples,38 and might also apply in the present case. At higher CO2 densities the 3200 2800 2400 2000 1600 1200 800 400 0 ,$ 720 640 560 480 400 320 240 160 80 0.00 197 198 199 200 201 202 203 204 i 197 198 199 200 201 202 203 204 Mass Fig. 2 ICP-MS spectra from methylmercury isolated from supercritical fluid extracts of (a) 0.5 g PACS-1 + 10 g unpurified sand spiked with 2.5 pg 199Hg (sample 1, Table 3); and (b) 0.5 g PACS-1 spiked with 2.5 pg I99Hg + 10 g purified sand (sample 3, Table 3).Black bars show the natural abundances of Hg isotopes. Peaks at mlz 203 are from the internal standard, 203Tl. At m/z 198 and 204 isobaric interferences from I98Pt and 2WPb reagent impurities are present, respectively. SFE parameters: flow rate, 4.0 ml min-I; temperature, 80°C; density, 0.8 g ml-I; 10 thimble volumes swept. solvating power of the supercritical fluid is enhanced,16 favouring transport of the analyte out of the extraction thimble. Fig. 4 illustrates the determined methylmercury concentra- tions in PACS-1 and the BCR-sediment as a function of the number of thimble volumes swept. Note that these data were obtained by summing the concentrations obtained following sequential extractions of duplicate sediment samples. After each sequence the pressure was reduced to 0.1 MPa before continuing the extraction.For PACS- 1 the cumulative con- centrations are 7.1 ng g-1 (extraction efficiency, 86%) and 7.9 ng g-1(96%) at 25 and 50 sweeps, respectively. The latter value is close to the concentration obtained by other methods13 and shows almost quantitative extraction and trapping of me- thylmercury. As can be seen in Fig. 4(a), for 25 sweeps (preceeded by 3 pressure reductions) the determined concentra- tion and extraction efficiency is significantly higher than those reported for PACS-1 in Tables 4 and 5 using 25 sweeps under the same temperature and density conditions without pressure reductions (4.0 ng g-1, 49%). Pressure reductions cause turbulent flows and some decompaction of the matrix, leading to enhanced release of sample components in the initial stages of the next extraction.The advantage of pressure reduction has previously been demonstrated for systems where stagnant mobile phase mass transfer is important for the overall mass transport process.39 However, this effect was negligible for the BCR-sediment. This might depend on differences in the strength of adsorption to the matrix, but also differences in pore size and distribution between the two sediments. In order to elucidate the apparently complex extraction mechanisms in- volving effects of sample support and pressure reductions in the SFE of methylmercury from various sediment matrices a more detailed study is currently being planned. By comparing the results obtained in the initial studies for PACS-1 using no sample support and no pressure reductions an extraction efficiency of 28% was obtained (10 sweeps).Whereas an extraction efficiency of 75% was obtained using addition of sample support and employing two pressure reductions as can be seen in Fig. 4(a), (10 sweeps). These results clearly demonstrate the importance of pressure reductions and use of sample support for enhanced extraction efficiencies of methyl- mercury from PACS- 1. Spectral Interference Dibutylsulfide As can be seen in Fig. l(a), a large peak can be observed between methyl- and inorganic mercury in the chromatogram. This is a spectral interference occurring when a large amount of some carbon-containing species enters the plasma. The back- ground correction system incorporated in the spectrometer can not correct for very intense broad-band emission, as discussed in previous work.28 Fig.5(a) shows the response from simultaneous monitoring of the carbon channel at 247.857 nm, a huge peak being detected at the same retention time as for the interference observed on the mercury channel. The compound causing this interference was studied using GC-MS under similar chromatographic condi- tions as for the GC-MIP-AES system, and was identified as dibutylsulfide. At a later retention time dibutyldisulfide was also found. It was concluded that these compounds were formed in the reaction between sulfur and the Grignard reagent, in agreement with the findings made by Yong et a1.2I These workers performed an in situ hexylation of organotin com- pounds in sediments using a Grignard reagent resulting in similar spectra interferences as reported here when employing a flame photometric detector. It can be seen from Fig.5(b) and (c)26 Analyst, January 1996, Vol. 121 that no peak for dibutylsulfide is present before the addition of the Grignard reagent. The correlation between the extraction characteristics for sulfur and methylmercury is apparent from the data shown in Figs. 3 and 4. However,. a somewhat lower supercritical fluid density (< 0.5 g ml-1) is optimal for the extraction of sulfur (Table 4). This was not expected and the reason remains unclear. Sulfur is probably extracted from the sediment in an inorganic The similar extraction profiles seen in Fig. 4 indicate that methylmercury is associated with sulfur during the extraction from the sediment.In this way, sulfur could act as a mediator for transport of methylmercury from the sediment. It is also noteworthy that inorganic mercury is found in the toluene extract which could be explained by a similar transport mechanism. The concentrations of inorganic mercury found ( < 5 ng g-l) are far below the certified total mercury content of 4.57 f 0.16 pg g-l present in PACS-1. Thus, the selectivity against inorganic mercury, when extracting methylmercury with supercritical CO2, is very good. The fact that conditions for optimal extraction of methylmercury and sulfur differ (Table 4 and Fig. 3) indicates that there may be at least two different forms of the latter present, both of which form dibutylsulfide following addition of Grignard reagent.However, it should be noted that Louie et al.,40 who studied the SFE of sulfur from coal samples, observed only the elemental form (S8) in GC- MIP-AES chromatographs obtained without any derivatization step. The concentration of sulfur in PACS-1 is certified as 1.32 f 0.08%, which would normally cause plugging problems when using SFE instrumentation equipped with capillary-type re- strictors.41 Frequently, granular elemental copper is added to the sediment to avoid this serious pr0blem.~9?~1 The in- strumentation employed here incorporates a variable restrictor which could cope with the high sulfur content of PACS-1. Nevertheless, addition of copper was evaluated and found to reduce the quantity of sulfur extracted and completely suppress methylmercury extraction from sediment.This observation provides further support for the postulation that sulfur mediates the SFE of methylmercury. Interestingly, Johansson et a1.19 did not observe any degradation in the SFE of alkyllead species following addition of copper powder, indicating that the extraction mechanism depends on the nature of the target analyte and its interaction with the matrix. Much less sulfur was extracted from the BCR-sediment as the peak for dibutylsulfide was smaller. No data is, however, available concerning how much sulfur this sediment contains. It has been found by Craig and Moreton4* that sulfur-rich sediments have a much lower ability for methylation due to the fact that inorganic mercury is tightly bound to sulfur and thereby not available for biomethylation.It may paradoxically be so that the SFE of methylmercury from sediments is favoured by a high content of sulfur provided that the restrictor does not clog, although less methylmercury will be formed in the environment in such a sediment. This can be verified if the extraction curves in Fig. 4 are compared. As can be seen in Fig. 4(a), 7.9 ng g-l of methylmercury is extracted from PACS-1, which is close to values reported by Horvat et al.13 and obtained in this work using distillation, i.e., close to 100% extraction efficiency. From the BCR-sediment, Fig. 4(b), approximately 50% (24.2 ng g-l) of the methylmercury is extracted under the same conditions. However, as pointed out by Horvat et al.,13 the amount of humic substances are also of fundamental importance for the availability of methylmercury, but unfortunately no information is available on the total organic carbon (TOC) content in the BCR sample.It must be concluded that it is very important that sediment reference materials for methylmercury are characterized with respect to sulfur, TOC and possibly minerals as well. Otherwise it will be difficult to know how well a reference material matches the matrix present in unknown sediment samples. Performance of the SFE GC-MIP-AES Method The results obtained for the recovery tests are reported in Table 5. It can be seen that for the SFE of the spiked BCR-sediment a recovery of 50.4 f 3.8% is found. The extraction efficiencies for native methylmercury and the recoveries for spiked sediments are very much the same.This suggests that, at least for the BCR- sediment, native methylmercury might not be significantly more strongly bound to the matrix. For the BCR-sediment, it is Table 4 Design and results for the optimization of SFE conditions using the model sediment PACS-1 (average methylmercury concentration 8.23 ng g-l for two independent sets of results from Horvat et ~21.~3) Design and results- Dibutylsulfide peak area Methylmercury/ng g-' (arbitrary units) Density/g ml-l TemperaturePC (coded level) (coded level) Observed Predicted Observed Predicted 0.50 (- 1) 0.50 (- 1) 0.80 (+1) 0.80 (+1) 0.50 (-1) 0.50 (-1) 0.80 (+1) 0.80 (+I) 0.65 (0) 0.65 (0) 40 (-1) 40 (-1) 40 (-1) 40 (-1) 80 (+1) 80 (+1) 80 (+1) 80 (+1) 60 (0) 60 (0) 0.94 0.79 1.10 1.18 2.32 2.55 4.04 3.98 2.00 1.87 0.83 0.83 1.10 1.10 2.40 2.40 3.97 3.97 2.08 2.08 0.42 0.42 1.50 1.80 1.94 2.17 1.93 1.36 1.34 1.09 0.37 0.37 1.60 1.60 2.01 2.01 1.60 1.60 1.40 1.40 Model parameters- Methylmercury Dibuty lsulfide Regression Regression coefficient Probability coefficient Probability Variable (coded) level (%) (coded) level (%) Constant 2.077 > 99.9 1.397 > 99.9 Density (d) 0.4625 > 99.9 0.205 95.0 T X d 0.325 > 99.9 -0.41 99.7 Temperature (T) 1.11 > 99.9 0.4075 99.7Analyst, January 1996, Vol.121 27 therefore possible to use spiked sediments to evaluate the extraction procedure using the SFE-parameters given in Table 5. The recovery data for the distillation are much higher and consistent with values reported by other workers.13~30 However, for the sulfur-rich sediment reference material PACS-1, the cumulative extraction efficiency at 50 sweeps is 96%, higher than the value obtained using distillation.As the BCR-sediment sample was part of an extensive laboratory intercomparison exercise, the performance of the SFE method could be critically evaluated. It is clear that this methodology can provide accurate results as shown in Table 6 . However, recovery tests should be performed or standard additions used for calibration. One negative feature of the SFE 10.00 1 2-oo 1 a 0.00 J O.OE+O I I l l Thimble volumes swept 0 10 20 30 40 50 60 0 A 300000 200000 100000 0 h c v) C 3 .- 0 20 40 60 Thimble volumes swept Fig. 4 Cumulative methylmercury concentration, detected by GC-MIP- AES (e) and extractable sulfur, detected by GC-FID (A) expressed as summed peak areas from detected dibutylsulfide derived from sequential SFEs of ( a ) PACS-1, and (b) the BCR-sediment as a function of the total number of thimble volumes of COz swept.SFE parameters: flow rate, 4.0 ml min-I; temperature, 80°C; density, 0.8 ml min-I; support, purified Fig. 3 Response surface models displaying the effects of temperature and COz density on (a) the determined concentration of methylmercury; and (6) the peak area for dibutylsulfide (arbitrary units) obtained from the SFE of PACS-1. SFE parameters: flow rate, 4.0 ml min-I; 25 thimble volumes swept; support, purified sand. sand. Table 5 Summary of the calculated extraction efficiencies and recoveries obtained in this work for PACS-1 and the BCR sediment using SFE and distillation. Error terms represent one standard deviation of the mean for the number of replicates given in parenthesis.For all extractions the flow was 4.0 ml min-1 Extraction efficiency (%) Recovery* (%) Experimental conditions- PACS- 1' BCR PACS- 1 BCR SFE: 80 "C, 0.8 g ml-l, SFE: 125 "C, 0.65 g ml-l, SFE: 80 "C, 0.8 g ml-I, - - 10 sweeps, no support 28.5 k 1.1 ( 5 ) - - - 10 sweeps, no support 38.8 _+ 3.2 (3) - 25 sweeps, with support, no pressure reductions 48.7 k 0.5 (3) 47.6 f 8.0* (5) 65.0 f 11 (3) 50.4 * 3.8 ( 5 ) Z 50 sweeps,$ with support, SFE: 80 "C, 0.8 g ml-l, 4 pressure reductions 93.9; 98.2 52.6; 46.9 - - Distillation 80.4 k 8.7 (3) 68.9 k 4.7 (5) 80.7 k 6.9 (4) 78.0 k 7.3 (5) * Calculated from spiked sediment samples.t Calculated on the basis of determined methylmercury concentrations in final extracts and reported values * Calculated from the mean value submitted by 11 independent laboratories.29 reported by Horvat et al. given at the beginning of the Results and Discussion. Based on cumulative methylmercury concentrations following sequential SFE, data from Fig. 4.28 Analyst, January 1996, Vol. 121 GC-MIP-AES results is the fairly large s, value compared with other methods. With a longer sweep time the extraction efficiency can be increased which should improve the un- certainty. As pointed out in an earlier section, there is a difference in the extraction efficiency between the BCR- sediment and PACS- 1, which is attributable to differences in the t ~ E! LL 0 z LL i I I I I I 1 2 3 4 5 0 1 2 3 4 5 Time/min I I 2 3 4 5 Time/min Fig.5 Chromatograms for supercritical fluid extracts of 0.5 g PACS-1 + 10 g purified sand obtained using (a) the carbon channel of the GC-MIP- AES system, and (b) and (c) GC-FID instrumentation. In (a) and (b) the extracts were derivatized, but were not in (c). Arrow in (b) identifies the dibutylsulfide peak. SFE parameters: flow rate, 4.0 ml min-1; temperature 80 "C; density, 0.8 g ml-1; 25 thimble volumes swept. sulfur contents, even when employing 50 sweeps during the SFE. As can be seen in Table 6, the methylmercury concentra- tion determined for PACS-1 is in good agreement with values reported by Horvat et al.13 and that obtained by distillation in combination with GC-MIP-AES. It should be stressed that the SFE method is not general in the respect that extraction efficiencies may vary for various sediments and that character- ization of the sample is very important for the understanding of various matrix effects.The LOD for methylmercury in sediment using the developed method is estimated to be 0.1 ng g-l (three times the blank level), based on a 20 pl injection into the GC-MIP-AES instrument, a sample intake of 0.5 g for the SFE and elution of the analyte from the ODS trap into 1.5 ml toluene. It is further assumed that the extraction efficiency is 50%, as was the case for the BCR-sediment (Table 5 ) . Conclusions The developed SFE GC-MIP-AES procedure provides a fast and accurate means of quantifying methylmercury in sediments. The major advantages lies in the high level of automation, low consumption of chemicals and rapidity compared to other currently used methods.One particularly attractive feature is that the final toluene eluate from the SFE system is directly amenable to Grignard derivatization, reducing the number of steps, and thus potential sources of error, in the analytical method. All these combined positive aspects would make this methodology particularly useful when large numbers of sedi- ments samples are to be analysed for methylmercury, for example in environmental monitoring programmes. Never- theless, a number of potential limitations have been identified which must be given due consideration. Matrix effects must be investigated prior to applying SFE to sediments having varying concentrations of sulfur and, perhaps, organic carbon due to pronounced differences in the extraction efficiency as a function of sulfur content.Care must be taken to ensure that the SFE system is free from inorganic mercury contamination (in the support sand or as residues from previous Table 6 Results for the determination of methylmercury in various sediments by SFE or distillation in combination with GC-MIP-AES and comparison with literature data. Error terms represent one standard deviation of the mean for the number of replicates given in parenthesis. All values reported here have been corrected for recovery as reported in Table 5 unless otherwise stated. Methylmercury concentrationhg g- 1 Method PACS- 1 BCR SFE GC-MIP-AES 7.9 f 0.3 (2)* 46.8 f 6.9 (6) Distillation GC- MIP-AES 8.2 f 1.8 (4) 40.5 k 3.4 (7) Distillation GC- CV-AFSt 8.47 f 0.63 (8) - Alkaline digestion GC-CV-AFSt 7.99 & 0.42 (3) - Interlaboratory comparison* - 49.4 * 7.9 (1 1) Total mercury 4570 f 1605 91000 f 11000~ * SFE parameters: flow, 4.0 ml min-1; temperature, 80 "C; density, 0.8 g ml-I; support, purified sand; 50 thimble volumes swept.Not corrected for recovery, values obtained from Fig. 4(a). f Results for laboratory 1, from Horvat et ~ 1 . ~ 3 * Mean of combined means from 1 1 laboratories that reported acceptable 5 Certified value k 95% confidence interval. 1 Mean of combined means from 13 laboratories that reported acceptable results.29 results.29Analyst, January 1996, Vol. 121 29 samples in the extraction thimbles) since spurious formation of methylmercury has been observed under certain conditions.The amount of sulfur in the sediments may be both beneficial and detrimental. Methylmercury is probably more easily extracted from sediments rich in sulfur as a result of its transport mediating behaviour. However, it is well known that sulfur may lead to serious clogging of capillaries if used as restrict or^.^^ The detection limit of 0.1 ng g-1 is sufficiently low for the determination of methylmercury in most natural sediments even though lower concentrations may occasionally be encoun- tered. 1 7 1 3 As the results indicate that improved extraction efficiencies are obtained by increasing both the temperature and super- critical fluid density, future work will be performed to further study the effects of these parameters and a more detailed study of the extraction mechanisms involving pressure reductions and the effect of sample support.This work was supported by the Swedish Environmental Protection Board and the Centre for Environmental Research in UmeA. Dr. C. Pontkr, Svensk Grundamnesanalys AB, LuleA, Sweden, is thanked for permission to perform the DIN-ICP-MS measurements. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Craig, P. J., and Brinkman, F. E., in Organometallic Compounds in the Environment, ed. Craig, P. J., Longman, London, 1986, ch. 2, pp. Slmer, F., Schuster, G., and Seiler, W., J. Atmos. Chern., 1985, 57, 2638. Lindqvist, O., Johansson, K., Aastrup, M., Bringmark, L., Hovsenius, G., HAkanson, L., Iverfeldt, A., Meili, M., and Timm, B., Water Air Soil Pollut., 1991, 28, 66.Jensen, S., and Jernelov, A., Nature (London), 1969, 223, 753. Xun, L., Campbell, N. E. R., and Rudd, J. W. M., Can. J. Fish. Aquat. Sci., 1987, 44, 750. Kudo, A., and Mortimer, D. C., Environ. Pollut., 1979, 19, 239. Hultberg, H., Iverfeldt, A., and Lee, Y.-H., in Mercury Pollution: Integration and Synthesis, ed. Watras, C. J., and Huckabee, J. W., Lewis, Boca Raton, FL, 1994, pp. 313-321. Ebinghaus, R., and Wiken, R. D., Appl. Organomet. Chem., 1993,7, 127. Amyot, M., Mierle, G., Lean, D. R. S., and McQueen, D. J., Environ. Sci. Technol., 1994, 28, 2366. Nagase, H., Ose, Y., Sato, T., and Ishikawa, T., Sci. Total. Environ., 1982, 24, 133. Alli, A., Jaffe, R., and Jones, R., J. High Resolut. Chromatogr., 1994, 17, 745.Hintelmann, H., and Wiken, R.-D., Appl. Organomet. Chem., 1993, 7, 173. Horvat, M., Bloom, N. S., and Liang, L.,Anal. Chim. Acta, 1993,281, 135. Horvat, M., MandiC, V., Liang, L., Bloom, N. S., Padberg, S., Lee, Y- H., Hintelmann, H., and Benoit, J., Appl. Organomet. Chem., 1994,8, 553. 65-1 10. 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Lee, Y.-H., Munthe, J., and Iverfeldt, A., Appl. Organornet. Chern., 1994, 8, 659. Analytical Supercritical Fluid Chromatography and Extraction, ed. Lee, M. L., and Markides, K. E., Chromatography Conferences, Provo, UT, 1990. Barnabas, I. J., Dean, J. R., and Owen, S. P., Analyst, 1994, 119, 2381. Chester, T. L., Pinkston, J. D., and Raynie, D. E., Anal. Chem., 1994, 66, 106R. Johansson, M., Berglof, T., Baxter, D. C., and Frech, W., Analyst, 1995,120,755. Dachs, J., Alzaga, R., Bayona, J. M., and Quevauviller, Ph., Anal. Chim. Acta, 1994, 286, 319. Yong, C., Alzaga, R., and Bayona, J. M., Anal. Chem., 1994, 66, 1161. Liu, Y., Lopez-Avila, V., Alcaraz, M., and Becket, W. F., Anal. Chem., 1994,66, 3788. Chau, Y. K., Yang, F., and Brown, M., Anal. Chim. Acta, 1995,304, 85. Wai, C. M., Lin, Y., Brauer, R., Wang, S., and Becket, W. F., Talanta, 1993,40, 1325. Wang, S., Elshani, S., and C. M., Wai, Anal. Chem., 1995, 67, 919. Box, G. E. P., and Draper, N. R., Empirical Model-building and Response Surfaces, Wiley, New York, 1987. Bulska, E., Baxter, D. C., and Frech, W., Anal. Chim. Acta, 1991,249, 545. Emteborg, H., Baxter, D. C., and Frech, W., Analyst, 1993, 118, 1007. Quevauviller, Ph., Fortunati, G. U., Filipelli, M., Baldi, F., Bianchi, M., and Muntau, H., Appl. Organomet. Chem., 1995, submitted. Horvat, M., May, K., Stoeppler, M., and Byrne, A. R., Appl. Organomet. Chem., 1988, 2, 515. Magos, L., Analyst, 1971, 96, 847. Lind, B., Holmgren, E., Friberg, L., and Vahter, M., Fresenius' J. Anal. Chem., 1994,348, 815. Langenfeld, J. J., Hawthorne, S. B., Miller, D. J., and Pawliszyn, J., Anal. Chem., 1995, 67, 1727. Longerich, H. P., Fryer, B. J., and Strong, D. F., Spectrochim. Acta, Part B, 1987,42B, 39. Statistical Methods in Research and Production, eds., Davies, 0. L., and Goldsmith, P. L., Longman, London, 4th edn., 1980. Environmental Analysis Using Chromatography Inte$aced With Atomic Spectroscopy, ed. Harrison, R. M., and Rapsomanikis, S., Ellis Horwood, Chichester, 1989. Liu, Y., Lopez-Avila, V., Alcarez, M., and Beckert, W. F., J. High Resolut. Chrornatogr., 1993, 16, 106. Langenfeld, J. J., Hawthorne, S. B., Miller, D. J., and Pawliszyn, J., Anal. Chem., 1993,65, 338. Bjorklund, E., Turner, C., Karlsson, L., Mathiasson, L., Sivik, B., and Skogsmo, J., J. Supercrit. Fluids, submitted for publication. Louie, P. K. K., Timpe, R. C., Hawthorne, S. B., and Miller, D. J., Fuel, 1993, 72, 225. Pyle, S. M., and Setty, M. M., Talanta, 1991, 38, 1125. Craig, P. J., and Moreton, P. A., Mar. Pollut. Bull., 1983, 14, 408. Paper 510541 6H Received August 14, I995 Accepted September 18,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100019

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, January 1996, Vol. 121 (31-35) 31 Determination of Ultra-trace Amounts of Selenium(iv) by Flow Injection Hydride Generation Atomic Absorption Spectrometry with On-line Preconcentration by Co-precipitation with Lanthanum Hydroxide Part II. On-line Addition of Co-precipitating Agent Steffen Nielsen, Jens J. Sloth and Elo H. Hansen Chemistry Department A, Technical University of Denmark, Building 207, DK-2800 Lyngby, Denmark A flow injection procedure for the determination of ultra-trace amounts of selenium(rv) is described, which combines hydride generation atomic absorption spectrometry (HGAAS) with on-line preconcentration of the analyte by co-precipitation-dissolution in a filterless knotted Microline reactor. Based on a previously published procedure that requires the off-line premixing of sample and co-precipitating agent, the present approach facilitates on-line addition of the co-precipitant to the time-based aspirated sample.The sample and the coprecipitating agent (lanthanum nitrate) are mixed on-line and merged with an ammonium buffer solution of pH 9.1, which promotes precipitation and quantitative collection on the inner walls of an incorporated knotted Microline reactor. The SeIV preconcentrated by coprecipitation with the generated lanthanum hydroxide precipitate is subsequently eluted with hydrochloric acid, allowing an ensuing determination via hydride generation. At different sample flow rates, i.e., 4.8, 6.4 and 8.8 ml min-1, enrichment factors of 30, 40 and 46, respectively, were obtained at a sampling frequency of 33 samples h-l.The detection limit (3s) was 0.005 pg 1-1 at a sample flow rate of 6.4 ml min-1 and the precision (relative standard deviation) was 0.5% (n = 11) at the 0.1 pg 1-1 level. Keywords: On-line co-precipitation-preconcentration; co-precipitation with lanthanum hydroxide; on-line addition of the co-precipitant; flow injection hydride generation atomic absorption spectrometry; selenium(rv) assay Introduction Flow injection (FI) on-line pre-concentration procedures by collection and dissolution of precipitates were first described by Jimenez et al. in 1987 for the indirect determination of anions.',* In these experiments a stainless-steel filter was used for the collection of the precipitates. Based on this conceptual idea, the authors proceeded with the pre-concentration of organic constituents,36 and subsequently with the preconcen- tration of trace elements,7-10 primarily employing flame atomic absorption spectrometry (AAS) as the means of detection.However, the use of an incorporated filtering sytem may not only limit the efficiency of the on-line collection of the precipitate formed, but it might also, to some extent, influence the ensuing on-line dissolution process, because the presence of the filtering device can give rise to the generation of considerable back-pressure. As a consequence, the sample loading is reduced which, in turn, affects the enrichment factor (EF)" of the system. Besides, the filtering device often detracts from the reproducibility of the determination. To achieve an effective on-line preconcentration an approach is therefore needed that promotes effective collection of the precipitate formed, allows its ensuing instantaneous dissolution and obviates the problems associated with generation of unwanted back-pressure.In 1991 Fang et a1.12 introduced a simple and most elegant solution to this problem, demonstrating that the collection of precipitate could effectively be achieved on the inner walls of a tube provided that it was tightly knotted. This device, the so-called knotted reactor (KR), which thus eliminates the need for the use of a filter, has proved to be very effective, providing optimum conditions for on-line precipita- tion-dissolution manipulations. 1 1 The applicability of the approach has been demonstrated in a series of papers by Fang and co-workers,12-14 Welz et al.15 and Min and Hansen16 for the determination of low levels of various metal ionic species by means of co-precipitation in an FI-AAS system.In all instances, the on-line collection was effected by means of an organic co-precipitating agent, which in turn required the use of an organic solvent for the ensuing dissolution procedure. For obvious reasons, it would be preferable if an inorganic solvent, such as a mineral acid, could be employed, because many organic solvents exhibit toxic or unpleasant properties. Although one can take advantage of the fact that the FI system is a closed one, the necessity of using organic solvents entails additional problems. Thus, these chemicals cannot be handled by most pump tubes and therefore require the use of displacement facilities, which makes the system more complicated.A system based on employing co-precipitation and ensuing dissolution in an inorganic solvent was recently described by Tao and Hansen.17 Used for the determination of SeIV, the authors employed an FI manifold combining hydride generation and atomic absorption spectrometry (HGAAS) with time based sampling, where on-line preconcentration of Se"' was effected by co-precipitation with La1'' at a pH around 9.1. The dissolution procedure was made by hydrochloric acid (Fig. I), which is the medium for the hydride generation process. In the procedure, the co-precipitant was premixed off-line with the sample, which, of course, is a severe practical limitation of the approach, and it would therefore be of interest to explore the possibilities of accomplishing this procedure on-line.This is not only because the sample manipulations would be significantly reduced in the actual assay, but also, and most importantly,32 Analyst, January 1996, Vol. 121 because such an investigation would reveal if this approach had general applicability. The present methodological study is thus devoted to these matters, using as its basis the approach of Tao and Hansenl7 (Fig. 1) but extending it to the determination of trace-level amounts of Sew via the on-line addition of co-precipitating agent. Because Tao and Hansen demonstrated that the La co- precipitation procedure as such was successful for practical assays of a number of 'real samples', the present procedure was for the same reason confined to aqueous standards, and hence it was found redundant to apply it to other types of sample.However, as it turned out, and as one intuitively might expect by considering the dynamic conditions under which the precipitat- ing reaction takes place, this investigation proved to be anything but a trivial exercise. Experimental Apparatus A Perkin-Elmer (Norwalk, CT, USA) Model 2100 atomic absorption spectrometer was used in combination with a Perkin- Elmer Model FIAS-200 flow injection unit (equipped with two individually controlled peristaltic pumps and a five-port FI- valve), with a hydride generation accessory (the gas-liquid separator used in the chemifold was a Perkin-Elmer W- configuration unit). A selenium hollow cathode lamp (S.& J. Juniper, Harlow, Essex, UK) was used at a wavelength of 196.1 nm with a spectral bandpass of 2.0 nm, and was operated at 7 mA. The temperature of the quartz atomizer cell was set at 9OO0C, and the argon carrier flow rate was fixed at 100 ml min-l. The output signals were processed with a time constant of 0.5 s in the peak-height mode and recordings from the graphics screen were printed out by an Epson Model FX-850 printer. The acutation times of the injector valve and the two pumps were programmed with the use of the FI software of the Model 2 100 atomic absorption spectrometer. The filterless knotted reactor precipitate collectors were made from 0.5 mm id, 1.8 mm od Microline tubing (cross-linked ethyl vinyl acetate) by tying interlaced knots (the optimum length of the knotted reactor, LKR, was 100 cm).The knots were made with approximately 5 mm diameter loops. All the other reaction coils, connections and conduits in the FI-manifold (Fig. 2) Fig. 1 Schematic diagram of the on-line co-precipitation-dissolution HGAAS system with off-line addition of precipitating agent (La'", added to individual samples), shown in (a) the loading (precipitating) sequence and (b) the elution stage. QTA, quartz cell; Ar, argon; RC, reaction coil; SP, gas- liquid separator; P1 and P2, peristaltic pumps; V, valve; W, waste; and KR, filterless knotted reactor (Redrawn from ref. 17, with permission from the Royal Society of Chemistry). Fig. 2 Schematic diagram of the FI-HGAAS system for the on-line co- precipitation-dissolution HGAAS system with on-line addition of La"'.MC, mixing coil (6 cm); other symbols as in Fig. 1 . The pumping rate of the sample solution (X) was investigated at three different levels (i.e., 4.8, 6.4 and 8.8 ml min-1, respectively; for details, see text).Analyst, January 1996, Vol. 121 33 consisted of 0.5 mm id PTFE (polytetrafluoroethylene) tub- ing. Reagents and Standard Solutions All the reagents were of analytical-reagent grade, and distilled water was used throughout. Sodium tetrahydroborate solution [0.3% (m/v) in 0.05 mol 1-1 sodium hydroxide solution] was prepared freshly daily. Lanthanum nitrate hexahydrate solution, 0.5% m/v, was made by dissolving 0.6662 g of lanthanum nitrate hexahydrate in 100 ml of distilled water. The buffer solutions, freshly prepared every day, were in all instances 0.2 moll-' ammonium chloride adjusted to the appropriate pH buffer (9.1-9.2) by addition of 0.2 mol 1-1 ammonia solution (the optimum pH value depended on the sample flow rate; see Results and Discussion).Standard solutions of selenium(1v) for calibration purposes were prepared by three-stage aqueous dilutions of a 1000 mg 1-1 stock solution, which was made by dissolving 1.4053 g of selenium dioxide in 1000 ml of 1.0 mol 1-1 hydrochloric acid. All glassware was soaked for at least 24 h in 1 moll-' nitric acid, and finally rinsed in distilled water before use. Operational Procedure The time-based FI-HGAAS system with on-line addition of the co-precipitant, Lar1*, is shown in Fig. 2, depicting the optimized experimental parameters.In the precipitation procedure (Fig. 2a), the sample, hydrochloric acid and sodium tetrahydroborate solutions were introduced by pump 1, and the buffer and lanthanum nitrate solutions were delivered via pump 2. During the precipitation sequence both pumps were activated for a period of 99 s. The sample and La'r1 were pre-mixed in the mixing coil (MC) of length (LMC) 6 cm, and subsequently merged with the buffer solution at the entrance to the knotted reactor (KR) which had a length (LKR) of 100 cm. The precipitate, which was formed instantaneously after the merging point of KR, was collected on the inner walls of the knotted reactor. The effluent emerging from the reactor was discarded. Simultaneously, the acid was pumped through the by-pass of the rotor of the valve and directed into the hydride generation system.During this stage, the baseline for the final readout was established. At the end of the precipitation period pump 2 was stopped, and the valve was actuated automatically from the fill-mode to the inject-mode for a period of 10 s (dissolution-hydride- generation procedure; Fig. 2b), by which means the acid was introduced to the knotted reactor where the precipitate adhering to the inner walls of the knotted reactor was dissolved. This concentrated zone was directed from the knotted reactor to the hydride generating system, where the analyte was merged with a reducing solution of sodium tetrahydroborate. After passing through a reaction coil (RC) (LRC = 35 cm), the gas-liquid mixture was guided into the gas-liquid separator (SP) in which the hydrogen selenide and the evolved hydrogen were separated and swept into the atomizer cell by a steady argon carrier flow.The absorption signal was then recorded. The waste from the gas-liquid separator, which included unreacted sodium tetrahydroborate, acid and some argon, was removed by aspiration. After 10 s in the inject mode, the valve was returned to the fill mode, allowing the sample solution to be interchanged so that the remainder of the previous sample in the sample pump tube could be effectively washed out and the next sample kept ready for precipitation. Results and Discussion Preliminary Investigations In their experiments with off-line addition of co-precipitant, Tao and Hansen17 anived at the optimized operational par- ameters shown in Fig.1, pointing out that the pH of the precipitation reaction was very critical. When the present authors tried to reproduce the assay, difficulties were encoun- tered in achieving comparable characteristics, i.e., the peak heights obtained were somewhat lower (about 25%) than those previously reported. While this might possibly be due to the 'knotting efficiency' of the incorporated reactor in the two systems, it was nevertheless decided to make a closer scrutiny of the flow rate of the ammonia buffer used. It was thus found that with all other parameters fixed as indicated in Fig. 1 (the concentration of La added to each sample being 20 mg 1-l, and the pH of all sample solutions adjusted to 3), progressively lower flow rates of the buffer (pH 9.1) yielded increasingly higher signals, the optimum flow rate (QBuffer) being 0.5 ml min-1, that is, one third of the flow rate used by Tao and Hansen. With these conditions, comparable results were obtained, and therefore the low buffer pumping rate was used in the following on-line approach.Optimization of the On-line Co-precipitation-Dissolution Preconcentration System with On-line Addition of the Co-precipitant The FI-manifold used for the on-line addition of precipitation agent is shown in Fig. 2. As a first approximation, the optimization procedure might be effected by reproducing the favourable kinetic conditions prevailing in the off-line system of Tao and Hansen, that is, by ensuring that the presentation of the samples to the co-precipitation reactor and that the reaction parameters within it (ix., concentrations of constituents and the pH of the buffer) are identical in the two instances.To reproduce the sample presentation of the off-line system, the aqueous sample solutions were therefore pre-mixed with an Larr1 stream, the pH of which was adjusted to 3 by means of the addition of hydrochloric acid. Thus, with the individual pumping rates used and with due consideration to the dilutions of the merging streams, the concentration of the La stream yielding the optimum concentration, as found for the off-line system (20 mg 1-I), might readily be calculated. However, with different lengths of the added mixing coil MC (varied between 50 and 200 cm), it was only possible to reach about 35% of the results that Tao and Hansen achieved with off-line addition of Lar1'.This proved that the conversion from the off-line to the on- line addition of co-precipitating agent was no trivial task, very likely because the precipitation reaction takes place under dynamic conditions, where it is very critical how the precipitate is formed in order to be entrapped in the knotted reactor. Thus, it is essential that the hydrophilic precipitate consist of small, curdy particles which willingly adhere to the hydrophilic Microline tube and that the formation of larger particles, which might be flushed through the reactor, is prevented. A feasible avenue to improve the performance of the system, and possibly reach higher enrichment factors (EF, the ratio between calibration curves with and without preconcentration), would therefore be to increase the delivery of sample solution during each cycle.This could be carried out either by prolonging the time for the aspiration of sample, which would have a negative effect on the sampling frequency, or by increasing the flow rate of the sample solution. It should be noted that in the latter instance, which would be operationally preferable, it is a condition that quantitative entrapment of precipitate is still achieved, and therefore a conscientious optimization procedure is called for. However, with increasing accumulated precipitate in the knotted reactor, the back pressure of the system (as expected) tended to increase, and in order to avoid troubles due to potential disruptions of tube connections Tao and Hansen therefore limited the sample flow rate (Qs) to 4.0 ml min-1.In the present investigation, where higher pumping rates were also34 Analyst, January 1996, Vol. 121 attempted, care was taken to fasten the acid and sample tubes securely with metal wire tighteners, which therefore permitted us to overcome possibly higher back-pressures and hence to increase the sample flow rates, which in turn allowed this parameter to be altered in order to achieve higher sensitiv- ities. The sample flow rate was tested at levels of Qs = 4.8,6.4 and 8.8 ml min-l, respectively, with the length of the knotted reactor ( L K ~ ) affixed at 100 cm. For each sample flow rate used, the FI system was optimized accordingly, that is, the parameters which have any influence on the co-precipitation ([LaTrr], pHBuffer, QL~III and QBuffer) were investigated, while the experimental parameters for the hydride generation procedure were identical with those described previously17 (Fig.1 b). As it turned out, the experimental parameters for the co-precipitation procedure were rather similar to those reported by Tao and Hansen, except that it was necessary to increase the pH of the buffer slightly as the sample flow rate was increased. The optimized parameters are depicted in Fig. 2 and the results obtained are shown in Table 1. While it should be expected that the limit of determination is improved with increasing pumping rate of the sample, it is significant that the sensitivity of measurement increased with increasing sample consumption.Thus, at Qs = 8.8 ml min-l, the sensitivity was 0.776 specific absorbance (pg l-l)-l, or almost twice that achieved with the off-line addition. During the optimization procedure of the co-precipitation reaction it was found that it was unnecessary to add acid to the lanthanum nitrate solution. Therefore, for a given sample flow rate, optimum signals could be achieved simply by optimizing the pH of the buffer solution (see Table 1). Thus, by ensuring that it had sufficient buffering capacity (0.2 mol 1-l), identical 0.4 0.39 0.38 0.37 0.36 0.35 0.34 0.32 0.33 ~ 170 180 190 200 210 Po 230 240 2% La"' (pprn) Fig. 3 Effect of the lanthanum nitrate concentration on the peak absorbance of a 0.5 pg 1-1 Se'" standard. The pumping rate (X) of the sample solution was 8.8 ml min-1 (pH 9.17); all other experimental parameters as in Fig.2. signals were achieved for the aspirated sample solutions even if the pH of these solutions was varied within a pH range of 3-7. Therefore, hydrochloric acid was not added to the lanthanum nitrate solutions in further work. For a sample flow rate of Qs = 8.8 ml min-l the optimization of the LarrT concentration is shown in Fig. 3. For all the sample flow rates used the optimum concentration of La(NO& was found to be 210 ppm. This is not surprising because the premixing of the sample and the Larr1 solution in the very short premixing coil, MC (LMc = 6 cm), is minimal. Rather, the mixing of the sample and the lanthanum nitrate solution was optimized by optimizing the length of the knotted reactor, L m , which was established to be 100 cm.In the preconcentration stage, the main objective is to ensure that the precipitation is quantitative or that as much as possible of the precipitate formed is entrapped in the knotted reactor. Therefore, it should be expected that it would be preferable to utilize relatively small flow rates of the buffer and the lanthanum nitrate solutions to minimize the total flow rate through KR and thereby avoid unnecessary disturbance of the co-precipitation process. This was, in fact, confirmed in this investigation. Furthermore, it was found, which in retrospect is not surprising, that the optimized flow rates of the buffer and lanthanum nitrate solutions were constant at all sample flow rates used, and that the optimum pH of the buffer increased with increasing sample flow rates.However, as was shown experi- mentally, the optimized pH value of the buffer is a very critical parameter for each individual sample flow rate; this is noted in Table 1. Indeed, if the pH was varied between 9.1 and 9.2 at the various flow rates tested, variations up to 2630% in the recorded signals were actually obtained. Yet, at fixed values, which are readily maintained by ensuring a sufficient buffering capacity of the buffer, the system is very robust. 0.8 0.7 S 0 0.6 0.5 8 0.4 0.3 h .- c - v 0 5 0.2 0.1 0 1 2 3 4 5 6 7 8 9 Flow rate/rnl rnin-' Fig. 4 Graphical representation of the slopes of the individual calibration curves as a function of the flow rates of the sample solution.Experimental parameters as in Fig. 2 and as shown in Table 1. Table 1 Characteristics for the FI on-line co-precipitation-preconcentration HGAAS system with on-line addition of Lallr at different sample flow rates Sample flow-rate/ml min-1 pH of buffer Concentration of La(N03)3/mg 1-1 Calibration range/pg 1-1 Regression equation in calibration range (6 standards, n = 3, CS, in yg 1-I) Sample volume per assaylml (loading time, 99 s) Sample frequency cf) (samples h-1) Relative standard deviation Limit of detection (3 s)/pg I-' Enrichment factor (EF) Concentration efficiency (n = 11; 0.1 pg 1-1) (%) (CE = EF fl60) 4.8 9.13 0.01-0.30 210 0.510 Cs, + 0.005 (Y = 1.000) 7.9 33 0.5 0.006 30 16.5 6.4 9.15 0.01-0.30 (r = 0.999) 210 0.678 Cs, + 0.003 10.6 33 0.5 0.005 40 22.0 8.8 9.17 0.005-0.30 ( r = 0.999) 210 0.776 Cs, + 0.009 14.5 33 0.8 0.004 46 25.3Analyst, January 1996, Vol.121 35 By depicting the slope of each calibration curve versus the sample flow rate (Fig. 4), it is feasible to obtain useful information as to the capacity of the knotted reactor of LKR = 100 cm. Thus, according to this figure, there is a tendency to a linear relationship until a sample flow rate of 6.4 ml min-1 is reached, which strongly indicates quantitative co-precipitation. Above this level, the curve tends to bend off, indicating a lack of sufficient capacity of the knotted reactor at these sample flow rates (and the co-precipitation period of 99 s). Very likely this ‘breakthrough’ of the capacity of the reactor can be explained by an insufficient adherence of the precipitate on the inner walls of the knotted reactor because of a too high flow through the knotted reactor at high sample flow rates.This explanation was, in fact, verified by halving the sample flow rate and doubling the time for the co-precipitation while maintaining all other experimental parameters as in Fig. 2. Conclusion The conceptual idea of Tao and Hansenl7 and the work presented in this paper demonstrate that it is possible to arrange an effective on-line FI procedure for preconcentration of trace levels of elements by co-precipitation-dissolution in inorganic media. The performance of the previously described system has been improved from off-line addition to on-line addition of the co- precipitant to the sample. The optimization by on-line addition yielded very satisfactory results, notably allowing higher sensitivity of measurement and yielding better enrichment factors (EF) and improved concentration efficiencies (CF = EF f/60, see Table l), the former being twice as high as previously obtained.17 These advantages primarily resulted from the on- line system allowing the consumption of higher sample volumes at higher sample flow rates, which, in turn, was enabled by mechanical improvements to the FI system, notably the use of metal wire tighteners to fasten the sample and acid pump tubes.With the principles of this FI-HGAAS system for precon- centration by on-line co-precipitationdissolution with on-line addition of the co-precipitant, it could be of interest to include other hydride forming elements.Further, it would be interesting to test the possibilities for developing the FI system by incorporating on-line speciation of the hydride forming ele- ments. Work in this direction is currently being conducted at this laboratory. The authors wish to extend their appreciation to Julie Damm’s Foundation (Denmark) for partial financial assistance of this research programme. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Martinez-JimCnez, P., Gallego, M., and Valcircel, M., J . Anal. At. Spectrom., 1987, 2, 21 1. Martinez-JimCnez, P., Gallego, M., and Valcircel, M., Anal. Chem., 1987,59,69. Montero, R., Gallego, M., and Valcircel, M., J . Anal. At. Spectrom., 1988,3,725. Montero, R., Gallego, M., and Valcircel, M., Anal. Chim. Acta, 1988, 215, 241. Martinez Calatayud, J., and Garcia Mateo, J. V., J . Pharm. Biochem. Anal., 1989, 7 , 1441. Du, K. P., Wang, Y. Z., and Fang, Z. L., Shenyang Yaoxueyuan Xuebao, 1992,9, 130. Martinez-JimCnez, P., Gallego, M., and Valcircel, M., Analyst, 1987, 112, 1233. Santelli, R. E., Gallego, M., and Valcarcel, M., Anal. Chem., 1989, 61, 1427. Santelli, R. E., Gallego, M., and Valcircel, M., J . Anal. At. Spectrom., 1989, 4, 547. Esmadi, F., Kharoaf, M., and Attiyat, A. S . , Microchem. J., 1989,39, 71. Fang, Z. L., Flow Injection Separation and Preconcentration, VCH, Weinheim, Germany, 1993. Fang, Z. L., Sperling, M., and Welz, B., J . Anal. At. Spectrom., 1991, 6, 301. Fang, Z. L., and Dong, L., J . Anal. At. Spectrom., 1992, 7 , 439. Chen, H., Xu, S., and Fang, Z. L., Anal. Chim. Acta, 1994, 298, 167. Welz, B., Xu, S., and Sperling, M., Appl. Spectrosc., 1991, 45, 1433. Min, R. W., and Hansen, E. H., Chem. Anal. (Warsaw), 1995, 40, 243. Tao, G. H., and Hansen, E. H., Analyst, 1994, 119, 333. Paper 51031 99K Received May 19,1995 Accepted September 7,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100031

出版商:RSC    

年代:1996

数据来源: RSC