摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 79 Determination of Sodium, Potassium, Calcium, Magnesium, Iron, Copper and Zinc in Cerebrospinal Fluid by Flow Injection Atomic Absorption Spectrometry J. L. Burguera and M. Burguera Departamento de Quimica, Facultad de Ciencias, Universidad de Los Andes, Apartado Postal 542, Merida 5 10 1 -A, Venezuela and 0. M. Alarcon Departamento de Bioquimica, Facultad de Medicina, Universidad de Los Andes, Merida, Venezuela A method is proposed for the determination of sodium, potassium, calcium, magnesium, iron, copper and zinc in cerebrospinal fluid by flow injection atomic absorption spectrometry. The sample is injected as a discrete slug into a carrier stream of doubly distilled, de-ionised water. Standards are prepared in solutions containing physiological concentrations of various chemical species.Recoveries are 91-1 06%. The procedure is simple, quick, reliable and reproducible. Keywords: Metal determination; flow injection; atomic absorption spectrometry; cerebrospinal fluid Cerebrospinal fluid (CSF) is a liquid that fills the ventricles (cavities) of the brain and the spinal cord and acts as a lubricant and a mechanical barrier against shock.1.2 During certain diseases that affect the meninges or the central nervous system, or both, the CSF may change significantly in physical characteristics, cytological constituents and chemical content; there may also be microbiological and serological alterations to the fluid.3 Careful examination of the CSF in such situations may be useful in differential diagno~is.~.5 As a normal adult only has 100-150 ml of CSF,6 it is necessary to devise methods of determining the chemical constituents of CSF using small amounts of sample.Colorimetry,7-9 flame photometry,7 atomic absorption spectrometry ,1@15 argon plasma emission spectrometry16 and atomic absorption spectrometry with electrothermal atomisa- tion17J* are techniques that have been widely used for the determination of metal species in CSF, with satisfactory results. However, these methods are time consuming to perform and some of them require the use of millilitres of sample for the determination of each element. Flow injection (FI) coupled with atomic absorption spec- trometry (AAS) has recently been used for the determination of elements such as zinc, copper, lithium, calcium, iron and magnesium in blood serum.lS23 This method proved to be simple, rapid, versatile, precise and accurate.In addition, the determination of different elements can be performed using yl of sample. This paper describes a combination of FI and AAS for the direct determination of sodium, potassium, calcium, mag- nesium, iron, copper and zinc in yl samples of CSF. Our results compare well with those obtained using other methods, and as the FI-AAS method can be performed using micro- samples (5 pl for sodium, potassium, calcium and magnesium, 20 yl for zinc and 100 pl for iron and copper) it is thus ideally suited for CSF analysis. Experimental Apparatus The flow system used was essentially the same as that described previously by Fukamachi and Ishibashi.24 The peristaltic pump was a Sage 375-A, furnished with Solvaflex pumping tubes.The atomic absorption spectrometer (Varian Model AA-1475) was used for detecting the elements under the operating conditions shown in Table 1. The injector consisted of a rotary valve (Rheodyne Model 7125) to which a loop of specific volume was attached. PTFE tubing (0.5 mm i.d.) was used throughout. Reagents All reagents were of analytical-reagent grade and de-ionised , doubly distilled water was used throughout. Stock solutions (1000 mg 1-1) of each of the metals were prepared from the following reagents (BDH Chemicals Ltd.): sodium chloride, potassium chloride , calcium carbonate, magnesium metal powder, ammonium iron(I1) sulphate, copper nitrate and zinc nitrate.Working solutions were prepared daily by appropriate dilution. Preliminary tests and optimisation of the experimen- tal parameters were carried out with 3200,100,50,30,0.3,0.1 Table 1. Atomic absorption spectrometer operating conditions for different elements* Species Parameter Na K Ca Mg Fe c u Zn Wavelength/nm . . . . . . 589 769 422.3 285 248.3 324 213.5 Lampcurrent/mA . . . . . . 5 5 4 3 5 4 5 Slit widthhm . . . . . . . . 0.5 1 .o 0.5 0.5 0.2 0.5 1 .o Flame(air-acetylene)/lmin-'. . 9.0+1.5 10.8+1.5 12.0+1.7 10.8+2.0 8.5+1.8 12.5+1.75 10.0+1.3 Burner? height below axishm 15 15 11 9 8.5 11 11.5 Recorder range (FSD)/mV . . 5 X 100 5 x 100 5 x 100 20 x 100 2 x 1 0 0 50 x 1 1 x loo * No background correction was needed. j. A single-slot burner was used for each element.80 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL.1 - Table 2. Experimental parameters used in the FI-AAS system for the determination of sodium, potassium, calcium, magnesium, iron, copper and zinc in CSF. Spectrometer conditions as specified in Table 1 Species Parameter Na K Ca Mg Fe c u Zn Samplesize/pl . . . . . . 5 5 5 5 100 100 20 Dispersion tube lengthkm . . 100 110 60 60 60 20 60 Carrier flow-rate/ml min-1 . , 4.0 3.0 2.5 2.5 2.0 1.5 1.5 Species added to the standard solutions * . . . . . . . . - Na Albumin Na Albumin Albumin Albumin * At the physiological concentrations indicated in Table 3. +Na + P + Na + Na + Na 0 25 50 75 100 !5irc a Sample volume/pI Fig. 1. Effect of sample volume on peak height: A, Na; B, K; C, Mg; D, Ca; E, Zn; F, Cu; and G, Fe.Amounts of elements taken are as given in text. Other conditions as in Tables 1 and 2 10 E 2 UJ a r .- u 5 a a 0 C ~ 20 40 60 80 100 Tube lengthkm Fig. 2. C, Mg; D, Ca; E, Zn; F, Cu; and G, Fe. Conditions as for Fig. 1 Effect of dispersion tube length on peak height: A, Na; B, K; and 0.1 mg 1-1 of sodium, potassium, calcium, magnesium, iron, copper and zinc, respectively. For the determination of the metals present in the CSF, stock solutions were used to prepare working solutions containing physiological amounts of sodium, albumin and phosphorus. 10 . Eo 2 0 a c .- r 5 a a - A I I I I 0 2 4 6 8 Flow-rate/ml min-’ Fig. 3. Effect of carrier solution flow-rate on peak height: A, Na; B, K; C, Mg; D, Ca; E, Zn; F, Cu; and G, Fe.Conditions as for Fig. 1 Procedure The carrier stream of water was pumped continuously, at a constant rate, through the rotary valve injector and the dispersion tube into the spectrometer nebuliser . Samples of the standards or of CSF were injected directly into the continuously flowing carrier stream, and the resulting tran- sient peaks recorded. When a steady base line had been reached a slug of sample was injected into the carrier stream. When the trace returned to the base line another slug of sample could be injected. The minimum time between samples was in all instances ca. 10 s. Sample Collection We studied fasting normal adults; the median age was 40 years (range 38-43 years). This was a population in whom complete neurologic evaluation failed to reveal any evidence of cur- rently active neurological pathology.Consent was obtained from each subject and it was explained that lumbar puncture was an unusual, but safe, procedure and that the results were of scientific interest. From each individual about 4 ml of CSF were obtained by lumbar puncture; the samples were stored in pre-rinsed plastic tubes. This amount of CSF was taken in order to perform all our experiments, however, for routine analysis such a sample volume is unnecessary, and by using the proposed method the analysis can be carried out on 1 ml of CSF. Before use, each plastic tube and syringe was soaked overnight in nitric acid - water (20 + 80 V/V) then rinsed five times with distilled water and allowed to drain dry.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL.1 81 Table 3. Changes in the absorbance (YO) of each element determined in the presence of various species at physiological concentrations Species* ~~ ~ Element Pt Albumin Na K Mg Ca c u Na . . . . -1.5 +1.4 - +2.6 +1.2 +1.5 +1.6 K . . . . -1.8 +1.0 +23.8 - +1.6 +1.7 +1.5 . . . . +1.5 Ca -40.0 +33.5 + 17.6 +3.8 +2.0 - Mg . . . . -5.0 +3.0 +28.6 +2.7 - +3.5 +1.6 Fe . . . . -3.5 + 10.8 + 18.2 +1.5 +3.2 +2.7 +2.1 c u . . . . -7.6 +8.5 +18.5 +1.8 +2.8 +3.0 Zn . . . . -6.8 +6.0 +14.7 +2.2 +3.5 +2.6 +1.3 * Physiological concentrations: P, 15; Na, 3240; K, 98; Mg, 27; Ca, 45; Cu, 0.15; Fe, 0.35; and albumin 500 mg 1 - 1 . t P added as phosphate. - ~ ~ ~- Fe +1.3 +1.2 +1.2 +2.0 +1.8 +1.5 - Table 4. Table 2 Variation in results for the determination of sodium, potassium, calcium, magnesium, iron, copper and zinc in CSF.Element Na . . . . K . . . . Ca . . . . Mg . . . . Fe . . . . c u . . . . Zn . . . . Mean of 30 daily determinations +SD/mg1-1 3330 5 51 90.5 5 1.8 47.0 5 1.1 24.7 k 0.4 0.308 k 0.009 0.074 k 0.002 0.068 & 0.002 Coefficient of variation, Y O 1.5 1.9 2.3 1.6 2.9 2.7 2.9 Mean of 10 determinations on 1 d +SD/mg 1 - 1 3334 k 49 91.3 k 1.5 47.5 k 1.0 24.9 k 0.5 0.310 k 0.009 0.072 k 0.001 0.067 k 0.001 Coefficient of variation, Y O 1.5 1.6 2.1 2.0 2.9 1.5 1.9 Conditions as in Table 5. Determination of the metal species in CSF by various methods * Mean/mg 1- * Species Present method? Na . . . . . . . . 3313 K . . . . . . . . . . 89.95 Ca . . . . . . . . 46.9 Mg . . . .. . . . 28.8 Fe . . . . . . . . 0.300 c u . . . . . . . . 0.081 Zn . . . . . . . . 0.061 * Ten pooled sample specimens. t Conditions as given in Table 2. $. As specified in the text. ~ ~~ Other methods$ 91.00 46.5 27.5 3320 0.302 0.078 0.060 Results and Discussion In our attempts to find a simple and precise method for the determination of sodium, potassium, calcium, magnesium, iron, copper and zinc in CSF, we optimised the spectrometer operating conditions (Table 1) and the FI parameters in order to obtain linearity and good sensitivity for the ranges at which these elements are reported to be found in CSF.25-27 FI Variables When a sample is injected into the flowing carrier stream a peak response is obtained. The height, width and shape of the recorded peaks are affected by four inter-related factors: length and internal diameter of the dispersion tube, carrier flow-rate and sample volume.We did not investigate the effect of varying the internal diameter of the dispersion tube, because tubes wider than those used (0.5 mm i.d.) could not be fitted into the spectrometer nebuliser. The highest peaks with least tailing were obtained under the optimum conditions listed in Table 2. Sample Size It is evident that the volume of the injector loop will influence both the peak height and the peak shape for a constant flow-rate of carrier and a constant amount of element. Fig. 1 illustrates the influence of different sample volumes on the peak heights. In all instances the peak height increased with sample volume injected until a steady-state signal was reached.The best sensitivity was achieved by injecting sample volumes of 5 pl for sodium, potassium, calcium and mag- nesium, 20 pl for zinc and 100 pl for iron and copper determinations. The injection of larger volumes than these increased the peak width, thus reducing sample throughput. The differences in the responses (Fig. 1) of the elements under study are due firstly to the amount of each element used to optimise the experimental parameters and secondly to the recorder span ranges used in each instance. The amounts of Na, K, Ca, Mg, Fe, Cu and Zn used to optimise the sample size were 16.0, 0.5, 0.25, 0.15, 0.03, 0.01 and 0.001 pg, respectively, regardless of the sample volume. For each element the recorder span range was set (Table 1) to obtain good reproducibility; peak heights could be varied by adjust- ing the recorder mV range.We found good precision with the recorder span ranges given in Table 1. The recorder span may be adjusted in order to obtain better sensitivity, although the precision is consequently reduced (the coefficient of variation increases to 5-6%, regardless of the element under study). Dispersion Tube Length The effect of the dispersion tube length was investigated. Initially, an increase of the dispersion tube length slightly increased the peak heights, but the use of larger tube lengths (> than 110 cm for potassium and magnesium; > than 100 cm for sodium; > than 60 cm for calcium, iron and zinc; and > than 20 cm for copper) decreased the peak heights (Fig.2). The fact that the dispersion tube length was found not to be critical could be due to the characteristics of the FI system used, where the sample is not involved in a chemical reaction while being transported to the spectrometer.82 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 Table 6. Recovery of sodium, potassium, calcium, magnesium, iron, copper and zinc added to pooled CSF.* Conditions as given in Tables 1 and 2 Foundt/mg 1- Recovery, % Species addedlmg I-’ Na K Ca Mg Fe Cu Zn Na K Ca Mg Fe Cu Zn Na, 400; K, 20; Ca, 10; Mg, 5; Na, 800; K, 50; Ca, 20; Mg, 10; Fe,0.05;Cu,0.03;Zn,0.03 . . 3804 112.11 56.4 29.4 0.340 0.097 0.088 102 101 99 98 95 94 91 Fe,0.10;Cu,0.08;Zn,0.10 . . 4337 148.05 66.3 34.6 0.396 0.148 0.159 106 105 99 99 97 97 95 * The “known amounts” of elements already present in the pooled CSF samples are those given in Table 4.t Each result is the mean of ten determinations. Table 7. Results for metal concentrations in CSF of 12 normal subjects. Conditions as given in Tables 1 and 2 Parameter Na K Ca Mg Fe c u Zn 3387 90.8 46.1 27.3 0.300 0.062 0.071 Standard deviatiodmg 1-1 . . 95 7.3 2.2 2.3 0.040 0.030 0.037 Meadmg 1-1 . . . . . . . . Range@ 1-1 , . . . . . . . 3278-3521 74.0-100.1 40.8-56.0 23.5-30.8 0.190-0.320 0.030-0.124 0.025-0.144 Carrier Solution Flow-rate Changes in the carrier solution flow-rate within the interval studied (1-8 ml min-1) caused small variations in the measured signals for all the metal species (Fig. 3). However, sharper but smaller peaks were obtained by increasing the rate of flow of the carrier stream above 6 ml min-1.Reproducibil- ity also deteriorated with the higher flow-rates, probably because they affected the dispersion of the sample in the carrier stream, the nebuliser efficiency and the rate of volatilisation of droplets in the burner. Interferences The effect of physiological concentrations of various species26 on the determination” of each element under study was investigated (Table 3). An interference was defined as significant if a change of more than two standard deviations in the measurements was observed. Physiological amounts of albumin affected the determination of all the metals investi- gated, except for sodium, potassium and magnesium. Physio- logical amounts of sodium affected the absorbance signals for all the metals under study, whereas phosphorous only affected the signal for calcium.Therefore, physiological concentra- tions of the above-mentioned species were included in the metal standards for which they were found to interfere (Table 2) * Reproducibility and Precision The metal species were determined in pooled specimens of CSF taken daily for 30 d and also 10 consecutive times on 1 d (Table 4). It was observed that the day to day variation of copper and zinc was larger than the within-day variation over 30 d because of the lack of precision for the analysis at low sample concentrations. Accuracy To check the accuracy of the method, ten pooled specimens of CSF were also analysed by flame photometry (Na and K7) and AAS (Ca,lO Mg,lO Fe,28 Cull and Znll).The results given in Table 5 show good agreement with the present results, which is an indication that the accuracy of the proposed method is satisfactory. Recovery Recoveries of the different metal species added to a labora- tory pooled CSF that contained “known amounts” of the species are summarised in Table 6. Recoveries ranged from 91 to 106%. Calibration Graphs Calibration graphs were prepared for each element using the spectrometer and optimum FI conditions given in Tables 1 and 2. The maximum peak heights increased linearly with element concentration, as expressed by the equations: ANa = 1.95 + 0.003XNa, r = 0.9990; AK = 0.30 + 0.07XK, r = 0.9998; Aca = 0.20 + 0.05Xca, r = 0.9994; AMvIg = 0.40 + 0.20XMg, r = 0.9990; AFe = 0.16 + 6.64XFe, r = 0.9998; Acu = 0.29 + 23.50Xcu, r = 0.9999; and Az, = 0.82 + 14.24Xz,, r = 0.9996, where A represents the peak heights (cm) and X N a (2000-4500 mg 1-11, XK (40-200 mg 1-11, Xc, (5-100 mg 1-l), XM, (10-40 mg l-l), XF, (0.1-0.5 mg l-l), Xc, (0.05-0.5 mg 1-1) and Xz, (0.01-1.0 mg 1-1) represent sodium, potassium, calcium, magnesium, iron, copper and zinc con- centrations, respectively.Normal Values With the technique described, the metal concentrations were determined in the CSF of 12 normal subjects, six women and six men, who were 38-43 years old (mean 40 years). The results are summarised in Table 7. It should be pointed out that no sex-related significant differences of the element concentrations in the CSF were found. The results obtained by the proposed method compared well with those of other investigators for each element studied.25-z7 Conclusion The FI-AAS technique described here is quicker (ca.120 measurements per hour), simpler and at least as accurate as the flame photometric7 and conventional direct aspiration AASlO>llJ8 methods. However, with this method much less sample is required ( ~ 1 ) for the analysis and the CSF samples did not require initial dilution or pre-treatment. 1. 2. 3. References Merrit, H. H., and Fremont-Smith, T. J., “The Cerebrospinal Fluid,” W. B. Saunders, Philadelphia, 1937, p. 1. Fishman, R. A., in Baker, A. B., and Baker, L. H . , Editors, “Clinical Neurology,” Harper & Row, New York, 1975, p. 1. Davson, H., and Welch, K., in Siesjo, B. K., Editor, “Ion Homeostasis Brain, Proceedings of the Third Alfred Benzon Symposium, London, 1970,” Munskaard, Copenhagen, 1971, p.9.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 83 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Bogden, J. D., Troiano, R. A., and Joselow, M. M., Clin. Chem., 1977, 23, 485. Vara, T. R., Garranza, A. G., and Vara, T. C., Rev. Clin. ESP., 1972, 126, 491. Kandel, E. R., Editor, “Handbook of Physiology,” Section 1, Volume 1, American Physiological Society, Williams and Wilkins, Baltimore, 1977, p. 80. Dean, J. A., “Flame Photometry,” McGraw-Hill, New York, 1960, p. 227. Woodbury, J., Lyons, K., Carretta, R., Hahn, A., and Sullivan, J. F., Neurology, 1968, 18, 700. Watson, D., Clin. Chem., 1964, 10, 412. Decker, C. F., Aras, A., and Decker, L.E., Anal. Biochem., 1964, 8, 344. Meret, S., and Henkin, R. I., Clin. Chem., 1971, 17, 369. Montoya, C. M. A., Lopez, M. G., Ruiz, G. A., Juarez, A. G., Jaimes, M. A , , Falcon, D. O., and Polanco, J. L. L., Arch. Invest. Med., 1982, 13, 235. Palm, R., Sjoestroem, R . , and Hallmans, G., Clin. Chem., 1983, 29, 486. Dobbie, J. W., and Smith, M. J. B., Scott. Med. J., 1982, 27, 17. Davson, H., “Physiology of the Cerebrospinal Fluid,” Little Brown, Boston, MA, 1967, p. 187. Hershey, C. O., Varnes, A. W., Hershey, L. A., and Strain, W. H., Neurotoxicology, 1983, 4, 157. McGahan, M. C., and Bito, L. Z., Anal. Biochem., 1983,135, 186. Mazzucotelli, A., Minoia, C., and Frache, R., in Braetter, P., and Schramel, P., Editors, “Trace Elemental Analysis in Chemistry, Medicine and Biology, Proceedings of the Second International Workshop, Berlin, 1982,” de Gruyter, Berlin, 1983, p. 975. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. Rocks, B. F., Sherwood, R. A., and Riley, C . , Clin. Chem., 1982,28, 440. Rocks, B. F., Sherwood, R. A., Bayford, L. M., and Riley, C., Ann. Clin. Biochem., 1982, 19,338. Burguera, J. L., Burguera, M., Gallignani, M., and Alarcon, 0. M., Clin. Chem., 1983, 29, 569. Burguera, J. L., Burguera, M., and Gallignani, M., An. Acad. Bras. Cienc., 1983, 20, 72. Rocks, B. F., Sherwood, R. A., Turner, Z. J., and Riley, C., Ann. Clin. Biochem., 1983, 20,72. Fukamachi, K., and Ishibashi, N., Anal. Chim. Acta, 1980, 119, 383. Rinaldi, M. R., Ardanaz, 0. S., and Notario, R. D., “Liquid0 Cefalorraquideo,” Editorial Medica Panamericana, Buenos Aires, 1980, p. 84. Altan, P. L., and Dittmer, D. S., Editors, “Blood and Other Body Fluids,” Third Edition, Federation of American Societies for Experimental Biology, Bethesda, 1971 , pp. 315-320. Long, C., King, E. J., and Sperry, W. M., Editors, “Bio- chemists’ Handbook,” Van Nostrand, New York, 1961, p. 892. Olson, A. D., and Hamilin, W. B., Clin. Chem., 1969,15,438. Paper J515 Received May 31st, 1985 Accepted July 16th, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100079

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 85 SHORT PAPER A Comparison of Curve Fitting Algorithms for Flame Atomic Absorption Spectrometry Stephen R. Bysouth and Julian F. Tyson Department of Chemistry, University of Technology, Loughborough, Leicestershire LE 7 7 3TU, UK A comparison of six commercially available curve fitting algorithms for calibration in flame atomic absorption spectrometry (FAAS) has been made. Programs using the algorithms were written in BASIC for a microcomputer and, where possible, tested against the commercial version. Calibration curves for magnesium, chromium and nickel produced by the algorithms were compared, on the basis of the sum of squares of the percentage deviations and its root mean square, with each other and with linear interpolation and manual fitting.The standard deviations of the goodness of fit parameters were calculated t o indicate significant differences between the fits obtained. With the exception of the simple parabola, linear interpolation and manual methods, which were significantly poorer, the algorithms performed similarly. In general the errors in curve fitting were well below 5% for the commercial algorithms. Keywords: Calibration; flame atomic absorption spectrometry; curve fitting With the introduction of automatic data handling techniques, several curve fitting algorithms have appeared for calibration in AAS.1 As the processes that cause the bending of the calibration curve are numerous,2 many different models for the curve have been used. De Galan et aZ.3 have compared the use of manually drawn graphs, polynomials of degrees one to four, a rational function (as used by Perkin-Elmer) and a cubic spline function. Miller-Ihli et aZ.4 compared linear, cubic spline and Stineman interpolation methods with linear and quadratic functions and a rational function (as used by Baird Atomic), fitted by a least-squares procedure.Both of these rational functions are described later. In this paper, a comparison is presented of the existing algorithms that are commercially available , using typical calibration data. The algorithms are compared with each other and with two additional techniques: (i) linear interpolations4 and (ii) graphical plots by three colleagues using various manual curve fitting methods. Experimental Calibration and Test Data Nickel, chromium and magnesium were chosen as test elements, as these three elements give characteristically different calibration curve shapes. The calibration for nickel was curved over the absorbance range 0-0.8 and asymptotic to absorbance 0.8.The calibration for chromium was initially linear but curved towards an asymptote at absorbance 1.2. The calibration for magnesium was virtually linear over the absorbance range 0-1.2. Eight standards, including a blank, were made up for each element, by serial dilution of 1000 mg 1-1 stock solutions (BDH Chemicals Ltd.) , using calibrated glassware. These were then presented to either a Shandon Southern A3300 or a Pye Unicam SP90A atomic absorption spectrometer. An air - acetylene flame was used and operating conditions were optimised for maximum sensitivity, The output was monitored with a Philips PM 8251 chart recorder.The standards were aspirated, in order of increasing concentration , several times to account for drift. The blank was aspirated between each standard, and the blank absorbance value subtracted from each standard absorbance value. The calibration data points were distributed evenly along each curve enabling compari- sons to be made between each element for four and five point calibrations. All seven standards (the blank being excluded) were used to calculate a parameter for “goodness of fit” in each instance. Goodness of Fit Parameters The parameters suggested by Miller-Ihli et aL,4 namely the sum of squares of the percentage deviations (SSPD), equation ( l ) , and its root mean square (RMSPD), equation (2), were used where Cci is the ith concentration calculated by the algorithm, c k i is the ith actual concentration and N is the number of data points tested.As the percentage deviation for a blank ( c k i = 0) is infinite, only seven data points ( N = 7) were used in the tests. To give an indication of any significant difference between algorithms, calculations of the standard deviations (SD) of the parameters were made using equations (3) and (4) (derived from the rules for the propagation of random errorss), assuming 1% relative standard deviation in Cci and 0% relative standard deviation in Chi. Algorithms and Programs The eight different algorithms and calibration methods tested are summarised in Table 1.The programs were written in BASIC for a Sharp MZ700 microcomputer, which incorporates a plotter - printer for the production of graphs and “hard copy” of results. The basic curve fitting program described by Miller6 was modified to produce the required algorithms.7-14 Where possible the results from these algorithms were compared with results from the appropriate commercial instrument. The comparison of86 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 Table 1. Summary of algorithms Equations* used and method of curve fitting C = a + bA + cA2, solved for three data points or fitted by least squares for more data points. Coefficients reduced for less than three data points . . . . . . . . . . . . . . . . A/C = a + bA + cA2, fitted as above C = aA + bA2 + cA3, fitted as above .. . . . . . . . . . . . . . . . . . . . . . . C = (klA + k42)/(k2A - l), if the top standard absorbance is within 15% of that predicted by the bottom standard, k3 is set to zero. When the number of standards is three or less, the equation is solved with the appropriate number of coefficients. Otherwise the equation is fitted by least squares points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AIC = a + bA + cA2 solved for each set of three calibration C = a + bA + cA2. A straight line is calculated between the blank and lowest standard. A quadratic is then applied between the next two data points, a third point being calculated using extrapolated slopes . . . . . . . . . . . . . . . . .. C = a + bA, solved for every two points . . . . . . . . . . . . Manufacturers who have adopted Name assigned to algorithm in the algorithm Fig. 1 Baird Atomic (Data-comp system) Baird Atomic (Alpha-star system) Instrumentation Laboratory IL Cubic Baird Quadratic Baird Rational Perkin-Elmer Varian Associates Pye Unicam - PE 3 Coefficient PE 2 Coefficient Varian Rational PU Quadratics Linear Interpolations * Where C is concentration, A is absorbance and a, b, c, k l , k2 and k3 are coefficients to be found during the fitting pro- cedure. Table 2. Calibration and test data Data used for calibration Nickel Chromium Magnesium Concentration/ 4 point 5 Point mg 1-1 X X 0 X X 5 10 X 20 X 30 X 40 50 X X 60 Absorbance 0 0.198 0.317 0.472 0.560 0.617 0.662 0.701 Concentration/ mg I-' 0 5 10 15 20 30 40 50 Absorbance 0 0.296 0.510 0.696 0.824 1.002 1.140 1.198 Concentration/ mg 1-1 0 0.1 0.3 0.4 0.5 0.7 1.0 1.3 Absorbance 0 0.092 0.283 0.384 0.484 0.656 0.926 1.156 the Pye Unicam SP9 algorithm showed that it was not the same as the one written for the microcomputer, so the SP9 computer was used for subsequent experiments.Adjustment of the burner height was used to produce the required absorbance values. The other algorithms written for the microcomputer showed no significant difference from those used in commercial instruments. Results and Discussion The calibration data used is shown in Table 2 and the RMSPD values and their standard deviations are presented as output from the microcomputer in Fig. 1. Values for the associated SSPDs are not given, as they varied by powers of ten, and can easily be calculated from the RMSPD values.The large value for the Perkin-Elmer three-coefficient fit to the four point magnesium calibration is due to the curve becoming discon- tinuous and asymptotic to absorbance 0.6, as shown in Fig. 2. In this instance, the commercial algorithml'J2 would choose the two-coefficient version of the function. Clearly, a quadratic function (Baird Quadratic) does not provide a satisfactory model for calibration curves, except when the curve is virtually linear. A parabola can produce the required asymptote but will also give curvature at the blank level where, in practice, the calibration is often linear. The results for the manual plots are the average RMSPD and SDRMSPD values for the three analysts involved.Although these average parameters compare favourably with those of the computer fitted algorithms, individual results varied and no one person was consistent. Linear interpolation only appears to be useful if curvature is slight, or if many calibration points are employed. As might be expected, all the algorithms perform well for the situation in which the calibration is nearly linear but differences become apparent as curved calibrations are encountered. It is interesting to note that fitting a curve by eye is no better, and often worse, than fitting by computer. There are several ways in which the goodness of fit of a curve to a set of points can be assessed. A commonly used parameter for straight lines is the correlation coefficient.However this tends to give values close to unity if the fit is good for higher absorbances and concentrations even if the fit is poor further down the calibration curve. In addition, it cannot be used for algorithms that either solve equations explicitly for a number of data points, or use interpolation methods. The sum of squares of the percentage deviations and its root mean square are thus more appropriate measures of the goodness of fit as they can be used for all types of algorithm if intermediate data points are used between calibration points. These parameters represent the fit over all of the curve and, being based on percentage deviations in the concentrations, are unbiased towards any part of the curve. The RMSPD represents the likely error in the concentration, calculated by the algorithm, due to the curve fitting method. The parameter used by Knegt and Stork15 is very similar but is based on the absorbance values rather than concentration values.This does not give any indication of the error in concentrations due to the curve fitting method. No attempt was made to compare the calibration algorithms with the standards distributed differently along the curve, which may have improved the performance of some algorithms. This would require detailed knowledge of the curve shape, which may not be available in a real analytical situation.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 87 10 - 5w - - - - 0 \ H E w Fig. 1. RMSPD values for each algorithm and element. The dotted lines represent k half the standard deviations of the values: (a) nickel, 5 point; ( b ) nickel, 4 point; (c) chromium, 5 point; ( d ) chromium, 4 point; (e) magnesium, 5 point; and u> magnesium, 4 point.A, Baird Quadratic; B, Baird Rational; C, IL Cubic; D, Varian Rational; E, PE 3 Coefficient; F, PE 2 Coefficient; G , PU Quadratics; H, Linear Interpolation; and I, Manual 1.2 0 1.3 Concentration, p.p.rn. Fig. 2. The discontinuity observed for the four-point magnesium calibration, when fitted with the Perkin-Elmer 3 Coefficient function Conclusions It is apparent that the performance of a particular algorithm depends on the shape of the calibration curve. There are situations in which the errors introduced by the poor fit of the calibration function can be quite serious.In such instances, manual plotting of the calibration curve is no better than the computer fitted curves. Given the difficulty of predicting what the shape of a particular element’s calibration curve will be for a particular instrument under a particular set of operating conditions, it seems likely that whatever algorithm is used there will always be errors due to the lack of fit. These errors will generally be worse, the fewer the calibration points that are used. As calibration is a non-productive part of an instrument’s operating period, there is naturally a tendency to reduce the number of standards required to the minimum. Some improvements in the fits obtained with the various algorithms may be possible if appropriate weighting is given to each calibration point.It is well known that the calibration points are heteroscedastic. One possible way round the errors introduced by fitting curves to a limited number of calibration points is to use an alternative calibration strategy. Two such alternative strategies, based on the use of a microcomputer, are: ( a ) to produce a continuous concentration - time profile from an exponential dilution flask16 and to collect, in effect, a very large number of calibration points over the required working range; and ( b ) to dilute automatically, by a known factor, a concentrated standard to give the same absorbance as the unknown.” Neither of these two methods requires any knowledge of what the shape of the absorbance concentration plot is. We thank the Trustees of the Analytical Chemistry Trust Fund of the Royal Society of Chemistry for the award of an SAC Research Studentship. 1.2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Tyson, J. F., Analyst, 1984, 109, 313. Price, W. J . , “Spectrochemical Analysis by Atomic Absorp- tion,” Heyden, London, 1979. de Galan, L., van Dalen, H. P. J., and Kornblum, G. R., Analyst, 1985, 110, 323. Miller-Ihli, N. J., O‘Haver, T. C., and Harnly, J. M., Spectrochim. Acta, Part B , 1984, 39, 1603. Miller, J. C., and Miller, J. N., “Statistics for Analytical Chemistry,” Ellis Horwood, Chichester, 1984, p. 46. Miller, A. R., “BASIC Programs for Scientists and Engineers,” Sybex, Berkeley, CA, 1981. Baird Atomic, “Handbook for the Model A5195 Datacomp,” Baird Atomic, Braintree, Essex. Hall, J. R., “Notes on Alphastar-2 Atomic Absorption Curve Correction,” Baird Atomic, Braintree, Essex, 1982. Limbek, B. E., and Rowe, C . J., “Varian Lecture Transcript,” Varian Techtron, Victoria, Australia, 1977. “Some Applications of the IL751 Atomic Absorption Spec- trometer,” Reprint 84, Instrumentation Laboratory, Wilming- ton, MA, 1977. Perkin-Elmer, US Patent 4238830, 1980. Barnett, W. B., Spectrochim. Acta, Part B, 1984,39, 829. Whiteside, P. J., Stockdale, T. J., and Price, W. J., Spectro- chim. Acta, Part B , 1980, 35, 795. Whiteside, P. J., and Stockdale, T. J., “Signal and Data Processing for Atomic Absorption Spectrophotometry,” Pye Unicam, Cambridge. Knegt, J., and Stork, G., Fresenius 2. Anal. Chem., 1974,270, 97. Tyson, J. F., and Appleton, J. M. H., Anal. Proc., 1985,22,17. Bysouth, S. R., and Tyson, J. F., work in progress. Paper J.5116 Received July 12th, 1985 Accepted September 2nd, 1985

ISSN:0267-9477    

DOI:10.1039/JA9860100085

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 89 COM M U N ICATIO N Material for publication as a Communication must be on an urgent matter and be of obvious scientific importance. Rapidity of publication is enhanced if diagrams are omitted, but tables and formulae can be included. Communications should not be simple claims for priority: this facility for rapid publication is intended for brief descriptions of work that has progressed to a stage at which it is likely to be valuable to workers faced with similar problems. A fuller paper may be offered subsequently, if justified by later work. Manuscripts are usually examined by one referee and inclusion of a Communication is at the Editor's discretion. Calibration in Continuum-source Atomic Absorption by Curve Fitting the Transmission Profile John Kindevater and Thomas C.O'Haver Department of Chemistry, University of Maryland, College Park, MD 20742, USA Keywords: Continuum-source atomic absorption spectrometry; curve fitting; computer model; microcomputer The objective of this work was to demonstrate a new method of signal measurement in continuum-source atomic absorption spectrometry that greatly extends the range of calibration graph linearity. The conventional method of absorbance measurement leads to linear calibration graphs only if the well-known requirements for adherence to Beer's Law are met. In continuum-source atomic absorption spectrometry,l stray light and non-zero spectral band pass result in serious deviations from linearity, even if a high-resolution kchelle spectrometer is used.In earlier work we overcame this problem by measuring the absorbance at several points across the wavelength profile of the absorption line and maintaining a separate calibration graph for each point .* This procedure is complex and requires a large number of standard solutions for best results. In the work reported here, we ensemble-average the transmission profile of the analytical line and fit it to a computer model that takes into account the inherent shape and width of the line, the background intensity (lo), instrumental broadening due to finite resolution and unab- sorbed stray light. This yields a value for the intrinsic absorbance, that is, the absorbance at the line centre which would in principle have been observed in the absence of instrumental broadening and stray light.The intrinsic absor- bance is then plotted against solution concentration to give a single calibration graph covering several orders of magnitude of concentration with much greater linearity than has previ- ously been achieved. Experimental Optical System In this work we used a single-channel continuum-source AA system similar to that described previously.3 The system consists of an echelle spectrometer (Spectrospan 111, Spec- trametrics Division of Beckman Instruments Andover, MA, USA) fitted with a single adjustable exit slit and a photomultiplier tube detector. The spectrometer is modified for wavelength modulation by means of a 6 mm thick quartz refractor plate positioned behind the entrance slit. The plate is mounted on the shaft of a limited rotation torque motor (Model G300 PD, General Scanning Inc., Watertown, MA, USA).Entrance and exit slits of the spectrometer are set at 50 pm wide and 300 pm high. A standard Techtron 10-crn slot burner supporting a stoicheiometric air - acetylene flame is used as the atomisation source. The primary continuum source is a Cermax 300-W xenon lamp operated at a current of 15 A. For the line-source measurements, a hollow-cathode lamp operated at the manufacturer's recommended current is used. Because the purpose of this investigation was to compare the behaviour of different measurement and calibration methods, no special effort was made to optimise the nebuliser adjustment or flame conditions to obtain the best possible sensitivity .Control and Data Acquisition An Apple I1 microcomputer (64K memory) drives the torque motor via a 12-bit digital to analog converter (Model DA 101, Tecmar Inc., Cleveland, OH, USA). Photocurrent from the photomultiplier tube is amplified and low-pass filtered by means of a laboratory-constructed operational amplifier circuit and is read by the microcomputer via a 12-bit analog to digital (ADC) converter (Model A113, Interactive Structures Inc., Bala Cynwyd, PA, USA). A machine language sub- routine generates a 40 Hz triangular wavelength modulation waveform and performs the data acquisition. The modulation amplitude is adjusted so that the modulation interval is approximately three times the spectral band pass of the spectrometer, i.e. , about 0.07 nm for the Mg resonance line at 285.2 nm.Transmitted intensity data are taken at the rate of 1.6 KHz (40 points per modulation cycle) and ensemble- averaged for 10 s. Owing to the hysterisis of the torque motor, the forward and reverse scans across the profile do not coincide exactly; for that reason, only the data taken in one direction are used. The machine language data acquisition routine is called from a control program written in BASIC. This program allows the operator to change modulation frequency, ADC gain and90 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 > In a c .- c ,E the number of modulation cycles to be averaged directly from a control menu. In addition, the intensity data are plotted on the computer’s graphic display immediately after the data acquisition is complete.By monitoring these plots the operator may detect problems that cause systematic changes in the data, such as wavelength or order drift. The data can then be stored in a disk file to be used later by the profile modelling program. + ++ 1Opg ml-1 Reagents Standard solutions of copper and magnesium were prepared by dilution of 4000 and 1000 pg ml-1 stock solutions with 1% mlV HN03 and HCl, respectively. Procedure The analytical wavelength is selected and the photomultiplier high voltage and preamplifier gain are adjusted so that the I0 signal is about 3500 counts on the (12-bit) ADC. Solutions are aspirated into the flame as usual and the data acquisition routine is activated. The transmission profile is ensemble- averaged for 10 s and stored to the disk while the next solution is introduced.Curve fitting is done off-line after all the standards have been run. Construction of the Model The model that is used to fit the experimentally measured transmission profiles is based upon the following parameters: ( i ) the intrinsic peak shape of the atomic absorption line; (ii) the intrinsic peak width of the atomic absorption line; (iii) the intrinsic peak height, which we refer to as the intrinsic absorbance. This is the absorbance that would theoretically be measured at the line centre with a measurement device of infinitely great resolution, zero stray light and infinite dynamic range; (iv) the background intensity of the light source, from which the atomic absorption is observed, this we refer to as 1,; ( v ) the slit function of the spectrometer, which describes its resolution; and (vi) the stray light of the spectrometer.It is assumed that the observed transmission profile is the result of the absorption of light from the background source by the intrinisic atomic line, which is then broadened by the slit function of the instrument. The intrinsic peak shape and width are characteristics of the atomic line. The instrument slit function and the stray light are characteristics of the instru- ment system at a given wavelength. These parameters are assumed to be constants for all concentrations of a given element and need not be determined for every individual sample. For each sample to be measured, only the background intensity Zo and the intrinsic absorbance are varied to obtain a best fit to the observed transmission profile.For the initial work reported here, we make the simplifying assumption that the intrinsic line shape is Lorentzian. Thus: Ai 1 + [(h - h,)/6h]* A(h) = . where A(h) is the absorbance profile of the line, Ai is the intrinsic absorbance at the line centre ho and 6h is the intrinsic line width. The transmission (intensity) profile is given by I’(h) = Z,lO-A(V . . . . . . where I, is the intensity of the primary continuum-source light source at wavelength ho. The observed transmission profile I@) results from the convolution of I’(h) with the spectrometer slit function S(h) and the addition of stray light Z,: Z(h) = I’(h)*S(h) + I, . . . . . . (3) where * denotes convolution.Both I. and I , are assumed to be constant across the modulation interval. The slit function can be estimated by observing the profile of a hollow-cathode lamp. The stray light is simply the intensity observed at the line centre at high concentrations, where the intrinsic absorbance is greater than, say, 5. The line width 6h is then determined by fitting the profile of a moderate concentration (0.1 < absorbance < 1.0). For each sample, standard and blank, the intrinsic absor- bance Ai and background intensity I. are adjusted to give a best fit to the observed transmission profile. The intrinsic absorbance is then used just like the conventional absorbance would be used to construct calibration graphs, with the important difference that the intrinsic absorbance gives usable results over a much larger dynamic range.The I. value is a measure of background absorption. The modelling and fitting is carried out with the help of a second BASIC program. This program is compiled for reasons of speed and memory use. Currently, the modelling and fitting is an interactive process between the computer and the operator: that is, the program loads the appropriate data and plots it on the screen. The operator then inputs values for Ai and lo (assuming the other parameters have been set in earlier trials) and the computer generates and plots a model superimposed on the data plot. A standard deviation is also calculated, which is used as a measure of the goodness of fit. The operator then uses the visual information and the fit information to choose a better value of Ai and/or lo and the process repeats until a best fit is obtained.Results and Conclusions Fig. 1 shows a series of experimentally observed transmission profiles for Mg at 285.2 nm for different concentrations. The data points shown are ca. 0.0026 nm apart; the total wavelength region displayed in each profile is ca. 0.036 mm. To 0.2 pg ml-1 Model Ai = 0.08 +++++ 1 Ai = 1.64 I 1 +++ ++ I +++++ 1000 pg ml-1 r--1 1 Ai = 37.4 I Ai = 85.8 Wavelength Fig. 1. Transmission profiles (transmitted intensity versus wavelength) for Mg (285.2 nm) atomic absorption in an air - acetylene flame. The data points shown are ca. 0.0026 nm apart; the total wavelen th region displayed in each profile is ca. 0.036 nm. The left-hanf column shows the experimentally measured profiles for four concentrations of Mg.The right-hand column shows the computer- modelled profiles that give a best fit to the experimental data for each concentration. The value of the intrinsic absorbance Ai (see text) is given for each modelJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 91 100 10 al C (0 +! x 1.0 0 m .- .- L 4 d C - 0.1 0 0 0 0 0 0 n 0 O R A 0 0 0 0 I 1 I I I 1 I 0.01 0.1 1 .o 10 100 1000 Concentration/vg ml- Fig. 2. Atomic absorption analytical curves for Mg (285.2 nm). 0, Intrinsic absorbance Ai derived from the curve-fitting method; 0, conventional continuum-source absorbance measured at the centre of the profile; and A , line source absorbance, measured with an Mg hollow-cathode lamp the right of each experimental profile is the best-fit model obtained by the modelling program.Each model is labelled with the intrinsic absorbance Ai needed to give the best fit. In this instance the model is based on a Lorentzian intrinsic peak shape. Based on some initial curve-fitting experiments, the intrinsic peak width is taken as 0.0026 nm, and the slit function of the spectrometer is approximated by a triangle with a half-width of 0.0078 nm. Despite these rather drastic simplify- ing assumptions, the models are seen to be good fits to the experimental data. Fig. 2 shows the analytical curve for Mg formed by plotting Ai versus the solution concentration. The conventional conti- nuum-source analytical curve, based on the measured absor- bance in the centre of the profile, and the line-source analytical curve measured with a hollow-cathode lamp, are also shown for comparison.The intrinsic absorbance plot clearly covers a much wider concentration range with useful response. It is not, however, perfectly linear over the whole range, nor is the slope on log - log co-ordinates exactly unity. We can expect perfect linearity only if the model is completely rigorous (which it clearly is not in this preliminary work) and if the atomic concentration in the observed portion of the flame is exactly proportional to the solution concentration. Fig. 3 shows similar data for Cu at 324.7 nm. Note that in both instances the sensitivity for the curve-fitting method is greater 0 0 0 0 0 0 0 o o 0 0 O n i D V. I 10 100 1000 10000 Concentration/pg ml-I Fig. 3. Atomic absorption analytical curves for Cu (324.7 nm). 0, Intrinsic absorbance Ai derived from the curve-fitting method; and 0, conventional continuum-source absorbance measured at the centre of the profile than that of the conventional continuum-source measure- ment. This is expected, because the curve-fitting method takes into account the instrumental broadening and stray light, which are the major reasons why the sensitivity of continuum- source AA measurement is less than that of line-source measurement. As can be seen in Fig. 2, the sensitivity of the curve-fitting method at low concentrations is comparable to that of line-source measurement. Work is now continuing in our laboratory to develop a more refined model, taking into account Doppler broadening, the damping factor and hyperfine splitting of the atomic line. We are also seeking to speed up and automate the curve-fitting process and to characterise the precision and reliability of the method. References 1. 2. 3. O’Haver, T. C., Analyst, 1984, 109, 211. Harnly, J. M., and O’Haver, T. C., Anal. Chern., 1981, 53, 1291. O’Haver, T. C., Harnly, J. M . , Marshall, J . , Caroll, J., Littlejohn, D., and Ottaway, J. M . , Analyst, 1985, 110, 451. Paper J5l41 Received October 2nd, I985

ISSN:0267-9477    

DOI:10.1039/JA9860100089

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 93 INSTRUCTIONS TO AUTHORS The Journal of Analytical Atomic Spectrometry (JAAS) is an international journal for the publication of original research papers, short papers, communications and letters concerned with the development and analytical application of atomic spectrometric techniques. The journal will be published bimonthly, will also include comprehensive reviews on spe- cific topics of interest to practising atomic spectroscopists and will incorporate the literature reviews which were previously published in Annual Reports on Analytical Atomic Spectro- scopy (ARAAS) . Manuscripts intended for publication as papers or commun- ications must describe original work related to atomic spectrometric analysis. Papers on all aspects of the subject will be accepted, including fundamental studies, novel instrument developments and practical analytical applications. As well as atomic absorption, atomic emission and atomic fluorescence spectrometry, papers will be welcomed on atomic mass spectrometry and X-ray fluorescence/emission spectrometry.Papers describing the measurement of molecular species where these relate to the characterisation of sources normally used for the production of atoms, or are concerned for example with indirect methods of analyses will also be acceptable for publication. Papers describing the development and applications of hybrid techniques involving atomic spec- trometry ( e . g . , GC coupled AAS and HPLC - ICP) will be particularly welcome.Manuscripts on other subjects of direct interest to atomic spectroscopists including sample prepara- tion and dissolution and analyte pre-concentration proce- dures, as well as the statistical interpretation and use of atomic spectrometric data will also be acceptable for publication. There is no page charge for papers published in JAAS. The following types of papers will be considered. Full papers, describing original work. Short papers, also describing original work, but shorter and of limited breadth of subject matter; there will be no difference in the quality of the work described in full and short papers. Communications, which must be on an urgent matter and be of obvious scientific importance. Rapidity of publication is enhanced if diagrams are omitted, but tables and formulae can be included.Communications should not be simple claims for priority: this facility for rapid publication is intended for brief descriptions of work that has progressed to a stage at which it is likely to be valuable to workers faced with similar problems. A fuller paper may be offered subsequently, if justified by later work. Communications will normally be examined by one referee. Reviews, which must be a critical evaluation of the existing state of knowledge on a particular facet of analytical atomic spectrometry. Every paper (except Communications) will be submitted to at least two referees, by whose advice the Editorial Board of JAAS will be guided as to its acceptance or rejection. Papers that are accepted must not be published elsewhere except by permission. Submission of a manuscript will be regarded as an undertaking that the same material is not being considered for publication by another journal. Copyright.The whole of the literary matter (including tables, figures, diagrams and photographs) in JAAS is copyright and may not be reproduced without permission from the Society or such other owner of the copyright as may be indicated. Manuscripts. Papers should be typewritten in double spacing on one side only of the paper. Three copies of text and illustrations should be sent to the Editor, Journal of Analytical Atomic Spectrometry,The Royal Society of Chemistry, Burling- ton House, London, W1V OBN, and a further copy retained by the author. Papers from North America should be submitted to Dr.J. M. Harnly, US Associate Editor (JAAS), US Department of Agriculture, Beltsville Human Nutrition Center, Bldg 161, Barc-East, Beltsville, Maryland 20705, USA. These papers will be refereed within the United States and Canada. Proofs. The address to which proofs are to be sent should accompany the paper. Proofs should be carefully checked and returned immediately (by Air Mail from outside Europe). Reprints. Fifty reprints of each paper are supplied free on request. Additional reprints can be purchased if ordered at the time of publication. Details are sent to authors with the proofs. Notes on the Writing of Papers for JAAS Manuscripts should be in accordance with the style and usage shown in recent copies of The Analyst. Conciseness of The recommended order of presentation is as indicated below: expression should be aimed at: clarity is increased by adopting a logical order of presentation, with suitable paragraph or section headings.To facilitate abstracting and indexing by Chemical Abstracts Service, and other abstracting organisations, it would be helpful if at least one forename could be included with each author’s family name. Descriptions of new methods should be supported by experimental results showing accuracy, precision and selectiv- ity. ( a ) Title. This should be as brief as is consistent with an adequate indication of the original features of the work. The particular aspect of the subject being discussed should be mentioned in the title. ( b ) Synopsis. A synopsis of about 100 words, giving the salient features and drawing attention to the novel aspects, should be provided for all papers.(c) Keywords. Up to 5 keywords, indicating the topics of importance in the work described, should be included after the synopsis. ( d ) Aim of investigation. An introductory statement of the object of the investigation with any essential historicalJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 background, followed, if necessary, by a brief account of preliminary experimental work. Description of the experimental procedures. Working details must be given concisely. Analytical procedures should preferably be given in the form of instructions; well known operations should not be described in detail. Results. These are best presented in tabular form, followed by any statistical evaluation, which should be in accordance with accepted practice.Discussion of results. This section will comment on the scope of the method and its validity, followed by a statement of any conclusions drawn from the work. The accuracy and precision of any analytical method des- cribed should, wh,ere possible, be discussed with respect to real samples and the scope of the method indicated. Nomenclature. Current internationally recognised (IUPAC) chemical nomenclature should be used. Common trivial names may be used, but should first be defined in terms of IUPAC nomenclature. SZ units. The SI system of units should be used. These units are summarised in Appendix I. The effect on current style of papers for JAAS includes the following: ( a ) dimensions should preferably be given in metres (m) or ( 6 ) temperatures should be expressed in K or “C (not OF); (c) wavelengths should preferably be expressed in nano- metres (nm) (not mp), but angstroms (A) will be allowed; ( d ) frequency should be expressed in Hz (or kHz, etc.), not in c/s or c.P.s.; rotational frequency can be denoted by use of s-1; in millimetres (mm); (e) the micron (p) will not be used; 10-6 m will be 1 pm. Abbreviations.SI units should be used. Molarity is generally expressed as a decimal fraction (e.g., 0.375 M). Abbreviational full stops are omitted after the common contractions of metric units (e.g., ml, g, pg, mm) and other units represented by symbols. Abbreviations other than those of recognised units should be avoided in the text.Percentage concentrations of solutions should be stated in internationally recognised terms. Thus the symbols “m” for mass and “V” for volume are to be used instead of “w” for weight and “v” for volume. The following show the manner of expressing these percentages together with acceptable alterna- tives given in parentheses, where applicable: YO m/m (g per 100 g); O% m/V (g per 100 ml); YO V/V. Further implications of the use of the term “mass” are that “relative atomic mass” of an element (A,) replaces atomic weight, and “relative molecular mass” of a substance (M,) replaces molecular weight. Concentrations of solutions of the common acids are often conveniently given as dilutions of the concentrated acids, such as “dilute hydrochloric acid (1 + 4) ,” which signifies 1 volume of the concentrated acid mixed with 4 volumes of water.This avoids the ambiguity of 1 : 4, which might represent either 1 + 4 or 1 + 3. Dilutions of other solutions can be expressed in a similar manner. Tables and diagrams. The number of tables should be kept to a minimum. Column headings should be brief. Tables consisting of only two columns can often be arranged horizontally. Tables must be supplied with titles and be so set out as to be understandable without reference to the text. Either tables or graphs may be used but not both for the same set of results, unless important additional information is given by so doing. The information given by a straight-line calibration graph can usually be conveyed adequately by an equation or statement in the text.The style used in headings to tables and in labels on the axes of graphs, where the numbers represent numerical values, is, for example: Volume/ml. The diagonal lines (solidus) will not be used to represent “per.” In accordance with the SI system, units such as grams per millilitre are already expressed in the form g ml-1. For a table (or graph), this would appear as: Concentration of solution/g ml-1. It should be noted that the “combined” unit, g ml-1, must not have any “intrusive” numbers. To express concentration in grams per 100 milk litres, the word “per” will still be required: Concentratiodg per 100 ml. It may be preferable for an author to express concentrations in grams per litre (g 1-1) rather than grams per 100 ml. Most diagrams will be retraced and lettered in order to achieve uniform line thicknesses and lettering size and style, so it is not essential to prepare specially traced drawings. However, all diagrams should be carefully and clearly drawn on good quality paper and should be clearly lettered.If possible, complicated flow charts, circuit diagrams, etc., should be supplied as artwork for direct reproduction in order to avoid time-consuming and expensive redrawing. Three sets of illustrations should be provided, two sets of which may be made by any convenient copying process for transmission to the referees. All diagrams should be accompanied by a separately typed set of captions, Wherever possible, extensive identifying lettering should be placed in the caption rather than on lines on graphs, etc.Photographs. Photographs should be submitted only if they convey essential information that cannot be shown in any other way. They should be submitted as glossy or matt prints made to give the maximum detail. Colour photographs will be accepted only when a black-and-white photograph fails to show some vital feature and can be supplied either as prints or transparencies. References. References should be numbered serially in the text by means of superscript figures, e.g., Foote and Delves,l Burns et a1.2 or Hirozawa,3 and collected in numerical order under “References” at the end of the paper. They should be listed, with the authors’ initials, in the following form (double-spaced typing): 1. Foote, J. W., and Delves, H. T., Analyst, 1983, 108, 492. 2.Burns, D. T., Glockling, F., and Harriott, M., 1. Chromatogr., 1980, 200, 305. 3. Hirozawa, S. T., in Kolthoff, I. M., and Elving, P. J., Editors, “Treatise on Analytical Chemistry,” Part 11, Volume 14, Wiley, New York, 1971, p. 23. Journal titles should be abbreviated according to the Chemical Abstracts Service Source Index (CASH). For books, the edition (if not the first), the publisher and the place and date of publication should be given, followed by the page number. Authors must, in their own interest, check their lists of references against the original papers; second-hand references are a frequent source of error. The number of references must be kept to a minimum.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 95 Appendix I The SI System of Units In the SI system there are seven base units- Some derived SI units that have special names are as follows- Physical quantity length mass time electric current thermodynamic temperature amount of substance luminous intensity Name of unit metre kilogram second ampere kelvin mole candela Examples of other derived SI units are- Physical quantity area volume density flow-rate concentration wavelength velocity magnetic field strength Symbol for unit m kg S A K mol Cd Physical quantity energy force power electric charge electric potential difference electric resistance electric capacitance frequency magnetic flux density pressure (magnetic induction) Name of unit square metre cubic metre kilogram per cubic metre millilitre or litre per minute microgram or milligram per gram nanometre metre per second ampere per metre Certain units will be allowed in conjunction with the SI system, e.g.- Physical quantity volume magnetic flux density temperature, t energy (magnetic induction) Name of unit litre gauss degree Celsius electronvol t Symbol for unit 1 G "C eV Name of unit joule newton watt coulomb volt ohm farad hertz tesla pascal Symbol for unit m* m3 kg m-3 ml min-1 or 1 min-1 pg g-1 or mg g-1 nm ms - 1 A m-1 Definition of unit 10-3 m3 = dm3 10-4 T t/"C = TIK - 273.16 1.6021 x lO-19J Symbol for unit J N w C V Q F Hz T Pa The common units of time (e.g., minute, hour, day) and the angular degree (") will continue to be used in appropriate contexts.Appendix II Abbreviations Whenever suitable, elements may be referred to by their chemical symbols and compounds by their formulae.provided that they are defined at the first place of mention. The following abbreviations will be used extensively in the Atomic Spectrometry Updates and may be used in original papers a.c. AA AAS AE AES AF AFS APDC ASV CMP CRM cw d.c. DCP alternating current atomic absorption atomic absorption spectrometry atomic emission atomic emission spectrometry atomic fluorescence atomic fluorescence spectrometry ammonium pyrrolidinedithiocarbamate (ammonium tetramethylenedithiocarbamate) anodic-stripping voltammetry capacitively coupled microwave plasma certified reference material continuous wave direct current d.c. plasma DMF DNA EDL EDTA ETA FAAS FAES FAFS FI GC GDL HCL h.f. HPLC IBMK N, N-dimethylformamide deoxyribonucleic acid electrodeless discharge lamp ethylenediaminetetraacetic acid electrothermal atomisation flame AAS flame AES flame AFS flow injection gas chromatography glow discharge lamp hollow-cathode lamp high- frequency high-performance liquid chromatography isobutyl methyl ketone (4-methylpentan-2-one)96 ICP IR LC LTE MECA MIP MS NAA NaDDC NTA OES PMT p.p.b. p.p.m. PTFE inductively coupled plasma infrared liquid chromatography local thermal equilibrium molecular emission cavity analysis microwave-induced plasma mass spectrometry neutron-activation analysis sodium diethyldithiocarbamate nitrilotriacetic acid optical emission spectrometry photomultiplier tube parts per billion (109) parts per million polytetrafluoroethylene JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY, FEBRUARY 1986, VOL. 1 r.f. REE RM RSD SBR SEM SNR SSMS TCA TLC TOP0 u.h.f. uv VDU vuv XRF radio frequency rare earth element reference material relative standard deviation signal to background ratio scanning electron microscopy signal to noise ratio spark-source mass spectrometry trichloroacetic acid thin-layer chromatography trioctylphosphine oxide ultra-high-frequency ultraviolet visual display unit vacuum ultraviolet X-ray fluorescence

ISSN:0267-9477    

DOI:10.1039/JA9860100093

出版商:RSC    

年代:1986

数据来源: RSC