摘要:

Analyst, July 1996, Vol. 121 l l l N Meeting Report The Royal Society of Chemistry’s Initiative for Emerging Professionals in Analytical Science, 1996 The Royal Society of Chemistry held the first initiative for emerging professionals in analytical science in Buxton on April 26-28. The meeting took place in the Palace Hotel, right in the heart of one of the world’s most popular national parks. The location was excellent, with the hotel well placed for an exploratory trip around Buxton. What a pity that we did not have time for such a trip, since the delegates were fully occupied with scientific discourse (and perhaps the occasional sip of refreshment). The format of the meeting was different to all the previous ones I have attended, but very successful never- theles s. Friday evening was a relaxed affair, beginning with drinks prior to an excellent dinner.The meeting was formally opened by the (retiring) President of the Analytical Division, Dr. Bryan Pierce. He outlined the importance of the meeting, but also stressed the need to enjoy ourselves during the event. Following the formal opening, we had an opportunity to address questions to interesting answers and subsequent lively discussions. They ranged from ‘Which scientist, living or dead, would you most like to have been a research assistant for and why?’ to ‘Why does the EPSRC seem so reluctant to fund analytical science?’. The time allotted for the session soon passed, and we had only got through half of the intended number of questions. The evening was rounded off by informal discussions over a few drinks.The organizers were kind enough to remind us that we had an early start the next day, but that did not curtail the discussion. Saturday morning started with a 7.30 am alarm call for all delegates. However, much to the surprise of the hotel staff (and perhaps the organizers as well), the majority of delegates were down for breakpdst by 7.35 am. The format of the event involved four sessions, each of which was started by a presentation from a guest speaker. Syndicate groups when then formed to discuss the topic. The composition of the syndicate groups was changed regularly, adding to the variety of views and opinions we encountered. In reality, the group dynamics reached equilibrium rapidly, making each of the syndicate sessions very lively. Friendly rivalry developed between some of us, which added a little extra ‘spice’ to some discussions.The first session focused on miniaturization and the guest speaker was Dr. Peter Fielden from the Department of Instrumentation and Analytical Science, UMIST. Miniaturi- zation was described as volume size reduction, and that enhanced performance can be observed at the microscopic level. The need for a multi-disciplinary response was clearly identified. The problems associated with miniaturization were highlighted, namely funding, widespread acceptance and some remaining technical difficulties. The key technical difficulty identified was sampling at the micro-level and the problem of getting a representative sample. The syndicate groups then discussed the subject, and many divergent views were ex- pressed.However, all felt that this was an important area, and that we must ensure further work takes place since there is much fundamental science yet to be investigated at the microscopic level. The second session focused on bioanalytical technology, and the guest speaker was Dr. David Cullen from the Cranfield Biotechnology Centre at Cranfield University. Here, the potential benefit and challenges of bioanalytical technology were introduced. It was shown how surface science, molecular engineering and fabrication methods all play an important part in the development of biosensor devices. This is particularly important when trying to develop miniaturized bioanalytical methods such as arrays of sensors to act as artificial ‘noses’ and ‘tongues’. The syndicate groups were then formed to discuss the topic further.The discussion within groups was varied, with some groups taking a very positive view of how progress could be made, and others, a somewhat pessimistic view of the potential contributed that smaller groups of researchers could make compared to multinational groups. The general view was that bio- and chemical sensors offer many exciting opportunities to broaden the scope of analytical science in everyday life. This concluded the morning session. The guest speaker for the afternoon session Professor Keith Bartle from the School of Chemistry at the University of Leeds, on the subject of separation science. The major development areas of separation science were identified, namely micro- column separations, selective GC detection employing the atomic emission detector, trace analysis by large volume injection GC, and true 2-D chromatography such as GC-GC. There was an interesting explanation of the problems which held back the rapid development of SFC, and that its advantages had been overlooked.Chiral separations were identified as a potential growth area for the future. The syndicate groups then formed to discuss this issue further. There was clear agreement that separations were not obsolete, but that more work had to focus on the basic science of separation processes rather than applications driven research. The final session of the day provided an opportunity to reflect on general discussions that had been held and try to solidify the ideas raised.There was much discussion required to reach any clear consensus, and on many topics we found several views existed. Dinner gave us a chance to relax after the sustained activity throughout the day. Someone had a strange sense of humour and decided at 7.30 am alarm call was too late. At 3.30 am the fire siren was activated; personally I felt it was a little too early for a Sunday start. After standing in the grounds of the hotel for 30 min while the fire brigade inspected the hotel for a fire we were allowed back inside. After this false start, Sunday started for real with an 8.00 am breakfast (for those who survived the merry-making of the night before). This was followed by the only session of the day, which was chemometrics. The guest speaker was Dr.Roy Tranter from Glaxo Wellcome operations, Barnard Castle. We found the presentation clarified the subject to a large extent and started to shred the mystical veil that surrounded chemometrics. The talk focused on how useful chemometrics is to the analytical scientist, and why it is important to understand the background of how chemometrics works. Students find chemometrics a very difficult subject to understand, and a number of reasons were proposed for this. The question of whether chemometrics should be included in all practical sessions was raised, since it would provide students more chance to use the technique, and hence become more familiar with it. The syndicate groups then discussed the subject, arriving at a popular conclusion of how little we understand about it, but realise that it is a valuable112N Analyst, July 1996, Vol.121 method and must be better understood. Many of us felt that most books written on the subject were incomprehensible to the non- specialist, hence preventing many from starting to understand the subject. The event concluded with a presentation by the syndicate groups of their views on the important issues in analytical science. The general views were that analytical science is multi- disciplinary, and its contribution to the maintenance of, and improvement to, the quality of life has been underestimated. Other issues, such as increasing the status of analytical science and analytical scientists was raised. Once the over-all views were obtained, the issue was narrowed down to two primary areas of concern and action. These are to be released as a formal communication to the analytical science world in general in due course. Tom McCreedy School of Chemistry University of Hull, UK NOMINATIONS FOR THE 1997 BENEDETTI-PICHLER AWARD The American Microchemical Society is soliciting nominations for the prestigious 1 997 Benedetti-Pichler Award. The award, established in 1966, is given annually to recognise mtstanding achievements in microanalytical chemistry. The award consists of a plaque and 2xpenses to attend the Eastern Analytical Symposium in Somerset, New Jersey, in November 1997 :o receive the award at a session to honour the awardee. Nominations or fbrther information, including at least two supporting letters should be sent no later than October 30, 1996 to: Dr Robert G. Michel Department of Chemistry University of Connecticut Storrs, CT 06269 USA Tel : +1 203 486 3143; Fax : +1203 486 2981; E-mail : MICHEL@UCONNVM.UCONN.EDU

ISSN:0003-2654    

DOI:10.1039/AN996210111N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 113N Future Issues Will Include Determination of Phosphate as Aggregates of Ton Associates by Light Scattering Detection and Application to Flow Injection- Mitsuko Oshima, Nobutake Goto, Joko P. Susanto, Shoji Motomizu Determination of Lithium as a Chemical Tracer and its Application to Flow Rate Measurements-Chang J. Park Determination of Complex Mixtures of Airborne Isocyanates and Amines. Part 1. Liquid Chromatography with Ultraviolet Detection of Monomeric and Polymeric Isocyanates as their Di-n-butylamine Derivatives-Gunnar Skarping, Marten Spanne, Hakan Tinnerberg, Marianne Dalene Deconvolution and Spectral Clean-up of Two Component Mixtures by Factor Analysis of Gas Chromatographic Mass Spectrometric Data-R. G. Brereton, Peter Hindmarch, Cevdet Demir Preparation of a Water-soluble Macrocyclic Ether and Its Application to the Improved Extraction-Separation of Alkaline Earths and Lanthanides as an Ion Size Selective Masking Reagent-Shigeo Umetani, Takayuki Sasaki, Quyen T.H. Le, Masakazu Matsui, Shigekazu Tsurubou Harmonization of Air Quality Measurements at European Union Level-Annette Borowiak, Emile De Saeger Development of a Tube Enzyme Immunoassay for On-site Screening of Urine Samples for the Presence of Beta- Agonists-Willem Haasnoot, Lucia Streppel, Geert Ca- zemier, Martin Salden, Piet Stouten, Martien Essers, Piet Van Wichen Bifunctional Cryptand Modifier for Capillary Electrophoresis in Analyses of Inorganic-Organic Anions and Inorganic Cations-Jeng-Shong Shih, Chyow-San Chiou Performance Improvements in the Determination of Mercury Species in Natural Gas Condensate Using an On-line Amalga- mation Trap or Solid-phase Micro-extraction with Capillary Gas Chromatography-Microwave Induced Plasma-Atomic Emission Spectrometry- James Snell, Wolfgang Frech, Yngvar Thomassen Polymer Analysis by Column Liquid Chromatography Coupled Semi-on-line with Fourier Transform-Infrared Spectrometry- Govert W.Somsen, Eduard J. E. Rozendom, Cees Gooijer, Nel H. Velthorst, U. A. Th. Brinkman Measurement Methods and Strategies for Non-infectious Microbial Components in Bioaerosols at the Workplace- Wijnand Eduard The NIR Optical Detection of Acids in Atmospheric Air by Phthalocyanine Dyes in Polymer Films-Luis E. Norena, Frank Kvasnik Certification of Reference Materials Related to the Monitoring of Aldehydes in Air by Derivatization with 2,4-Dinitrophe- nylhydrazine-Jan-Olof Levin, Roger Lindahl, Carola E.M. Heeremans, Koos Van Oosten On-line Solid-phase Extraction-Liquid Chromatography- Particle Beam Mass Spectrometry of Carbamate Pesticides- Jaroslav Slobodnik, S. J. F. Hoekstra-Oussoren, M. E. Jager, M. Honing, Ben L. M. Van Baar, U. A. Th. Brinkman Intercomparison of Tube-type Diffusive Sampling for the Determination of Volatile Hydrocarbons in Ambient Air- The0 L. Hafkenscheid, Jacques Mowrer Spectrophotometric Flow Injection Determination of Lead in Port Wine Using In-line Ion-exchange Concentration- Antonio 0. s. s. Rangel, Teresa I. M. s. Lopes, Raquel P. Sartini, E. A. G. Zagatto Selective Determination of the Holmium in Rare Earth Mixtures by Second Derivative Spectrophotometry with Benzoylindan- 1,3-dione and Cetylpyridinium Chloride-Wang Naixing, Si Zhikun, Jiang Wei, Qi Zhong-Cheng What is the Best Sorbent for Thermal Desorption?-R.H. Brown Effect of Redox State on the Response to Organic Vapours of Poly-N-(2-cyanoethyl)pyrrole Coated Thickness-shear Mode Acoustic Wave Sensors-David C. Stone, Zhiping Deng, Michael Thompson High Sensitivity Conducting Polymer Sensors-A. C. Partridge, P. Harris, M. Andrews Photolytic Spectroscopic Quantification of Residual Chlorine in Potable Waters-Mark Johnson, Paul Melbourne Macrocycles as Host Molecules and as Catalytic Sites-Ian 0. Sutherland, Mark Dolman, A. J. Mason, K. R. A. S. Sandanayake, Andrew Sheridan, Alastair F. Sholl Flow Injection-Fourier Transform Infrared Determination of Oil and Grease After Microwave-assisted Extraction-Miguel de la Guardia, Yasmina Daghbouche, S.Garrigues Extraction of Polycyclic Aromatic Hydrocarbons from Environmental Matrices: Practical Considerations for Super- critical Fluid Extraction-J. R. Dean Deconvolution of Overlapping Analytical Peaks by Means of the Fast Hartley Transform-P. R. Fielden, A. Economou, A. J. Packham COPIES OF CITED ARTICLES The Royal Society of Chemistry Library can usually supply copies of cited articles. For further details contact: The Library, Royal Society of Chemistry, Burlington House, Piccadilly, London WlV OBN, UK. Tel: +44 (0)171-437 8656. Fax: +44 (0) 17 1-287 9798. Telecom Gold 84: BUR210. Electronic Mailbox (Internet) LIBRARY@RSC.ORG. If the material is not available from the Society’s Library, the staff will be pleased to advise on its availability from other sources. Please note that copies are not available from the RSC at Thomas Graham House, Cambridge.

ISSN:0003-2654    

DOI:10.1039/AN996210113N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

114N Analyst, July 1996, Vol. 121 Technical Abbreviations and Acronyms The presence of an abbreviation or acronym in this list should NOT be read as a recommendation for its use. However, those defined here neednot be defined in the text of your manuscript. AAS ac A P ADC ANOVA AOAC ASTM bP BSA BSI CEN CPm CMOS c.m.c. CRM CVAAS cw CZE dc DRIFT DELFIA DNA EDTA ELISA emf ETAAS EXAFS EPA FAAS FAB dPm FAO-WHO FIR FT FPLC FPD GC GLC HGAAS HPLC ICP id INAA IR ISFET iv im IGFET ISE LC LED LOD LOQ atomic absorption spectrometry alternating current analogue-to-digital analogue-to-digital converter analysis of variance Association of Official Analytical Chemists American Society for Testing and Materials boiling point bovine serum albumin British Standards Institution European Committee for Standardization counts per minute complementary metal oxide silicon critical micellization concentration certified reference material cold vapour atomic absorption spectrometry continuous wave capillary zone electrophoresis direct current disintegrations per minute diffuse reflectance infrared Fourier transform spectroscopy dissociation enhanced lanthanide fluorescence immunoassay deoxyribonucleic acid ethylenediaminetetraacetic acid enzyme linked immunosorbent assay electromotive force electrothermal atomic absorption spectrometry extended X-ray absorption fine structure spectroscopy Environmental Protection Agency flame atomic absorption spectrometry fast atom bombardment Food and Agriculture Organization, far-infrared Fourier transform fast protein liquid chromatography flame photometric detector gas chromatography gas-liquid chromatography hydride generation atomic absorption high-performance liquid inductively coupled plasma internal diameter instrumental neutron activation infrared ion-selective effect transistor intravenous intramuscular insulated gate field effect transistor ion-selective electrode liquid chromatography light emitting diode limit determination limit of quantification World Health Organization spectroscopy chromatography analysis mP MRL mRNA MS NIR NMR NIST od OES PBS PCB PAH PGE PIXE PPt PPb PPm PTFE PVC PDVB QC QA REE rf RIMS r m S rpm RNA SCE SE SEM SIMS SIMCA s/N SRM STM STP TIMS TLC TOF TGA TMS tris TRIS uv UV/vIS VDU XRD XRF YAG Commonly Used Symbols M Mr r S U melting point maximum residue limit messenger ribonucleic acid mass spectrometry near-infrared nuclear magnetic resonance National Institute of Standards and Technology outer diameter optical emission spectrometry phosphate buffered saline polychlorinated biphenyl polycyclic aromatic hydrocarbon platinum group element particle/proton-induced X-ray parts per trillion (1012; pg g-1) parts per billion ( lo9; ng g-' parts per million (106; yg g-1) poly (tetrafluoroethylene) poly(viny1 chloride) poly(diviny1 benzene) quality control quality assurance rare earth element radio frequency resonance ionization mass spectrometry root mean square revolutions per minute ribonucleic acid saturated calomel (reference) electrode standard error scanning/surface (reflection) electron microscopy secondary-ion mass spectrometry soft independent modelling of class signal-to-noise ratio Standard Reference Material scanning tunnelling (electron) standard temperature and pressure thermal ionization mass spectrometry thin-layer chromatography time-of-flight thermogravimetric analysis trimethylsilane 2-amino-2-( hydroxymethy1)- propane-l,3-diol (ligand) 2-amino-2-(hydroxymethyl)- propane- 1,3-diol (reagent) ultraviolet ultraviolet-visible visual display unit X-ray diffraction X-ray fluorescence yttrium aluminium garnet emission analogy microscopy molecular mass relative molecular mass correlation coefficient standard deviation atomic mass

ISSN:0003-2654    

DOI:10.1039/AN996210114N

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (889-896) 889 Nitrogen Factors for Sheepmeat Part 2.* Lamb Meat Analytical Methods Committee7 Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK W1 V OBN Nitrogen factors (the percentage nitrogen on a fat-free basis) for lamb meat were estimated using a total of 81 lamb carcases from the UK. The samples included carcases representative of the EU class ranges for conformation and fatness. The sampling period extended over one year in order to include the three age groups under which lamb meat is marketed. The results obtained were compared with those reported for mutton in Part 1 of this study and a factor of 3.50 for the whole side, on a fat-free basis, for the lean with intermuscular fat was recommended as an overall figure for mutton and lamb.There were significant differences between the factors for some joints and different factors were recommended for use when the type of joint used was known. Keywords: Sheepmeat; mutton; lamb meat; nitrogen ,factor The Analytical Methods Committee has received and has approved for publication the following report from its Nitrogen Factors Sub-committee. Report The constitution of the Sub-committee responsible for the preparation of this report was: Professor R. A. Lawrie (Chairman), Miss I. A. Agater (until May 1995), Dr. R. A. Allen (from May 1995), Mr. R. A. Evans (until May 1995), Mr. D. J. Favell, Mr. M. W. Fogden, Mr. J. Grant (from May 1995), Mr. A. J. Harrison, Mr. N. Harrison, Mrs. D. B. Homer, Dr. R. B. Hughes, Dr. A.J. Kempster, Mr. R. S. Kirk, Mr. C. R. Morrison (from August 1994), Mr. T. O’Dea (from August 1994), Dr. K. Pickering (from May 1995), Dr. R. Wood (MAFF Project Officer), Dr. M. L. Woolfe (MAFF Project Officer) and Mr. J. J. Wilson (Secretary). Introduction The UK is the largest producer of sheepmeat in the European Union and has a high per capita consumption at over 6 kg per year. Recent surveys show an increase of the use of sheepmeat by the manufacturing and catering sectors, 10% of the sheep slaughtered in the UK now being used in processed products. Limited information is available, however, on the composition of British sheepmeat. Although data on the composition of sheepmeat from various countries has been published, these cannot be related confidently to British breeds, production systems and populations, especially as the methods used for sample preparation and subsequent chemical analysis are unclear or non-standard. Since the Analytical Methods Com- mittee (AMC) had recently published updated values for the nitrogen factors for pork2 and beef,3 it was deemed appropriate ‘ For Part 1, see ref.1. i- Correspondence should be addressed to the Secretary, Analytical Methods Committee, Analytical Division, The Royal Society of Chemistry, Burlington House, Piccadilly, London, UK WlV OBN. that the Committee should obtain data for the nitrogen fdctors of sheepmeat under similar, strictly controlled, conditions. In assessing the nitrogen factors for sheepmeat, the Nitrogen Factors Sub-committee identified two categories, viz, mutton and lamb.The term ‘mutton’ herein refers to meat from female sheep which have lambed or which have reached a stage of pregnancy which depreciated the value of the carcase. The term ‘lamb’ refers to three categories of animals as follows: ( a ) milk-fed, i.e., lambs slaughtered before weaning, typi- cally under 4 months of age (and may also include some concentrate-fed lambs of similar age); (h) main season, i.e., lambs slaughtered in the IJK during the main slaughtering period, July to November, typically 6-10 months of age; ( c ) hoggets, i.e., lambs of about 1 year of age (when showing the emergence of the first pair of permanent incisor teeth). The investigation was carried out in two parts. In the first the composition of mutton was determined and the findings were published.1 The present paper gives findings on the composition of lamb meat, compares these with the values for mutton and gives recommendations on nitrogen factors for sheepmeat.Experimental The Nitrogen Factors Sub-committee worked closely with the Meat and Livestock Commission (MLC) and four specified independent laboratories on the design and implementation of Table 1 Number of carcases selected on each of the visits. (Each group of 3 made up a batch) Fat class* Lamb type Visit 2-3L 4L-4H Main season 1 3F, 3Mi’ 3F 2 3M 3F, 3M 3 3F, 3M 3M Hoggets 1 3F, 3M 3M 2 3F 3F, 3M 3 3M 3F, 3M Milk-fed/early season 1 3F 3F, 3M 2 3F, 3M 3M 3 3F, 3M 3F * Relationship between the fat classes used in commercial classification and a visual assessment of carcase subcutaneous fat content to the nearest percentage unit (SFe) SFe (%) EU fat classes Range 1 < 6.0 2 6.0-9.9 3L 10.0-11.9 3H 12.0-13.9 4H 16.0-17.9 5 > 18.0 4L 14.0-1 5.9 t M, male carcases; F, female carcases.Mean 4.0 8.0 11.0 13.0 15.0 17.0 20.0890 Analyst, July 1996, Vol. 121 this study, in all respects, as it related to nitrogen factors for Services, Bristol; and Tayside Regional Council, Public Analyst sheepmeat. The four laboratories, chosen as representative of Department, Dundee. industry, government and enforcement interests, were: Unilever Particular attention was paid to full compliance with written Research, Sharnbrook, Bedford; Laboratory of the Government detailed protocols (approved by the Nitrogen Factors Sub- Chemist, Teddington, Middlesex; Avon County Scientific committee) for carcase selection, dissection, sample prepara- Table 2 Chemical composition of lamb lean tissue* (% of tissue) Lamb type Main Joint Milk-fed season Leg and chump- Total fat 4.3 5.3 Moisture 74.6 73.7 Nitrogen 3.28 3.24 Fat-free nitrogen 3.43 3.42 Ash 1.10 1.08 Hydroxyproline 0.19 0.17 Total fat 7.1 8.0 Moisture 72.0 70.8 Nitrogen 3.29 3.31 Fat-free nitrogen 3.54 3.60 Ash 1.07 1.03 Scrag, shoulder-middle neck and breast- Total fat 7.6 9.2 Moisture 72.2 71.0 Nitrogen 3.15 3.15 Fat-free nitrogen 3.41 3.48 Ash 1.05 1.03 Hydroxyproline 0.22 0.23 Loin and best-end neck- Hydroxyproline 0.21 0.22 Hoggets 5.5 73.4 3.23 3.41 1.05 0.26 8.5 70.6 3.23 3.58 0.99 0.20 9.2 71.1 3.07 3.38 1 .oo 0.22 Differences between type means SE 0.23 0.24 0.029 0.03 1 0.017 0.052 0.33 0.36 0.029 0.041 0.016 0.012 0.45 0.48 0.032 0.041 0.017 0.015 Significance? ** ** ns ns ns ns * * ns ns ** ns * ns ns ns ns ns Overall mean 5.1 73.8 3.25 3.42 1.08 0.19 7.9 71.0 3.29 3.58 1.03 0.21 8.9 71.3 3.13 3.44 1.03 0.23 SE 0.13 0.14 0.017 0.018 0.010 0.030 0.19 0.21 0.017 0.024 0.009 0.007 0.26 0.28 0.018 0.024 0.010 0.008 * In Tables 2-9, the statistical significance of differences between the lamb type means was determined by analysis of the variance ratio of mean squares (Student’s t test).t Significance: ns, no significant difference between lamb types, *P < 0.05, **P < 0.01, ***P < 0.001. Table 3 Chemical composition of lamb total fatty tissue* (% of tissue) Lamb type Differences between type means Main Joint Milk-fed season Hoggets Leg and c h u m p Total fat 71.0 76.1 74.4 Moisture 20.2 16.9 18.2 Nitrogen 1.35 1.16 1.23 Fat-free nitrogen 4.58 4.86 4.8 1 Ash 0.34 0.28 0.26 Hydroxyproline 0.48 0.46 0.48 Total fat 79.6 82.7 80.5 Moisture 15.2 11.8 14.0 Nitrogen 0.90 0.82 0.90 Fat-free nitrogen 4.37 4.84 4.69 Ash 0.27 0.21 0.24 Hydroxyproline 0.34 0.27 0.26 Total fat 73.5 76.4 75.3 Moisture 19.7 17.2 18.5 Fat-free nitrogen 4.01 4.27 4.04 Ash 0.32 0.26 0.27 Hydroxyproline 0.34 0.34 0.3 1 Loin and best-end neck- Scrag, shoulder-middle neck and breast- Nitrogen 1.07 1.01 1 .oo * Intermuscular and subcutaneous fat.t Significance, as Table 2. SE 1 S O 1.09 0.094 0.176 0.025 0.037 1.87 1.32 0.087 0.290 0.027 0.039 1.37 1.10 0.07 1 0.134 0.014 0.025 Significance? ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns * Overall mean 74.8 17.8 1.21 4.80 0.29 0.47 81.6 12.9 0.85 4.72 0.23 0.28 75.6 17.9 1.01 4.17 0.27 0.33 SE 0.87 0.63 0.054 0.102 0.014 0.021 1.08 0.76 0.050 0.167 0.016 0.022 0.79 0.63 0.041 0.077 0.008 0.014Analyst, July 1996, Vol.121 89 1 tion, packaging and labelling, analytical test methods and quality control. Eighty-one lamb carcases were included in the trial, these were selected at three abattoirs. These carcases were rep- resentative of milk-fed lambs (and may include some concen- trate-fed lambs), main season lambs and hoggets in equal proportions. Carcases were selected at the abattoir in batches of three, to fall into specific fat-class cells (Table 1). Fat classes are based on the amount of subcutaneous fat cover on the carcase.Conformation, a visual assessment of carcase shape (a charac- Table 4 Chemical composition of lamb lean and intermuscular fat* (% of tissue) Differences between Lamb type type means Main Joint Milk-fed season Leg and chump- Total fat 8.4 10.6 Moisture 71.3 69.4 Nitrogen 3.16 3.08 Ash 1.05 1.02 Hydroxyproline 0.21 0.19 Total fat 15.8 18.6 Moisture 65.1 62.5 Nitrogen 3.00 2.96 Fat-free nitrogen 3.56 3.64 Ash 0.97 0.92 Scrag, shoulder-middle neck and breast- Total fat 18.3 22.4 Moisture 63.7 60.4 Nitrogen 2.81 2.73 Fat-free nitrogen 3.44 3.52 Ash 0.93 0.88 Hydroxyproline 0.24 0.25 Fat-free nitrogen 3.45 3.45 Loin and best-end neck- Hydroxyproline 0.22 0.22 Hoggets SE 10.3 0.39 69.6 0.35 3.09 0.031 3.44 0.031 0.99 0.016 0.28 0.048 18.3 1.13 62.9 0.97 2.91 0.041 3.56 0.035 0.89 0.019 0.20 0.012 20.6 1.09 62.0 0.95 2.71 0.042 3.42 0.041 0.87 0.017 0.24 0.015 Significancet *** *** ns ns ns * ns ns ns ns ns * * ns ns ns ns ns Overall mean 10.1 69.8 3.10 3.45 1.02 0.21 18.0 63.0 2.95 3.61 0.92 0.22 21.3 61.3 2.74 3.48 0.89 0.24 SE 0.22 0.20 0.018 0.018 0.009 0.028 0.65 0.56 0.024 0.020 0.01 1 0.007 0.63 0.55 0.024 0.024 0.010 0.009 * Combining the chemical composition of the lean tissue and fatty tissue, in the proportions of lean tissue and intermuscular fat tissue only, present in each joint.t Significance, as Table 2. Table 5 Chemical composition of lamb lean and total fatty tissues* (% of tissue) Joint Differences between Lamb type type means Main Milk-fed season Hoggets SE Significance? Leg and chump- Total fat 15.1 18.6 17.8 1.07 ns Moisture 65.8 63.0 63.6 0.88 Nitrogen 2.96 2.85 2.86 0.045 ns Fat-free nitrogen 3.49 3.50 3.49 0.033 ns Ash 0.98 0.93 0.91 0.019 * Hydroxyproline 0.23 0.22 0.30 0.043 ns Total fat 26.3 30.9 29.6 2.47 ns Moisture 57.0 52.8 54.1 1.98 ns Nitrogen 2.65 2.54 2.55 0.083 ns Fat-free nitrogen 3.59 3.69 3.62 0.046 ns Ash 0.86 0.78 0.77 0.030 ns Hydroxyproline 0.24 0.23 0.21 0.012 ns Total fat 23.9 28.8 26.5 1.65 ns Moisture 59.2 55.3 57.3 1.37 ns Nitrogen 2.63 2.53 2.52 0.057 ns Fat-free nitrogen 3.46 3.55 3.44 0.042 ns Ash 0.87 0.81 0.81 0.020 ns Hydroxyproline 0.25 0.26 0.24 0.015 ns * Loin and best-end neck- Scrag, shoulder-middle neck and breast- Overall mean 17.8 63.6 2.87 3.50 0.94 0.24 29.8 53.8 2.56 3.66 0.79 0.23 27.4 56.4 2.54 3.51 0.82 0.25 SE 0.62 0.5 1 0.026 0.019 0.01 1 0.025 1.43 1.14 0.048 0.027 0.017 0.007 0.95 0.79 0.033 0.024 0.012 0.009 * Combining the chemical composition of lean tissue and total fatty tissue present in each joint. t Significance, as Table 2.892 Analyst, July 1996, Vol.121 teristic in the national carcase classification scheme) was measured on the slaughter line but did not form part of the selection criteria. Dissection and Chemical Analysis The left side of each carcase was divided into seven joints: leg; chump; loin; breast; best-end neck; shoulder-middle neck; and scrag (see Fig. 1 of Part 1). Each joint was separated into its component soft tissues (i.e., lean, intermuscular fat and subcutaneous fat) waste (see Appendix) and bone.The macerated lean of the leg and chump, from each of the three carcases that .formed a batch, were combined and homogenized to form a single sample. Similarly the lean from the loin and best-end of neck joints, and that from the shoulder-middle neck, scrag and breast joints of the three carcases of each batch, were combined and homogenized. Fatty tissue (subcutaneous and intermuscular together) from each of the three (multiple) joints was homogenized from the three carcases in each batch. There was thus a total of three lean and three fat samples from each batch of three carcases. The samples of lean and fat were stored at -18 "C and subsequently analysed for total fat, moisture, ash, hydroxy- proline and nitrogen by the four laboratories listed above.The methods used were those of BS 44014 or approved variants thereof which had been shown to give equivalent results. The estimated chemical composition of the lean with fatty tissue was constructed mathematically from the components of the original tissues. Similarly the chemical composition of the forequarter, hindquarter and side was calculated. Results The national slaughter population of lambs consists of ca. 59% of main season lambs, 17% of milk-fed and early-season lambs and 24% of hoggets. The data designated 'overall' in Tables 2-9 below were obtained by weighting the values for each type of lamb in proportion to its contribution to the national popula- tion. The chemical composition of each of the three (composite) joints, from each of the three types of lamb, and for the lambs overall, is given for lean tissues, for total (subcutaneous and intermuscular) fatty tissue, for lean with intermuscular fatty tissue and for lean with total fatty tissues in Tables 2-5, respectively.Since the principal purpose of this investigation was to identify currently appropriate values for nitrogen factors in the meat, comments on the results will be mainly confined to this parameter. Although the fat-free nitrogen values, for lean with inter- muscular fat, were similar between types, those for the main season lambs tended to be higher in all three joints. The fat-free nitrogen contents of the lean with intermuscular fat of the loin with best-end neck joint was significantly greater than those for the other two joints. When the data from the joints were mathematically combined in proportion to their contribution to the total meat in the carcase (Table 6), it was again evident that Table 6 Chemical composition of lamb: combined carcase meat (% of tissue) Lamb type Tissue Lean tissue- Total fat Moisture Nitrogen Fat-free nitrogen Ash Hy droxyproline Total .futty tissue- Total Tat Moisture Nitrogen Fat-free nitrogen Ash H ydroxyproline Main Milk-fed season 6.3 7.5 73.0 72.0 3.22 3.21 3.44 3.48 1.07 1.05 0.21 0.21 74.2 77.8 18.8 15.9 1.10 1 .OO 4.23 4.53 0.3 1 0.25 0.37 0.35 Hoggets 7.7 71.9 3.15 3.42 1.02 0.24 76.3 17.4 1.03 4.38 0.26 0.3 1 Lean and interniusculur.fat portion of totulJatty tissue- Total fat 15.2 18.3 17.5 Moisture 65.9 63.4 64.1 Nitrogen 2.94 2.87 2.85 Ash 0.97 0.93 0.91 Hydroxyproline 0.23 0.23 0.25 Total fat 21.4 25.7 24.1 Moisture 61.0 57.5 58.8 Nitrogen 2.75 2.64 2.64 Ash 0.90 0.84 0.83 Hydroxyproline 0.24 0.24 0.26 Fat-free nitrogen 3.47 3.52 3.45 Lean and total fatty tissue- Fat-free nitrogen 3.49 3.55 3.49 Differences between type means SE 0.28 0.3 1 0.022 0.026 0.010 0.023 1.46 1.12 0.07 1 0.107 0.014 0.027 1.38 1.16 0.044 0.027 0.0 17 0.022 1.62 1.33 0.054 0.027 0.019 0.020 Significance* ** ns ns ns ** ns ns ns ns ns h ns ns ns ns ns * ns ns ns ns ns Overall mean 7.3 72.1 3.20 3.46 1.05 0.22 76.8 16.8 1.02 4.44 0.26 0.34 17.6 64.0 2.88 3.49 0.93 0.23 24.6 58.4 2.66 3.53 0.85 0.24 SE 0.16 0.18 0.01 3 0.0 15 0.006 0.013 0.84 0.65 0.041 0.062 0.008 0.015 0.80 0.67 0.025 0.0 15 0.010 0.012 0.94 0.77 0.03 1 0.016 0.01 1 0.01 1 * Significance, as Table 2.Analyst, July 1996, Vol.121 893 the fat-free nitrogen values for the lean with intermuscular fat did not differ significantly between lamb types, although the values for main season lambs were again somewhat greater than those for milk-fed lambs and hoggets. For the lamb carcases overall, the mean fat-free nitrogen content for lean with intermuscular fat was 3.50%. This compares with the values of 3.50 and 3.65% for pork2 and beef3 carcases, respectively, previously published by the AMC. In Part 1 of this investigation,’ the joints comprising the forequarters (viz., breast, best-end neck, shoulder-middle neck and scrag) and those comprising the hindquarters (viz., leg, chump and loin) were combined before chemical analysis.To permit ready comparison with the results for mutton, the results for lamb joints were mathematically combined to give the results for forequarters and hindquarters. These are shown in Tables 7 and 8, respectively. It is evident that there were no significant differences between the quarters in the values for fat- free nitrogen with intermuscular fat which approximated to 3.50% in each case. The physical composition of the lamb cuts, quarters and overall carcases are given in Table 9. Milk-fed lambs had a higher proportion of lean and bone, in each of the cuts, than the other two types. In addition, the hoggets had a higher proportion of lean, and less fat and bone than main season lambs. Overall, ca. 63% of the side was lean tissue. In Part 1 of this investigation, the chemical composition of the lean, of the fatty tissues and of the lean with fatty tissues of the mutton were given in Tables 3,4 and 5, respective1y.l ‘Fatty tissues’ in Table 4 of that publication referred to the combined subcutaneous and intermuscular fats, but in Table 5 the ‘lean and fatty tissue’ referred to the mathematical recombination of the lean with only the subcutaneous fat portion of the total fatty tissue.To clarify the position, to conform with the style of the reports on beef and pork, and to permit ready comparison of the data on lamb with those on mutton, the chemical composition of the lean with the intermuscular portion of the fatty tissues, and of the lean with total fatty tissues (subcutaneous and inter- muscular) are given for mutton in Tables 10 and 11, re- spectively.The fat-free nitrogen contents for the lean tissue with intermuscular fat, and for the lean tissue with intermuscular and subcutaneous fats, of the forequarters, hindquarters and side, for both mutton and lamb, and of the joints of lamb have been collected in Table 12. No significant differences were found between the nitrogen factors for lamb and mutton. The nitrogen factor for sheepmeat, based on the results given in Parts 1 and 2 of this investigation, is 3.50. This figure is similar to the mean (3.44) for the fat-free nitrogen content derived from 1000 analyses of sheepmeat published in the scientific literature, and based on values from various countries throughout the world (but for which the methods of analysis were not standardized).Recommendations On the basis of these results The Nitrogen Factors Sub- committee makes the following recommendations, to be applied as appropriate: (i) A nitrogen factor of 3.50 (on a fat-free basis) for the lean with intermuscular fat, is appropriate for sheepmeat (lamb and Table 7 Chemical composition of lamb forequarter meat* (% of tissue) Differences between Lamb type type means Tissue Lean tissue- Total fat Moisture Nitrogen Fat-free nitrogen Ash Hy drox yproline Total fatty tissue- Total fat Moisture Nitrogen Fat-free nitrogen Ash H ydrox yproline Main Milk-fed season 7.6 9.0 72.1 70.9 3.17 3.18 3.43 3.49 1.05 1.03 0.22 0.23 74.6 77.5 18.9 16.3 1.04 0.98 4.05 4.34 0.3 1 0.25 0.34 0.33 Hoggets 9.1 71.0 3.09 3.40 1 .oo 0.22 76.3 17.7 0.98 4.13 0.26 0.29 Lean and intermuscular fat por Total fat 18.4 Moisture 63.6 Nitrogen 2.82 Fat-free nitrogen 3.46 Ash 0.93 Hydroxyproline 0.24 Lean and total fatty tissue- Total fat 24.8 Moisture 58.5 Nitrogen 2.62 Fat-free nitrogen 3.48 Ash 0.86 Hydroxyproline 0.25 ,tion of total fatty tissue- 22.5 20.8 60.2 61.7 2.74 2.72 3.54 3.44 0.88 0.87 0.25 0.23 29.7 27.5 54.5 56.4 2.5 1 2.5 1 3.57 3.46 0.80 0.80 0.26 0.24 SE 0.42 0.45 0.029 0.037 0.015 0.0 13 1.42 1.12 0.069 0.118 0.015 0.025 1.12 0.96 0.041 0.037 0.017 0.014 1.78 1.47 0.060 0.038 0.021 0.014 Significance+ * ns ns ns ns ns ns ns ns ns ns * * ns ns ns ns ns ns ns ns ns ns * 0 v e r a 11 mean 8.8 71.1 3.16 3.46 1.03 0.23 76.7 17.1 0.99 4.24 0.26 0.32 21.4 61.1 2.75 3.50 0.93 0.24 28.3 55.6 2.53 3.53 0.8 1 0.25 SE 0.24 0.26 0.017 0.02 1 0.009 0.007 0.82 0.65 0.040 0.068 0.009 0.014 0.64 0.55 0.024 0.02 1 0.0 10 0.008 1.03 0.85 0.034 0.022 0.012 0.008 * Breast, best-end neck, shoulder-middle neck and scrag joints.+ Significance, as Table 2.894 Analyst, July 1996, Vol. 121 mutton) generally and should be used in the analysis of sheepmeat products. (ii) The factors shown in Table 12 are applicable to specific joints, quarters and sides, respectively, when the source of the sheepmeat is known. The AMC gratefully acknowledges the financial support given by MAFF and the MLC to the work of the Sub-committee. Appendix Waste Material in Carcase Dissection In all publications in The Analyst produced from August 1986 to date, on new or up-dated nitrogen factors for pork, beef and ~heepmeat,l-37~ it has been stated that carcase dissection, for the production of samples for analysis, was carried out in accordance with detailed protocols laid down by the MLC and approved by the Nitrogen Factors Sub-committee.Copies of these are available from the MLC.6 During the course of these dissections, in which cuts from carcases are further reduced to lean meat, intermuscular fatty tissues and subcutaneous fatty tissues, a small amount of material, other than bone, is produced and categorized as ‘waste’. Its nature is described in the specifications for each of the three species, and it consists mainly of sinews, heavy connective tissues (such as Zigamentum nuchae), glands and readily accessible blood vessels. The proportion of this waste material, when expressed as a percentage of the lean meat (plus intermuscular fat) of a carcase or cut, varies between approximately 1% for pork, 1.5% for lamb to 2% for beef, being rather higher in the forequarter part of the carcase due to the greater presence of cartilagenous protein in this area.During a recent series of analyses on lamb (hogget) carcases, the opportunity was taken to examine the chemical nature of this waste material. Analyses, carried out by two of the independent participating laboratories, showed that its nitrogen content was approximately 6% (fat-free basis), a figure which reflects its high content of cartilagenous material. From the data available it is possible to calculate the effect of the inclusion of such waste material on the nitrogen factors for the lean (plus intermuscular fat) of the various cuts for all species, In the case of clean forequarter beef, where the influence of its inclusion would be the greatest, the effect would be to raise the nitrogen factor by ca.0.05. The effects with sheepmeat would be smaller, and in pork even less, since the amount of waste material in the latter is only about half of that Table 8 Chemical composition of lamb hindquarter* meat (% of tissue) Differences between Lamb type type means Tissue Lean tissue- Total fat Moisture Nitrogen Fat-free nitrogen Ash H ydrox yproline Total fatty tissue- Total fat Moisture Nitrogen Fat-free nitrogen Ash H y drox yproline Main Milk-fed season 4.9 5.9 74.0 73.0 3.28 3.26 3.45 3.46 1.09 1.07 0.19 0.18 73.6 78.3 18.7 15.3 1.21 1.06 4.52 4.86 0.32 0.26 0.44 0.40 Hoggets 6.1 72.8 3.23 3.44 1.03 0.25 76.4 16.9 1.12 4.79 0.25 0.41 Lean and intermuscular fat portion of total fatty tissue- Total fat 11.1 13.8 13.2 Moisture 69.1 66.8 67.2 Nitrogen 3.09 3.01 3.01 Ash 1.02 0.98 0.96 Hydroxyproline 0.21 0.20 0.27 Total fat 17.3 21.0 20.0 Moisture 64.1 61.0 61.7 Nitrogen 2.90 2.79 2.81 Fat-free nitrogen 3.5 1 3.54 3.51 Ash 0.95 0.90 0.88 Hydroxyproline 0.23 0.22 0.29 Fat-free nitrogen 3.48 3.50 3.47 Lean and total fatty tissues- )r Leg, loin and chump joints.t Significance, as Table 2. SE 0.23 0.24 0.026 0.027 0.0 13 0.040 1.62 1.18 0.086 0.141 0.022 0.034 0.82 0.70 0.033 0.028 0.015 0.038 1.41 1.16 0.05 1 0.029 0.020 0.035 Significance? ** ** ns ns ns * ns ns ns ns ns ns ns ns ns ns ns * ns ns ns ns * ns Overall mean 5.7 73.2 3.26 3.45 1.07 0.20 77.0 16.3 1.10 4.78 0.27 0.41 13.1 67.3 3.03 3.49 0.98 0.22 20.1 61.8 2.82 3.53 0.91 0.24 SE 0.13 0.14 0.015 0.016 0.007 0.023 0.94 0.68 0.050 0.081 0.013 0.020 0.47 0.40 0.019 0.016 0.009 0.022 0.8 1 0.67 0.029 0.017 0.011 0.020 IAnalyst, July 1996, Vol.121 895 Table 9 Physical composition of lamb, tissues as a percent of joints and side Joint, tissue Leg, lean- Intermuscular fat Subcutaneous fat Bone and waste Intermuscular fat Subcutaneous fat Bone and waste Intermuscular fat Subcutaneous fat Bone and waste Best-end neck, lean- Intermuscular fat Subcutaneous fat Bone and waste Intermuscular fat Subcutaneous fat Bone and waste Shoulder-middle neck, lean- Intermuscular fat Subcutaneous fat Bone and waste Intermuscular fat Subcutaneous fat Bone and waste Forequarter, lean- Intermuscular fat Subcutaneous fat Bone and waste Hindquarter, lean- Intermuscular fat Subcutaneous fat Bone and waste Intermuscular fat Subcutaneous fat Bone and waste Chump, lean- Loin, lean- Breast, lean- Scrag, lean- Carcase, lean- * Significance, as Table 2.Lamb type Differences between type means Milk-fed 71.0 3.9 7.3 17.8 65.9 6.2 13.6 14.2 66.8 13.2 6.5 13.4 56.4 10.4 18.6 14.6 53.6 14.8 15.9 15.8 64.3 6.1 11.0 18.6 58.9 8.0 7.1 25.9 60.8 11.5 9.3 18.4 69.0 5 .O 9.9 16.1 64.4 8.6 9.6 17.4 Main season 69.5 4.7 8.3 17.5 61.1 8.1 15.0 15.8 64.1 15.8 7.5 12.6 52.9 12.4 17.8 16.9 50.3 17.8 17.2 14.7 61.3 6.9 13.2 18.6 56.6 9.8 8.8 24.8 57.5 13.9 10.7 17.9 66.5 6.0 11.4 16.0 61.5 10.4 11.0 17.0 Hoggets 69.8 4.7 8.6 17.0 63.9 6.9 13.9 15.3 65.4 14.8 7.3 12.5 55.7 11.9 16.1 16.4 52.8 16.4 16.0 14.9 64.2 6.5 11.8 17.5 61.9 7.4 7.7 23.0 60.4 12.6 10.0 17.0 67.7 5.6 10.9 15.7 63.7 9.5 10.4 16.4 SE 0.45 0.19 0.46 0.28 0.81 0.29 0.94 0.43 1.13 1.12 0.39 0.54 1.06 0.54 0.65 1.16 1.12 0.66 1.04 0.52 0.53 0.42 0.40 0.34 0.97 0.59 0.58 0.95 0.71 0.41 0.53 0.40 0.60 0.18 0.68 0.32 0.65 0.30 0.64 0.34 Significance* * ** ns ns *** *** ns L ns ns ns ns * * ns ns ns ns * ** *** *** ns * *r* * ns ns ** *** ns * *** ns ns ** *** ns ns Overall mean 69.8 4.6 8.2 17.4 62.6 7.5 14.5 15.4 64.9 15.1 7.3 12.7 54.2 11.9 17.5 16.4 51.5 16.9 16.7 14.9 62.5 6.7 12.5 18.3 58.3 8.9 8.2 24.5 58.8 13.2 10.3 17.8 67.2 5.7 11.0 15.9 62.5 9.9 10.6 16.9 SE 0.26 0.11 0.26 0.26 0.47 0.17 0.54 0.25 0.65 0.65 0.22 0.3 1 0.61 0.31 0.37 0.67 0.65 0.38 0.60 0.30 0.3 1 0.24 0.23 0.20 0.56 0.34 0.33 0.55 0.41 0.24 0.30 0.23 0.35 0.11 0.39 0.18 0.38 0.17 0.37 0.20 Table 10 Chemical composition of mutton: lean and intermuscular fat* (% of tissue) Forequarter Mean SE Total fat 23.1 1.03 Moisture 59.6 0.80 Nitrogen 2.67 0.033 Ash 0.8 1 0.013 H y drox yproline 0.25 0.007 Fat-free nitrogen 3.48 0.220 Hindquarter Side Mean SE Mean SE 11.6 0.57 18.0 0.83 68.0 0.44 63.4 0.67 3.09 0.022 2.86 0.027 3.50 0.016 3.49 0.017 0.96 0.010 0.88 0.008 0.22 0.010 0.24 0.007 * Combining the chemical composition of the lean tissue and fatty tissue in the proportions of the lean and intermuscular fat only, present in each quarterlside.896 Analyst, July 1996, Vol.121 Table 11 Chemical composition of mutton: lean with fatty tissues' (% of tissue) Forequarter Hindquarter Side Mean SE Mean SE Mean SE Total fat 29.1 1.42 19.6 1.22 24.8 1.34 Moisture 55.0 1.11 61.8 0.99 58.1 1.07 Nitrogen 2.41 0.044 2.83 0.040 2.64 0.041 Ash 0.75 0.016 0.87 0.013 0.81 0.013 Hydrox yproline 0.26 0.008 0.24 0.01 1 0.25 0.008 Fat-free nitrogen 3.50 0.029 3.52 0.020 3.52 0.021 * Combining the chemical composition of the lean tissue and fatty tissue, present in each quarter/side.Table 12 Fat-free nitrogen contents of lean and associated fats (% of tissue) Mutton- Forequarter Hindquarter Side Lamb- Forequarter Hindquarter Side Leg and chuoxp- Loin and best end neck Scrag, shoulder-middle neck and breast Lean and intermuscular fat 3.48 3.50 3.49 3.50 3.49 3.49 3.45 3.61 3.48 Lean and inter- muscular and subcutaneous fats 3.50 3.52 3.52 3.53 3.53 3.53 3.50 3.66 3.51 in beef. This material would not normally be included in butchered meat cuts and, where it did occur, the amount would be variable and difficult to quantify. While noting that the inclusion of such waste with the edible meat portion of the carcase could have a small but insignificant effect on the nitrogen factors for the three species, the Sub- Committee believes that, because of its minor and variable contribution, it would not be realistic to include its effect when calculating the nitrogen factor. The Sub-committee further concluded that the above considerations do not in any way invalidate or depreciate the status of the nitrogen factors as published. References Analytical Methods Committee, Analyst, 1995, 120, 1823. Analytical Methods Committee, Analyst, 199 1, 116, 76 1. Analytical Methods Committee, Analyst, 1993, 118, 1217. British Standards Institution, BS4401: Analytical Methods for Meat and Meat Products: Part I . 1980, Determination of Ash; Part 2. 1980, Determination of Nitrogen; Part 3. 1980, Determination of Moisture; Part 4. 1970, Determination of Total Fat; and Part 11. 1979, Determination c$ r,(-)-H~idroxyprolipe, BSI, Milton Keynes, Buckinghamshire. Analytical Methods Committee, Analyst, 1986, 111, 969. Meat and Livestock Commission, PO Box 44, Winterhill House, Snowdon Drive, Milton Keynes, MK6 IAX. Paper 6/02 7236 Accepted April 18, I996

ISSN:0003-2654    

DOI:10.1039/AN9962100889

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (897-900) 897 Determination of Mercury in Fluorescent Lamp Cullet by Slurry Sampling Electrothermal At om ic Absorption Spectrometry* R. Dobrowolski and J. Mierzwa? Central Laboratory, M . Curie-Sktodowska University, PL-20-031 Lublin, Poland A method for the determination of mercury in fluorescent lamp cullet by slurry sampling ETAAS is proposed. Palladium nitrate was applied as an effective matrix modifier. The furnace programme was optimized. The stability of slurries, mercury partitioning between solid and liquid phases for different media and the influence of palladium modifier were investigated. Using the optimized conditions, a detection limit (based on a 30 criterion) for mercury of 24 ng g-1 was achieved. For 1.55 and 2.16 pg 8-1 of mercury in the phosphor sample the RSDs (n = 9) were 8.3 and 8.7%, respectively. Calibration was performed using the standard additions method.The results obtained using this method suggest that the determination of mercury in fluorescent lamp cullet can be carried out effectively and rapidly. Keywords: Electrothermal atomization atomic absorption spectrometry; slurry sampling; mercury; phosphors; glass Introduction Mercury is a cumulative and very toxic element for humans and animals.l.2 One of the sources of environmental pollution by mercury is the manufacture and disposal of fluorescent lamps (e.g., ref. 3). Mercury in fluorescent lamp tubes is adsorbed mainly on the phosphor (halophosphate Ca/Sb, Mn) layer. Long-term storage of fluorescent lamp cullet causes uncon- trolled desorption of mercury.As a result of this phenomenon, mercury pollution of air, soil and plants is observed. Therefore, the utilization of waste or damaged lamps is necessary. Thermal pre-treatment is a widely used method. Moreover, this method creates the possibility of phosphor recycling. The main parameters of the recycling process are estimated on the basis of mercury determination in phosphors. In a previous paper,4 two AAS techniques for mercury determination in fluorescent lamp cullet were described. The first was based on the determination of mercury by CVAAS after phosphor acid wet digestion. The second was based on mercury determination by direct atomization of a solid sample of phosphor in a special 'ring chamber graphite tube'.5 However, the second method cannot be widely used because this tube was designed for an AAS-3 spectrometer (Carl Zeiss, Jena, Germany) only.Analysis involving slurry formation is simpler than direct solid sample analysis. The slurry technique allows the sample decomposition step to be omitted. Moreover, conventional atomizers and autosamplers can be applied. From the analytical point of view, slurry sampling ETAAS combines positive features of solution injection and direct solid sampling.6.7 - Presented in part at the 3rd Polish Analytical Seminar, Poznan, Poland, April 27-28, 1994. To whom correspondence should be addressed. The slurry sampling technique was successfully applied by Slovak* for the ETAAS determination of 1-50 ng of mercury in aqueous solutions after preconcentration on a chelating ion- exchange resin.Mercury was bound on a resin with thiol groups. This method offers the possibility of preparing standards and samples in the same matrix and removing calibration problems. However, the method is limited to the simple analysis of liquid samples because complete desorption of mercury from the solid to the liquid phase (depending on, e.g., pH, mercury form and solution composition) must be achieved prior to the preconcentration procedure. More recently, Bermejo-Barrera et al.9 proposed palladium nitrate as a chemical modifier for the determination of mercury in marine sediment slurries. They studied palladium, magne- sium and palladium-magnesium modifiers and found that use of palladium nitrate (at a concentration of 15 mg 1 - 1 ) pre-mixed with slurries was satisfactory for mercury stabilization at 200 "C and successful determination using calibration by the standard additions method.In the present study, a procedure for mercury determination in phosphors from fluorescent lamp cullet by slurry sampling ETAAS has been proposed. Experimental Apparatus The measurements were carried out using an AAS-3 spec- trometer interfaced to an IBM PC-386/DX compatible micro- computer, EA-3 electrothermal atomizer and MPE autosampler. A program written in Turbo Pascal was used to calculate, display and store the transient signal data. The primary source of radiation was a mercury hollow-cathode lamp (Juniper, Harlow, Essex, UK). A deuterium lamp background correction system was used.All investigations were carried out with pyrolytic graphite-coated graphite tubes equipped with pyrolytic graph- ite-coated L'vov platforms. Temperatures of the graphite furnace were calibrated with an EP-7 optical pyrometer (PZO, Poland) and/or a thermocouple (up to 1000 "C). Samples were weighed with a Model R-200 D analytical balance (Sartorius, Germany). Slurries were mixed with a UD- 20 ultrasonic disintegrator (Venpan, Poland) (maximum output power 180 W and ultrasonic amplitude 1-8 pm) with a titanium probe or else vortex mixed. Samples and Reagents Samples of phosphor scraped from fluorescent lamp cullet were analysed. The phosphor samples investigated had fine grains of diameter < 20 pm so they were ground for only a short time (8 min) using an MM-2 Spectro Mill (Retsch, Germany).Jt was verified for phosphor slurries by means of an optical micro- scope that after grinding the average grain diameter was < 15 pm and the homogeneity of the samples was improved.898 Analyst, July 1996, Vol. 121 For the preparation of the slurries, doubly distilled water, nitric acid of spectral purity (Merck, Darmstadt, Germany) and sodium sulfide (pure for analysis) were used. Standard mercury solutions were prepared from a mercury stock standard solution (Merck) and 1 moll-' nitric acid of spectral purity. More dilute solutions of mercury were prepared by diluting the stock standard solutions with doubly distilled water immediately before measurements. Palladium nitrate (AAS chemical modi- fier solution from Merck) was used as the modifier.Pure argon (99.999%) was used as an inert gas. Analytical Procedure A thin layer of phosphor was scraped manually from pieces of fluorescent lamp cullet and the powder was transferred directly into a closed vessel. The sample was thoroughly mixed by rotating the vessel and in this way about 100-150 rng of phosphor contaminated by mercury were collected. Slurries were prepared by weighing amounts of phosphor from 10 to 40 mg in the polyethylene vessels of autosampler tested earlier for their purity. Then 0.05% nitric acid was added as a liquid medium. The final volume of slurry was about 1 ml. A v/v factor (volume of solid/volume of diluent) as proposed by Miller-IhlilO was taken into account. The slurries were pre- mixed with palladium nitrate (palladium mass 50 pg per gram of slurry) and homogenized before each measurement with an ultrasonic disintegrator (10 s) or vortex mixed (30 s).Mixing was performed in the vessels of the autosampler directly prior to the sampling of the slurry, which permitted maximum elimina- tion of the error resulting from sedimentation of solids. A slurry volume of 20 pl was injected into the electrothermal atomizer. All measurements were based on background-corrected peak areas (integrated absorbances). The gas stop mode was used during the atomization step to prevent loss of the analyte. Calibration by the standard additions method was finally applied. The basic experimental conditions and parameters for the time-temperature programme with application of the palladium modifier are shown in Table 1.Results and Discussion Stability of Slurry The stability of slurries (time dependence of the absorbance signal) was studied. The slurry sample was ultrasonically agitated and then injected into the graphite furnace after the specified period of time, immediately after agitation, after 5 s, after 10 s, etc. Each measurement was repeated three times and the average values of the analytical signal (percentage of highest signal value) are presented in Fig. 1 for slurries prepared in 0.1 mol 1-1 Na2S and in 2 mol 1-l HN03. For a slurry in 0.05% HN03 the results of stability tests were similar. The Table 1 Instrumental parameters for Hg determination by slurry sampling ETAAS Wavelengthhm Spectral bandpass/nm Hg lamp current/mA Purge gas Drying temperatures/"C (drying times in parentheses) Atomization temperaturePC Cleaning temperaturePC Integration time/s Sample volume/yl Range of slurry concentration (% m/v) 253.6 0.7 3.5 Argon 105 (30 s) 200 (40 s) 1900" 2450 10 20t 1-4 * Ramp time 2 s.t Aliquot of the slurry suspension or supernatant. slurry of the phosphor shows a great tendency for rapid sedimentation, irrespective of the kind of liquid medium used. The use of stabilizing agent, e g . , poly(viny1 alcohol) or polyethylene glycol (agents used in the manufacture of fluorescent lamps), was not possible because of a too high background signal, making their correction impossible. The heavy organic matrix of poly(viny1 alcohol) or poly(ethy1ene glycol) was probable not completely pyrolysed. Mechanical homogenization of slurry (vortex mixing) is not very efficient because of fast sedimentation of the phosphor.Ultrasonic homogenization, directly before slurry sampling, increases the repeatability of the analytical signal through more effective mixing. The time of ultrasonic agitation was optimized experimentally. It is probable that too long a time of ultrasonic homogenization causes repeated agglomeration processes and decreases the repeatability of the analytical signal (see Table 2). It was found experimentally that for the ultrasonic disintegrator used here and ultrasonic agitation directly in an autosampler vessel (slurry volume approximately 1 ml), the optimum time of ultrasonication is about 10 s at a low power level. The use of a higher power of the probe was not favourable because of very fast heating of the slurry.It must be taken into account that the power output to the probe is not constant from unit to unit and tuning is usually subjective.l Mercury Partitioning and Liquid Media Application of different liquid media (see Table 3) results in different partitioning of mercury between the liquid and solid phases. For example, for sodium sulfide as liquid medium, extraction of mercury from the solid to the liquid phase was observed, probably because of the formation of sulfide complexes. On the other hand, sodium sulfide causes a significant shift and decrease of the mercury analytical signal (Fig. 2). Bermejo-Barrera et aZ.9 observed total extraction of mercury from a marine sediment sample with 17.5% nitric acid. For the less concentrated 2 mol 1-1 (approximately 12.5%) I O o i io 15 io 25 30 Delay of sampling time/s Fig.1 Stability tests for phosphor slurries (approximately 2% m/v, without Pd addition): A, slurry in 2 mol 1-1 nitric acid; B, slurry in 0.1 mol 1-1 sodium sulfide. Table 2 Effect of ultrasonic mixing time on repeatability of the phosphor slurry signal (slurries approximately 3.5% m/v) Average Signal Time of analytical repeatability mixing/s signal (A s) (RSD, %)* 5 0.115 12.5 10 0.182 11.1 15 0.181 13.2 20 0.183 14.5 25 0.179 14.9 * RSDs were evaluated from nine individual measurements of the separate slurries.Analyst, July 1996, Vol. 121 899 nitric acid, extraction of mercury from the phosphor to the liquid phase was also observed (Table 3) but at a lower level.The use of 2 mol 1-1 nitric acid as a liquid medium led to deterioration of the mercury peak shape (Fig. 2), probably because of the extraction of some components of the phosphor matrix to the liquid phase, which causes more interferences. The high concentration of nitric acid also influences the background level and graphite tube lifetime, and reduces the efficiency of the palladium modifier.12 Therefore, 0.05% nitric acid was finally applied as the liquid medium. Palladium Modifier A palladium modifier has been used previously for the ETAAS determination of mercury in solution, permitting the use of a higher temperature of thermal pre-treatment (up to 400 "C) and enhancing the analytical signaL13-'6 A series of measurements were carried out to determine the optimum amount of palladium modifier.No significant varia- tion of the peak area with variation in the amount of palladium in the range 1.0-2.5 pg (in one slurry sampling) was observed. An amount of 1.0 yg of Pd per 20 yl of slurry was selected for further experiments. The addition of the palladium nitrate modifier (both for an aqueous standard and a slurry in 0.05% nitric acid) led to analytical signals that were more symmetric and regularly shaped and descended to the baseline (Fig. 2). The application of palladium does not shift the thermal stability of mercury (Fig. 3), in contrast to 0.1 mol 1-1 Na2S4 (Fig. 3) or thiol groups.8 A halophosphate phosphor chemical modifies the mechanism of mercury atomization from the phosphor and a higher atomization temperature (1 900 "C in this study with Pd) is necessary.It is very likely that a higher temperature is required with the slurry because part of the mercury has to diffuse from the solid lattice of phosphor. This temperature is very different to the optimum atomization temperature (in the range 1200-1300°C) for an aqueous mercury standard (Fig. 4) with palladium. Palladium modifier delayed the atomization of mercury from the phosphor until higher temperatures and more isothermal conditions are attained. The presence of a palladium modifier also caused a better separation of the background signal from the analytical signal. In fact, the mechanism of modification by palladium in the analysis of slurries is very complex17 and the chemical modification mechanism needs a separate study. Considering the results of the thermal pre-treatment graphs (Fig.3), the maximum temperature of thermal pre-treatment of the phosphor sample was 200 "C. By this means, the possibility of eventual loss of the analyte was minimized. Stabilization of mercury at the same maximum temperature as reported by Bermejo-Barrera et al.9 for marine sediment slurries mixed with the palladium nitrate modifier was observed. Calibration Calibration graph and standard addition methods were com- pared. The lines were not parallel, and the slope of the standard addition graph (0.282) was smaller than that of the aqueous standard calibration graph (0.352). The correlation coefficients were 0.9989 for the calibration graph method and 0.9994 for standard addition method.This means that some suppression of the mercury signal, due to the halophosphate matrix effect, exists. Hence the method of standard addition based on spiking the slurries with aqueous standard solutions was applied. The results obtained for three real samples of cullet (taken from different places) are presented in Table 4. Table 3 Analyte partitioning in slurries 0.61 -4 A I Slurry Concentration Concentration in liquid in solid fraction (%)* fraction (%) In water 0 100 In 0.05% HN03t 0 100 In 0.1 rnol 1-l NazS 41 59 In 2 mol I-' HN03 16 84 * The liquid fraction concentration was examined after 12 min of centrifugation. + With palladium modifier. 0.2 w 0 d cd 0.1 v1 s 4 0.0 0 4 8 12 Time/s Fig. 2 Analytical signals of mercury obtained from the atomization of A, phosphor slurry (approximately 2.2% m/v) in 2 rnol I-' nitric acid; B, phosphor slurry (approximately 2.2% m/v) in 0.1 mol 1-I sodium sulfide; C, aqueous mercury standard (200 ng ml-I), with palladium; and D, phosphor slurry (approximately 2.2% m/v) in 0.05% nitric acid, with palladium.o J I S O 200 250 300 350 4 k 450 500 6h ~ 700 Temperature/"C Fig. 3 Thermal pre-treatment graphs for mercury: A, slurry (approx- imately 4% m/v) in 2 mol 1-1 nitric acid; B, slurry (approximately 2.8% m/v) in 0.1 moll-' sodium sulfide; C, slurry (approximately 2.2% m/v) in 0.05% nitric acid with Pd modifier; and D, slurry (approximately 2.5% m/v) in 0.05% nitric acid without modifier. 1200 1400 1600 1800 2000 2200 0.11, ' ' ' ~ , , ' I ' Temperaturel-C Fig.4 Atomization graphs obtained with palladium (1 pg g-l) for: A, an aqueous mercury standard (200 ng ml-1); and B, phosphor slurry approximately 1.6% m/v).900 Analyst, July 1996, Vol. 121 Figures of Merit The sensitivity of the proposed method expressed by means of the characteristic mass was 140 pg of Hg. The limit of detection for mercury (calculated using the 30 criterion) was 24.3 ng g-1. This limit of detection was calculated for 12 replicate injections (RSD = 2.1%) of a 3% m/v slurry of the synthetic matrix (calcium halophosphate). The repeatability of the mercury determination by proposed technique was also calculated (see Table 4). The RSD is approximately is 8-9% (for nine individual measurements of slurry) and is comparable to that for the ETAAS determination of mercury in the same material directly from a solid sample (see also Ref.4). In this case the RSD was in the range 9.6-9.9%. No statistically significant difference for the mercury determination results for separate slurries of the same sample repeated after two or three days were observed . Conclusions The use of the standard additions procedure and the addition of palladium nitrate as a chemical modifier were necessary for the effective determination of mercury in the halophosphate matrix. Table 4 Comparison of results of mercury determination in fluorescent lamp cullet by slurry sampling ETAAS and solid sampling ETAAS Concentration of mercury/mg kg-1 Cullet taken ETAAS with ETAAS with from dump slurry sampling solid sampling* CVAASt Sample A 1.55 1.48 1.59 Sample B 1.90 1.82 1.85 Sample C 2.16 2.10 2.19 (RSD 8.3%)* (RSD 9.6%) (RSD 3.4%) (RSD 8.6%) (RSD 9.8%) (RSD 3.3%) (RSD 8.7%) (RSD 9.9%) (RSD 3.7%) * Determination of Hg by solid sampling ETAAS using the ring chamber graphite furnace (according to ref.4). t Determination of Hg by CVAAS after wet dissolution of the sample with concentrated nitric acid in a closed system (according to ref. 4). * RSDs were evaluated from nine individual measurements. The homogeneity of the slurry would be greatly improved if ultrasonic mixing was used and optimized properly. The repeatability of the mercury determination method is considered to be acceptable. The results for mercury determination in real samples correspond well with those obtained by CVAAS after wet dissolution of the sample4 and by another ETAAS technique, solid sampling of the phosphor directly into the ring chamber graphite tube.4 The proposed method is simple and fast and can be used successfully to control the thermal utilization of fluorescent lamp cullet (using an AAS spectrometer equipped with a standard graphite furnace).It may also be used for the rapid detection of mercury pollution of cullet dumps. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Taylor, D., Mercury as an Environmental Pollutant, Imperial Chemical Industries, Birmingham, 4th edn., 1975. Chapman, P., Romberg, G., and Vigers, G., J . Water Pollut. Control Fed., 1982, 54, 292. Surdel, M., Rocz. Glebozn., 1991, 42, 237. Dobrowolski, R., and Mierzwa, J., Analyst, 1992, 117, 1165. Schmidt, K. P., and Falk, H., Spectrochim. Acta, Part B, 1987, 42, 431. Stephen, S. C., Littlejohn, D., and Ottaway, J. M., Analyst, 1985,110, 1147. Miller-Ihli, N. J., Spectrochim. Acta, Part B , 1989, 44, 1221. Slovak, Z., Anal. Chim. Acta, 1979, 110, 301. Bermejo-Barrera, P., Moreda-Pineiro, J., Moreda-Pineiro, A., and Bermejo-Barrera, A., Anal. Chim. Acta, 1994, 296, 181. Miller-Ihli, N. J., At. Spectrosc., 1992, 13, 1. Miller-Ihli, N. J., Fresenius’ J . Anal. Chem., 1993, 345, 482. Voth-Beach, L. M., and Schrader, D. E., J . Anal. At. Spectrom., 1987, 2,45. Ping, L., Fuwa, K., and Matsumoto, K., Anal. Chim. Acta., 1985,171, 279. Matsumoto, K., ACS Symp. Ser., No. 445, 1991, 278. Bortoli, A., Dell’Andrea, E., Gerotta, M., Marin, V., and Moretti, G., Acta Chim. Hung., 1991, 128, 573. Welz, B., Schlemmer, G., and Mudakavi, J. R., J . Anal. At. Spectrom., 1992, 7, 499. Qiao, H. C., and Jackson, K. W., Spectrochim. Acta, Part B, 1992,47, 1267. Paper 61001 94G Received January 3,1996 Accepted March 29, 1996

ISSN:0003-2654    

DOI:10.1039/AN9962100897

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (901-904) 90 1 Determination of Isotope Enrichments of Magnesium in Microwave-d igested Biological Samples by Thermal Ionization Mass Spectrometry Using a Direct Loading Technique Werner StegmannaTb, Steven L. Goldstein" and Michael Georgieffb a Max-Planck-Institut fiir Chemie (Otto-Hahn-Institut), Postfach 3060, 0-55020 Mainz, Germany. E-mail: stegepalesmpch-mainz.mpg.de b Universitatsklinik fiir Anasthesiologie, Klinikum der Universitat Ulm, Sektion Experimentelle Anasthesie, Parkstrasse 11,0-89073 Ulm, Germany. The isotope ratios of magnesium were determined in isotopically normal and 26Mg-enriched samples of human blood, blood plasma, urine and faeces and bovine muscle. The measurements were made with a magnetic sector, thermal ionization mass spectrometer (TIMS) equipped with a multiple ion collector system for simultaneous detection of the ion currents.The samples were decomposed using microwave digestion with HN03 and HCl. Without further chemical treatment, the mineralized samples were deposited together with silica gel and boric acid on rhenium filaments, which served as thermal ionization source filaments. This method, called the direct loading technique (DLT), results in stable ion signals of the magnesium isotopes with isotope ratios indistinguishable from those of natural Mg standards within experimental error. Fractionation-corrected 26MgPMg ratios of natural Mg standards were determined with a relative external precision of 0.02 %. The magnesium recoveries for all of the analysed matrices were 297%; 26Mg was added to calibrated sample solutions to produce isotopic enrichments within a range typically appearing in samples of human tracer studies. Linear regression analysis of measured versus expected per 1000 (%o) enrichments yields y = 0.998~ + 0.79.The DLT described here is a simpler and quicker method than other methods reported hitherto. It has the advantage of avoiding magnesium separation and purification steps prior to TIMS analysis for all of the analysed biological samples and thus reduces contamination and guarantees optimum magnesium recovery. The reported method improves the applicability of stable isotopes of magnesium in human tracer studies. Keywords: Thermal ionization mass spectrometry; magnesium; direct loading technique; stable isotopes; biological material Introduction There is increasing interest in the use of stable isotopes for studying mineral and trace element metabolism in humans.l.2 Because the risks associated with radioactive tracers are absent, stable isotopes are especially valuable in studies of neonates, infants and pregnant and lactating women and in clinical s tudies.3 The only radioisotope of magnesium that can be used for tracer studies is the short-lived 28Mg (tl,2 = 21.3 h). The stable magnesium isotopes 25Mg and 26Mg are not only radiation-free alternatives to 28Mg but also represent the only tools for long- term tracer studies of the element.4-7 However, the high natural abundances of 25Mg (10.00%) and 26Mg (11.01%) reduce the sensitivity with which magnesium stable isotope tracers can be measured in body fluids and tissue samples.Of the various methods that have been used to determine magnesium stable isotope tracers in biological samples,"-' measurements made by magnetic sector mass spectrometry using thermal ionization ion sources (TIMS) show the highest precision and accuracy and alone are capable of detecting magnesium stable isotope tracers in humans for weeks after tracer admini~tration.~ The preparation of organic samples for TIMS classically requires chemical extraction of the element under investigation. Magne- sium chemical extraction is mostly performed using wet ashing decomposition methods and ion-exchange chromatogra- phy.7>12.13 Alternative preparation methods trying to avoid the decomposition step have recently been reported. These methods result in low and varying total magnesium recoveries and have limited applicability.14 We describe here a sample preparation method for TIMS for human total blood, blood plasma, faecal, urine and bovine muscle samples, using solely microwave digestion for sample mineralization in order to avoid contamination and recovery problems. Magnesium isotope ratio measurements were made by TIMS using a multi-collector ion detection system and a direct loading technique (DLT) for mineralized biological samples. A similar DLT has been in use for many years in cosmochemical research and has been used to determine magnesium isotope abundance anomalies in single crystals of extraterrestrial materials.15.16 We also report on the advantage of using silica gel and boric acid with direct loaded mineralized biological samples instead of the conventionally used silica gel and phosphoric acid.Whereas the ion currents produced with silica gel and phosphoric acid were low or unstable and poorly reproducible, silica gel and boric acid provided an efficient, stable and reproducible ion emitter for all of the direct loaded mineralized biological samples. The results obtained with the proposed method demonstrate that magnesium isotope ratios can be determined in normal and 26Mg-enriched biological samples with high precision and accuracy in an easy and rapid way. Experimental Chemistry Highly enriched 26Mg as MgO (obtained from Isotec, Miamisburg, OH, USA) was dissolved in 4 mol 1-1 HCL. It was analysed for isotopic composition (99.55% 26Mg,902 Analyst, July 1996, Vol.121 0.35% 24Mg and 0.10% 25Mg), and was also analysed for magnesium concentration via isotope dilution mass spectrome- try (IDMS) using a natural magnesium standard. (In IDMS, a known amount of enriched isotope, e.g., 26Mg, is added to the sample. After equilibration of the added enriched isotope with the natural element in the sample, the altered isotopic composition in the mixture is determined mass spectro- metrically and is used to calculate the concentration of the element in the sample.l7 In the reverse way, IDMS is used with standard solutions of the element with natural isotope abun- dances for the determination of the concentrations of tracer solutions with the enriched isotopes.) Magnesium standards were prepared from Titrisol solutions (Merck, Darmstadt, Germany) using 4 mol 1-1 HCl.The HC1 used was purified twice by sub-boiling distillation from analytical-reagent grade acid. Boric acid (Suprapur) was obtained from Merck. High- purity water with a metered resistivity of 18.2 MQ (Millipore, Be.dford, MA, USA) was used to prepare the acid dilutions. A previously described 2sMg tracer7 (96.82% ZSMg, 2.15% 24Mg, 1.03% 26Mg) was used in IDMS experiments for the blank determinations, the recovery experiments and the measurement of the magnesium concentrations in aliquots of mineralized sample solutions. The apparatus used for sample mineralization was a high- temperaturehigh-pressure microwave digestion system (PMD, Kiirner, Rosenheim, Germany) with two 35 ml quartz vessels in which the pressure control mode is used and 100% power corresponds to 750 W.The analytical procedures were similar for all of the biological samples under investi- gation: 0.1-0.2 g of the wet sample was weighed into the quartz vessel and 2 ml of 65% m/m HN03 and 0.5 ml of 32% m/m HC1 (Merck, analytical-reagent grade) were added. Microwave digestion was performed at 60% power for 10 min. After cooling, the digested samples were transferred into 15 ml TEFLON-PFA vials (Savillex, Minnetonka, MN, USA) and evaporated to dryness. Evaporation was carried out in a clean box with a horizontal laminar air flow. For further treatment, the sample residues were dissolved in 5 ml of 4 moll-' HCl. To check the feasibility and reliability of the DLT, the magnesium isotopic composition was measured in a set of samples artificially enriched in 26Mg.The 26Mg enrichments were achieved in the following way: ( i ) aliquots of the solutions of microwave-digested samples (human blood, urine and faeces and bovine muscle) were analysed for magnesium concentra- tion via IDMS; (ii) aliquots of the sample solutions or of the magnesium standard solution were combined with 26Mg tracer solutions containing known amounts of 26Mg to obtain samples with 26Mg enrichments of approximately 1-1000 per 1000 (%o) in excess of natural abundance. For samples with 26Mg enrichments of < 5%0, diluted tracer was used in order to reduce the weighing error. Decomposition experiments on biological samples using microwave digestion usually show high recoveries of magne- sium.18 To determine the Mg recovery, in an IDMS experiment 25Mg tracer solution was added (i) to two sets of all of the analysed biological samples prior to the microwave digestion step and (ii) to two sets after the microwave digestion step.Taking the results of ( i ) as the true Mg concentrations of the samples, the Mg recoveries calculated from the results of (ii) were 97-102% with an error (20) of <4%. Total system blanks were determined by IDMS and were found to be 13-19 ng of magnesium, which was d0.5% of the magnesium content of each of the analysed biological samples. Memory effects from the microwave digestion apparatus were avoided by performing a cleaning run after each digestion run. The silica gel-boric acid filament loading blank was 0.005-0.01 ng of magnesium.~~ ~ ~ ~~~ Mass Spectrometry For the DLT, 1 pl of a suspension of silica gel was pipetted stepwise with a microsyringe on to an outgassed rhenium ribbon (thickness 0.03 mm, width 0.7 mm, length 10 mm, zone-refined Re, from Sandvik, Elyria, OH, USA) of a single-filament ion source and was dried by increasing the filament current to approximately 0.8 A. The silica gel suspension (60 g 1-1) was made from gaseous Sic14 and triply distilled water. 16 An aliquot of the sample solution containing 50-100 ng of magnesium was added stepwise to the silica gel followed by drying steps (at about 0.8 A), then 2-4 pl of 0.5 moll-1 boric acid were added stepwise as described for the sample solution. Finally, the filament current was increased gradually until the silica gel was cemented on the rhenium ribbon (at approximately 1.6 A).The entire sample loading procedure was performed in a laminar flow bench. Magnesium isotope measurements were performed with a magnetic-sector TIMS system (Model 26 1, Finnigan MAT, Bremen, Germany). It has a rotating magazine that holds 13 ion- source filament assemblies, a reference pyrometer which measures filament temperature and a multi-collector ion detection system consisting of seven Faraday cups each connected with an electrometer amplifier with a 1011 Q feedback resistor. This instrument routinely measures isotope ratios of Sr and Nd to 20-30 ppm 20 external reproducibility, corrected for mass fractionation (e.g., ref. 19). After the sample magazine had been inserted in the mass spectrometer, the rhenium filaments were heated by increasing the filament current over 15 min until the filament temperature reached 1500-1600 "C.In this temperature interval the magnesium ion beam intensity entered a stable phase with slowly increasing beam current (3 X 10-l2 A d 24Mg+ d 5 X 10-l1 A). These conditions were reproducible for all of the biological samples analysed using the DLT with silica gel and boric acid. We also tested the DLT using silica gel and phosphoric acid as an ion emitter. With samples of microwave-digested total blood the magnesium ion current never exceeded 3 X 10-12 A and the beam intensity never reached a stable plateau. For microwave-digested urine and faecal samples the magnesium ion currents reached maxima of approximately 5 X 10-11 A at filament temperatures of about 1650 "C. However, the ion beam intensities showed an erratic behaviour and never reached a stable phase.The ion currents of the three stable isotopes 24Mg, 25Mg and 26Mg were measured simultaneously in the static mode as soon as the 24Mg signal exceeded 3 X 10-l2 A. During magnesium analysis no interfering species at the Mg masses were observed. The only ion signals appearing in the magnesium mass region were 23Na+ and 27Al+, which gave no detectable contribution to the magnesium masses. The Mg ion intensities were measured using an integration time of 16 s. Baseline measurements were made at the beginning of each run and after sets of 10-20 intensity measurements. Usually 200-400 simultaneous triple ion current integrations were taken per run.Gain calibration measurements of the amplifier channels were made during the runs. The drifts in the relative gains of the amplifiers were < 1 X 10-5 within a run and < 5 X 10-5 during all of the runs of this study. The 24Mg+ signal was used as a reference peak for calculating the 25MgPMg and 26Mg/24Mg abundance ratios. Because of the occurrence of mass-dependent isotope fractionation during thermal ionization, the measured isotope ratios were corrected in the following way.21 The measured (m) isotope ratio (25Mg/24Mg), was normalized to the reference (0) value for terrestrial magnesium, (25MgPMg)o = 0.12663, given by Catanzaro et aZ.,20 to obtain the correction factor a = (25Mg/24Mg),/(25Mg/24Mg)O.a was used to calculate the fractionation-corrected (c) ratio (26Mg/24Mg), = (26Mg/ 24Mg),/a2. This fractionation correction is useful only if theAnalyst, July 1996, Vol. 121 903 magnesium stable isotope tracer used is highly enriched in 26Mg so that the admixture of the tracer to normal magnesium causes neglegible disturbances to the 25Mg/z4Mg ratio. For the means of the fractionation-corrected ratios (26Mg/z4Mg), of each of the analysed Mg standards and 26Mg- enriched samples, deviations (8) in %O relative to the reference value (26MgPMg)o = 0.139 733 +_ 0.000 028, i.e., the mean for all natural magnesium standard runs, are reported: 826Mg = [(26Mg/24Mg)c/(26Mg/24Mg)0 - 11 X 103. Results and Discussion The results for isotopically normal samples are compiled in Table 1.The data include both natural Mg salts and microwave- digested biological samples. The fractionation-corrected 26Mg/ 24Mg ratios obtained for both sets are identical within experimental error. The grand mean of the Mg standard runs is 0.139733 k 0.000028. This result lies within the range of previously published TIMS data for normalized 26Mg/24Mg ratios of Mg standards, which show a spread of more than 3.5%.21 The main reason for such differences is the occurrence of variable mass fractionation effects in different mass spec- trometers, e.g., the present result is about 0.3% lower than the result obtained for the same Mg standard using a different mass spectrometer (0.140 16 k 0.000 1216). The external precision of the fractionation-corrected 26Mg/ 24Mg ratios of the Mg standards and of the microwave-digested samples from Table 1 is <0.02%.It is better than previously reported precisions (0.03%13 and 0.1 %14) for Mg isotope ratio measurements which were performed to evaluate the biological applicability of TIMS using chemically extracted Mg and the silica gel-phosphoric acid loading technique or no ion emitter at all. Table 1 Isotopically normal samples: natural Mg standards and microwave- digested biological- samples Mg standards* Total bloods Blood serums Urine$ Muscle (bovine)§ Faecess (26Mg/24Mg),* 0.139670f11 0.139780f 18 0.139 732 f 34 0.139741 +24 0.139728+31 0.139 720 + 32 0.139727.t 14 0.139 810 f 6 0.139 690 k 26 0.139 755 k 30 0.139741 f 18 0.139788k35 0.139 714 f 28 0.139 734 f 8 1 0.139788f27 0.139 764 f 54 0.139721 f 7 3 0.139789k 18 0.139730f 12 0.139 744 + 71 0.139756k 16 626Mg (%o).'- -0.45 f 0.22 0.34 f 0.24 0.06 f 0.27 -0.01 f 0.32 -0.03 f 0.30 -0.10 f 0.3 1 -0.04 f 0.23 0.55 f 0.21 0.16 k 0.29 0.05 k 0.24 0.40 f 0.32 -0.13 f 0.29 0.01 + 0.22 0.39 + 0.28 0.22 + 0.44 0.40 * 0.24 0.08 f 0.24 0.16 f 0.23 -0.31 f 0.27 -0.09 + 0.21 -0.03 f 0.22 * (26Mg/24Mg), are means of fractionation-corrected ratios (see text). The errors correspond to the last figures shown and are 20 of the mean of all ratios of a single analysis (0 = absolute standard deviation).t 626Mg = [(26Mg/24Mg)c/(26Mg/4Mg)0 - 11 X 103, where (26Mg/24Mg)o = 0.139 733 f 0.000 028 is the mean of all fractionation-corrected Mg standard runs. The errors are calculated using the error propagation law and are 20.Natural Mg standards used were all MgC12 from Merck (Titrisol). 50-100 ng of Mg were loaded together with silica gel and 0.5 moll-' boric acid on rhenium filaments. Microwave-digested samples were loaded as described in the previous footnote. Fig. 1 shows the respective means of the measured 25Mg/ 24Mg and 26Mg/24Mg ratios of the normal runs. The data are aligned along the trend predicted by mass-dependent isotope fractionation. The relative range of fractionation of the 25Mg/ 24Mg ratios of approximately 4% and of the 26Mg/24Mg ratios of approximately 7%0 demonstrates that the described method can be used in dual tracer studies, i.e., the simultaneous administration of 25Mg and 26Mg in the same individual. The relative shifts of the isotope ratios caused by fractionation (as shown in Fig.1) are comparable to recently reported values of external precisions of Mg isotope ratio measurements per- formed with inductively coupled plasma mass spectrometry using a quadrupole mass analyser.loJ1 The results for the 26Mg-enriched samples are summarized in Table 2. The measured 6 26Mg values are compared with the 826Mg values calculated from gravimetry of calibrated solu- NORMAL RUNS UNCORRECTED RATIOS F 0.1 356 a! I d 0.1354 s, $' El-' '- 0 NATURALMg I 1 0 MICROWAVE DIGESTED BIOLOGICAL SAMPLES I 0.1348L ' ' ' ' I I I ' I ' 0.1244 0.1245 0.1246 0.1247 0.1248 0,1249 25Mg / 24Mg Fig. 1 Three-isotope representation of the uncorrected Mg isotope ratios of the normal runs. Each data point represents the respective means of the measured 25MgPMg and 26MgPMg ratios of a run. The line is an ideal fractionation line through (25Mg/24Mg)o = 0.126 63, i.e., the normalization value taken from Catanzaro et a1.,20 and 26MgPMg = 0.139 733, i.e., the mean of the natural fractionation-corrected Mg standard runs, using the fractionation law given in the text.Table 2 Isotopically enriched samples: calculated (CAL) and measured (MEA) 26Mg enrichments Sample* M U U S B S B U M B F M B U S F 626MgCAL (%o)t 0.78 1.12 1.40 2.41 4.97 20.99 22.17 52.43 83.74 179.13 211.63 489.64 803.59 829.28 837.25 1028.45 626MgMEA (%o)t 0.92 f 0.25 1.04k0.32 . 1.65 f 0.30 2.44 f 0.25 4.80 f 0.33 22.28 f 0.28 23.79 k 0.35 5 1.27 k 0.29 85.47 k 0.26 182.74 f 0.49 213.83 k0.31 485.84 f 0.33 800.39 f 0.49 833.51 f 0.63 836.69 k 0.48 1024.29 k 0.61 * Enriched samples were made up using mixtures of calibrated isotopically normal sample solutions (S = natural Mg standard; B = blood; F = faeces; U = urine; M = bovine muscle) and 26Mg tracer solutions.where (26Mg/24Mg)CAL,MEA are fractionation-corrected means. Errors given are 20. i' 626MgCALMEA = [(26Mg/24Mg)CAL,M"A/(z6Mg/24Mg)~ - 11 x 103,904 Analyst, July 1996, Vol. I21 tions. The relationship between the measured (y) and the calculated (x) 626Mg values was analysed using linear regres- sion analysis, yielding y = 0.998~ + 0.79 (correlation coefficient r2 = 1 .OOO). The excellent linearity is not dependent on the matrices analysed and demonstrates the accuracy of the method over the wide range of enrichment. To test how small an 26Mg enrichment could be detected, five samples with a2“g values of < 5%0 were analysed (see Table 2).Linear regression analysis (y-intercept = 0) of these data yield y = 0.987~ (r2 = 0.989). Deviations of the data from the regression line are <(0.3 f 0.3)%0. It is apparent from these results that 26Mg enrichments of 2 1%0 over the normal 26Mg/ 24Mg value can be resolved in biological samples by the method described here. The following is an example of the potential of the method if it is used in a human tracer experiment. The magnesium content of the blood plasma of a 70 kg human is about 60 mg. An intravenous administration of about 6 mg of 26Mg would lead to a 6”6Mg of approximately 1000%0. Because of the rapid turnover of plasma magnesium, the 626Mg declines to about 20%0 at 50 h and to the quoted detection limit after about 600 h.7 Hence measurements of the isotope appearance-disappearance pattern in blood, urine and faeces are easily resolvable with this method.The high precision and accuracy of the method make it possible to use it for cellular uptake studies of magnesium stable isotope tracers in intact humans. The results of our study demonstrate that the DLT reported here provides an easy, rapid and very precise method for the magnesium isotope analysis of biological samples. The DLT is especially appropriate for the analysis of samples with small Mg stable isotope enrichments. The possibility of analysing such small enrichments has great potential for extending the application of Mg stable isotopes in human tracer studies.We are very much indebted to Professor V. Krivan and P. Barth (Sektion Analytik und Hochstreinigung, Universitat Ulm) for the possibility of using their microwave digestion facility and sharing with us their valuable analytical advice. We thank Professor F. Begemann, Professor A. W. Hofmann and Professor H. Wanke for their interest and support and H. Feldmann for his excellent technical assistance. We benefitted considerably from discussions with J. Vogt and U. Wachter. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Sandstrom, B., Fairweather-Tait, S., Hurrel, R., and van Dokkum, W., Nutr. Res. Rev., 1993, 6, 71. Crews, H. M., Ducros, V., Eagles, J., Mellon, F. A,, Kastenmayer, P., Luten, J. B., and McGaw, B.A., Analyst, 1994, 119, 2491. Abrams, S. A., J . Pediatr. Gastroenterol. Nutr., 1994, 19, 1.51. Currie, V. E., Lengemann, F. W., Wentworth, R. A., and Schwartz, R., Int. J . Nucl. Med. Biol., 1975, 2, 159. Schwartz, R., Fed. Proc., Fed. Am. Soc. Exp. Biol., 1982, 41, 2709. Schuette, S. A., Ziegler, E. E., Nelson, S. E., and Janghorbani, M., Pediutr. ReA., 1990, 27, 36. Stegmann, W., and Karbach, U., Biol. Mass Spectrom., 1993, 22, 441. Schwartz, R., Spencer, H., and Wentworth, R. A,, Clin. Chim. Acta, 1978, 87, 265. Schwartz, R., and Giesecke, C. C., Clin. Chim. Acta, 1979, 97, 1. Schuette, S., Vereault, D., Ting, B. T. G., and Janghorbani, M., Analyst, 1988, 113, 1837. Cary, E. E., Wood, R. J., and Schwartz, R., J . Micronutr. Anal., 1990, 8, 13. Garner, E. L., Machlan, L. A., Gramlich, J. W., Moore, L. J., Murphy, T. J., and Barnes, 1. L., Nat. Bur. Stand. Spec. Publ., 1976, No. 422, 951. Turnlund, J. R., and Keyes, W. R., J . Micronutr. Anal., 1990, 7, 117. Vieira, N. E., Yergey, A. L., and Abrams, S. A., Anal. Biochem., 1994, 218, 92. Lee, T., Papanastassiou, D. A., and Wasserburg, G. J., Geochim. Cosmochim. Acta, 197’7, 41, 1473. Stegmann, W., and Begemann, F., Earth Planet. Sci. Lett., 198 1, 55, 266. Fassett, J. D., and Paulsen, P. J., Anal. Chem., 1989, 61, 643A. Krushevska, A., Barnes, R. M., and Amarasiriwaradena, C., Analyst, 1993,118, 1175. Barling, J., Goldstein, S. L., and Nicholls, I. A., J . Petrol., 1994, 35, 1017. Catanzaro, E. J., Murphy, T. J., Garner, E. L., and Shields, W. R., .I. Res. Natl. Bur. Stand. ( U S . ) , 1966, 7QA, 453. Esat, M. T., Nucl. Instrum. Methods Phys. Res., Sect. B , 1984, 233, 545. Paper. 6100060F Received January 3, 1996 Accepted March 8, 1996

ISSN:0003-2654    

DOI:10.1039/AN9962100901

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (905-908) 905 Uranyl Photophysics on Colloidal Silica: an Alternative Lumi nescence-en hanci ng Medium for Uranyl Assay Martin Lopez and David J. S. Birch* Department of Physics and Applied Physics, Strathclyde University, 107 Rottenrow, Glasgow, UK G4 ONG The luminescent behaviour of uranyl (UOz2+) in aqueous colloidal silica solution is illustrated and its biexponential decay with time investigated. The two components have lifetimes of 240 and 55 ps and after about 1.5 h the decay becomes monoexponential with a lifetime of 240 ps and luminescence quantum yield approaching 25 % . The absorption and emission spectra are red shifted and greatly enhanced compared with those in dilute aqueous perchloric acid solution. The suitability of aqueous colloidal silica as a dispersed medium for enhancing luminescence to detectable levels for concentration measurements is also shown, with uranyl concentrations down to 10 nmol dm-3 (0.2 ppb) being routinely measurable.The luminescence of uranyl on dried silica gel is found to be almost identical with that of uranyl in aqueous colloidal silica, illustrating that the uranyl is extremely closely associated with the silica surface and thereby sheltered from quenching by water molecules. This colloidal enhancing medium has advantages over traditional methods used in that a longer lifetime is achieved, diffusional quenching by anions is considerably reduced and the uranyl can be easily extracted from bulk solution. Keywords: Uranyl luminescence; colloidul silica; uranyl assay; photophysics Introduction The absorption and luminescent properties of uranyl (U0z2+) have long been used to detect and determine the concentration of uranyl in solution by means of steady-state and time-resolved measurements.1-5 Absorption techniques have the inherent disadvantages of the lack of sensitivity because of uranyl's low molar absorptivity and interference by any other absorbers present. Time-resolved luminescence measurements have the advantage over steady-state measurements that the presence of diffusional quenchers can be compensated for since the experimentally determined concentration relies only on the initial intensity of the decay, regardless of how much the lifetime is quenched and background fluorescence due to any organic impurities present (typically of nanoseconds duration) can be time-gated out.The concentration of uranyl is then usually determined by reference to a standard of known concentration in the same medium. Since the luminescence of uranyl in dilute aqueous acidic solution is weak and inefficient (a =: 0.1%)6 owing to quenching by water molecules,7 it is normally detected in the steady-state or time-resolved domain by the addition of an enhancing agent of some sort. The luminescence of uranyl can be enhanced in aqueous solution by complexation, which can increase both the molar absorptivity at ' To whoin correspondence should be addressed. the excitation wavelength and the lifetime of decay by shielding the uranyl molecule from the surrounding water molecules. Some traditional aqueous enhancers of uranyl luminescence are phosphoric acid, phosphate complexes, sulfates and fluo- rides.3.4.8--1" Although the useful ion-exchange properties of silica gel with regard to uranyl in aqueous solutions have long been known,11-14 little attention has been paid to the effect of silica gel15 and colloidal silica'b on the spectroscopic properties of uranyl and its analytical potential in aqueous colloidal silica as a means of concentration determination.l77l8 Silica colloids dispersed in aqueous solution are discrete uniform spheres of silica with no internal area or detectable crystallinity. In alkaline solution the colloids have an overall negative charge, which results in a mutual repulsion and formation of stable products in solution. In the colloidal silica used here (Ludox AM), some of the surface silicon atoms have been replaced by aluminium, which creates a fixed negative charge independent of pH and stabilizes the colloids in the neutral pH range.19 The structure of the colloidal silica used is represented in Fig.1. The internal structure of the colloids is made up of tetragonal silicate units, and at the surface the oxygen atoms which are not required to make up the Si-0-Si bridges are free to bind with other ions from the solution. For an alkaline solution obtained using sodium hydroxide, the counter ion Na+ will be associated with the negatively charged silica surface. These sodium ions can be replaced by other cations with a greater affinity for the oxygen. This is what happens with uranyl, which is attracted to and binds with the silica surface.Colloidal silica can also be turned into a solid or gel when a three-dimensional gel network forms as the colloid particles interconnect due to the formation of siloxane bonds (Si-0-Si). This gelling process is useful for trapping the uranyl in a stable matrix. In addition, pre-made silica gel can also extract the uranyl from solution and can then be filtered out and dried. The colloidal silica medium has the advantages over conventional enhancers mentioned above of easy separation from bulk solution and, as is shown in this paper, that the negatively charged silica surface reduces the effects of anionic quenchers. OH X I bulk water Fig. 1 Surface structure of Ludox AM.906 Analyst, July 1996, Vol.121 Experimental All time-resolved measurements were taken with a time- resolved fluorimeter20 which used a pulsed nitrogen laser excitation source [LSI (Newton, MA, USA) 337ND, he, = 337 nm, 1-20 Hz, 250 pJ per pulse] and photomultiplier detector [Hamamatsu (Hamamatsu City, Japan) R29491 and the decays (Aern > 450 nm] were averaged by a transient digitiser [LeCroy (Chestnut Ridge, NY, USA) 93101 over 100 pulses before being transferred to a PC for analysis with IBH (Glasgow, UK) software as a 500-point data curve. Steady-state emission measurements were made using a Shimadzu RF540 (Kyoto, Japan) spectrofluorimeter and steady-state absorption with a Perkin-Elmer (Norwalk, CT, USA) Lambda 2 spectrometer. The colloidal silica used was Ludox AM, manufactured by DuPont (Wilmington, DE, USA), which consists of particles of 12 nm diameter.The solution bought was diluted by a factor of 2 to give a 15% silica solution by weight at pH = 9. The gel used was purchased from Aldrich JMilwaukee, WI, USA) (70-270 mesh, average pore size 60 A). The uranyl was added from a stock solution made from uranyl nitrate crystals (EML, Reading, Berks., UK) dissolved in de-ionized water. The cuvettes used were standard 4 cm3 disposable plastic cuvettes (Hughes and Hughes, Wellington, Somerset, UK). All measure- ments were made at 25 "C. For measurements in aqueous colloidal silica at low uranyl concentrations, the background of a blank without any uranyl present, due to the large scattering effect and luminescence of the colloidal silica itself, was subtracted.Results Absorption and Emission Spectra Uranyl on the silica surface experiences a very different environment from that of aqueous uranyl and this is reflected by the large changes in its absorption and emission spectra (Fig. 2). The absorption is increased across the whole range and, in the visible absorption region, it can be seen that there is a large increase in the molar absorptivity from a maximum of about 9.0 to 24.4 dm3 mol-1 cm-1 and that the maximum red shifts by 10 nm from 414 to 424 nm. In addition to being greatly enhanced in intensity, the emission is red shifted with the first peak increasing by 20 nm from 487 to 507 nm. Stability with Time The behaviour of uranyl on silica is not stable with time and at first exhibits biexponential luminescent decay with a longer 350 400 450 500 550 600 650 Unm Fig.2 Absorption and emission spectra of uranyl on colloidal silica. a, Uranyl absorption (colloidal silica); b, uranyl emission (colloidal silica); c, uranyl absorption (perchloric acid); and d, uranyl emission (perchloric acid). (major) component of lifetime 240 ps and a shorter (minor) component of lifetime about 55 ps (see Fig. 3 for 0.67 mmol dm-3 uranyl). Both lifetimes are considerably longer than that measured for simple hydrated uranyl in dilute aqueous acidic solution (1.4 ps). This decay eventually becomes completely monoexponential after about 1.5 h with a lifetime of 240 ps (Fig. 4). This lifetime is even longer than that achieved by addition of phosphoric acid8 ( < 170 ps). The change in relative amplitude of the longer component with time is illustrated in Fig.5 , which was calculated using a biexponential decay fit and fixing the lifetime of one of the components at 240 ps. The shorter component is probably due to uranyl ions present as hydrolysed polymers in solution or to uranyl in a different association with the silica surface where it is more accessible to water quenching. This biexponential behaviour seems to occur at all concentrations, since, when studied for concentrations in the range 10-1000 pmol dm-3, the uranyl at first exhibited biexponential decays in all cases with similar lifetimes, the majority component making up about 90% of the decay (Table 1). Enhancement of Emission Intensity The increase in steady-state emission intensity can be attributed to the increase in molar absorptivity and quantum yield.The expected relative increase for weakly absorbing solutions, R, in the integrated emission intensity for the same concentration of uranyl adsorbed on silica and in its simple hydrated form with the same excitation intensity should be21 If (9)dV I@ (1) -- R = - lorn If0 (WV Eo@o where I f ( 9 ) is the intensity at wavenumber 9, I is the molar absorptivity at the excitation wavelength, @ is the quantum yield and the subscript o indicates the hydrated uranyl. As an example, R for uranyl on colloidal silica where all the uranyl is emitting with a monoexponential lifetime (240 ps) in compari- son with the pure uranyl ion in 10-2 mol dm-3 perchloric acid was 665. From eqn. (l), taking E = 24.4 dm3 mol-1 cm-1 and I, = 9 dm3 mol-1 cm-1, this implies that the quantum yield has increased by a factor of 245.The value of Q0 for simple hydrated uranyl in dilute aqueous acidic solution is about 0.1 % and so @ for uranyl on the silica is approaching 25%. Wheeler and Thomas16 examined the behaviour of uranyl on NALCO silica No. 11 15 (pH = 10.4, diameter = 80 8, at 20% dilution) and observed a discrepancy between the measured 420 360 30 0 240 180 120 60 n - I ' I ! 1 ~ 1 1 I 0 i a 20 30 40 50 Time/ min Fig. 3 Biexponential behaviour of uranyl on colloidal silica.Analyst, July 1996, Vol. 121 907 increase in emission intensity and that expected due to the increase in lifetime, and attributed this to emission from a different energy level with a different lifetime and quantum yield.However, they did not consider the effect of stability with time and compared their results with hydrolysed uranyl in water, In this instance, with Ludox AM, no such discrepancy was observed. Linearity of Intensity with Concentration Uranyl was added via a microsyringe to a 3 cm3 sample of colloidal silica solution in a cuvette with mixing and the steady- state intensity was measured immediately. The intensity was found to vary linearly with concentration, as can be seen in Fig. 6. This illustrates that despite being unstable with time (biexponential decay), on a practical level this could still be used as a means for concentration determination via fast measurement procedures and the large enhancement of intensity with respect to the uncomplexed uranyl ion increases detection limits.In the time domain, which is normally more useful if there are dynamic quenchers present since they can be corrected for, there is a corresponding linear increase of initial intensity of the decay for the 240 ps component with concentration, and this is also shown in Fig. 6. The enhancement of luminescence allowed measurements at much lower concentrations than normal to be made and detection limits with the current instrumental arrangement were lowered by a factor of 100 with t l o o ! ' I ' I ' I ' 1 ' I ' ! 0 40 80 120 160 200 240 time/ min Fig. 4 Change in average lifetime (monoexponential fit) with time. 100 95 90 85 a o f , , I . I , . 0 20 40 60 80 100 time/ rnin Relative amplitude of 240 ps component.Fig. 5 solutions down to about 10 nmol dm-3 (about 0.2 ppb) being routinely measurable. This is of a similar sensitivity to that obtained using other aqueous enhancers in equivalent experi- mental set-ups9 and could be further improved through instrumental improvements such as optimization of emission signal collection and delivery to the photomultiplier detector and use of an interference or bandpass filter rather than a cut-off filter, which would reduce the effect of background lumines- cence. Quenching by Ions in Colloidal Silica Solution The nature of the silicate surface should affect the quenching rates in comparison with free ions in acidic solution. Since the surface is negatively charged, positively charged ions which are not firmly bound to the surface should quench more strongly because they are electrostatically attracted towards the silica surface and negatively charged ions should quench less since they are repelled.In order to test this, silver and iodide ions were used, both of which are very strong uranyl quenchers and the Stern-Volmer quenching rates were calculated (see Fig. 7). Higher positively charged ions will be highly hydrolysed and also probably interact strongly with the surface. At higher concentrations of silver than those used for the experiment ( > 40 pmol dm-3), the behaviour becomes non-Stern-Volmer and a precipitate forms. A uranyl concentration of 1.7 X 1 0-4 mol dm-3 was used in both cases (monoexponential T unquenched = 240 ps). The calculated Stern-Volmer quenching rates were (4.57 f 0.48) x 108 and (7.72 k 0.47) X 106 dm3 mol-1 s-1 for silver and iodide ion respectively which compare with (2.02 k 0.13) X 109 and (4.78 f 0.28) X lo9 dm3 mol-1 s-1 for the same ions Table 1 Biexponential behaviour of uranyl at different concentrations (taken about 10 min after mixing) Concentration1 pmol dm-3 Fit %Ips SIPS zdps sips x2 10 Monoexp.50 Monoexp. 100 Monoexp. 500 Monoexp. 1000 Monoexp. B iexp . Biexp. Biexp. Biexp. B iexp . 169 2.19 1.864 251 20.9 64.9 3.84 1.258 181 1.65 2.125 244 12.0 64.0 2.97 1.119 181 1.63 2.016 223 7.47 54.5 3.77 1.301 188 1.43 1.683 238 9.12 65.6 3.55 1.010 177 1.20 2.121 246 12.2 67.6 2.44 1.279 2 2.5 2 ki 1.5 c.. I - n Y z r .- 0.5 P) Y c m .- u - o -0.5 I I I I I I I I I I -8.5 -7.0 -5.5 -4.0 -2.5 log ( [U0,2+]/ rn~ldrn-~) Fig.6 Graph of luminescent intensity versus uranyl concentration.908 Analyst, July 1996, Vol. 121 in perchloric acid. These results agree with the expected trend. The decreased value for the silver ion can be explained by the increased viscosity of vicinal water, the electrostatic interaction of the silver ions with the silica and the decreased mobility of the uranyl ion, which is bound to the silica surface. The very large difference for the iodide ion (about 1OOOX) reflects the strong repulsive effect of the colloid surface. Uranyl on Dried Silica Gel To establish the effect of uranyl on dried silica gel, a known volume of silica gel was mixed with aqueous uranyl solution for 2 h, during which time all the uranyl present was absorbed. The solution was filtered and the silica gel left to dry for 24 h at room temperature in a fume cupboard.The average concentration of the uranyl in the silica gel was 10-3 mol dm-3. The measured monoexponential lifetime of uranyl on the dried silica gel was 248 3 4.74 ps (standard deviation), which is the same as the longer component in aqueous colloidal silica solution, and the emission spectrum is also very similar. This lifetime of about 240 ps is comparable to the lifetimes found for uranyl in silicate glasses,**3*3 which indicates that the uranyl is virtually inaccessible to quenching by water molecules once it is bound to the silica surface. Conclusions The aqueous colloidal silica used has been shown to have a very strong effect on the spectroscopic properties of uranyl, enhancing its absorption and emission and increasing its lifetime by over two orders of magnitude from 1.4 to 240 ps.The biexponential behaviour observed eventually becomes monoexponential with just the 240 ps component present. Once adsorbed, the uranyl experiences an environment almost 30 -? “0 20 10 independent of the surrounding aqueous medium, which is shown by its high quantum yield and the similarity with its luminescence on dried silica gel. Colloidal silica appears to have considerable potential as an alternative to phosphoric acid/ phosphate addition as a luminescence enhancer for measuring the concentration of uranyl in solution because it has the advantages of being easily extracted from solution by physical means such as by filtration or gelation and of reducing the effects of any anionic quenchers present. We acknowledge the financial support of the Engineering and Physical Sciences Research Council and British Nuclear Fuels plc.References 1 2 3 4 .5 6 7 8 9 10 11 12 13 14 1.5 16 17 18 19 20 21 22 I 0 ’ I I I 0 1 2 3 4 [I-]/mmol dm” or [Ag’]/104mol dm” Quenching of uranyl luminescence by silver and iodide ions. Fig. 7 23 Matsui, T., Kitamori, T., Fujimori, H., Suzuki, K., and Sakagami, M., J . Nucl. Sci. Technol., 1992, 29(7), 664. Fujimori, H., Matsui, T., and Suzuki, K., J . Nucl. Sci. Technol., 1988, 25(10), 798. Jia, W. J. He, A. O., Wang, Z. L., and Chin, C. T., J . Radioanal. Nucl. Chem. Lett., 1986, 108(1), 33. Kenney-Wallace, G. A., Wilson, J. P., Fanell, J. F., and Gupta, B. K., Talanta, 1981, 28, 107.Ishibashi, K., Sakamaki, S., Imasaka, T., and Ishibashi, N., Anal. Chim. Acta, 1989, 219, 181. Jorgensen, C. K., and Reisfeld, R., Strurt. Bonding (Berlin), 1982,50, 121. Moriyasu, M., Yokoyama, Y., and Ikeda, S., Inorg. Nucl. Chem., 1977,39, 2211. Zhou, P., Wang, Z. L., Xu, Y. Z., Zhang, L., and Wang, Y. B., .I. Radioanal. Nucl. Chem. Lett., 1993, 175(2), 81. Cheng, C. K., Wang, Z. L., Liu, X. N., Tang, F. X., Pan, X. X., and Zheng, C. F., IEEE J . Quantum Electron., 1986, 22(7), 998. Moriyasu, M., Yokoyama, Y., and Ikeda, S., J . Inorg. Nucl. Chem., 1977, 39, 1199. Ahrland, S., Grenthe, I., and Nor&, B., Acta Chem. Scand. 1960, 14, 1059. Ahrland, S., Grenthe, I., and Norin, B., Acta Chem. Scand. 1960, 14, 1077. Hafez, M. B., and Hafez, N., Isotopenraxis, 1990, 26, 7. Milonjic, S. K., CokeSa, M., and Stevanovic, R. V., J . Radioanal. Nucl. Chem. Articles, 1992, 158(1), 79. Thomas, J. K., and Wheeler, J., .I. Photochem., 198.5, 28, 285. Wheeler, J., and Thomas, J. K., J . Phys. Chem., 1984, 88, 750. Romanovskaya, G. I., Gorpenko, V. A., Lebedeva, N. A., and Trifonova, 1. A., Zuvod. Lab., 1992, 58(8), 26. Romanovskaya, G. I., Zakharova, G. V., and Chibisov, A. K., Zh. Anal. Khirn., 1984, 39(5), 930. Ludox@ Colloidal Silica: Properties, Uses, Storage and Handling, Du Pont, Wilmington, DE, 1987. Lopez, M., PhD Thesis, Strathclyde University, 1995. Birks, J. B., Photophysics of Aromatic Molecules, Wiley, Chichester, 1970. Allsopp, S. R., Cox, A., Kemp, T. J., Reed, W. J., Carassiti, V., and Traverso, O., J . Chem. Soc., Faraday Trans. I , 1979, 75, 342. Pandey, K. K., Indian J . Pure Appl. Phys., 1991, 29, 362. Paper 6/01 251 E Received February 2 I , I996 Accepted April 3, 1996

ISSN:0003-2654    

DOI:10.1039/AN9962100905

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (909-912) 909 Second-derivative Synchronous Fluorescence Spectroscopy for the Simultaneous Determination of Naproxen and Salicylic Acid in Human Serum Dimitrios G. Konstantianos and Pinelopi C. Ioannou* Laboratory of Analytical Chemistry, University of Athens, Panepistimiopolis, 15771 Athens, Greece Second-derivative synchronous fluorescence spectrometry was used to develop a simple, rapid and sensitive spectrofluorimetric method for the simultaneous determination of naproxen and salicylic acid in human serum. The method is based on the intrinsic fluorescence of naproxen and salicylic acid in chloroform-1% acetic acid solution. A A1 of 130 nm was used for the direct measurement of salicylic acid in the binary mixture, whereas naproxen was determined from direct measurements at AA = 60 nm and by means of a correction equation which incorporates the concentration of salicylic acid.The range of application is 0-14 mg 1-I for naproxen and 0-13 mg 1-1 for salicylic acid. The detection limits for naproxen and salicylic acid are 0.003 and 0.01 mg 1-1, respectively. Serum samples are extracted into chloroform-1 % acetic acid solution prior to instrumental measurement. Analytical recoveries range from 97 to 105% (mean 102%) for naproxen and from 97 to 112% (mean 103%) for salicylic acid. The within-run precision (RSD) for the method for four naproxen-salicylic acid mixtures varied from 1.2 to 6.7% and the day-to-day precision for mixtures varied from 2.1 to 5.0%. Keywords: Naproxen; salicylic acid; second-derivative synchronous fluorescence spectrometry; serum Introduction A variety of interesting applications of synchronous (scanning) derivative fluorescence spectrometry (SDSFS) have been reported during the last decaub.14 The theoretical aspects of the SDSFS technique and its advantages and limitations have been the topics of several reports.576 Most of the applications of this technique are concerned with agricultural, environmental and food chemistry and also pharmaceutical analysis for the direct resolution of mixtures, which usually involves time-consuming and expensive techniques.Naproxen, [(+)-2-(6-methoxy-a-methyl-2-naphthaleneacetic acid] (NAP), is a non-steroidal anti-inflammatory drug, often preferred to acetylsalicylic acid (aspirin) because of its better absorption following oral administration and fewer adverse effects.In plasma or serum, naproxen is the predominant species? Among the methods recommended for the determina- tion of unchanged naproxen in serum, gas chromatographic methods8 are tedious, time consuming and relatively in- sensitive. Other chromatographic methods, such as high- performance liquid chromatography (HPLC), although pro- viding high sensitivity when fluorimetric detection is used,9 are * To whom correspondence should be addressed. also time consuming and require special and expensive apparatus. Non-chromatographic methods, such as UV spec- trophotometry,lO mass fragmentography 1 1 and spectrofluor- imetry,12 have also been reported. Despite the excellent sensitivity of fluorimetric methods, they are not entirely specific.In particular, the most serious interference in the fluorimetric assay of NAP in serum occurs from salicylic acid (SA), the main metabolite of acetylsalicylic acid in man, because of severely overlapping spectral bands. Therefore, this method is not suitable for the determination of serum naproxen in patients receiving salicylates simultaneously. Here, we report a method for the simultaneous determination of NAP and SA in serum using the SDSFS technique. The method is based on the intrinsic fluorescence of NAP and SA in chloroform-1% acetic acid solution. NAP and SA are deter- mined in the organic layer after extraction with chloroform-1 % acetic acid solution. The method is as sensitive as the previously described fluorimetric method12 but, owing to its simplicity and greater specificity, it can easily be applied to routine clinical or pharmacokinetic studies.Experimental Apparatus We used a Model 5 12 fluorescence spectrometer (Perkin-Elmer, Norwalk, CT, USA) interfaced to an IBM-PC 386DX micro- computer.l3 The spectrometer was equipped with a 150 W xenon arc lamp and a magnetic stirrer under the sample-cell holder. All measurements took place in a standard 10 mm (pathlength) quartz cell, thermostated at 25.0 * 0.5 "C. The instrument settings were energy mode, dynode voltage 750 V and excitation and emission bandwidth 20 nm. Excitation and emission monochromators were locked together and scanned simultaneously with a constant difference Ah = he, - hex. The scan speed and response time of the spectrometer were set at 4 nm s-l and 'Fast' mode, respectively.Reagents Spectroscopic grade chloroform (Merck, Darmstadt, Germany) was used to prepare a 1% v/v solution of acetic acid in chloroform. Henceforth this mixture will be referred as 'mixed solvent'. Stock standard solutions of NAP (Sigma, St. Louis, MO, USA) and SA (Fluka, Buchs, Switzerland) containing 400 mg 1-1 were prepared in the mixed solvent. Working standard solutions were prepared by dilution with the mixed solvent. The stock standard solutions of NAP and SA were stable for several months at room temperature. Aqueous stock standard solutions of SA and NAP containing 1000 mg 1-l were prepared in distilled, de-ionized water at pH 11.6. These solutions were stable for at least 1 month at room temperature, and were used for recovery experiments in albumin solutions and in normal910 Analyst, July 1996, Vol.121 control serum (Wellcome Clinical Chemistry, QA Programme). A stock standard solution of bovine serum albumin (Sigma) containing 100 g 1-I was prepared in distilled, de-ionized water. Sample Treatment To 0.100 ml of serum containing 0.50-12.0 pg of NAP and 0.50-10.0 pg of SA, add 0.400 ml of the mixed solvent, sonicate the mixture for 5 min and centrifuge for 3 min at 1500g. Use the organic layer for the measurement of SA and NAP. Spectrofluorimetry Transfer 0.100 ml of the organic layer into a cuvette, add mixed solvent to a total volume of 2.00 ml and start the stirrer. Record the synchronous fluorescence spectra by scanning both mono- chromators simultaneously at a constant difference Ah = 130 nm (hex = 280-380 nm).Hereafter all wavelengths referred to synchronous spectra are taken as equal to those of the corresponding excitation wavelengths. Evaluate the second- derivative signal for SA, AZSA, within the spectral range 316-344 nm. Calculate the total Concentration of SA in serum, CSA, Semm, directly from the calibration curve obtained by plotting AZSA versus SA concentration of control serum spiked with SA (concentration range 5-100 mg 1-1) and treated similarly. Record the synchronous fluorescence spectra of the same sample by scanning both monochromators simultaneously at a constant difference Ah = 60 nm (hex = 230-380 nm). Evaluate the second-derivative signal for NAP in the mixture, AZNAp(SA), within the spectral range 264-284 nm and the signal contributed to both compounds, AZNAp + SA, within the spectral range 330-350 nm.Calculate the true signal of NAP in the cuvette, AZNAP, true, using the equation AJNAP, true = A~NAP(SA) (0.3 1 AINAP + SAIAINAP(SA) + 0.95) ( 1) Calculate the concentration of NAP in serum, CNAP(serum), directly from the calibration curve obtained by plotting the AZNAp versus the concentration of NAP in control serum spiked with NAP (concentration range 5-120 mg 1-I) and treated similarly. Results and Discussion Comparison of Spectra The fluorescent properties of salicylic acid in alkaline aqueous (hex = 300 nm, he, = 390 nm) and in chloroform-acetic acid (hex = 314 nm, he, = 444 nm) solutions have been reported previously.’ The fluorescent properties of NAP in alkaline aqueous solutions have also been reported, and a method based on these properties was proposed for the determination of naproxen in serum.12 To clarify the possibility of the simultane- ous determination of SA and NAP using their native fluores- cence, we examined the spectral characteristics of NAP and SA in aqueous and chloroform-acetic acid solutions. The excitation and emission spectra of NAP and SA in these solvents are shown in Fig. 1A and B, respectively. As can be seen from Fig. 1, the fluorescent characteristics of NAP in both solvents are similar, showing two excitation maxima at 284 and 330 nm (curves a and c) and an emission maximum at 355 nm (curves a’ and c’). The best spectral resolution of the mixture is observed in chloroform-1% acetic acid solution.However, none of the solvents is suitable for the simultaneous determination of NAP and SA by conventional fluorimetry because of the broad overlapping of the excitation band of SA and the emission band of NAP. Fig. 2 shows the synchronous fluorescence spectra of SA-NAP mixtures in the mixed solvent at different concentration ratios of the com- ponents and at different Ah values corresponding to the difference between the emission and excitation maxima for NAP (Ah = 60 nm), for SA (Ah = 130 nm) and an intermediate AIL value of 95 nm. The derivatives of these spectra show two AZ signals which can be related to the concentration of NAP and SA. It can also be seen from Fig. 2 that the analytical signal of NAP is influenced by the presence of SA, whereas the analytical signal of SA at Ah = 130 nm is not dependent on the concentration of NAP.A Wavelength Inm B 500 Wavelength /nm. Fig. 1 Fluorescence excitation (a-d) and emission (a’-d’) spectra of A, aqueous solutions (pH 11.6) of salicylic acid, C = 2.0 mg I - ] , he, = 300 nm, he, = 400 nm (a, a’), and naproxen, C = 2.0 mg 1-1, he, = 284 nm, he, = 355 nm (b, b’); and B, chloroform-1% acetic acid solutions of salicylic acid, C = 2.0 mg 1-l, he, = 314 nm, he,, = 444 nm (c, c’), and naproxen, C = 2.0 mg 1-I, he, = 282 nm, he, = 355 nm (d, d’). - 6 0 - 9.2 .$ 0 9.2 -0.2 I K P 240 320 400 240 320 400 240 320 400 Wavelength /mu Fig. 2 ( a x ) Synchronous fluorescence spectra of A, NAP; B, SA; and C, NAP-SA mixture and (d-f) their second-derivative spectra in the mixed solvent.(a) CSA/CNA~ = 12: 1, Ah = 60 nm; (b) CSA/CNAP = 1 : 8, Ah = 130 nm; and (c) CsAICNAp = 1 : 1 , Ah = 95 nm.Analyst, July 1996, Vol. 121 91 1 Selection of AIL In order to select the optimum wavelength difference between excitation and emission monochromators (Ah), we obtained the total fluorescence spectra of NAP and SA in the mixed solvent. The two-dimensional (contour plots) of these spectra are shown in Fig. 3. The parallel diagonal lines superimposed on the spectrum shown in Fig. 3 represent the scan paths through the excitation- emission matrix that would be obtained with synchronous scans at the wavelength interval shown. It is evident that at Ah = 60 nm, which corresponds to the difference between the emission and excitation maxima of NAP, the signal for NAP is the maximum and the influence from SA is minimal.The maximum analytical signal for SA is obtained at Ah = 130 nm and is not influenced by the presence of NAP. The Ah values between 60 and 130 nm are not suitable for measuring NAP-SA mixtures because of the smaller analytical signals for both compounds and the strong influence of each acid on the signal of the other. Fig. 3 Two-dimensional (contour plots) ‘total’ fluorescence spectra (background corrected) of (a) NAP and (b) SA. The dashed lines on the contour plots are the trajectories followed during synchronous scans. Therefore, for measuring NAP-SA mixtures, Ah values of 60 and 130 nm were selected for the determination of NAP and SA, respectively.General Analytical Characteristics General analytical characteristics for the determination of SA at Ah = 130 nm have been reported previously.1 The linear concentration range for NAP at Ah = 60 nm was 0-14 mg 1-1. Pearson’s correlation coefficient l 4 (Y) for the calibration graph was 0,9992 for NAP. The detection limit, defined as three times the standard deviation of the lowest concentration, was 0.003 mg 1-l. The relative standard deviations (RSDs), covering the range of interest for NAP (0.040, 0.40 and 8.0 mg 1-1) were found to be 2.4, 2.3 and 2.8%, respectively. 2 0.24 0.16 0.08 10 20 30 AISA+NAP I AINAP 15 45 80 100 130 ‘SA 1 ‘NAP Fig. 4 Effect of the analytical signal of salicylic acid on the signal of NAP (expressed as the signal of NAP) in the presence of SA (a) and the relationship of the correction factor (b) for NAP, V ~ Y S U S the ratio AZsA + NAP/ AINAP.Table 1 Determination of 0.40 mg 1- mixtures by the SDSFS technique of NAP in synthetic NAP-SA NAP found*/mg 1 - I SA : NAP (mass ratio) Uncorrected? Corrected+ (corrected) Recovery f s (%) 0.5 1 2 3 4 5 6 7 0.39 0.39 0.37 0.37 0.35 0.35 0.33 0.3 1 0.41 103 f 2 0.40 l 0 0 f 1 0.41 103k 1 0.40 99k 1 0.42 104 -1 2 0.39 98k2 0.38 96-14 Mean 100-12 0.39 97-12 . * Average of three measurements. + Without using correction. 4 After using correction eqn. (1). Table 2 Precision data (n = 10) KSD (%) Mean concentration/ mg 1-1 Within-run Day-to-day NAP SA NAP SA NAP SA 10.3 82.6 6.7 2.9 - - 23.4 39.3 5.4 2.9 2.3 5.0 52.4 10.1 3.2 3.8 - - 91.9 29.7 1.2 5.5 2.1 3.8 -____912 Analwt, .luly 1996, Vol.121 Determination of NAP and SA in Binary Mixtures We performed a detailed study on the influence of the signal of each acid on the analytical signal of the other at Ah = 60 and 130 nm. It was lound that the analytical signal of SA (Ah = 130 nm) was not affected by the presence of up to a 100-fold excess of NAP. Hence the determination of SA in the presence of NAP at the ratios found in serum could be performed with satisfactory accuracy and sensitivity. The effect of increasing concentration of SA on the analytical signal of NAP, A I N A p (264-284 nm), obtained by SDSFS at Ah = 60 nm is shown in Fig. 4. As can be seen, the analytical signal of NAP decreases with increasing concentration of SA (or the ratio CSA/CNAP).Moreover, the ratio of the signal of NAP in pure solutions, AZNAp,,,ue, to its signal in the presence of SA, AZNAP(SA), expressed as a correction factor, CF, is linearly related to the ratio of the signal AZNAp+ s A to the signal AZNAp(SA) according to the equation CF = 0.31 (f0.02) A l ~ p , p + ~ A / A l p ~ . , q y ~ A j i- 0.95 (kO.01) (1. = 0.9990) (2) The true (or corrected) signal for NAP, AZNAp, trL,e, is given by eqn. (1). Eqn. (2) was obtained by measuring 36 binary mixtures of NAP and SA covering the concentration range from 0.2 to 2.0 mg 1-I for each component and at 15 different concentraton ratios. As can be seen from Fig. 4, the linearity of eqn. (2) is extended up to a signal ratio of 25 or up to a 110-fold mass excess of SA. Table 1 summarizes results for the determination of NAP in the presence of SA in synthetic binary mixtures.These results were obtained by calculating the concentration of NAP from a calibration curve after correction of the analytical signal of NAP using eqn. ( 1 ) . As can be seen from Table I , NAP-SA mixtures could be resolved satisfactorily by using the SDSFS technique at Ah = 60 nm over a wide range of CSA/CNAp ratios. Also, NAP could be measured without any correction at CS,&'sAp ratios < 0.2. Serum Samples Serum samples and albumin solutions containing NAP and SA gave signals similar to those obtained with aqueous standard solutions. Recovery experiments on albumin solutions con- taining NAP and SA at several albumin and component concentrations gave values of 97 + 4% ( n = 5 ) and 99 f 3% ( n = 10) for NAP and SA, respectively, and these values were not dependent on the albumin concentration over the range 30-50 g 1-1.The detection limit of NAP in serum by the proposed procedure was 0.1 pg m - I . To determine the within-run precision, we measured three serum pools with different NAP and SA concentrations, ten times each. To assess the day-to-day Table 3 Determination of NAP and SA in synthetic serum mixtures Concentration/nig 1-1 Recovery Added Found" NAP SA NAP SA NAP SA 60.0 30.0 59.9 30.9 100 103 90.0 30.0 91.9 29.7 101 99 34.0 6.0 34.2 6.6 101 110 100.0 10.0 97.1 11.2 97 112 10.0 85.0 10.3 82.6 103 97 22.5 40.0 23.4 39.3 104 98 50.0 10.0 52.4 10.1 105 101 Mean 102 103 precision, we performed repeated analyses of two serum samples during 2 weeks (Table 2).The RSDs obtained were satisfactory for trace analysis. Recovery data for synthetic NAP-SA mixtures added to serum are shown in Table 3. The selected concentrations for binary mixtures are typical for NAP and SA levels in serum during the first 12 h from a typical subject following an oral dose of 650 mg of aspirin15 and 500 mg of napr~xen.~ Interference Studies We tested 15 commonly used drugs for their potential interference in the method by supplementing pooled serum containing NAP and SA with high concentrations of each drug (final concentration in serum 400 mg 1-I). Our criterion for interference was an analytical signal differing by more than +S% from that expected for NAP and SA alone. No interference was noted with any of the following drugs: antipyrine, amiloride, amitriptyline, amoxycillin, caffeine, carbamazepine, chlorpromazine, dithranol, ibuprofen, imipramine, indometha- cin, levodopa, diflunisal, phenacetin, theophylline, ketoprofen, tolmetin and azapropazone.Significant interference was ob- served only from 6-desmethylnaproxen (6-DMN), the main metabolite of NAP in serum, at 6-DMN/NAP ratios > O . l , which are not expected in serum samples. Conclusions The results obtained for the simultaneous determination of NAP and SA prove the usefulness of derivative synchronous fluorescence spectrometry in resolving mixtures with strongly overlapped emission excitation spectra. NAP-SA mixtures over a wide range of ratios that cannot be measured by conventional tluorimetry can be readily quantified by the combined technique using the appropriate equation which involves the influence of salicylic acid on the analytical signal of naproxen.The method is very sensitive, thus requiring only a small sample volume (100 pl), rapid, simple and inexpensive, since it requires no sophisticated detection equipment. We gratefully acknowledge the financial support from the Greek Ministry of Industry, Energy and Technology. References I 2 3 4 5 6 7 8 9 10 11 12 13 14 IS Konstantianos, D. G., and Ioannou, P. C., Analyst, 1992, 117, 877. Izquierdo, M. C . , Gutierrez, M. C., Gomez-Hens. A., and Perez- Bendito, D., Anal. Lett., 1990. 23, 487. Lianidou, E. S . , Ioannou, P. C., and Sacharidou, E.. Anal. Chin?. Acta, 1994, 290, 159. Munoz. de la Pena, A., Salinas, F., and Duran-Meras, I., Anal. Chem., 1988, 60, 2493. Vo-Dinh, T., Anal. Chem., 1978, 50, 396. Rubio, S . , Gomez-Hens, A., and Valcarcel, M.. Talantu. 1986, 33, 633. Calvo, M. V.: Lanao, J. M., and Dominguez-Gil, A., Int. J . Phai-m., 1987, 38, 117. Wan, S. H., and Matin, S. B., J . Chroniutop.., 1979, 170, 473. Wanwimolruk, S., J . Liq. Chromutogi-., 1990, 13, 161 1. Mahrous, M. S., Abdel-khalek, M. M., and Abdel-hamid, M. E., J . Assoc. Off. Anal. Chem., 1985, 68, 535. Larsen, N. E., and Marinelli, K., J . Chromutogr., 1981, 222, 482. Anttila. M., J . Phurm. Sci., 1977, 66, 433. Lianidou, E. S.. Ioannou, P. C., Polydorou, C. K., and Efstathiou, C. E., Atial. Chim. Acta, 1996, 320, 107. Miller, C., and Miller, J. N., Statistics jui. Analytical Chrmistry, Ellis Horwood, Chichester, 2nd edn.. 1988. Amick, E. N., and Mason, W. D., Anal. Lett., 1979, 12. 629. Paper 6100600K Received January 25, 1996 Accepted March 20, I996 Average of three measurements.

ISSN:0003-2654    

DOI:10.1039/AN9962100909

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, J u l ~ 1996, Vol. 121 (913-922) 913 Observation of Albumin Resonances in Proton Nuclear Magnetic Resonance Spectra of Human Blood Plasma: N-Terminal Assignments Aided by Use of Modified Recombinant Albumin Roy Harris", Sunil U. Patelb, Peter J. Sadlerbl" and John H. Vilesb a Delta Biotecbhnology Limited, Castle Court, 59 Castle Boulevard, Nottinglzam, UK NG7 IFD h Department of Chemistry, Bir-kheck College, University of London, Gordon House, 29 Gordon Square, London, UK WCIH OPP Two-dimensional total shift correlation spectroscopy (TOCSY) and double-quantum-filtered phase-sensitive homonuclear shift-correlated spectroscopy (DQF-COSY) 'H NMR spectra are used to assign peaks for about one sixth of the amino acids residues of isolated human serum albumin (67 kDa) to amino acid types.Sequential assignments are presented for 'H NMR resonances of the N-terminal residues Aspl, Ala2 and His3 of human serum albumin (HSA). These are based on pH-dependent chemical shifts reflecting the titrating N-terminal NH2 and the His3 imidazole ring, in addition to DQF-COSY and TOCSY experiments. Studies of variant recombinant human albumin with Aspl deleted, rHA(2-585), aided the assignments. The structural nature of the N-and C-termini of HSA are discussed and pK, values of 7.9 and 6.3 were determined for the N-terminal amino group and His3 imidazole ring, respectively. About 20 spin systems for albumin, including those for the N-terminal amino acids, were assigned in 1H NMR spectra of blood plasma by comparison with isolated albumin. Resonances for lipids within lipoproteins and also several low molecular mass components can also be assigned in 2D TOCSY IH NMR spectra of plasma.Keywords: Blood plmmu; humun serum alhunzin; ' H NMR; N-terminus Introduction Blood plasma is a heterogeneous mixture of lipoprotein particles, plasma proteins and low molecular mass components such as amino acids, metal ions and anions.' 1H NMR studies of plasma have been largely confined to 1D spectra and most previous studies have used Hahn or Carr-Purcell-Meiboom- Gill (spin lock) (CPMG) spin-echo sequences to filter out broad resonances associated with plasma proteins, such as albumin, which have usually been considered to be broad and un- resolvable.z--5 Resonances for approximately SO low-M, mass components of blood plasma have been assigned in lH NMR ~ p e c t r a , ~ .~ and also peaks for lipids within lipoproteins*-1 1 and the N-acetyl groups of mobile carbohydrate side-chains of glycoproteins. 12 Human serum albumin (HSA) (67 kDa) is the most abundant protein in blood with a plasma concentration of about 0.63 mmol I- 1. It is involved in the binding, transport and delivery of a range of endogcnous small molecules and metal ions, and also ' To whom correspondence \hould be addressed. drugs and xenobiotics.13314 It is a single polypeptide chain of 585 amino acids with a triple domain structure, which is largely helical.15 The N-terminus of HSA is of particular interest as it is the primary binding site for both Cu2+ and Ni2+ in blood plasma. 16-19 Cu-HSA may represent the exchangeable portion of the Cu2+ in blood plasma, and for which transport2() and storage21 roles have been suggested, while Ni-HSA has been shown to be responsible for nickel-induced allergic re- sponses.22J3 The binding of the anti-arthritic gold drug auranofin at Cys34 is reflected in a change of environment at the N-terminus of albumin.24Js IH NMR assignments have been reported for approximately one fifth of the amino acid spin systems of bovine serum albumin (BSA).26 Specific assignments for the N-terminal residues, Asp 1, Thr2 and His3 of BSA,26 and His3 ECH of HSA have been made.27 The binding of metals to the N-terminus of albumin has recently been investigated by *H NMR spectro- scopy,16 and 1H NMR studies of Ni2+ binding to albumin in intact blood plasma have also been reported.28 The stoi- chiometry of Nix+ binding appeared to be less than 1 : 1 , suggesting that natural albumin may be partially modified or clipped at the N-terminus. Partial degradation of the N-terminus has recently been confirmed.29 In this work, we have used recombinant human albumin with both N- and C-terminal amino acid deletions to aid the assignment of the iH NMR resonances.We show that 2D techniques can be used to assign 20 individual residues for HSA in spectra of blood plasma in addition to low-M, components. The 1H NMR detection of albumin signals from plasma will provide a method for studying the binding of metal ions, pharmaceuticals, xenobiotics and endogeneous molecules to albumin in its natural environment.30 Experimental Samples Human serum albumin was purchased from Sigma (St.Louis, MO, USA) as essentially globulin free and fatty acid free. Any residual salts and unbound small molecules were removed using a PD-10 gel filtration column, purchased from Pharmacia (Uppsala, Sweden). Recombinant human albumin (rHA) was produced by secretion from Sacchar-omyces cerevisiue essen- tially as described by Sleep et ~ 1 . 3 1 DNA sequences encoding rHA variants with the N-terminal Aspl residue deleted, rHA(2- 585), or the C-terminal residue Leu585 deleted, rHA( 1-584), were constructed using synthetic oligonucleotides correspond- ing to part of the HSA cDNA (Ballance, D. J., unpublished work). The variant rHAs were expressed in S. cerevisiae as for914 Analyst, July 1996, Vol. 121 full-length rHA.The proteins were purified by ion-exchange and affinity chromatography and then characterized by electro- spray mass spectrometry, N-terminal sequence analysis and peptide mapping. All protein samples were dissolved in 0.1 moll-' phosphate buffer in D20, to give an albumin Concentration of 2 mmoll-l. The pH* (uncorrected meter reading; pD = pH* + 0.432) was measured within the NMR tube using an ultra-thin combination electrode purchased from Aldrich (Milwaukee, WI, USA) standardized with buffers at pH 4, 7 and 10, together with a Coming 145 meter. Small adjustments to pH* were made using microlitre additions of NaOH or DCI (0.1 mol 1-I). Heparinized blood was obtained from a healthy human volunteer. The plasma was separated by centrifugation at 277 K, freeze-dried and reconstituted in half the volume of 0.1 moll- deuterated phosphate buffer (pH* 7.0), giving an approximate albumin concentration of 1.2 mmol 1-1.A low molecular mass extract was obtained using a Centricon ultra filtration device (Amicon, Beverly, MA, USA) with a 5 kDa cut-off. The low molecular mass extracts were freeze-dried and redissolved in the same volume of D20, containing 0.1 mol 1-1 phosphate buffer (pH* 7.4). NMR The NMR experiments were carried out using Varian Unity VXR600 and JEOL GSX5OO spectrometers at 3 10 K. Typically for 1D 'H spectra, 256 transients were acquired using a pulse repetition time of 3 s, with a flip angle of 50" and a spectral width of 8000 Hz centred on the water peak, which was suppressed using pre-saturation.For processing, the free induction decays were zero-filled to 32 K, and resolution enhancement was performed using combined sine-bell and exponential (line broadening of 0.5 Hz) function^.^^^^^^^^ 2D 'H total shift correlation spectroscopy (TOCSY)3s spectra were acquired using the MLEV-17 sequence with a spin-lock time of 55 ms. Nuclear Overhauser effect spectroscopy (NOESY)36 spectra were obtained using various mixing times between 20 and 200 ms. Typical parameters for both TOCSY and NOESY were spectral width 8000 Hz in F [ and F2, 4096 data points in the t2 dimension with 256 increments of tl and relaxation pulse delay 1.6 s, during which the residual water signal was irradiated. Double-quantum-filtered phase-sensitive homonuclear shift-correlated spectroscopy (DQF-COSY)37 spectra were acquired with 400 increments in tl; typically the data sets were zero-filled to 4096 by 1024 points.Two- dimensional spectra obtained in H20-D20 (90 + 10) used the simulated cross-peaks under bleached alpha carbons (SCUBA)3* pulse sequence for solvent suppression. Two- dimensional spectra were processed using optimum shifted Gaussian window functions. Increasing the shift of the Gaussian window function caused broad resonances with short transverse relaxation times to be attenuated to a greater extent. All spectra were obtained at 310K and are referenced to sodium 3-(trimethylsily1)propionate-2,2,3,3-d4 (TSP) via internal diox- ane at 3.764 ppm. Calculations The program KaleidaGr~ph~~ supported by Apple Macintosh computers was used to fit the pH-chemical shift profiles to the following equation: where 6obs = observed chemical shift, 6, and a0 = limiting chemical shifts of fully protonated and deprotonated residues, respectively, n = Hill coefficient and K, = dissociation constant.Results and Discussion First, assignments of 'H NMR resonances in spectra of isolated albumin are described together with the problems associated with the *H NMR spectra of a 67 kDa protein. Second, sequence-specific assignments of the N-terminal residues of isolated albumin are discussed, and the final section describes the detection of albumin in blood plasma. Human Serum Albumin Assignments of spin systems in albumin were obtained by recording 2D TOCSY and DQF-COSY spectra of HSA at various pH* values. Fig. 1 shows the region of aCH to methyl connectivities of TOCSY and DQF-COSY spectra of HSA at pH* 6.80.Relay peaks for three Thr a y spin systems are observed in the 2D TOCSY spectra. Approximately one sixth of the total spin systems for HSA were resolved in 2D spectra. Seventy-five spin systems were assigned to an amino acid type and are labelled with the appropriate three-letter code and an arbitrary Roman numeral, as shown in Table 1. Where two similar spin systems have overlapping chemical shift ranges,40 all likely amino acid types are given, for example Val/Leu or Lys/Arg. The assignment of His resonances was aided by their characteristic titration behaviour with pH*. Two-dimensional spectra of isolated albumin in HzO-D20 (90 + 10) were also recorded, but only one strong aCH to NH through-bond connectivity for AlaVIT was clearly observed in both TOCSY and DQF-COSY spectra at pH 5.5 (1.43, 4.14, 7.72 ppm).NOESY spectra of HSA in H20-D20 (90 + 10) were also obtained using mixing times of 20, 50 and 200 ms. Extensive overlap of signals in NH to aCH and NH to PCH regions was observed, even when short mixing times were used. The use of resolution enhancement for NOESY spectra resolved many side-chain nuclear Overhauser effects (NOES), but not NOES from backbone NHs. The relaxation properties of albumin (2 rnmoll-l,310K, 600 MHz, 0.1 mol 1-1 phosphate buffer) were also investigated. Inversion-recovery experiments indicated that the bulk of the protein had a longitudinal relaxation time of approximately 0.94 s. CPMG spectra indicated a short transverse relaxation time, 15 ms, which implies a linewidth at half-height of 21 Hz for the bulk of the protein resonances.Relative to the bulk, resonances assigned to the N-terminal residues have shorter longitudinal relaxation times, approximately 0.87 s, and longer transverse relaxation times, 18 ms, giving narrower linewidths. Many histidine ECH and 6CH resonances have longitudinal relaxation times of about 1.6 s and transverse relaxation between 18 and 26 ms. Sequence-specific assignments could not be obtained using NOESY data in the standard manner.41 Hence unequivocal assignments of a particular spin system to a residue type within HSA are difficult. Ambiguities in spin system assignments may arise from (1) the occurrence of incomplete spin systems (missing cross-peaks), e.g., the aCH proton resonance from the less flexible peptide backbone is often too broad to be observed, (2) the overlap of two different spin systems, (3) the possible occurrence of degeneracy of resonances within the same spin system and (4) numerous possible AMX spin systems.The assignments to residue types given in Table 1 are based on spin- system types seen in TOCSY and DQF-COSY spectra and a knowledge of the usual range of observed shifts for a particular amino acid in proteins.40 The relatively high molecular mass of albumin means that it tumbles slowly in solution and consequently many signals arc too broad to be resolved. Apart from the N-terminal residues, no particular domain or subdomain of albumin appears to be more mobile than the bulk of the protein.The X-ray structure indicates that individual domains are connected by helices and packed close together to form a heart-shaped m ~ l e c u l e . ' ~Analyst, July 1996, Vol. 121 915 Charged residues commonly on the surface of proteins are not necessarily more easily observed in spectra of HSA than hydrophobic residues. Residues with methyl groups are favoured for NMR observation because of their additional proton intensity and resonances due to protons at the end of long side-chains, for example Lys E protons, are also more easily observed. The limited number of aCH to backbone NH connectivities reflects the inherent lack of mobility of the NH protons situated on the peptide backbone. In addition, water pre-saturation resulted in cross-saturation of NH resonances; saturation transfer causes broadening of albumin resonances and general loss in intensity.The overcrowding in the NOESY spectra of HSA may be caused by the presence of cross-peaks due to spin diffusion, a phenomenon particularly pronounced with proteins of high molecular mass.42 To reduce this problem, short mixing times of between 50 and 20 ms were employed, since the build- up of.spin diffusion is slower than that of direct NOEs.43 In addition, strong resolution-enhancing window functions were used to reduce the overlap in NOESY spectra by filtering broader resonances. However, even so, it was not possible to resolve inter-residue backbone NOES, perform a standard NOESY walk and sequentially assign portions of the protein.In future studies, problems associated with water suppression may well be eliminated by the use of pulse field gradients within pulse sequences such as Watergate,44 which do not rely on irradiation of the water signal. Some of the problems of spin diffusion could be reduced by the use of 60-70% random deuteration .45 N-Terminal Assignments of Isolated Albumin In this section, 'H NMR sequence-specific assignments for the N-terminal amino acids of isolated HSA, Aspl-Ala2-His3- Lys4,46 are dealt with in turn. Aspl The cross-peaks assigned to Aspl a(3 correlations in the TOCSY spectrum (Fig. 2) form an easily resolvable AMX spin DQF-COSY HSA '"i NAcNeu system with typical Asp chemical shifts.40347 The observation of one easily resolvable AMX spin system with relatively narrow linewidths in the 2D 1H NMR spectra of HSA suggests that this residue has a higher mobility than the bulk of the protein, typical of a terminal or relatively unstructured portion of the protein.The pH dependences of the chemical shifts of the aCH, (3CH and (3'CH resonances are shown in Fig. 3 and the derived pK, values, Hill coefficients and limiting shifts are listed in Table 2. In addition, this spin system has an associated pK, of 7.9 at 37OC, a value expected for an N-terminal amino group. The titration curve also has a low Hill coefficient, possibly reflecting cooperativity with the titrating His3 imidazole ring (pK, 6.3). The lowering of Hill coefficients due to interacting titrating groups has been discussed by Markley.48 An aspartate p- carboxylic acid group would typically have a pK, of 3.9.49 However, the titration of this group cannot be followed below pH 4.5, because of a marked structural transition of albumin at this pH, known as the Neutral-to-Fast transition.50-51 These resonances are not observed in one- or two-dimen- sional spectra of recombinant human albumin with Aspl deleted, rHA(2-585), Fig.2, confirming the assignment of Aspl resonances. Ala2 The resonances assigned to Ala2 aCH and pCH3 protons of HSA have shifts and spin system patterns typical of an alanine residue, Fig. 2. They also have relatively narrow linewidths. However, there are many other resolvable resonances assign- able to alanine residues. The pH titration curve for the pCH3 alanine protons is U- shaped, Fig. 3, and appears to arise from two overlapping titration curves, one tending to high field as the pH is raised from 5 to 7, the other tending to low field as the pH is raised from 7 to 9.The Ala2 aCH resonance was fitted to a pK, of 6.64 with a Hill coefficient of 0.7, as shown in Fig. 3 and Table 2. The titration curve for the PCH3 alanine resonance reflects its proximity to the titratable N-terminal amino group and the neighbouring imidazole of His3. This results in a U-shaped titration curve, with shifts to high field as the N-terminal amino TOCSY HSA 1 u 2.0 1.5 1.0 2.0 1.5 1 .o Fig. 1 600 MHz 2D DQF-COSY and TOCSY 'H NMR spectra of HSA in DzO, pH* 6.8 in 0.1 mol 1-* phosphate buffer, showing the cvCH to methyl connectivities. A number of the resolved Ala and Thr spin systems are labelled with arbitrary Roman numerals.The N-acetylneuraminic acid (NAcNeu) peaks arise from residual amounts of globular proteins containing glycan chains within the sample of commercial HSA.916 Analyst, July 1996, Vol. 121 group protonates (pK, 7.9) and shifts back to low field as the imidazole ring protonates (pKa 6.3). The changes in chemical shift with pH of both the aCH and PCH3 resonances are small, with only 0.03 ppm between protonated and deprotonated forms, consistent with these protons being several bonds away from the titrating group. The deletion of Asp 1 from HSA causes a dramatic shift in the resonances assignable to Alal (formally Ala2 in rHA) in the spectrum of rHA (2-585), which is expected since this becomes the N-terminal residue.The Ala PCH3 resonance shifted to low field by 0.14 ppm and the Ala2 aCH resonance shifted to higher field by 0.32 ppm, as shown in Fig. 3. A summary of the chemical shifts is given in Table 3. His3 Resolution-enhanced spectra of rHA(2-585) were obtained at a number of pH values and the chemical shifts of the His ECH resonances are compared with those of intact HSA in Fig. 4. The pH-chemical shift profiles for all resolvable His ECH reso- nances of HSA were used for this purpose. All but one of the His ECH resonances of HSA are present in the spectra of rHA(2- 585) and have similar chemical shifts at comparable pH values. The resonance present in the spectra of HSA but not in those of rHA(2-585) can be assigned to His3 ECH. In addition, one new, sharp His ECH resonance was observed for rHA(2-585) which was not present in spectra of full-length HSA and is assigned to His2 (previously His3 in full-length albumin). A comparison of the aromatic regions of 1D 1H NMR spectra of HSA and rHA(2-585) is shown in Fig.4. As for Ala2, the resonances assigned to the imidazole ring protons of hi53 are strongly perturbed in the spectra of rHA(2- 585) compared with HSA owing to the deletion of Aspl and the formation of an N-terminal alanine. Between pH 6 and 7 the His2 ECH peak of rHA(2-585) is consistently to lower field compared with that of His3 ECH of intact HSA. The closer proximity of the His residue to the positive charge of this N- terminal amino group would be expected to lower the pK, of the imidazole ring48 and explain the observed change in shift.All but one of the NMR-observable His ECH resonances of rHA(2-585) have similar shifts to those of intact HSA over a range of pH* values. This suggests that the folding of HSA is very similar to that of rHA. Indeed, Carter and He15 have reported no significant difference between the X-ray crystal structures of HSA and rHA. The good alignment of the resonances suggests that the removal of Aspl causes little conformational change in the protein. The resonance-labelled His-XI11 in the spectra of rHA(2-585), Fig. 4, has a slightly different chemical shift to the analogous peak for full-length HSA. This could be because it is close in space to the N- Table 1 Spin system assignments for HSA in D20, pH* 6.80 in 0.1 mol 1-I phosphate, 37 "C * The spin systems in bold with an asterisk are also observed in spectra of blood plasma. Residue Thr I Thr 11* Thr I11 Ala I Ala2* Ala 111 Ala IV* Ala V Ala VI* Ala VII Ala VIII* Ala IX Ala X Ala XI Ala XI1 Ala XI11 Ala XIV Ala XV* Ala XVI Ala XVII* Ala XVIII* Ala XIX Ala XX* Ala XXI Ala XXII Ala XXIII Val I* Val I1 (Leu) Val 111 (Leu) Val IV (Leu) Val/(Leu) V Val/(Leu) VI Leu I* (Val) Leu I1 (Val) Leu I11 (Val) Leu IV (Val) Lys I1 (Arg) Lys I11 (Arg) Lys IV (Arg) Lys4* a 3.92 4.26 4.48 4.03 4.28 3.88 4.02 4.28 4.21 3.73 4.12 4.22 4.29 4.18 3.83 4.19 4.36 4.17 4.48 4.16 4.20 4.3 1 4.46 4.85 4.63 4.61 4.13 - - - - - - - - - - - - - B 4.3 1 4.04 4.06 1.27 1.30 1.29 1.32 1.34 1.40 1.44 1.44 1.46 1.47 1.49 1.51 1.52 1.51 1.56 1.55 1.56 1.63 1.62 1.73 1.42 1.46 1.53 2.02 2.06 2.25 1.85 1.75 1.76 - - - - - - - - Y 6 E 1.23 1.08 1.19 0.93, 0.91 1.02, 1.03 1.05, 1.16 0.90, 0.97 0.64 1.02 1.49 0.81, 0.85 1.57 0.91, 0.84 1.58 1.01, 1.02 1.51 0.96 1.37, 1.44 1.65 2.96 I .43 1.73 3.04 1.54 1.75 3.04 1 S O 1.57 2.89 Residue Arg I* (Lys) Aspl* Glu/Gln I Glu/Gln I1 Gly I Gly I1 Ser I Ser I1 LYS v (Arg) Phe I Phe I1 Phe I11 (Tyr*2) Tyr I (Phe)* Tyr I1 (Phe)* Tyr I11 (Phe)* Tyr IV (Phe) Tyr V (Phe) Tyr VI (Phe) Tyr VII (Phe) Tyr VIII (Phe) Tyr IX (Phe) Tyr X (Phe) Tyr XI (Phe) Tyr XI1 (Phe) His I* His II* His III* His IV* His 3* His VI His VII His VIII His IX His X His XI11 His XX (x B Y 5 & - - - 1.86 3.10 - 1.76 1.64 3.21 4.14 2.75, 2.62 - 2.03, 2.1 1 2.34 - 2.06, 2.12 2.40 3.89, 3.67 4.10, 3.81 4.31 4.03 3.91 3.59 33-H 4-H 2,6-H 7.16 7.35 6.93 7.12 7.30 6.93 7.74 7.5 1 7.42 Aromatic ring 5.92, 7.03 6.32, 6.71 6.20, 6.39 6.47, 6.62 6.53, 6.98 6.73, 6.84 6.80, 7.10 6.85, 7.05 6.87, 7.22 6.97, 7.1 1 7.17, 7.34 7.26, 7.44 ECH 6CH 8.34 7.22 8.18 7.22 8.13 7.20 7.98 6.93 7.90 7.84 - 8.74 - 8.26 - 8.37 - 8.27 - 7.81 - 7.93 -Analyst, July 1996, Vol.121 917 terminus; from published views of the crystal structure of HSA,l5 His67 and His9 are possibilities. Like the Aspl resonances, the pH profile of His3 ECH for HSA gives a low Hill coefficient of 0.7, which possibly reflects the titrating N-terminal amino group. The His3 ECH proton has an associated pK, of 6.3. A solvent-accessible histidine residue in a peptide, not affected by neighbouring charged or titrating groups, would be expected to have a pK, of 6.9.48 Lower pK, values can be caused by the close proximity of positively- charged The broadness of the His3 ECH resonance compared with other histidine resonances of HSA is surprising.In the ‘H NMR spectra of BSA, His3 ECH gives rise to two peaks throughout much of the pH titration. This has been associated with the in this case the N-terminal amino group. ‘) 4 3.5 i 3.5 4.0 4.5 . . . - 1.5 1 .o 3.2 2.2 2.0 (b) 6 3.0 3.5 4.0 4.0 4.5 , , , , . . . 2.2 2.0 1.5 3.2 2.2 2.0 1.5 1 .O 3.2 6 6 Fig. 2 600 MHz 2D TOCSY spectra (MLEV-17, spin-lock of 55 ms) of ( a ) recombinant human albumin rHA(2-585) (2 mmol l - l ) , (h) and (c) human blood plasma in 0.1 mol I-’ phosphate buffer (pH* 6.8), showing the Aspl a(3 and Ala ap regions. The spectra of HSA but not rHA(2-585), contain an AMX spin system assignable to Aspl.The spectrum of rHA(2-585) shows the absence of a cross-peak in the spectra of HSA and plasma. However, a new, sharp cross-peak at 1.44/3.96 ppm (now Alal) is observed. I I HSA (2 mmol I-’) and blood plasma, for Ala2, observed918 Analyst, July 1996, Vol. 121 1.28 heterogeneity of Cys34 (free SH or blocked), resulting in His3 ECH experiencing two different magnetic environments.24.25 The His3 ECH resonance of HSA is also split, but this splitting is less pronounced than in the spectra of BSA. 0 I 1 I I i I I Lys4 There is a number of cross-peaks in the 2D TOCSY spectra assignable to Lys ~ / 6 or Arg 6/y resonances. In particular, one set is assigned to Lys4, based on Ni2+-binding experiments, ring current shifts and modelling studies.16 As with Aspl and Ala2 resonances, the resonances for the side-chain of Lys4 are relatively narrow and the 6 / y , 8/y', E/Y and E/Y' resonances are also observed in 2D TOCSY spectra.The pH titration curve of the Lys4 E C H ~ resonance shows a gradual small shift to high field with increasing pH*, as shown in Fig. 3. The removal of Aspl has no significant effect on the chemical shifts of the ECH~, 6CH2 or yCH2 resonances of Lys4, as shown in Table 3, perhaps because these protons are too many bonds away from Aspl to be significantly affected. Leu585 Two-dimensional spectra for HSA and for rHA with the C- terminal residue Leu585 deleted were also compared (data not shown). Ten Leu/Val spin systems are resolved in the 2D 3 .6 1 , 4 5 6 7 8 9 1 0 1 1 2.0 ::: 2.4 2m51 $Asp1 PCH 2.3- 4 5 6 7 8 9 1 0 1 1 PH* TOCSYDQF-COSY spectra of HSA (Table 1) and similar cross-peaks are observed in the spectra of rHA( 1-584). All the clearly resolvable Leu spin systems observed for HSA were also observed in the spectra of rHA( 1-584). It was hoped that the comparison between rHA( 1-584) and HSA would assist in the assignment of the C-terminal resonances. However, no reso- nances typical of a Leu spin system (Leu585) present in the 1H NMR spectra of intact HSA but absent in spectra of rHA( 1-584) were observed. This suggests that the C-terminus has structural rigidity in solution and that 1H NMR resonances for Leu585 are broad and unresolved. Indeed, the X-ray crystal structure of HSA15 indicates that the C-terminus is helical right up to the final residue. The N-terminus Tentative 1H NMR assignments of the first three amino acids of BSA (Asp l-Thr2-His3) have been reported previously.26 These were based largely on comparisons with other albumins containing sequence mutations (rat, porcine and human).It was assumed that particular differences in resonances were due to mutations in the N-terminus of the albumins (e.g., Thr2fAla2 for BSA/HSA and Aspl/Glul for BSARSA) and not the result of variations in sequence elsewhere in the protein. Assignments made in the present work are based on the removal of only Aspl and are therefore less ambiguous. Chemical shift-pH profiles 4.28 4.27 4.26 4 5 6 7 8 9 1011 Ala2 PCH, 0 0 1.3 0 0 I OOO 0 7 . 4 7 , 4 5 6 7 a 9 1 0 1 1 2-g1 2.9 0 0 0 2 .8 8 " 1 , 4 5 6 7 8 9 1 0 1 1 PH* Fig. 3 pH dependence of the chemical shifts for Aspl aCH, BCH and B'CH, Ala2 aCH and BCH3, His3 ECH and Lys4 E C H ~ resonances of HSA in 0.1 mol 1-1 phosphate buffer at 37 "C. The curves are computed best fits for the pK, values listed in Table 2. Table 2 pK, values, Hill coefficients ( n ) and chemical shifts of Aspl, Ala2 and His3 'H NMR resonances of HSA. 6 , is the shift of the fully protonated form, 6,, that of the fully deprotonated form. A standard error is given (fo-, 68% confidence limits) Resonance pK, n a+ (ppm) 60 (PPd Asp1 aCH 7.89 f 0.03 0.71 k 0.03 4.209 f 0.004 3.669 & 0.008 Asp1 P'CH 7.87 f 0.10 0.52 f 0.05 2.788 f 0.006 2.531 f 0.011 Asp1 PCH 7.81 f 0.08 0.56 f 0.06 2.675 f 0.007 2.340 f 0.013 Ala2 aCH 6.64 f 0.16 0.74 k 0.1 1 4.297 f 0.002 4.263 f 0.001 His3 ECH 6.34 f 0.02 0.76 f 0.02 8.598 f 0.014 7.573 f 0.006Analyst, July 1996, Vol.121 919 for His ECH resonances of HSAS2 are not the same as those of BSA,Sl even though 14 of the 16 His residues are conserved in BSA. In spite of over 80% sequence homology, there is little correlation between the chemical shifts of individual spin systems resolved in BSA26 and those of HSA given in Table 1. Despite this lack of correlation for the majority of the resonances, the pK, values, Hill coefficients and pH-chemical shift profiles of the Aspl and His3 resonances reported for BSA26 are almost identical with those of HSA. A lH NMR assignment for His3 ECH HSA has also been proposed by Bos et ~ 1 .~ 7 although no assignments for Aspl or Ala2 were made. Their assignment for His3 assumed that paramagnetic Cu2+ broadening is specific for His3 resonances, and agrees with that presented here. It is interesting to compare, in Table 3, the shifts of the N- terminal resonances of HSA with that of the model peptide Asp-Ala-His-N-methylamide (DAH-nma) described by Laus- sac and Sarkar,s3 at a comparable pH. The resonances associated with Aspl, Ala2 and His3 in HSA are consistently shifted to higher field compared with the unstructured model peptide, as are the Lys4 resonances compared to random coil shifts47 This suggests that the magnetic environment of the N- terminus is perturbed by the bulk of the protein. The crystal structure of HSA indicates that the protein does not form a helix before residue Lys4.15 Furthermore, the N-terminal residues Aspl , Ala2 and His3 were not observed in the X-ray crystal structure,lS presumably because of the inherent mobility of the N-terminus.This agrees well with the observation of relatively sharp NMR signals for the N-terminal residues. Blood Plasma This final section describes the detection of albumin in intact blood plasma by 'H NMR spectroscopy and, in particular, the resonances for N-terminal amino acids are assigned. Other components of blood plasma are also discussed. Prior to resolution enhancement, the 1D lH NMR spectrum of blood plasma is dominated by signals from triglycerides and phospholipids within lipoproteins and plasma proteins (Fig. 5). After the use of resolution enhancement, the broader resonances are filtered out, so that resonances with narrower linewidths can be observed.In the aliphatic region of the 1H NMR spectra of plasma, resonances for the low molecular mass components are observed and in the aromatic region the ECH resonances of histidine residues associated with HSA and peaks for free amino acids, r,-Phe, L-Tyr and L-His, are observed. For comparison, the spectra of isolated HSA and a low molecular mass ( < 5 kDa) extract of plasma are shown in Fig. 5. Two-dimensional TOCSY and DQF-COSY spectra of plasma, processed using unshifted Gaussian functions, are also dominated by lipid signals, and even without the use of Table 3 IH NMR chemical shifts (in ppm) of the N-terminal residues of HSA (pH* 6.70), rHA(2-585) (pH* 6.7.5) and rHA(1-584) (pH* 6.70) and, for comparison, the model peptide Asp-Ala-His-N-methylamide (DAH- nma) at pH* 7.0,53 together with random coil shifts for Lys47 Resonance HSA rHA(2-585) rHA( 1-584) DAH-nma Aspl PCH Aspl P'CH Aspl aCH Ala2 BCH3 Ah2 aCH His3 ECH LYS-4 E C H ~ LYS-4 6CH2 LYS-4 yCH2 2.75 - 2.61 - 4.13 - I .30 1.44 4.28 3.96 7.91 7.79 2.93 2.93 1.65 1.65 1.36, 1.42 1.36 * Random coil shifts for Lys.2.75 2.61 4.13 1.29 4.30 7.86 2.93 1.65 1.36, 1.42 2.77 4.17 I .34 4.34 8.01 [3.02*] [ 1.7*] [ 1.45*] resolution enhancement functions some of the broad plasma protein signals are lost owing to transverse relaxation during delays in the pulse sequences. Sugar resonances with narrow linewidths dominate the region between 3.5 and 4.5 ppm and cause ridges in the f, dimension of the spectra.The use of shifted Gaussian functions to process both dimensions resulted in the further filtering out of broad unresolved features arising from, in particular, many AMX systems, methyl to methylene correlations and Lys E C H ~ correlations, of plasma proteins. A range of TOCSY spin-lock times from 30 to 65 ms was investigated. For the majority of scalar couplings, a spin-lock time of between 50 and 60 ms resulted in the most intense cross- peaks. DQF-COSY spectra tended to be less resolved, partic- ularly for signals close to the diagonal and for lipids signals. The spin-lock sequence in TOCSY eperiments has a filtering effect on broad signals with short transverse relaxation times, similar to that observed in ID CPMG experiments.54 In CPMG experiments, half the bulk of the signals from albumin is lost in <20 ms.In spite of this, the strongest cross-peaks for the majority of resolvable resonances of albumin are observable l T " T l r " i I T , , , I I I , I r-1 9 .o 8.5 8.0 7.5 9.0 8.5 8.0 7.5 6 rHA(2-585) 3 His:f.l I C His'Y I I 0 H l s t N i l l HSA h E, Q v 00 8.5 1 4 5 6 7 8 9 PH* Fig. 4 (a) The aromatic region of 500 MHz 1H NMR spectra of HSA and rHA(2-585), pH* 6.7 in 0.1 moll-I phosphate buffer, at 37 "C. The Roman numerals indicate His ECH resonances for which pH-chemical shift profiles are presented in (b). The His3 resonance is shifted to high field on the deletion of Aspl. (b) Graphs of pH* versus chemical shift for His ECH 'H NMR resonances of HSA and rHA(2-585).Continuous curves and dots represent data for intact HSA; data for rHA(2-585) are shown by various symbols at four pH* values: 6.33, 6.41, 6.74 and 7.88. Corresponding resonances for HSA in spectra for rHA(2-585) are observed apart from His3. A resonance observed for rHA(2-585) but not for HSA is shown as a filled circle and is assigned to His2.920 Analyst, July 1996, Vol. 12 I with spin-lock times of 50-60 ms. In DQF-COSY spectra, broad signals are also filtered out during the evolution period. Two-dimensional TOCSY spectra have certain advantages over DQF-COSY spectra, particularly for albumin and other high molecular mass components of plasma. DQF-COSY and TOCSY spectra both have the advantage over absolute value COSY or COSY -45 spectra of giving non-dispersive signals,ss but in TOCSY spectra all signals are upright and so do not suffer from phase cancellation.Cancellation is a particular problem in the spectra of large proteins as the linewidths often greatly 4 7.'S 6 6 4 HOD I 1 ' " ' 1 " ' ' 1 " ' ' 1 " " 1 " " 1 " " 1 " " " " " " 0 6 8 6 4 2 1 ' ~ ~ ' 1 ' " 1 ' ' " I ~ " ~ I ' ~ ~ I ' ' ~ ' I ' ' ' ~ I ' ' " I ' ' 8 6 4 2 0 6 Fig. 5 1D 'H NMR spectra of ((0 isolated HSA, (h) plasma, before and after the use of resolution enhancing, sine-bell/exponentiaI window functions and (c) low molecular mass (< 5 kDa) extract of plasma. Cit, citrate; Diox, added dioxane; For, formate; Glc, glucose; and Lac, lactate. exceed the J couplings. The crowded region close to the diagonal is particularly complicated in DQF-COSY spectra, as are the cross-peaks for fatty acids which result from the superposition of many signals with slightly different chemical shifts.Albumin in blood plasma When the TOCSY and DQF-COSY spectra of intact blood plasma are compared with those of isolated albumin, it is clear that intense spin systems observed for isolated HSA are also observed for blood plasma (Fig. 6). Spin systems exhibiting identical chemical shifts (at the same pH) are indicated with an asterisk in Table 1. The intensities of cross-peaks assigned to HSA in the spectra of plasma are weaker than the comparable peaks for isolated albumin, when spectra are acquired and processed under similar conditions, but 20 spin systems can be resolved. Attempts to resolve albumin signals in the spectra of blood plasma present a number of difficulties.In addition to the broadness of albumin resonances, there is interference from other components present at much higher concentrations, in particular sugars and triglycerides. These give rise to pro- nounced ridges in thefl dimension. The future use of pulsed field gradients56 to acquire the 2D data is likely to reduce t l noise. It is clear that resonances for the first two N-terminal residues of albumin, Asp1 ap and Ala2 ap, are also resolved in the spectra of intact blood plasma. Cross-peaks are present with similar linewidths, intensities and chemical shifts identical with those of isolated albumin, at a comparable pH (Fig. 2). The possibility that these resonances are from other components of blood plasma can be ruled out.It is unlikely that these relatively intense signals come from plasma proteins other than albumin, as these are considerably less concentrated in blood plasma. The most abundant plasma proteins are albumin (67 kDa) approx- imately 630 pmol 1-1, immunoglobulin g (150 kDa) 100 pmol 1-1, al-antitrypsin (51 kDa) 36 pmol I-' and transferrin (80 kDa) 36 pmol l-1.1 Free L-amino acids can also be ruled out as their chemical shifts are well doc~mented.~.~ The assignment of peaks for ECH and 6CH of His3 of albumin in the spectra of plasma by direct comparison with isolated albumin is difficult as their chemical shifts are highly sensitive to pH around pH 7. Reliable assignments can be made only by a comparison of the pH+hemical shift profiles for all observed His ECH resonances observed in plasma and isolated HSA.Such studies indicate that many His ECH resonances observed for isolated albumin have matching pH-chemical shifts profiles for blood plasma and the details of these titrations have recently been reported.30 Lys 86 correlations result in a large group of overlapping cross-peaks centred around 3.0 and 1.6 ppm. This region of the spectrum of plasma is very similar in appearance to that of isolated albumin. The most intense Lys spin system observed is shifted to higher field than the bulk of the Lys correlations and is assigned to Lys4 (Table 3). A similar high-field-shifted Lys spin system is observed in the spectrum of blood plasma. Overlapping cross-peaks for Lys ECH correlations have also been observed in 1HJT HMQC spectra of blood plasma.h Low molecular mass components of plasnza Figs.5 and 6 show that peaks for a number of low molecular mass components observed in ID and 2D spectra of plasma are also observed in the low molecular mass extracts ( < 5 kDa) of blood plasma. The chemical shifts of around 50 low molecular mass components of plasma have already been determined from spin-echo and J-resolved spectra, in which they are readily observable owing to their high mobility and relatively highAnalyst, July 1996, Vol. I21 92 1 I Lpprpmp- 3.0 2.0 1 .O 0.0 [ I 1 .o 2 .o 3.0 4.0 5.0 I Fig. 6 (u) 600 MHz 2D TOCSY (MLEV- 17. spin-lock 50 ms) spectra of the aliphatic region of isolated HSA and (h) blood plasma. The spectrum of isolated albumin is plotted with a higher threshold so that the intensities of cross-peaks are comparable to those of albumin peaks from plasma.For expansions of the methyl to alpha region, see Figs. 1 and 2. (c) The same 2D TOCSY spectrum of intact blood plasma plotted with a very high threshold, showing lipid spin systems from lipoprotein particles. Under these plotting conditions, peaks for other components of plasma are not observed because of their weak intensities relative to the lipid signals. Cross-peaks for albumin in plasma are indicated by three letters and a Roman numeral, and free I*-amino acids are indicated by conventional single-letter codes. Lac, lactate: Chol (CH2N to POCH2), choline; Gn, glycan chains; HB, hydroxybutyrate; p-glc, (J-glucose; V, valine; L. leucine; A, alanine; 7'.threonine; 1. isoleucine; Q, glutamine; H, histidine; N, asparagine; F, phenylalanine; TG, signals from triglyceride backbone of lipid molecules.922 Analyst, July 1996, Vol. 121 concentrations.6.7 These include free amino acids, acetate, acetone, acetoacetate, acetylcarnitine, betaine, carnitine, choline, citrate, citrulline, creatine, creatinine, dimethylamine, ergothioneine, formate, fructose, a-glucose, 13-glucose, 3 - ~ - hydroxybutyrate, p-hydroxyphenyllactate, isobutyrate, lactate, methylamine, 1-methylhistidine, 3-methylhistidine, myo-in- ositol, 2-oxoglutarate, pyruvate, succinate, taurine, trimethyl- amine and trimethylamine oxide (TMAO). Two-dimensional TOCSY experiments provide reliable assignments as con- nectivities between resonances are observed, allowing identi- fication of complete spin systems, rather than just the observation of individual resonances.Lipoprotein particles Lipid signals from lipoproteins dominate the 1D and 2D spectra of plasma owing to the high concentration of lipids. TOCSY spectra correlate different parts of fatty acid esters of phospho- lipids and triglycerides via their direct and relay peaks, as shown in Fig. 6(b). Resonances for the lipid signals from VLDL, LDL and HDL have previously been assigned in 1D and Hahn spin- echo 'H NMR spectra.8 Both the CH2 and CH3 peaks shift slightly to high field as the density of the particle increases. Various approaches have been used to simulate the CH2 and CH3 regions of 1D spectra of blood plasma to determine lipoprotein concentrations.+-' 1 Conclusions In spite of the size of human albumin, 67 kDa, it has been possible to obtain sequence-specific 1H NMR assignments for the resonances of the N-terminal amino acids without the use of NOE data.The application of 2D 'H NMR methods applied to the study of blood plasma, in particular 2D TOCSY, has allowed the detection of cross-peaks associated with albumin. The assignment of resonances for the N-terminal residues, their associated pK, values and chemical shift-pH profiles, provide the basis for probing metal binding sites of isolated albumin and albumin in intact blood plasma, and also studies of carbamyla- tion and degradation of the N-terminus. The 1H NMR detection of albumin signals in the spectra of plasma will also allow studies of drug binding to albumin in its natural biological medium.We thank the EPSRC and MRC for support and the MRC Biomedical NMR Centre, Mill Hill, and ULIRS for the provision of NMR facilities. We also thank Dr. J. K. Lodge (Birkbeck College) for helpful discussions. References 1 2 3 4 5 6 7 8 9 10 11 12 Geigy Scient$ic Tables, ed. Leutner, C., Ciba-Geigy, Basle, 1984, vol. 3. Bock, J. L., Clin. Chem., 1982, 28, 1873. Nicholson, J. K., Buckingham, M. J., and Sadler, P. J., Biochem. J., 1983,211, 605. Rabenstein, D. L., Millis, K. K., and Strduss, E. J., Anal. Chem., 1988, 60, 1380. Nicholson, J. K., and Wilson, I. D., Nucl. Magn. Reson. Spectrosc., 1989, 21, 449. Nicholson, J. K. Foxall, P. J. D., Spraul, M., Farrant, R. D., and Lindon, J. C., Anal. Chem., 1995, 67, 793.Bell, J. D., and Sadler, P. J., Encyclopedia of Nuclear Magnetic Resonance, eds. Grant, D. M., and Harris, R. K., Wiley, New York, in the press. Bell, J. D., Sadler, P. J., Macleod, A. F., Turner, P. R., and La Ville, A., FEBS Lett., 1987, 219, 239. Otvos, J. D., Jeyarajah, L. W., and Bennett, D. W., Clin. Chem. (Winston-Salem, N.C.), 1991, 37, 377. Otvos, J., D., Jeyarajah, L. W., Bennett, D. W., and Krauss, R. M., Clin. Chem. (Winston-Salem, N.C.), 1992, 38, 1632. Ala-Korpela, M., Hiltunen, Y., and Jokisaari, J., NMR Biomed., 1993, 6, 225. Bell, J. D., Brown, J. C. C., Nicholson, J. K., and Sadler, P. J., FEBS Lett., 1987, 215, 3 11. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 Peters, T., Jr., Adv.Protein Chem., 1985, 37, 161. Carter, D. C., and Ho, J. X., Adv. Protein Chem., 1994, 45, 153. Carter, D. C., and He, X.-M., Nuture (London), 1992, 358, 209. Sadler, P. J., Tucker, A., and Viles, J. H., Eur. J . Biochem., 1994,220, 193. Laussac, J., and Sarkar, B., Biochemistry, 1984, 23, 2832. Camerman, N., Camerman, A., and Sarkar, B., Can. J . Chem., 1976, 54, 1309. Glennon, J. D., and Sarkar, B.. Biochem. J., 1982, 203, 15. Lau, S., and Sarkar, B., J . Biol. Chem., 1971, 246, 5938. Laurie, S. H., and Pratt, D. E., Biochem. Biophys. Res. Commun., 1986, 135, 1064. Dolovich, J., Evans, S. L., and Nieboer, E., Br. J . I d . Med., 1984,41, 51. Nieboer, E., Evans, S. L., and Dolovich, J., Br. J . Ind. Med., 1984,41, 56. Ni Dhubhghaill, 0.M., Sadler, P. J., and Tucker, A., J . Am. Chem. Soc., 1992, 114, 1118. Christodoulou, J., Sadler, P. J., and Tucker, A., Eur. J . Biochem., 1994,225,363. Sadler, P. J., and Tucker, A., Eur. J. Biochem., 1992, 205, 631. Bos, J. M., Labro, J. F. A., Fischer, M. J. E., Wilting, J., and Janssen, L. H. M., J . Bid. Chem., 1989, 264, 953. Patel, S. U., Sadler, P. J., Tucker, A,, and Viles, J. H., .I. Am. Chem. Soc., 1993, 115, 9285. Chan, B., Dodsworth, N., Woodrow, J., Tucker, A., and Harris, R., Eur. J . Biochem., 1995,227, 524. Christodoulou, J., Sadler, P. J., and Tucker, A,, FEBS Lett.. 1995, 376, 1. Sleep, D., Belfield, G. P., Ballance, D. J., Steven J., Jones, S., Evans, L. R., Moir, P. D., and Goodey, A. R., BiolTechnology, 1991, 9, 183. Glasoe, P. K., and Long, F. A., J . Phys. Chem., 1960, 64, 188. Lindon, J., and Ferrige, A. G., Prog. Nucl. Magn. Reson. Spectrosc., 1980, 14, 27. Bell, J. D., Kubal, G., Radulovic, S., Sadler, P. J.. and Tucker, A., Analyst, 1993, 188, 241. Braunschweiler, L., and Emst, R. R., J . Magn. Reson., 1983, 53, 521. Kumar, A., Ernst, R. R., and Wiithrich, K., Biochem. Biophys. Res. Commun., 1980, 95, 1. Aue, W. P., Bartholdi, E., and Emst, R. R., J . Chem. Phys., 1976,64, 2229. Mueller, L., Schiksnis, R. A., and Opella, S. J., J . Magn. Reson., 1986, 66, 379. KaleidaGruph, Synergy Software, Reading, PA USA, 1991. Gross, K. H., and Kalbitzer, H. B., J . Magn. Reson., 1988, 79, 87. Wiithrich, K., NMR of Proteins and Nucleic Acids, Wiley, New York, 1986. Neuhaus, D., and Williamson, M. P., The NOE in Structural and Conformational Analysis, VCH, New York, 1989. Bauer, C. J., Frenkiel, T. A., and Lane, A. N., .I. Magn. Reson., 1990. 87, 144. Piotto, M., Saudek, V., and Sklenar, V., J . Biomol. NMR, 1992, 2, 661. LeMaster, D. E., Q. Rev. Biophys., 1990, 2, 133. Dugiaczyk, A., Law, S. W., and Dennison, 0. E., Proc. Natl. Acad. Sci. USA, 1982, 79, 7 1. Bundi, A., and Wuthrich, K., Biopolymers, 1979, 18, 285. Markley, J. L., Acc. Chem. Res., 1975, 8, 70. Critical Stability Constunts, ed. Martell, A. E., and Smith, R. M., Plenum Press, New York, 1982, vol. 5. Foster, J. F., in Albumin Structure Function and Use, ed. Rosenoer, V. M., Oratz, M., and Rothschild, M. A., Pergamon Press, New York, Sadler, P. J., and Tucker, A., Eur. J . Biochem., 1993, 212, 81 1. Viles, J. H., PhD Thesis, University of London, 1994. Laussac, J., and Sarker, B., Can. .J. Chern., 1980, 58, 2055. Meiboom, S., and Gill, D., Rev. Sci. Instrum., 1958, 29, 688. Derome, A. E., Modern NMR Techniques jbr Chemistiy Research, Pergamon Press, Oxford, 1987. Davis, A. L., Laue, E. D., Keeler, J., Moskau, D., and Lohman, J., J . Magn. Reson., 1991, 94, 637. Paper 6100705H Received January 30, 1996 Accepted April 4, 1996 , 1977, pp. 53-84.

ISSN:0003-2654    

DOI:10.1039/AN9962100913

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, July 1996, Vol. 121 (923-928) 923 Vapour Generation-Fourier Transform Infrared Direct Determination of Ethanol in Alcoholic Beverages Amparo Perez-Ponce, Salvador Garrigues and Miguel de la Guardia* Department of Analytical Chemistry, University of Valencia, 50 Dr. Moliner Street, 461 00 Burjasot, Valencia, Spain A procedure is proposed for the direct determination of ethanol in alcoholic beverages. The method is based on the injection of small volumes of untreated samples into a heated Pyrex glass reactor in which, at a temperature between 80 and 100 "C, the ethanol is volatilized and introduced by means of a nitrogen carrier flow into a gas cell of an FTIR spectrometer. The measurement of the area of the flow injection recording, obtained from the absorbance of the transient signals, in the wavenumber range between 1150 and 950 cm-I, allows the direct quantification of ethanol without water background problems and free from interferences from sugars, providing a limit of detection of 0.02% v/ v and typical RSDs between 0.11 and 0.5% for five analyses of the same sample containing between 5 and 30% v/v ethanol.The sampling frequency of the method is 51 h-1 and accurate results were obtained for different types of alcoholic beverages, from low-alcohol beers to wines and spirits. Keywords: Ethanol determination; alcoholic beverage analysis; vapour generation; Fourier transform infrared spectrometry; flow injection Introduction FTIR spectrometry is not only a fast analytical technique that provides very interesting qualitative information about the composition of samples and chromatographic effluents,l-3 but it can also, as has been clearly established in recent years,&7 be useful for accurate quantitative analysis, avoiding the traditional drawbacks found with the different sample preparation tech- niques commonly employed for IR analysis.However, some limitations remain in the general applicability of IR measure- ments. The analysis of dissolved samples requires the use of cells of known thickness and previous selection of a suitable solvent. Thus, limitations in the transparency of solvent and window materials often preclude the application of IR methods for quantitative determinations,g especially, in the analysis of aqueous solutions owing to the strong absorbance of water and its incompatibility with several of the window materials employed in IR spectrometers. For the analysis of aqueous samples by FTIR, it is necessary to use adequate transmittance or reflectance cells equipped with water-resistant crystals. For transmittance measurements, it involves the use of a very small pathlength (of the order of 25 pm) owing to the water absorbance.On the other hand, the use of attenuated total reflectance (ATR) cells involves poor sensitivity and so, both measurement techniques transmittance and ATR, are not very appropriate for determining low concentrations of analytes. * To whom correspondence should be addressed. In contrast FTIR spectrometry is very useful for carrying out determinations in gaseous or vapour samples owing to the high transparency of gases, the low background values and the possibilities offered by the use of multiple-pass cells, which can provide good sensitivity.'? l o Our research group has developed a simple system for the generation of vapour phases, which can be adapted to any FTIR spectrometer equipped with a gas cell, and permits us to obtain FTIR transient signals in the gaseous state from liquid samples injected directly into a heated Pyrex glass reactor." The aforementioned methodology has been employed for the direct determination of ethanol in chloroform and, in this work, we applied the vapour generation method for the FTIR determination of ethanol in alcoholic beverages.The determination of ethanol in beverages is commonly carried out by official methods, based on physical measure- ments, often after a previous distillation,12J3 or by a redox volumetric procedure. l 4 Instrumental methods based on the use of gas chromatography,15J6 liquid chromatography,17,1* po- tentiometry 19 and differential-pulse polarography20 offer good alternatives for the fast and accurate determination of ethanol. Of the various instrumental methods available, IR spec- trometry, in both the middle and near ranges, provides interesting possibilities for the determination of ethanol, using transmittance21-26 and absorbance measurement^,^^-^ in the latter case giving limits of detection of the order of 0.025% VJV.The aim of this work was to develop a simple method for the determination of ethanol in alcoholic beverages by means of injection of samples into a volatilizer reactor and measurement in the vapour phase, in order to solve sensitivity problems and to avoid the interference of sugars, the absorbance of which is close to the measurement band of ethanol and can cause serious limitations for direct analysis in the presence of the sample matrix.Experimental Apparatus and Reagents A Magna 550 FTIR spectrometer (Nicolet, Madison, WI, USA), equipped with a temperature-stabilized DGTS detector, a long- lasting Ever-Glo source and a KBr beamsplitter, was employed for spectral measurements with a nominal resolution between 1 and 32 cm-1 using a Wilmad (Buena, NI, USA) ultramini long- path cell, Model 3.2, with a volume of 100 ml and a permanently aligned multiple bandpass of 3.2 m, equipped with a ZnSe window.OMNIC software was used to control the instrument, for data acquisition and also for processing the analytical results. The FTIR spectra of liquid samples were obtained using a micro flow-through cell (Spectra Tech, Stamford, CT, USA) with ZnSe windows and a pathlength of 0.025 mm. The manifold employed for the vapour generation FTIR measurements (Fig. 1) was a single-channel manifold with a nitrogen Carrie1 flow which includes a volatilization reactor of924 Analyst. July I996, Vol. 121 2.5 ml internal volume with a gas inlet and a gas outlet. Samples were injected inside the reactor through a septum using Hamilton microsyringes of different volumes. The temperature of the reactor was controlled by means of a thermocouple and operated using a laboratory-made electrically controlled heater. The vapour phase generated inside the reactor was passed through the IR gas cell using nitrogen carrier gas.The spectra of the transient signals obtained for samples were processed against a calibration line obtained from aqueous solutions of ethanol. Analytical-reagent grade absolute ethanol (99.5%) from Panreac (Barcelona, Spain) was employed for the preparations of standards and N2 (C-45) from Carburos Metalicos (Barcelona, Spain) was employed as the carrier gas. General Procedure Inject 1 pl of untreated sample into the reactor, previously heated to 80 or 100 OC, and transport the vapour phase generated into the gas cell of the FTIR spectrometer using a carrier nitrogen flow of 300 ml min-1. The Gram-Schmidt plot, which is a representation of the light intensity arriving at the detector as a function of time, is obtained for the transient signals generated.Each Gram- Schmidt point corresponds to an FTIR spectrum in which, an appropriate spectral range corresponding to the absorption of ethanol can be selected. With the absorbance data from 1150 to 950 cm-1, a chemigram can be constructed. The chemigram provides a flow injection (FI) recording which represents the absorbance, in a selected wavenumber range, as a function of time. The areas of the chemigram peaks were employed as the analytical quantitative variable using both uncorrected and corrected data, taking into consideration a baseline correction established between the two valleys found before and after the sample injection.Data for samples were interpolated on the corresponding calibration graph obtained for aqueous standard solutions of ethanol injected in the same way as the samples. Results and Discussion Vapour-phase FTIR Spectra Versus Liquid-phase FTIR Spectra The FTIR spectra obtained for pure ethanol and aqueous solutions of ethanol, obtained using a micro flow liquid transmittance cell of 0.025 mm pathlength with ZnSe windows, indicate that the use of a water blank creates background problems except in the wavenumber range between 3000 and 800 cm-1, and additional problems appear between 2500 and 1500 cm-1, due to the strong absorbance of water, which provides negative bands for aqueous solutions of ethanol or aqueous samples recordered in front of a blank of pure water.Around 1050 cm-1, ethanolic solutions provide an intense band which appears clearly defined in both standards and real samples of wine and beer. This band can be chosen for analytical purposes, but the presence of sugars, such as glucose, sucrose or maltose, strongly affects this band and hence a careful baseline correction or first-order derivative spectra must be used in order to obtain accurate results in the analysis of samples using aqueous solutions of ethanol as standards. In contrast, the spectra obtained in the vapour phase, for both samples and aqueous standards, in front of an N2 blank (see Fig. 2), provide well established bands reducing the background problems to the wavenumber ranges 4000-3400 and 2000-1200 cm- 1. The sensitivity obtained for the injection of 1 pl of sample inside the volatilization reactor and the recording of a single transient signal, using a 3.2 m pathlength, is comparable to that found for batch measurements in the liquid state but, as can be seen in the inset of Fig.2, the spectra for real samples in the wavenumber range 1100-800 cm--l are well shaped and free from the interference of matrix components such as sugars or water. Effect of Instrumental Parameters The effect of the instrumental parameters such as nominal resolution, number of scans and background conditions on the quality of the analytical signals obtained when working in the vapour phase was evaluated. Concerning the number of scans required to establish the background, 10 is sufficient to obtain an appropriate and stable background of the gas cell when passing an N2 flow con- tinuously through the cell.A greater number of scans does not improve the stability of the reference spectra and increases the time of analysis. The number of scans employed to obtain each spectrum, working in the series mode, was considered in order to Iiave an adequate number of points per series to depict the correspond- ing chemigram and to obtain a good S/N. On the other hand, the effect of the nominal resolution is closely related to the number of scans which can be made in a fixed period of time, because the use of a poor nominal resolution involves a low time acquisition per scan. Hence, to obtain the best S/N, both parameters must be balanced for a fixed acquisition time. The use of resolution values 1 8 cm-1 decreases the noise of the blank without affecting the area of peaks obtained for ethanol solutions, thus improving the S/N.Under these conditions the measurements are carried out faster than using good nominal resolution values and clear recordings of transient signals can be obtained. Table 1 summarizes the effect of the nominal resolution and the number of scans on the area and height of the FI recording for an acquisition time of the order of 2 s. As can be seen, an increase in the number of scans does not lead to significant variations of the sensitivity of the determination of ethanol and sensitive and reproducible values can be obtained, under all the different conditions assayed. Effect of Experimental Parameters The carrier gas flow is a critical parameter which clearly affects the speed of sample introduction into the measurement cell and, I Thermocouple L Ic-.0 0 U controlled Waste FTI R Fig. 1 Manifold employed for vapour generation-FTIR analysis.Analyst, July 1996, Vol. 121 925 as can be seen in Fig. 3, an increase in the carrier flow rate decreases the sensitivity linearly when height values of the chemigram peaks are considered, and according to a power law for area values. The relationship between FI recording and carrier gas flow rate is so clear that it is possible to normalize the area values with respect to the lowest flow, and this normalization provides comparable values for the area of the chemigrams obtained for the same concentration of ethanol injected under different conditions, when the flow values are accurately measured.Temperature is the most important parameter introduced by the vapour generation FTIR technique. The appropriate selec- tion of the volatilization temperature provides the best sensitiv- ity, owing to the complete volatilization of the analyte in the shortest time, and also the best selectivity, owing to the possibility of elimination of the matrix effect by means of the discrimination, in terms of relative volatility, of all the compo.unds present in the same sample. Fig. 4 shows the effect of the volatilization temperature on the chemigram of ethanol established between 1150 and 950 cm-1 and that of the water chemigram established between 1586 and 1480 cm-1 for a series of injections of a 10% v/v ethanol standard solution.When the temperature increases, the differ- ence in time between the maximum of the chemigrams for ethanol and water decreases, thus indicating that at high temperatures both components are volatilized simultaneously. However, when the reactor temperature is between 70 and 80"C, the water vapour is generated with a certain delay compared with the generation of ethanol vapour. On the other hand, for a temperature S50"C the chemigrams for ethanol provide very wide peaks and the water vapour is slowly generated inside the reactor and provides an increase in the background. Hence, temperatures 280 "C must be recom- mended in order to obtain sensitive and reproducible measure- ments for ethanol. In general, an increase in temperature involves an increase in the height of the chemigrams, the area 0.30 - 0.30 Wave n u rn be rkm-' 0.20 a, S 10% ethanol standard e 8 a L2 0.10 0.00 3500 2000 Waven u m berkm-' 500 Fig.2 FTIR spectra of ethanol and alcoholic beverage samples obtained in the vapour phase. 10% v/v aqueous standard of ethanol (-), water (---). Inset: 10% v/v aqueous standard of ethanol (A), wine sample (B), beer sample (C), free-alcohol beer sample (D) and water (E). All spectra were obtained using a gas cell of 3.2 m pathlength with ZnSe windows for an injection volume of 1 1.11 and a carrier gas flow rate of 300 ml min-I. N2 was employed to establish the background. The spectra depicted correspond to the maximum absorbance of the transient signals obtained after injecting samples or standards in the manifold shown in Fig.1. Table 1 Effect of both nominal resolution and accumulated number of scans on the FI recording obtained for ethanol by vapour generation, using in all cases an acquisition time of the order of 2 s Area of chemigram peaks Height of chemigram peaks Sensitivity Sensitivity Resolution/ No. of Time of Value RSD [area LOD Value RSD [area 4 2 1.96 8.41 f 0.04 0.52 0.841 0.06 21.9 f 0.5 2.45 2.19 8 4 2.02 8.23 k 0.1 1 1.30 0.823 0.04 22.2 f 0.4 1.90 2.22 16 7 I .88 8.21 f 0.10 1.00 0.821 0.04 23.5 f 0.5 2.09 2.35 32 13 1.95 8.26 * 0.06 0.78 0.826 0.07 23.2 * 0.3 1.10 2.32 cm-1 scans acquisition/s obtained (%) (% v/v)-l] (% v/v) obtained (%) (% v/v)-'1926 Analyst, July 1996, Vol. 121 Uncorrected height I values remaining constant in the temperature interval studied.This enables one to work in a wide range of temperature without loss of sensitivity and thus offers interesting possibilities of reducing the contribution of water to the spectra obtained during analysis by vapour generation-FTIR. An additional parameter that must be taken into consideration in order to obtain appropriate sensitivity and precision values in analysis by vapour generation-FTIR spectrometry is the volume of sample injected. Some experiments (see Fig. 5 ) 0 2 4 6 8 1 0 1 2 Timelm i n -.- Height 2o h, + Uncorrected area v) - Normalized area G c .- 0 400 800 1200 1600 Carrier gas flow/ml min-' Fig. 3 Effect of the carrier flow on the FI recording. Experimental conditions: ethanol concentration, 10% v/v; injection volume, 1 pl; reactor temperature, 100 "C; and nominal resolution, 8 cm- l.Uncorrected values are direct experimental data and normalized area data were obtained with respect to the lowest flow value employed. 12 j ,I-- a O F . 1 ' I ' I ' I ' 1 ' I 0 1 2 3 4 5 6 Volume of aqueous solution of ethanol /MI Fig. 5 Effect of the injection volume on the chemigram area of a solution of 10% v/v ethanol. Experimental conditions: temperature, 80 "C; carrier flow rate, 359 ml min-I; and resolution, 8 cm-l. indicated that on increasing the volume of sample injected the area of the chemigram increases, but the reproducibility of the measurements decreases. However, the fact that results ob- tained with increasing volumes of ethanol provide a typical regression line of A = 0.28 + 6.15 V ( A = area, V = volume in pl), with a regression coefficient Y = 0.9999 1, indicates that the system works correctly for different volumes of sample injected, thus providing a direct way for improving the analytical sensitivity.Additionally, the method can be standard- ized by using both the same volume of different standards or different volumes of a single standard. Effect of Sugar The presence of sugars strongly affects the IR spectral bands of ethanol when the measurements are carried out directly in the liquid phase, thus creating interferences. Vapour-phase generation avoids spectral interference from sugars because, when working at low temperature, only ethanol and water are volatilized in the reactor and, for this reason, sugars are not transported by the nitrogen carrier through the FTIR gas measurement cell.Results obtained by studying the effect of increasing concentrations of sucrose on the area of the chemigram peaks obtained, in the spectral range 1150-950 cm-1, for a fixed ethanol concentration of 10% v/v using a volatilization temperature of 100 "C, indicated that sugar has no effect on the area values and only sugar concentrations higher than 15% can affect the reproducibility of measurements. In fact, the injection of highly concentrated sugar samples provides a residue of decomposed sugar, indicating that this compound is not volatilized at the low working temperatures employed. Analytical Features of Vapour Generation FTIR Spectrometry for the Determination of Ethanol in Alcoholic Beverages. Table 2 summarizes the main figures of merit of the methodo- logy developed for the determination of ethanol using vapour generation-FTIR, carrying out the measurements of areas of chemigram peaks, both corrected and uncorrected with respect to a baseline established between the two valleys found before and after the ethanol chemigram or, alternatively, for blank solutions, between the measurements obtained at different injection times.Calibration lines were established for different, concentration ranges for the different kinds of samples. As can be seen, the limit of detection for the determination of ethanol by vapour generation FTIR corresponds to 0.02% v/v, which is appropriate for the quantitative analysis of all types of alcoholic beverages, also including regular beers and low- alcohol beers.Additionally, note that the aforementioned procedure is very fast, providing a sample frequency of 5 1 h-l and not requiring any preliminary treatment of samples. Analysis of Real Samples Ethanol was determined in a series of real samples of alcoholic beverages by vapour-phase FTIR and, as can be seen in Table 3, the values found are comparable to those obtained by reference procedures based, in the case of spirits and beers, on near-IR determination,28,31 and for wines on a pycnometric procedure. In all instances, analyses were carried out without any sample preparation, except in the case of beers, for which prior de- gassing was carried out by filtration using a filter-paper. The regression between values found by the developed procedure (y) and the reference values (x) provides the regression equation y = 0.09 + 0.997~ with a regression coefficient of Y = 0.99997 for measurements carried out at 80°C and y = -0.21 + 1.004~ with r = 0.998 forAnalyst, July 1996, Vol.121 927 Table 2 Figures of merit for the determination of ethanol in alcoholic beverages* Type of Dynamic range Temperature/ LOD S sample (% v/v) "C Calibration equation r (% v/v) (% v/v) (n, C ) (5, 30) Spirits 20-50 80 (5, 5) Beers 1-12.5 80 ( 5 , 5 ) (6, 10) A, = -2.3 + 0.894C 0.9997 0.06 0.4 A, = -2.3 + 0.901C 0.9998 0.06 0.5 A, = -0.09 + 1.064C 0.9998 0.02 0.13 A, = -0.07 + 1.089C 0.9996 0.07 0.2 A, = 0.0953 + 0.7077C 0.999 0.02 0.14 A, = 0.0323 + 0.7055C 0.9997 0.02 0.1 1 A, = 0.059 + 0.75C 0.99994 0.06 0.06 A, = 0.058 + 0.75C 0.9998 0.2 0.05 100 Wines 1-20 100 * Calibration equation: A,, corrected area taking into consideration a baseline established between the two valleys found before and after the ethanol chemigram; A,, uncorrected area; C, concentration of ethanol (% v/v); r, regression coefficient; LOD, limit of detection ( n = 5 ) for a probability level 99.86%; s, standard deviation for n analyses of a sample with a concentration C (% v/v).Table 3 Determination of ethanol in alcoholic beverages by vapour generation-FTIR Ethanol (% v/v) Sample Rum Whisky Cider Low-alcohol beer Free alcohol beer Beer 1 Beer 2 Cider Low-alcohol beer Free alcohol beer Beer 1 Beer 2 Wine 1 Wine 2 Wine 3 Wine 4 Wine 5 Wine 6 Wine 7 Wine 8 Reference value* 40.3" 42.6" 4 0.748 0.03" 5.70d 7.29" 4 0.74" 0.038 5.70" 7.29" 12.75b 22.44b 12.10b 17.65b 11.50b 10.2b 12.24b 15b Found+ Corrected Uncorrected 40.1 f 0.6 41.9 f 0.7 4.19 f 0.04 40.1 f 0.8 42.7 f 0.6 4.28 f 0.04 0.626 f 0.006 0.621 f 0.016 0.168 f 0.007 5.81 f 0.09 7.31 f 0.09 4.11 f 0.09 0.192 f 0.019 5.88 f 0.11 7.39 f 0.1 1 4.20 k 0.05 0.57 f 0.04 0.59 f 0.05 0.00 f 0.03 5.45 f 0.07 7.05 f 0.08 12.69 f 0.17 23.01 f 0.05 11.3 f 0.3 17.66 2 0.1 1 11.46 f 0.14 10.22 f 0.03 12.3 f 0.3 13.68 f 0.23 0.133 f 0.019 5.49 f 0.15 7.12 f 0.09 12.91 f 0.19 23.41 f 0.02 11.4 f 0.3 18.02 f 0.21 11.50 f 0.08 10.13 f 0.05 12.33 f 0.22 14.28 f 0.19 Tempera- ture/"C 80 80 80 80 80 80 80 100 100 100 100 100 100 100 100 100 100 100 100 100 * Reference values were obtained by (a) NIR determination following the method proposed by Gallignani et a1.28,3' and (b) in the case of wine samples by a pycnometric procedure. t Mean f s (n = 5).measurements carried out at 100 OC, which indicates that the method does not require a blank correction, because the intercept is statistically comparable to zero in both cases, and it does not have constant relative errors, because the slope is statistically equal to 1. In general, results obtained working at 100 "C provide the best sensitivity and precision. Conclusions The method developed is a fast and accurate alternative for the direct determination of ethanol in all types of alcoholic beverages. Compared with previously developed procedures involving IR measurements, vapour generation-FTIR allows direct analy- sis without any sample pretreatment, not requiring spectral baseline corrections and providing sugar-free interference measurements. The excellent limit of detection found for extremely low injection volumes, the absence of water background problems and the avoidance of the interference of low-volatility sample components offer exciting possibilities for developing simple procedures for the direct determination of ethanol in complex matrices. Studies are in progress in our laboratory in order to determine ethanol in blood.The authors acknowledge the financial support of the Conselleria de Educaci6n y Ciencia de la Genaralitat Valen- ciana (project GV 1021/93) and that of the Spanish DGICYT (project PB 92-0870). References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Dien, M., Introduction to Modern Vibrational Spectroscopy, Wiley, New York, 1993.Infrared and Raman Spectroscopy, ed. Schrader, B., VCH, Wein- heim, 1995. Eiceman, G. A., Clement, R. E., and Hiel, H. H., Jr., Anal. Chem., 1994,64, 70R. Garrigues, S. and de la Guardia, M., Analyst, 1991, 116, 1159. Curran, D. J., and Collier, W. G., Anal. Chim. Acta, 1985, 177, 259. GuzmBn, M., Ruzicka, J., and Christian, G. D., Vib. Spectrosc., 1991, 2, 1. de la Guardia, M., Garrigues, S., and Gallignani, M., Anal. Chim. Acta, 1992, 261, 53. McDonal, R. S., Anal. Chem., 1984, 56, 349. Hanst, P. L., Fresenius' J. Anal. Chem., 1986, 324, 579. Janatuinen, T., and Byckling, E., Int. Lab., 1985, 15(7), 12. L6pez-Anreus, E., Garrigues, S., and de la Guardia, M., Anal. Chim. Acta, 1995, 308, 28. Offical Methods of Analysis of the Association of Official Analytical Chemists, AOAC, Washington, DC, 1990. Kreneger, E., and Weber, S., Monatsschr. Brau., 1981, 34, 70. Pilone, G. J., J . Assoc. Ofl. Anal. Chem., 1985, 68, 188. Caputi, A. J., and Mooney, D. P., J . Assoc. Off. Anal. Chem., 1983, 66, 1152. Cutia, A., J . Assoc. Ojf Anal. Chem., 1984, 67, 192. Iwachido, T., Ishimaruk, K., and Toei, K., Anal. Sci., 1986, 2, 57. Morawski, J., Dicer, A. K., and Ivie, K., Food Technol., 1983. 37, 57. Kakabadse, G. J., Lab. Pract., 1990, 39, 51. Chan, W. H., Lee, A. W. M., and Cai, P. X., Analyst, 1992, 117, 1509. Glenn, A. L., J . Pharm. Pharmacol., 1963 15, Suppl., 123T. Agwu, J. V., and Glenn, A. L., J . Pharm. Pharmacol, 1967, 19, 76s. Heisz, O., Labor Praxis, 1989, 13, 402. Hasley, S. A., Anal. Proc., 1986, 23, 126. Coventry, A. G., and Hunston, M. J., Cereal Food World., 1984,29, 715. Dumoulin, E. D., Azain, B. P., and Guerain, J. T., J . Food. Sci., 1987, 52, 626. L6pez-Mahia, P., Simal Ghndara, J., and Paseiro Losada, P., J . Vib. Spectrosc., 1992, 3, 133. '928 Analyst, July 1996, Vol. 12 1 28 Gallignani, M., Garrigues, S., and de la Guardia, M., Af?cr/ Chin1 Artu., 1994, 287, 275. 29 Gallignani, M., Garrigue5, S., and de la Guardia. M., Andyst, 1994, 119, 1773. Paper 61004 79B 30 Gallignani, M., Garrigues, S., and de la Guardia, M., Analyyt, 1993, Rec-ei Lied Junuui-y 22, I996 118, 1167. Accepted Mascli 29, 1996 31 Gallignani, M., Gamgues, S., and de la Guardia, M., Anul Chrm Artu. 1994, 296, 155.

ISSN:0003-2654    

DOI:10.1039/AN9962100923

出版商:RSC    

年代:1996

数据来源: RSC