摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 101 ASU Highlights This review is the second Atomic Spec- trometry Update and describes develop- ments in the analysis of clinical samples, foods and beverages. By far, the most active area of research is the determina- tion of elements present in body fluids and tissues at the microgram per litre level. The element which could possibly be a candidate for the “Element of the Year” award is aluminium, which is dis- cussed in depth in the reviews. Although flow injection methods of analysis have been used extensively in clinical labora- tories only now is it becoming accepted as a viable approach to sample introduction into flames or plasmas for elemental determinations. Although the total con- centration of elements in body fluids can provide important toxicological or nutri- tional information more research is being conducted into the distribution of these elements in various samples.Therefore, chromatography coupled to atomic spec- trometry is seen as an expanding hybrid technique. The use of methods to increase the sensitivity of flame AAS has also attracted interest. In clinical laboratories with increasing sample throughput the major disadvan- tage of graphite furnace AAS has been seen to be the relatively long analysis time. This problem has been addressed by various workers and methods to speed-up the analysis time of this technique have been proposed. In contrast little work has been published on the analysis of foods and beverages. Alistair A. Brown Pye Unicam, UK102 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.I I FEATURES SAC 8 6 / 3 ~ ~ BNASS Parallel lecture streams will be held on the Monday, Tuesday, Thursday and Friday on SAC themes. The BNASS Atomic Spectrometry streams will be held on the Wednesday, Thursday and Friday. Update courses will be held on the Wednesday, as will several all-day tours. An exhibition of scientific equipment is planned. Full social and accompanying persons programmes will be organised. The Conference language will be English and the full programme will be published in the March 1986 issue of Analytical Proceedings. ~ AN INTERNATIONAL CONFERENCE ON ANALYTICAL CHEMISTRY AND ATOMIC SPECTROSCOPY IUPAC and The Federation of European Chemical Societies (FECS event No.78) ~ sponsorship has been obtained. UNIVERSITY OF BRISTOL, 20-26 JULY, 1986 Organised by the Analytical Division, Royal Society of Chemistry, in conjunction with The Spectroscopy Group of The Institute of Physics Further details and registration forms may be obtained from: Miss P. E. Hutchinson, Secretary of The Analytical Division, Royal Society of Chemistry, Burlington House, London W1V OBN, UK. ~ Plenary, invited and contributed lectures and posters. Special symposia on particular themes organised by RSC Groups and associated bodies. Workshops for the demonstration of new apparatus and techniques. One-day update courses (Wednesday), visits and social programme. LECTURERS Plenary: J. H. Knox (UK) ‘Advances in Columns and Packings for HPLC” M. Bonner Denton (USA) “Concepts for lmproved Automated Laboratory B. V. L’vov (USSR) “New Advances in Furnace Atomic Absorption Spectrometry” G. Tolg (FRG) “Extreme Trace Analysis of the Elements-The State of the Art Produc tivity ” Today and Tomorrow” Invited: J. F. Alder (UK), M. S. Cresser (UK), A. R. Date (UK), L. de Galan (NETH), J. Goldsmith (UK), J. G. Grasselli (USA), D. A. Hickman (UK), W. Horwitz (USA), R. D. Snook (UK), V. Sychra (CZECH), A. Thorne (UK), A. Townshend (UK), G. Werner (GDR), T. S. West (UK) STR UCTU RE SPONSORSHIP

ISSN:0267-9477    

DOI:10.1039/JA986010101b

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 103 Conferences and Meetings SAC 8613rd BNASS July 2&26, 1986, Bristol, UK This conference, which will be held in the University of Bristol, is organised by the Analytical Division of the Royal Society of Chemistry in conjunction with the Spectroscopy Group of the Institute of Physics. The Plenary Lectures will be given by Professor J. H. Knox (Edin- burgh), Professor M. Bonner Denton (Tuscon, AZ, USA), Professor B. V. L'vov (Leningrad, USSR) and Professor G. Tolg (Dortmund, FRG) and the Invited Lecturers will include J. F. Alder, industrial and metallurgical, pharmaceut- ical, plastics, rubbers and textiles, sur- faces, water and effluents, microcom- puters and microprocessors, optical- fibres, quality control, sensors, the integrated laboratory.In addition there will be update courses, workshops, an exhibition and a social programme. For further information contact Miss P. E. Hutchinson, Secretary of the Ana- lytical Division, Royal Society of Che- mistry, Burlington House, Piccadilly, London W1V OBN. M. Barber, M. S. Cresser, A. R. Date, J. Analytiktreffen 1986 man, W. Horwitz, R. D. Snook, V. 'ychra9 A* Thorne9 A. G* This meeting consists of Analytiktreffen Werner and T. S. West. There will also be and Ix CANAS, conferences on research an Association of British Spectroscopists and analytical applications of atomic spec- Professor L* de In troscopy and specifically analytical atomic J* G. Grasselli, D* A. Hick- September 15-19, 1986, Neubrandenburg, DDR to the above contributed papers and spectroscopy, respectively.The confer- been Offered in the ence languages will be German, England and Russian. The lectures, posters and posters have Techniques: fluorescence and emlSS1on discussions will cover the theory and sPectroscoPY 7 and kinetic analytical applications of atomic absorp- methods, chromatography 7 e'ectroana'y- tion spectrometry, atomic emission spec- methods, enzyme techniques, flow trometry (arc, spark, laser, high fre- injection methods, immunoassay masS quency, microwave and glow discharge spectrometry, microanalysis, techniques), atomic fluorescence spec- spectroscopy, probe methods, radio- trometry and X-ray fluorescence spec- chemistry, sample preparation, pre- trometry. concentration and separation, thermal F~~ further information contact D ~ ~ .methods, X-ray emission and other X-ray D ~ . sC. K. ~ i t t f i ~ h , K ~ ~ ~ - M ~ ~ ~ - methods. Materials and Areas of Applica- Universit5t ~ ~ i ~ ~ i ~ , sektion Chemie, tion: agricultural, atmospheric, biological DDR-7010 Leipzig, Talstr. 35, GDR. and microbiological, clinical, environ- mental, food and drink, geological, indus- FACSS '86 trial and metallurgical, pharmaceutical, September 28-October 3, 1986, St. Louis, plastics, rubbers and textiles, surfaces, MO, USA water and effluents. Other Aspects: auto- The Federation of Analytical Chemistry mation and robotics, biotechnology , che- and Spectroscopy Societies will hold its mometrics, data processing, education, 1986 meeting at the Cervantes Conven- historical, microbiological, clinical, envi- tion Center and Sheraton Hotel, St.ronmental, food and drink, geological, Louis. The Symposia will cover atomic spec- troscopy (including, for example, ICP- MS, ICP excitation mechanisms and applications, new plasmas for elemental analysis, X-ray fluorescence elemental analysis and fundamentals and applica- tions of ETA), chromatography, NMR, molecular and mass spectroscopy and diverse topics (such as automated sample preparation, chemiluminescence and environmental). The Royal Society of Chemistry is jointly arranging symposia on furnace atomic emission spectroscopy, background correction in AA, fitting calibration curves for atomic spectro- scopy and multi-divisional fluorescence. Workshops and short courses will be offered prior to, during and after the conference.The FACSS Employment Bureau will again be available. Details of these activities will be given in the Preli- minary Program brochure, which will be available in mid-summer 1986. There will be an exhibition of scientific instrumentation, services and publica- tions: for exhibition details contact Dr. E. G . Brame, Jr . , FACSS Exhibit Director, 133 North Cliffe Dr., Wilmington, DE 19089-1623, USA. For additional information contact Dr. M. Fishman, FACSS XI11 General Chair, USDA-ERRG, 600 E. Mermaid Lane, Philadelphia, PA 19118, USA. Second International Colloquium: Solid Sampling with Atomic Spectroscopic Methods October 13-15, 1986, Wetzlar, FRG The 1986 colloquium in this bienniel series, jointly organised by GDCh, Fach- gruppe Analytische Chemie and Arbeitskreis fur Mikro- und Spuren- analyse der Elemente (A.M.S.El.), is intended to focus on the state of the art of solid sampling and will consist of a bal-104 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 anced mixture of concise oral papers and posters, interspersed with discussion periods for themes of particular impor- tance. The main topics will be: Theory and Instrumentation; Methodology (Proce- dures); Biological Applications; Environ- mental Applications; and Product and Quality Control. The registration fee is DMlOO before June 1 1986 and DM125 after this date. For further information contct Dr. M. Stoeppler, Institut fur Chemie, KFA- Postfach 1913, D-5170 Julich 1, FRG. Ninth Australian Symposium on Ana- lytical Chemistry April 27-May 1, 1987, Sydney, Australia This meeting will be held at the Centre- point Exhibition and Convention Centre and organised by the Analytical Chem- istry Division of the Royal Australian Chemical Institute.The Symposium will consist of plenary lectures, contributed papers, workshops and a trade exhibition, and the lecture programme will cover agricultural, clinical and biological, com- puter and laboratory automation, en- vironmental, food and carbohydrate, mining and metallurgical, occupational hygiene and pharmaceutical areas of analytical chemistry. For further information contact Mr. J. E. Eames, Secretary, 9AC, P.O. Box 137, North Ryde, NSW 2113, Australia. Euroanalysis VI September 7-11, 1987, Paris, France Euroanalysis VI will be held at the Centre International de Conferences in Paris. The plenary, keynote and contributed lectures will cover all aspects of analytical chemistry but special sessions are planned to discuss: the use and construction of analytical probes; applications of analy- tical methods for solving environmental problems; analysis of solid-state samples; and new methods of teaching analytical subjects (poster session). There will also be an exhibition and a social programme. Further information is available from GAMS, 88 Boulevard Malesherbes, 75008 Paris, France. ECASIA 87 October 19-23, 1987, Fellbach, FRG The 1987 European Conference on Appli- cations of Surface and Interface Analysis will take place in Fellbach, near Stuttgart. For details contact U. Nagorny, Max- Planck-Institut fiir Metallforschung, Insti- tut fur Werkstoffwissenschaften, See- strasse 92, D-700 Stuttgart 1, FRG.

ISSN:0267-9477    

DOI:10.1039/JA986010103b

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

104 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Future Issues will Include- Thermal Vaporisation for Inductively Coupled Plasma Optical Emission Spec- trometry: A Review-Henryk Matu- siewicz A Predictive Model of Plasma Matrix Effects in Inductively Coupled Plasma Atomic Emission Spectrometry-Michael H. Ramsey and Michael Thompson Matrix Effects in Electrothermal Vapori- sation Inductively Coupled Plasma Spec- trometry-Henryk Matusiewicz, Fred L. Fricke and Ramon M. Barnes Laser-excited Atomic Fluorescence (LAFS) as a Real Analytical Method. Part 1. Design of a Graphite Tube Atom- iser for the Determination of Trace Studies of a Low-noise Laminar Flow hnounts of Lead by LAFS-Klaus Dit- Torch for Inductively Coupled Plasma trich and Hans-Joachim Stark Atomic Emission Spectrometry- John Davies and Richard D. Snook Application of a Simplified Model for Atom Formation by a Tungsten-strip Determination of Inorganic Anti- Heater in Atomic Absorption Spec- mony(V) and Antimony( 111) Species in trometry-S. Nakamura Natural Waters by Hydride Generation Comparative Study of the Sputtering Pro- Atomic Absorption SPectr0metrY-S- c* cess in the Conventional and Microwave- Apte and A- G* Howard coupled Hollow-cathode Discharge- Sergio Caroli, 0. Senofonte, N. Violante, 0. Falasca and A. Marconi mination of the Elemental Contamination The Update in the June issue is-Instru- of Vegetables-M. Ottaviani and P. mentation-John Marshall, Stephen J. Magnatti Haswell and Richard D. Snook Atomic Spectrometry Atomic Absorption Spectrometric Deter- Update

ISSN:0267-9477    

DOI:10.1039/JA986010104b

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 105 Constant-temperature Atomiser = Computer Controlled Echelle Spectrometer System for Graphite Furnace Atomic Emission Spectrometry Erik Lundberg, Douglas C. Baxter and Wolfgang Frech Department of Analytical Chemistry, Universityrof Ume8, S-90 1 87 Ume8, Sweden A constant-temperature atomiser (CTA) was combined with an echelle spectrometer for graphite furnace atomic emission spectrometry (GFAES). The spectrometer incorporated a quartz refractor plate, which provided wavelength modulation background correction, under computer control. In order to investigate various parameters for GFAES measurements we developed a versatile computer program (for data acquisition and waveform generation) and a CTA in which tube dimensions and heating modes could be varied.Optimum detection limits and dynamic calibration ranges were obtained for tubes of medium length having spatial and temporal isothermality. A three-step square waveform in combination with appropriate data acquisition parameters was used, which also rendered good background correction capability. Detection limits for 16 elements and analytical results for five elements in two NBS reference materials are presented. Keywords: Graphite furnace atomic emission spectrometry; echelle spectrometer; constant-temperature atomiser; wavelength modulation background correction; computer control Graphite furnace atomic emission spectrometry (GFAES) offers, in comparison with conventional, line-source graphite furnace atomic absorption spectrometry (GFAAS), the possi- bility of performing multi-element determinations with similar detection limits for a number of elements.'.* In spite of this advantage, the technique has only been used on a minor scale.This is largely due to the fact that line-source AAS is already well established3 and that AES has a number of intrinsic limitation^.^ With respect to GFAES, the major disadvantage is the relatively low temperature of the graphite furnace as an emission source,s which precludes the sensitive determination of several elements of biological and environmental interest because of their high excitation potentials (e.g., As, Cd and Se). Of paramount importance for optimum GFAES perfor- mance, is the design of the atomiser. Owing to the dependence of the emission intensity on the gas-phase temperature as experienced by the analyte atoms, it is desirable to vaporise the sample when the excitation temperature over the entire tube length is at an optimum with respect to signal to noise ratio (SNR).This can be achieved by use of a constant- temperature atomiser.6 Alternative ways to approach con- stant-temperature conditions include the L'vov platform technique2 and probe atomisation7 in Massmann-type fur- naces. However, the existance of temperature gradients in these types of furnace is well established's and such conditions are particularly unfavourable for GFAES, because of self- absorption at higher analyte concentrations.9 Additionally, as has been demonstrated elsewhere for GFAAS, the use of high gas-phase temperatures during atomisation normally reduces spectral, as well as non-spectral interference effects,lOJl and this is also to be expected for GFAES.In order to avoid distortion of the signal transients and resultant misinterpretation of the analytical data,l* as well as to extract maximum information from the peak shape, it is essential to use a computer based data acquisition system in GFAAS and in GFAES.13J4 For GFAES the computer is also necessary for maximum flexibility with respect to wavelength modulation background correction. This flexibility is essential for good correction capability as well as for achieving optimum SNRs. The computer also offers the possibility of extending the useful concentration range by making intensity measure- ments in the wings of the emission line profile.' This paper describes the optimisation of (i) an isothermal atomiser with respect to tube dimensions and heating mode and (ii) the parameters for wavelength modulation and data acquisition (e.g., waveform and frequency), Experimental Instrumentation A block diagram of the instrumentation is shown in Fig.1 and the spectrometer and computer components used are listed in Table 1. The Cchelle monochromator was modified for wavelength modulation by installing a quartz refractor plate mounted on the shaft of a scanner torque motor situated behind the entrance slit. Initially, a 3-mm thick plate was used, but this was later replaced by one of 6-mm thickness. A scanner controller was used to operate the torque motor and provide a plate positional feedback output to the storage oscilloscope (Tektronix Inc.Model 564) for waveform moni- toring. A single-element exit slit cassette was installed; matched entrance and exit slit dimensions, with the maximum Furnace programmer I I Scanner controller monochromator PMT I H T l J Amplifier Oscillo- I I I - ' 1 I In t I Computer Fig. 1. Schematic diagram of GFAES instrumentation: CTA, constant-temperature atomiser (see Fig. 2 L, lens; QP, quartz refractor plate; PMT, photomultiplier tube; and IT, cup and tube temperature, output from temperature control photodiodes106 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Table 1. Spectrometer and computer components Monochromator . . . . Spectrometrics SMI 111 Cchelle monochromator with 20 PMT power supply nm mm-1 at 400 nm Reciprocal linear dispersion: 0.124 Slit widths: 0.05 mm Slit heights: 0.50 mm General Scanning Inc.controller (CCX 653) and torque motor (G-300 PD) with 3- or 6-mm quartz plate (Suprasil I, Heraeus-Schott Quartzschmelze, Hanau, FRG) Wavelength modulation Lens . . . . . . . . f= 100 mm, quartz Detector . . . . . . Hamamatsu R292photomultiplier Signal processing . . . , Alpha LSI minicomputer with 64 kbyte memory; 2 x 256 kbyte floppy disk; 16-channel 12 bit ADC and 2-channel 12-bit DAC; 10 kHz real-time clock; Beehive DM5 terminal; Houston DMP-2 digital plotter 11 Fig. 2. Constant-temperature atomiser. 1, Graphite tube; 2, graph- ite cup; 3, tube contact electrode, spring-loaded; 4, cup contact electrode; 5, silicone-rubber gasket; 6, connection for power lead to tube; 7, connection for power lead to cup; 8, connection for cooling water supply; 9, connection for outer gas flow supply; 10, removable window assembly; 11, connection for upper gas flow supply; 12, fibre-optic cable inlet (connected to IR-sensitive photodiode); 13, base plate with electrical insulation.A and B, cup terminal blocks; C and D, tube terminal blocks. B and D are spring-loaded 1 . 1 5 mm 1 Fig. 3. Atomiser confi urations. Ty e 1 electrodes used with pyrolytic graphite tubes &)-(c). Type 5 electrodes used with total pyrolytic graphite tube (d). Dashed lines indicate positions of cup contact electrodes available slit heights of 0.5 mm, were used throughout. The photomultiplier tube (PMT) was operated using the spec- trometer power supply, and the signal generated was ampli- fied by a home-made current to voltage converter with variable gain to produce a maximum output of 10 V, thereby ensuring that the full range of the analogue to digital converter (ADC) was utilised.For measurements of the calibration graphs, it was neces- sary to reduce the PMT voltage at high concentrations to avoid exceeding the 10-V ADC range. Consequently, all calibration graph data were normalised to a PMT setting of 1000 V by using appropriate conversion factors (determined by means of a hollow-cathode lamp). In this respect, however, the range of the ADC is limiting in terms of the effective concentration interval that can be measured using a given PMT voltage. The construction of the constant-temperature atomiser, which is shown in Fig.2, has been described previously,6 as has the principle of the optical feed-back temperature control of the tube and cup.15 The tube length and number of contact electrodes were varied to investigate the effects on GFAES performance, see Fig. 3. Type 1 electrodes, which were used for configurations (a)-(c), were useful up to wall temperatures of ca. 2600 K for configuration ( c ) . Above this temperature, excessive deterioration of the tube lifetime occurred, due to sparking between electrodes and tube. In conjunction with the thin total pyrolytic graphite (TPG) tubes, configuration (d), it was necessary to use type 2 electrodes in order to avoid compressing the tubes. With configuration ( d ) , it was possible to reach wall temperatures of ca.2950 K with lifetimes of approximately 50 firings. Tube temperature measurements referring to the outer surface were made using a disappearing filament pyrometer (Keller Spezialtechnik Pyro Werk Model PBO 6AF3). Spec- troscopic temperature measurements were made using “two- line” atomic absorption methods with lead16 and nickel” as the thermometric species, and also by an iron atomic emission “slope” method. l8 The atomic absorption measurements were made using instrumentation described earlier.12 In all instances, the reported spectroscopic temperatures are based on peak-area signal evaluation. The furnace temperature programme was varied through- out the experiment. However, in all instances the tube temperature was ascertained to be constant before the cup heating was initiated.Thus constant-temperature excitation conditions were always established. Ashing temperatures of 670 K (cup) and 870 K (tube) for 30 s were used throughout. Heating rates for the tube and cup were of the order of 2000 and 8000 K s-1, respectively. Purge gas flow-rates of 140 (upper flow) and 1200 ml min-1 (outer flow) were used for all measurements. Data Acquisition Data were acquired from the pre-amplifier output (at the rate of ca. 9400 readings per second for the 6-mm quartz plate) by means of the minicomputer-based data system listed in Table 1. The time relationship between the wavelength modulation and the sampling of intensity data is shown in Fig. 4. The computer drives the scanner motor controller through one of its digital to analogue converters (DAC).The modulation waveform is the three-stepped waveform shown in the upper part of Fig. 4. The middle step is centred on the atomic line of interest so that intensity readings can be made on the line (0 V output from DAC) and off the line on both sides (maximum k 5 V output from DAC, corresponding to an angle of deflection of the quartz plate of k 10’). The time spent by the quartz plate at each wavelength position is divided into a delay time, which is necessary for the plate to move into position, and a sampling time. The delay time and the number of readings taken during the sampling interval (which determines the sampling time) are specified in the user program. In order to reduce noise, theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 107 Displacement are averaged and called I , . those taken on the line are -5 V 0 +5 v 1 Emissibn line B? profile Wavelength modulation inierval Plate position relative to light path I % t 4 Fig. 4. Timing scheme for wavelength modulation and data acquisi- tion. D, Delay time before sarnplin ; S, sampling interval; llt, detection frequency; Zl-Z4, see text; Ed, background largest possible number of analogue to digital readings should be taken, without distorting the emission peak shape, and averaged, at each wavelength position. To save memory space, up to 32 readings are accumulated in one memory location. If more readings per position are required, corre- spondingly more memory space is needed. Intensity data are stored in prime memory as a one-dimensional array, with 4000 memory locations being available for this purpose.This means that if, for example, 96 readings per position are taken and the detection frequency is 49 Hz (optimum parameters for the 6-mm plate, see Table 4), then sampling times during atomisation as long as 13.6 s are possible before the 4k array is filled up. The assembly language program used for emission intensity data acquisition and waveform generation has the following input arguments: delay time before sampling, number of readings at each wavelength position, quartz plate start and return positions (determining the modulation interval), plate increment (determining the number of steps in the square wave) and number of sweeps. By means of these parameters the SNR, as well as the modulation interval, can be optimised and, in addition, profile scans with arbitrary resolution can be performed.To accomplish the three-dimensional plot of the copper emission line as a function of time seen in Fig. 10, a 23-step square wave was used. The number of readings per wavelength position was eight, the delay time was 1 ms and the modulation interval k 4.6 V (corresponding to k 0.0283 nm), resulting in a time resolution of 33 ms per scan (i.e., per half-cycle of the modulation waveform). Data Processing Because of the high data rate, it is impossible to compute background corrected emission values in real time from the intensity data. Therefore, the processing starts when the atomisation step is finished. The program is written in BASIC with assembly language sub-routines for, amongst others, emission intensity data acquisition, tube and cup temperature data acquisition, refractor plate waveform generation and access to raw intensity data.During each half-cycle of the modulation waveform, a number of readings are taken on and off the line. The intensity readings that are taken off the line averaged and called Z2 (see-Fig. 4). For the next half-cycle, corresponding averaging is performed, and these mean intensity values are called 13 and 14, respectively. The background corrected emission value for the first half-cycle is then calculated as 12 - (I, + Z#2. In this way, one corrected value is calculated for every half-cycle of the modulation waveform, and so the detection frequency will be llt (see Fig.4). Subtracting the mean value of I I and I3 is necessary in order to reduce the effect of rapid changes in background emission with time. The main menu of the BASIC program includes the following options: modify input parameters (data acquisition and waveform generation), modify plot parameters (scale expan- sion in x and y directions), start new run, plot specific or background emission signal as a function of time (smoothed or unsmoothed), plot tube or cup temperature - time profiles, calculation of peak height and peak area and plot emission line profiles as a function of time. After the atomisation step is finished, plotting of the signals commences directly, as no time-consuming data reduction takes place. 19 Instead, the values to be plotted are calculated from the stored raw intensity data immediately before they are fed to the plotter (i.e., calculations are included in the “plotting loop”).This saves time, as only the raw data required for a particular plot are converted into emission (or temperature) values. During plotting, the peak-height and peak-area values are also calculated for subsequent display on the printer. Reagents and Materials Standard solutions were prepared in 0.01 M acidic media from reagents of the highest available purity. Acids were taken from an all-quartz sub-boiling still (Acidest ; Heraeus Quartzschmelze). For all determinations of potassium, 2 pg of caesium chloride were added as an ionisation buffer. In the final list of detection limits, recommended matrix modifiers were added to lead [200 pg NH4H2P04 + 10 pg Mg(N03)2] and thallium (1y0 H2S04) solutions.Pyrolytically coated graphite (PCG) tubes ( i d . 4.0 mm, 0.d. 6.0 mm and length 8.0 or 18.0 mm) and cups (i.d. 2.8 mm, 0.d. 5.2 mm, length 8.0 mm and depth 7.0 mm) were manufactured from RWOl high-density graphite (Ringsdorff- Werke). Total pyrolytic graphite tubes [i.d. 5.0 mm, wall thickness 320 pm and length 18.0 mm; Pye Unicam (Cam- bridge, UK)] were used in conjunction with PCG cups for some of the experiments. A mixture of 5.5% methane in argon (SR grade) was used throughout as the purge gas. Procedure Sample Preparation and Determination The NBS Standard Reference Materials, bovine liver 1577 and oyster tissue 1566, were digested in duplicate in concentrated distilled nitric acid using a similar procedure to that described previously for tissue samples.20 Two blank solutions were digested in parallel and all samples diluted 1 + 19 with Millipore water.All sample manipulations were performed in a laminar flow “clean” bench (Ultramare AB) provided a class 100 working environment. Where required, further dilutions were carried out to bring the analyte concentrations within the linear range of the calibration graphs. Sample aliquots of 2-10 p1 were pipetted into the graphite cup through the injection port in the tube. Standard additions were made by dispensing appropriate amounts of analyte into the cup. Aqueous calibration graphs were also prepared using identical measurement parameters to assess the magnitude of any matrix related interference effects.108 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 Table 2. Effects of optical baffles and furnace windows on GFAES signals measured at 2300 K with configuration (b). Indium, wavelength 410.17 nm IntensityN Optical components* Peak to peak In signal B LH RH noise (250pg) SNR X 0.08 2.37 29.6 X X X 0.35 2.11 6.0 X X 0.42 2.20 5.2 X X 0.10 1.90 19.0 X X 0.33 2.28 6.9 * B = 2 mm i d . optical baffles; LH = left-hand window, closest to monochromator; RH = right-hand window. Results and Discussion Optimisation of Some System Parameters As in all emission techniques, the GFAES detection limit is dependent on the signal to noise ratio. Consequently, optimum GFAES performance is to be expected when the optical configuration permits effective discrimination against the background signal component, as this factor determines the magnitude of the observed noise. In GFAES, the intense blackbody radiation generated by the heated graphite tube itself constitutes the major source of background emission.21 Previous studies have shown that even after careful optical baffling to prevent radiation from the furnace wall from directly entering the monochromator, a residual background is still observed.22 This residual background arises principally from Rayleigh scattering of tube wall continuum, with some contributions from optical components and aberrations in the optical system.Therefore, the effects of optical baffling and furnace windows on the SNR were investigated, see Table 2.The baffles do not significantly attenuate the emission intensity, but do reduce the background and hence the noise level, thus enhancing the SNR. It is evident that a major source of noise is the window (right hand) located on the furnace side remote from the monochromator.23 This window reduces the SNR by reflecting tube-wall radiation directly into the optical path. In contrast, the left-hand window degrades the SNR chiefly through absorption of specific radiation, with perhaps some scatter contribution to the noise. As can be seen, the SNR ratio was improved more than five-fold by replacing the windows with optical baffles, which demon- strates the unsuitability of windows in terms of the SNR in GFAES. A number of, to some extent, inter-related variables ultimately define the SNR in GFAES.In order to optimise several of these variables simultaneously, a simplex procedure was carried out, considering the following five factors: upper gas flow-rate, tube temperature, slit width, photomultiplier voltage and lens position. The results of this optimisation can be seen in Table 3. Optimisation of Data Acquisition Parameters The use of a minicomputer confers great flexibility on the data acquisition and subsequent data-handling routines. The former is particularly attractive in relation to the wavelength modulation background corection technique, as the computer can be used to generate the modulation waveform best suited to the measurement requirements.14 This versatility was exploited in order to investigate the background correction capability in the wavelength modulation GFAES system described here, as well as attainable SNRs.With wavelength modulation devices, the best SNK 1s generally achieved using a three-step square wave,19 as used here, of sufficient amplitude to closely bracket the wavelength distribution of the analyte signa1.24 However, such a wave- form does not easily lend itself to high-frequency operation owing to the finite response time of the torque motor.25 The use of a very narrow modulation interval in order to increase frequency will be limiting in terms of the analyte concentration range that can be accommodated before line broadening causes the signal profile to develop beyond the edges of the modulation interval and hence introduce background correc- tion errors.9 When operating the 3-mm quartz refractor plate at k3.0 V, it was therefore ascertained that the modulation interval was sufficient to make the background correction measurements at wavelengths beyond the extremities of the emission line profile, at concentrations more than two orders of magnitude above the detection limit.The upper part of Table 4 shows the effects of varying the delay time and the number of data points averaged during the sampling interval on the SNR for the 3-mm plate (see also Fig. 4). The first column of SNR results are based on peak to peak noise from calcium nitrate, which under the experimental conditions used gave rise to an almost steady-state back- ground. Such background signals are frequently observed in GFAES ~ystems.14~22~25 Fig.5 illustrates the specific and background signals giving rise to the best and worst SNRs. The last two columns of SNR results in Table 4 are based on peak to peak noise from sodium chloride, which gave rise to a transient background signal. From the data in the upper part of Table 4 it can be concluded that increasing the number of data points taken enhances the SNR (as long as the peak shape is not distorted), and that the delay time should be long enough to avoid sampling in the wings of the emission profile whilst the plate is moving into position. Although the SNR results for a 4.5-ms delay suggest that this value is quite adequate, oscilloscope monitoring of the galvanometer response indicated that the quartz plate was not, in fact, completely stationary after this time, see Fig.6. It would therefore seem that some small positional uncertainty is not so critical in limiting the SNR, but beyond this point the SNR does ultimately suffer, as shown by the results for a 3.5-ms delay time. Fig. 7(a) shows the background and corrected signals for 50 pg of sodium chloride using the 3-mm plate. In order to investigate if better background correction could be obtained at higher modulation frequencies, the quartz plate used in the above measurements was replaced by a 6-mm plate. As the wavelength displacement is directly proportional to plate thickness, doubling the latter permits the same modulation interval to be obtained with only half the rotational movement of the plate.26 Practically, this reduces the time required for the plate to move to each wavelength position, and conse- quently the sampling delay time can be decreased and the modulation frequency increased.As can be seen from Fig. 7(b), the background correction is good considering the magnitude of the sodium chloride signal and the fast rise time, 100 ms. An advantage of the 6-mm plate is that profile scans covering as much as 0.1 nm can be implemented. The lower part of Table 4 shows that the SNR is somewhat improved at higher frequencies, which is in accordance with the results reported by Harnly.27 Optimisation of Furnace Configuration Recent developments in graphite furnace technology have been directed towards overcoming the limitations inherent in Massmann-type atomisers. It is now recognised that the L'vov platform technique offers some degree of temporal isother- mality, thereby significantly reducing interference effects.28 Nevertheless, there still exists a considerable temperature gradient along the tube length,s resulting in problems withJOURNAL OF AN4LYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 109 Table 3. Simplex optimisation of some GFAES instrumental parameters, using configuration (b). Aluminium, wavelength 396.15 nm Parameter* A1 A.2 A3 A4 A.5 A6 B Upper gas flow-rate ml min-1 170 210 210 110 55 50 140 Slit TemperatureK widthimm 2530 0.05 2530 0.1 2640 0.1 2580 0.2 2470 0.2 2470 0.025 2500 0.05 PMT voltageN 800 950 850 850 850 lo00 900 Lens positiont/mm 194 194 190 190 198 202 198 SNR$ 41 70 19 15 40 26 126 * A = Sets of parameters used to create initial simplex; B = projected parameters generated by simplex, used in subsequent measurements t Distance from entrance slit to lens centre.Centre of furnace was ca. 400 mm from entrance slit. # Signal to noise ratio based on peak height from 300 pg A1 versus peak to peak noise of blank measurement. except where otherw:.se indicated. Table 4. Optimisation of SNR for 3- and 6-mm quartz refractor plates. No ashing step Sampling Number of data delay time/ms points averaged 3-mm plate, displacement f 3.0 V- 8.0 16 8.0 32 8.0 64 5.7 64 4.5 64 4.5$ 96 3.5 64 3.5 96 Detection A1 : Ca(NO&* frequency/Hz SNR 52 45 38 44 50 40 55 42 3.6 4.1 6.3 7.3 6.6 8.8 6.2 7.1 Mn : NaClt SNR Average Minimum 2.6 2.8 3.6 3.0 4.0 4.1 2.9 3.9 1.6 2.1 2.4 2.2 3.2 3.5 1.9 2.5 6-mm plate, displacement f 1.5 V- 2.2 3.4 3.7 3.4 3.1 3.2 3.6 4.2 2.7 - - 4.5 32 67 4.5 64 50 4.5 96 40 4.0 96 41 3.5 96 43 3.0 96 45 2.5 96 47 2.04 96 49 1.5 96 51 - - - - - - - - - - - - - - - - * Peak height for 80 pg A1 versus peak to peak noise from 40 pg Ca(N03), at 396.15 nm.Tube wall temperature 2400 K, configuration (b). t Peak height for lC0 pg Mn versus peak to peak noise from 50 pg NaCl at 403.08 nm. Tube wall temperature 2930 K (3-mm plate) and 2780 K (6-mm plate), configuration ( d ) . Average SNR based on averaged noise over 2-s measurement, and minimum SNR based on noise under maximum background peak (“worst case”). $ Used for all experiments except determination of reference materials. § Used for determination of reference materials. 4 0 I n I 1 I I I I I 5 6 7 6 7 Time/s Fig.5. GFAES signals for 80 pg A1 in 40 pg Ca(N03)2 using (a) non-optimised and (b) optimised data acquisition parameters. 1, Background emission; 2, Ca(N03)2 blank; and 3, specific emission. Wavelength 396.15 nm; excitation temperature 2400 K in atomiser configuration (b)110 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 condensation, molecular formation and scattering effects in the cooler extremities of the furnace. This is equally true of systems utilising probe atomi~ation,~ an alternative modifica- tion of the Massmann-type furnace. In relation to the GFAES technique, a thermal gradient along the furnace length presents an additional drawback, as the presence of a Table 5. Vapour-phase temperatures experienced by thermometric species of different volatilities in three atomiser configurations. Nickel and lead line pairs in absorption and iron in emission using peak-area evaluation Spectroscopic temperature/K Atomiser Wall configuration temperature/K* Fe Ni Pb 2590 2570 2620 2200 2980 - 2300 2930 2440 2430 2450 2190 2640 2610 2420 2190 2530 2510 2510 2490 (a) ( b ) (4 * Measured at injection port using optical pyrometer.-4 - 0 25 50 Tirne/rns Fig. 6. Oscilloscope tracing of modulation waveform, using parameters given in Table 4, row 6 4 0 significant amount of free atoms in the cooler parts of the furnace will result in self-absorption .97*9 With the design of the constant-temperature atomiser used in this work it was possible to investigate the effects of furnace dimensions and temperature distribution on various aspects of GFAES.The vapour-phase excitation temperatures, measured using iron,18 nickel17 and lead16 as thermometric species, are listed in Table 5 for the atomiser configurations (a)-(c) shown in Fig. 3. Preliminary experiments showed that configurations (c) and ( d ) were very similar with respect to vapour-phase temperature and so only the results for the former are reported. Tube configurations (a) and (c) were heated uniformly along their entire lengths, as determined by optical pyrometer measurements, within ca. k 50 K. Configu- ration ( b ) , on the other hand, displays a marked temperature gradient comparable to that reported for Massmann-type furnaces,8 with the tube ends being 700-800 K cooler than the 2600 K final temperature at the injection port.The vapour- phase temperature measurements are, however, of greater significance for GFAES. Table 5 shows that in all instances the iron spectroscopic temperature agrees very closely with the measured outside wall temperature. This follows from the fact that the observed emission intensity is dominated by the contribution from atoms experiencing the highest tempera- ture, with correspondingly lower contributions from emitting atoms in lower temperature regions.9318 The lead spectro- scopic temperatures measured in configurations (a) and (b) are probably low due to the volatility of this element. It should be observed that even in the two-step atomiser the mean tube temperature will be somewhat dependent on the temperature of the cup through development of a temperature gradient between tube and cup.This gradient will be more severe at higher temperatures and for volatile elements, see also Table 5 . However, nickel is a non-volatile element, and no tempera- ture gradient problems were encountered using this element 1 1 I 1 1 I I 3 4 3 4 Time/s Fig. 7. Effect of modulation frequency on background correction capability using (a) 3-mm plate at 40 Hz and (6) 6-mm plate at 49 Hz. 1, Background and 2, corrected emission from 50 pg NaCl. Wavelength 403.08 nm; excitation temperature 2780 K in atomiser configuration (d) Table 6. Peak-area and peak-height detection limits at a tube-wall temperature of 2580 K for different tube configurations Peak-area detection limit*/pg Peak-height detection limit*/pg Element A1 .. . . . . Cr . . . . . , c u . . . . . . Ga . . . . . . In . . . . . . K . . . . . . Mn . . . . . . Pb . . . . . . Unm 396.15 425.43 324.75 403.29 410.17 404.41 403.08 405.78 ~ (a) 9.7 11 200 37 260 15 1300 7.2 ( b ) 3.8 2.8 5.4 6.3 6.9 25 110 400 ~~ - (c> 1.1 2.3 1.7 2.4 3.2 15 63 240 (4 1 .o 1.8 6.0 - - - 2.6 160t (a> 9.8 15 120 31 54 530 37 1820 ( b ) 2.7 4.6 44 15 11 130 410 9.0 (c> (4 1.4 1.6 2.9 1.8 14 5.6 4.0 - 4.6 - 82 - 3.9 3.4 230 2003. * Detection limit defined as three times the standard deviation of analyte signal for a concentration 5-10 times the detection limit t 283.33-nm analytical line. (n 3 6).JOURNAL O F ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 111 The detection limit data shown in Table 6 are consistent with the observations made concerning the temperature characteristics of the various tube configurations.All measurements were made with the same tube surface temper- ature, and so the detection limits should be better in configurations (c) and ( d ) by virtue of the isothermality of phase temperatures over the tube lengths. The benefits of a constant-temperature measurement zone are evident by comparing configurations (b) and ( c ) as both have 18 mm long tubes. The advantages of increasing the geometric length of the furance are also apparent, as configuration (a) using a short 8-mm tube gives the poorest detection limits of all. The detection limit data for configuration (d) is not complete, as the major purpose of the atomiser design was to investigate higher temperature excitation conditions.It might at first seem surprising that the detection limits for (c) and (d) are similar, considering the larger internal diameter of tube (d) [5.0 mm as compared with 4.0 mm in tube (c)]. Although the generated atomic vapour is more dispersed as a result of th’e larger cross-sectional area of tube (d), greater optical separ- tion of the tube wall radiation at the monochromator entrance slit is possible.22 Consequently, the SNRs were about the same limits are slightly better than those obtained using peak-height 2 - ? .c 1 c Q) - .- 1 B o C these tubes, and the correspondingly higher mean vapour- - 1 Log(mass Cu/ng) Fig. 8. Copper eak-height calibration graphs using the atomiser confi urations in Fig. 3.Wavelength 324.75 nm; excitation tempera- ture k80 K [for configuration ( d ) slightly higher] 2 u) in the two tube types. Generally, the peak-area detection measurements. In addition , peak-area evaluation is more widely applicable to practical analytical situations, owing to $ 1 c Q) - .- -I g o the peak-height method.30 u) c c .- the greater immunity from interference effects compared with Figs. 8 and 9 show calibration graphs for copper, using peak-height and peak-area evaluation, respectively. For all tube configurations, the peak-area graphs provide longer useful calibration intervals, as peak-height values are limited by the pronounced effect of self-absorption. In all the calibration graph comparisons, the onset of curvature occurs absorption at the cooler ends of this tube.It can thus be concluded that for analytical applications, covering a concen- tration interval of about four orders of magnitude, the combination of peak-area measurements and a spatially isothermal atomiser [configurations (c) and (d)] is most suitable. At higher analyte concentrations, GFAES suffers from emission line broadening. With increasing concentration the emission signal rises to a maximum intensity when the atomic density is so great as to approximate a blackbody emitter at that wavelength.31 Beyond this point, the wavelength inte- grated intensity increases due to emission over a wider wavelength interval, i.e. , line broadening. At such concentra- tions in the graphite furnace, self-absorption is also signifi- cant,9J2 so that only radiation in the wings of the broadened emission line is totally transmitted, and the profile takes on a self-reversed shape.This phenomenon is illustrated by the three-dimensional plot shown in Fig. 10. At the point of maximum atom concentration throughout the furnace, self- absorption is so severe that no intensity is transmitted at the line centre, as shown by the dip in the middle of the plot. The corresponding intensity - time signal recorded for the same amount of copper at the line centre is shown in Fig. ll(a). broadening becomes so severe that the profile is wider than the modulation interval. When this happens, the background measurements are taken in the wings of the analyte emission line profile and so background correction errors are made. As the intensity at these points is greater than at the line centre, where self-absorption is then complete, the background corrected intensity becomes negative9.32 [see Fig.11(6)]. To avoid correction errors, the modulation interval must be increased, but this in turn degrades the SNR at the lower concentration region (as for a certain frequency, less data acquisition time is then spent at the line centre) and reduces - 1 Log(mass Cuhg) Fig* 9* Copper Peak-area calibration graphs. Conditions as given in earliest in configuration (61, because of significant self- Fig. 8 Fig. 10. Three-dimensional plot of emission signal from 50 ng of When even greater sample amounts are atomised, line copper using tube configuration ( c ) . Conditions as given in Fig. 8. Every second scan plotted as a thermometric species.As peak-area evaluation was used, the reported nickel spectroscopic temperatures should reflect the vapour-phase conditions experienced by the atomic species throughout the duration of the signal pulse. As expected, configurations (a) and (c) display nickel vapour- phase temperatures close to those measured by the optical pyrometer, whereas in configuration ( b ) the mean vapour temperature is lower due to the temperature gradient along the tube length.112 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 :: - . * ...................................... ? > v) C c c .- c - I I I I I 1 Time/s 4 7 10 0 5 10 Fig. 11. GFAES signals for (a) 50 n and (b) 20 pg of copper using tube configuration (c). 1, Background emission; and 2, specific emission.In (b), the cup (3) and tube (47 temperature profiles are shown. Conditions as given in Fig. 8 Table 7. Comparison of detection limits Detection limit/pg Peak height Peak area Element Unm Te,,/K* Best reported? CTFS Wall§ CTFS ZAASI Ag . . . . 328.07 2670t t 2.6 2.2 - 1.9 0.5 co . . . . 345.35 2580 80 74 - 39 2.0 Cr . . . . 425.43 2800tt 0.60 1.7 2.8 0.9 1 .o 1.7 4.6 7.4 1.8 1 .o - 10 2.0 Cu . . . . 324.75 28Mt t Fe , . . . 371.99 2580 9.0 18 Ga . . . . 403.29 2580 2.0 4.0 - 1.7 2580 0.3" 4.6 - 2.4 - In . . . . 410.17 K . . . . 404.41 2580 100 80 Mg . . . . 285.21 2580 15 11 - 6.1 0.5 Mn . . . . 403.08 28Wt t 0.40 2.0 2.8 1.1 1 .o Na . . . . 330.23 2580 180 37 Ni . . . . 341.48 28Wt t 105 41 36 18 10 230 120 1200 230 5 .O - 80 - 110 - Pb , .. . 405.78 2800 t t - 25 10 283.33 2 8 W t TI . . . . 377.57 26207 t 506 30 V . . . . 437.92 2580 75 25 - 30 20 * Tex, = excitation temperature used for CTF measurements. t Best reported GFAES peak-height detection limits (SNR = 1) from reference 2 for platform atomisation, except a probe (reference 7) and 4 Detection limits defined as in Table 6. 0 Tube wall atomisation values from reference 14. 7 Zeeman atomic absorption spectrometry values (2a criterion) from reference 34. 77 Furnace configuration (d), all others using configuration (c). A1 . . . . 396.15 2580 1.5 1.4 - 1 .o 4.0 - - 65 - - 51 - b tube wall (reference 33) atomisation. Table 8. Analysis of NBS Standard Reference Materials by GFAES, using optimised data acquisition parameters, configuration (d), and the 6-mm quartz refractor plate Concentration/pg g-1 Peak height Peak area Aqueous Standard Element calibration additions Ag .. . . 0.08f0.02 0.08 * 0.02 Cr . . . . 0.19f0.04 0.24 f 0.05 Cu . . . . 170f20 220 f 25 Mn . . . . 11.3f 1.4 12.6 f 1.6 Ni . . . . 0.25 fO.06 0.16 f 0.04 Ag . . . . 0.45f0.11 0.72 +_ 0.18 Cr . . . . 0.6OkO.20 0.65 f 0.22 c u . . . . 68 f 13 144 f 28 Mn . . . . 11.2f3.0 12.2 f 3.2 Ni . . . . 2.1 f 0 . 7 1.9 f 0.7 Bovine liver- Oyster tissue- * Information value. t Not certified, value from reference 1. Aqueous Standard calibration additions 0.08 f 0.01 0.12 f 0.02 201 k 15 12.1 f 0.8 0.22 f 0.02 0.08 k 0.01 0.11 f 0.02 200 f 15 11.8 k 0.8 0.21 k 0.02 0.65 k 0.09 0.71 f 0.08 67 k 10 15.6 * 1.9 1.2 k 0.2 0.68 f 0.09 0.62 f 0.07 68 f 10 16 f 1.9 1.1 f 0.2 Certificate (0.06) * (0.088 f 0.012)t 193 f 10 10.3 f 1.0 - 0.89 4 0.09 0.69 4 0.27 63 f 3.5 17.5 f 1.2 1.03 f 0.19JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 113 the advantages of performing background correction in close proximity to the analytical line. For optimum performance, it is therefore essential to adapt the modulation interval to the concentration range of interest. Also evident in Fig. ll(b) is the problem of removing all analyte from the atomiser within one heating cycle to avoid carry-over contamination. System Performance TO assess the performance of the present system, the detection limits for a number of elements of varying volatilities and excitation potentials were measured, and these are compared with previously reported GFAES297J4-33 and Zeeman-effect GFAAS (ZAAS)34 values in Table 7.Lead detection limits were measured at two lines having excitation potentials of 4.38 eV, the better performance at the shorter wavelength being due to a much lower background and hence reduced noise level. It should be observed that the values reported for measurements made in tube configuration (c) at 2580 K are not necessarily the optimum values, as no investigations were made using higher excitation temperatures for these elements. Of more importance than detection limits in aqueous solutions is the performance of the system with respect to real samples. Table 8 gives the results of GFAES analysis of the NBS Standard Reference Materials, bovine liver and oyster tissue.Except for silver in oyster tissue, acceptable agreement is demonstrated between GFAES and certified values using peak-area evaluation, even when the standardisation was made against aqueous solutions. It should be stressed that the conditions for correct area evaluation are fulfilled for the constant-temperature furnace used here. Considering peak- height evaluation, erroneous results were frequently obtained regardless of standardisation mode. It should also be noted that the peak height in contrast to the peak area, is a function of the rate of atom formation.4 This rate is strongly matrix dependent, which explains why relatively good agreement with certified values can be obtained only occasionally with peak-height evaluation. Conclusions In this paper we have established the optimum conditions for wavelength modulated background corrected GFAES with respect to data acquisition parameters as well as furnace dimensions, heating mode and excitation temperatures.The best furnace configuration was that providing spatial as well as temporal isothermality. Using this furnace, the optimum excitation temperatures were found to be within a very narrow range for all elements investigated. This means that the system easily lends itself to multi-element determinations, as opti- mum conditions can be used for most elements. As has been shown, detection limits in aqueous solutions are only slightly better than those obtained by other GFAES systems using platform or probe volatilisation. This indicates that existing GFAES systems are fairly well optimised, at least with respect to aqueous solutions.A comparison with respect to real samples has not been carried out, as results for such samples close to the detection limit are not available from other systems. In spite of the fact that the detection limits for some elements of interest are poorer than those attainable with ZAAS (see Table 7), GFAES is a potential multi-element technique and, in addition, provides an alternative method for trace element determinations. This work was supported by the Swedish Natural Science Research Council. The authors are indebted to Bruno Hutsch, Ringsdorff- Werke, FRG, for pyrocoating the graphite parts of the constant-temperature furnace. We are also grateful to Jerker Ohman, who carried out the simplex optimisation calcula- tions.1. 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. References Marshall, J., Littlejohn, D., Ottaway, J. M., Harnly, J. M., Miller-Ihli, N. J., and O’Haver, T. C., Analyst, 1983, 108, 178. Bezur, L., Marshall, J., Ottaway, J. M., and Fakhrul-Aldeen, R., Analyst, 1983, 108, 553. Walsh, A., Anal. Chem., 1974,46, 698A. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” Hilger, London, 1970. Ottaway, J. M., Hutton, R. C., Littlejohn, D., and Shaw, F., Wks. 2. Karl-Marx-Univ. Leipzig, 1979, 28, 357. Baxter, D. C., Frech, W., and Lundberg, E., Analyst, 1985, 110, 475. Giri, S. K., Littlejohn, D., and Ottaway, J. M., Analyst, 1982, 107, 1095. Falk, H., Glismann, A., Bergann, L., Minkwitz, G., Schubert, M., and Skole, J., Spectrochim. Acta, Part B , 1985, 40,533. Marshall, J., Littlejohn, D., Ottaway, J. M., Miller-Ihli, N. J., O’Haver, T. C., and Harnly, J. M., Spectrochim. Acta, Part B, 1984,39, 321. Lawson, S. R., Dewalt, F. G., and Woodriff, R., Prog. Anal. At. Spectrosc., 1983, 6 , 1. Frech, W., Cedergren, A., Lundberg, E., and Siemer, D. D., Spectrochim. Acta, Part B, 1983, 38, 1435. Lundberg, E., and Frech, W., Anal. Chem., 1981,53, 1437. Frech, W., Zhou, N. G., and Lundberg, E., Spectrochim. Acta, Part B, 1982, 37, 691. O’Haver, T. C., Harnly, J. M., Marshall, J., Carroll, J., Littlejohn, D., and Ottaway, J. M., Analyst, 1985, 110, 451. Lundgren, G., Lundmark, L., and Johansson, G., Anal. Chem., 1974,46, 1028. Siemer, D. D., Lundberg, E., and Frech, W., Appl. Spectrosc., 1984, 38, 389. van den Broek, W. M. G. T., de Galan, L., Matousek, J. P., and Czobik, E. J., Anal. Chim. Acta, 1978, 100, 121. Littlejohn, D., and Ottaway, J. M., Analyst, 1978, 103, 595. Harnly, J. M., O’Haver, T. C., Golden, B., and Wolf, W. R., Anal. Chem., 1979, 51, 2007. Frech, W., Cedergren, A., Cederberg, C., and Vessman, J., Clin. Chem., 1982,28,2259. Ottaway, J. M., and Shaw, F., Appl. Spectrosc., 1977, 31, 12. Littlejohn, D., and Ottaway, J. M., Analyst, 1977, 102, 553. Littlejohn, D., and Ottaway, J. M., Anal. Chim. Acta, 1979, 107, 139. Epstein, M. S., and O’Haver, T. C., Spectrochim. Acta, Part B, 1975, 30, 135. Bezur, L., Marshall, J., and Ottaway, J. M., Spectrochim. Acta, Part B, 1984,39,787. Zander, A. T., O’Haver, T. C., and Keliher, P. N., Anal. Chem., 1976,48, 1166. Harnly, J. M., Anal. Chem., 1982, 54, 876. L’vov, B. V., Spectrochim. Acta, Part B, 1978,33, 153. Epstein, M. S., Rains, T. C., and O’Haver, T. C., Appl. Spectrosc., 1976,30, 324. Frech, W., Lundberg, E., and Cedergren, A., Prog. Anal. At. Spectrosc., 1985,8,257. Winefordner, J. D., McGee, W. W., Mansfield, J. M., Parsons, M. L., and Zacha, K. E., Anal. Chim. Acta, 1966,36, 25. Epstein, M. S., Rains, T. C., Brady, T. J., Moody, J. R., and Barnes, I. L., Anal. Chem., 1978, 50, 874. Ottaway, J. M., and Hutton, R. C., Analyst, 1976, 101,683. Slavin, W., Carnrick, G. R., Manning, D. C., and Pruszkow- ska, E., At. Spectrosc., 1983, 4, 69. Paper J5iS2 Received November lst, 1985 Accepted November 18th, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100105

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 115 Hydride Generation Atomic Absorption Determination of Arsenic in Marine Sediments, Tissues and Sea Water with ln Situ Concentration in a Graphite Furnace* Ralph E. Sturgeon,t Scot N. Willie and Shier S. Berman Analytical Chemistry Section, Division of Chemistry, National Research Council of Canada, Ottawa, K7A OR9, Canada Methods are described for the determination of total As in sea water and in marine tissues and sediments based on the generation of AsH3 using NaBH4 with its subsequent trapping in a graphite furnace at 600 "C. Calibration is achieved with a simple aqueous calibration graph having a sensitivity of 0.18 k 0.03A ng-1 and yields an absolute detection limit of 40 pg. Sample volumes of 10-500 pl produce total method concentration detection limits (3a blank) of 60 pg 9-1 in sea water, 0.2 pg g-1 in sediments and 0.15 pg 9-1 in tissues.Corresponding precisions of 2-3% are typical for analyses of these samples. Results are reported for the determination of As in a suite of marine reference materials. Keywords : Arsenic determination; marine sediments; sea water; h ydride generation atomic absorption spectrometry; graphite furnace trapping Brooks et al. 1 have recently reviewed progress in instrumental methods used for quantitative measurements of As. The most popular means of determining this element is atomic absorp- tion spectrometry (AAS) , with hydride generation methods being preferred at low concentrations.2.3 However, the relatively large sample volume required for each measure- ment is a major disadvantage for the analyses of some (biological) materials.Efforts to improve sensitivity and detection limits have generally made use of batch hydride generation - collection techniques employing cryogenic con- densation@ or trapping in balloons9 or other devices,10.11 followed by rapid introduction of the arsine, as a discrete plug of vapour, into the atomisation cell. This approach, whilst eliminating the problems encountered with direct transfer of arsine due to its low rate of evolution,lo tends to be slow and complex for routine use. Although the combination of trapping the arsine and releasing it into a pre-heated furnace is attractive,5.12 the full potential of this arrangement was not realised until Drasch et a1.13 proposed that the furnace serve as both the hydride trapping medium and atomisation cell.Experimental implementation of this for the determination of As in biological tissues was rather cumbersome and no data were given.13 We have recently reported the use of a similar arrangement for the hydride generation graphite furnace AAS (GFAAS) determination of Sb in sea water14 and Se(1V) in sea water and total Se in marine sediments and tissues.15 The high efficiency with which these elements could be trapped in the furnace and subsequently atomised permitted in situ pre-concentration of Se, for example, from as much as 50 ml of sea water.15 An ,increase in relative detection power of 2500-fold over a conventional 2 0 4 aqueous sample could thus be obtained, permitting routine determinations of Se(1V) at the 20 pg g-1 level.This methodology offers substantial advantages over con- ventional furnace or hydride generation techniques including superior detection limits, simplicity of operation and the use of small sample volumes. We report here on the application of in situ metal trapping to the determination of total As in sea water and in marine biological tissues and sediments by hydride generation GFAAS. * NRCC No. 25284. Experimental Apparatus Atomic absorption measurements were made using a Perkin- Elmer Model 5000 spectrometer fitted with a Model HGA-500 graphite furnace and a Zeeman effect background correction system. Peak absorbance signals were recorded with a Perkin-Elmer PRS-10 printer - sequencer. An As electrode- less discharge lamp (Perkin-Elmer) operated at 8 W was used as the source.Absorption was measured at the 193.7-nm line with the spectral band pass set at 0.7 nm, Perkin-Elmer pyrolytic graphite-coated tubes were used with minor modification. The coating of pyrolytic graphite was removed from the interior surface of the tube using sandpaper wrapped on a glass rod. Alternatively, old, well used graphite tubes were used. Arsine was generated in a similar, but smaller (ca. 10-ml volume) reaction cell than that described earlier14J5 for use with Sb and Se. The internal purge gas supply line to the furnace was routed through a FTFE stopcock, which permit- ted the operator to select gas flow into either the bottom of the hydride cell or into the furnace, as shown in Fig.1. With this arrangement, an Ar flow could be used to strip the generated hydride from solution and carry it out the top of the cell where it was directed, via a 1 mm i.d. x 1.5 mm 0.d. quartz tube, into the sample introduction port of a pre-heated furnace. Sodium tetrahydroborate(II1) solution was pumped into the cell using a rack-mounted Ismatic peristaltic pump (Cole Parmer Instrument Co.). Power supply cel I t TO whom correspondence should be addressed. Fig. 1. Schematic diagram of gas distribution system116 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Table 1. Furnace programme Operation Programme Temperature/ Internal purge/ No. Step "C Time/s ml min- Generation - collection . . 1 1 600 10 0 2 600 15 0 3 600 60 100 Atomisation .. . . . . 2 1 2600 4 0 2 2700 2 300 Reagents As standard solutions. Stock solutions (1000 mg 1-l) of As(II1) and As(V) were prepared by dissolution of As203 and Na2HAs04 in dilute HCl and HN03, respectively, and standardised by conventional GFAAS procedures. Working standards of lower concentration were prepared by serial dilution of the stocks using distilled, de-ionised water (DDW). Acids. High purity sub-boiling distilled HCl, HN03 and HC104, prepared in-house, were used for sample decomposi- tions. Sodium tetrahydroborate(III) solution, 1% mlV. Prepared daily, or more frequently if required, by dissolution of 0.25-g pellets of NaBH4 (Alfa Inorganics) in DDW. No additional stabiliser (i.e., NaOH) was used.16 Samples Several marine samples were analysed for As, including National Research Council of Canada (NRCC) open ocean sea water (NASS-l), NRCC near shore sea water (CASS-l), NRCC marine sediment (BCSS-1) , NRCC lobster hepatopan- creas (TORT-1) and NBS oyster tissue (SRM 1566).Procedure All sample and analytical manipulations were conducted in a routine laboratory environment. Aliquots of NASS-1 and CASS-1 (500 pl) were transferred directly into the hydride cell containing 2 ml of 0.5 M HC1. Nominal 0.5-g samples of BCSS-1 sediment were decom- posed by mixed acid digestion in a PTFE bomb according to the procedure described by Siu and Berman17 and diluted to 50 ml using 1 M HC1. Total As was determined using lo-@ aliquots of these solutions delivered into the hydride cell containing 2 ml of 0.5 M HC1.Nominal 0.5-g samples of TORT-1 and SRM 1566 were dry ashed according to the procedure described by Siu and Berman18 using Mg(N03)2 as an ashing aid. The sample was diluted to 50 ml with 1 M HC1. Total As was determined using 10-pl aliquots of these solutions delivered into the hydride cell containing 2 ml of 0.5 M HCl. Reagent blanks were processed through identical steps for decompositions of sediment and biological tissues. The sequence of operations describing AsH3 generation, collection and atomisation is outlined below. During AsH3 collection the stopcock was closed to direct internal purge gas through the hydride cell and the furnace was pre-heated at 600 "C for 10 s. The NaBH4 solution was then pumped into the cell for 15 s at a rate of 4 ml min-1, during which time the AsH3 was swept, via the generated stream of hydrogen (ca.50 ml min-1 under these conditions) into the furnace, where it was trapped. Internal purge gas flow was automatically initiated at the end of the NaBH4 addition and the cell purged for 60 s at a setting of 100 ml min-1 of Ar. At the end of this cycle, thermal programming of the furnace was terminated, the quartz transfer line removed from the sample introduction port and the stopcock opened to permit internal purge gas to flush the furnace. The sample was then atomised at 2600 "C using maximum power heating and internal gas stop, followed by a cleaning cycle at 2700 "C with a 300 ml min-1 internal purge gas flow. The furnace programme is shown in Table 1. Internal purge gas was again diverted through the hydride cell, which was emptied and rinsed with DDW.The NaBH4 solution was withdrawn from its injector tip by reversing the direction of the peristaltic pump. The next sample aliquot was then added to the cell and the measurement process repeated. Replicate measurements could be made every 3-4 min. Calibration graphs prepared from spikes of either As( 111) or As(V) added to 2 ml of 0.5 M HC1 were constructed and used for sample analysis. Results and Discussion The efficiency of generation and transfer of AsH3 from the hydride cell was not affected by the presence of active sites on the cell surface19 and it was therefore unnecessary to silylate the cell. In contrast to SeH2, arsine could be introduced into the pre-heated furnace using either the internal purge gas lines, as described for SbH3,14 or directly from the cell into the furnace via its sample introduction port.This relative stability of AsH3 was consistent with the need to utilise a graphite tube with a well developed surface area and/or surface concentration of active sites on to which the arsenic could deposit. Pyrolytic graphite-coated tubes decreased the trapping efficiency to 1&15%0 of that obtained with older, used tubes in which the coating of pyrolytic graphite was worn away. These observa- tions are consistent with the studies undertaken by Koreckova et al. ,20 which clearly demonstrated the enhanced tenacity with which As was bound to ordinary graphite compared with glassy carbon. The latter material is characterised by a comparatively low surface area and few active surface sites.The typical lifetime of a graphite tube exceeded 1200 firings and permitted "used" tubes, not suitable for other analyses, to be put to good service. The trapping efficiency of the tube increased from approxi- mately 25% of the optimum at 200 "C to 90% at 500 "C. The signal recovery rose to 100% at 600 "C and remained constant up to 900 "C, falling to 80% at 1000 "C. The retention of As in the furnace at these temperatures is comparable to that noted by Ediger21 for the charring of aqueous As samples in the absence of matrix modifiers. It is not clear, however, in which form the AsH3 is collected in the furnace. Atomisation temperatures above 2400 "C (maximum power heating) gave a constant, optimum response. Analytical Blanks Absolute blanks were assessed for each of the sample matrices.Although this was easily implemented for sediment and biological materials, for which a real reagent blank was processed with the samples, only a "synthetic" blank was available for sea water. For this purpose, 2 ml of DDW served as a "carrier" medium for the HC1 and NaBH4 reagents. Absolute blanks of 0.06 k 0.01 ng were obtained. It was verified that the DDW contributed insignificantly to theJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 117 measured As, the source being traced to the HCl. Reagent- grade HCl, used without clean-up, produced a blank of 0.14 ng. Addition of NaBH4 solution to dilute HCl, in an effort to sparge the acid free of As, reduced the blank to its stated, acceptable level.Absolute blanks for the analyses of the remaining samples were BCSS-1 250 k 18 ng, TORT-1 115 k 12 ng and SRM 1566 30 k 30 ng. Blank corrections to the analytical results were acceptable, contributing 6% to the sea water, 5% to the total As in BCSS-1,2% to the total As in TORT-1 and 0.5% to the total As in SRM 1566. Speciation Studies Arsenic commonly occurs in most environmental materials in the arsenite (+3) and arsenate ( + 5 ) oxidation states. However, several organoarsenic species have also been identified in natural waters and biological tissues, including monomethylarsonic acid (MMAA) , dimethylarsinic acid (DMAA), trimethylarsine oxide (TMAO) and possibly mono-, di- and trimethylar~ines.6~22-25 Utilising the method- ology described here, it is possible that the latter species, volatile arsines, if originally present in sea water samples, are swept from the cell and trapped in the furnace.The pH (ca. 0.3) used for hydride generation in this study is sufficiently low to promote the formation of the volatile arsines of all of the above-mentioned species.22 Specific studies to investigate the relative response from several As species were undertaken. Generation of AsH3 from spikes of As(II1) and As(V) produced identical signals. Sensitivity differences of 2040% have been previously reported26.27 for the generation of AsH3 from these different oxidation states, probably because of the use of lower sample acidities26 or the slower kinetics of AsH3 formation from As(V) as opposed to As(III).28 The latter effect is completely eliminated for generation systems utilising AsH3 trapping prior to atomisation,12 as described in this study.The responses to the generation of AsH3 from spikes of DMMA and MMAA (Pfaltsz and Bauer, Stanford, CT) were 84 and loo%, respectively, of that obtained from As(II1). Efficient generation and trapping of CH3AsH2 and (CH3)2AsH suggest that other, volatile arsenic species may also be sequestered. The methodology, as outlined, is Table 2. Analysis of sea water, based on generation of AsH3 from 5OO-pl sample volumes Concentrationhg ml-1 Run No. 1 2 3 4 5 6 7 8 Mean NASS-1 1.53 1 S O 1.56 1.61 1 SO 1.55 1.55 1.54 k 0.04 CASS-1 1.07 1.15 1.06 1.06 1.08 1.07 1.05 1.09 1.08 f 0.03 Table 3. Analytical results Sample Units Determined* Accepted value NASS-1 .. ng g- 1 1.54 -t 0.04 ( 7 ) 1.65 5 0.19 1.04 k 0.07 11.1 ? 1.4 24.6 f 2.2 NBS 1566 . . pgg-1 13.0 k 0.2 ( 5 ) 13.4 k 1.9 replicates shown in parentheses. 1.08 f 0.03 (8) . ’ ngg-: CASS-1 BCSS-I . . pgg- 11.2 k 0.2 (4) TORT-1. . . pg g-1 25.5 2 0.8 ( 5 ) * Precision expressed as standard deviation based on the number of therefore suitable for total As only. Speciation may be possible through selective pH control in the generation Organoarsenic compounds not reducible with NaBH4 have been detected in some marine organisms ( i . e . , arseno- betaine30). However, the hot, mixed acid decomposition procedures used in this study should result in the effective decomposition of all forms of arsenic.29 ce11.6.29 Analytical Results Analytical results are summarised in Tables 2 and 3.Table 2 presents blank-corrected data for the determination of total As in NASS-1 and CASS-1 sea water, illustrating the precision of replicate determination that can be achieved with this technique using such small sample volumes. Results for the determination of total As in all samples are summarised in Table 3. In all instances, calibration was achieved with a simple graph prepared by generating AsH3 from 2-ml aliquots of DDW spiked with As(1II). The accuracy of the method is evident from a comparison of the results with accepted or certified values for these materials. Interference Effects It is well known that the generation of arsine is susceptible to interferences from various elements,7J6731 in particular those which react with NaBH4 to yield volatile hydrides or the reduced element.Interferences from such species as the transition metals of Groups VIII and Ib and elements in Groups IVa and Va have been found not to depend on the analyte to interferent ratio but on the concentration of interfering element in the sample solution.32 The data in Table 3 substantiate the conclusion that no interferences are present in this study. Interference levels of Fe, Cu, Ni and Se selected for study were arbitrarily established by first identifying the “worst case” ratio of As to interferent that might be encountered with this suite of samples and subsequently examining the effect of an interfer- ent mass excess 50-fold greater than this ratio. In this manner, the influence of 200 pg of Fe, 6 pg of Cu, 2 pg of Ni and 2 pg of Se on the determination of 1 ng of As in 2 ml of 0.5 M HC1 was studied and found to have no effect on the signal.No attempt was made to establish the maximum interfer- ence concentration levels and, in view of this, the above levels are considered to be tolerable. The advantage of using the furnace to trap AsH3 is that it eliminates the effects of those interferences which cause variable rates of hydride evolution. Additionally, the range of interference-free determination can probably be significantly increased when the hydride genera- tion is undertaken in 5 M HC1 instead of in the 0.5 M acid used here.31 The methodology should therefore prove suitable for the determination of trace amounts of As in a wide variety of biological and geological materials.~~~~ ~ Tables 4. Figures of merit Sensitivity/ LOD*/ Precision?, Linear Matrix A ng-I ng g-’ Yo range/ng Sea water . . . . 0.18 k 0.03 0.006 3(25) 3 Sediments . . 0.18kO.03 220 2(50) 3 3 Biologicals . . 0.18 k 0.03 150 2(90) the procedural or method blank. the LCD at which the sample analyses were run. * Limit of detection defined as three times the standard deviation of t The numbers in parentheses are the concentration factors above118 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Analytical Figures of Merit Figures of merit are presented in Table 4. The absolute sensitivity, as determined from the slopes of calibration graphs obtained using a number of tubes over several months, is 0.18 k 0.03 A ng-1, ( i .e . , 24 pg per 0.0044 A), comparable to that obtained by direct injection of aqueous solutions. This suggests that the generation - trapping process is ca. 100% efficient. Calibrated detection limits, based on the variabilitry of the blank (3a), are 60 pg g-1 in sea water and ca. 200 ng g-1 in sediments and biological materials. These numbers are strictly “methodological” detection limits and were based on a linear extrapolation of the blank measurements made on 0.5-g samples up to 1 g as reported in Table 4. A detection limit of 37 pg was calculated from linear regression analysis of standard calibration graphs consisting of five points (each determined in triplicate). This figure represents a more realistic evaluation of the detection limit of the measurement process itself (i.e . , generation - trapping and atomisation) rather than the blank-dependent methodological value given above. The precision of determination is approximately 3% (rela- tive standard deviation) on determinations 50-fold above the detection limits. The linear working range spans two decades, extending to 3 ng. Higher analyte concentrations are accessible by simply working with smaller sample aliquots or by introducing a purge gas flow during atomisation. Conclusions The combination of hydride generation with subsequent trapping in the graphite furnace provides a rapid, simple, accurate and precise method for the determination of As in a variety of environmental samples. Utilisation of the furnace as an atomisation cell is advantageous in that a single system can be used for both hydride samples and conventional aqueous samples. References 1.2. Brooks, R. R., Ryan, D. E. and Zhang, H., Anal. Chim. Acta, 1981, 131, 1. Godden, R. G., and Thomerson, D. R., Analyst, 1980, 105, 1137. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. Robbins, W. B., and Caruso, J. A., Anal. Chem., 1979, 51, 889A. Maher, W. A., Anal. Chim. Acta, 1981, 126, 157. Knudson, E. J., and Christian, G. D., Anal. Lett., 1973, 6, 1039. Braman, R. S., and Foreback, C. C., Science, 1973,182,1247. Howard, A. G., and Arbab-Zaver, M. H., Analyst, 1981,106, 213. Crecelius, E. A., Anal. Chem., 1978,50, 826. Manning, D. C., At. Absorpt. Newsl., 1971, 10, 123. Chapman, J. F., and Dale, L.S., Anal. Chim. Acta, 1979,111, 137. Narasaki, H., Fresenius 2. Anal. Chem., 1985,321, 464. Uthus, E. O., Collings, M. E., Cornatzer, W. E., and Nielson, F. H., Anal. Chem., 1981, 53, 2221. Drasch, G., Meyer, L. V., and Kauert, G., Fresenius 2. Anal. Chem., 1980,304,141. Sturgeon, R. E., Willie, S. N., and Berman, S. S., Anal. Chem., 1985,57,2311. Willie, S . N., Sturgeon, R. E., and Berman, S. S., Anal. Chem., accepted for publication. Bye, R., Talanta, 1982,29,797. Siu, K. W. M., and Berman, S. S., Anal. Chem., 1983, 55, 1603. Siu, K. W. M., and Berman, S. S., Talanta, 1984, 31, 1010. Reamer, D. C., Veillon, C., and Tokousbalides, P. T., Anal. Chem., 1981,53,245. Koreckova, J., Frech, W., Lundberg, E., Persson, J.-A., and Cedergren, A., Anal. Chim. Acta, 1981, 130,267. Ediger, R. D., At. Absorpt. Newsl., 1975, 14, 127. Andreae, M. O., Anal. Chem., 1977,49, 820. Chapman, A. C., Analyst, 1926, 51, 548. Penrose, W. R., Conacher, H. B. S., Black, R., Meranger, J. C., Miles, W., Cunningham, H. M., and Squires, W. R., Environ. Health Perspect., 1977, 19, 53. Johnson, D. L., and Braman, R. S., Deep-sea Res., 1975,22, 503. Welz, B., and Melcher, M., Anal. Chim. Acta, 1981, 131, 17. Nakahara, T., Anal. Chim. Acta, 1981, 131, 73. Agemian, H., and Cheam, V., Anal. Chim. Acta, 1978, 101, 193. Crecelius, E. A., Environ. Health Perspect., 1977, 19, 147. Edmonds, J. S., Francesconi, K. A., Cannon, J. R., Raston, C. L., Skelton, B. W., and White, A. H., Tetrahedron. Lett., 1977, 18, 1543. Welz, B., and Melcher, M., Analyst, 1984, 109, 573. Welz, B., and Melcher, M., Wasser, 1984, 62, 137. Paper J.5145 Received October I7th, 1985 Accepted October 31st, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100115

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

119 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Determination of Arsenic, Selenium and Cadmium in Marine Biological Tissue Samples Using a Stabilised Temperature Platform Furnace and Comparing Deuterium Arc with Zeeman-effect Background Correction Atomic Absorption Spectrometry Bernhard Welz and Gerhard Schlemmer Department of Applied Research, Bodenseewerk Perkin-Elmer & Co. GmbH, 0-7770 Uberlingen, FRG Arsenic, selenium and cadmium are determined in three marine biological tissue samples from an inter-calibration exercise and NBS Standard Reference Material oyster tissue using graphite furnace atomic absorption spectrometry. Deuterium arc background correction is plagued by severe spectral interferences, which cause substantial signal distortion. Highly time-resolved signal display is a useful diagnostic tool for this type of interference.Zeeman-effect background correction appears to work satisfactorily for the routine determination of these elements in marine biological tissue samples. Calibration against acid standard solutions can be used for selenium and cadmium, when stabilised temperature platform furnace conditions are applied, but not for arsenic. Agreement with accepted average values of the inter-calibration exercise is generally good. Standard reference materials are only useful in part for methods development because their analytical behaviour is sometimes different from that of the samples investigated. Keywords: L Vov platform; marine biological tissue samples; spectral interferences; stabilised temperature platform furnace; Zeeman-effect and deuterium arc background correction Graphite furnace atomic absorption spectrometry is one of the most sensitive techniques for the determination of trace elements in a variety of matrices. However, the literature cites numerous difficulties and interferences that are observed when complex samples such as biological tissues are analysed. Besides the numerous “chemical” interferences, some work- ers report spectral interferences mainly in the determination of arsenic,’ cadmium2 and selenium1.3 when continuum- source background correction is used.Separation of the analyte element from the matrix is therefore frequently proposed using precipitation or extraction and back- extraction. Such procedures, however, can create new prob- l e m ~ , ~ and excessive sample handling can lead to analyte losses as well as to a positive systematic error due to contamination.5 Recent experience has shown that most non-spectral interferences described earlier disappear when using the stabilised temperature platform furnace (STPF) concept ,6 which includes the application of a matrix modifier, atomisa- tion from a L’vov platform with maximum power heating and with interrupted inert gas flow.It also includes an instrument with fast electronics and the use of integrated absorbance values for signal evaluation. A dramatic reduction of interfer- ences has been shown for a number of trace elements in synthetic matrices,7 biological materials,s fish tissues9 and in natural waters,lO using this STPF concept.Excessive background attenuation due to large amounts of matrix volatilised during atomisation of the analyte element and spectral interferences caused by erroneous background correction are further problems in the determination of trace elements in biological tissues. Zeeman-effect background correction11712 is the most effective technique currently available to overcome these limitations. A combination of the stabilised temperature platform furnace with Zeeman-effect background correction has been shown to allow interference- free determination of the most “difficult” elements in complex matrices. Selenium is determined in biological materials, l3 cadmium in sea water14 and urine,l5-17 as are several additional trace elements in biological18 and other materials.l9 Under optimised conditions, calibration can be carried out directly against acidified reference solutions containing only the same amount of matrix modifier. In the present work we have analysed three unknown reference materials, based on marine biological samples, that were used in a recent international inter-calibration exercise and compared our results with those obtained by other participants using a variety of procedures and techriiques.20 We have applied the STPF concept to the determination of arsenic, selenium and cadmium, and investigated to what extent Zeeman-effect background correction is necessary or helpful for this type of application. Highly time-resolved signal display was used as an additional diagnostic tool for this purpose.Part of the investigation was to determine “recom- mended conditions” that could be used for the reliable routine analysis of samples such as those supplied in the inter- calibration exercise. Also analysed was the NBS Standard Reference Material oyster tissue, to investigate to what extent reference materials can be used for the development of analytical methods that have to be applied to the wide variety of naturally occurring tissue samples. Experimental Apparatus A Perkin-Elmer Model 3030 atomic absorption spectrometer equipped with an HGA-500 graphite furnace and an AS-40 autosampler and a Perkin-Elmer Model Zeeman/3030 atomic absorption spectrometer equipped with an HGA-600 graphite furnace and an AS-60 autosampler were used throughout this work. Both instruments were equipped with graphics display, and highly time-resolved signals were printed on a PR-100 printer.Pyrolytically coated graphite tubes (Perkin-Elmer Part No. B010-9322) with an inserted L’vov platform (Perkin- Elmer Part No. B010-9324) were used exclusively. Unless otherwise stated, graphite furnace programmes and matrix modifiers were used according to the manufacturer’s recom- mended conditions.21 Electrodeless discharge lamps were used for arsenic and selenium, and a hollow-cathode lamp for cadmium. All lamps were operated under the conditions recommended by the manufacturer. A Perkin-Elmer Auto- clave-3 was used for pressure decomposition. The combustion in a stream of oxygen was carried out in a Trace-0-Mat (Kuerner Analysentechnik, Rosenheim, FRG), which has been described in detail elsewhere.**120 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Reagents and Calibration Standard stock solutions (1000 mg 1-1) for all elements were prepared from Titrisol concentrates (Merck, Darmstadt, FRG) by dilution with 0.2% mlV HN03.Calibration against acidified standard solutions was used for cadmium and selenium determinations. Except for arsenic, analyte addition was used for control purposes only during methods develop- ment. Aliquots of 10 pl of both the sample and the matrix modifier solutions were used for all determinations. The same amount of matrix modifier was added to the sample and standard solutions in the graphite tube. The matrix modifiers used in this work were: 10 pg of Ni as the nitrate for arsenic, 100 pg of NH4H2P04 for cadmium and a mixed modifier with 15 pg of Cu(N03)2 (corresponding to 5 pg of Cu) and 10 pg of Mg(NO& for selenium? Matrix modifier solutions were prepared from analytical-reagent grade salts.The NH4H2P04 modifier solution was purified by extraction with a solution of ammonium pyrrolidinedithiocar- bamate (APDC) in isobutyl methyl ketone (IBMK). Nitric acid was purified by sub-boiling distillation (sub-boiling still, Kuerner Analysentechnik, Rosenheim, FRG). Samples The samples examined were the reference materials used in the seventh inter-calibration exercise for marine biological tissues organised by the International Council for the Explora- tion of the Sea (ICES). The samples were prepared and distributed by the National Research Council, Canada.The samples are identified as lobster (spray-dried, acetone extrac- ted tomalley paste: lobster hepatopancreas), scallops (freeze- dried scallops aductor) and plaice (freeze-dried plaice). Forty eight laboratories reported results for cadmium, 22 for arsenic and only five for selenium. A t-test at the 95% confidence level was applied to the means of all values submitted. Means were successively rejected until a homogeneous set of results was obtained.20 The mean, and standard deviation of these remaining values, is referred to as the accepted mean in Tables 1 and 3. The mean and standard deviation of all submitted values, except for one obvious outlier, are used for selenium in Table 2. Also investigated was NBS Standard Reference Material No.1566 oyster tissue (US Department of Commerce, National Bureau of Standards, Gaithersburg, MD 20899, USA). Sample Treatment Determination of arsenic and selenium Pressure decomposition was used for the determination of arsenic and selenium. Approximately 1 g of sample, 10 ml of purified nitric acid and 10 ml of de-ionised water are placed in the PTFE beaker of the Autoclave-3. The system is closed tightly, heated to 140 "C on a hot-plate and kept at this temperature for 1 h. After cooling to room temperature, the Autoclave-3 is opened, the solution quantitatively transferred into a 50-ml calibrated flask and diluted to volume with de-ionised water. Determination of cadmium Combustion in oxygen was used for the determination of cadmium. Approximately 0.1-0.2 g of sample is placed in a small cup made of ash-free filter-paper (Schleicher & Schuell No.589'), which is fixed in the quartz sample holder of the Trace-0-Mat, the apparatus is closed and the cold finger filled with liquid nitrogen. The IR radiators are focused on the upper edge of the filter-paper cup to ensure slow combustion of the sample after ignition. A stream of oxygen (70 ml min-1) is started simultaneously with the ignition. The sample is completely ashed after about 1 min. The residual liquid nitrogen is removed from the cold finger, and the IR radiators focused on the nitric acid in the test-tube. After about 30 min of boiling under reflux, the apparatus is allowed to cool. The nitric acid solution is transferred into a 25-ml calibrated flask and diluted to volume with de-ionised water.Results and Discussion Arsenic The recommended conditions21 for the determination of arsenic are the addition of 10-20 pg of nickel matrix modifier as the nitrate, a thermal pre-treatment temperature of 1300 "C and an atomisation temperature of 2300 "C. The use of somewhat lower thermal pre-treatment temperatures has essentially no influence on the integrated absorbance signal of arsenic, but reduces the risk of pre-atomisation losses of arsenic in samples with high concentrations of total dissolved solids. A thermal pre-treatment of 1000 "C and an atomisation at 2200 "C was therefore preferred for routine applications. The NBS SRM oyster tissue has a high arsenic content (13.4 k 1.9 mg kg-l) and the decomposition solution could be diluted five-fold for the measurement.The highly time- resolved signals show very little background attenuation with a deuterium background corrected instrument (Fig. 1). Appearance time and shape of the absorbance signals are very similar for the oyster tissue sample and the reference solution, and there is no apparent peak distortion due to the matrix. This is confirmed by measurement with Zeeman-effect background correction, which results in the same peak shape, indicating the absence of spectral interferences. We were unable, however, to get rid of residual non-spectral interfer- ences in the determination of arsenic in oyster tissue. Consistently low values of 9.7 k 0.6 mg kg-1 were obtained with the standard calibration procedure, whereas the analyte addition technique resulted in 14.2 k 0.9 mg kg-1.The behaviour of the ICES reference samples was found to be distinctly different from that of the oyster tissue when deuterium arc background correction was used. Fig. 2 shows the highly time-resolved signals recorded for the lobster sample using thermal pre-treatment temperatures of 1000, 1100 and 1200 "C, respectively, and an atomisation tempera- ture of 2200 "C. A maximum background attenuation of 0.1415 A was observed under these conditions and this should present no problems for a deuterium arc background corrector, as it did not for the oyster tissue SRM. Neverthe- less, a signal depression and distortion was found for arsenic when thermal pre-treatment temperatures of less than 1200 "C were applied.It was shown in a later experiment with Zeeman-effect background correction that this signal depres- sion is due to a spectral interference from the matrix. The arsenic signal appears to be relatively free from 0 Timeis 3.0 Fig. 1. Highly time-resolved si nals for arsenic after deuterium arc background correction (solid line? and for the background attenuation (broken line). (a) Reference solution containing 0.5 ng As; and ( b ) oyster tissue sample solutionJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 121 L m o;AvJ .:.::. .. : ....... .?.., ,+. ~ ~ ~ / . . , ~ ; ~ : . , ~ . . . , i , , . . , I "L..: '..,?::.. ...... ........ . . .... .... : ........ 0 . . . . . . L I 0 4.0 Timeis Fig. 2. Determination of arsenic in lobster using deuterium arc background correction.Highly time-resolved signals for the analyte element after background correction (solid line) and for the back- ground attenuation (broken line), after thermal pre-treatment at (a) 10o0, (b) 1100 and (c) 1200 "C; atomisation temperature 2200 "C ............ ..:- ............. . , _ . . .,:..:-, .. ' .. :: 0.2 (a) a 1 ..... .. : : . . ..... I . . C m ......... ..... . . . . . . . . . . . . . . . . . ......... -. - 0 2.5 Tim eis ZAA 0.3, 0 C m 9 s n 4 0 0 3.0 Tirne/s Fig. 3. Determination of arsenic in plaice using (a) deuterium arc and ( b ) Zeeman-effect back round correction techniques (ZAA, Zeeman AA; BG, backgrouncfonly). Highly time-resolved signals for the analyte element after background correction (solid line) and for background only (broken line) interferences when a thermal pre-treatment of 1200 "C is applied, part of the arsenic is lost, however, when only a slightly higher temperature of 1300 "C is chosen.Selecting the proper temperature for thermal pre-treatment appears to be fairly critical for the determination of arsenic in this type of sample when deuterium arc background correction is used. While the same temperature programme could be used for the scallops reference sample, it was ineffective for the plaice reference sample. Fig. 3 shows a background signal of about 0.25 A and a background corrected signal for arsenic that is substantially distorted when a deuterium arc background corrector is used. The difference becomes clearly visible when the signal is compared with that obtained after Zeeman-effect background correction.The difference between the two background correction systems becomes even more apparent when the signals are evaluated quantitatively. It is impossible to use integrated absorbance values after deuterium arc background correction because the peak area depends very much upon the integra- tion time, and can even give negative values. In addition, if o.2 t O.l t 1 1 I V I I I I 1 40 20 0 20 40 60 80 100 Arsenic concentrationipg I-' $ 0.3 c (b) R 40 20 0 20 40 60 80 100 Arsenic concentrationipg 1-1 Fig. 4. Determination of arsenic in plaice. Peak-height values from deuterium arc background correction (a) (1.9 mg kg-l) are much lower than eak-area values from Zeeman-effect background correc- tion ( b ) (5.grng kg-I), both using the analyte addition technique.A, Calibration graph; and B , analyte addition graph peak-height evaluation and the analyte addition technique are used, as frequently practised in graphite furnace AAS, the results with deuterium background correction are lower by a factor of almost three compared with those obtained with Zeeman-effect background correction and evaluation of integrated absorbance values (Fig. 4). This demonstrates the obvious, but frequently ignored fact, that spectral interfer- ences cannot be compensated for by the analyte addition technique. The reason for this is that spectral interferences do not affect the slope of the calibration graph but cause a positive or, more frequently, a negative offset. It should also be stressed that such a peak-shape distortion due to a spectral interference is frequently obscured completely, or at least partly, in an ordinary recorder tracing and can only be detected with the help of highly time-resolved graphics.The spectral interference disappears when Zeeman-effect back- ground correction is used. The temperature dependence of the arsenic signal, as shown in Fig. 2 for the lobster sample also disappears, so that thermal pre-treatment between 1000 and 1200 "C can be selected without any influence on the signal and on the result. In spite of these substantial improvements, there is a residual non-spectral interference in the determination of arsenic in the marine biological materials investigated here, as already observed for the oyster tissue SRM.The analyte addition technique consistently gave results that were higher by 10-35% compared with those obtained by direct calibration against reference solutions. The characteristic mass was typically mo = 16 pg (0.0044 A s)-1 for the standard solutions and ca. mo = 18 pg (0.0044 A s)-1 for the sample solutions. The analyte addition technique was therefore applied to the determination of arsenic in all samples investigated. It should be stressed once again that the analyte addition technique is applied only after the spectral interferences are removed by the use of Zeeman-effect background correction to eliminate the residual non-spectral interferences. The results obtained in the three marine biological tissue samples are summarised in Table 1, and compared with the accepted mean values from the inter-calibration exercise20 and with those from hydride122 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 0.2 Table 1. Determination of arsenic in marine biological tissue samples (average of three independent decompositions per sample) using the analyte addition technique. All values are in mg kg-l Hydride 0 Lobster 25.0 f 5.0* 24.6 f 0.7 23.0 4 0.8 0.2 Scallops . . . . . . 7.1 f 2.1" 7.0 f 0.3 6.5 k 0.4 Plaice . . . . . . 4.641.6* 4.7k0.1 5.140.5 Accepted generation Sample mean AAS This work . . . . . . a, C 14.2 f 0.9 NBSoyster tissue . . 13.4 f 1.9t (D e * From inter-calibration exercise.20 v) Q ( t Certificate value. - D 0 ..... . . . . . . . . . . . . . 0 3.0 Timeis Fig. 6. Determination of selenium in three marine biological tissue sam les using deuterium arc background correction.Highly time- resoyved signals for the analyte element after back round correction (solid line) and for the background attenuation ?broken line). (a) Lobster; (6) scallops; and ( c ) plaice ZAA 0.2 I I I J 0 3.0 Time/s 0 Fig. 5. Determination of selenium in NBS Standard Reference Material oyster tissue. Highly time-resolved signals for: (a) reference solution containing 0.2 ng Se; (6) sample solution using deuterium arc background correction; and ( c ) sample solution using Zeeman-effect background correction. Solid line, analyte element absorbance after background correction; and broken line, background attenuation generation AAS.24 Table 1 also gives the value found for the oyster tissue SRM together with the certificate value.All results were obtained using Zeeman-effect background cor- rection and the analyte addition technique. Selenium The recommended conditions for selenium determination are a thermal pre-treatment temperature of 900 "C and an atomisation temperature of 2100 "C. The recommended matrix modifier is 15 pg of copper nitrate mixed with 10 pg of magnesium nitrate .21 All of the samples investigated show more or less severe signal distortion due to spectral interfer- ences when these conditions are applied and deuterium arc background correction is used. Fig. 5 shows the highly time-resolved signals for a standard solution and for the oyster tissue sample using deuterium arc and Zeeman-effect back- ground correction, respectively. A deflection of the signal to negative absorbance values is observed in spite of the relatively low background attenuation when deuterium arc background correction is used, which clearly indicates a spectral interference that disappears with Zeeman-effect background correction.3.0 Timeis Fig. 7. Determination of selenium in plaice using Zeeman-effect background correction. Highly time-resolved signals for the analyte element after background correction (solid line) and for background attenuation (broken line) Low values of 1.5 k 0.2 mg kg-1 are obtained for selenium in oyster tissue with standard calibration as well as with the analyte addition technique when deuterium arc background correction is used. With Zeeman-effect background correc- tion a value of 2.0 f 0.2 mg kg-1 is obtained with both calibration techniques, which compares very favourably with the certificate value of 2.1 k 0.5 mg kg-I.The severe double peaking that can be seen in Fig. 5 has apparently no influence on the accuracy of the determination when platform atomis- ation and peak-area integration are used. All the samples from the inter-calibration exercise exhibited more or less severe signal distortion and deflection to negative absorbance values when deuterium arc background correction was used (Fig. 6 ) . Varying thermal pre-treatment and/or atomisation temperatures, or changing to nickel as the matrix modifier, had no significant influence on the over-all situation. This means that deuterium arc background correction in essence cannot be used for the determination of selenium in the marine biological materials investigated here. Spectral interferences in the vicinity of the selenium resonance line are caused predominantly by phosphate.12 Molecular PO, bands exhibit a rotational fine structure, which almost inevitably leads to wrong results with a continuum source background corrector.123 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 No problems due to spectral interferences were observed Cadmium when the same samples were analysed using Zeeman-effect background correction. Fig. 7 shows the undistorted signal for selenium in the plaice sample obtained under these condi- tions. All determinations could be carried out directly against standard solutions that contained the same amount of matrix modifier (copper and magnesium nitrates) as the sample solutions.The characteristic mass was found to be around mo = 23 pg (0.0044 A s)-1 for sample and standard solutions. The results obtained for the three marine biological tissue samples are summarised in Table 2 and compared with the data from hydride generation AAS24 and the mean values from the inter-calibration. The latter are less useful for this element because only five participants submitted results for selenium and one of them is an obvious outlier.20 The agreement of the remaining four sets of data, using three different techniques, graphite furnace AAS, hydride generation AAS and flu- orimetry, however, is fairly good. The accuracy is further confirmed by the result obtained for the oyster tissue Standard Reference Material, as discussed earlier.Table 2. Determination of selenium in marine biological tissue samples (average of three independent decompositions per sample). All values are in mg kg-1 Inter- Hydride calibration generation Sample mean AAS This work Lobster . . . . . . 6.9k0.8 6 . 2 f 0 . 4 5 . 7 k 0 . 3 Scallops . . . . . . 0.6 2 0.3 0.71 k 0.08 0.73 f 0.04 Plaice . . . . . . 2 . 5 f 0 . 3 2.6k0.2 2.7f0.3 NBSoyster tissue . . 2.1 f 0.5* - 2.0 f 0.2 * Certificate value. ................................................................ 0.2 (a) 0.2 8 + ?J Q O m ........................................... L-----J .............................................. .... ... . . . . .... :': .... ...... . . ................................0.2 1 0 2.5 Timeis Fig. 8. Optimisation of the thermal pre-treatment and atomisation temperatures for the determination of cadmium in plaice using deuterium arc background correction. Highly time-resolved signals for the analyte element after background correction (solid line) and for backeround attenuation (broken line). Thermal re-treatment and atomisation temperatures: (a) 800 and 1700 "C, (by 900 and 1600 "C and (c) 900 and 1400 "C, respectively A thermal pre-treatment temperature of 700 "C, an atomis- ation temperature of 1600 "C and 0.2 mg of ammonium phosphate as the matrix modifier are recommended for the determination of cadmium.21 To avoid pre-atomisation losses of cadmium in the presence of high concentrations of total dissolved solids, the thermal pre-treatment temperature was lowered to 550 "C for the analysis of the oyster tissue sample.In spite of this relatively low thermal pre-treatment tempera- ture essentially no background attenuation was observed and the determination of cadmium could be carried out without interferences using deuterium arc or Zeeman-effect back- ground correction. A value of 3.63 * 0.05 mg kg-1 was obtained using the standard calibration method, which com- pares favourably with the certified value of 3.5 * 0.4 mg kg-1. No reasonable signals, however, could be obtained under these conditions for any of the marine biological tissue samples of the inter-calibration exercise when deuterium arc background correction was used. A background attenuation of almost 1 A, and severe signal distortion and over- compensation, was observed for the plaice sample.It is demonstrated in the highly time-resolved signals of Fig. 8 that an increase in the thermal pre-treatment temperature to 900 "C and a decrease of the atomisation temperature to 1400 "C reduces the spectral interferences substantially but cannot completely eliminate them. A thermal pre-treatment temper- ature of 900 "C for cadmium may also include some pre-atomisation losses of this element, so that it cannot be considered a "safe" temperature for routine investigations. Moreover, lengthy investigations for each individual type of sample to determine which temperature programme is best suited to minimise interferences is usually not acceptable in routine analysis. Similarly to arsenic and selenium, this problem disappears when Zeeman-effect background correction is used in the determination of cadmium.In addition, Zeeman-effect back- ground correction allows the recommended conditions, a thermal pre-treatment temperature of 700 "C and an atomisa- tion at 1600 "C, and the more effective ammonium phosphate - magnesium nitrate mixed matrix modifier to be used for all samples. The characteristic mass was typically ca. mo = 0.6 pg (0.0044 A s)-1 for samples and reference solutions, and all determinations could be carried out directly against references that contained the same amount of matrix modifier. The results obtained in the three marine biological tissue samples are summarised in Table 3. For all three reference samples, and for the oyster tissue SRM, the values obtained after combustion in a stream of oxygen and determination using Zeeman-effect background correction agree satisfac- torily with the accepted mean values from the inter-calibration exercise20 and with the certificate value, respectively.Conclusion Deuterium arc background correction 2 s not capable of compensating for the background attenuation found in the determination of arsenic, selenium and cadmium, at least in Table 3. Determination of cadmium in marine biological tissue samples (average of six independent decompositions). All values are in mg kg-* Accepted This work Sample mean Lobster . . . . . . 26+2* Scallops . . . . . . 0.75 f 0.10* Plaice . . . . . . . . 0.06 k 0.02* NBSoystertissue . . . . 3.5 f 0.4t * From inter-calibration exercise.20 t Certificate value.29 k 2 0.74 k 0.04 0.04 f 0.02 3.63 2 0.05124 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 some of the biological tissue samples investigated. Only References Zeeman-effect background correction guarantees that the results for these elements are not affected by spectral interferences. Standard reference materials such as the NBS oyster tissue sample may show an analytical behaviour that is substantially different from that of other samples. No spectral interferences were found for this SRM in the determination of arsenic and cadmium, whereas severe interferences occurred with reference samples used in an inter-calibration exercise. Standard Reference Materials, while undoubtedly ideal for use in quality assurance procedures,25 cannot therefore necessarily be used to decide if a certain procedure is acceptable for routine analysis of a wide variety of natural samples.Owing to the usually insufficient time resolution and too high time constant of ordinary recorders, the effect of peak distortion due to spectral interferences is completely, or at least partly, obscured when used for signal recording. Only the highly time-resolved display of electronically undistorted signals can be used as a diagnostic tool for this kind of interference. It is obvious, but frequently ignored, that the analyte addition technique cannot identify and/or correct for spectral interferences. Provided the analyte element added shows the same analytical behaviour as the analyte species present in the sample, the analyte addition technique can compensate only for those interferences that influence the sensitivity, i.e., the slope of the calibration graph. Spectral interferences, however, cause a positive or negative offset, i.e., add or subtract a constant absorbance value to each measurement in a certain matrix.This means that spectral interferences shift the calibration graphs without affecting their slope, resulting in too high or too low values for the analyte element even when calibration graphs are parallel with and without the interferent . Arsenic and cadmium could be determined in some of the samples using deuterium arc background correction after careful optimisation of the temperature programme. This procedure, however, does not appear to be acceptable for routine purposes where the same graphite furnace programme should be applicable for all samples.Only Zeeman-effect background correction really solves the problem because it eliminates the spectral interferences. This technique therefore allows a reliable determination of arsenic, selenium and cadmium in all samples. In addition, it gives more freedom in selecting thermal pre-treatment and atomisation temperatures as well as matrix modifiers. Zeeman-effect background correction, in combination with the STPF concept, substan- tially increases the reliability of graphite furnace atomic absorption spectrometry for routine applications. 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. Saeed, K., and Thomassen, Y., Anal.Chim. Acta, 1981,130, 281. Feitsma, K. G., Franke, J. P., and de Zeeuw, R. A., Analyst, 1984, 109,789. Manning, D. C., At. Absorpt. Newsl., 1978, 17, 107. N&ve, J., Hanocq, M., Molle, L., and Lefebvre, G., Analyst, 1982, 107, 934. de Goeij, J. J. M., Kosta, L., Byrne, A. R., and Kutera, J., Anal. Chim. Acta, 1983, 146, 161. Slavin, W., Manning, D. C., and Carnrick, G. R., At. Spectrosc., 1981, 2 , 137. Kaiser, M. L., Koirtyohann, S. R., Hinderberger, E. J., and Taylor, H. E., Spectrochim. Acta, Part B, 1981, 36, 773. Koirtyohann, S. R., Kaiser, M. L., and Hinderberger, E. J., J . Assoc. Off. Anal. Chem., 1982,65, 999. May, T. W., and Brumbaugh, W. G., Anal. Chem., 1982,54, 1032. Manning, D. C., and Slavin, W., Appl. Spectrosc., 1983,37,1. Fernandez, F. J . , Bohler, W., Beaty, M. M., and Barnett, W. B., At. Spectrosc., 1981, 2 , 73. Fernandez, F. J . , and Giddings, R., At. Spectrosc., 1982,3,61. Carnrick, G. R., Manning, D. C., and Slavin, W., Analyst, 1983, 108, 1297. Pruszkowska, E., Carnrick, G. R., and Slavin, W., Anal. Chem., 1983,55, 182. Pruszkowska, E., Carnrick, G. R., and Slavin, W., Clin. Chem., 1983,29,477. Dungs, K., and Neidhart, B., Analyst, 1984, 109, 877. Slavin, W., Manning, D. C., Carnrick, G. R., and Pruszkow- ska, E., Spectrochim. Acta, Part B, 1983,38, 1157. Voellkopf, U., and Grobenski, Z., At. Spectrosc., 1984,5,115. Slavin, W., Carnrick, G. R., Manning, D. C., and Pruszkow- ska, E., At. Spectrosc., 1983, 4, 69. Berman, S. S., “ICES Seventh Round Intercalibration for Trace Metals in Biological Tissue, ICES 7/TM/BT,” Prelimi- nary Report, 1984. Perkin-Elmer, “Analytical Methods for Furnace Atomic Absorption Spectrometry,” Perkin-Elmer, Uberlingen, Publi- cation B332, 1984. Knapp, G., Raptis, S., Kaiser, G., Toelg, G., Schramel, P., and Schreiber, B., Fresenius Z. Anal. Chem., 1981, 308, 97. Welz, B., Schlemmer, G., and Voellkopf, U., Spectrochim. Acta, Part B, 1984, 39, 501. Welz, B., and Melcher, M., Anal. Chem., 1985, 57, 427. Keith, L. H . , Crummett, W., Deegan, J., Libby, R. A., Taylor, J. K., and Wentler, G., Anal. Chem., 1983,55,2210. Paper J5l24 Received August 14th, 1985 Accepted September 23rd, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100119

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 125 Determination of Arsenic and Antimony by Hydride Generation Atomic Absorption Spectrometry Using a Small Hydride Generator Ping-kay Hon, Oi-wah Lau and Shu-ki Tsui Department of Chemistry, The Chinese University of Hong Kong, Shatin, N. T., Hong Kong A sensitive method for the determination of arsenic and antimony by hydride generation atomic absorption spectrometry, using a glass vial (internal volume, 8.6 ml) as the hydride generator, is described. The optimum flow-rate of the carrier gas was 0.2 I min-1 and the hydrochloric acid concentration used was 0.25 M. The absolute sensitivities were 0.06 and 0.10 ng for arsenic and antimony, respectively. The relative standard deviations were around 2% for both elements.A masking solution consisting of thiourea and ascorbic acid proved to be effective for suppressing the effect of a number of interferents. The accuracy of the proposed method was checked by determining the contents of arsenic and antimony in two standard reference materials. Keywords: Arsenic determination; antimony determination; h ydride generation atomic absorption spectrometry; small h ydride generator; standard reference materials Chemical generation of volatile hydrides with subsequent atomisation is a commonly used method for the determination of elements such as arsenic, bismuth, germanium, selenium, antimony, tin and tellurium. Chu ef al.2 first reported the use of an electrically heated tube as an alternative to a flame for the determination of arsenic using hydride generation atomic absorption spectrometry.Thompson and Thomerson3 used sodium tetrahydroborate(II1) as a reducing agent coupled with a flame heated silica tube as a detection device. Similar systems have been described by various workers.4’ In the work reported to date, very little consideration has been given to the effect of the size of the hydride generator on the sensitivity attainable, and the hydride generators des- cribed have usually been quite large (50-60 ml). It is believed that a smaller hydride generator would result in higher sensitivities for the determination of hydride-forming ele- ments as the hydride generated will be diluted less. The purpose of this work was to develop an atomic absorption method to determine arsenic and antimony after hydride generation using an electrically heated silica atomising tube and a small hydride generator.A capped glass vial with an internal volume of 8.6 ml was chosen as the hydride generator. The optimum conditions for the determinations, and the sensitivities, precision and accuracy of the proposed method are reported. The interferences from some foreign ions were also studied, and the effectiveness of a mixture of thiourea and ascorbic acid to minimise the various interferences was assessed. Experimental Apparatus A Perkin-Elmer Model 360 atomic absorption spectrometer, equipped with an arsenic electrodeless discharge lamp and Varian Techtron antimony hollow-cathode lamp and a recorder in the 10-mV range, was used. Automatic background correction was not used as the corrector was not available.The burner head of the instrument was removed and replaced by an aluminium plate. The absorption cell was an electrically heated silica T-tube (16.5 cm X 0.7 cm i.d.) with a side-arm (7.5 cm X 0.4 cm i.d.). The silica tube assembly was mounted on the aluminium plate (Fig. 1). It was necessary to pre-condition the silica tube by rinsing it in 38% m/V hydrofluoric acid for 45 min. The hydride generator was an ordinary glass vial (1.5 cm i.d.) with an internal volume of 8.6 ml. It was sealed with a screw-cap whose liner was removed and replaced with a layer of silicone-rubber. Three pieces of PTFE tubing passed through the screw-cap (see Fig. 1). Tubing J (23 cm x 2 mm 0.d.) connected the vial to the side-arm of the silica tube and was the outlet for the hydride.Tubing L (44 cm x 2 mm 0.d.) had 3 cm of its length inside the vial and was connected to the gas flow meter, which, in turn, was connected to a cylinder of nitrogen. Tubing K (9 cm x 1 mm 0.d.) had about 3.5 cm of its length inside the vial and was used for the injection of the reductant with a syringe into the vial. A Matheson Model 603 gas flow meter was used to measure the nitrogen flow-rate. Reagents All reagents used were of analytical-reagent grade. Sodium tetrahydroborate(II1) solution (1 % m/V) was prepared by dissolving 1 g of sodium tetrahydroborate(II1) in 100 ml of distilled water with prior addition of two pellets of potassium hydroxide. Stock solutions containing 1000 pg ml-1 of arsenic and antimony were prepared from the oxide and tartrate, respec- tively, by standard procedures.Standard solutions were prepared by appropriate dilution with 0.25 M hydrochloric acid. C I - D NaBH4 Fig. 1. Schematic diagram of the apparatus. A, Absorption cell; B, side-arm; C, heating coil; D, box made of insulating material; E, aluminium plate; F, hydride generation vial; G, silicone-rubber; H, magnetic stirring bar; I, magnetic stirrer; and J, K and L, PTFE tubing126 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Results and Discussion Table 1. Recommended experimental conditions Element Parameter . As Sb Wavelengthhm . . . . . . . . 193.7 217.6 Band width/nm , . . . . . . . 0.7 0.2 Lamp current/mA . . . . . . . . 8.6* 20 Atomising temperaturePC . .. . 1100 1100 Nitrogen flow-rate/l min-1 . . . . 0.20 0.20 Acidityth . . . . . . . . . . 0.25 0.25 * Power supply (in W) for the electrodeless discharge lamp. t Hydrochloric acid was used. Table 2. Effect of the volume of nitrogen inside the vial on the absorbance of 1 ml of arsenic(II1) (8 ng) and antimony(II1) (15 ng) solutions reduced separately by 1 ml of 1% m/V sodium tetra- hydroborate(II1) solution with a nitrogen flow-rate of 0.20 1 min-1 Absorbance Volume of nitrogen As Sb displaced/ml (193.7 nm) (217.6 nm) 6.6 0.58 0.67 5.0 0.56 0.65 3.0 0.55* 0.67* 2.0 >1 .Ot >1.0t * A few solution droplets entered tubing J. t Solution droplets entered the absorption tube. Procedure The silica tube was aligned properly to allow maximum intensity of the light beam to pass through and was then heated electrically, its temperature being determined by a thermo- couple.The optimum conditions for each element are given in Table 1. An aliquot (1 ml) of the sample, standard or blank solution was pipetted into the vial, which was then capped. The magnetic stirrer was turned on, and nitrogen was passed through for 10 s. Next, 1 ml of 1% m/V sodium tetrahydro- borate(II1) solution was injected into the vial through tubing K with a syringe. The peak height of the absorbance was then recorded using the meter or a pen recorder. Minimum electrical damping was used. The screw cap of the vial was removed and the vial rinsed three times with distilled water before the procedure was repeated. Steel and orchard leaves samples were dissolved by stan- dard procedures.About 0.07 g (for As) or 0.5 g (for Sb) of the orchard leaves sample (NBS SRM 1571), previously dried at 85 "C, was ashed using magnesium oxide and nitrate as ashing aids.8 The ash was dissolved in 8 ml of 1 + 1 V/V(for As) or 10 ml of 1 + 4 V/V (for Sb) sulphuric acid. The resulting solution was diluted to 50 ml with distilled water. About 0.03 g (for As) or 0.10 g (for Sb) of the steel sample (NBS SRM 661 AISI 4340 steel) was dissolved in the minimum amount of 2 + 1 V/V HzS04, and the resulting solution was diluted to 100 ml with distilled water. The arsenic and antimony content were determined by the proposed method using standard additions calibration. The test solutions were prepared by adding, to 1 ml of the sample solution, 3 ml of the masking solution (5% m/V in thiourea and 10% m/V in ascorbic acid), 0.5 g of KI and the appropriate amounts of the standards, and diluting to 25 ml with 0.25 M hydrochloric acid.Optimisation of Conditions Hydride generator The effect of the volume of nitrogen inside the vial on the peak absorbance for 8 ng of arsenic and 15 ng of antimony was studied separately by adding 1 ml of lo/' sodium tetrahydro- borate(II1) solution to several vials each containing 1.0 ml of a standard solution of the respective element. The internal volumes of the vials having the same diameter were varied by varying their heights. The volume of nitrogen inside the vial or the dead volume was deduced from the difference between the internal volume of the vial and the volume of solution.The results shown in Table 2 suggested that the peak absorbance did not vary significantly with the dead volume. The unusually large signal observed using the smallest vial under study can be attributed to the absorption due to the solution droplets, which were observed to be carried through the outlet PTFE tubing into the absorption tube. No attempt was made to dry the hydrides, as moisture traps are known to produce undesirable memory effects. The vial with an internal volume of 8.6 ml is large enough to avoid droplets of solution being carried into the absorption tube. It has also been found that the best position for the carrier gas inlet tubing L was just above the solution surface. Absorption tube and atomisation temperature It was observed that the untreated silica tube gave a very low response for both arsenic and antimony. The silica tube was then treated with hydrofluoric acid according to the method of Welz and Melcher,9 and this treatment was believed to remove active spots that could speed up H radical recombi- nation.The signals were significantly increased after treat- ment with hydrofluoric acid; the values for 1 ml of arsenic(II1) (8 ng) were 0.05 (before treatment) and 0.58 (after treatment) and those for antimony(II1) (15 ng) were 0.06 (before treatment) and 0.66 (after treatment). The absorbances of both arsenic and antimony were observed to increase rapidly with increases in the atomisation temperature up to 900 "C, and above this temperature the absorbances increased slowly. When the temperature was above 1200 "C, a distinct "pop" sound was heard, and the background absorption increased suddenly, possibly owing to the burning hydrogen inside the absorption tube.The optimum atomiser temperature was thus chosen to be around 1100 "C, which gave the highest signals without igniting the hydrogen. Band width For arsenic, the most commonly used band width of 0.7 nm was used. However, the sensitivity for antimony was found to be markedly affected by the band width. The signals obtained using a band width of 0.7 nm were only about half of the corresponding values using a band width of 0.2 nm. Thus, for the determintion of antimony, the band width used was 0.2 nm, despite that fact that a large lamp current (i.e., 20 mA) was needed. Flow-rate of the carrier gas The effect of the flow-rate of the carrier gas on the absorption signals was studied and the results are shown in Fig.2. It can be seen that lower flow-rates led to higher absorbance signals. However, the flow-rate around 0.15 1 min-1 was not recom- mended as a "popping" sound, possibly owing to burning hydrogen, was heard, and high background signals were observed. For both elements, the chosen flow-rate was 0.20 1 min-1. Note that the flow-rates recommended are lower than those used by Thompson et af.3 (Le., 1.2 1 min-1) and by Hon et af.7 ( i e . , 1.0 1 min-1).JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 127 1 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 Nitrogen flow-rate/l min-1 Fig. 2. Effect of the flow-rate of nitrogen on the absorbance of arsenic(II1) (8 ng) and antimony(II1) (15 ng).Other conditions as listed in Table 1 0.6 8 0.5 C m + 53 n a 0.4 0.3 t.' 0 0.2 0.3 0.2 Concentration of hydrochloric aCid/M Fig. 3. Effect of hydrochloric acid concentration on the absorbance of arsenic(II1) (8 ng) and antimony(II1) (15 ng). Other conditions as listed in Table 1 Table 3. Effect of hydride-forming elements and some metal ions on the determination of 8 ng of arsenic(II1) and 15 ng of antimony(II1) in separate 1-ml aliquots of 0.25 M hydrochloric acid reduced by 1 ml of 1% mlV sodium tetrahydroborate(II1) solution As( 111) Sb(II1) Tolerance limit of Tolerance limit of Tolerance limit of Tolerance limit of Interferent Added as unmasked solution* masked solution*+ unmasked solution* masked solution*t .. . . . . . . 0.1 pg (6.7) 1.0 pg (67) As H3AsO3 - - Bi B13+ 0.1 vl3 (13) 1 (125) 0.5 (33) 3 .o Pg (200) Ge . . . . . . . . Ge02 0.3 vg (38) 3.2 pg (400) 34 pg (2 267) 40 pg (2 667) Pb . . . . . . . . Pb2+ 40 pg (5 000) 80 pg (10 000) 20 pg (1 333) 40 pg (2 667) Sb . . . . . . . . SbO+ 6.5 ng (0.8) 30 ng (3.8) Se H2Se03 6.0 ng (0.8) 0.1 P8 (13) 5 ng (0.3) 1.5 Yg (100) Sn Sn2+ 12 ng (1.5) 0.15 pg (19) 5 ng (0.3) 0.5 Pg (33) Te Te02 50 ng (6) 1.5 pg (188) 30 ng (2) 0.3 pg (20) . . . . . . 20 pg (2 500) 20 pg (2 500) 40 pg (2 667) 300 pg (20 ow . . . . . . 2 Pg (250) 18 pg (2 250) 40 ng (2.7) 1.5 Pg (loo) . . . . . . . . - - . . . . . . . . . . . . . . . . . . . . . . . . Ag+ . . . . . . . . - 0.7 Pg (88) 60 pg (7 500) 0.7 pg (47) 50 pg (3 333) Cd2+ co2+ cu2+ .. . . . . - 20 Yg (2 5 w 20 pg (2 500) 20 pg (1 333) 40 pg (2 667) Fez+ . . . . . . - 2 (250) 30 pg (3 750) 0.3 mg (20 000) 0.5 mg (33 333) Ni2+ . . . . . . . . - 0 * 4 I % (50) 1 Yg (125) 1.5 Pg (loo) 200 pg (13 333) between the mass of the interferent and the mass of the analyte is given in parentheses. the mass of the analyte is given in parentheses. - - * The tolerance limit is defined as the mass of the hydride-forming element or metal ion per ml causing a 5% signal depression. The ratio t The analyte solution contained 1.2% mlV of ascorbic acid and 0.6% mlV of thiourea. The ratio between the mass of the interferent and Table 4. Effect of anions on the determination of 8 ng of arsenic(II1) and 15 ng of antimony(TII)* As(II1) Sb(II1) Tolerance limit of Tolerance limit of Tolerance limit of Tolerance limit of unmasked masked unmasked masked Interferent solution? solution? solution? solution? Br- .. . . . . 0.5 mg c1- 0.5 mg c104- . . . . . . 0.5 mg c 2 0 4 2 - . . . . . . 0.5 mg Cr2072- . . . . 0.25 pg (31) I- 0.5 mg Mn04- . . . . . . 6 ng (0.8) N03- . . . . . . 0.5 mg po43- . . . . . . 0.5 mg S2- . . . . . . 20 ng (2.5) so32- . . . . . . 0.5 mg so42- . . . . . . 0.5 mg s~o3~- . . . . . . 0.5 mg . . . . . . . . . . . . . . - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg - 0.5 mg 10 pg (1 250) 4 ng (0.3) 10 pg (1 250) 2ng(O.l) 20 ng (2.5) 10 ng (0.7) * The conditions used were the same as those in Table 3 except that for the tolerance limit the mass of the interfering anions was used.t The ratio between the mass of the interferent and the mass of the analyte is given in parentheses.128 JOURNAL OF Acid concentration The effect of the concentration of acid was studied using 0.1-0.3 M hydrochloric acid. The results are shown in Fig. 3. For both elements, the absorption signals increased with increasing acid concentrations. However, the vigour of effervescence also increased with acid concentration. Water droplets began to enter the connecting tubing when the acid concentration was 0.30 M, and when it was above 0.35 M, the effervescence was so vigorous that solution droplets were splashed into the absorption tube, causing very high back- ground absorption. Thus an acid concentration of 0.25 M was chosen as this gave the highest possible signals for arsenic and antimony without the sample solution being carried into the absorption tube.Note that the acid concentration used is much lower than the corresponding values (i.e., 1.5 M) used by other workers, and the range of acid concentrations that can be used is also greatly reduced. Sodium tetrahydro borate( III) concentration The concentration of sodium tetrahydroborate(II1) solution was not critical in the range 1-2% mlV and a 1% sodium tetrahydroborate(II1) solution was used because it is relatively stable and can be kept for 2 d if stored in a refrigerator. Interference Studies Depressive interferences were observed for the determination of six hydride-forming elements, as the hydrides, using the argon - hydrogen flame10 or heated quartz tube atomisation.11 A number of transition metals in Groups VIII and IB interfered with the determinations, and nickel and copper caused severe signal suppression.12 Pierce and Brown11 reported that a number of oxidising anions also caused depressive interferences on arsenic.However, relatively little information is available on the mutual interferences of volatile hydride-forming elements. 12 The range of interference-free determinations for the volatile hydride-forming elements were frequently found to be extended by one to several orders of magnitude when the concentration of the acid was increased.13 However, the concentration of hydrochloric acid used in the proposed method was kept to a relatively low value of below 0.25 M, and it is thus necessary to study the interferences from some metal ions and the mutual interferences of hydride-forming ele- ments at low acidity.The results are shown in Table 3. The results for the effects of a number of anions are shown in Table 4. The tolerance limit was defined as the mass of interferent per ml causing a 5% signal depression. The results of the interference study showed that all observed interferences were suppressive in nature. Nearly all hydride-forming elements caused serious interferences. The ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 tolerance limits for antimony in the determination of arsenic and those for selenium, tin and tellurium in the determination of both arsenic and antimony were at the ng level. However, the interferences caused by lead and germanium were slight.Smith10 ascribed the mutual interference of virtually all the volatile hydrides on each other to compound formation, which could occur in a cool argon - hydrogen flame. Inter-elemental compound formation may also occur in the electrically heated atomiser maintained at 1100 "C in this study, and may be responsible for some of the mutual interferences between the hydride-forming elements. The interference of selenium on arsenic was assumed to be caused by the earlier volatilisation of selenium hydride, thus decreasing the amount of radicals necessary to cause atomisation of arsenic hydride.12 Pre- sumably, the same mechanism may be operative for the serious interference of selenium on antimony. The compara- tively lower interferences of lead and germanium in the determinatlon of arsenic and antimony may be ascribed to the extremely low conversion efficiency into the hydrides in the absence of oxidising agent and at low acidity, respectively.Compared with many hydride-forming elements, all metal cations under study caused much less severe interference. For arsenic(III), those causing slightly more severe interferences were nickel and silver, and those for antimony were cobalt and silver. Welz and Melcherl4 believed that the interference from transition metal ions is caused by capture and decomposition of the evolved gaseous hydride by the finely dispersed metal precipitate, as already discussed by Smithlo and proposed later by Kirkbright and Taddia.15 In this system, the same mechanism was also believed to be operative for the inter- ferences of the transition metal ions.All anionic species caused negligible interference except for three, namely, sulphide, dichromate and permanganate, all of which caused serious interferences for both arsenic and antimony. The mechanism giving rise to sulphide interference is uncertain. A possible mechanism for this is that the concen- trations of arsenic and antimony were decreased by the forma- tion of sulphides that are insoluble in dilute hydrochloric acid. 16 The other two interfering anions were oxidising agents. One possible cause of the interferences was the consumption of sodium tetrahydroborate(III),ll making the latter less avail- able to the analyte. However, this was unlikely to be the major cause of the interferences in view of the fact that a large excess of tetrahydroborate(II1) was used.It might be possible that the analytes were oxidised to the pentavalent states by the interfering oxidising anions. The +V oxidation states of both arsenic and antimony were reported to have a lower analytical sensitivity than the corresponding +I11 oxidation states. 1.7 Table 5. Comparison of sensitivities, detection limits and precision data Absolute Absolute sensi tivi ties */ng detection limitthg Precision, % Reference As Sb As Sb As Sb 7 0.30 (5) 0.33 (3.3) 1.4 1.8 3.9 4.6 3 0.52 (8.7) 0.61 (6.1) 0.8 0.5 5.7 4.9 4and5 5.8t(97) 0.3t(3) 10 - - - 22 0.14$ (2.3) - 0.031 - 4.3 - O.l(l.7) - 0.02 - 0.9 - 23 This work 0.06 0.10 0.22 0.26 2.1 2.2 * The ratios of sensitivities compared with the present method are indicated in parentheses.t Estimated values. $ Data in ng ml-1 using 15 ml of sample solution. Remarks Volume of generator, ca. 50 ml Volume of generator, ca. 50 ml Volume of generator, ca. 60 ml Method involving solvent Using a continuous flow Volume of generator, 8.6 ml extraction generatorJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986. VOL. 1 1.0 129 - / L I I I I 0 5 10 15 20 25 Concentrationhg ml-' Fig. 4. Calibration graphs for the determination of arsenic(II1) and antimony(II1). Conditions as listed in Table 1 Table 6. Contents of arsenic and antimony in standard reference materials Value obtained by present method/ Certified value/ Sample Pgg-' Pg g-' Orchard leaves (SRM 1571) . . As 10.6,10.3 1 0 k 2 Sb 2.7,2.9 2.9 f 0.3 Steel (SRM 661) .. . . . . As 175,179 170 k 10 Sb 40,41 40 f 2 Suppression of Interference Although the use of masking agents to minimise the interfer- ence from transition metal ions has been well documen- ted,15J7-*0 relatively few applications of masking agents to deal with the interference from hydride-forming elements or interfering anions could be found. Yu et al.20 used a mixture of thiourea and ascorbic acid to minimise the effects of large amounts of iron and gold on the determination of antimony in geochemical samples. The mixture of thiourea and ascorbic acid was expected to reduce the interferences from all three classes of interferents under study, namely, hydride-forming elements, transition metal ions and oxidising anions. Thiourea was effective in over- coming the interferences from the softer transition metals in the determination of arsenic.19 The speculation for the effectiveness of thiourea to minimise interferences from the hydride-forming elements stemmed from a knowledge of its reactions with some of these elements.21 It was hoped that thiourea would bind with these interferents and prevent them from interfering in the analysis.Ascorbic acid was expected to reduce the interfering oxidising anions and thus minimise their interference. The masked sample solution was prepared by adding to the unmasked solution a mixture of thiourea and ascorbic acid such that the final solution contained 0.6% mlV of thiourea and 1.2% mlV of ascorbic acid. The tolerance limits for the various interfering species in the presence of the masking solution were again determined and the results are shown in Tables 3 and 4.The results showed that except for a few interferents, the tolerance limits were all increased. The effectiveness was especially pronounced for the oxidising anions, and good for many hydride-forming elements, where the tolerance limits were raised from the ng to the pg levels. Although the tolerance limits for cadmium(I1) and copper(I1) in the determination of arsenic and for germanium in the determination of antimony were unaffected or affected only slightly by the masking solution, the tolerance limits for these ions were in the range 20-40 pg, and they are expected not to interfere in the analysis of many real samples. The interfer- ence from sulphide was not reduced by the masking solution, yet no attempt was made to suppress its interference by other means as it is expected that sulphide would be removed by oxidation during the acid digestion commonly used for sample preparation.In summary, the masking solution has been shown to be effective in reducing the effects of many interferents. However, many species still interfere but to different extents. Hence, it is best to use a standard additions calibration procedure for the determination of arsenic and antimony in real samples unless the matrix interference effects are known to be low. Calibration Graphs, Sensitivities, Detection Limits and Precision The calibration graphs are shown in Fig. 4. The linear range was 0-10 ng ml-l for As(II1) and 0-20 ng ml-1 for Sb(II1).The slopes were 0.067 and 0.043 p.p.b.-l and the correlation coefficients were 0.9990 and 0.9999 for As(II1) and Sb(III), respectively. The absolute sensitivities, detection limits and precision for the determination of arsenic and antimony are shown in Table 5. Here the absolute sensitivity is the mass of the analyte in solution that will produce a 1% absorption. The absolute detection limit is defined as the mass that yields an absorbance of twice that of the standard deviation of the absorbance of a blank. Comparison with Other Methods This method is compared with some recently reported methods in Table 5. The sensitivities for arsenic and antimony obtained by the proposed method are much better than those obtained by other workers. The detection limit for antimony is better using the proposed method and that for arsenic is also much better than that obtained by Thompson et a1.,3 Hon et al.7 and Fleming et al.,4 but worse than that obtained by Fasching et al.,22 whose method involved solvent extraction, and that by Ikeda,*3 who used a continuous flow method. The precision obtained in the determination of the elements under study is around 270, which is lower than the correspond- ing values reported by previous workers, except for the precision for arsenic reported by Ikeda.23 The experimental results obtained clearly indicated that improvements in sensitivity can be achieved by using a small hydride generator and connecting tubes of correspondingly smaller diameter. It seems unlikely that the dilution effect is the only important factor affecting sensitivities.However, it is possible that a low flow-rate of the carrier gas combined with a low acidity increases the residence time of the generated hydride inside the absorption tube, and thus increases the sensitivities for the determinations. This can explain why the absorption signals were insensitive to the changes in the volume of nitrogen inside the vial over the narrow range of 3-6.6 ml, when the flow-rate of the carrier gas was kept constant at 0.2 1 min-1. A small hydride generator is necessary to increase the sensitivities of the determinations because at low flow-rates the carrier gas can then sweep the generated hydride out completely and quickly into the absorption tube with little dilution.130 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 Our Droposed explanation for the increase in sensitivity References using a :mill hydride generator is in line with the atomisation mechanism of volatile hydride-forming elements suggested by Welz and Melcher? the atomisation of gaseous hydrides in a heated quartz cell is caused by collision with free H radicals. When the residence time is increased, the generated hydrides would have more chance to react with the H radicals and become atomised, resulting in higher absorption signals. Analysis of Standard Reference Materials The accuracy of the proposed method was checked by determining the contents of arsenic and antimony in orchard leaves (NBS Standard Reference Material 1571) and a steel sample (NBS Standard Reference Material 661), using stan- dard additions calibrations and the results are presented in Table 6.The contents of both elements in the two standard reference materials determined by the proposed method agree well with the certified values, demonstrating the accuracy of the proposed method. Conclusion The proposed hydride generation atomic absorption pro- cedure for the determination of arsenic and antimony uses a glass vial with an internal volume of 8.6 ml as the hydride generator. Precise and accurate results can be obtained. However, an improvement in the sensitivity using a small hydride generator was achieved using low acid concentrations and a low flow-rate of the carrier gas. The effects of foreign ions on the determination of arsenic and antimony using the proposed method have been studied, and a masking solution consisting of thiourea and ascorbic acid proved to be effective for suppressing the interference from oxidising anions, some hydride-forming elements and metals ions, which interfered seriously with the determination of arsenic and antimony.The authors thank Ms. C. Y. Cheung and Mr. S. F. Luk for experimental assistance. 1. Godden, R. G., and Thomerson, D. R., Analyst, 1980, 105, 1137. 2. Chu, R. C., Barron, G. P., and Baumgarner, P. A. W., Anal. Chem., 1972,44, 1476. 3. Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 595. 4. Fleming, D. E., and Taylor, G. A., Analyst, 1978, 103, 101. 5. Collett, D. L., Fleming, D. E., and Taylor, G. A., Analyst, 1978, 103, 1074. 6. Evans, W. H., Jackson, F. J., and Dellar, D., Analyst, 1979, 104, 16. 7. Hon, P. K., Lau, 0. W., Cheung, W. C., and Wong, M. C., Anal. Chim. Acta, 1980, 115, 355. 8. Bock, R., “A Handbook of Decomposition Methods in Analytical Chemistry,” International Textbook Co., London, 1979, p. 132. 9. Welz, B., and Melcher, M., Analyst, 1983, 108, 213. 10. Smith, A. E., Analyst, 1975, 100, 300. 11. Pierce, F. D., and Brown, H. R., Anal. Chem., 1976,48,693. 12. Welz, B., and Melcher, M., Anal. Chim. Acta, 1981, 131, 17, and references cited therein. 13. Welz, B., and Melcher, M., Spectrochim. Acta, Part B, 1981, 36,439. 14. Welz, B., and Melcher, M., Analyst, 1984, 109, 569. 15. Kirkbright, G. F., and Taddia, M., Anal. Chim. Acta, 1978, 100, 145. 16. Vogel, A. I., “A Textbook of Macro and Semi-micro Qualitative Inorganic Analysis,” Fourth Edition, Longman, London, 1964, pp. 222,237 and 247. 17. Guimont, J., Pichette, M., and Rheaume, N., At. Absorpt. Newsl., 1977, 16,53. 18. Aggett, J., and Aspell, A. C., Analyst, 1976, 101, 341. 19. Peacock, C. J., and Singh, S . C . , Analyst, 1981, 106, 931. 20. Yu, X. A., Dong, G . X., and Li, C. X., Talanta, 1984,31,367. 21. Welcher, F. J., “Organic Analytical Reagents,” Volume IV, Van Nostrand, New York, 1955, pp. 177-188. 22. Amankwah, S. A., and Fasching, J. L., Talanta, 1985,32,111. 23. Ikeda, M., Anal. Chim. Acta, 1985, 167, 289. Paper J6l7 Received July 2nd, 1985 Accepted November 21st, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100125

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

13 1 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Determination of Lead by Graphite Furnace Atomic Absorption SDectrometrY with Argon - Hydrogen as the Purge Gas Using Low-temperature Atomisation Zhe-ming Ni,* Heng-bin Han and Xiao-chun Le Institute of Environmental Chemistry, Academia Sinica, P. 0. Box 934, Beijing, China A method is described for the determination of lead in environmental samples by graphite furnace atomic absorption spectrometry with argon - hydrogen as the purge gas using low-temperature atomisation. The addition of hydrogen to argon helps to reduce the interference caused by the matrix in the sample, and a significant enhancement of the peak absorbance was achieved for lead in environmental sample analysis. The suppression of the lead atomic signal for a real sample in argon was probably due to the presence of halide or perchlorate in the sample.The decrease in the level of interference when hydrogen was added to argon may be ascribed to the formation of HCI, which has a relatively high dissociation constant. The analytical results obtained for lead in environmental reference materials indicate that the method is accurate and reliable. Keywords: Lead determination; graphite furnace atomic absorption spectrometry; argon - hydrogen gas; lo w-tempera ture a tom isa tion Graphite furnace atomic absorption spectrometry is now a well established method for the determination of trace elements in biological and environmental samples. However , severe interferences involving incomplete release of analyte and vapour-phase interactions are frequently encountered in pulse-operated commercial electrothermal atomisers, causing a depression of the atomic signal. To alleviate these problems, matrix modification,1-3 a graphite p l a t f ~ r m , ~ a constant- temperature furnace5 and capacitive discharge heating6 have been used for determinations of trace elements in complex matrices either by chemical treatment of the sample or by isothermal atomisation in the furnace. A simple and con- venient approach for minimising interferences is to introduce reactive gases into the protective atmosphere of an elec- trothermal atomiser during a portion of the furnace heating cycle.Frech and Cedergren7 reported the successful use of hydrogen in a CRA 63-type furnace to eliminate chloride interference in the determination of lead in steel dissolved in aqua reqia.A number of interferences in the determination of lead have been studied and the use of a hydrogen flame in conjunction with a heated carbon rod reduces the interference considerably when an argon - hydrogen or nitrogen - hydrogen flame burns simultaneously with sample vaporisation from the rod.* Heinrichsg reported that chloride suppression of the lead signal can be diminished by using nitrogen - hydrogen as the purge gas in a commercial (Perkin-Elmer HGA-70) graphite furnace. Ottaway and Huttonlo used hydrogen to eliminate the interference of the calcium molecular oxide band emission in the determination of barium and to reduce the effect of strontium band emission on the determination of lithium. The use of hydrogen was also effective in reducing the background absorption from sodium chloride and calcium oxide and improved the precision of the determination of seleniumin the presence of background absorption.11 In a tungsten tube atomiser, a significant increase in sensitivity was achieved for some elements and the appearance temperatures were much lower with an argon-hydrogen mixture than with pure argon.12 Other workers, however, reported that the addition of hydrogen to the nitrogen purge gas to eliminate lead suppres- sion in non-saline waters was unsuccessful13 and hydrogen was not sufficiently effective as a suppressor for the Pb - NaCl ~~ * To whom correspondence should be addressed. system. l4 The inconsistent results obtained by various workers may be ascribed to the use of different types of atomisers or to working under different experimental conditions.The object of this work was to investigate the effect of hydrogen on the determination of lead using a Perkin-Elmer HGA-400 atom- iser. The results indicate that by introducing hydrogen (10% VIV) into the argon purge gas the reduction conditions within the atomiser are improved and the suppression effect on lead caused by complex matrices is reduced. The method was applied to the determination of lead in environmental standard materials, including NBS coal fly ash and bovine liver. The results obtained are in good agreement with the certified values. Experimental Apparatus All results were obtained using a Perkin-Elmer Model 4000 or Model 3030 atomic absorption spectrometer in the automatic background mode fitted with an HGA-400 electrothermal atomiser and a Perkin-Elmer Model 056 chart recorder.Lead absorbance was measured at the 283.3-nm resonance line. The lead hollow-cathode lamp (Shanghai Electro-optic Instrument Works) was operated at 5 mA. Signals were obtained with the internal purge gas interrupt mode and using maximum power heating during the atomisation stage. The graphite tube was purged with argon or 10% hydrogen in argon at an inner flow-rate of 50 ml min-1 during the drying and charring stages rather than in the atomisation stage in order to prevent the diffusional loss of analyte atoms. An Eppendorf pipette fitted with disposable polypropylene tips was used to introduce the sample solution into the graphite furnace.Volumes of 10 or 20 pl were subjected to the following programme: drying at 110 "C for 30 s, charring at 500 "C for 30 s and atomisation at 950 and 1100 "C for 6 s in argon - hydrogen and argon, respectively. Finally, the tube was cleaned at 2500 "C for 3 s. A spectral band width of 0.7 nm was used. For measurement of the atomisation temperature, the absorbance - time profile for lead and the temperature - time profile for the HGA-400 graphite furnace were obtained simultaneously by using a laboratory-made photoconductive device and a double-pen recorder (Hitachi QD-25). The temperature scale was established by using the temperature programme on the HGA-400 power supply. The heating rate setting of the atomiser was 0.1 K ms-l.132 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 A 0 n " " V A 0 Reagents All reagents were of analytical-reagent grade. A stock solution of lead (1000 mg 1-1) was prepared by dissolving the appropriate mass of pure lead metal (99.99%) in nitric acid and diluting with de-ionised water. Standard solutions were prepared by serial dilution of this stock solution with 0.01 M nitric acid. Sample Decomposition Procedure Accurately weigh 10&200 mg of sample into a 30-ml PTFE crucible (from a pressure decomposition vessel) and add 1 ml each of 72% perchloric acid, 68% nitric acid and 35% hydrofluoric acid. Follow the procedure reported previ- ously,15 but use a sample dissolution time of 6-7 h. The final contents of the pressure vessel were dissolved by adding 0.01 M nitric acid and diluted to 10 or 25 ml, depending on the amount of lead in the sample.Results and Discussion Influence of the Amount of Hydrogen Added on the Lead Absorption Signal The proportion of hydrogen added to the argon was varied from 5 to 40%. The addition of 5% of hydrogen increased the peak absorption for lead by 20% compared with that in pure argon, but a further increase in the proportion of hydrogen gave no apparent difference in signal. Therefore, 10% of hydrogen in argon gas was used in the subsequent study. Influence of Hydrogen on the Atomisation and Char Tempera- ture for Lead The introduction of hydrogen in argon as the sheath gas lowered the atomisation temperature for lead. As shown in Fig. 1, lead atomised at 1100°C in argon and at 950°C when argon - hydrogen was used.This phenomenon may be due to the reducing property of hydrogen and a change in atomisa- tion mechanism compared with that observed with pure argon gas. Sturgeon et al. 16 reported the use of mass spectrometry to detect both atomic and molecular forms of the vaporising analyte and the results showed that both Pb and PbO species were detected during sample atomisation. This may indicate that PbO(g) may give rise to Pb(g) via a dissociation process. Probably the addition of hydrogen would enhance the thermal dissociation of lead oxide according to the following reac- tion": PbO(s,l) + H2(g> -+ Pb(s,l,g) + H20(g) 0 10 20 30 40 Time/s Fig. 1. Ar. C, Temperature - time profile Absorbance - time profiles for lead: A, in Ar - H,; and B, in The change in the atomisation temperature of lead has the advantage that atomisation would occur before the bulk of the inorganic matrix vaporises, thus allowing a better time resolution of the atomic peak from the matrix volatilisation.Low-temperature atomisation has the further benefit of extending the lifetime of the graphite tube. The tolerable char temperature for lead in hydrogen was found to be 750 "C, which is 50 "C lower than that in argon. Fig. 2 depicts the decomposition - atomisation curves for lead in argon and argon - hydrogen. The left-hand curves show the peak absorbance values obtained as a function of the thermal decomposition temperature with atomisation at 1100 "C in argon and 950 "C in argon - hydrogen. The right-hand curves give the peak-height absorbances as a function of the atomisation temperature with charring at 500 "C, which was used in all determinations. Effect of Hydrogen on the Peak Absorbance of Lead Although the increase in peak absorbance for lead when using argon - hydrogen was not pronounced compared with that using pure argon when standard lead solution was used, the peak absorbance doubled for lead in real sample analysis.Fig. 3 shows the lead signal and background absorption at the lead wavelength obtained in the analysis of an 83-401 soil (China) sample using argon and argon - hydrogen. The results are similar to those observed by Culver and Shraderlg using a 0 ' I I I 400 800 1200 1600 Temperature/"C Fig. 2. Decomposition/atomisation graphs for 0.4 ng of lead: A, in Ar - Ht; and B in Ar 1 I 0 4 8 12 Ti me/s Fig.3. Absorption signals for lead and background in soil: A, lead in Ar; B, lead in Ar - H,; C, background in Ar; and D, background in Ar - H2JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 133 1 Table 1. Results of lead recovery study. Volumes of sample solutions containing 200 mg of NBS SRM 1577a bovine liver and 100 mg of soil 83-401 (China) were 10 and 50 ml, respectively 0.20 0) m 0.15 e 2 0.10 a 0.05 fi " 0 - - - - 0.05 Pb contentlpg g-1 - studied in the argon - hydrogen atmosphere. It was found that there were no significant interferences on 0.4 ng of lead from I I 1 I I Sample Present method NBS SRM 1633a coal fly ash . . 72 NBS SRM 1577a bovine liver . . 0.14 81-101 river sediment (China) .. 75 82-201 coal fly ash (China) . . . . 33 83-401soil(China) . . . . . . 13 82-301 peach leaves (China) . . 1 .o Certified value 72.4 k 0.4 0.135 f 0.015 79 * 12 33.8 f 4.4 0.99 f 0.08 13* Fig. 5. Effect of MgC12 on the atomic absorption of 0.4 ng of lead: A, in Ar - H2; and B, in Ar Varian CRA-90 atomiser. Hydrogen did not enhance the atomic signal in aqueous standards, but when lead was determined in a nickel-based alloy, a 1300% enhancement was obtained. Our experiment also demonstrates that the argon - hydrogen atmosphere results in a reduction in the non-specific absorbance of the sample. Analytical Results for Standard Reference Materials In order to evaluate the applicability of the method to real samples, studies of the recovery of lead from reference materials were undertaken.The results in Table 1 show that when lead was determined using pure argon, only 50-69% of lead was recovered, depending on fhe type of sample, whereas a recovery of 90-100% was achieved for lead using argon - hydrogen purge gas. Evidently, the hydrogen purge resulted in a marked improvement in the recovery of lead. To obtain accurate results, analyses were carried out by the standard additions method because of the large matrix variation between samples. The relative standard deviation was 2.7% Effects of Chlorides and Perchlorates on Lead Absorption The suppression of lead atomic absorption in real sample analysis is probably caused by chloride present in the sample solution or perchloric acid used in the sample decomposition procedure. The loss of lead in the presence of chloride and perchlorate is probably due to the formation of volatile lead monochloride.As the formation of halide has been recognised as the major cause of interference, a study on the effect of for seven repiicate determinations of lead in soil containing a lead concentration of 13 pg g-l. The analytical results obtained for lead in environmental reference materials using argon - hydrogen purge gas are given in Table 2 and are compared with the certified values. The satisfactory agreement indicates that the method is accurate for the samples examined.134 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. References Ediger, R. D., Peterson, G. E., and Kerber, J.D., At. Absorpt. Newsl., 1974, 13,61. Ediger, R. D., At. Absorpt. Newsl., 1975, 14, 127. Shan, X.-Q., and Ni, Z.-M., Acta Chim. Sin., 1981,39,575. L’vov, B. V., Peliva, L. A., and Sharnopolskii, A. I., Zh. Prikl. Spektrosk., 1977, 27, 395. Hageman, L., Mubarak, A., and Woodriff, R., Appl. Spec- trosc., 1979, 33, 226. Chakrabarti, C. L., Wan, C. C., Hamed, H. A., and Bertels, P. C., Anal. Chem., 1981, 53, 444. Frech, W., and Cedergren, A., Anal. Chim. Acta, 1976,82,93. Amos, M. D., Bennett, P. A., Brodie, K. G., Lung, P. W. Y., and Matousek, J. P., Anal. Chem., 1971,43,211. Heinrichs, H. Z., Fresenius 2. Anal. Chem., 1979,295, 355. Ottaway, J. M., and Hutton, R. C., Analyst, 1977, 102,785. Beaty, R. D., and Cooksey, M. M., At. Absorpt. Newsl., 1978, 17, 53. 12. 13. 14. 15. 16. 17. 18. 19. Sychra, Y., Kolihova, D., Vyskocilova, O., Hlavac, R., and Puschel, P., Anal. Chim. Acta, 1979, 105,263. Thompson, K. C., Wagstaff, K., and Wheatstone, K. C., Analyst, 1977, 102,310. Frech, W., and Cedergren, A., Anal. Chim. Acta, 1977,88,57. Shan, X.-Q., Ni, 2.-M., and Zhang, L., At. Spectrosc., 1984,5, 1. Sturgeon, R. E., Mitchell, D. F., and Berman, S. S., Anal. Chem., 1983,55, 1059. Vyskocilova, O., Sychra, V., Kolihova, D., and Puschel, P., Anal. Chim. Acta, 1979, 105, 271. Culver, B. R., and Shrader, D. E., Am. Lab., 1976, March, 59. L’vov, B. V., Spectrochim. Acta, Part B , 1978, 33, 153. Paper J5l33 Received September 2nd, 1985 Accepted October 21st, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100131

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 135 The Problem of Background Correction in the Determination of Chromium in Urine by Atomic Absorption Spectrometry with Electrothermal Atomisation David J. Halls and Gordon S. Fell Trace Metals Unit, Biochemistry Department, Glasgow Royal Infirmary, Castle Street, Glasgow G4 OSF, UK A previously described positive interference in the determination of chromium in urine was ascribed to the failure of the deuterium arc background correction system to correct fully for background absorption. In this work, it is shown that the interference is dependent on the gain setting of the photomultiplier and on the atomisation temperature. At temperatures below 2400 “C the interference is absent, but on increasing the atomisation temperature from 2400 to 2700 “C the interference increases in severity. The interference is thought to be emission caused by chromium together with potassium and sodium in the matrix.Lowering the atomisation temperature to 2400 “C allows the determination of chromium in urine without significant interference using conventional deuterium arc background correction. Keywords: Chromium in urine; atomic absorption spectrometry; interference; electrothermal atomisation; background correction Most recent attempts to measure chromium in urine have used atomic absorption spectrometry with electrothermal atomisa- tion. The measurement is complicated by the very low concentration of chromium (normally of the order of 1 pg 1-1) and the difficulty of background correction at the relatively long chromium wavelength (357.9 nm) where the emission intensity of a deuterium arc is low.Early attempts at determinationl-3 tried to avoid the use of background correction completely. The values obtained for normal concentrations were higher than more recent esti- mates, however. Guthrie et al.4 described the complications they found in trying to measure chromium in urine with deuterium arc background correction. Distorted or multiple peaks were obtained and they claimed that the deuterium arc failed to remove the background correction completely. The apparent chromium concentration was proportional to the background absorption. In later work,s they used the special- ised research technique of continuum source, Cchelle mono- chromator, wavelength-modulated atomic absorption spec- trometry (CEWM-AAS) to overcome the problem of back- ground correction.Kayne et a1.6 modified a spectrometer to take a quartz - halogen light source for background correction. This gave more intense emission at the chromium wavelength and gave satisfactory background correction in the determina- tion of chromium in serum and urine. Spectrometers with quartz - halogen sources for background correction later became available commercially. Such a spectrometer, the Perkin-Elmer 5000, has been used by Veillon et al.7 to measure chromium in urine with a standard-additions tech- nique. This method has been evaluated by other groups and has been published as a Proposed Selected Method.3 These papers have created the impression that it is impossible to make measurements of chromium in urine with conventional deuterium arc background correction.The work of Guthrie et al.4 has often been quoted (for example, by Chester# and Versiecklo) as an example of the failure of the deuterium arc background correction system. In the Proposed Selected Method for chromium in urine,8 an atomic absorption spectrometer with “enhanced background correction capabili- ties” was listed as a necessary requirement. Equipment that fulfilled this requirement was listed as (a) the Perkin-Elmer 4000 or 5000 with the HGA-500 furnace, (b) the CEWM-AAS system of Harnly and O’Haver,ll (c) the modified Perkin- Elmer 603 with a quartz - halogen source as described by Kayne et al. ,6 (d) the Perkin-Elmer Zeeman system and (e) the Varian miniature furnace using the procedures described by Routh.12 In the last instance, Routh12 had found that the background could be reduced by sheathing the furnace in a nitrogen - hydrogen diffusion flame.The remaining back- ground was removed by a deuterium arc background correc- tion system without any of the problems found by Guthrie et In our work on the determination of chromium in urine with deuterium arc background correction, we encountered prob- lems similar to those described by Guthrie et al.4 In this paper, the nature of this effect is described and a straightforward solution to the problem is offered, which allows the deter- mination of chromium in urine with atomic absorption instrumentation available in many laboratories. a1.4 Experimental Instrumentation Most of the work was carried out with a Perkin-Elmer HGA-500 furnace installed in a Model 2280 spectrometer with deuterium arc background correction.This was equipped with an AS-1 autosampler and a Model 56 recorder. Confirmation of the effects was obtained by using the same furnace in a more advanced spectrometer, the Perkin-Elmer 3030, which has much faster electronics and VDU presentation of the peak shape. Reagents Standards were prepared from a BDH Spectrosol 1000 mg 1-l solution of chromium as chrornium(II1) nitrate. Acids used were of BDH Aristar grade and salts used in interference studies were of AnalaR grade. Water was first de-ionised and then distilled. Control of Contamination All urine containers, glassware, pipettes, micropipette tips and autosampler cups were washed with 20% V/V nitric acid and then rinsed thoroughly with distilled water before use.Instrumental Conditions The spectrometer settings and furnace programmes used with standard and pyrolytically coated graphite tubes are shown in136 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY. APRIL 1986, VOL. 1 Table 1. Instrumental conditions for the measurement of chromium in urine Wavelength . . . . . . . . 357.9 nm Background correction . . . . On Slit width . . . . . . . . 0.7 nm Volume injected . . . . . . 20 pl Inert gas . . . . . . . . Argon Furnace programme- Standard graphite tube Pyrolytic graphite tube Temperature/ Ramp Hold Temperature/ Ramp Hold time/s timels Stage "C timels time/s "C Dry 110 7 20 120 7 20 7 20 1000 7 20 Ash .. . . . . 1200 Atomise . . . . 2400 0 5* 2400 0 5* Clean 1 2 2700 1 2 . . . . . . 2700 . . . . . . * Internal flow-rate, 30 ml min-1; Autozero at -2 s and Record at 0 s. Table 1. Maximum power heating was used in the atomisation step of both programmes. In normal operation, to achieve a balance between the hollow-cathode lamp energy and the deuterium arc energy, the lamp current was turned down to about 11 mA (normal operating current 25 mA) and the deuterium arc energy was increased to its maximum value. Measurements were taken as the peak height from the recorder trace. Argon was used as the carrier gas throughout. Results and Discussion Choice of Heating Programme The conditions for the drying stage were set to give complete drying of the urine sample without boiling.The maximum ashing temperature for aqueous standards and urine samples was found to be 1200 "C for the standard tube and 1000 "C for the pyrolytic tube. These were temperatures indicated by the programmer and not absolute temperatures. Above these indicated temperatures, losses of chromium in the ashing stage were observed. To obtain the greatest sensitivity, maximum power heating was used in the atomisation stage. The peak absorbance of a chromium standard solution increased with increasing atom- isation temperature. Normally the maximum practical tem- perature (about 2700 "C for a standard tube) would be chosen to give the greatest sensitivity. The choice of atomisation temperature was determined by the nature of the interference found, however.The Interference Fig. 1 illustrates the nature of the interference when an atomisation temperature of 2700 "C was used. No difficulty was experienced with aqueous standards that gave similar peak heights with and without background correction. The particular urine sample in Fig. 1 showed a small total absorbance without background correction. Some of that signal would be due to background absorption. With back- ground correction on, instead of the signal becoming smaller it becomes much larger. The example shown is an extreme example as the interference peak swamps the chromium peak. The enhancement is usually more modest and the signal may appear as a double peak as found by Guthrie et al.4 (see also Fig. 5). It is important to point out the steps involved in using background correction.Without background correction, the recommended operating current (25 mA) for the hollow- cathode lamp can be used. To achieve background correction with the deuterium arc source, the hollow-cathode lamp current had to be turned down to about 11 mA to balance the output from the deuterium arc. A correspondingly higher gain setting had to be used. That this interference was primarily D I 0' I Fig. 1. Nature of the interference. A and B, duplicate injections of a 10 pg I-' chromium standard without and with background correction, respectively; C and D, a urine sample without and with background correction. A and C used the chromium hollow-cathode lamp at its recommended current (25 mA) whereas in B and D it was reduced to about 11 mA to allow balance of energy with the deuterium arc.A standard graphite tube was used Table 2. Effect of lamp current and slit width on the total absorbance of a urine sample at 357.9 nm. Background correction not used Absorbance Lamp current/ Slit width Slit width mA 0.7 nm 0.2 nm 10 >o. 100 0.022 15 0.028 0.007 20 0.011 0.003 25 0.008 0.002 30 0.006 0.002 influenced by the change in gain setting was shown by recording the total absorbance signal from a urine sample without background correction at different hollow-cathode lamp current settings (Table 2). The lower the hollow-cathode lamp current chosen, the higher was the gain setting required. As the lamp current was decreased, the total absorbance increased as a result of the interference. Decreasing the slit width reduced the effect.137 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 01 ' I I 2000 2200 2400 2600 Atomisation temperaturei'c Fig. 2. Effect of atomisation temperature on total eak height for a solution of 6 pg 1-1 of chromium in 0.1 M nitric acidP(A) and a urine sample (B). Both were measured with background correction using a standard graphite tube The effect of gain suggested an emission interference, which would be expected to be affected by temperature. Fig. 2 shows the effect of atomisation temperature on the total peak height recorded with background correction for an aqueous standard and a urine sample. As previously indicated, for the aqueous standard a smooth increase in total peak height with increasing atomisation temperature was observed.The urine followed this trend up to 2400 "C. At higher temperatures, the interference became prominent and the total absorbance increased rapidly. Similar effects were shown with pyrolytic- ally coated tubes. Effect of Potassium on the Chromium Signal A number of inorganic constituents of urine were examined separately to see whether they would give the interference shown by urine. The interference was shown by sodium and potassium, the latter giving the greatest increase. The effect of potassium on the chromium signal was examined in more detail in order to understand better the nature of the interference found with urine. The interference is shown in Fig. 3 for a standard graphite tube. At an atomisation temperature of 2700 "C, the signal for chromium was apparently enhanced by potassium [Fig.3(a)]. When background correction was used with its correspond- ingly higher gain setting, the effect was greater [Fig. 3(b)]. The signal for potassium alone does not account for the difference. At an atomisation temperature of 2400 "C, the interference was absent [Fig. 3(c and d ) ] . The influence of gain and temperature on the interference showed that it was the same in nature as that from the urine matrix. The effect of potassium was seen at three different chromium wavelengths (357.9, 359.4 and 425.5 nm) and was present also when pyrolytically coated tubes were used. An increase in interference with an increase in slit width was seen. At a constant gain setting, the effect of background correction was to decrease the interference. The influence of temperature and gain on the interference suggests an effect due to emission, covering at least the wavelength range 357-426 nm.Both potassium and chromium are necessary for the interference as the effects were not shown by chromium solutions alone (Fig. 1) or by potassium alone (Fig. 3). Verification of the Interference with a More Advanced Spectrometer In case the problems experienced were an artifact of the Model 2280 spectrometer, the following experiments were repeated with the furnace in a Perkin-Elmer 3030 spec- trometer, which has faster data capture electronics and VDU Time - Fig. 3. Effect of potassium on the chromium signal at atomisation temperatures of 2700 [(a) and (b)] and 2400°C [(c) and ( d ) ] , respectively.Du licate injections of: A, 2 pg 1-l of Cr; B, 2 p I-' of Cr + 1.8 g 1-1 ofK as KC1; C, 1.8 g 1-1 of K as KC1; and D, istilled water without [(a) and (c)] and with [(b) and ( d ) ] background correction. A standard graphite tube was used display of peak shape: (1) the recording of signals of chromium standards and urine (diluted 1 + 1 with distilled water) at 2400 and 2700 "C with and without background correction; (2) the effect of atomisation temperature (2300- 2700 "C) on the signal for a urine diluted as above; (3) the effect of lamp current on the signal; and (4) the effect of potassium chloride on the chromium signal at atomisation temperatures of 2400 and 2700 "C. The results generally confirmed previous findings but did give a clearer understanding of the interference.In Fig. 4(a), the signal for a simple chromium solution at an atomisation temperature of 2700 "C does show a small disturbance on the tail of the main peak, but this does not exceed the height of the main peak. The disturbance is seen also in the record of the background signal. In Fig. 4(b), the presence of potassium chloride causes a much larger disturbance, with both positive- and negative-going components, suggesting sudden tempor- ary noise at that time. The positive component gives a peak that exceeds that of the main peak. The background absorp- tion reaches its true peak before the chromium signal appears. The background signal also displays the sudden onset of noise. Fig. 5 shows the effect of atomisation temperature on the signal for chromium in urine.At 2300 and 2400"C, the phenomenon is not seen. It begins in a small way at 2500 "C and increases in severity up to 2700 "C. Again, it can be seen that the maximum background signal occurs before the chromium peak and the phenomenon is unrelated to back- ground absorption.138 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Time/s Fig. 4. Peak shapes for ( a ) a simple a ueous solution of 6 pg 1-l of Cr and ( b ) 6 pg I-' of Cr in the presence of 3.7 g 1-1 of K as KC1 at an atomisation temperature of 2700 "C. Hollow-cathode lamp current, 11 mA. Continuous line shows background-corrected signal; broken line displays background absorbance 0 I I 0 2 0 2 I 0 I I 1 0 2 Ti me/s Fig. 5. on. Atomisation temperature: (a) 2300; ( b ) 2400; (c) 2500; ( d ) 2600; and ( e ) 2700 "C Effect of atomisation temperature for chromium in urine.Urine diluted 1 + 1 with a 12 pg 1-l Cr standard. Background correctionJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 139 Implications for the Determination of Chromium in Urine The interference is readily overcome by decreasing the atomisation temperature to 2400 “C, as is apparent from Figs. 2 and 5. Direct determination can then be effected after dilution of the samples 1 + 1 with distilled water, as has been previously described.13 This method has been in use for over 3 years and has shown good correlation with probe electrother- mal atomisation atomic emission spectrometry (ETA-AES)14 and with electrothermal atomic absorption spectrometry (ETA-AAS) using Zeeman-effect background correction. 15 A three-way inter-laboratory comparison between the ETA- AAS method using deuterium background correction at Glasgow Royal Infirmary, the ETA-AES method at the University of Strathclyde and the ETA-AAS method7 with quartz - halogen source background correction at the labora- tory of C.Veillon and colleagues at the US Department of Agriculture, Beltsville, MD, USA, has shown satisfactory agreement. 14 Conclusions The interference described by Guthrie et aZ.4 has led to the widespread assumption that it is impossible to determine chromium in urine with deuterium arc background correction. It is normally assumed that a quartz - halogen source or the Zeeman effect should be used for background correction, as indicated in the Proposed Selected Method for chromium in urine.s Background absorption at the chromium wavelength is not very high and does not coincide in time with the onset of the interference in the atomisation stage.It is apparent from the observations that the interference is emission (possibly overloading the photomultiplier) caused by chromium in conjunction with potassium and sodium in the matrix. The observation of Guthrie et aZ.4 that the apparent chromium concentration (signal peak height) was linearly related to the background absorbance is to be expected as both the emission interference signal and the background absorbance are dependent on the concentration of potassium and sodium in the matrix. The direct relationship in this instance was misleading.The increased gain settings necessary when using the deuterium arc background correction system at the 357.9 nm chromium wavelength accentuate the interference. Results have shown that, even at the normal gain setting (that is, with the hollow-cathode lamp operated at its normal operating current), this interference can still be observed. Users of quartz - halogen background correction should therefore check whether this interference is occurring in their systems. The interference is readily removed by lowering the atomisation temperature to 2400 “C with only a small reduction in sensitivity. This allows chromium in urine to be determined with atomic absorption instrumentation with deuterium arc background correction, which is available in many more laboratories than the specialised or expensive equipment said to be necessary in the Proposed Selected Met hod.* We are grateful to Dr. D. Littlejohn of the University of Strathclyde for helpful discussions on emission from graphite furnaces. 1. 2. 3. 4. 5 . 10. 11. 12. 13. 14. 15. References Davidson, I. W. F., and Secrest, W. L., Anal. Chem., 1972,44, 1808. Ross, R. T., Gonzalez, J. G., and Segar, D. A., Anal. Chirn. Acta, 1973, 63, 205. Schaller, K. H., Essing, H. G., Valentin, H., and Schacke, G., At. Absorpt. Newsl., 1973, 12, 147. Guthrie, B. E., Wolf, W. R., and Veillon, C., Anal. Chern., 1978,50, 1900. Guthrie, B. E., Wolf, W. R., Veillon, C., and Mertz, W., in Hemphill, D. D., Editor, “Trace Substances in Environmental Health XII, Proceedings of the 12th Annual Conference on Trace Substances in Environmental Health,” University of Missouri, Columbia, MO, 1978, p. 490. Kayne, F. J., Komar, G., Laboda, H., and Vanderlinde, R. E., Clin. Chem., 1978, 24, 2151. Veillon, C., Patterson, K. Y., and Bryden, N. A., Anal. Chim. Acta, 1982, 136, 233. Veillon, C., Patterson, K. Y., and Bryden, N. A., Clin. Chem., 1982,28,2309. Chesters, J. K., in Bratter, P., and Schramel, P., Editors, “Trace Element Analytical Chemistry in Medicine and Biol- ogy,” Volume 2, Walter de Gruyter, Berlin, 1983, p. 155. Versieck, J., Trace Elements Med., 1984, 1, 2. Hardy, J. M., and O’Haver, T. C., Anal. Chern., 1977, 49, 2187. Routh, M. W., Anal. Chem., 1980, 52, 182. Halls, D. J., and Fell, G. S., in Bratter, P., and Schramel, P., Editors, “Trace Element Analytical Chemistry in Medicine and Biology,” Volume 2, Walter de Gruyter, Berlin, 1983, p. 667. Baxter, D. C., Littlejohn, D., Ottaway, J. M., Fell, G. S., and Halls, D. J., J . Anal. At. Spectrom., 1986, 1, 35. Baxter, D. C., MSc Thesis, University of Strathclyde, 1984. Paper J.5143 Received October loth, 1985 Accepted November 13th, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100135

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 141 Determination of Metals in Poly(viny1 chloride) by Atomic Absorption Spectrometry Part I. Determination of Calcium, Aluminium and Antimony in Samples of Poly(viny1 chloride) with a High Content of Alkaline Earths Miguel A. Belarra, Felix Gallarta, Jesus M. Anzano and Juan R. Castillo Department of Analytical Chemistry, Science faculty, University of Zaragoza, Zaragoza, Spain The treatment of PVC with H2S04 and H202 causes difficulties if the PVC contains lead or alkaline earths because of the precipitation of the corresponding sulphates. This precipitate can be dissolved with EDTA in ammonia. The procedure is fast and simple and the determination of calcium, aluminium and antimony by atomic absorption spectrometry give a good degree of accuracy.Keywords: Poly(vin yl chloride) analysis; calcium, aluminium and antimony determination; flame atomic absorption spectrometry In spite of its instability, poly(viny1 chloride) (PVC) is one of the most widely used plastics because stabilisers are incorpor- ated that significantly reduce the degradation processes. Certain compounds of tin, lead, calcium, barium, etc., have been widely used as stabilisers for PVC. Although it is obviously important to be able to determine these elements in PVC samples, there are not many analytical methods for doing so, and atomic absorption spectrometry has hardly been used. Usually, the most problematic and time- consuming part of the analytical process is the dissolution of the plastic. The classical method of calcination and subsequent dissolution of the residue in acid1>2 has the disadvantage of possible losses of volatile components, and the treatment with concentrated nitric acid proposed by Rombach et al.3 is too long, as the attack by the acid takes 8 h.One elegant, simple and reasonably fast method is dissolution of PVC in a suitable organic solvent such as dimeth~lacetamide.~ However, a number of practical constraints considerably limit its use: solutions with a concentration of more than 1% mlV are too viscous, and this presents problems in the introduction of the sample to the spectrometer; on the other hand, more dilute solutions may not be suitable for the determination of trace amounts. This approach has not been considered further since the work of Olivier.5-7 After a report by the Analytical Methods Committees on general methods for the destruction of organic matter, Taubinger and Wilson9 published a study of the use of 50% hydrogen peroxide together with concentrated sulphuric acid to treat various organic samples, including PVC.The Ana- lytical Methods Committee extended the study of this type of attack10Jl and Down and Gorsuchl2 made a detailed study of the loss of elements in various organic matrices on treatment with the above-mentioned reagents. This method has been used by Anwar and Marrl3 to determine tin in PVC, and by Mendiola and CO-workers14-17 to determine lead, calcium, barium, cadmium, zinc and tin using 30% hydrogen peroxide, which they found to be equally effective. Mendiola and co-workers tested other methods of attack, and concluded that the most satisfactory results were obtained with sulphuric acid and hydrogen peroxide.One method that may give better results in the future is the direct introduction of the solid into an electrothermal atomiser, which has been used by Girgis- Takla and Chroneosl8 to determine lead. However, the treatment with sulphuric acid causes difficul- ties if the PVC contains lead or alkaline earths because of the precipitation of the corresponding sulphates. Mendiola et al.14 recommended centrifuging the precipitate and dissolving it in concentrated nitric acid, but this procedure is long and entails the nebulisation of two solutions. In this paper, the first of a series on the determination of metals in different types of PVC, we propose the dissolution of the precipitate in EDTA in ammonia solution without the necessity to separate it.This is a faster and more convenient method of determining calcium, aluminium and antimony in samples of PVC with a high content of lead or alkaline earths. Experimental Reagents All solutions were prepared with analytical-reagent grade chemicals and re-distilled water. They were kept in poly- ethylene containers. Concentrated sulphuric acid, sp. gr. 1.84. Hydrogen peroxide, 30% mlm. Concentrated ammonia solution, sp. gr. 0.89. EDTA solution, 4% mlV. Dissolve 4.00 g of EDTA disodium salt in 100 ml of water containing a few drops of ammonia solution. Calcium standard solution, 500 yg ml-1. Prepared by dissolving 1.26 g of calcium carbonate in 50 ml of 1 M HC1 and diluting to 1 1 with water.Aluminium standard solution, 1000 pg ml-1. Prepared by dissolving 1 .OO g of aluminium in 25 ml of 6 M HCl and diluting to 1 1 with 1% V/V HCl. Antimony standard solution, 1000 yg ml-I. Prepared by dissolving 2.74 g of potassium antimony1 tartrate hemihydrate in water and diluting to 1 1 with water. This solution is stable for at least 2 months. Solutions of lower concentrations of these three metals were prepared each day by diluting the standard solutions. Apparatus A Perkin-Elmer Model 3030 atomic absorption spectrometer, fitted with Pye Unicam (calcium), Perkin-Elmer (aluminium) and Instrumentation Laboratory (antimony) hollow-cathode lamps, was used. The standard system of nebulisation and the corresponding burners for air - acetylene (10-cm slit) and dinitrogen oxide - acetylene (5-cm slit) flames were used. The instrument parameters used in measuring the atomic absorption of calcium, aluminium and antimony are given in Table 1.142 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL.1 Table 1. Instrument parameters used in the measurements by atomic absorption spectrometry Distance of burner Lamp current/ below optical axis/ Working range/ Sensitivity/ Element W avelengt h/nm mA Flame mm pg ml-1 pg ml- Calcium . . . . 422.7 6 Air - acetylene 22.5 0-30 0.17 (blue, oxidising) (blue, oxidising) (red, reducing) (pale blue, oxidising) 422.7 6 Air - acetylene 7.5 0-5 0.12 Aluminium . . 309.3 25 N20 - acetylene 10 0-80 1.70* Antimony . . .. 217.6 15 Air - acetylene 10 0-25 0.65 * In the presence of 1.5% KCI. Preparation of the Sample Weigh approximately 0.1 g of the sample and place it in a 100-ml Kjeldahl flask. Add 2 ml of concentrated sulphuric acid and heat for 15 min, first gently and later more strongly. Then add, drop by drop, 5 ml of 30% hydrogen peroxide; the solution is immediately decoloured and a white precipitate appears. Boil the mixture vigorously to eliminate the hydrogen peroxide and leave to cool. Dilute with about 25 ml of distilled water. Add 10 ml of concentrated ammonia solution and 10 ml of the 4% mlV EDTA solution and shake the mixture until the precipitate has dissolved, heating gently if necessary. Dilute to 100 ml with water. This solution is diluted further with water for the determinations.Procedure Determination of calcium using a calibration graph Prepare from the calcium standard solution, by diluting with distilled water, a series of solutions containing between 1 and 20 pg ml-1 of calcium. The standard solution is diluted between 10- and 100-fold with distilled water. Determination of antimony using a calibration graph Prepare from the antimony standard solution, by diluting with distilled water, a series of solutions containing between 4 and 24 pg ml-1 of antimony. The standard solution is diluted between 2- and 5-fold with distilled water. Determination of aluminium using a calibration graph Prepare from the aluminium standard solution, by diluting with distilled water, a series of solutions containing between 3 and 25 pg ml-1 of aluminium.A solution of potassium chloride is added to all these solutions to give a concentration of 1.5% mlV. The standard solution is diluted between 2- and 5-fold with distilled water. Determination of calcium by standard additions The standard solution is diluted 10-fold. In a series of 25-ml calibrated flasks place 20 ml of the diluted solution and add 0, 1, 2, 3 and 4 ml of 125 pg ml-1 calcium solution. Dilute to volume with distilled water. Determination of antimony by standard additions The standard solution is diluted 5-fold. In a series of 25-ml calibrated flasks place 20 ml of the diluted solution and add 0, 1,2,3 and 4 ml of the 100 pg ml-1 solution of antimony. Dilute to volume with distilled water. Determination of aluminium by standard additions The standard solution is diluted 2-fold.In a series of 25-ml calibrated flasks place 20 ml of the diluted solution and add 0, 1, 2, 3 and 4 ml of the 150 pg ml-1 solution of aluminium. Dilute to volume with distilled water. Measurement conditions The above solutions are nebulised and the atomic absorptions of calcium, antimony and aluminium are measured under the conditions given in Table 1. For calcium the distance of the burner below the optical axis was 22.5 mm. Results and Discussion Attack of the Sample Using the method proposed here, samples of PVC with high contents of alkaline earths and lead can be effectively dissolved. Although 50% hydrogen peroxide allows a rapid attack of the sample, if it is not available the use of 30% hydrogen peroxide instead of the recommended 50% does not pose any special problems; it is sufficient to add slightly more reagent and prolong the attack by a few minutes. The treatment with EDTA in ammonia solution to achieve complete dissolution of the precipitate takes less than 5 min, and therefore decomposition of the sample is achieved in 30-35 min.This time is significantly shorter than that needed to centrifuge the precipitate and subsequently dissolve it in nitric acidl4; further, in the procedure proposed here, the determi- nation of the different elements is carried out on one solution. As a test of the effectiveness of the attack of the sample, recovery tests were carried out. As samples of PVC containing known amounts of the metals were not available, these were added to the sample, before the attack, in the form of 1 ml of a hydrochloric solution of the metal.The attack was carried out in the normal way and the results shown in Table 2 were obtained from the corresponding calibration graphs. The recoveries of antimony and calcium are almost loo%, but that of aluminium is lower. However, as will be shown below, the low recovery of aluminium is not due to any incorrect treatment of the sample, but rather to the effect of interferences in the determination. The use of EDTA in ammonia solution is not limited by the presence of cations that give highly insoluble hydroxides under these pH conditions, even at high concentrations. To confirm this, the procedure was applied to a polyolefin containing over 20% of aluminium, and entirely satisfactory results were obtained.Effect of the Reagents on the Atomic Absorption of the Elements The possible interference of the reagents used to attack the sample (sulphuric acid, ammonia and EDTA) in the atomic absorption of the elements under study was investigated. With antimony, if the sample is made up to 200 ml or more the presence of the reagents does not modify the signal. UnderJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 143 Table 2. Results of recovery assays Element Calcium . . . Antimony . . , Aluminium . . Metal found in sample/mg 12.94 8.17 14.57 13.77 3.36 3.53 3.38 3.32 1.91 2.06 2.03 1.97 1.95 2.18 Metal added/mg 6.25 12.50 12.50 12.50 2.43 2.43 2.43 2.43 1.51 1.51 1.51 1.51 1.51 1.51 Metal found/mg 19.15 21.03 26.85 26.11 5.81 5.99 5.80 5.74 3.35 3.49 3.49 3.40 3.36 3.61 Recovery of added metal, o/o 99.4 102.9 98.2 98.7 Mean: 99.8 100.8 101.2 99.6 99.6 Mean: 100.3 95.4 94.7 96.7 94.7 93.4 94.7 Mean: 94.9 Table 3.Determination of calcium, antimony and aluminium in PVC Calcium Antimony Standard Standard Calibration graph additions Calibration graph additions Parameter Dilution of the original solution . . . . . . 1 + 9 1+99 1 + 9 1 + 1 1 + 4 1 + 4 Number of determinations . . . . 6 4 6 4 6 6 Average value/ g per l00g . . . . . . . . 9.75 9.57 9.69 2.56 2.57 2.60 Range/g per 100 g . . . . 9.56-9.87 9.45-9.76 9.47-9.96 2.53-2.59 2.44-2.63 2.41-2.70 Standard deviation/ Relative standard g per 100 g . . . . . . 0.13 0.17 0.25 0.03 0.08 0.10 deviation, YO . . . . 1.3 1.8 2.6 1.1 3.3 3.9 Aluminium Standard Calibration graph additions 1 + 1 1 + 4 1 + 1 10 6 8 1.90 1.93 1.99 1.88-1.94 1.78-2.13 1.95-2.03 0.025 0.133 0.029 1.30 6.88 1.46 the same conditions, the atomic absorption of aluminium is increased by about lo%, which can be attributed to a reduction in its ionisation (estimated by Willis19 to be 14% when a dinitrogen oxide - acetylene flame is used) owing to the presence of sodium from the EDTA.To counteract this effect, 1.5% of potassium chloride can be added to the standard solutions of aluminium. In determining calcium under the optimum conditions of sensitivity (Table 1, distance of the burner below the optical axis 7.5 mm), the above-mentioned reagents produce a significant reduction in the atomic absorption of calcium (about 40°/0), even when the sample is diluted to 1000 ml.However, this interference can be eliminated by increasing the distance of the burner below the optical axis (to 22.5 mm). Although under these conditions the sensitivity of the determination is reduced, this is not a problem because of the high concentration of calcium in the sample. Analysis of PVC Samples The method proposed here has been used to determine calcium, aluminium and antimony in a sample of commercial PVC used for covering cables, with a known calcium content of about 10% (present as calcium carbonate to prevent the production of hydrogen chloride if combustion occurs), and about 2% of aluminium and 2.5% of antimony. The solution obtained after attacking the sample was diluted in such a way that the concentration of the elements to be determined was within the optimum working range and that the reagents used would not cause interference.The results obtained from the calibration graphs and by the standard-additions method are given in Table 3. For calcium and antimony, these results agree fairly well for all the conditions under which they were determined. However, with Table 4. Influence of calcium on the atomic absorption of 30 pg ml-I of aluminium Absorbance Concentration of In the absence of In the presence of EDTA - NH3 calcium/pg ml-1 EDTA - NH3 0 0.079 0.081 3 0.080 0.082 30 0.079 0.082 300 0.074 0.067 3000 0.064 0.041 aluminium, the standard-additions method gives higher results than those obtained from the calibration graph, indicating the possible existence of a matrix effect that reduces the atomic absorption of that element.The influence of the different components of the sample on the atomic absorption of aluminium was studied, and it was found that calcium diminishes the aluminium absorption if the Ca: A1 ratio is more than 1, as can be seen from the results shown in Table 4. In the sample analysed, this ratio is approximately 5. Hence in this instance the aluminium has to be determined by the standard-additions method. Accuracy and Precision In all instances the results obtained agree with the approxi- mate known levels of these metals in the PVC analysed. The precision of the determinations depends on the dilution of the sample and on the technique used. The more dilute solutions give higher standard deviations, which become unacceptable for aluminium at a dilution of 1 + 4.The results obtained by the standard-additions method (which must be144 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, APRIL 1986, VOL. 1 Table 5. Determination of aluminium in polyolefins Without treatment With treatment with Parameter with EDTA - NH3 EDTA - NH3 Number of Average value/g per determinations . . 6 6 loog . . . . . . 23.3 23.5 Range/g per lOOg . . 22.5-24.0 23.1-23.8 Standard deviationlg Relative standard perlOOg . . . . 0.51 0.36 deviation, % . . 2.78 1.53 used with aluminium) are less precise than those obtained from the calibration graph. In any event, the three elements can be determined with a relative standard deviation of less than 1.5%. Further, treatment with EDTA in ammonia solution does not adversely affect the precision, as is shown in Table 5, which gives the results for the determination of aluminium in a polyolefin that does not contain alkaline earths or lead, with and without treatment with EDTA in ammonia solution.Conclusion Treatment with EDTA in ammonia solution gives good results for dissolving the precipitate formed when samples of PVC with a high content of alkaline earths and lead are attacked using concentrated sulphuric acid and hydrogen peroxide. The procedure is rapid and simple, the effects of these reagents on the atomic absorption of calcium, antimony and aluminium are easily compensated for and good accuracy and precision are obtained. This study was financed by the Cornision Asesora de Investigacion Cientifica y Tkcnica, Project 3378183. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. References Druckmann, D., At. Absorpt. Newsl., 1967, 6 , 113. Fassy, H., and Lalet, P., Chim. Anal. (Paris), 1970, 52, 1281. Rombach, N., Apel, R., and Tschochner, F., GZTFachz. Lab., 1980,24, 1165. Musha, S., Munemori, M., and Nakanishi, Y., Bunseki Kugaku, 1964, 13, 330. Olivier, M., Fresenius 2. Anal. Chem., 1969, 248, 145. Olivier, M., At. Absorpt. Newsl., 1971, 10, 12. Olivier, M., Fresenius 2. Anal. Chem., 1971, 257, 135. Analytical Methods Committee, Analyst, 1960, 85, 643. Taubinger, R. P., and Wilson, J. R., Analyst, 1965,90, 429. Analytical Methods Committee, Analyst, 1967, 92, 403. Analytical Methods Committee, Analyst, 1976, 101, 62. Down, J. L., and Gorsuch, T. T., Analyst, 1967, 92, 398. Anwar, J., and Marr, I. L., Talanta, 1982, 29, 869. Mendiola, J. M., Gonzalez, A., and Arribas, S . , Afinidad, 1980,37, 39. Mendiola, J. M., Gonzalez, A., and Arribas, S . , Afinidad, 1980, 37,251. Mendiola, J. M., and Gonzalez, A., Rev. Piast. Mod., 1981, 289,413. Mendiola, J. M., and Gonzalez, A., Rev. Plast. Mod., 1981, 299,550. Girgis-Takla, P., and Chroneos, I., Analyst, 1978, 103, 122. Willis, J. B., Appl. Opt., 1968, 7 , 1295. Paper J5110 Received June 24th, 1985 Accepted October lst, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100141

出版商:RSC    

年代:1986

数据来源: RSC