ISSN:0144-557X    

DOI:10.1039/AP99330FX005

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

February 1993 ANPRDI 30(2) 65-124 (1993) Analytical Proceedings Proceedings of the Analytical Division of The Royal Society of Chemistry 94 96 98 101 103 106 110 Productivity Ehhancement in Atomic Spectroscopy 110 CONTENTS 65 Reflections of SAC '92 67 Reports of Meetings 67 VAM Viewpoint 'EURACHEM Workshop on Comparability and Traceability' by Ronald F. Walker 'The Who, What, Why and How of Waste Analysis' by Neil B. Tytler and S. J. Tully 'Ultraviolet Absorption and Luminescence Spectrophotometry for the Analysis of Pesticide and Herbicide Residues and Traces' by Jean-Jacques Aaron 69 SUMMARIES OF PAPERS 69 Air, Earth, Fire and Water-The Analysis of Waste 69 72 The Analyses of Fungicides, Herbicides and Insecticides 72 'Biosensors for the Analysis of Pesticide Residues' by Canh Tran-Minh 73 74 'Analysis of Captan in Fruits by a Magnetic Particle Based lmmunoassay' by Janet S.Colliss 'Interlaboratory Comparability of Accuracy in Residue Analysis' by Nigel A. Smart 'New Directions in Affinity Chromatography' by K. Jones 'The Sievers SCD 350 Detector' by M. Dyson 'Geochemical Applications of Pyrolysis Gas Chromatography with an Atomic Emission 90 75 78 New Directions in Chromatography 78 79 87 Detector' by Steve Rowland, Roger Evens, Les Ebdon and Andy Rees 'Supercritical Fluid Chromatography; Recent Developments and New Directions' by 89 Hans-Gerd Janssen and Carel A. Cramers 'Photoionization Detection is 30 Years Old. The Story So Far Plus "Son of Photoionization Detection": Far-ultraviolet Adsorption' by J.S. Hayhurst and J. N. Driscoll 93 Research and Development Topics in Analytical Chemistry 93 'Flow Injection Studies of Cyclodextrin-enhanced Fluorescence in the Amino Acid-Thiol- o-Phthalaldehyde Reaction' by Muhammad Y. Khokhar and James N. Miller 'Design and Application of Chiral Liquid Chromatography for Drug Metabolism Studies' by G. J. Furlonger, Anthony F. Fell and B. Kaye 'Comparison of Absorbance Ratios and the Peak Purity Parameter for the Verification of Peak Homogeneity in High-performance Liquid Chromatography' by Daemon Lincoln and Anthony F. Fell 'Assessment of the Performance of a New Protein-based Phase in the Chiral Liquid Chromatography of Drugs' by M. C. Banks, Anthony F. Fell and R. D. McDowall 'Drug and Analogue Structural Relationships in Chiral Separations with Mobile-phase Additives' by P. Mitchell and B. J. Clark 'Optimization of the Separation of Anthracyclines and Their Metabolites Using Reversed- phase Liquid Chromatography' by G. Nicholls, B. J. Clark and J. E. Brown 'Liquid Chromatographic Studies on the Potential Degradation of Preservatives in Formulated Drug Products' by Melanie J. Thompson, B. J. Clark, Anthony F. Fell and M. L. Robinson 'Appropriate Precision: Matching Analytical Precision Specifications to the Particular Application' by Michael H. Ramsey 113 Equipment News 118 Analytical Division Distinguished Service Award 118 SAC Silver Medal (Rules) 119 Conferences and Meetings 121 Courses 121 Publications Received 122 Analytical Division Diary 0144-557x1199312; 1-A ANALYTICAL PROCEEDINGS, FEBRUARY 1903, VOL 30 ... I l l

ISSN:0144-557X    

DOI:10.1039/AP99330BX007

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 65 Reflections of SAC ‘92 SAC ‘92 took place in Reading from September 2Oth-26th, 1992, and attracted an international group of delegates from a wide range of disciplines related to ana- lytical chemistry. The meeting was organized in three streams, ensuring that almost everyone could miss or attend some presentation of interest! Sessions were devoted to molecular spectroscopy (Spectroscopy Across the Spectrum), environmental analysis, chromatography and separation science, analysis of phar- maceuticals and in biotechnology, atomic spectroscopy, chemome trics, electro- chemistry, automation and Row injection, sensors, forensic methods and analytical QC and accreditation. By any standards this constituted an ambitiously compre- hensive programme.science by analytical developments, in, for instance, medical diagnosis, were in evidence at Reading. I t is not the inten- tion here to provide a comprehensive overview of the meeting, or even a representative sample, but merely a personal recollection of some aspects of the conference. Others will have alterna- tive memories of their time in Reading with new contacts made and, perhaps, old ideas refreshed. Abstracts of presentations and posters and selected papers have been published in Analytical Proceedings. ‘ 7 ’ Combined this year with the 150th anniversary of the Laboratory of the Government Chemist (LGC), the meet- ing underlined some of the traditional roots of analytical science and thereby Dr. B. King a sound scientific facility.The LGC’s promotion of Valid Ana- lytical Mcasurement fitted well with one of the underlying themes of the meeting, which stressed and demonstrated the importance of the integrity of measure- ments. Sessions relating to Method Vali- dation and Proficiency Testing, Refer- ence Materials and Traceability and Analytical QC and Accreditation provided a forum for those who depend upon the accuracy of analytical data. Amongst the many visits during the week that to the LGC’s new premises at Teddington was well attended and appre- ciated by the delegates. It was here that the theme of the contribution of analysis to an expanding global trade, and the The plu!forrn p r t y prior to the Opening Ceremony. Buck row (L-K): Dr. J. S. Gow (Secretury- Gc.ncrtrl, RSC‘), Prof?ssor G.Den Boef (IUPAC) und Professor J . Munn (University of Reading). Front row (I.-K): Pro.fessor c‘. W . Rees (President, RSC), Dr. G . Robinson (DTI), Dr. E. J . Ncwmuri (Prc.sidc>nt, Anulyticul Division), Dr. R. D. Worswick (The Government Chemist) and MY. P. G. W . Cohh (Cliairmun, SAC Y2 Executive Committee) Triennial conferences, of which SAC is one, highlight the progressive changes that continue to occur in analytical chemistry. The time between such meet- ings is sufficient for advances to be significant, both in the detailed techno- logy and in the developments of preferred organizational procedures. Delegates heard from speakers of progress made in the subject and would have been able to consider the technological changes that were influencing chemical developments and the demands that science rightly requires of analysis.Both the technological pull of new, more-demanding applications, as, for example, in some environmental ana- lyses, and the technological push given to emphasized its continuing and indeed growing importance and relevance in both the scientifically obscure and the day by day life of the nations of the world, indeed of the very world itself. An interesting personal note was drawn to the notice of the delegates by Dr. R. Worswick, The Government Chemist, in that the first Principal of what was to become the LGC, George Phillips, received a per- formance related pay award and addi- tional funding for the laboratory by virtue of the value attributed to his analytical work. Perhaps there is a lesson for many of us here in that demonstrating the value of analytical contributions is an important duty and responsibility so as to ensure adequate support for the maintenance of Profcxsor A.F. Fell66 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 Professor M . Valcarcel inevitably increased regulatory require- ments that go with it, was developed in the CHEMAC meeting. CHEMAC is the body providing the UK input to Eura- chem, which is the European forum for analytical chemistry, and the meeting dealt with aspects of the European perspectives of analytical quality assur- ance and the European Commission’s programmes on measurement and testing. The presentations and visits formed but part of the conference for it is in the informal contacts and the relating of ideas in different areas that much of the new impetus for fundamental developments and new applications is provided.As with most major conferences the Plenary lectures provided focal points at the meeting and the contributions were all well received, giving opportunities for the invited lecturers either to survey major advances in a given subject area or to present new advances in their particular areas of specialization. Dr. Bernard King gave the first Plenary with a description of the progress towards ‘The Development of an International Measurement System’. Wide ranging reviews of ‘Hyphenated Techniques’ and ‘Flow- through Biochemical Sensors’ were provided by Professor A. Fell and Pro- fessor M. Valcarcel, respectively, whilst Professor Callis and Professor Klockow presented some of the impressive advances in their areas.Professor Callis addressed himself to the application of NIR spectroscopy, with an associated and appropriately selected chemometric tech- nique, to the analysis and examination of the determination of biomass in fer- menters and the assessment of skin burns (this being an accurate, non-invasive, means of treatment diagnosis). Professor Klockow discussed the analytical oppor- tunities and difficulties associated with ‘Speciation in Atmospheric Trace Analysis’. The usually independent meeting of Spectroscopy Across the Spectrum, or- ganized by the Molecular Spectroscopy Group, was included within the SAC programme at Reading, providing oppor- tunities for delegates to hear more of these topics than would otherwise be the case.Contributions in the Chemometrics ses- sion at the meeting illustrated well that information-rich data have for many years been less than adequately sifted and used; however, there is a growing awareness that visual or electronic inspection of univariate data leaves much to be desired and multivariate approaches can provide hitherto unrealized dimensions of infor- mation. The applications of robust statis- tics and the use of chemometrics in imaging was particularly well received by the delegates. One of the fundamental stages in any analytical technique is the perturbation of analyte molecules and there is a growing body of knowledge relating to our ability to interact selectively in the molecular domain. Laser techniques coupled with mass spectroscopy provided one such example of these developing manipula- tive capabilities in a presentation by Colin Creaser from the University of East Anglia.A number of posters were presented at the conference, thus allowing a wider incorporation of topics to be included in the meeting. The one-to-one interactions fostered by such sessions can be especially fruitful if the atmosphere is created cor- rectly. It was in this form that the Educational Swapshop provided oppor- tunities for the interchange of valuable teaching experiences. Despite the eloquence of lectures and the imaginative preparation of visual aids analytical science and chemistry itself remain practical pursuits. This was most elegantly demonstrated by Dr. M. Professor D. Klockow Profrssor J .Cullis Leonard of Queen’s University, Belfast, in his presentation to the meeting of ‘Trace Fluorine Absorptivity’. Long after the audience of this particular lecture have forgotten the details they will remember the impressive colour changes that reflect our ingenuity for developing selective molecular and ionic interactions that are so often the basis of sensitive analytical procedures. Registration is a topic of concern to chemists in the development of their careers. The registration procedure is designed to establish recognition of the equivalence of different qualifications and accumulated experience. The develop- ment of the European Chemist designa- tion ,3 which goes towards classifying chemists from diverse backgrounds within Europe, was explained by Dr.J. S. Gow, the Secretary General of the Royal Society of Chemistry. Analytical chemistry is more than ever before the domain of the specialist and therefore it was appropriate that the procedural con- siderations relating to the Royal Society of Chemistry’s Indicative Register of Analytical Chemists3 were presented and discussed in a special session. As is usual at SAC Conferences, the social programme was varied and enjoy- able. In particular, the barbecue on the Friday evening was extremely well received, with huge amounts of food and wine available and also great entertain- ment from the Tadley Concert Band, who were frequently joined by the SAC conferees. Reading was memorable as a meeting and a venue but future meetings are likely to be different. The science will have changed by the time this international gathering convenes in Hull in 1995 and so may our approaches to effective com- munication processes that are aimed at increasing awareness of developing areas and promoting more participative discussion.67 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 References 3 For details of European Chemist and Royal Society of Chemistry, Thomas 1 Anal. P r o c . , 1992, 29, 309-384. Indicative Register of Analytical Graham House, Cambridge CB4 4WF. 2 Anal. P r o c . , 1993, 30, in the press. Chemists: Membership Department, JOHN GREEN

ISSN:0144-557X    

DOI:10.1039/AP9933000065

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 67 Reports of Meetings East Anglia Region The twenty-fifth Annual General Meeting of the Region was held at 1.45 p.m. on Tuesday, November 17th, 1992, at Warren Spring Laboratory, Gunnels Wood Road, Stevenage. The Chair was taken by the Chairman of the Region, Dr. C. S. Creaser. The following office bearers were elected for the forthcoming year: Chairman-Mr. P. R. Brawn. Vice- Chairman- Mr. P. Snowdon. Honorary Secretary-Mr. A. Anderson, Depart- ment of Analytical Chemistry and Phar- macy, Huntingdon Research Centre, P.O. Box 2, Huntingdon, Cambridge- shire PE18 6ES. Honorary Treasurer- Mr. B. W. Woodget. Honorary Assistant Secretary-Mrs. I. S. Dawe. Members of Committee-Mr. P. E. Anderson, Dr. C. S. Creaser (ex ofJicio), Professor M.B. Evans, Dr. A. R. Griffiths, Dr. A. Roberts-McIntosh and Mr. M. Scotter. Mr. A. G. Croft and Mr. G. M. Telling were re-appointed as Honorary Auditors. Thermal Methods Group The twenty-eighth Annual General Meet- ing of the Group was held at 12.30 p.m. on Thursday, November 19th, 1992, at the Scientific Societies’ Lecture Theatre, London. The Chair was taken by Mr. P. J. Haines, a past Group Chairman. The following office bearers were elected for the forthcoming year: Chairman-Dr. R. S. Whitehouse. Vice-chairman- Dr. D . J . Morgan. Honorary Secretary - Dr . C. J. Keattch, Industrial and Laboratory Services, P.O. Box 9, Lyme Regis, Dorset DT7 3BT. Honorary Treasurer- Dr. R. H. Still. Members of Committee-Dr. B. Cantor, Mr. 3 . P. Davies, Dr. J. Deeny, Dr. J . P. Gupta, Mr. S. R. du Kamp, Dr. M. Odlyha, Mrs. J. A. Hider (ex oficio), Dr. G. M. Clark (co-opted), Dr. F. W. Wilburn (co-opted), Dr. R. C. MacKenzie (co-opted). Dr. A. Dyer and Dr. D. Griffiths were re-appointed as Honorary Auditors.

ISSN:0144-557X    

DOI:10.1039/AP993300067a

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 69 Air, Earth, Fire and Water-The Analysis of Waste The following is a summary of one of the papers presented at a Joint Meeting of the South East and Western Regions and the Micro & Chemical Methods Group held on March 25th, 1992, in HarweII Laboratory, Didcot, Oxfordshire. The Who, What, Why and How of Waste Analysis Neil B. Tytler and S. J. Tully Tavwood Environmental Consultancy, Green ford House, 309 Ruislip Road East, Green ford, Middlesex UB6 9BQ The recycling or disposal of waste is a science in its own right. Waste products arising from the whole spectrum of British industry, whether manufacturing, service or commerce is dealt with. There is probably only one type of waste not handled by the waste management industry, that being intermediate- and high-level radioactive waste.The chemical and physical tests employed in controlling wastes range from simple old- fashioned ‘0’ Level wet chemical methods to modern sophisti- cated techniques using state-of-the-art equipment. The three main procedures employed by the waste industry are essentially analogous to those used in chemical manufactur- ing. Firstly, in manufacturing, is evaluation of a supplier’s product to assess whether it is suitable for use, whereas in the waste industry this is the assessment of waste from a customer to allocate a suitable process route. Secondly, in manufacturing there is quality assurance of this product when it is delivered to check whether it conforms to the specification ordered. For waste materials, this is the evaluation on reception to ensure that the original specification agreed is complied with.Lastly, in manufacturing there are process controls which monitor each stage of a chemical process in order to optimize production and to ensure that any environmental discharge standards are being complied with. This is exactly the same with waste processing. Waste Assessment Assessment of a customer’s waste generally involves a rep- resentative (usually technical j visiting the customer to evaluate the waste. As previously mentioned, this waste can literally be anything, and comprehensive interrogation of the customer, usually by means of a questionnaire, is necessary before a possible disposal option can be formulated. Wastes are first evaluated to see whether they have any recovery potential: for example, a high metal content where the yield would be sufficient to cover processing costs and generate a reasonable profit.Organic solvent wastes are similarly evaluated. If there is no reclamation potential, then possibly an on-site treatment option may be suitable, e.g., in the case of contaminated soil. Ultimately, for many wastes a disposal facility must be found as all other recovery or treatment options are unsatisfactory. The three main disposal methods are landfill, treatment and incineration. Waste Licences All disposal facilities, regardless of which of the three disposal methods they offer, must have a licence to receive waste issued under the Control of Pollution Act 1974. I t is planned that this Act should be superseded by the Environmental Protection Act (EPA) 1990 in April 1993, with the exception of chemical incinerators which will have separate Integration Pollution Control (IPC) authorization under Section I of the EPA 1990.Each site licence is comprehensive and details the nature and composition of wastes each site is allowed to receive. The analytical assessment of a customer’s waste, therefore, is primarily to decide which disposal facility would be licensed to receive that waste, or how the waste might be modified so as to be acceptable to a given facility. Waste Analysis Inorganic Wastes Generally tests of the following properties would be carried out on inorganic wastes: pH; density; total acidity; total alkalinity; heavy metal content, i.e., Cu, Cr, Zn, Pb, Cd, Ni, Hg, Al, As, Sb, Ca, Se, P; anion content, i.e., SO4’-, CN-, C1-, NO3-, po43-.Organic Wastes For organic wastes the following properties represent an initial screening: flash point (Fpj; viscosity; calorific value (Cv); rate of combustion; mutual compatibility; pesticide residues; presence of PCBs, C1, Br, S, P; ash content; heavy metal content. A list of the different analytical techniques employed for determining all of these properties is given in Table 1. In summary, therefore, the main objectives in analysing samples of a customer’s waste are: (i) to confirm the customer’s original description/specification; (ii) to establish whether a recovery or disposal option is to be followed; (iii) to evaluate whether on-site treatment or part treatment is a possibility; (iv) to ascertain which facility is licensed to receive the waste; (v) to agree a specification for the waste with the customer before the material is transported. In the course of an assessment it is not uncommon to find all manner of unusual things.What the plant manager believes is the nature of his waste can be far from reality. Waste Reception The objective of testing waste upon arrival at the disposal facility is exactly the same as the testing of a raw material from a supplier; it is a quality control procedure. The nature of the tests, therefore, is more that of spot checks of selected determinands depending on the waste composition expected. These tests may include those of: colour; viscosity; Cv; CI70 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 Table 1 Example of analytical methods used in the waste industry Determinands Pesticidcs Organic solvents Halogenated organics Metals, carbonyls Phosphorus Organometallics Azides Oxidizing agents Polyhalogenated dioxins and furans Calorific value Particulate Flash point Total acidityhotal alkalinity COD BOD Ash Analytical techniques suitable Gas chromatography (GC) using thc following detectors: electron capture detector (ECD); nitrogen-phosphorus detector (NPD) GC with flame ionization dctector (FID) or coupled to a low resolution mass spectrometer (MS) GC-MS.Some may be better analysed by high-performance liquid chromatography (HPLC) or X-ray fluorescence Atomic absorption spectrometry (AA) or inductively coupled plasma (ICP) with an optical emission spectrometer (ES) or mass spectrometer (MS) Ion chromatography (IC) As metals, see above, or as compounds then HPLC-MS or HPLC-ICP-ES or HPLC-ICP-MS.Also colorimetric applications may be suitable for certain compounds so ultraviolet (UV) and visible spectroscopy may be used HPLC or infrared spectroscopy (IR) One or more of the above techniques would be suitable dependent upon the chemical nature of the compound involved High rcsolution mass spectrometry (HRMS) Isoperibol calorimeter BECURA or Andersons probes Abel closed cup Standard titration methods Potassium dichromate digestion Standard BODS.day apparatus Simple muffle furnace methods Table 2 Third International Conference on the protection of the North Sea, (list of hazardous substances according to priority) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1s 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Substance Mercury Cadmium Copper Zinc Lead Arsenic Chromium Nickel Drins HCH DDT Pentachlorophenol Hexachlorobenzene Hexachlorobutadiene Carbon tetrachloride Chloroform Trifluralin Endosulfan Simazine Atrazinc Tributyltin compounds Triphenyltin compounds Azi n phos-et h yl Azinphos-methyl Feni t ro t h ion Fenthion Malathion Parathion Parathion-methyl Dichlorvos Trichloroethylene Tctrachloroethylene Trichlorobenzene 1.2-Dichloroethane Trichloroethane Dioxins Water * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Air * * * * * * * * * * * * * * * * * content; ash content; pH; CN- content; phenol content; total alkalinity; and total acidity.Compatibility trials and a DOE flame test may also be performed, plus facility specific tests related to the site licence. The predominant difference between the reception of a waste and reception of a raw material is that invariably the waste is going to be processed immediately, whilst raw materials tend to be stored prior to use. Therefore, waste reception tests must be comprehensive but rapid and com- pleted within 4-1 h at most, the greater part of the analytical effort being expended at the beginning, in the evaluation of the customer’s waste before transportation. Process Control At the process control point the analytical work does vary significantly with the type of disposal facility, although in respect of certain environmental tests there is overlap.Landfill The successful operation of a landfill site is related in part to the proper management of groundwater, leachate and landfill gas, each of which require separate analysis. Leachate The aim is to see how waste is decaying and if any toxic substances are being leached out. Typical analysis undertaken would include: heavy metals (Cu, Cr, Zn, Hg, Cd, Ni, Pb, As, Se); anions (CN-, phenolics, S042-, Cl-); pH and possibly fatty acids. Ground water Detecting any changes in groundwater chemistry through contamination by leachate is the main reason behind the analysis. Measurement of pH, chemical oxygen demand (COD), biochemical oxygen demand (BOD), total organic carbon (TOC), hardness, Ca, K, Na, Mg, NO3-, P043-, SO4*-, conductivity and possibly pesticides is usually undertaken.Landfill gas The concentrations and volumes of landfill gas produced give an indication of how the microbial degradation of the waste is progressing. Testing for CH4, CO2 and 0 2 is generally carried out on-site as a matter of routine, with tests for H2, N2, 14C and chlorofluorocarbons occasionally being undertaken in the laboratory. Treatment and Incineration Two main types of analysis are carried out at these waste disposal facilities: those concerned with controlling the actual process, and those relating to environmental monitoring. As the range of actual processes used in the treatment and incineration of wastes is extensive, it is not intended to detail the various analyses carried out. However, they all fall into one of two categories: blending and mixing of wastes to produce a suitable feedstock for processing? or monitoring of the various reactions to completion with final tests to ensure that the desired outcome is achieved.All these controls are exercised by a combination of spot sampling and continuous monitoring. The final products of any treatment or incineration process are wastes in themselves, liquids, solids and gases. Liquids are either produced as filtrates from solid concentration processes, or as scrubbing liquids from gas-cleaning plants. They areANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 71 analysed for compliance with a discharge consent which details limits on ‘Red List’ substances (see Table 2), mostly metals, pesticides, chlorinated solvents and dioxins.Solids are pro- duced either as filter cakes or incineration ash and are destined for disposal to landfill. Again, heavy metals; residual hydro- carbons, PCBs and dioxins; and pH are the main determi- nands. Gases result from emissions from fume extraction plants for incinerators. Methods laid down by Her Majesty’s lnspectorate of Pollution (HMIP) in their Guidance Note BPM I 1 detail how particulate and acid gases (SO2, HCI, NO,) should be measured on a routine basis. Finally, other forms of environmental monitoring are also carried out both on- and off-site. Operators of landfill sites are required to monitor airborne dust and asbestos (when permit- ted for disposal) together with assessment of ground and surface waters for some distance beyond the site boundary.Treatment and incineration plants carry out extensive Control of Substances Hazardous to Health (COSHH) testing on-site for control of toxic fumes, as well as vegetation and water monitoring off-site. Conclusion The extent of analytical work carried out in pursuance of waste disposal has grown enormously in the last decade, and rightly so. Some techniques such as inductively coupled plasma optical emission (ICP-ES) have revolutionized the work that in the past had to be undertaken by slow atomic absorption (AA) methods. The detection of PCBs and dioxins is carried out to lo-’ standards as a matter of course using gas chromatography high resolution mass spectrometry (GC-HRMS) techniques. There is no doubt that as further advances are made in analytical methods the waste management industry will be expected to demonstrate further that their activities are removing pollution and not causing it.Only in this way can public confidence in waste disposal be gained. THE ROYAL SOCIETY OF CHEM ATOM IC SPECTROMETRY U PDATES STRY: ANALYTICAL DIVISION ATOMIC SPECTROMETRY GROUP A Joint Meeting on INSTRUMENTAL TECHNIQUE DEVELOPMENTS will be held on March 25, 1993 in Fitzwilliam College, Cambridge The programme of lectures for this meeting will be as follows. ’Laser excited atomic fluorescence spectrometry in a graphite furnace-femtogramme detection limits, but single element analyses; future prospects?’ by Professor R. G. Michel (University of Connecticut, USA); ’Recent developments in resonance ionization mass spectrometry’ by Dr. A. W. McMahon (AEA Environment and Energy); ’Thermal ionization mass spectrometry for the measurement of stable isotopes in environmental studies‘ by Dr. J. R. Bacon (Macaulay Land Use Research Institute); ’Novel approaches to microwave sample preparation and sample introduction t o microwave induced plasma atomic emission spectrometry‘ by Dr. H. Matusiewicz (Polytechnicka Poznanska, Poland); ’Measurement uncertainties in quadrupole ICP-MS’ by Dr. S. T. Sparkes (University of Plymouth); ’The development of AAS instrument control software’ by Dr. I. L. Shuttler (Perkin- Elmer Bodenseewerk); ’Every pciture tells a story-ICP signals captured’ by Mr. A. Batho (Thermo Electron). The registration fee, including coffee, buffet lunch and tea, is: f40.00 for RSC members; f50.00 for non- members; f25.00 for students and retired members. Overnight accommodation is available at the college at an additional cost of f40.00 for bed and breakfast. For information and booking forms please contact Ms J. M. Cook, British Geological Survey, Keyworth, Nottingham NG12 5GG.

ISSN:0144-557X    

DOI:10.1039/AP9933000069

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

72 ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 The Analyses of Fungicides, Herbicides and Insecticides The following are summaries of four of the papers presented at a Meeting of the Analytical Division held on March 27th, 1992, in the Scientific Societies' Lecture Theatre, London W1. Ultraviolet Absorption and Luminescence Spectrophotometry for the Analysis of Pesticide and Herbicide Residues and Traces Jean-Jacques Aaron lnstitut de Topologie et de Dynamique des Systemes de I'Universite Paris 7, Associe au C.N.R.S., I rue Guy de la Brosse, 75005 Paris, France Because of their widespread use, pesticides are often present in the environment at polluting levels. Pesticide residues there- fore constitute a subject of considerable concern. Recently, it has been emphasized that, in many instances, multi-residue analytical methods routinely used in the United States for monitoring contaminants in foods detect only about 40% of the pesticides likely to leave residues.This paper describes the usefulness of optical methods which, because of their relatively high sensitivity and their specificity, can determine a single pesticide or pesticide class. These techniques are also important for improving detection in thin-layer chromatographic (TLC) and high-performance liquid chromatographic (HPLC) analysis of pesticides at trace Optical methods used for pesticide analysis include ultraviolet (UV) absorption, fluorescence and phosphor- escence spectrophotometry. The main features, advantages and limitations of these three particular techniques are presented.UV Absorption UV absorption spectrophotometry has been applied to the determination of several classes of pesticides, such as or ano- Spectroscopic and Structural Requirements Absorption bands must be in the range 190-400 nm with E ~ ~ , ~ >lo3 1 mol-' cm-I. I t is preferable that an aromatic and/or a heterocyclic moiety be present in the pesticide structure. Alternatively, chemical derivatization may lead to a UV- absorbing structure. Application to HPLC and TLC Analysis In spite of its moderate sensitivity limits of detection (LOD) between about 0.01 and 7 pg ml- 1, this technique has the advantage of being inexpensive, versatile, easily interfaced with HPLC and TLC, and applicable to a large variety of pesticides. Derivative Spectra Second-derivative UV spectra can be used to identify herbi- cides such as dinitroaniline derivatives in soil and plants.8 phosphates,'T5 carbamate insecticides,' and herbicides. 7.a \ Luminescence Fluorescence Spectrometry Fluorescence spectrometry has been utilized for directly quantifying several types of pesticides, such as carbamate insecticides."-" Furthermore, fluorescence detection has been combined with TLC for the qualitative or quantitative analysis of organophosphates and carbamate and Absorption bands of pesticides studied by fluorescence are generally in the 250-3.50 nm region.Emission bands are in the 300-450 nm region. Aromatic and/or heterocyclic structures with fluorochrome substituents are recommended for improv- ing the sensitivity of the fluorimetric method.Fluorimetric HPLC (or TLC) detectors are extremely sensitive (LOD in the ng ml-I range) and selective, but the number of fluorescent pesticides is relatively small. For the latter reason, several fluorogenic reactions have been pro- posed. 2 3% 15 Fluorogenic labelling reactions using dansyl chlor- ide and fluorescamine have been developed, leading to LOD in the ng range (per TLC spot). Fluorogenic reagents such as chelates and flavones have also been proposed, with LOD in the 20-100 ng range (per TLC spot). A photochemically induced fluorescence technique has been developed recently for determining several classes of natural non-fluorescent pesticides. "-'* Because of its sensitivity (LOD at the ng m - ' level) and its broad range of application, this method, which is especially valuable for HPLC detection, has been described in some detail.I t is mainly applicable to herbicide derivatives of dinitroanilineIx and to nitrogenous pesticides . I "-' ' with HPLC for the analysis of herbicides and fungicides.---. 3 3 16.17 Phosphorescence Spectrometry Low-temperature (77 K) phosphorescence (LTP) and room- temperature phosphorescence (RTP) of a large number of pesticides have been studied and compared."-" Polychlori- nated phenols, carbamates and naphthalenic pesticides have been analysed by these phosphorimetric methods. Whereas LTP requires the use of cryogenic equipment, RTP facilitates detection of pesticides on solid surfaces (generally, cellulose). A drawback of the latter method, however, is the relatively large background phosphorescence due to the solid substrate.LTP has been shown to be a sensitive technique, with LOD values ranging between 0.001 and 30 pg ml-',-3 according to the pesticide structure. The RTP method presents several advantages over LTP for pesticide analysis, such as the small volume ( 5 PI) of sample required, the simplicity of the procedure and the specificity. For example, the detection of aromatic pesticides at the nanogram level is feasible in the presence of organochlorinated pesticides without significant interference by the latter species.'".''*''ANALYTICAL PROCEEDINGS, FEBRUARY 1993. VOL 30 73 1 2 3 4 S 6 7 8 9 10 I 1 12 References Newman, A. R . , Anal. Chem., 1989, 61, 861A. Shcrma. J . , A n d . Chem., 1989, 61, 153R. Shcrma, J ., Anal. Chem., 1991, 63, 118R. Mallet. V. N . , Duguay, M.. Bernier, M.. and Trottier, N.. Int. J . Environ. A n d Ciiem., 1990. 39. 271. Singh, R., Anulyst, 1989. 114, 425. Cabras, P.. Mcloni, M., Plumitallo, A., and Gennari, M.. J . Clirornatogr., 1989, 462, 430. Stahl. M., Luehrmann, M., Kicinski. H. G., and Kettrup, A.. %. Wussc>r Ahwusser Forsch., 1989, 22, 124; Chem. Absrr., 1989, 111, 83769s. Traorc. S . . and Aaron. J . J . , Anulyst, 1989, 114, 609. Argaucr, R. J . , in Anulyticul Methods f o r Pesticides and Plant Growth Regulutors, cd. Zweig, G.. Academic Press, New York. 1977, vol. IX, p. 119. Argaucr, K. J . , in Pesticides Analytical Melhodology, eds. Harvey, J . . and Zwcig, G., American Chemical Society, Washington, DC, 1980, p. 103. Larkin, M.J . . and Day, M. J . . Anul. Cliim. Acta, 1979, 108. 325. Aaron, J . J . , and Some, M., Analusis, 1982, 10. 481. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Lawrence, J . F., Renault, C., and Frei, R. W., J . Chromatogr., 1976. 121, 343. Volpe, Y ., and Mallet. V. N., Anal. Chim. Acta, 1976.81. 11 1. Traore, S . , and Aaron, J. J . , Talanta, 1981, 28, 765. Gauch, R., Leuenberger, V., and Mueller, U., 2. Lehensm- Unters. Forsch., 1989, 188, 36; Chem. Abstr., 1989, 110, 179 164w. Brenneckc. R., Ppan~enscliutz-Nuclir. Buyer (Engl. Ed.), 1988, 41, 113; Chem. Abstr., 1990. 112, 15861 Ig. Traore, S . , and Aaron, J. J . , Anal. Lett., 1987, 20, 1995. Miles, C. J . , and Moye. M. A., Anal. Chem., 1988, 60, 220. Patel, B. M., Moye, M. A., and Weinberger, R., J .Agric. Food Chem., 1990, 38, 126. Patel, B. M., and Moyc, M. A., Tulantu, 1991. 38, 913. Moye, M. A., and Winefordner, J . D., J . Agric. Food Chem., 1965, 13. 516. Aaron, J . J., Kalecl, E. M.. and Winefordner. J . D., J . ASric. Food Cliem., 1979, 27. 1233. Aaron, J . J . , and Winefordner. J . D., Anulusis, 1979, 7, 168. Aaron. J . J . , Ward, J . L., and Wincfordncr, J . D.. Analusis, 1982, 10, 98. Vo- Dinh, T.. Room- Temnperuture Phosphorirnetry f o r Clietnicul Analysis, Wilcy, New York, 1984, p. 279. Biosensors for the Analysis of Pesticide Residues Canh Tran-Minh Laboratory of Biotechnology, Ecole Nationale Superieure des Mines de St Etienne, 158 Cours Fauriel, 42023 St Etienne Cedex, France The primary purpose of controlling environmental pollution is to protect human health and the ecological integrity of the biosphere on which humans and other organisms depend.Great concern is being expressed over the effects of pesticides on public health and the survival of species of fish and wildlife. sensitive to pesticide residues since these compounds inhibit the activity of the enzyme on its substrate. The degree of inhibition can be expressed in terms of a percentage of total inhibition which is defined as the relative change in the biosensor response with and without inhibitor for the same substrate concentration. Enzymatic Techniques for Pesticide Residue Analysis The ability to detect and measure pesticides as they exist in plants, soils, natural waters and animals is essential in understanding the impact of these biologically active chemicals o n the environment.These residues are widespread although the amounts detected usually are fairly small. I n order to analyse large numbers of samples of unknown treatment history, analytical methods that can determine a number of pesticide residues simultaneously are most often used. The fact that many pesticides inhibit enzymes has led to the introduction of biological techniques for the detection and determination o f these various compounds. The basic action of organophosphorus pesticides is asso- ciated with their ability to inhibit acetylcholinesterase (AChE) in the central and peripheral nervous systems where the enzyme plays an important role in the transmission of nerve impulses. The reaction can be written as follows: hC Ilk Acetylcholine ---+ choline + acetic (1) Organophosphates and carbamates react with cholinesterase in a similar manner and this decreases the activity of AChE to its substrate acetylcholine.The resulting pH change, caused by the release of acetic acid, decreases as the concentration of pesticide increases. This can be measured by a potentiometric transducer associated with the immobilized enzyme. Immobilization Technique Three different techniques for immobilization of the enzyme butyrylcholinesterase and their use in association with pH glass electrodes for the determination of certain insecticides (carbo- furan, carbaryl and paraoxon) by inhibition of the enzyme, have been studied and the results compared. The immobiliz- ations are performed by (i) cross-linking a mixture of the enzyme and human serum albumin with glutaraldehyde on to nylon nets, (ii) fixing the enzyme on to commercially available polyamide pre-activated PALL membranes and (iii) physically retaining the enzyme in a synthetic prepolymer of acrylamide- methacrylamide hydrazide which is cross-linked with glyoxal.It is notable that the inhibitor detection limits are a direct function of the immobilized enzyme activity. In the construc- tion of sensors for inhibitor determination, the control of the amount, and hence the activity, of the immobilized enzyme is a determining factor. Low detection limits can be obtained with low enzyme activities. This is in contrast to enzyme sensors developed for substrate determination, where the aim is to achieve total conversion of substrate to product, by using a maximum amount of enzyme.Biosensor Response The dynamic behaviour of the sensors in terms of their response times to substrate and pesticide concentrations depends on the thickness and degree of reticulation of the enzyme membranes in each case. Sensor reproducibility for measurements of pesticides depends on the reproducibility of immobilization of the number of enzyme-active sites on the membranes. Therefore Biosensors Based on Enzyme Inhibition The immobilization of a cholinesterase (acetyl or butyryl cholinesterasc) on a pH electrode gives rise to a biosensor74 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 membranes or sensors produced in bulk generally tend to show more reproducible results compared with those made indi- vidually.The technique of immobilization on to nylon nets, therefore, produced more reproducible results compared with the other two techniques, since the enzyme solution was homogenized and spread evenly on to the net before it was finally cut into small discs. For both the nylon net and copolymer membrane sensors, the loss of enzyme activity was less when the sensors were maintained at the same temperature (25 "C), than when they were used at 4 and 25 "C intermittently. The PALL mem- branes, on the other hand, showed the opposite effect. This was probably due to a slower enzyme desorption process which occurred at lower temperatures and in unstirred conditions. Storage stability, as opposed to operational stability, is simply the ability of the enzyme to retain its activity under a specified set of storage conditions.It has been observed that, contrary to their stability in solution, nylon net and PALL enzyme membranes, when stored dry at 4 "C, showed no loss of their original enzyme activity even after a period of three years. This, however, was not possible with the copolymer electrodes as the gel had a tendency to dry up and crack. A comparative study of the costs involved in the various immobilization techniques (considering essentially the costs of the labour and materials) indicated that a ratio of 1:2:3 existed, with the nylon net membranes being the cheapest and the copolymer membrane the most expensive. The high cost of the copolymer membrane electrode can be attributed to the skilled labour required in the preparation of the copolymer and the need for several chemical reagents.In a study of the three different immobilization techniques for the determination of insecticides (carbofuran, carbaryl and paraoxon), the nylon net membranes appeared most promising in terms of reproducibility, stability and cost. However, it should be remembered that both nylon and PALL membranes have the tremendous advantage of maintaining the same enzyme activity for long periods (over three years) when stored dry. This feature, along with the fact that both nylon and PALL membranes are cheap and can be mass-produced and hence ensure better reproducibility, make them an appropriate choice for the development of 'disposable' sensors. Flow Injection Analysis (FIA) A biosensor can be used in conjunction with a flow injection system for the determination of enzyme substrates, mainly glucose, cholesterol and lactate.' In this instance, the biosensor acts as a detector which should provide an electrical signal proportional to the concentration of the analyte.The measure- ments are made before steady-state conditions are established, thus high sample throughput is possible. Sample (inhibitor) Injection - Was Pump L 2 m valve \ 0 m 4- .- ite Sensor Waste Detection cell Buffer -u (Substrate) Fig. 1 Experimental apparatus for pesticide determination using FIA Where the determination of enzyme inhibitors is concerned, the flow injection system offers the facility of separating the biocomponent from the basic sensor itself.2 Therefore, the various parameters involved in the enzyme inhibition kinetics can be investigated and studied independently.The enzyme acetylcholinesterase used for the determination of pesticides is immobilized in a single-bead string reactor (SBSR). The detector is a simple pH electrode with a wall-jet entry (Fig. I). Variations in enzyme activity due to enzyme inhibition are measured from pH changes when the substrate acetylcholine is injected before and after the passage of the solution containing the pesticide. The percentage inhibition of enzyme activity is correlated to the pesticide concentration. This FIA system has been applied to the determination of some organophosphorus (azinphos-ethyl, azinphos-methyl, bromophosphomethyl, dichlorovos, fenitrothion, malathion, paraoxon, parathion-ethyl and parathion-methyl) and carba- mate insecticides (carbofuran and carbaryl).Experiments were conducted in N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) buffer and a synthetic sea-water preparation. Several parameters influencing the performance of the system were studied and discussed. The detection limits of the insecticides ranged from 0.5 to 275 ppb. The enzyme reactor could be regenerated after inhibition with a dilute solution of pyridine-2-carbaldehyde oxime methiodide (2-PAM) and reused for analysis. The immobilized enzyme did not lose any activity up to 12 weeks when stored at 4 "C. This FIA system developed for cholinesterase inhibitors offers the advantages of being inexpensive, compact, simple to construct and sensitive for this application.References 1 Bertermann, K . , Eke, P., Scheller, F.. Pfeiffer, D., and 2 Kurnaran. S . , and Tran-Minh, C., Anal. Biochem., 1992, 200, Janchen, M., Anal. Lett., 1982, 15, 397. 187. Analysis of Captan in Fruits by a Magnetic Particle Based lmmunoassay Janet S. Colliss J. T. Baker UK, Wyvols Court, Basingstoke Road, Swallowfield, Berkshire RG7 IPY Captan is a non-systemic surface fungicide used to control a broad range of fungal diseases of fruits and vegetables. Analytical determinations of captan are important because of its potential carcinogenic properties and persistence on crops of up to 21 days after spraying. Classic analytical techniques for captan include gas-liquid chromatography and thin-layer chromatography. Although these techniques are accurate, they are time-consuming and tedious, requiring trained staff, specialized instrumentation and hazardous solvents such as benzene.Enzyme-linked immunosorbent assay (ELISA) offers an attractive alternative analytical method which is simple, cost effective and rapid, requiring minimal staff training and providing multiple results within an hour. In the captan RaPID (Ohmicron, Newtown, PA, USA) assay developed by Herzog et al. ,' captan and added enzyme-ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 75 labelled captan compete for binding sites on a polyclonal antibody, covalently bound to magnetic particles. The use of magnetic particles as the solid support and means of separation provides a large surface area for rapid reaction kinetics and improved sensitivity compared with other immunoassay techniques. Experimental Captan was extracted from samples of peaches or apple juice into acetone using a homogenization process.After filtration, the sample was concentrated by a solid-phase extraction procedure using BAKERBOND (J. T. Baker, 7Phillipsburg, NJ, USA) CI8 solid-phase extraction columns,- eluting the captan into methanol. Aliquots of the diluted methanol samples were added to disposable test-tubes, together with enzyme-labelled captan and magnetic particles with captan specific antibodies attached. After incubating for 30 min at room temperature, the unbound captan was separated from the bound fraction using the magnetic rack. After washing, colour reagent containing a substrate for the enzyme was added to the bound fraction and the absorbance measured after 20 min using the RaPID photometric analyser set at 450 nm.Captan levels were calculated by the programmed software from a range of standards included in the assay. A quick-screening procedure for fruit juice was investigated using neat and captan-spiked apple juice, omitting the solid- phase extraction concentration prior to the immunoassay procedure. The RaPID immunoassay method for captan was evaluated for precision, recovery, cross-reactivity and correlation with a standard gas chromatographic electron-capture detection (GC- ECD) procedure.3 Cross-reactivity Table 1 shows the cross-reactivity for captan and related phthalimide fungicides, expressed as the least detectable dose (LDD) and as a percentage. Table 1 Cross-reactivity Compound LDD (ppm) Cross-reactivity (%) Captan 0.01 100 Captafol 1 4.7 Folpex 8.6 0.1 THPl 10 4 .1 Phthalimide No result <0.001 Precision The within- and between-run precision, determined using three levels of captan water samples, assayed singly 10 times over five days, are shown in Table 2. ~~ ~ Table 2 Precision Mean (PPm) 0.10 0.59 2.31 Coefficient of variation within- run (YO) 19.73 9.51 11.31 Coefficient of variation between- 10.71 run (Yo) 16.01 8.77 Recovery The average recoveries of the assay, determined using seven samples spiked with captan levels of 1 and 5 ppm were 92 and 104%, respectively. Corresponding recoveries for the quick- screening apple juice method were 118 and 113%. All unspiked samples were negative.Correlation A comparison of the RaPID assay with the GC-ECD method' gave a correlation coefficient of Y = 0.9646. Discussion The current Canadian, USA and EEC maximum residue limits (MRLs) for captan on peaches are 5 , 50 and 2 pg g-', respectively (Pesticides Directorate, 1991). The RaPID assay method detects levels of captan at less than or equal to 5% of these tolerances. Similar findings apply to the usefulness of the procedure for apples. It was concluded that the RaPID magnetic particle immuno- assay technique has potential as a screening method to control pesticide residues in fruits, being ideally suited for analysing large numbers of samples in short periods of time. References 1 Herzog. D. P., Itak, J. A.. Selisker, M. Y.. Leister, W., and Fleeker, J .R., Ohmieron Corporation, 375 Pheasant Run. Newtown, PA 18940, USA. J . T. Baker Inc., 227 Red School Lane. Phillipsburg, NJ 08865, USA. North Dakota State University (Biochemistry Dept). Fargo, ND 58105, USA. 2 3 lnterlaboratory Comparability of Accuracy in Residue Analysis Nigel A. Smart Consultant, 5 1 Gilpin Green, Harpenden, Hertfordshire A15 5NP Intertaboratory comparability of analytical results is a part of general quality assurance. It is an important aspect of both official and company validation of analytical measurements. The European Community (EC) requires estimates of inter- laboratory precision for official Community methods. ’ Accuracy has become an omnibus term covering nearness of results to ‘true’ mean value whether deviations are: (i) a constant per cent.o r amount, often termed bias, or more recently ‘trueness’; and/or (ii) arbitrary variations of a random nature, perhaps described by a standard deviation as a measure of precision. These are, in outline, the current definitions in the British and international standards for statistical terminology.2 Accuracy is a function of the analytical pro- cedure or analytical system used, although it may refer to a laboratory as an organization. A method is a detailed procedure setting out all necessary information for an operator to obtain reliable results. This is a straightforward concept when using simple apparatus. With increasing complexity of instrumentation, and differences in performance of instruments, it is difficult to detail exactly parameters for setting up apparatus; laboratories use different protocols, although based on the same general principles. It is76 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 then appropriate to think in terms of the ‘analytical system’, the written procedure together with the instrumentation used.Interlaboratory comparability can be of two main types: (i) a collaborative study, in which the same detailed method is used in a number of laboratories on identical samples distributed from a common source; and (ii) a co-operative or check-sample study, in which identical samples are analysed but participating laboratories are free to use their own method. Regular co- operative studies have become known as proficiency testing. Co-operative studies compare analytical systems rather than analytical methods.Collaborative studies examine the same method(s) in different laboratories. Reasons for interlaboratory comparisons of accuracy in analysis of pesticide residues in water, food and environmental substrates include: (i) to obtain estimates of repeatability and reproducibility and, sometimes, of trueness; (ii) to enable the significance of series of results in surveys carried out in several laboratories to be evaluated statistically; (iii) for proficiency testing of individual analysts and laboratories at a professional level; (iv) to enable official approval of the competence, or otherwise, of laboratories for public screening, statutory or legal work; and ( v ) to examine candidate standard methods. EC and UK legislation on pesticides and increased public awareness of possible food contamination emphasize some of these requirements.The International Standard IS0 5725 Concepts of how to plan out and interpret results from interlaboratory work on pesticide residues methods have developed markedly since these started in the 1950s. The IS0 agreed International Standard 5725 on Precision of Test Methods, published in 1986 (replacing the intitial 1981 version), is the main reference document on setting up and evaluating collaborative studies. I S 0 5725 has again been revised and is being extended to set out the many practical applications of the estimates of precision, and trueness, obtainable from collaborative studies. Four parts of the new six-part standard are awaiting publication.The basic concepts in the collaborative arrangements and statistical analysis of the I S 0 5725 approach are: (i) the measurement method under investigation has been standard- ized in detail; (ii) the method is used under constant operating conditions; (iii) participating laboratories are representative of the spectrum of user laboratories; (iv) all the laboratories analyse identical samples; and ( v ) every single test result ( y ) is the sum of three components: where rn represents the general average (expectation), B represents the laboratory component of trueness under repeat- ability conditions, and e represents the random error occurring in every test under repeatability conditions. The data obtained by experiment are considered to form a single distribution.The statistical analysis deduces information on: possible aberrant or outlying values or laboratories; repeatability and reproducibility of the method; trueness of the method (when reference materials or a reference method are available); and, also, when interlaboratory tests are done at a sufficient number of levels of determinand, on the variation of parameters with level. The new revision of the Standard contains recommendations for use of graphical plots in addition to the rigorous calculation. Graphical presentations can provide a useful over-all view of the data, particularly to non-specialists in the type of analyses undertaken. The h and k consistency statistics of Mande13 are of particular interest. Programmes for computerized calculation of statistical parameters in I S 0 5725 from the raw data, as well as for testing normality of distributions of series of results, are available.The Analytical Methods Committee of the Royal Society of Chemistry has published a recommendation for evaluating y = r n + B + e t 1) results from co-operative studies.4 That Committee has, more recently, suggested use of robust statistics for dealing with data from collaborative studiess and has made a recommendation for carrrying out and evaluating proficiency testing.‘ Samples for Interlaboratory Comparisons Substrates should preferably have incurred residues for exam- ination in practical ‘field’ situations. Certain basic require- ments must be met for stability of both pesticide and substrate. Animal fats, vegetable fats, fish and grain have been suitable for interlaboratory studies in the short term. The Community Bureau of Reference (BCR) have milk powder and pork fat reference materials containing organochlorine pesticides.Rapid transport of frozen sample in dry ice has been used in check-sample programmes. Availability, drawbacks and use of samples containing pesticide residues for purposes of inter- laboratory comparison have been reviewed.’ Further suitable substances are clearly required, particularly for standardiz- ation programmes. Development costs are high and such work does not attract a high priority. Collaborative and Co-operative Work on Methodology in Various Countries Collaborative studies of pesticide residues methods for particu- lar chemicals have been carried out in the USA, UK, Denmark, The Netherlands and by the European Community.More recent collaborative studies have tended to examine multiresidue procedures, such as sweep codistillation, Luke’s procedure and the application of gel-permeation chroma- tography. The Association of Official Analytical Chemists (AOAC) continues to arrange collaborative studies of methods for pesticide residues analysis in the USA (and elsewhere) through its sub-committee E. The working group on pesticide residues of the Gesellschaft Deutscher Chemiker has carried out collaborative work on methods and, more recently, used check-samples to assess the capability of laboratories.’ There is a general trend towards increased use of check-samples for screening pesticide residues work. Three sets of results have recently been obtained from laboratories active in residues analysis in the UK, through the Food Science Laboratory FAPAS arrangements.The Water Research Centre, at Medmenham, monitors laboratory perfor- mance in determining pesticides in water through its AQUA- CHECK programme. Results from Joint Food and Agriculture Organization/ World Health Organization (FAOWHO) Food and Feed Contamination Monitoring Programme studies involving pesti- cide residues are published periodically. Accuracy Found in Practice In individual laboratories, relative standard deviations (RSDs) of results from analysis of residues of pesticides under repeatability conditions are usually at the lower end of the range 5-10%. Precision varies with the level of pesticide to be determined, as well as from substrate to substrate.Guidelines for analysts indicate a limit of determination should be about a quarter, or less, of the statutory residues limit and such sensitivity requirements are met in most regulatory laboratories. The over-all (reproducibility) coefficient of variation between laboratories assessed by collaborative study of meth- ods in the UK averages about 15% .9 Generally, differences and errors between laboratories are greater than those within a laboratory. The ratio of reproducibility to repeatability standard deviation probably averages about three. Relative values for bias are obtained each time an interlab- oratory trial is carried out. However, absolute values for the trueness of residues methods, especially as studied in a representative number of user laboratories, are almost non- existent. Small groups of laboratories, as well as individualANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 77 laboratories, have analysed the same sample(s) by more than one method, demonstrating the equivalence of some estab- lished procedures. Larger errors and variances than are indicated in the work cited have been apparent in the Joint FAONHO Food and Feed Contamination Monitoring Programme results. AQUA- CHECK results, where the level of pesticide analysed is much lower, show less precise estimates of chemical added. Conclusions The accuracy of methods and of analytical systems estimated through interlaboratory studies of residues analysis varies widely. The lack of a sufficient range of suitable standards, costs of setting up adequate check or reference samples for the range of substrates in which residues are found, and the non- standarization of methodology have all contributed to a relatively modest interlaboratory validation of analytical measurements for pesticide residues. References Council Directive (85/591/EEC) concerning the introduction of Community methods of sampling and analysis for the monitoring of foodstuffs intended for human consumption, Off. I . , 1985, No. L372, pp. 50-51. British Standard 5532, British Standards Institution. London, Part 1. 1978; I S 0 Standard 3534, International Standards Organization, Geneva, 1977. Mandel, J . , personal communication. Analytical Methods Committee, Analysr, 1987, 112. 679. Analytical Methods Committee, Analyst, 1989, 114, 1699. Analytical Methods Committee, Analyst, 1992, 117, 97. Smart, N . A., Residue Rev., 1985, 96, 1 . Thier, H.-P., Stijve, T.. and Specht, W . . Fresenius’ Z . Anal. Chern., 1989, 334, 546. Smart. N . A.. Analyst, 1984, 109, 781.

ISSN:0144-557X    

DOI:10.1039/AP9933000072

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

78 ANALYTICAL PROCEEDINGS, FEBRUARY 1993. VOL 30 New Directions in Chromatography The following are summaries of five of the papers presented at the Analytical Symposium of the RSC Annual Chemical Congress, held on April 13th-15thr 1992, in the University of Manchester Institute of Science and Technology. New Directions in Affinity Chromatography K. Jones Affinity Chromatography Ltd., Freeport, Ballasalla, Isle of Man When faced with the problems of separating and purifying proteins with little or no history biochemists inevitably turn to the well tried low-selective low-performance techniques of ion exchange, size exclusion and hydrophobic interaction. These traditional methods have one major attraction: when operated as a sequence a reasonable degree of success can be expected.That the use of such a sequence results in a very expensive process if scaled-up is of little immediate interest to the analyst. It is shown here that a more careful selection in the very earliest stages of investigation can eliminate this latent problem. Basic Mechanisms A common feature of all biological macromolecules is their ability to recognize and bind to other molecules, often in a highly specific manner. It is this binding ability which allows all proteins to be separated and purified by the affinity process. If only one of the basic intermolecular binding forces (selected from electrostatic, hydrophobic, hydrogen bonding and van der Waals) is utilized between the ligand and the protein of interest, then the medium can only exert a non-specific effect. The more interactions that are utilized the better the selectivity.When all four interactions are brought into play the uniquely selective process of affinity chromatography can be used to achieve one- step separations of all proteins from any source. By a combination of computer-assisted molecular modelling and extensive experimental experience it has proved possible to design geometrically perfect backbones which can be fitted exactly into the folds and fissures that are intrinsic in all protein structures. These frameworks can then be designed to carry the chemical entities which maximize the four binding effects into the appropriate areas of the protein. This results in very high specificity for the target protein. In contrast, non-specific media can only perform a limited part of the purification routine at each step, and several stages in series have to be utilized.Some 5040% of the total cost of a therapeutic protein is incurred at the purification stage’ and the replacement of a multi-step process by a single affinity step can have a revolutionary economic effect on product value. Scale-up As always the key to any successful chromatographic process lies with the medium and how well it performs. When large- scale strategies are considered the cost and chemical stability of the medium becomes of major importance even at the very earliest stage. There is little point in developing a successful analytical strategy if it is later found that the medium is unstable under depyrogenating conditions, is too expensive, has a short lifetime, is not available in bulk quantities and is not manufactured in a sterile environment.This latter factor is of little interest to most analysts but one which becomes of very major interest when scale-up is considered. Any process which will eventually require depyrogenation and sterility, immedi- ately excludes silica-based supports from further consider- ation. The favoured regenerationldepyrogenating reagent 1 mol I-’ NaOH destroys silica in minutes. Consequently, it has been accepted that although reversed-phase protein separations are now well established for analysis, alternatives to silica have to be reviewed for scale-up purposes. Of all available supports beaded agaroses in their various forms are favoured. Although having only low pressure capability they have a major advantage: a relatively low price.Many proteins occur only in very low concentrations in the mother liquor. This provides a unique platform for affinity chromatography. It can uniquely concentrate very dilute solutions whilst simultaneously stabilizing the adsorbed protein on the column. Despite these obvious advantages the develop- ment of affinity chromatography has been hampered by two problems: the high cost of current affinity media and the difficulties of making such media operationally stable in a multi-cycle pyrogen-free environment. These complications have been compounded by the extensive use of ‘natural’ ligands, i.e., other biomolecules such as cofactors and anti- bodies. Such ligands are difficult to identify, tend to be rare, difficult to purify, are catalytically and enzymically unstable and are generally very expensive (Protein A adsorbents cost upwards of f 5 per ml).Fortunately it has proved possible to overcome these difficulties by the use of novel synthetic structures which have their origins in textile dye chemistry. Natural versus Synthetic Affinity Ligands One of the major advantages of working in the affinity mode i s the vast range of potential ligands available, theoretically offering an infinite range of ligands capable of separating all proteins. However, of the many affinity ligands used over the last decades relatively few are commercially available. Militat- ing against the ‘natural’ proteinaceous ligands is their general instability and high cost.In contrast the accidental discovery that many different proteins could be separated by textile dyes led to the current development. Despite the success of dyes they did not prove to be the breakthrough so eagerly awaited. An essential feature of all chromatographic processes is the necessity for exact repeatability from column to column, year after year. Textile dyes are bulk chemicals, most of which contain many by- products which are co-produced at every stage of the dye manufacturing process. This fact alone makes reproducibility problematic. Of equal significance is that the bonding process between dye and matrix has been insufficiently researched. For these reasons, all commercially available textile dye-based ligands leak to varying degrees, the leakage rate beingANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 79 dependent upon the degree of impurity present and the bonding technology used.It is, however, clear that dyes do have an underlying power to separate a very diverse range of proteins. They also have sufficient structural complexity to enable extensive manipulation of their structures, providing an opportunity to enhance these basic properties. By making design changes and by simultaneously eliminating leakage the way was clear towards the creation of a new generation of stable and inexpensive ligands. These not only replace the current generation of unstable media but also greatly widen the scope of such media to separate and purify a very diverse range of proteins. After synthesizing many structures a range of broadly specific structures were selected and introduced as the MIMETIC range of ligands.Ligand Design New ligands can be created by designing in two ways: modifying existing structures, or by the total de novo design of new structures. The best possible opportunity for the design of de novo structures is presented when X-ray crystallographic structural data are available. Without such data initial targeting of specific binding groups is very difficult. Where X-ray data are not available, fragmentation data, produced for example by tryptic mapping, can be useful. In some instances substantial data are already available on an appropriate modelling database. By interchange of molecular modelling facilities such as Evans and Sutherland, the Brookhaven protein database and the interactive MACROMODEL (and SYBYL) pro- grams, protein models can be projected and manipulated to expose those segments which may be binding centres.Once a protein structure is identified and an appropriate backbone designed which inserts into a fissure, the necessary ionic, polar and non-polar groups can be incorporated and juxtaposed with equivalent groups on the protein of interest. There are six stages of development: design; synthesis; immobilization; characterization; in-depth testing; and prod- uct evaluation. However, the embryonic nature of this technology offers no guarantee that a suitable structure will result even after following this logical approach. Conclusions The difficulties of separating and purifying proteins is reflected in the very high proportion of total production costs incurred in the final stage of manufacture.It is widely accepted that the use of current separations media significantly contributes to these adverse costs, primarily caused by unavoidable yield reduction at each successive stage. Replacement of non-selective affinity media using ‘natural’ ligands is technically successful, but in general they are unstable under the harsh regeneration procedures commonly used, and are also uneconomic. Re- agents based on textile dyes have had some degree of success but suffer from certain disadvantages, such as leakage and lack of reproducibility. Until recently computer-assisted molecular modelling technology capable of creating designed synthetic highly selective ligands was unavailable.Once a synthetic stable state-of-the-art ligand has been selected by the use of the simple but rapid screening PIKSI module, which contains 10 MIMETIC ligands, powerful and sophisticated design techno- logies can redesign any of the base structures and custom- specified affinity media can be synthesized. References 1 Pfund, N. E.. and Charles, K. G., The Wheut from the Chaff, Hambrecht and Quist Report, USA, 1987. The Sievers SCD 350 Detector M. Dyson Dyson Instruments Ltd., Hetton Lyons Industrial Estate, Hetton, Houghton-le-Spring, Tyne & Wear DH5 ORH The Need for Sulfur Detection Sulfur is a very important element. It is responsible for many of the more exciting flavours and tastes in food and drink. Garlic, onions and horse-radish sauce derive their strong tastes from allyl sulfide; mustard oils contain allyl isothiocyanate. It is emerging that many beverages such as coffee and fruit juices owe their 'freshness' to the presence of volatile sulfur compounds; when these evaporate the drinks become stale.Beers, wines and spirits contain sulfur compounds which add to or detract from their flavour in appropriate concentration. Flavour is a two-edged sword: many of the repellant smells from decaying food are sulfur based, in particular, H2S, the smell of rotten eggs and stink bombs is well known. Deadlier repellants (mustard gas) containing bis(2-chloroethyl) sulfide exist for battlefield use. The pharmaceutical industry makes sulfur-containing drugs which are used in everything from ointments to medication and diet.Certain amino acids contain sulfur. Since natural gas has no smell of its own, sulfur containing mercaptans, the smelliest substances known, are added to North Sea gas by the Gas Boards to warn of hazardous gas leaks. The petroleum industry is dependent on the use of catalysts in feedstock reactions and these are sensitive to poisoning by sulfur compounds at very low levels, as low as 1 ppm. Sulfur compounds also tend to be reactive and corrode the pipework. Individual oilfields can be characterized by the profile of sulfur compounds in the crude naphtha from them. This arises out of the unique prehistoric flora which formed the subterranean oil reservoirs of those regions. Analysis of this profile in lumps of tar found floating in the oceans has helped to draw the noose around those who dumped it there.The analysis of sulfur compounds is therefore very important and the emergence of a new detector to detect them specifically has generated much interest. Description of the Sievers Sulfur Chemiluminescence Detector The sulfur chemiluminescence detector (SCD) is a mass sensitive detector' (Fig. 1) that can be fitted to all the major gas chromatographs (GC). It can be tuned for sensitivity or selectivity but it is usually used as a selective detector. The basic operation is that peaks emerging from a chromato- graphic column are burned in a standard flame ionization detector (FID), in a hydrogen-rich (reducing) flame. Those peaks containing sulfur compounds are burned to sulfur monoxide (SO) , which is strongly oxidized to chemilumi- nescent sulfur dioxide (SOZ") by ozone.The chemilumi- nescence is measured by a photomultiplier and is used as a measure of the sulfur peak. Residual ozone which would harm the vacuum pump (and any other corrosive species) is removed by a hopcalite trap.80 Regulator Filter air I Tr a nsf e I - gas " Hydrogen line FID j Gas I I restrictor u e Ozone nerato r Photomultiplier I! chromatograph 1 I ~ L . T . + 10 El Recorder ANALYTICAL PROCEEDINGS. Probe interface - assembly FEBRUARY 1993, VOL 30 Set screw A Mounting base for all GC models- except 8A For GC mpdel 8A \ Collector assembly __ Set screw B I I 1 I I i i Flush here '-- I Fig. 1 Schematic diagram of thc Sicvers SCD 350 Fig. 3 Shimadzu probc mounting bracket The Ceramic FID Probe The SO is removed from the FID and transferred into the ozone reaction cell down the vacuum line terminated by a silica tube positioned in the flame, 3-5 mm above the FID jet (see Fig.2). It is held in place by a bracket that fits on t o the top of the GC. There are different mounting brackets for different GCs. Fig. 3 shows one for the Shimadzu GCs. The ceramic probe is only intended t o be an inert, fireproof conduit that draws the flame contents. However, when there is excessive column bleed, SCD sensitivity can be rapidly lost over a few injections. This sensitivity is immediately restored by scraping and re-exposing the inside surface of the probe with a piece of wire. I t appears that there is some catalytic surface effect on SO abundance which is not yet fully understood.The catalysis is disabled when the surface is coated by column bled silica (because of the temperature involved it can hardly be anything else), but it is restored by exposing the original probe surface. Fig. 2 Diagram of SCD probc Before use, new probes are conditioned for a few hours by turning up the gas flows and burning in. The Chemistry of Sulfur Inside the SCD When sulfur compounds are burned inside an FID flame, the major products formed are SO?, SO, S?, HS and H2S.' The relative abundances are strongly dependent on the H2 : air ratio, which is set in the SCD t o maximize SO formation. S-compound + H2/02 - SO' H 2 0 + other products (1) + SO' + 0 3 - SO?" + 0 2 (2) SO?" - SO2 + hv ( 3 ) Reaction (2) is highly exothermic and excites the SO2 t o chemiluminescence even though the concentration of ozone in air is only about 2?40.~ The light emitted by the relaxation process has wavelengths in the region 260-480 nm, indicating several species of excited The maximum intensity is at 350 nm.The optimum flow rates to maximize SO formation are about 200 ml min-' of H2 and 400 ml min-' of air, which yields about 20% SO.4 These flows are tuned for individual GCs t o maximize the sulfur peak height. Of the two flows, air has the greatest effect on response. If there is insufficient air, selectivity suffers and hydrocarbon peaks are detected. If there is too much air the flame chemistry favours the direct formation of SO? which is not chemiluminescent (see Fig. 4). Note at this stage that the GC carrier gas, usually helium, does not interfere with the flame chemistry.Connection of the SCD t o high-performance liquid chromatographs (HPLC) and supercritical fluid chromatographs (SFC) will be mentioned later; in both applications there is evidence that the mobile phase perturbs the flame chemistry. Flame Ionization Detector Signal At these air and H2 flow rates the FID still functions as a detector but at about 10% of its optimized sensitivity.4 This remains a very usable signal and the analyst therefore has both FID and SCD signals from the same injection. Effluent splitting or dual column injections are unnecessary. Resistance to Signal Quenching The SCD signal is resistant t o quenching by co-eluting hydrocarbons and most common solvents if the residence timeANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 81 I 2 m c*l b) o? 1 c A 2 u__ I C) 2 h, Fig.4 Effcct o f air and hydrogen on response. Chromatogram of test mixture with low air ( a ) : with correct hydrogcn:air ratio (6); with cxccss of air ( c ) . I , Dimethyl sulfidc; 2, thiophenc of the S species in the flame is <7.5 ms,' though there may be some band broadening of the sulfur peak due to the co-elutant acting as part of the stationary phase. Quenching mechanisms are possible in principle. They would involve reactions to remove the SO radical formed in the FID flame before it can be transformed into chemiluminescent SO?, or they would be caused by a lack of ozone. The SO radical is quickly stabilized and its abundance preserved by transferring it into a vacuum (ca.10 Torr") where intermolecular collisions are much less frequent. This low pressure also prevents H 2 0 condensing in the line between the FID and hopcalite trap. The reaction between SO and ozone takes place in a cool cell remote from the end of the column and the FID. The ozone would be decomposed by the FID flame heat,' or it would react preferentially with hydrogen. * I Torr = 133.322 Pa Band Broadening in the Detector The large volume of the transfer line and ozone reaction cell has the potential for band broadening by simple volume dilution or by adsorption of SO on to the walls. Fast transfcr flow rates from the FID to the cell offset both effects but there is still some peak broadening4 and this has been reported to be larger than that of a flame photometric detector (FPD).s Fig.5 shows the overlay of the same peak detected simultaneously by FID and SCD. The broadening is acceptably small. Key Properties of the Detector Sensitivity When tuned for sensitivity the SCD can detect down to levels as low as 4 pg s-l of S which is equivalent to about 0.2 ppb (by volume) or 10-20 pg of sulfur compound on column at S/N = 3. Typical limits of detection are shown in Table The SCD is up to 100 times more sensitive than the FPD. Selectivity When tuned for selectivity the response against hydrocarbons is greater than lo'. Even when tuned for sensitivity the selectivity is still about lo".' Fig. 6 shows no SCD response to hydrocarbons in reformate or against alkenes. Fig. 7 shows no response against terpenes and other natural organics in lime oil.120 100 > & 80 2 60 5 40 20 > c. .- a, c I I I I I I I 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 Tirneirn i n FID fi 4.50 4.55 4.60 4.65 4.70 Tirnehin Fig. 5 Band broadening of ethyl disultide.' ( a ) SCD response to 8 ng ethyl disulfide jn hexane; ( b ) FID and SCD rcsponscs to 8 ng cthyl disulfidc on an expanded time scale Table 1 Limits of detcction (ppb) FPD SC D Hydrogen sulfide 130 3 Carbonyl sulfide 90 3 Methanethiol 240 4 Ethanethiol 290 3 Dimethyl sulfidc 120 2 Dimethyl disulfide 180 282 ANALYTICAL PROCEEDINGS, FEBRUARY 1993. VOL 30 Linearity The SCD has been shown to be linear over three decades of response.',437 For example, with H2S Fig. 8 is obtained. Interestingly, if calibration graphs are related to the molar sulfur content of the species rather than to the species itself, a common calibration graph can be drawn' as in Fig.9. Relative Response Factors The common calibration graph shown in Fig. 9 arises because all sulfur species are converted to SO' and the sulfur is detected by means of the oxidation of SO' to SOZ". The first reaction strips away intermolecular differences, the second reaction is thus similar for all species and so their molar responses are similar to each other (see Table 2). This factor has great analytical importance. It means that sulfur compounds can be estimated without knowing their identity (although assumptions are made about the number of S atoms per molecule). Like using an FID the relative response factors can all be assumed to be equal to 1 (*lo%), at least to (a) SCD " 2 3 4 I ,i I I I I I 0 5 10 15 Timelmin Fig.6 Selectivity against hydrocarbons in reformate ( a ) and alkcncs ( h ) . Benzene (1); toluene (2); xylcncs (3); 1,2,4-trimcthylbcnzenc (4); 1,2,3,S-tetramethylbenzene (5); naphthalene (6); 2-mcthylnaphtha- lcne (7); I-methylnaphthalene (8) SCD 0 5 10 15 20 25 Timelmi n Fig. 7 Selectivity against hydrocarbons in lime oil begin with. Fig. 10 shows chromatograms of sulfur standards. Note the annotated quantities: they can be related by eye to the peak sizes. Comparison of SCD with Flame Photometric Detector Inevitably, the SCD is compared with the flame photometric detector which it seeks to replace. Table 3 summarizes the main differences. Selectivity When tuned for selectivity, the SCD response is greater than lo7 better than hydrocarbons.Even when tuned for maximum sensitivity the selectivity is still about LO6. 12 10 8 al c 0 $ 6 2 m $ , 2 000 000 000 000 000 000 000 000 000 000 000 000 1 I I I I I 0 100 000 10 000 (TJ : 1000 100 50 100 150 200 250 300 350 Hydrogen sulfide/ng Linear response of H2S Fig. 8 0.001 0.01 0.1 1 10 Amount S lng Calibration graphs for pesticides: dursban (0). malathion (@). Fig. 9 EPTC (0) and cycloate (W) Table 2 Relative SCD response factors Compound Thiophenc Methyl disulfidc 2-Methylthiophcne 3-Mcthyl thiophene 2-Ethylthiophenc 2.5-Dimcthylthiophene Bcniothiophcne Di benzothiophenc 0.92 0.95 0.98 1 .oo 1.01 1 .05 1.07 1.14ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 83 The selectivity of the FPD is not so clear cut.Response can be about 50000 times better than hydrocarbonsy but it varies with resolution, i.e., the signal can be quenched. It is also a function of S concentration, the smaller the concentration, the less the selectivity, which is a nuisance when determining trace quantities. * O Sensitivity The SCD is about 100 times more sensitive than the FPD in sensitive mode and offers about the same or slightly better sensitivity in selective mode. Linearity The response of the SCD is linear over three decades. The FPD has nominally a quadratic response but this varies with the species and with dissociation of the S2 molecule because of detector conditions. In practice the response can vary from linear to quadratic. ") Relative Response Factors The relative response factors are similar for most S com- pounds.This means that, as with an FID trace, observed peak 0 10 C '2 20 . Q, I- E .- 30 40 rM-2 (AMOCO) Standard METS-3 Standard 37 PPm b 32 PPm 30 PPm d h 32 PPm 23 PPm 53 PPm 7 2 0 PPm MES-2 Stand; E- w d Fig. 10 Chromatograms of sulfur standards: a , carbon disulfidc; b, thiophcne; c, 3-methylthiophcnc; d , n-propyl sulfide; e , isopropyl disulfide; f , n-propyl disulfide; g, bcnzothiophene; h, n-decancthiol: i , n-pcntyl disulfide; j, phenyl sulfide; k , dibenzothiophene; 1. n-octyl sulfidc; m, n-decyl sulfide; n, n-dodecyl sulfide; 0. methancthiol; p. cthancthiol; q, dimcthyl sulfide; r, methyl ethyl sulfide: s, dicthyl sulfide; t , dimethyl disulfide; u, 2-ethylthiophene: v. unknown; w, dicthyl disulfide; x, bcnzencthiol (impure standard); y, from bcnzencthiol Table 3 Comparison of SCD and FPD Fcaturc SCD Sclcctivity > 107 Limit of dctection Lincarity Linear.1000-fold 2 pg (app,rox.) 400 fg s- Response factors Predictable response similar resp. factors Qucnching Insensitive to C02 and hydrocarbons S-FPD 50000 (approx.) but varies with resolution and concentration 200 pg (approx.) Non-linear, approx. quadratic Unpredictable. Can't quantify without solute identify Quenched by C02 and hydrocarbons 50 pg s-' size is a good indication of solute quantity. The analyst can quantify solutes without knowing their identity. Relative response factors for FPD measurement are all different' and unless the analyst knows the solute's identity, he (or she) cannot apply the appropriate response factor and so cannot quantify unknown solutes.Observed peak size is not a reliable guide to solute quantity. Quenching The operation of the SCD protects the flame reaction mechanism for quenching when used with a GC. The co-eluting mobile phases of isocratic HPLC and SFC require additional compensation from the FID gas flows but otherwise the SCD can resist quenching here too. The single-flame FPD is notorious for quenching. In dual- flame mode the FPD resists quenching and is more truly quadratic in response but there is a trade off, it loses 8690% of its sensitivity.' Detector for Supercritical Fluid Chromatography and High-performance Liquid Chromatography Many sulfur compounds are either thermally unstable or non- volatile and are not suitable for analysis by GC; successful attempts have been made to couple the SCD to SFCs" and HPLCs13 but on both occasions the mobile phase affects the chemistry inside the flame.High-performance Liquid Chromatography Fig. 11 depicts the combination of an SCD with HPLC.13 Isocratic Analysis The SCD response is sensitive to and varies with mobile-phase concentration (in Fig. 12, methanol) which alters the flame chemistry. The effect can be offset by increasing the air flow rate sufficiently to meet the methanol demand and optimize SO formation (see Fig. 12). Since the air flow rate cannot be programmed, applications have been limited to isocratic analyses. Optimum Flame Ionization Detection The column effluent is fed into an FID through a short length of deactivated fused silica tubing (see Fig.11). There is an optimum FID temperature (see Fig. 13): if it is too high it will evaporate the mobile phase and precipitate solutes inside the 14 - Fig. 11 Use of SCD with HPLC. 1, Helium gas; 2, HPLC pump; 3, gas chromatographic oven; 4, flamc ionization detector; 5 , flamc sampling probe; 6, transfer line; 7, oxygen; 8, ozone generator; 9, reaction cell; 10. vacuum pump; 11, chemical trap; 12, photomultiplier tube; 13. integrator; 14, SCD 350 main body84 I ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 transfer line. If the FID temperature is too low, the mobile phase can alter the flame chemistry and reduce SO formation. When the detector temperature is set optimally, vaporization and aerosol transport efficiently transfer the solute from column to FID.The quenching effect from a cool FID can be offset by increasing the hydrogen flow rate, but at optimum flow the response is less than when the FID is at its optimum temperature. Examples of HPLC analyses include the analysis of thiocar- bamates (Fig. 14) and the analysis of phenyl(thiohydant0in) (PTH) amino acids (Fig. 15). Supercritical Fluid Chromatography The SCD has been used with capillary SFC. Detection limits of about 12 pg of S were achieved together with a selectivity versus hydrocarbons >lo7 and a linear dynamic range of lo3. Signal quenching has been observed with decompressed flow rates of C02 above 5 ml min-'. Increased quenching (com- pared with pure COz) is observed with 2% methanol modified C 0 2 at decompressed flow rates above 4 ml min-'.This limits the maximum column size used so far to 250 pm i.d. (packed capillary SFC). Column Selection and Preparation For Gas Chromatography Analysis methods developed in gas chromatography for the FPD have had to acknowledge the quenching of sulfur peaks by hydrocarbons and other species. Useful columns are limited to those which isolate the sulfur peaks from those others which perturb measurement. With the SCD, this restriction is removed. In GC-SCD, capillary columns are preferred over packed for two basic reasons: (i) the quality of manufacture of capillary columns compared with packed columns reduces the amount of 14 1 0 250 270 290 31 0 330 Air flow rates/ml min 1 Fig. 13 350 and B, 58 "C SCD sensitivity versus air flow rate at FID temperature of A, bleed to desensitize the ceramic probe (though conditioning is nevertheless essential to reduce the bleed to a minimum); (ii) the most useful analyses tend to be complex substances such as naphtha, food extracts, etc., and these are normally chromato- graphed on WCOTs anyway.The best combination of sample capacity, adequate resolu- tion and reasonable analysis times is obtained by using a 30 m column with 0.32 mm internal diameter, and 1 pm film thickness of bonded methyl silicone phase to minimize bleeding effects. Columns must be very thoroughly conditioned too. It is best 3, 4 6 I 0 a Fig. 14 HPLC/SCD Ti me/m i n Analysis of seven thiocarbamates by packed capillary column 1 2 3 4 6 I 0 16 Ti m e/m i n Fig. 15 Analysis of PTH amino acids by packed capillary column HPLC/SCD.1 , Alanine; 2, proline; 3, methionine; 4, phenylalanine; 5 , leucine; 6, norleucineANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 85 41 I n-C4 n-C5 / ( ,n-C6 il I T - ~ ~ FID 0 10 20 30 40 Time/m i n Fig. 16 3 . bcnzothiophene; 4, dibenzothiophene SCD analysis of crude oil. 1 , Carbon disulfide; 2 , thiophene; to avoid those stationary phases that bleed a lot, either by their volatile nature or their high-phase loading or at the selected operating temperature. For Supercritical Fluid Chromatography and High-perform- ance Liquid Chromatography The quenching of peaks which can occur in SFC and HPLC due to the mobile phases attenuating the flame chemistry is clearly reduced by having smaller flows of mobile phase. If these smaller flows are to preserve optimal separation performance of the columns, the columns must be capillary or megabore.Packed columns of larger diameters would lack resolving power at flow rates optimized for detector performance. Analytical Applications Petrochemicals Petrochemicals is a major application area for the SCD. Sulfur compounds present in the original crude oil (see Fig. 16) poison catalysts in feedstock reactions. Early versions of the SCD were unable to detect some inorganic species such as H2S and COS; however current Sievers production models can. Foods, Flavours and Drinks This is a major application area which is yet to develop fully. Examples of products for analysis include coffee, chocolate and galbanum oil (Fig. 17). Galbanum oil is used as a fixative in the perfumery business and in formulations to flavour ice creams, sweets, etc.Wine Adulterants Recently some major stores and a number of wine merchants here and in Germany had to withdraw wines from Northern Italy when traces of methyl isocyanate were found to have contaminated them. Fig. 18 shows FID and SCD traces for this contaminant. ’‘ Two hop oils have similar FID traces but are seen to be clearly different (Fig. 19) when their sulfur profiles are compared. The determination of volatile sulfur compounds is also of interest to the brewing industry because of their effect on flavour even at very low levels. They also indicate problems - 1000 mV Standards SCD - 500 mV Galbanum oil SCD JL n -250 mV FI D 0 5 10 15 20 25 Time/min Fig. 17 enethioate; 2.S-sec-butyl-3-methylbut-2-enethioate SCD analysis of galbanum oil. 1, S-isopropyl-3-methylbut-2- _. .. .. t . , I . I ... I . . ” 8 L L I ’ ’ SCD JAALn- Fig. 18 in Italian wine SCD analysis of wine Contaminant (methyl isocyanate, MISC) during fermentation. Fig. 20 shows sulfur compounds in beers and lagers.ls Other Applications Other application areas for the SCD include the tobacco industry and pesticides. New Developments At Pittcon 1992 two new products were launched. Model 355 Flameless Burner This is a self-contained detector that does not utilize the FID; it is an alternative to the SCD 350. Flame gases mix with column86 FID ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 0 0, C 0 v) 2 L c 0 a c 0" Sample 1 ,J SCD 350 Sample 2 SCD 350 Samples - Sulfur detection - Data processing- Results 10 20 30 40 Ti me/mi n Fig.19 SCD analysis of two hop oil samples Fig. 21 Diagram of SCD 357 can detect and measure S compounds in places and at levels not previously possible. C D A 10 Ti me/mi n 0 Fig. 20 Sulfur compounds in beers and lagers. A , H,S; B, CH,SH; C, dimethyl sulfide; D, ethyl methyl sulfide (internal standard); E, CH$SCOCH3 effluent between two sealed, concentric silica tubes and the reaction products are drawn into the ozone chamber down the vacuum line. Sensitivity is increased by a factor of five to 10. Linearity is extened to five decades. In the absence of the FID it is easier to use. Model 357 On-line Total Sulfur Chemiluminescence Analyser (TSCA) The total sulfur chemiluminescence analyser (TSCA) is a stand-alone process control instrument for measurement of total sulfur in C02 or other inorganic gases (see Fig. 21).The analyser uses the 355 detector and is aimed at the fizzy drinks market and their C02 suppliers. The Environmental Market The environmental market for the SCD will be legislation driven. New applications will emerge as analysts realize they Conclusion The sulfur chemiluminescent detector arrives at a time when the world is becoming aware of the importance of trace quantities of sulfur in biochemistry: everything from taste to metabolism, and as concern grows about the element's role in environmental pollution. The SCD has a much better performance than its rival the flame photometric detector. It is easier to use than the FPD and has wider applicability.It is therefore expected to displace the FPD in due course. The market for the SCD is much greater than that of the FPD and an increase in the analysis and study of sulfur compounds made possible by this new detector is confidently predicted. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 References Benner, R. L., and Steadman. D. H., Anal. Chem., 1989, 61, 1268. Farwell, S. 0.. and Barinaga, C. J., J. Chromatogr. Sci., 1986. 24, 483. Hutte. R. S . , Johansen, N. G., and Legier, M. F., J . High Resolut. Chromatogr., 1990, 13, 421. Shearer, R. L., O'Neal, D. L., Rios, R., and Baker, M. D . , J . Chromatogr. Sci., 1990, 28, 24. Gaines, K. K., Chatham, W. H., and Farwell, S . O., J. High Resolut. Chromatogr., 1990, 13, 489. Technical Note on SCD 350, Sievers.Johansen, N. G., Plams, M., and Legier, M. F., Brewers Digest, April 1990, 36. Johansen, N. G., personal communication. Guiochon, G., and Guillemin, C. L., Quantitative Gas Chro- matograph y , Elsevier Journal of Ch romatography Library, Elsevier, Amsterdam. 1988, vol. 42. Dressier, M . . Selective Gas Chromatographic Detectors, Elsevier Journal of Chromatography Library, Elsevier, Amster- dam, 1986. vol. 36. Varian (UK) Ltd., personal communication. Chang, H. C. K., and Taylor, L. T., J . Chromatogr., 1990,517, 491. Chang, H. C. K., and Taylor. L. T., Anal. Chem., 1991, 63, 486. MacNamara, K., Irish Distillers Ltd., Dublin, personal comm unica t ion. Burmeister, M. S., Drummond. C. J., Pfisterer, E. A., Hysert, D. W., Sin, Y. O., Sime. K. J., and Hawthorne, D.B., Technical Note, Molson Breweries, Ontario, Canada, 1991.ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 87 Geochemical Applications of Pyrolysis Gas Chromatography with an Atomic Emission Detector Steve Rowland," Roger Evens, Les Ebdon and Andy Reest Department of Environmental Sciences, University of Plymouth, Drake Circus, Plymouth, Devon PL4 8AA Although gas chromatography-atomic emission detection (GC-AED) is a long established analytical technique (see ref. 1 for a review) it is only the recent availability of a commercial instrument that has seen the widespread application of the method.' The multi-element (>25 elements) nature of the detector, which is based on the principles of microwave- induced plasma spectroscopy with photodiode array detection, is suitable for application to a wide variety of environmental analytical problems, such as pesticide analysis and the screen- ing of water and biota for organometals (e.g., tributyltin derivatives).'-5 The method promises to be an equally useful tool in other areas; for example a recent article described the application of GC-AED to the analysis of metalloporphyrins in crude oils, with selective analysis of iron, vanadium and nickel species.' In py-GC-AED, use of a pyrolysis (py) probe as the GC sample introduction method further extends applicability to the analysis of polymeric material^,^.^ including geo- polymers."-' ' Indeed, petroleum geochemistry in general is a challenging area to which GC-AED should be applied.This paper summarizes some of our preliminary experiments in this field.Experimental Experiments (py-GC-AED and py-GC-MS) were performed with a CDS 120 pyroprobe fitted with a platinum coil. Dry, powdered samples (<200 pg, accurately weighed) were placed in quartz tubes blocked with quartz wool and pyrolysed for 20 s at a maximum temperature of 610 "C. The pyroprobe was introduced directly into the GC injector ports." GC ovens were cooled with liquid carbon dioxide to -40 "C. GC columns were: for GC-AED, 50 m X 0.3 mm HP-1 'Ultra'; for GC- MS, 30 m x 0.3 mm DB-1 (J & W). Helium was the carrier gas and oven temperature was programmed from -40 "C (held for 5 min) to 300 "C (held for 15 min) at 5 "C min-'. The GC- AED instrument was a Hewlett-Packard 5890A interfaced with a HP5921A AED via a HP203 heated transfer line.Helium was the plasma support gas. The GC-MS instrument was a Carlo Erba 'Mega' chromatograph coupled to a Kratos MS 25 spectrometer. Electron impact spectra (mlz 19-500 u) were acquired at 38 eV, 400 FA emission current. Results and Discussion The geochemical samples chosen for study were organic-rich Pleistocene, diatomaceous sediments collected from 11" 03.90'S, 78" 04.67'W in the upwelling zone off Peru.".13 Samples were collected as part of the Ocean Drilling Program Leg 112 and spanned a depth range of 1-22 m below sea floor. Interest in the chemistry of the macromolecular organic matter in these sediments results partly from the similarities of the present day upwelling environment to that which is thought to have resulted, in palaeotimes, in important oil producing strata such as the Monterey formation of California.The petroleum industry has particular interest in the fate of the sulfur- containing macromolecules'5 and the py-GC-AED/MS tech- niques are particularly well suited to an investigation of such compounds. * 'To whom correspondence should be addressed. t Present address: National Rivers Authority, Rivers House. St. Mellons. Cardiff, Wales. 5000 4000 3000 2000 1000 0 0 10 20 30 40 50 60 Ti m e/m i n 100 (b) 200 400 600 7:51 15:45 23 : 39 Scan real time/min : s Fig. 1 ( a ) Pyrolysis-GC-AED (S emission line 181 nm) of diato- maceous Pleistocene sediment from Peru upwelling zone. ( b ) Pyroly- sis-GC-MS (summed mass fragmentogram mlz 34,48,76,84,98,112) of diatomaceous Pleistocene sediment from Peru upwellling zone.R.T. = retention time. For analytical conditions, see text. Note close correspondence between AED and MS data allows identification of even the most volatile pyrolysis products Pyrolysis of five Leg 112 sediments produced very complex distributions comprising over 400 different pyrolysate products ranging from gases (e.g., methane, carbon dioxide) to com- pounds up to C3(). The cryogenic cooling facility ensured that even the most volatile products were trapped and that most were resolved by GC. Comparisons of the py-GC-AED carbon emission line chromatograms showed that all the samples were similar to one another, as expected for sediments which have received a fairly uniform input of organic matter, and similar to those of the ancient organic matter of the Monterey kerogen.l3 However, the complexity of the py chromatograms was greatly reduced when the selectivity of GC-AED was used to monitor sulfur-containing pyroproducts (Fig. 1). Indeed, the88 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 complementary use of py-GC-AED sulfur detection and py- GC-MS mass fragmentography allowed most of the sulfur- containing compounds to be identified on the basis of retention index and mass spectral interpretations. The major products included HZS, methanethiol, carboxysulfide and carbon disul- fide with minor amounts of thiophene and CI-3 alkylated homologues (Fig. 1). Neither method alone would have allowed these identifications to be made, since GC-AED provides only confirmation of the presence of sulfur (emission spectra) and retention index data, whilst GC-MS mass fragmentography is impractical without some prior indication, from py-GC-AED, of the expected distribution and number of components.The data clearly show the usefulness of GC- AED as a screening A number of diagenetic changes in the sulfur-containing pyroproducts were revealed by GC-AED and will be the subject of a more detailed report. This has already proved possible for the nitrogenous products in these samples.’ ’ The data obtained thus far are at best only semi-quantitative. However, as LarterIs and others have shown, py-GC data can yield valuable quantitative results if attention is given to methodology. Figues of merit are available for py-GC-AED’ so the method should be useful for quantitation.For all the py- GC studies a CDS 120 pyroprobe was used with a coiled platinum filament. With care, samples could be weighed on a six figure balance and reweighed after pyrolysis to give semi- quantitative information. This was not possible, we found, if a platinum ribbon was used. We were concerned too that selective pyrolysate losses might occur in the commercial py- GC interface box and for this reason we employed a direct insertion py-GC technique.” We found that both GC-AED and GC-MS injector ports allowed a snug fit for the pyroprobe and that suitable lock nuts could be milled from aluminium or brass with graphite tape as a gas-tight ferrule. The GC column could then be brought close to the end of the pyrolysis filament for maximum transfer of pyrolysate, with the injector port swept with carrier gas and heated in the normal way.A slight crimping of the platinum filament was usually sufficient to retain the sample holder without adversely affecting pyrolysis. When a known amount of a man-made polymer (poly-t- butylstyrene) was mixed with the sample, the resulting GC- AED carbon emission chromatogram continued the t-butyl- styrene monomer, against which carbonaceous components could be measured (cf. ref. 15). However, in order that the equally important products in the sulfur and nitrogen emission py-GC-AED chromatograms could be quantified, man-made polymers producing, respectively, sulfur and nitrogen-contain- ing pyrolysis products (e.g., monomers) in a reliable and reproducible way, are needed.We addressed this problem by investigating the py-GC-AED behaviour of polyphenylene sulfide (PPS; Aldrich) and polyvinylpyrollidone (Aldrich) . Neither proved suitable as an internal standard, although the data do show the general applicability of the approach. For example, Fig. 2 shows the pyrolysis products (sulfur emission chromatogram) resulting from py-GC-AED of PPS. Whilst the complexity of components is too great to allow the use of PPS as an internal standard, the major product (tentatively identified as benzenethiol) does have a suitable retention index, response and general GC characteristics for quantitation. ’ Extension of this approach to other polymers promises to be a major contribution to the development of multi-element py- GC-AED as a quantitative geochemical analytical tool.Conclusion Atomic emission is a sensitive multi-element specific detection system which has great promise for py-GC studies of organic 12 000 10 000 8000 6000 4000 2000 0 0 10 20 30 40 50 60 70 80 Time/rn i n Fig. 2 Py-GC-AED (S emission line 181 nm) of polyphenylcne sulfide. Chromatogram indicates unsuitability of polymer as an S- containing internal standard due to complexity of pyroproducts geochemical samples but py-GC techniques require some modification before they can be used to obtain reliable quantitative data. Elimination of the pyroprobe-GC interface resulted in no depreciation of the quality of py-GC-MS and py-GC-AED data and facilitated transfer of pyrolysate to the GC column. Morc polymers need to be examined by py-GC-AED and py-GC-MS in order to provide suitable heteroatomic internal standards for pyrolysis. The authors are grateful to the Natural Environment Research Council (NERC) for a postdoctoral research associateship (to A.W.G.R.) and to Hewlett-Packard and British Petroleum International for, respectively, provision of, and funding for, GC-AED. We thank A. Douglas, Newcastle Research Group, and P. Lyne, BP Analytical Division, for t h e loan of CDS pyroprobes. 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 References Uden, P. C.. Young. T., Wang, T., and Cheng. Z., J . Chroma- togr., 1989, 468, 319. Wylic, P. L., and Quimby, R . D., J . High Reyolut. Chroma- togr., 1989, 12, 813. Ebdon, L., Even$, R.. Hill, S. J., and Rowland, S. J . , Abstr. Pap. Am. Cfzern. Soc., 1990, 199, 193. David, F., and Sandra, P., Proc. l3th Syrnp. Capillary Clirornatogr. ( A h . ) , 1991, 1453. David, F., and Sandra, P., Proc. 13th Symp. Capillary Chromatogr. ( A h . ) , 1991. 1464. Quimby. B. D., Dryden. P. C., and Sullivan, J. J . , J . High Resolut. Chrornatogr., 1991. 14, 110. Riska, G. D., Estes, S. A., Beyer, J . O., and Uden, P. C., Spectroclrim. Acta, Part R, 1983, 38, 407. Oguchi, R., Shimizu. A., Yamashita, S . , Yamaguchi, K.. and Wylic, P.. J . High Redut. Chrornatogr., 1991, 14, 412. Sceley, J . A., Zeng, Y ., Eglinton. T. I.. Ericsson, I . . and Uden, P. C., J . Anal. At. Specfrom., in the press. Sinninghe DamstC, J . S . . Eglinton. T. I., and de Leeuw, J . W., Geoclrim. Cosmochirn. Actu, 1992, 56, 1743. Patience, R. L.. Baxby. M., Bartle, K. D . , Perry, D. L.. Rees, A.W. G., and Rowland, S. J.. Org. Geocliem., 1992. 18, 161. Whiton, R., and Morgan, S . , Anal. Cliem., 1985, 57, 778. Patience. R. L., Clayton, C. J., Kearsley. A. T., Rowland, S. J., Bishop, A. N., Rees, A. W. G., Bibby, K. G., and Hopper, A. C., Proc. ODPSci. Results, 1990, 112, 135. For a review see, Geochemistry of Sulfur in Fossil Fuels, ACS Symposii~m Series, eds. Orr, W. L., and White, C. M., American Chemical Society, Washington, 1990, p. 429. Lartcr, S . , in Petroleum Geochemical Exploration of the Norwegian SMf, Graham and Trotman, London, 1985. pp. 269-286.

ISSN:0144-557X    

DOI:10.1039/AP9933000078

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

89 ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 Supercritical Fluid Chromatography; Recent Developments and New Directions Hans-Gerd Janssen and Carel A. Cramers Laboratory of lnstrurnental Analysis, Eindhoven University of Technology, P. 0. Box 513, 5600 MB Ein d ho ven, The N e th e rlan ds Within analytical chemistry, chromatography is by far the most widely used analytical technique. Gas chromatography (GC) and (high-performance) liquid chromatography (HPLC) have gained widespread acceptance in numerous application areas. As both gases and liquids can be used as the mobile phase in chromatography, extending the range of mobile phases to the supercritical region is but a logical step. Since supercritical fluids combine many characteristics of gases and liquids it is not surprising that SFC can be seen as an intermediate technique between GC and HPLC.Potential advantages of (carbon- dioxide based) SFC in comparision with LC include the compatibility with various GC detectors and the increased speed of analysis. In comparison with GC, SFC is advan- tageous for the analysis of high relative molecular mass or thermally labile components. History of SFC SFC is by no means a new technique. The first experiments using supercritical fluids as the mobile phase were performed by Klespcr, Corwin and Turner as far back as 1962,’ well before the introduction of HPLC. After the initial period of interest in SFC in the 1960s, the progress of SFC slowed down. The developments in SFC doubtlessly continued but clearly did not reach the exponential development curve that is character- istic for the development of new techniques and methods.In part, the slow development was due to early experimental problems, the lack of commercially available instrumentation and the fact that SFC development was overshadowed by the simultaneous development of HPLC and capillary GC. In the 1970s, SFC was in a dormant state. Research interest in SFC was limited. A strong revival of the interest in SFC occurred in the early 1980s. Two important aspects were the introduction of the first commercial instrument by Hewlett- Packard and the introduction of open-tubular columns in SFC by Novotny rt af. in 1981.’ From then on SFC developed along two lines, i.e., the old line of packed columns and the newer line of open-tubular columns.Packed and Open-tubular Columns After the introduction of open-tubular columns, considerable debate arose on which of the two column types should be preferred for SFC. Later, the consensus was reached that both column types have their own unique advantages and disadvantages. In general, open columns possess a high efficiency. As diffusion in supercritical fluids is much slower then in gases, the inner diameter of open columns in SFC has to be much smaller than in GC. Typically, 50 pm open columns are used. The use of these narrow columns imposes severe restraints on the instrumentation. Injection and detection are highly critical. Apart from the limitation imposed by extra column band broadening, the sensitivity of the detection device requires special consideration due to the strongly reduced sample capacity of narrow-bore open columns.Packed columns are generally much easier to operate. Furthermore, columns packed with sub-10 pm particles are more time efficient than contemporary open columns. A fundamental problem of packed columns in SFC is the inherently high pressure drop which limits the maximum obtainable plate number. The choice of the column type in SFC is determined by a number of parameters. The most important of these are the required plate number and analysis speed, sample loadability , detection limits and the injector and detector compatibility. Packed columns are superior over open-tubular columns with regard to the speed of analysis, the sample capacity, the injector compatibility and the detection limits.Open columns are to be preferred in terms of the maximum obtainable plate number. In addition, open columns are more favourable for combination with various detectors because a large variety of components can be eluted with pure carbon dioxide as the mobile phase. Instrumental Developments Numerous improvements have been published on various aspects of instrumentation for SFC. As instrumentation for open-tubular SFC is far more complicated than for packed- column SFC, technological improvements have centred on open-tubular SFC. Areas of special interest were injection techniques and detector couplings. Nowadays, a wide variety of injection devices is available for open-tubular SFC. Flow split, timed split and combined split injection techniques allow the introduction of nanolitre sample sizes on to open columns with inner diameters below 50 pm.3 Total injection without splitting enables the introduction of several microlitres of sample4 but is not yet applicable in routine analysis.Further research is needed to develop easy to operate and reliable systems for the introduction of large sample sizes in SFC. Over the past several years a number of fixed restrictor designs have been develo ed. In particular, the polished ‘integral’ tapered restrictor and the frit restrictor‘ are now widely used. Variable restrictors have been described.’ These systems are, however, not yet applicable in daily practice. Progress in detection techniques for SFC has been extremely rapid. Existing detectors, such as the ultraviolet (UV) detector and the flame ionization detector (FID), have been further developed and adapted to suit the specific requirements of SFC.Low-volume detection cells have been developed for UV detection and photodiode-array UV detection in open-tubular SFC. Other detectors which have proved extremely useful in GC, such as the electron capture detector, the nitrogen/ phosphorus detector and the flame photometric detector, have been introduced in SFC. Powerful identification possibilities for unknown compounds are provided by the compatibility of SFC with mass spectrometry (MS)8 and Fourier transform infrared spectroscopy (FTIR).9 Research work in the future should focus on the development of more sensitive and modifier-compatible detectors. P Mobile and Stationary Phases During recent years a large number of potential mobile phases for SFC has been thoroughly investigated.Key features in these studies were: (i) the applicability of the mobile phase for the elution of polar solutes and (ii) the detector compatibility. Despite the problems often experienced when trying to elute polar solutes, carbon dioxide is still, by far, the most widely used mobile phase in SFC. In open-tubular SFC, fairly polar solutes can be eluted with pure C 0 2 as the mobile phase. In90 ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 packed-column SFC, however, most of the separations require modified mobile phases. The development of stationary phases for packed-column SFC has concentrated and will continue to concentrate on preparing more homogeneous materials which exhibit a reduced silanol influence.Although still far from perfect, polymeric stationary phases are definitely an improve- ment over conventional hydrocarbonaceous packing materials. ‘O Open-tubular SFC clearly benefited from the progress in GC column technology. Nowadays a wide variety of stationary phase material is available. The selectivity can be optimized by choosing from a series of stationary phases with various polarities. In the case of extremely complex samples, multi- dimensional SFC with series coupled columns of different selectivities provides enhanced separation power.” The use of mu1 tidimensional chromatographic techniques is expected to increase in the future owing to the ever increasing complexity of the samples to be analysed.Applications Despite the potential advantages of SFC listed above, the number of unique applications that can neither be solved using GC nor LC but can be solved using SFC is limited. I t is clear that this range of applications does not provide sufficient right to exist for SFC. A much larger number of applications exists. however, in which SFC should be the method of choice because it is simply easier, more sensitive, more rugged or faster than either GC or LC. A typical example of an analytical problem that can be solved using either of the three chromatographic techniques but where SFC is the most favourable technique is the analysis of polymer additives. These compounds cannot be analysed using normal gas chromatography but require the use of high- temperature GC.In general, if an analytical problem can be solved using GC, GC is very often the technique that should be chosen. This is, however, no longer true when high-tempera- ture GC is needed. High-temperature GC still suffers from a number of practical problems. First of all the number of stationary phases available for high-temperature GC is limited. Furthermore, on-column injection, which is difficult to auto- mate, and the use of a retention gap are mandatory. Coupling of the retention gap to the analytical column is by no means trivial. Last, but not least, the current generation of high- temperature GC columns are very susceptible to breakage. Especially in routine analysis, SFC is a better alternative for the analysis of high relative molecular mass components than is high-temperature GC.For the particular example of polymer additives, analysis by LC requires gradient elution. This leads to relatively long total analysis times. Also, the detection limits of UV detection are generally poor. The analysis of liquid crystals used in liquid crystal displays is a second example of an analytical problem that can be solved using (high-temperature) GC, LC and SFC. Again, SFC is the most reliable, most rugged and fastest method. An example where GC cannot be used is the analysis of poly(methy1hydro- siloxanes). At higher temperatures these components tend to react with the stationary phase and the column wall of the GC column. LC can be used but generally does not provide sufficient resolution to separate the individual isomers.More- over, detection in LC is cumbersome. For this particular example open-tubular SFC is preferable as a complete separation of the individual isomers requires an extremely high plate number. For many other applications packed columns are clearly advantageous over open columns. Conclusions The number of chromatographic applications that can neither be solved by GC nor by LC, but can be solved using SFC, is limited. For a fairly large number of applications in which high relative molecular mass components or components of limited thermal stability have to be determined, however, SFC is to be preferred over GC and LC because it is easier, faster, more rugged or more reliable. Hence it is clear that SFC is a useful technique which definitely deserves a place among the other chromatographic techniques.An application area where SFC holds remarkable potential is the separation of chiral samples. As a result of numerous instrumental improvements, experi- mental difficulties are now seldom a major obstruction to the application of SFC. 1 2 3 4 5 6 7 8 9 10 1 1 References Klesper. E.. Corwin. A. H . . and Turner, D. A., J . Org. Chern., 1962. 27. 700. Novotny. M.. Springston. S . R.. Pcaden. P. A.. Fjeldsted, J . C.. and Lee, M. L.. Anal. CIi~rn., 1981, 53. 407A. Richter. B. E . , Knowles. D. E . , Andcrson. M. R., Porter. N. L., Campbell, E. R., and Later, D. W., J . High Resolut. Clirorna~ogr. Clirornutogr. Cornrnun., 1988. 11. 29. Farbrot Buskhe, A.. Berg, B. E., Gyllenhaal, O., and Greibrokk, T.. J . High Kesolut.Clirornutogr. Clirornutogr. Cornrnun., 1988. 11. 16. Guthric, E. J., and Schwartz. H . E.. .I. Clrrornutogr. Sci., 1986, 24. 236. Jackson, W. P . . Markidcs, K . E.. and Lee. M. L., 1. High Kcsolut. Chrornatogr. Clirornutogr. Cornrnun., 1986. 9, 2 13. Bcrger, T. A., and Toney, C.. J . Chrornutogr., 1989,465. 157. Wright, B. W., Kalinoski, H . T.. Udseth, H. K., and Smith, R. D.. J . High Resolut. Chrornatogi-. Clirornutogr. Cornrnun. ~ 1986. 9. 145. Jinno. K.. Clirornatogruyhiu, 1987. 23, 55. Schocnmakcrs. P. J., Uunk. L. G . M., and Jansscn, H.-G., J . Chrornutogr., 1990, 506, 563. Juvancz, Z . . Payne, K. M., Markidcs, K. E . . and Lee, M. L., Anal. Chern., 1990, 62, 1384. Photoionization Detection is 30 Years Old. The Story So Far Plus 'Son of Photoionization Detection': Far-ultraviolet Adsorption J.S. Hayhurst and J. N. Driscoll HNU Systems Ltd., Warrington, Cheshire and Newton, MA, USA Photoionization, as a means of detection, has been with us for about 30 years. Robinson' first reported the development of a photoionization detector in 1957. At the same time, groups in various parts of the world' were working on the development of flame ionization techniques. This latter technique became very popular and was rather quickly licensed to a number of commercial gas chromatography (GC) manufacturers since the detector was very sensitive and easy to build. Lovelock3 became interested in the photoionization tech- nique and published a review of ionization techniques in 1961ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 29 91 This included the flame ionization detector (FID), the photo- ionization detector (PID), cross-section and electron capture detection (ECD).The FID has become the most popular in the group as a result of its wide dynamic range, its selective response to organic compounds, its sensitivity and its simplicity of con- struction. While the ECD has also maintained a significant place as a selective detector for halogenated species, the cross- section detector has been forgotten. During the 1960s and 1970s, a variety of paperssp9 on photoionization were published using glow or microwave discharges as the ionizing source. Both types involved flowing high purity helium or argon through an electrical or microwave discharge to generate the source of energetic photons required to ionize the sample. With these early PIDs, the light source and ion chamber were not separated and, as a result, neither section could be optimized.The maximum lamp intensity occurs at low pressures while the maximum sensitivity for the ion chamber is at near atmospheric pressure. As a result, these early detectors were very pressure (or flow-rate) dependent. I n general, these detectors were difficult to operate, mechanically complex, unstable and required a vacuum pump. The glow discharge detector had the continual problem of column bleed collecting on the discharge electrodes, resulting in problems with igniting the lamp. It is no wonder that the PID was replaced with the FID during the 1960s. By the late 1960s most researchers agreed that photoionization had become firmly entrenched as the detector of choice for analysis of carbon compounds.In the period from 1973 to 1974, a major breakthrough in photoionization technology [separating the ion chamber from the ultraviolet lamp was reported by Driscoll and Spaziani,“’ Sevcik and Krysul,“ as well as Ostojik and Sternberg.” This design also eliminated many of the deficiencies of the glow discharge detector by allowing the ionization chamber to operate at atmospheric pressure and the lamp to be maintained at low pressures. This improved the sensitivity and simplified the operation of the detector. Now, after half a decade, photoionization appeared to be on the upswing again. The PID described by Driscoll and Spaziani13 was offered commercially by HNU Systems in the spring of 1976.Since this was the first commercial PID, it attracted a great deal of interest because of its reported 50-fold improvement in sensitivity over the FID for aromatic hydrocarbons. Other features of the new PID that were interesting were its response to inorganic compounds, and its non-destructive nature. Within two years, however, the PID was replaced with a new modelI4 that eliminated some of the deficiencies, such as temperature limitations and decompo- sition of thermally labile compounds on the inlet, which were inherent in the first PID. Some recent applications of the PID for analysis of organic and inorganic compounds are described briefly. All appli- cations and additional references in the text are to the PIDs with separated lamp and ionization chambers.The period covered is from 1976 to mid-1991. Discussion The PID detector is now in its third generation, has lower dead volumes for easier use with low flows and narrow diameter capillary columns. Its construction also allows easy interchange of lamps (8.3,9.5, 10.2 and 11.7 eV) which facilitates use of the PID as a selective detector; only species with ionization potentials at or below the energy of the lamp being ionized are detected. The different lamps do, however, vary in photon output, the lowest detection limits being achieved with the 10.2 eV source. In reviewing the role the PID detector had found in GC analysis, it is interesting to concentrate on the less well known, but arguably potentially very significant, applications for PID, particularly in industrial hygiene and emissions analysis. This is an area which has become much more important through the enactment of legislation, e.g., in the UK Control Substances Hazardous to Health (COSHH) and the Environmental Protection Act with its focus on integrated pollution control.This is not just a UK phenomenon but is reflected in similar legislation throughout the industrial world. PID traditionally has found most use as an aromatics detector, e.g., in standard US Environmental Protection Agency (EPA) methods for waste and potable waters (EPA methods 601/602/501/503) due both to its selectivity and sensitivity. Lower detection limits for benzene are typically sub-pg. There is also new emphasis on this application with the development of portable GCs for soil gas and contaminated groundwater analysis with the instigation of such programmes as LUST (Leaking Underground Storage Tanks). The PID has been the detector of choice for these applications due to its selectivity, low detection limits, the lack of gases, and its non-support-destructive nature which allow series detection, typically TCD or ECD, for environmental applications.The almost non-destructive nature (typically only one in 10 molecules of the analyte are ionized) has also meant that many PIDs have been used in series with other detectors and other instrumentation, e.g., with FID to determine which analytes are unsaturated or aromatic by comparison of relative res- ponse, or to determine the number of double bonds in a fatty acid. A growing number of PID detectors have been linked in series with mass spectrometers (MS); capillary GC-MS often does not allow sufficient eluent to ‘split’ the flow into a ‘destructive’ detector such as FID.The largest growing applications area for PID has been the continuous analysis of workplace air for toxic compounds and emission/fence-line monitoring for compliance with regulatory limits. The PID has been successful in these areas, particularly as it has no support gas requirements and because of its long- term stability at high sensitivities. The chromatograms listed below are all relatively common PID applications but seem to be not so well known to the analytical chemist, who typically perceives the PID to be an ‘aromatics detector’. These applications are: (i) isocyanates to ppt levels; (ii) formaldehyde at ppb levels (new low ‘control’ levels are being set for this compound); (iii) low-level organic sulfur compounds [nuisance (Environmental Protection Act) and process/feedstock contamination]; (iv) ethylene oxide, a suspected carcinogen ( e .g . , use as sterilization medium for medical disposables now has to be ‘justified’); ( v ) arsine/ phosphine at low ppb levels (the PID can detect a whole range of inorganics at trace levels). In all the chromatograms mentioned so far, particularly of air samples, a negative peak can be observed; this is caused by strongly UV-absorbing compounds. The far-UV light from a PID source is in the range 105-145 nm (10.2 eV = 122 nm). ‘Son of PID’ The effect mentioned above has been translated into the development of a new detector: far-UV adsorption (‘Son of PID’). This new detector is now a second generation instru- ment, having been optimized by reduction in dead volume and the development of a new photodiode ‘tuned’ to the frequency range emitted by a 10.2 eV (122 nm) source.Chromatograms for formaldehyde, anaesthesia gases, hy- drocarbons, sulfur compounds and permanent gases have been obtained, together with data on initial lower levels of detection which are likely to be exceeded during the continuing development programme. Conclusion PID, one of the first GC detectors to be developed, is now undergoing a resurgence of interest due to its ease of use and low detection limits, for ‘environmental’ samples.92 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 Far-UV (FUV) detectors appear to be able to offer a more sensitive (one to two orders of magnitude) alternative for many methods, such as landfill gas, currently analysed by TCD. References 1 2 Robinson, J., US Pat., Application Filing Date 1957; cited in Austr. Pat., 40359J72, 1972. McWilliam, I., in Detectors in Chromatography, ed. Nicholson, A. J. C., Australian Scientific Industries Association, Mel- bourne, Australia, 1984, p. 5. Lovelock, J. E., Nature, 1961, 188, 401. Lovelock, J. E., Anal. Chem., 1961, 33, 162. 3 4 9 10 I 1 12 13 14 Yamane, M., J. Chrornatogr., 1962, 9, 192. Roesler, J. F., Anal. Chem., 1964, 36, 1900. Locke, D. C., and McLoan, C. E., Anal. Chem., 1965,37,389. Oruce, J . G. W., Fenimore. D. C., Simmonds, P. G., and Zlatkis, A., Anal. Chern., 1968, 40, 541. Freeman. R. R., and Wentworth, W. E., Anal. Chem., 1971, 43, 1987. Driscoll, J. N., and Spaziani, F. F., Anal. Znst., 1974; presented at the ISA Meeting, NYC, 1974. Sevcik, J., and Krysul, S., Chromatographia, 1973, 6, 375. Ostojik, N., and Sternberg, Z., Chromatographia, 1974, 7, 3. Driscoll, J. N . , and Spaziani, F. F., Res. Dev., 1976, 27, 50. Driscoll, J . N . , Ford, J., Jaramillo, L. F., Becker, J. H., Hewitt, G., Marshall, J. K., and Onishuk, F., Am. Lab., 1978,10, 137.

ISSN:0144-557X    

DOI:10.1039/AP9933000089

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 93 Research and Development Topics in Analytical Chemistry The following are summaries of seven of the papers presented at a Meeting of the Analytical Division held on July 9th-?Oth, 1991, in the University of Aberdeen. Summaries of thirty-one other papers from the Meeting were published in the January, February, March and June, 1992, issues. Flow Injection Studies of Cycloc Amino Acid-Thiol-o-Phthalalde lextrin-enhanced Fluorescence in the iyde Reaction Muhammad Y. Khokhar and James N. Miller Department of Chemistry, L oug h boroug h University of Tech nolog y, Lo ug h boro ug h, Leicesters h ire LEI1 3TU The a-, P- and y-cyclodextrins (CyDs) are cyclic, water-soluble receptor molecules with characteristic bucket-shaped cavities. They form inclusion complexes with compounds of size and geometry compatible with the dimensions of these cavities.Van der Waals interactions, hydrogen bonding, release of strain energy in the CyD ring and release of high-energy water molecules from the cavity are involved in complex formation. Various chemical effects of CyD inclusion complexes on guest molecules have been reported, including the sheltering of the concealed parts of and enhanced reactivity of the exposed parts of a guest molecule, and strong interactions between a pair of guest molecules. Many spectroscopic properties of the com- plexed molecules may also be affected, fluorescence enhance- ment phenomena being particularly common. o-Phthalaldehyde (OPA) forms rather unstable fluorescent isoindoles with amino acids in the presence of thiols.'*2 We have used flow injection analysis (FIA) to study the host-guest complex formation and the effects of CyDs on the stabilities and fluorescence intensities of these products.Experimental n-Phthalaldehyde (0.2 g) was dissolved in 2.0 ml of ethanol, 50 pI of a thiol [ethanethiol (ET), 2-mercaptoethanol (2-ME), -- 0 0.2 0.4 0.6 0.8 8 - 5 8" 1 2 3 4 ($-Cyclodextrin/mmol I ' Fig. 1 Fluorescence intensity, I f , of OPA-thiol-DL-lysine derivatives as a function of (3-CyD concentration. A, 2-Mercaptoethanol; B, ethanethiol. The inset diagram shows more detail of the measurements at low p-CyD concentrations etc.] were added, and the volume made up to 250 ml with 0.025 mol I-' borate buffer, pH 9.26. Cyclodextrin solutions were prepared daily in this OPA reagent solution, and where necessary diluted with it.Cyclodextrin-OPA-thiol solutions were used as the carrier stream in an FIA system comprising a single flow line of 0.5 mm i.d. PTFE tubing. Aqueous amino acid solutions (75 pl, 1 pmol I-' L-serine, 10 pmol I-' DL- lysine) were injected using a Rheodyne valve. Fluorescence intensities were measured at room temperature using a Perkin- Elmer LS2B filter fluorimeter fitted with a 7 pI illuminated volume flow cell. Results and Discussion P-Cyclodextrin (cyclomaltoheptaose) enhanced the fluores- cence of the isoindoles derived from DL-lysine and 2-ME or ET in steps (Figs. 1 and 2), over the CyD concentration ranges 10-4-3 x lop4 mol I - ' and 10-'-3 x mol I - ' . These isoindoles are probably too large to be incorporated into a single 0-CyD cavity: it may thus form both 1 : 1 and 1 :2 complexes with j3-CyD.However, the DL-lysine isoindole formed with ET also showed fluorescence enhancement in a y- CyD (cyclomaltooctaose) cavity, whereas that formed using 2- ME did not. The observed enhancements were by factors of two- to 10-fold, rather less than the effects found for other fluorophores. The DL-lysine isoindole formed using 2-methyl- propane-2-thiol did not show fluorescence enhancement in 10 8 6 \- 4 2 0 1 2 3 4 5 6 y-Cyclodextrin/mmol I - ' Fig. 2 Fluorescence intensity, If, of OPA-ethanethiol-riL-lysine derivative as a function of y-CyD concentration94 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 either CyD. With the exception of L-serine, whose isoindole with 2-ME showed an enhancement effect in y-CyD, no other enhancement effects were found with a range of amino acids, and a-CyD failed to enhance the fluorescence of any of the isoindoles.These results point to a combination of factors being responsible for the fluorescence enhancements, with specific interaction effects as well as the size compatibility of the CyD cavity and the isoindole influencing the outcome. The insta- bility of OPA-derived isoindoles in aqueous media is due to their In the present work, similar instability was demonstrated for those isoindoles whose inclusion in CyD cavities gave rise to enhanced fluorescence. The host-guest relationship does not, therefore, protect these fluorophores from hydrolysis. This finding, taken with the absence of enhancement effects in most cases, and the small extent of the enhancements in others, indicates that interactions between the isoindoles and the CyDs were generally weak.One of us (M. Y. K.) thanks the Ministry of Education of the Government of Pakistan for financial support, and the Bahauddin Zakariya University, Multan, Pakistan for study leave. References 1 2 3 4 5 Roth, M., Anal. Biochem., 1971, 43, 880. Simons, S. S. J., and Johnson, D. F., J. Org. Chem., 1978, 43, 2886. Simons, S. S. J., and Johnson. D. F., Anal. Biochem., 1977,82, 250. Kagan, J . , Tetrahedron Lett., 1966, 8097. Roth, M., and Hampai, A., J. Chromatogr., 1973, 83, 353. Design and Application of Chiral Liquid Chromatography for Drug Metabolism Studies G.J. Furlonger and Anthony F. Fell* Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford 807 7 DP B. Kaye Drug Metabolism Department, Pfizer Central Research, Sandwich CT73 9NJ Since the observation that a number of chiral drugs are associated with adverse side effects attributable to one of the enantiomers, thalidomide being a notable example,' pressure from the regulatory authorities has led to pharmaceutical companies choosing to develop single enantiomer drugs rather than racemates. As some classes of drug are known to invert ti-om the less active enantiomer to the therapeutically active isomer in vivo, the screening of new compounds for possible metabolic stereochemical inversion is therefore important. The profens, notably ibuprofen, are some of the few drugs known to invert, both in vitro and in vivo, in this manner.* Another compound shown to undergo in vivo inversion is the aryloxy propionate herbicide haloxyfop (Fig.l).3 There is now much experimental evidence4 for the mechanism of inversion as proposed by Nakamura et d5 This involves the formation of an ibuprofen-CoA (CoA = coenzyme A) thioester. Of the ibuprofen enantiomers, the less active enantiomer, (R)- ibuprofen, is the only substrate for the enzyme that catalyses this thioester formation. A racemase then catalyses the reaction to form (S)-ibuprofen-CoA thioester, which is hydrolysed back to the more active parent isomer. Thus there is a net increase in (S)-ibuprofen in vivo relative to the R isomer. Doxazosin is a postsynaptic a,-adrenoceptor antagonist and one of its derivatives, the racemic acid A, exhibits some structural similarity to ibuprofen and haloxyfop (Fig.1). This has been subjected to the in vitro conditions shown to elicit inversion in ibuprofen2 which was used as a positive control for the inversion. Chiral high-performance liquid chromatography (HPLC) assays were developed on an &'-acid glycoprotein (AGP) chiral HPLC column for the analysis of the enantiomers of A and of ibuprofen in order to study this inversion process. Experimental Sample Incubation with Liver Homogenates Whole-liver homogenates from rat o r dog, prepared in phosphate buffer and stored at -7O"C, were used in vitro at 37°C for 45-120 min under the following conditions:2 liver homogenate, 250 pl; adenosine 5'-triphosphate (ATP), 3 mmol I-'; reduced CoA, 0.4 mmol I-'; Tris-HCI, pH 7.7, 50 mmol I-'; MgCI2, 25 mmol I-'; drug, 0.2 mmol I-' (= 18.0 pg per 0.5 pl of A , 20.6 pg per 0.5 ml of ibuprofen).All concentrations are final incubation concentrations and the total incubation volume was 0.5 ml. The final drug concentration varied from 40 to 200 pmol I-' for ibuprofen and 10-20 pmol I-' for racemate A. Blanks were prepared by adding 200 pl of Tris-HCI (instead of the cofactor mixture) to the homogenate and drug. Each tube being incubated was subjected to gentle manual agitation approximately every 5 min. Sample Extraction and Clean-up After incubation the samples were acidified with HCI (3 mol I-', 0.5 ml) and extracted with 9 ml of solvent. The solvent used for ibuprofen was hexane; for A, insoluble in hexane, the solvent combination giving the cleanest extracts CH3 F 3 C + O e O - C - H I I COOH Ha I oxyfo p COOH 1 bu profen @.--O.--:ooH UK 2249 (A) * To whom correspondence should be addressed.Fig. 1 Structures of ibuprofen, haloxyfop and acid AANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 95 0 Y Table 1 Changes in the enantiomeric ratio for samples incubated with liver homogenate Incubate* RL no cofactors RL + cofactors 2 x RL + 2 x cofactors DL + cofactors DL no cofactors 2 x DL + 2 x cofactors Ethyl ucetute extraction: racemate A - Diethyl ether extraction: racemate A - RL no cofactors RL + cofactors RL + cofactors DL + cofactors DL + cofactors Amount recoveredhgt Amount spiked per enantiomerhg R S Ratio R : S 410 406 (99%) 414 (10lYo) 1 : 1.02 410 426 (104%) 394 (96Yo) 1 : 0.92 410 414 (101%) 406 (99%) I : 0.98 410 411 (100%) 408 ( 100~0) I :0.99 410 417 (102%) 302 (98%) 1 : 0.96 410 441 (107%) 379 (92%) 1 : 0.86 8220 82 12 (100Yo) 8228 (100%) I : 1 .oo 8220 8424 (102%) 8016 (98%) 1 : 0.95 8220 8478 (lO3Yo) 7962 (97%) 1 :0.94 8220 8255 (100%) 8185 (l000/,) I : 0.99 8220 8283 (101%) 8157 (99?'0) 1 : 0.98 Hrxune extraction: ibuprofen controls- RL + cofactors 4325 31 13 (72%) 5537 (1280/,) 1 : 1.78 DL f cofactors 4325 1328 (31%) 7322 (169%) 1 : 5.51 * RL = rat liver homogenate; DL = dog liver homogenate; 2X = double the amounts described in the text whilst adding the same amount of + Percentage relative to spiked amount of enantiomer after correcting for recoveries. drug.R I I I 0 5 10 15 0 5 10 15 5 10 15 Ti m e/m i n Fig. 2 Chromatograms showing ( a ) racemic ibuprofen; ( b ) ibuprofen after incubation with rat liver homogenate; (c) after incubation with dog liver homogenate. For chromatographic conditions see text was diethyl ether-chloroform (4 + 1 v h ) ; both ethyl acetate and diethyl ether were also tried with less success. The solvent layer was then transferred to a fresh sample vial and evaporated to dryness, reconstituted in fresh mobile phase and filtered through a 0.45 pm Millipore syringe filter immediately prior to chromatographic analysis. Solutions are either diluted or analysed directly. Chromatography Analyses of ibuprofen and A enantiomers were carried out on a chiral &,-acid glycoprotein column (100 X 4 mm) fitted with an &,-acid glycoprotein pre-column (ChromTech AB).The mobile phase comprised sodium dihydrogen phosphate (0.035 mol I-*); the optimum pH was 6.5 for ibuprofen and 4.0 for A. The flow rate was 0.9 ml min-' and ultraviolet (UV) detection was effected at 260 nm for ibuprofen and 225 nm for A. For these compounds the best resolution was obtained for aqueous eluent alone, without addition of an organic modifier. In order to maintain optimum performance of the AGP R S G I I L (b) R 10 15 0 5 10 150 5 Time/m in Fig. 3 Chromatograms showing racemic acid A before ( a ) and after ( h ) incubation with liver homogenate. For chromatographic conditions see text column despite the high throughput of biological extracts, the column was re ularl back-flushed overnight at a low flow rate to do this led to a loss of resolution after 20-30 injections. (0.05 ml min- F Y ) with propan-2-01-water (20 + 80 vh).Failure Results and Discussion Results for A with a single batch of homogenate are presented in Table 1. Solutions of the two compounds were chromato- graphed as external standards to calculate the concentration of the extracted enantiomers. For a racemic mixture the R : S ratio96 ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 should be 1 : 1; any deviation after incubation indicates that some form of enantioselectivity has occurred. For chiral inversion to be demonstrated it is necessary to show both that the R : S ratio changes and that one enantiomer has increased in absolute concentration. A simple change in the R : S ratio could be attributable to enantioselective metabolism alone.At this early stage in examining the possibility of chiral inversion, the experiments were carried out without an internal standard to see if there was any evidence for stereospecific metabolism taking place in vitro. With such a weakly eluting HPLC mobile phase, it has not been practicable to find a suitable internal standard for this early work. Thus it was necessary to assume that extraction losses would affect each enantiomer equally. With the increase in S-enantiomer con- centration in vitro, the ibuprofen results (see Fig. 2) gave clear evidence that inversion had occurred, thus validating the activity of the liver homogenate. Interestingly, the absolute concentra- tion of the S enantiomer recovered was greater than that originally added, thus confirming the earlier literature reports.’ The results for the racemic acid A were less clear cut, however (see Fig.3). There was a change in the enantiomeric ratio after incubation (this was confirmed several times), but the total recoveries were too low to detect reliably any increase in the R enantiomer Concentration above its original spiked level. However, after applying a correction factor to account for the extraction losses, there is some evidence that the R enantiomer is formed, with a concomitant decrease in S enantiomer, as the racemate A is metabolized by the rat liver homogenatc. Further work on this phenomenon and on a number of compounds which could potentially invert in vitro and/or in vivo is being carried out in these laboratories.The authors thank the Science and Engineering Research Council and Pfizer Central Research for their financial support of this research programme. References I 2 3 4 Blaschke, G., Kraft, H, P., Fickcntscher, K., and Kohler, F., Arzneim.-Forsch., 1979, 29, 1640. Knihinicki, R. D., Williams. K. M.. and Day, R. O., Biochem. Pharmacol., 1989, 38, 4389. Bartels, M. J., and Smith, F. A., Drug Metub. Disp., 1989, 17, 286. Sanins, S. M., Adams, W. J . , Kaiser, D. G., Halstead, G. W., Hosley. J., Barnes, H . , and Baillie, T. A., Drug Metab. Disp., 1991, 19, 405. Nakamura, Y . , Yamaguchi, T.. Takahashi, S . , Hashimoto, S . , Iwatani, K., and Nakagawa, Y., Pharmucobiodyn., 1981.4, S1. 5 Comparison of Absorbance Ratios and the Peak Purity Parameter for the Verification of Peak Homogeneity in High-performance Liquid Chromatography Daemon Lincoln and Anthony F.Fell* Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford BD7 IDP The introduction of linear photodiode array detectors (LPAI)) in liquid chromatography (LC) has seen the development of many data manipulation methodologies. Two basic types of LPAD differ in the number of photodiodes (35 or 210) and the spectral scan rate employed. The digital techniques include graphical methods (spectral overlays, isoabsorbance plots, three-dimensional projections of wavelength-time-absorbance data, higher derivative transformation of spectra) as well as numerical techniques (absorbance ratio, spectral suppression ,2 purity parameter,3 absorbance index, multiple absorbance ratio correlation‘).This paper compares a novel method developed recently, the ‘purity parameter’ (PUP), with the established absorbance ratio (AR) approach. In addition, a novel technique based on the PUP is described. This is the ‘PUP difference chromatography’ (PuPDC), which presents the differences between the target chromatogram and the reference chromatogram, in terms of the PUP expressed as a continuous function of elution time. The work described was conducted using a model system of solutes particularly susceptible to co-elution: paraxanthine (PX) and theophylline (TH), the correlation coefficient of the spectra being, Y = 0.9947. The elution conditions were designed to ensure specified degrees of peak overlap for PX and TH.Purity Parameter The purity parameter (PUP) is defined as ‘the average wavelength of a spectrum weighted by the absorbance at each individual ~avelength’:~ * To whom correspondence should be addressed. i = ti c AYhi i=O PUP$) = i = ri c A; i = 0 The advanced software for the purity parameter is currently only available on the Varian Polychrom 9060/9065 photodiode array detector. It relies on the linear additivity of overlapping spectra for detection of impurities at any point in the chromatographic peak profile. By comparing PUP values taken at specified points on the chromatographic peak with each other, it should be possible to evaluate peak homogeneity, since a pure peak should in principle have a constant PUP value at all points.Relative Difference in PUP This value, expressed as a percentage, allows comparison of the differences in PUP values taken at two different time points, f l and t2, in the chromatographic peak profile, obtained over a specified wavelength range ( R ) . It can be expressed as: where R = wavelength range (nm) and RD = relative distance. Experimental Instrumentation A Varian LC Star System Workstation operating on a Compaq Deskpro 286E computer was used to operate a Varian 9010ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 . . - @. ...‘ I I I I 1 I 97 0 Solvent Delivery System fitted with a Rheodyne 7125 manual loop injection valve (10 @I) and a Varian Polychrom 9065 detector (Varian Instrument Division, Walnut Creek, CA, USA). - 40 Column and Eluent A Spherisorb S5 ODS2 (25 cm x 4.6 mm i.d.) reversed-phase column was used.Two mobile phases were prepared to give two different values of the chromatographic resolution, R, = 0.4 and R, = 0.9, using: methanol-water (50 + 50, v/v) and methanol-water (30 + 70, v/v), respectively. Time Points Three PUP values were collected at sequential time points (st = 1 s), collectively described as ‘tl’, about the apex of the paraxanthine peak, for comparison with three values collected similarly about the predicted position of the later eluting impurity peak, t2. The wavelength ranges ( R ) chosen for assessment of the PUP data were: 220-367; 220-263; and 220- 287 nm. These three ranges allowed comparison of the PUP calculated under three conditions: the optimized wavelength range (220-287 nm); a broad range (220-367 nm); and a narrow (220-263 nm) wavelength range, respectively.Absorbance Ratio The wavelengths 220 and 278 nm were found to offer the greatest potential discrimination between the chosen com- pounds without compromising signal-to-noise ratio, as demon- strated by the difference spectrum (Fig. 1). Results For all three wavelength ranges, a statistically significant increase in the PUP value across the chromatographic peaks is observed. A corresponding decrease in absorbance ratios (AR) also occurs. The effect is dependent on the rate of change in concentration in the elution profile and is probably related to the phenomenon of spectral skewing. A linear correction algorithm within the software is designed to compensate for this time-related phenomenon (Fig.2). The displacement of the analyte spectrum results from the finite time taken to read the array of photodiodes as the chromatographic concentration profile changes. PUP Difference Chromatogram A more effective way to compensate for this effect (and coincidentally to allow purity assessment of the entire chro- I 40 250 matographic peak) is to subtract the PUP values obtained at a number of equidistant time points within a reference standard peak from the corresponding time points within the target peak. This procedure has been carried out to produce a new form of peak purity criterion, viz., the ‘PUP difference chromatogram’. Another graphical presentation is shown as the purity parameter plot in Figs. 3-5. Discussion By application of the standard PUP technique it was found to be possible to show spectral discrimination for the model system containing 0.5% m/v theophylline.The predicted optimal PUP wavelength range (220-287 nm) yields the steepest gradient in RD versus theophylline concentration plots for both resolution cases (R, = 0.4 and 0.9), confirming the optimal conditions selected. This compared well with detection of a similar concentration for the optimized absorbance ratio technique under the same conditions. Values were obtained under optimum conditions of time point selection and a chromato- graphic resolution of R, = 0.9. Visual inspection of the PUP difference chromatograms (and the respective PUP plots) confirms that 0.5% m/v and arguably 0.25% m/v of the peak impurity are detected at R, = 0.9, even though the spectra are strongly correlated (Fig.1). The data at R, = 0.4 are rather more difficult to interpret. Neither with the standard PUP nor the AR technique was it possible to detect confidently the impurity for this chromatographic resolution. The PUP difference chromatogram, and in particu- I Waveleng t h/ti me Fig. 2 Diagrammatic representation of the effect of spectral skewing. The error in absorbance values generated when the spectrum is skcwcd (A) is illustrated. The linear approximations of the absorbance baseline (B) and the true absorbance baseline (C) are also shown 2 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 Time/min Fig. 3 PUP difference chromatograms for the stated model system representing a 2.0% m/v added impurity at a chromatographic resolution, R, = 0.9ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 98 300 250 m $2 200 X Q) 150 $ a a 100 50 0 258 B C 2.0 2.1 2.2 2.3 2.4 2.5 Tirne/m i n Fig.4 Overlaid PUP plots for the stated model system at a chromatographic resolution. R, = 0.4. The cases shown represent the 2.0% m/v (A), 1.0% m/v (B), 0.5% m/v (C) and 0.25% m/v (D) impurity levels. The bracketting standards, Standard I (E) and Standard 2 (F), are also shown lar the multiple PUP plot (Fig. 4), shows clearly that the 2.0% m/v impurity can be detected and there is a strong case for detection at the 1.0% m/v level. Conclusions The purity parameter offers a number of advantages over the absorbance ratio technique. Although the AR based methods offer potentially the greatest sensitivity to co-eluting impuri- ties, in practice they are unlikely to achieve this due to the difficulties of appropriate wavelength and time point selection.Many numerical and graphical approaches have attempted to take into account the three-dimensional nature of the spec- trochromatographic data set and the inevitable ‘diluting’ of spectral and temporal differences between the primary compo- nent and any underlying impurities is a direct consequence of their general applicability. Although it is shown that it is possible to detect co-eluting impurities at very low levels, the methods employed (i.e., taking the PUP values from regions of the chromatographic peak known to contain the impurity) are not yet routinely applicable. A more successful approach would be to examine the PUP values collected at frequent time points across the - 3.5 3.6 3.7 3.8 3.9 4.0 4.1 4.2 4.3 Time/rn in 259.5 258.5 E t 3 m a 3 II \ - 257.5 256.5 Fig.5 Overlaid PUP plots for the stated model system at a Chromatographic resolution, R, = 0.9. The cases shown represent the 2.0% m/v (A), 1.0% m/v (B), 0.5% m/v (C) and 0.25% m/v (D) impurity levels. The bracketting standards, Standard 1 (E) and Standard 2 (F). are also shown chromatographic peak. However, in view of the trends observed with these values, comparison of a difference chromatogram may well prove to be more successful. For ‘absolute’ proof of purity, peak collection followed by analysis with a different chromatographic system, or preferably with an on-line ‘fingerprint’ spectroscopic system such as mass spec- troscopy, would be required.The authors thank the Science and Engineering Research Council and Sterling Research for their financial support of this research programme. References I Fell, A. F.. Scott, H . P., Gill, R., and Moffat. A. C., Clrromatographia, 1982, 16, 69. 2 Kirk, E. M.. and Fell. A. F., Clin. Chern.. 1989, 35, 1288. 3 Alfredson, T.. and Sheehan, T., Am. Lab., 1985, 17, 40. 4 Marr, J . D. G., Seaton, G. G . R., Clark, B . J . , and Fell, A. F., J . Clrrornatogr., 1990. 506, 289. 5 Chan, H. K., and Carr, G. P., J . Pharm. Riomed. Anaf., 1990.8, 271. Assessment of the Performance of a New Protein-based Phase in the Chiral Liquid Chromatography of Drugs M. C. Banks and Anthony F. Fell* Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford BD7 1 DP R.D. McDowall Department of Bioanalytical Sciences, Wellcome Research Laboratories, Beckenham, Kent BR3 3BS The usc of protein-based chiral stationary phases (CSPs) for the direct enantiomeric resolution of chiral drugs has increased rapidly over the past decade due to the broad applicability of these columns. This is because the aqueous buffered mobile phases used are compatible with biological samples, and because the inherent stereoselective nature of proteins in enzymes and receptors in vivo is exploited. * To whom correspondence should be addressed. The first of these protein-type CSPs appeared in 1982 comprising a silica-based bovine serum albumin (BSA) phase developed by Allenmark et a f .' Other protein-based CSPs then followed: &,-acid glycoprotein (AGP) by Hermansson in 1984,2 avidin by Miwa et af. in 1988,3 human serum albumin by Domenici et af. in lW0,4 and the glycoprotein, fungal cellulase, by Erlandsson et af. (1990).5 An additional protein-based CSP was developed in 1987 by Miwa et af.6 based upon ovomucoid (OVM), a protein found in chicken egg white. The propertiesANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 4 - 2 - 99 4 0.30 o f OVM differ significantly from those of earlier protein-based systems (Table l).7-9 By using ibuprofen as the test analyte, the parameters of mobile phase pH; organic modifier type and concentration; and buffcr concentration were evaluated to assess the perfor- mance of a commercially available immobilized OVM chiral stationary phase.The enantiomeric separation of two closely related benzo- diazepincs was also compared under identical conditions to assess the cffect of small structural changes within a molecule on resolution and performance. Experimental Apparatus The Ultron ES-OVM column (150 x 4.6 mm i.d.) obtained from Hichrom, Reading, contains ovomucoid immobilized onto a 5 pm silica support. The column can be used with aqueous mobile phases in the pH range 3.0-7.5 and with many of the commonly used organic solvents [e.g., methanol (MeOH), ethanol (EtOH), propan-2-01 (IPA) and acetonitrile (ACN)]. It has an upper operating temperature of 40 "C. The chromatographic system comprised LDC ConstaMetric 111 and ConstaMetric 1 pumps with a static mixer, controlled by an LDC Chromatography Control Module (CCM).Ultraviolet (UV) detection was with a Kratos Spectroflow Monitor SF 770 variable wavelength detector recording via the CCM to an LDC HI 2PP thermal chart recorder. Standards Kaccmic drug standards of ibuprofen and the benzodiazepines were obtained, respectively, from Sigma Chemical Company, Poole, and from Wyeth Laboratories, Maidenhead. Results and Discussion Effect of Organic Modifier Type on Capacity Factor, Resolution and Column Efficiency The effect o f changing the type of organic modifier used is illustrated in Fig. I . The concentrations were chosen to give comparable capacity factors for the first enantiomeric peak for ibuprofen (Fig. 1) at pH 5.0, one of the recommended values for this column. I t is evident that the type of organic modifier used greatly affects the observed stereoselectivity. The effect of organic modifier on resolution and column efficiency is shown in Table 2, from which it can be seen that a dramatic increase in column efficiency is observed for the addition of very small amounts of IPA.Ethanol clearly gives the highest resolution under these conditions. If IPA were then combined with the good resolving capabilities of EtOH, better resolution would he anticipated, as indeed is confirmed in Table 2 and Fig. 2. Effect of pH on Capacity Factor and Resolution Ethanol is the rccornmended organic modifier for OVM and since i t gave the best resolution in the preliminary experiments (Fig. l ) , the effects of pH and other parameters were explored on this system. Fig.3 shows that both the retention and resolution of ibuprofen enantiomers decrease with increasing pH. This would suggest favourable interactions with the less Table 1 Charactcristics of somc proteins used as chiral stationary phascs Property AGP OVM BSA Molecular wcight 4 I 000 28 800 66 000 lsoelcctric point 2.7 3.Y-4.5 4.7 Sialic acid* residues 14 0.3 Disulfidc bridges 2 8 17 YO Carbohydrate 4s 30 * A'-Acylncuraminic acid. 16 l 8 ~ Fig. 1 Effect of organic modifier type on capacity factor and resolution of separation of enantiomcrs of ibuprofen with OVM column. Conditions as for Tablc 2 with organic modifier in 10 mmol I - ' phosphate buffer. pH 5.0. as mobile phase. A. Capacity factor of first eluting peak; B , capacity factor of sccond eluting peak; C, resolution ionized, hydrophobic form of ibuprofen (pK,, == S ) , relative to the more polar anionic form.Such a significant change in resolution and retention for a small change in pH indicates that mobile-phase pH is a sensitive parameter for optimizing enantiomeric separations o n OVM, and must be carefully con trolled for day- to-day reproducibility . Effect of Organic Modifier Concentration on Capacity Factor and Resolution The effect of different concentrations of organic modifier on the capacity factor and resolution of ibuprofen on OVM is shown in Fig. 4. This shows a decrease in retention with increasing modifier concentration, typical of a reversed-phase system. Again, the importance of careful mobile-phase preparation is highlighted, with significant changes in retention caused by even a 1% change in EtOH.The unusual effect on resolution is interesting. This could be attributable to the presence of more than one different interaction site for ibuprofen on the immobilized protein. If EtOH concentrations greater than 2% were changing the conformation of the protein, it is possible that one of these sites could be rendered ineffective, leaving other sites to effect enantiomeric resolution. Effect of Buffer Strength on Capacity Factor and Resolution The phosphate buffer strength plays a much less significant role in the retention of ibuprofen on the OVM column; buffer concentrations in the range 10-20 mmol l--' are sufficient for retention control. Although resolution increases at buffer concentrations up to 30 mmol I - ' , the peak shape deteriorates at concentrations in excess of 20 mmol I-'.Effect of Molecular Changes on Retention and Resolution for Two Structurally Similar Benzodiazepines The increase in hydrophobicity from lorazepam to temazepam results from the change of the secondary amine to a tertiary amine and the loss of a single chlorine. This leads to a marked difference in retention and resolution, as shown in Fig. 5 . Under identical mobile-phase conditions, the more hydro- phobic temazepam shows greater affinity for the CSP than the more polar lorazepam. The dramatic increase in stereo- selectivity for temazepam enantiomers is all the more interest- ing since the chiral centre is the same for both drugs. The basis for this requires further clarification.One possible mechanism could be the presence of two, distinct chiral recognition sites or processes, one of which has a stronger binding constant for one of the temazepam enantiomers. For lorazepam, it is possible that the same binding site is involved in the retention of both enantiomers.100 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 - - - - - Table 2 Effect of combining organic modifiers in mobile phase on capacity factor. resolution and efficiency of separation of enantiomers of ibuprofen* Height of theoretical platdmm 1.50 1.20 C 0.90 .g 3 0 0.60 $ 11I - 0.30 Modifier( s) First eluting Second eluting EtOH : IPA : McOH k ' , ? k ' : i Resolution peak peak 100 : 0 : 0 6.71 8.82 1.20 0.23 0.52 50 : 50: 0 7.00 8.96 1.28 0.18 0.36 0 : 100 : 0 7.46 9.18 1.13 0.18 0.32 0: 50: 50 7.00 8.99 0.98 0.32 0.59 0:o: 100 8.09 11.15 1.08 0.49 0.83 50: 0: 50 7.36 9.95 1.12 0.36 0.71 33: 33 : 33 7.24 9.79 1.22 0.32 0.56 * Conditions: column, Ultron ES-OVM, 150 X 4.6 mm i.d.; mobile phase, 10 mmol I ' phosphate buffer, pH 5.0 with 4% ethanol.1.8% propan- 1- Capacity factor of first eluting peak. 4 Capacity factor of second eluting peak. 2-01 and 8% methanol; flow rate, 1.0 ml min-'; temperature, ambient; sample, 20 p1 of 50 pg ml-'; UV detector, 223 nm. I I I I 0 10 20 30 Ti me/mi n Fig. 2 Structure and separation of enantiomcrs of ibuprofen with OVM column. Conditions as for Table 2 with 2% ethanol, 1% propan- 2-01 in 10 mmol I-' phosphate buffer, pH 5.0, as mobile phase 0 ' I I I I \ ' 0 PH Fig. 3 Effect of pH on capacity factor and resolution of separation of enantiomers of ibuprofen with OVM column.Conditions as for Table 2 with 5% ethanol in 20 mmol I-' phosphate buffer. pH 4.5-6.5, as mobile phase. Labels as for Fig. 1 4.5 5.0 5.5 6.0 6.5 Conclusions The OVM column has been found to be widely applicable to the enantiomeric separation of many chiral pharmaceutical compounds.'0311 The study with ibuprofen has shown that a number of parameters need to be considered when attempting optimization of chiral separations with this CSP: (i) different organic modifiers will cause significant changes in enantioselec- tivity; in some cases a combination of organic modifiers can enhance the resolution; (ii) the mobile-phase pH has a strong 18 1.50 L 0 " " " " " ' 0 1 2 3 4 5 6 7 8 9 10 Ethanol concentration (% v/v) Fig.4 Effect of organic modifier concentration on capacity factor and resolution of separation of cnantiomers of ibuprofen with OVM column. Conditions as for Table 2 with &lo% ethanol in 20 mmol I-' phosphate buffer, pH 5.5, as mobile phase. Labels as for Fig. 1 Temazepam Loraze par !r- 0 10 20 30 0 5 10 15 Ti me/m in Fig. 5 Structure and separation of enantiomers of temazepam and lorazepam with OVM column. Conditions: column, Ultron ES-OVM, 150 x 4.6 mm i.d.;, mobile phase. 15% ethanol in 10 mmol I-' phosphate buffer, pH 5.0; flow rate, 1.0 ml min-'; temperature, ambient; sample. 20 p1 of 50 pg ml-'; UV detector, 231 nm. Impurity of temazepam (+); impurity artefact (*)ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 101 effect on the retentivity of the OVM column resulting from changes in the degree of ionization both of the analyte and the CSP itself, and possibly from conformational changes within the immobilized protein; (iii) organic modifier concentration also has a strong effect on the retentivity of the column; however, the effect on resolution is not predictable; (iv) buffer strength has a lcss significant effect on retention and resolution than other parameters. In general, the parameters of pH and organic modifier type and concentration are the most important when attempting to optimize enantiomeric resolution, and careful mobile-phase preparation is important for good reproducibility. The authors thank the Science and Engineering Research Council and Wellcome Research Laboratories for their finan- cial support of this research programme. References 1 Allcnmark.S., Bomgrcn, B.. and Boren, H.. J . Chrornutogr., 19x2, 237. 473. 2 3 4 5 6 7 8 9 10 1 1 Hermansson, J . . J. Chromatogr., 1984, 298, 67. Miwa, T., Miyakawa. T., and Miyake. Y., J . Clzromatogr., 1988. 457, 227. Domcnici, E., Bertucci. C., Salvadori. P . , Felix. G.. Cahagne, I . . Motellier, S . , and Wainer, I . W.. Chrornatograplziu, 1990, 29. 170. Erlandsson, P.. Marle, I . , Hansson. L.. Isaksson, K., Petters- son. G., and Pettersson, C., J . Am. Chem. Soc., 1990, 112, 4573. Miwa. T.. Miyakawa. T.. Kayano, M.. and Miyakc, Y.. J . Chromatogr.. 1987. 408, 316. Hermansson. J . . Trends A n d . Clzem., 1989, 8, 251. Mclamcd, M. D., in Glycoprotc.ins, ed. Gottschalk, A ., Elsevicr, New York, 1966, p . 317. Allenmark, S., J . Cliromatogr., 19x3, 264, 63. Miwa, T., Kuroda, H . , Sakashita, S . . Asakawa, N . , and Miyake, Y.. J . Chromarogr., 1990, 511, 89. Kirkland, K. M.. Ncilson, K. L., and MeCornbs. D. A.. J . Clirornutogr. , 199 1, 545, 43. Drug and Analogue Structural Relationships in Chiral Separations with Mobile-phase Additives P. Mitchell and B. J. Clark Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford BD7 'I DP The ability o f the body to discriminate between the enanti- omers of chiral drug substances has long been recognized.' Differences in potency and pharmacological action of drug enan tiomers, and also enantiomeric intereonversion of drugs in the body have all been widely reported.' Owing to the large number of enantiomeric drug substances regularly prescribed (>30%), considerable effort has been expended over the last decade on resolution and the analysis of individual enanti- omers,3 particularly for drug quality control and pharmaco- kinetic profiles of clinical samples.In recent years a major focus of analytical and bioanalytical chiral analysis has been towards chromatographic separation, and the majority of the work has centred o n high-performance liquid chromatography (HPLC). Of the three commonly used HPLX methods, two are based on operation with conventional reversed-phase column materials, through diastereomeric derivatization or the use of chiral selectors in the mobile phase. The former method was originally used in the resolution of chiral drugs and their metabolites in biological matrices4 but is now widely applied in a number of fields.In contrast, direct procedures, using chiral stationary phases, are favoured by the pharmaceutical analyst engaged in the determination of enantiomeric ratios for quality control and related activities. Numerous chiral stationary phases (CSPs) are commercially available and classification schemes based on their mechanism and use have been published.s However, it is still not possible to predict accurately whether the enantiomers of a specific chiral drug will separate on a particular CSP. In addition, as these phases are often very expensive, and sometimes have limited lifetimes, interest has been sustained in the alternative field o f chiral mobile phase (CMP) additives. The number and therefore flexibility of potential additives as chiral selectors and achiral support phases is large and this can lead to differences in selectivity with respect to the equivalent CMP.These additives can include: amino acid derivatives,'.' certain proteinsxT9 and inclusion eomplcxing agents, such as cyclodextrins"'." and crown ethers. '' As yet, however, there has been no significant attempt to classify their operation or to suggest possible resolution accruing from their use. It is proposed to investigate the rationale for chiral separations of drugs and their analogues using CMPs. As a fundamental part of this on-going investigation specific model compounds have been selected and in particular four basic chiral barbiturates have been examined. This paper reports o n the preliminary work involving the use of fbcyclodextrin (cyclomaltoheptaose) and a number of its derivatives as chiral selectors in the mobile phase.The barbiturates, hexobarbitone (HEX), methylphenobarbitone (METH), quinalbarbitone (QUIN) and secbutobarbitone (SEC) have the structures shown in Fig. I . Both HEX and METH are methyl-substituted at the N(3) position of the pyrimidine ring with all four disubstituted at position C(S). Owing to the absence of a methyl substituent at N(3), QUIN and SEC are not chiral at C(S), but do possess chiral aliphatic side chains. Both HEX and METH, however, are chiral at C(5). F' N-H 0 R ' R 2 R 3 CH3 CH3 Hexobarbitone Quinalbarbitone H CHZzCHCH2 CHB(CH~)ZCH I CH3 CH3CH2CH I CH3 Secbutobarbitone H C2H5 Fig.1 Structures o f the chiral barbiturates102 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 Although some of the chiral barbiturates have previously been separated o n cyclodextrin bonded chiral stationary phases13 and additionally with cyclodextrins as mobile-phase a d d i t i ~ e s , ' ~ little information is available on the mechanisms of any resolution obtained using derivatized p-cyclodextrins, or the use of different achiral stationary phases. In our pro- gramme, comparisons were made between those barbiturates with chiral centres at C(5) and those with a chiral centre in the C(5) side chain to establish the relationship between the type of substitution and the efficiency of the separation obtained. These results could yield further data on the mode of interaction with cyclodextrins and the resolution of chiral drugs with chiral selectors added to the HPLC mobile phase.Experimental Instrumental System The LC system consisted of a dual reciprocating pump (LDC/ Milton Roy ConstaMetric 3000, Milton Roy, Riviera Beach, FL, USA), a sample injector (Model 7125, Rheodyne, Cotati, CA, USA or Waters WISP 712B, Waters Associates, Stock- port) a UV detector (PU 4025 or LC-UV, Pye Unicam, Cambridge) and either a single pen recorder (Kipp and Zonen BD40, Delft, The Netherlands) or a recording integrator (Hewlett-Packard 3388A, Avondale, PA, USA). Chromatographic System All the HPLC columns were 100 x 4.6 mm i.d. in size, and a number of stationary phases were utilized, including Hypersil CIx, Hypersil BDS (base deactivated silica); Clx, Hypersil Cx, Hypersil C2 or Hypersil diol (supplied by Shandon Scientific, Runcorn, Cheshire) and a Spherisorb CN-RP (Jones Chroma- tography, Pencoed, Mid-Glamorgan); all of the above con- tained 5 pm packing materials.Also used was a porous graphitic carbon (PGC) column containing 7 pm packing (Hypercarb, Shandon Scientific, Runcorn). Reagents Methanol (HPLC grade) was obtained from Rathburn Chemi- cals (Walkerburn , Peeblesshire) and disodium hydrogen phos- phate (AnalaR grade) from BDH (Poole, Dorset). The mobile phases consisted of solutions of disodium hydrogen phosphate at a number of concentrations and varying pH values. To these were added a range of concentrations of chiral selectors, which consisted of P-cyclodextrin, or a derivatized (3-cyclodextrin (supplied by Sigma, Poole , Dorset) .The chiral drugs examined were quinalbarbitone, methyl- phenobarbitone, secbutobarbitone and hexobarbitone (all supplied by Sigma, Poole, Dorset). Solutions of these were prepared in a 1 + 1 mix of methanol and phosphate buffer (pH Results and Discussion In the initial experimental programme the separation of all four barbiturates was investigated using a range of additives and experimental conditions. However, these experiments showed that separation of the enantiomers of QUIN or SEC was not possible on any of the following stationary phases: Clx, base deactivated C18 (BDS), C8, C2, diol, nitrile or PGC. Neverthe- less, partial resolution of the enantiomers of HEX and METH was achieved (Fig.2) in these initial experiments using a Hypersil CI8 column. In order to enhance the chromatographic performance, attempts were made to optimize the separation. Apart from conventional adjustment of the mobile-phase parameters a number of column packings were also examined. Using a newly marketed base-deactivated Clx column, en- hancement of chromatographic peak shape and subsequent resolution was attained (Fig. 3). The method of preparation of this phase has not yet been disclosed but the interactions between the analyte and the polar silica support matrix of the stationary phase are believed to be reduced. 6)- I 35 38.5 1 I Time/min --+ Fig. 2 Separation of the enantiomers of hexobarbitone using a Hypersil ODS column, a mobile phase of 10 mmol I-' disodium hydrogen phosphate (pH 4.0) and methanol (80 + 20 v/v) (containing 5 mmol 1- b-cyclodextrin) at a flow rate of 1 ml min-' and a detection wavelength of 240 nm 26 29 '1 I, Time/min - Fig.3 Separation of the enantiomers of hexobarbitone (u) and methylphenobarbitone ( h ) using a Hypersil BDS column under the Conditions specified in Fig. 2 (a) I I 23.3 12.6 Time/min - Fig. 4 Separation of the enantiomers of hexobarbitone using a porous graphitic carbon column and a mobile phase consisting of 10 mmol I-' disodium hydrogen phosphate (pH 4.0) and methanol (1 + 1 v/v) containing either ( a ) 5 mmol I-' methyl (3-eyclodextrin o r ( h ) 5 mmol I-' hydroxyethyl (3-cyclodextrin. other conditions as in Fig. 2 With PGC as column packing material, separation was also accomplished, although it was necessary to vary the mobile- phase composition due to the high retention characteristics of this hydrophobic stationary phase. As (3-cyclodextrin has a very poor solubility in the mobile phase with a high proportion of non-aqueous solvents, the more soluble methylated derivative was used and the resolution of the enantiomers of both HEX [Fig.4(a)] and METH was achieved.ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 103 With polar column packings (diol, nitrile), loss of chiral specificity was observed and similar results were obtained with non-polar stationary phases containing shorter bonded chain lengths than CIS. Therefore, although the primary interaction between the barbiturate drugs and (3-cyclodextrin is thought to exist within the hydrophobic cyclodextrin cavity, together with the attractive or repulsive effects of the secondary hydroxyl groups on the rim, there are obviously effects resulting from interchanges between the achiral stationary phases and the chiral selector.When comparing the use of different derivatized cyclo- dextrins, using a PGC column packing, a marked difference was observed when moving from methyl (3-cyclodextrin to other derivatives. With the bulkier hydroxyethyl (3-cyclo- dextrin [average molecular substitution (AMS) 1 .O] consider- ably reduced enantiomeric resolution [ Fig. 4(h)] was achieved. Use of the more highly substituted hydroxyethylf3-cyclodextrin (AMS 1.6), or hydroxypropyl (3-cyclodextrins, resulted in loss of resolution. These effects are considered to be due to the sites of attraction or repulsion being spatially distanced from the points of interaction of the analyte which thus leads to a loss of chi ral specificity .Conclusion The interaction of the resolved chiral barbiturates with (3- eyclodextrin probably involves the 'docking' of the cyclic side chain of these drugs with the hydrophobic cavity of the cyclodextrins. It is also likely that interaction of the pyrimidine ring structure with the exposed primary hydroxyl groups of the selector completes chiral recognition. Further work in this field will be carried out to examine the selector behaviour of cyclodextrins and allow the relationship between the chro- matographic behaviour of a series of analytes and their interaction with (3-cyclodextrin and its derivatives to be elucidated.The authors thank the Science and Engineering Research Council and Sterling Research Group for their financial support of this research programme and are grateful to Shandon Scientific for their gift of columns used in this work. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 References Williams. K., and Lee. E., Drugs, 1985, 30, 333. Bjorkman, S . . Br. J . Clin. Pliarmacol., 1985, 20. 463. Pettersson. C., and Gioeli. C., J . Chromatogr., 1988,435, 225. Bjorkman, S . , J . Chromutogr. (Biomed. Appl.), 1087,414.465. Testa. B., Xenohiorica, 1986. 16, 265. Berthod, A.. Heng Liang Jin, Becsley, T. E.. Duncan. J . D., and Armstrong, D. W.. J . Pliarm. Biomed. Anul., 1090.8. 123. Waincr, I. W., and Doyle, T. D., J . Chrorncrtogr. (Biomcd.Appl.), 1984. 306, 405. Wainer. I. W., A Practical Guide to the Selection and LW of HPLC Clriral Szurionary Phuses, J . T. Baker, Phillipsburg. NJ, USA, 1988. Levin, S . , and Grushka. E., Anal. Clzem., 1985. 57, 1830. Hermansson, J . , J . Cliromutogr., 1984. 316. 537. Clark, B. J . , and Mama, .I. E., J . Phurm. Biotned. Anul.. 1989, 7, 1883. Gazdag, M.. Szepcsi, G., and Huszar, L.. J . Cliromutog r . . 1988, 436. 31. Shinbo. T., Yamaguchi, T . , Nishimura, K . , and Masaaki, S . , J . Chromatogr. , 1987, 405. 145. Sybilska, 11.. and Zukowski, J.. in Cliiral Seprutions by HPL C, Applicutions to Pliurmaceictirul C'ompowids, cd. Krstu- lovic, A. M . , Ellis Horwood, Southampton, 1990, p. 147. Optimization of the Separation of Anthracyclines and Their Metabolites Using Reversed-phase Liquid Chromatography G.Nicholls, B. J. Clark and J. E. Brown* Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford BD7 7 DP The anthracycline antibiotics constitute a class of compounds with potent antitumour properties, and include doxorubicin (Adriamycin), a clinically important antineoplastic agent effective against a wide spectrum of malignancies. I There are several synthetic analogues of doxorubicin,' including the 4'-epimer epirubicin. Both of these drugs are extensively metabolized, with the main metabolic pathway involving C-13 carbonyl reduction to yield doxorubicinol and epirubicinol; further metabolism generates the 7-OH and 7- deoxy aglycones (Fig. 1). In order to investigate the toxicity of the anthracyclines and their metabolites, an analytical tech- nique capable of determining low plasma and tissue concen- trations of drugs and metabolites was required.In this field, reversed-phase high-performance liquid chromatography (HPLC) is generally the method of choice; normal-phase HPLC cannot separate the relatively non-polar aglycones, and ion-pair chromatography and gradient elution offer no advan- tage over a simple isocratic reversed-phase m e t h ~ d . ~ In the preliminary stage of method development using the HPLC method of Maessen et af.' two metabolites, epirubicinol and doxorubicinol aglycone, were not resolved [Fig. 2(a)] and formal optimization strategies were employed to improve separation? (i) the solvent selectivity triangle, an approach based on a simplex lattice design, which enables the selection of a mobile phase composition that best resolves the desired * To whom correspondence should be addressed.components within a suitable capacity factor range; (ii) factorial design, a procedure which uses a statistical experimen- tal design to elucidate the relative importance and interactions of experimental variables; (iii) modified simplex procedure, a directed search technique, with movements of the simplex (a geometric figure, with n + 1 vertices for n variables) being dependent on the ranking of responses after each experiment. The first two procedures are simultaneous techniques: all experimental points are predetermined and experiments are completed before data analysis. In this experimental pro- gramme, the modified simplex strategy (a sequential tech- nique) was used to define the optimal global conditions following the application of the two simultaneous procedures.Experimental The HPLC system comprised: LKB 2150 pump, LDC/Milton Roy Spectromonitor 3000 detector, Hewlett-Packard HP3396A integrator and 150 X 4.6 mm i.d. column packed with 5 pm Spherisorb ODS I . The initial mobile phase was CH3CN-0.02 mol I-' NaH2P04 (35 + 65 v/v), containing 0.1% v/v triethylamine, pH 4. The flow rate was 1 ml min-' and the detection wavelength was 254 nm. Solvent Selectivity Triangle Seven experiments were performed using three solvents: CH3CN, CH30H and tetrahydrofuran (THF), from different1 04 Parent drugs: ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 0 OH 0 Doxoru bicin bCH3 OH Metabolites: 0 OH OH Doxorubicinol kCH3 OH Epirubicin Epirubicinol 0 OH OH 0 OH 0 OH OCH3 0 7-OH-Doxorubicinol aglycone 7-OH-Doxorubicin aglycone OH OH OCH3 0 OH OCH3 0 OH ? 7-Deoxydoxorubicin aglycone Daunorubicin (internal standard) OH Fig.1 Structures of the anthracyclines and their metabolites selectivity classes. Solutions were calculated to have approxi- mately the same solvent strength using Snyder's equation .' Retention data were analysed using a statistical mixture design strategy to obtain a resolution map. Factorial Design Three variables were investigated at two levels: pH (3.2 and 5.2); %THF (0 and 5% v/v); and buffer concentration (0.05 and 0.2 mol I-'). Buffer composition was modified to an appropriate pK,. Citrate buffer (Na2HP04-citric acid) was found to be suitable, with all molarities referred to the Na2HP04 component.Combinations of variables are de- scribed by a cubic diagram, the eight corners of the cube indicating the experiments to be performed. Results were analysed using the separation factor (a) for the worst- separated pair of peaks in the initial chromatogram. Modified Simplex The two most important interacting variables in the preceding technique were investigated, using an initial simplex defined by three points: pH 5.2, buffer concentration 0.2 mol I-'; pH 3.2, buffer concentration 0.2 mol I-'; and pH 4.2, buffer concen- tration 0.1 mol I-'. Further movements of the simplex were dependent on thc responses obtained using a suitable chro- matographic response factor (CRF).The experiments were continued until a global optimum was reached. Results and Discussion Solvent Selectivity Triangle This technique was used primarily as a 'scouting' exercise to determine the most suitable type of organic modifier and its general composition. Four compounds were used: a parent drug (doxorubicin); a polar metabolite (doxorubicinol); a non- polar aglycone (7-OH-doxorubicin aglycone); and the internalANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 6.0 5.5 I05 I - 12 1 0.005 Fig. 2 b) nj. 1 r 2 4 6 8 10 12 14 n8 I I I I I I 1 2 4 6 8 10 12 14 Elution time/min HPLC separation of the anthracyclines and their major metabolites by ( a ) ihe method reported by Maessen et al.' and ( b ) after the application of structured optimization stratcgies.Mobile phase: 35% v/v acetonitrile4.06 mol I-' disodium hydrogen phosphate4.03 mol I-' citric acid containing 0.05% v/v triethylamine (pH 4.6); all other conditions as in the text. Chromatographic peaks: I . doxorubici- no1;,2, epirubicinol; 3, 7-OH-doxorubicinol aglycone; 4, doxorubicin; 5. cpirubicin; 6 , 7-OH-doxorubicin aglycone; 7, daunorubicin; and 8, 7-dcoxydoxorubicin aglycone. The ordinate shows 0.005 a.u.f.s. in both instances standard (daunorubicin). Triethylamine (0.1% v/v) was incor- porated into all mobile phases to reduce chromatographic peak tailing through silanol end-capping. Analysis of the retention data gave a triangular resolution map, indicating solvent combinations giving good resolution of components. Two areas o f high resolution were implied near the acetonitrile and THF apices.However, although a high proportion of THF in the mobile phase resulted in good peak separation, retention was extended (fK = 32 min) and peaks were broad. Examination of the chromatograms clearly indicated that acetonitrile was the optimal organic modifier. However, small amounts of THF (<5%) in the mobile phase gave improved peak separation, and this was examined in the next stage of the optimization procedure. Factorial Design In this part of the study, results were evaluated using the resolution of epirubicinol and 7-OH-doxorubicinol aglycone as the separation factor a. The effect of each variable was calculated from the following equations:' A = 0.25[(Yl + Y2 + Y3 + Y4) - (Y5 + Y6 + Y7 + YS)] B = 0.25[(Yl + Y2 + Y5 + Y6) - (Y3 + Y4 + Y7 + YS)] C = 0.25[(Y1 + Y3 + Y5 + Y7) - (Y2 + Y4 + Y6 + YS)] where A , B and C are the three variables (pH, %THF and buffer concentration, respectively) and Yi refers to the result of experiment i.A second level of derived equations gave an estimate of the variable interactions. The most important individual effect was buffer concen- tration, followed by %THF and then pH. However, the greatest interaction was between buffer concentration and pH. These two parameters were examined in a two-dimensional matrix within the final part of the optimization strategy. Modified Simplex The previous experiments enabled an area of factor space to be defined for optimal response. By using CH3CN as organic modifier (from the solvent selectivity triangle), the effect of altering pH and buffer concentration (from the factorial design) was examined.The basic simplex method uses fixed-step movements of a geometric figure (for two variables, a triangle) across the response surface, with movements being determined by the results obtained after each step. In this case a modified approach was used.6 Ranking of responses was dependent on (a) the separation factor of any partially resolved peaks and ( h ) the number of chromatographic peaks observed; this was reflected in the CRF used. Initially, three widely spaced points were chosen, as previously described. As the experiments progressed, the simplex was forced away from regions of poor response towards an area of optimal response. Rapid location of the suspected global optimum was achieved (point 7 in Fig.3), leading to a final mobile phase composition of 35% v/v acetonitrile-0.06 mol I - ' disodium hydrogen phosphate-0.03 mol I-' citric acid, containing 0.05% v/v triethylamine at pH 4.6. A further six experiments confirmed the area of optimal response. The final chromatogram [Fig. 2(b)] shows virtual baseline separation of all eight peaks in 13.7 min. In this case, the modified simplex procedure was able to locate rapidly the position of optimal response within a defined factor space, thereby producing an improvement in resolution of the anthraeyclines and their metabolites in HPLC analysis. Conclusions This study has successfully demonstrated the utility of formal optimization techniques in the HPLC analysis of anthra- cyclines and their metabolites.The separation of eight components was achieved in a run time of 13.7 min, only 2 min longer than the original chromatogram, where only partial 3.5 - 'I I 1 L 3.0 I I I I I l * l I I I 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 Buffer concentration/mol I ' Fig. 3 Diagram of the movements of the simplex, which was based on the variables pH and buffer concentrations, during the modified simplex procedure. Optimal conditions were found at point 7. For HPLC parameters see Fig. 2106 ANALYTICAL PROCEEDINGS, FEBRUARY 1993. VOL 30 component resolution was observed. It was shown that the most important variables in anthracycline HPLC analysis are pH and buffer concentration; careful manipulation of these parameters using structured optimization techniques should enable the separation of any other metabolites of interest.G. Nicholls thanks the Science and Engineering Research Council for their financial support of this research programme, and Farmitalia Carlo-Erba (Italy) for the donation of drugs and metabolites. References I Bourna, J., Hcijnen, J . H., Bult, A., and Underberg. W. J . M., Phurm. Weekbl., Sci. Ed., 1986, 8, 109. 2 Arcamone, F., Med. Res. Rev., 1984, 4, 153. 3 Beijnen, J . H., and Underbcrg, W. J. M., J. fharm. Biomed. A n d . 1988, 6, 677. 4 Maesscn, P. A., Mross, K. B., Pinedo, H. M., and Van der Vijgh, W. J. F., J . Chromatogr., 1987, 417, 339. 5 Bcrridge, J. C., Techniques for the Automated Optimization of HPLC Sepurations, Wiley-Interscience, Chichester, 1985, p. 55.6 Ncldcr, J. A., and Mead, R., Comput. J., 1965. 7, 308. Liquid Chromatographic Studies on the Potential Degradation of Preservatives in Formulated Drug Products Melanie J. Thompson, B. J. Clark and Anthony F. Fell* Pharmaceutical Chemistry, School of Pharmacy, University of Bradford, Bradford BD7 I DP M. L. Robinson International Development Laboratories, Bristol-M yers Squibb, Moreton, Wirral, Merse yside L46 10 W The alkyl esters of p-hydroxybenzoic acid (paraben), e.g., methyl, ethyl, propyl or butyl, are widely used as preservatives by the pharmaceutical, cosmetic and food industries. They are often used as combinations of two or more alkyl esters because together they have a synergistic effect in increasing their anti- microbial activity. They degrade by hydrolysis under alkaline and acidic conditions to form p-hydroxybenzoic acid (PHBA) which has little preservative action.This reaction is well documented in the literature. '.' However, several unknown peaks have been observed by liquid chromatography (LC) in experimental formulations of a new acetylcholine esterase (ACE) inhibitor containing both sorbitol and glycerol in combination with methyl and propyl paraben~.~ These unknown peaks were not attributable to the hydrolysis product or to the active drug component. It was postulated that they could arise from degradation involving an interaction between the parabens and the polyols sorbitol and/ or glycerol. Therefore a series of experiments was designed to investigate the principal factors responsible for the appearance of these unknown peaks.The key objectives were: ( a ) to confirm the degradation profile originally observed for an experimental drug formula- tion containing both methyl and propyl parabens with the polyols, using a drug-free solution at 60°C; ( h ) to study the degradation of methyl p-hydroxybenzoate (MHB) in the presence and absence of either sorbitol (S) or glycerol (G); ( c ) to adapt isocratic and gradient LC methods to analyse methyl hydroxybenzoate (MHB) and its degradation products; (d) to examine the effects of pH, temperature and concentration ratio of polyol to paraben on the stability of MHB in sorbitol and in glycerol using a factorial design. Experimental Experimental Design A drug-free solution corresponding to the same formula as that for the experimental oral formulation containing the parabens and the polyols was prepared and incubated at 60°C.A three- variable two-level factorial design was developed using eight solutions for each set of experiments for sorbitol (Sigma, Poole, Dorset) and for glycerol (May and Baker, Dagenham, Essex) (see Table 1). The MHB (Sigma, Poole, Dorset) was dissolved in 0.1 mol I-' K2HP04 (AnalaR, Merck, Poole, * To whom correspondence should be addressed Dorset) and the appropriate amount of polyol (sorbitol or glycerol) added before making up to volume (10.0 ml). The pH was adjusted to the predetermined values as required. Samples were taken of each solution before incubation and then after 47 h , 5 days, 10 days and 28 days, respectively. Suitable control solutions containing MHB alone, or sorbitol, or glycerol, under each of the specified conditions of temperature and pH, were also prepared.The samples were analysed using isocratic and gradient LC on a HP-1090 diode array system (Hewlett-Paekard, Wald- bronn, Germany). Diode-array detection was used with a view to further work on peak purity. Data were acquired and evaluated using the HP-9000 series workstation (with HPLC Chemstation software). Chromatographic Conditions The mobile phases for gradient and isocratic separations comprised: solvent A, 1% m/v potassium chloride (AnalaR, Merck, Poole, Dorset) and 0.1% v/v orthophosphoric acid (Hipersolv, Merck, Poole, Dorset) at pH 2.0; and solvent B, 100% methanol (HPLC grade, Rathburn Chemicals, Peebles- shire). The three mobile phases comprised: ( a ) isocratic, solvent A-solvent B (78 + 7 v/v); ( b ) isocratic, solvent A- solvent B (50 + 50 v/v); ( c ) linear gradient, from 100% A to 100% B over 65 min.The column used was 5-pm Hypersil ODs, 150 x 4.6 mm i.d. (Chromex, Cheshire) with a flow rate of 1 ml min-I. The sample was introduced on to the column using a Rheodyne Model 7125 injection valve (Rheodyne, Cotati, CA, USA) fitted with a 20 pl sample loop. Detection was effected at 210 nm and column temperature was ambient (about 21 "C). Evaluation of Chromatograms The chromatograms from the samples collected at 28 days were integrated and the peak areas put into the factorial design equations4 to assess the effect of each variable on each chromatographic peak of significance and any interaction between variables.Results and Discussion The drug-free solution subjected to 60°C for four weeks displayed several known peaks (Fig. 1) with retention times corresponding to earlier data obtained for a commercialANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 107 Table 1 Three-variable, two-level experimental design for the degradation of sorbitol and glycerol (polyol) Concen t ra t ion Concentration Ratio Solution MHB (%m/v) polyol (%m/v) polyol : MHB pH* Temperature/"Ct Sor-hitol- SI s2 S3 s4 s5 S6 s7 S8 Glycerol - GI G2 G3 G4 G5 G6 G7 ti8 0.05 0.15 0.05 0.15 0.05 0.15 0.05 0.15 0.05 0.15 0.05 0.15 0.05 0.15 0.05 0. IS * pH at ambicnt temperature (about 21 "C). -1 * 1 "C. 25.0 15.0 25 .0 15.0 25.0 15.0 25.0 15.0 5.0 7.5 5.0 7.5 5 .o 7.5 5.0 7.5 500: 1 100: 1 500: 1 100: 1 500: 1 100.1 500: 1 100: 1 100: 1 50: 1 100: 1 50: I 100: 1 50: 1 l o o : 1 5 0 : 1 3.0 3.0 9.0 9.0 3.0 3.0 9 .o 9.0 3.0 3.0 9.0 9.0 3.0 3.0 9.0 9.0 25 25 25 25 60 60 60 60 25 25 25 25 60 60 60 60 800 700 v) 2 200 lo: t L 10 20 30 40 50 60 Time/m in Fig. 1 stored for 4 weeks at 60°C using gradient elution Chromatogram of degradant peaks in a drug-free formulation 5 160 -"i 140 8 3 20 - 0 - I I 10 20 30 40 50 Ti me/min Fig. 2 Chromatogram of degradant peaks in solution S5 with high conccntration ratio, low pH and high temperature (see Table 1). PHBA, peak 6; MHB, peak 8 f~rmulation.~ It can therefore be assumed that these peaks are not attributable to the active drug component. The factorial design equations4 enable assessment of the effects and interaction of pH, temperature and concentration ratio on the formation of any degradant peak, based on peak areas.For sorbitol (Fig. 2), seven degradation peaks were 200 180 160 3 140 r- 120 s 100 0 C 80 2 60 Q 40 20 0 0 I I I I I 10 20 30 40 50 Time/mi n Fig. 3 Chromatogram of degradant peaks in solution G4 with low concentration ratio, high pH and low temperature (see Table I ) . PHBA, peak 3; MHB, peak 4 detected and assessed and three degradation peaks for glycerol (Fig. 3). For each set of experiments the MHB and PHBA peaks were identified by running standards separately and then co-chromatographing them with the samples used in the factorial design. For both glycerol and sorbitol every sample showed a reduction in the peak area for MHB compared with the initial sample ( t = 0) and an increase in the hydrolysis product.However, a number of other peaks were also observed. From the factorial design results (Figs. 4-7) it can be seen that the rank of each factor differs markedly from peak to peak. Thus, for sorbitol, the concentration ratio of polyol to paraben appears to be the most influential factor overall, whereas for glycerol the temperature is the most influential. For a given polyol the patterns of priority for the effects, i.e., the rank order, of each factor on each peak show some differences. For sorbitol, the pH appears to have the least effect on the formation of all peaks, while for glycerol the concentration ratio has the least overall effect. Of the six peaks assessed for sorbitol, four of them show the same rank order where molar ratio is the most important factor and pH the least important.The two remaining peaks show the same pattern asANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 108 3 2 a, 3 m Y C m - a 1 0 1 2 3 4 5 6 7 8 Peak number Fig. 4 B. tcmperature; and C, concentration ratio Results from factorial equations for sorbitol (effccts). A, pH; I 1 2 3 4 5 6 7 8 Peak number Fig. 5 Results from factorial equations for sorbitol (intcractions) pH/concentration ratio; B, tempcrature/concentration ratio; pH/temperature 4, C. each other, where temperature is the most important and pH the least important. For glycerol four peaks were assessed. Two of these show the same rank order where temperature is the most important factor and molar ratio the least important.The other two peaks show completely different patterns in their rank order. This could indicate that several different reactions are occurring in each case. However to confirm this, kinetic experiments would have to be undertaken. Moreover, the unknown peaks should be isolated and identified by standard analytical techniques. One mechanism that has been postulated for this reaction is a trans-esterification reaction occurring between the paraben and the p01yols.~ In theory, in addition to the PHBA, this would yield six trans-esters for sorbitol and two for glycerol. In this work two unknown peaks were observed for glycerol and six for sorbitol. From the relative sizes of the peaks observed it would appear that the unknown peaks are formed at different 3 2 a, 3 m Y C m III - 1 0 --- 1 2 3 Peak number 4 Fig. 6 and C. as for Fig. 4 Results from factorial equations for glycerol (effects). A 3 2 a, 3 m > Y K m lr - 1 0 1 2 3 Peak number 4 Fig. 7 B and C, as for Fig. 5 Results from factorial equations for glycerol (interactions) B A , reaction rates and may themselves be subject to further degradation. One study on this trans-esterification reaction between sorbitol and parabens has been reported' and other work has also reported this type of reaction in similar compounds .6 Conclusions The experiments performed confirm the formation of unknown peaks in the presence of paraben and sorbitol and/or glycerol. The pattern of peaks observed is similar to that seen in the experimental formulation. The factorial design work enabled the key factors affecting the formation of these peaks to be elucidated, viz., temperature and concentration ratio for sorbitol, temperature and pH for glycerol. Further work is required to isolate and identify these products and the mechanism kinetics should also be studied.ANALYTICAL PROCEEDINGS, FEBRUARY 1993. VOL 30 Subscription Details JAASbase 1993 Updates f99.00/$218.00 JAASbase Backfile (1987-1992) f230.00/$506.00 Idealist Software f 2 10.00/$462.00 Six updates will be issued at regular intervals through 1993. 109 Special Introductory Offer Take out a subscription to JAASbase Updates, buy the JAASbase Backfile and receive Idealist absolutely free! Offer available only until July 1993. The authors thank the Science and Engineering Research Council and Bristol-Myers Squibb for their financial support of this research programme. 2 3 4 5 References 6 I Raval, N. N . , and Parrott, E . L., J . Plzurm. Sci.. 1967. 56. 274. Blaug, S. M., and Grant, D. E . , J . SOC. Cosmet. Chem.. 1984.25, 495. Robinson, M. L., personal communication. Berridge, J . C.. Techniques for the Automated Optimiiution of HPLC Sc?pul-utions, Wiley, Chichester, 1985. Runesson. B.. and Gustavii, K., Actu Phurm. Suec., 1986. 23. 151. Irwin, W. J . , Masuda, Q. N., and Li Wan Po. A . , In!. J . Pharm., 1984, 21, 35. JAASbase A unique database of atomic spectrometry reference information for the practising analyst JAASbase is a new PC-based product from the The database consists of listings of published Royal Society of Chemistry designed to meet every atomic spectrometry papers and conference atomic spectroscopist’s need for a comprehensive, papers, and includes tabulated information relating yet inexpensive source of current analytical atomic to the application of relevant techniques. The spectrometry information. It contains over 20,000 references are easily searched with the database regularly updated references compiled from the manager Idealist which also enables the addition of atomic spectrometry literature. personal data to the database. KOYAL 5 0 1 IETY O F C HEMISTRY JAASbase is an invaluable tool for all practising analysts - order your copy today! @- To order JAASbase and for further information please contact: Sales and Promotion Department, Royal Society of Chemistry, Thomas Graham House, Science Park, Milton Road, Cambridge CB4 4WF, United Kingdom. && Tel: +44 (0)223 420066. Fax: +44 (0)223 423623. Telex: 818293 ROYAL. Information Services

ISSN:0144-557X    

DOI:10.1039/AP9933000093

出版商:RSC    

年代:1993

数据来源: RSC

摘要:

110 ANALYTICAL PROCEEDINGS. FEBRUARY 1993, VOL 30 Productivity Enhancement in Atomic Spectroscopy The following is a summary of one of the papers presented at a Joint Meeting of the Atomic Spectroscopy Group and the North East Region held on February 5, 1992, in the University of York. Appropriate Precision : Matching Analytical Precision Specifications to the Particular Application Michael H. Ramsey Department of Geology, Royal School of Mines, Imperial College, London SW7 2BP The pursuit of improved precision in chemical analysis is one of the fundamental objectives of analytical method development. The ultimate precision required of a method is, however, hard to define from a purely analytical perspective. Moreover, the achievement of higher precision becomes increasingly expen- sive, both to pursue in development and to sustain in the routine application of the method.What is required is the concept of appropriate precision rather than high precision as a target for method development. Appropriate precision of analysis is that which does not contribute significant error to the interpretation of the system under investigation. If, for example, the ‘true’ analyte concen- tration in a system varies over a small range, for example lo%, then a precise description of this variation will probably require a high analytical precision such as +0.5%. If, however, the true analyte concentration varies over a large range, for example three orders of magnitude, then this variation can be described precisely with a lower analytical precision such as 2 10%.This paper demonstrates that targets for analytical precision can be related quantitatively to the intended application, rather than be set at an arbitrary level. Moreover, it shows that the precision of sampling must also be measured if appropriate precision is to be achieved. The effective communication of the relative importance of random errors (in sampling and analysis) is an important objective for giving justifiable confidence to the users of chemical analysis. The benefits of achieving appropriate precision are exemplified and shown to be both scientific and economic. A Problem of Communication In considering how a target of analytical precision can be specified, a generalized case can be examined of the communi- cation between an analyst and a user of chemical analysis.Prior to the analysis the analyst may discuss several questions with the user. ‘What precision is required for this analysis?’ can rarely be answered quantitatively with any justification and can often be countered, quite reasonably by the user, with the reply: ‘How do I know until I’ve seen the results?’. A second question from the analyst, to clarify the first, may be: ‘What is the problem that you are trying to solve?’. It is often only from the answer to this question that the analyst can estimate qualitatively what analytical precision may be required. A crucial third question may be: ‘How were the samples taken’. It is rare that details of sampling procedures are known, even less that the errors introduced by the sampling have been estimated. After the analysis, questions often arise from the user such as: ‘What confidence can I have in those analyses?’, to which the answer may be given that ‘the analytical quality was within specification but no comment can be made on the quality of the sampling’.To the question: ‘Would it be worth spending more on further samples or alternative methods of analysis’, the analyst might ask: ‘Have the analyses solved the original problem?’ Overall it is apparent from these exchanges that neither the analyst nor the user has enough information to decide whether the analytical precision has been sufficient for solving the original problem. There is a communication problem between the analyst and the user that requires an integrated approach. Given significant input from the analyst into the original experimental design, and the sampling protocol, the analyst can provide answers to most of the questions including a decision on whether the analytical precision was appropriate.Sampling and Analytical Quality Control Scheme (SAX) A basic prerequisite to the concept of appropriate precision is an integrated approach to the measurement of random errors in both the processes of sampling and analysis. In the sampling and analytical quality control schemes (SAQCS = ‘SAX’) duplicate samples are taken ‘in the field’ and both of these samples are analysed twice as analytical duplicates. The statistical technique called analysis of variance (ANOVA) is applied to the chemical analyses to separate the random errors (as variances) due to sampling and chemical analysis.’-3 A target for appropriate precision can be set by comparison between the relative proportions of the ‘true’ analyte variation and the variances originating in the processes of measurement.Classical ANOVA separates the variances that can be combined in an additive fashion to give the total variance stota12 by Stota]?- = sg2 + ss2 + sa2 (1) where s,’ is the variance origination in the chemical analysis and sS2 is the variance origination in the sampling procedure. The other component variance sg2 describes the ‘true’ varia- bility of the analyte in the system, the description of which is an objective of the experiment. The analytical and sampling variances can be combined to give an over-all ‘technical’ variance Stech2 = .Fa2 + ss2 (2) Four assumptions underlie the rigorous applications ofANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 111 ANOVA:“ (1) the errors are random ( i e ., not systematic); (2) the means of the component population are additive; (3) the variances are homoscedastic (e.g., they should not be a function of concentration) and independent of each other (e.g., the analytical error is not a function of sampling error); and (4) the distribution of the random errors is Gaussian (normal). The validity of some of these assumptions for certain analytical measurements has been questioned. The use of robust ANOVA has been recommended to overcome some of these limitations and provide more accurate estimates of the component variances. 3-s Acceptable Levels of Sampling and Analytical Precision Ideally both the sampling and analytical variances, should be zero.This would give the variance of data (st,,t,12) as an exact description of the ‘true’ variance of the analyte (sg2). Kealisti- cally, because of the nature of eqn. (1) small levels of technical variance have a negligible effect on the total variance. Although a limit for the technical variance cannot be calculated exactly, experience suggests that the technical variance should not contribute more than 20% to the total ~ a r i a n c e . ~ If this figure is exceeded then distortions of the true variation in the analyte concentration become increasingly severe. If, how- ever, the technical variance contributes less than l% to the total variance, then any further reduction in technical variance, such as improved analytical precision, will not be discernible in the over-all interpretation.Such improvements would, there- fore, not be cost effective. A target for analytical precision is that it should not contribute significantly to the technical variance. This target can be specified, again from experience, at the level at which the analytical variance contributes less than 20% to technical variance. If the analytical variance contributes more than 20%, then it will tend to become a limiting factor in the technical variance. Combining these two targets, the analytical variance should not contribute more than 4% (20% of 20%) to the total variance. If, however, the technical variance is not excessive (<20%) then the analytical component may dominate the sampling component without serious consequences for the over-ail interpretation.The communication of the relative importance of these measurement errors to the users of analytical data is best conveyed graphically using a pie chart called a SAX diagram (Fig. 1). Because the variances are additive they can be represented by areas, and the true targets for acceptable levels by angles on the pie chart. The suitability of the chemical analyses for subsequent interpretation can than be judged by a user not familiar with these statistics. Applications of Appropriate Precision and SAX The concepts of appropriate precision and SAX are considered generally applicable, but initial applications have both been in aspects of geochemistry. In the application of SAX to mineral exploration3 the ‘true’ analyte variability corresponds to the geochemical variance (sg2).Stream sediments taken in Cornwall were analysed for Cu, Pb and Zn by flame atomic absorption spectrometry (AAS) after dissolution with nitric acid. All 58 sites were sampled in duplicate, with the location of the duplicates being separated by 5 m. This separation represents the potential location error of the site on the map, and exemplifies a general principle. The sampling duplicates should not be taken as close as possible to ‘reduce’ the estimated sampling error. They should be taken in such a way as to give a realistic estimate of how different the sample could have been, due to all the sources of random variability in both space and time.Robust ANOVA of the chemical analyses showed that the analytical precision for Pb was appropriate to this application (Fig. 2). The analytical variance accounted for 1.1% of the total variance, well within the 4% limit. The estimate of analytical Pb precision is 8 pg g-’; when expressed relative to the mean lead concentration (90 pg g-’) it gives a value for relative precision of 18% (at 95% confidence). This ‘poor’ analytical precision is probably due to the proximity of the Pb concentration to the analytical detection limit of AAS (2 pg g-I). This precision would fail a traditional arbitrary target for relative precison, such as <10%, but is clearly adequate for this application. The technical variance (sampling plus analytical) contributes 4.3% of the total variance, which is again well within the target of 20%.This demonstrates that the spatial distribution of lead shown by the chemical analysis is predominantly that of the ‘true’ geochemistry. It has not, therefore, been disturbed significantly by random errors in the sampling or analysis. However, the analytical variance does contribute 26% to the technjcal variance, exceeding the limit of 20%, If the technical variance needed to be reduced, for example in an area of lower lead variability, then improvements would have to be made in the precision of both the analytical and sampling procedures. In a second application of SAX, to environmental geo- chemistry, the particular importance of measuring sampling error becomes evident.‘ In the survey, 35 urban park soils Fig. 1 SAX diagram showing the maximum proportions of analytical and sampling variances expressed as a proportion of the total variance.The total variance expresses the spread of the measured concentrations of the analyte over the range of test materials. For reliable interpretation of the chemical analyses, the combined sampling and analytical variances should not exceed 20% of the total variance. The ‘true’ analyte variance, excluding any measurement error, should account, therefore, for at least 80% of the variance Fig. 2 SAX diagram showing appropriate levels of both sampling and analytical precision for Pb in a mineral exploration survey. The analytical variance is less than the 4% target and the technical variance (sampling + analytical) is less than 20% of the total variance.The ‘true’ analytc variability is measured as the geochemical variance, which. at 95.7%. is clearly suitable for a reliable geochemical interpretation112 ANALYTICAL PROCEEDINGS, FEBRUARY 1993, VOL 30 were taken in two traverses across Hampstead Heath, London. Duplicate samples were taken at 40% of the sites, again separated by a distance of 5 m. Lead was solubilized using HN03 and HCIO4 and determined using inductively coupled plasma atomic emission spectrometry (ICP-AES). The fre- quency distribution of the lead concentrations was positively skewed, which again required the use of robust ANOVA for the variance estimations. The technical variance on this survey was not appropriate, being 43.3% of the total, well over the 20% target (Fig.3). The analytical precision was appropriate, however, with the analytical variance being 0.1% of the total, well below the 4% target. The problem clearly originates in the sampling pro- cedure, not in the chemical analysis. The environmental interpretation of the lead determination must proceed with extreme caution. There is no reason to doubt the robust estimate of the mean Pb concentration of the soil (374 pg g-’) but it does have a wide confidence interval (sg = 117 pg g-I). The spatial distribution of the lead across the Heath cannot, however, be reliably described. To design an experiment that will give a reliable description of the lead distribution, improvements must be made in the precision of the sampling and not in the precision of the chemical analysis.The Benefits of Appropriate Precision There are clear scientific benefits from the concept of appropriate precision. Primarily the analyst must become aware of the broader objectives of the chemical analysis, in terms of the whole problem to be addressed. Furthermore, the procedures of sampling, previously often neglected by both Fig. 3 SAX diagram for an environmental survcy for Pb in urban soils. Although the analytical variance is well within the 4% target. the over-all technical variance is over the 20% target. The problem clearly lies in the precision of the sampling procedure. Reliable interprctation of the spatial distribution of the lead concentration bctwccn the soils cannot be made using thesc chemical analyses analysts and user, must be integrated into the experimental design, and also into the procedures for error estimation.Research into methods of improving analytical precision can be justified when there is quantitative proof of the requirement for reducing analytical variance. Conversely, techniques which produce relatively poor precision, such as trace analysis using laser-ablation ICP-AES, can be applied with specified confi- dence in the reliability of the interpretation. Thc economic benefits of appropriate precision are that resources are not wasted, producing highly precise determi- nations when they are not required. Less precise, and therefore less expensive, methods can be shown to be appropriate for the interpretation required. There can also be large financial penalties incurred in industry when erroneous interpretations are made (e.g., the acceptance of a batch of material), when the precision of the analysis was not appropriate for the task.Conclusion Appropriate precision is that precision which does not contribute significant random error to the estimation of the ‘true’ variation of the elemental concentration. Target levels for appropriate precision should not be set as fixed analytical values, but related to the ‘true’ variance of the elemental variability. The analytical variance should not contribute morc than 4% of the total variance. Sampling error should be estimated by the analyst in an integrated sampling and analytical quality control scheme (SAX). The analyst can then decide not only if the analytical precision is appropriate for the application, but also whether the sampling precison is also appropriate. The combined sampling and analytical variance should not exceed 20% of the total variance. Robust ANOVA is generally preferable to classical techniques for the estimation of sampling and analytical variance. A SAX diagram showing the proportion of the measurement error within the total element variation is an effective means of communication to users of chemical analysis. Appropriate precision provides a realistic and cost-effective target for the development and application of chemical analysis. I t has been successfully evaluated for geochemical studies but it seems probable that it will have useful application over many areas of analytical chemistry. References 1 Micsch. A. T., in Computers in the Mineral Industry, Part I , Publications on Geological Science, vol. 9, cd. Parks, G. A. Stanford University Press. 1964, pp. 156-170. Ramscy, M. H., Thompson, M., and Hale, M . , 1. Geochem. Expl., 1992, 44, 23. Eisenhart. C., Riornetrics, 1947. 3, 1. Analytical Mcthods Committee, Analyst, 1989, 114, 1699. Ramscy, M. H., Appl. Geoclzem., in the press. 2 Garrett. R. (3.. Econ. Geol., 1969, 64, 568. 3 4 5 6

ISSN:0144-557X    

DOI:10.1039/AP9933000110

出版商:RSC    

年代:1993

数据来源: RSC