摘要:

Analyst, February 1996, Vol, 121 (105-109) 105 Raman Spectroscopic Determination of Thymidine Nucleoside Structures in Nucleotides* Rosario Escobara, Pedro Carmonab,? and Marina Molinab a Departamento de Quimica Analitica, Facultad de Quimica, Universidad de Sevilla, 41 OI2-Sevilla, Spain Instituto de Estructura de la Materia (CSIC), Serrano 121, 28006-Madrid, Spain Several spectral differences in Raman spectra were observed for thymidine (TM), 3’-O-acetylthymidine (3’0ATM), 3’,5’-di-O-acetylthymidine (3’5’DOATM), thymidine 5’-monophosphoric acid (H2TMP) and its Ca2+, Sr2+ and Ba2+ salts owing to nucleoside conformational differences in the thymidine moiety. The spectroscopic characterization of the nucleoside structures was made on the basis of previous crystallographic works relative to TM, 3’0ATM, 3’5’DOATM and Ba(TMP)-6H20.The crystal structures of the other compounds reported here are, as yet, unknown. However, their nucleoside structures were inferred from the position of the phosphoester stretching band. The results show that C(2’)-endo-anti7 C(3’)-endo-anti and C(2’)-endo/C(3’)-exo-anti puckers generate marker bands useful to characterize and titrate these thymidine nucleoside conformations. The furanose sugar puckering of the C(2’)-endo-anti geometry has been suggested to account for the enhanced activity of thymidine derivatives against human immunodeficiency virus-type 1. A Raman spectroscopic method is described for the determination of this structure in thymidine nucleosides and nucleotides. Keywords: Raman spectroscopy; thymidine nucleoside; thymidine nucleotide; nucleoside conformation Introduction Raman spectroscopy is a potentially powerful tool for nucleic acid structure analysis.The use of this technique for this purpose requires a detailed understanding of the vibrational modes corresponding to the Raman bands observed. However, less than 25% of the prominent bands in the spectra of nucleic acids have been well characterized by means of empirical correlations between X-ray crystallography and Raman spec- troscopy. For thymidine nucleosides, only the Raman spectra of crystals of dihydrate calcium salts of 5’-thymidylic acid, thymidine, 5’-acetylthymidine and double helical (dA-dT), polynucleotide have been recorded.1-2 We have extended the thymidine nucleoside structure-spectrum correlations to a number of other crystals from which high resolution diffraction data have been obtained and for which the molecular structures have been solved.These correspond to the hexahydrate of calcium 5’-thymidylate,3 where the deoxyribose was found to have a C(3’)endo-anti conformation, and 3’,5’-di-O-acetyl- thymidine and 3’-O-acetylthymidine of which molecules have the anti eonformation displaying the C(2’)endo and C(2’)endol c(3’)exo sugar pucker, re~pectively.~5 Also included are some *I Presented at The SAC ’95 Meeting, Hull, UK, July 11-15, 1995. + Author to whom correspondence should be addressed. nucleotide compounds, crystal stuctures of which are not yet known, but phosphodiester frequencies provide nucleoside structural information which supports the proposed structure- spectrum correlations.In this work the analytical characterization of some thym- idine nucleoside structures by Raman spectroscopy is re- ported. Each nucleoside residue in a nucleic acid chain adopts a local structure which may depend on the identities of neigh- bouring nucleotides. Also described here is a Raman spectro- scopic analytical method for the estimation of the percentages of thymidine residues having the C(2’)endo-anti nucleoside structure in polynucleotides in aqueous solution, as it has been shown that different thymidine nucleoside sugar puckers may be present in either single-stranded DNA or hybrid RNA-DNA structures.2.6 Moreover, the antiviral activity of thymidine derivatives seems to be attributable to the above nucleoside structure. Raman spectroscopy is ideally suited to this compar- ative analysis of crystal and solution conformations of nucleo- sides and nucleotides.Experimental Sample Preparations Thymidine (TM), 3’-O-acetylthymidine (3’0ATM), 3’,5’-di-0- acetylthymidine (3’5’DOATM) and thymidine 5’-mono- phosphoric acid (HZTMP) were supplied by Sigma (St. Louis, MO, USA) and used without further purification. Calcium, strontium and barium salts of thymidine 5’- monophosphoric acid were prepared as follows. Thymidine 5’- monophosphoric acid was dissolved in water and suspensions of the respective hydroxides were added to obtain a (1 + 1) acid : hydroxide mole ratio. The resulting solutions were stirred and heated at about 40 “C until they became transparent.The solutions were then slowly evaporated to obtain the crystalline salts. Spectroscopy Raman spectroscopic measurements were made by using a Jobin-Yvon (France) Model Ramanor U- 1000 spectrometer which was connected to a 386 personal computer. A Spectra Physics (Germany) Model 164 A+ laser (at 514.5 nm beam) was used for exciting samples. Samples were transferred into glass capillary tubes and their spectra were obtained from the average of at least three scans. The excitation power was controlled at about 100 mW for each powder sample, and about 200 mW for each aqueous solution. Quantitative analysis was carried out to evaluate the C(2’)- endo-anti structure (which is believed to possess antiviral activity against human immunodeficiency virus-type 1). For calibration, standard aqueous solutions containing H2(TMP) in106 Analyst, February 1996, Vol.121 0.1 moll-’ dimethylarsinic acid buffer (pH 8.0) were prepared. The concentration range used for H2(TMP) was between 2 and 20% (m/m). However, the concentration range appearing in the calibration graph is smaller owing to the fact that aqueous solutions of H2(TMP) contain a dynamic equilibrium between C(2’)-endo-anti and C(3’)-endo-anti forms, where the percent- age of the former7 is 70%. The intensity of the marker band characteristic of the C(2’)-endo-anti structure was measured relative to the 606 cm-1 band of dimethylarsinic acid buffer which was used as internal standard. Results and Discussion Raman spectra were obtained from thymidine derivatives, shown in Fig.1. Sugar rings in nucleosides and nucleotides are usually puckered with either C(2’) or C(3’) out of the plane formed by the other four atoms.7 The amount of displacement is about 0.5 A and the displaced atom may either be endo, i.e., lying on the same side of the sugar plane as atom C(5’) and the ring of the base, or em, lying on the opposite site of the sugar plane to C(5’) and the base. The most common conformation of the furanose ring is with C(2’)- or C(3’)-endo, and it has been shown7 that in solution C(2’) is endo in purine nucleosides and C(3’) is endo in pyrimidine nucleosides. In the crystalline state, however, this is not the case. It appears that the hydrogen bonding and packing arrangement determines the sugar pucker, because the same nucleoside may have a different type of pucker in different environments.In TM the conformation is C(3’)-exo-anti [sugar pucker- dependent pseudorotation angle (P) = 198’1, with a displace- ment of 0.56 8, from the least-squares plane through the other four ring atoms.8 However, the 3’0ATM molecule adopts an anti conformation about the N-glycosidic bond and exhibits a sugar pucker of C(2’)-endo/C(3’>-exos described by P = 137.9’. It may generally be expected that some features in the 0 TM 3’OATM 3’5‘OOATM HrTMP R, H Fig. 1 studied. Atomic numbering for the thymidine nucleosides and nucleotides 1 I I I t I 600 1000 I400 I800 Raman spectra of crystalline TM (upper) and 3’0ATM (lower). cm-’ Fig. 2 Raman spectrum of a nucleoside residue depend sensitively upon P of the ribose moiety.Some features may also depend upon the internal rotation angle, X, around the glycosidic bond.6 As long as X remains in the anti range, however, these spectral features seem to depend mostly on P rather than on X. This seems to be the case for the thymidine nucleosides studied here whose crystal structures were previously examined through X- ray diffraction. Fig. 2 shows the Raman spectra of TM and 3’0ATM. The C(3’)-endo-anti thymidine nucleoside markers previously reported9 at 668 f 2 and 791 f 2 cm-1 are close to the observed 675 and 793 cm-1 bands, respectively, of TM (Table 1). In 3’0ATM, however, the thymidine 675 cm-l band with medium intensity splits into two (659 and 686 cm-1). Neither of these two bands of 3’0ATM seems to be originated by the acetyl group, as the aqueous solution spectra of both TM and 3’0ATM in the 500-700 cm-1 range are very similar.The above band splitting is, then, attributable to nucleoside conformational features. The 3’5’DOATM structure reveals C(2’)-endo-anti conformers in the crystallographic unit4 which originate the strong band at 673 cm-l (Fig. 3, Table 1). Unlike 3’0ATM, only this band is visible in the 650-700 cm-* range of 3’5’DOATM, in spite of the fact that the acetyl group is also present in this compound. Taking into account that the orientations of the nucleobase relative to the sugar (angle X) and the orientation about the C(4’)-C(5’) bond (angle y) are similar in 3’OATM and 3’5’DOATM,4,5** the medium intensity bands of 3’OATM at 659 and 686 cm-1 can be attributed to the furanose endocyclic C(2’)-endo/C(3’)-exo-anti conformation.In the crystal structure of calcium thymidilate, Ca- (TMP).6H20, the deoxyribose ring is puckered with C(3’) 0.53 A from the plane of the other four atoms, the nucleoside residue adopting the C(3’)-endo-anti conformation.3 As to the strontium and barium salts of thymidine 5’-monophosphoric acid, Sr(TMp).3H20 and Ba(TMP)-6H20, their crystal structures are not yet known, but these show a phosphoester YO-P band at 800 cm-1. It is known that the phosphoester bands lying in the 800-820 cm-1 range are attributable to the C(3’)-endo-anti conformation, and the phosphoester bands appearing in the 820-840 cm- 1 interval result from C(2’)-endo-anti nucleo- sides.10-’1 Accordingly, we assign the 800 cm-1 band of Sr(TMP).3H20 and Ba(TMP)-6H20 to the C(3’)-endo-anti conformation.A medium intensity band in the 635-645 cm-l region is present in the Raman spectra of alkaline-earth salts of thymidine 5’-monophosphoric acid (Fig. 4, Table 1) assignable to C(3’)-endo-anti thymidine. Crystalline H2TMP shows a phosphoester YO-P band at 822 cm-1 (Fig. 3) which strongly suggests the presence of C(2’)-endo-anti thymidine moiety. This H2TMP crystal structure, not yet known through X-ray crystallography, produces a medium intensity band at 67 1 cm-1, as occurs for TM and 3’5’DOATM. On comparing the spectra of thymidine and acetylated thymidine derivatives (Table 1), it is clear that these compounds give, in the 735-775 cm-1 range, two medium-strong intensity bands, and HzTMP gives only one band at 751 cm-1.The alkaline-earth salts of H2TMP produce (in the above range) some bands which show weak intensity. Therefore, it may be concluded that one or two medium-strong intensity bands in the 735-775 cm-l interval can be indicative of high values of the pseudorotation angle, P. The Raman spectral feature in the 850-950 cm-l region is considered to reflect certain ring vibrations of the deoxyribose moiety. In fact, the 897 cm-1 band of TM is assigned to the ring breathing vibration of the deoxyribose residue.12 Another medium intensity band at 853 cm-l is assigned to another ring vibration of the deoxyribose ring.12 Apart from the appearance of the acetyl YO-C band at 889 cm-1 in the 3’0ATM and 3’5’DOATM spectra, it should be noted that there is a considerable perturbation of the 850-950 cm- spectral profile in going from TM to acetylated TM and alkaline-earth salts ofAnalyst, February 1996, Vol.121 107 H2TMP. However, the frequencies in this spectral range do not seem to depend greatly upon the P parameter value of the thymidine residue. Assignment of Raman bands to the 3’- and 5’-deoxy substituents of thymidine is complicated by antici- pated changes of sugar conformation which accompany sub- stitution. The following observations are cautiously suggested to be consistent with the effect of 3’- and 5’-deoxy substituents. Both 3’0ATM and 3’5’DOATM show an acetyl vC=O band at 889 cm-1 which is absent from the non-acetylated compounds studied here. These observations are true for both crystals and solutions.Accordingly, the 889 cm-1 band may be diagnostic of the 3’- and 5’-acetyl substituents. Some Raman lines located at higher frequencies are, furthermore, sensitive to the conformation given by the pucker of the ribofuranose ring. The crystals having C(2’)-endo-anti thymidine (3’5’DOATM, H2TMP) originate a medium intensity band in the 1205-1210 cm-1 range, this band being absent in the other nucleosides and nucleotides studied here. Table 1 Raman bands observed for thymidine derivatives. A, TM; B, 3’0ATM; C, 3’5’DOATM; D, H2(TMP); E, Ca(TMP)-6H20; F, Sr(TMP).3H,O; G, Ba(TMP)-6H20; s, strong; vs, very strong; sh, shoulder; m, medium; w, weak; vw, very weak; ring, thymine ring Compound A 564 w 632 vw 675 m 737 m 772 m 793 m 853 m 873 vw 897 m 961 vw 974 vw 1001 vw 1015 m 1052 vw 1066 w 1101 w 1123 w 1173 w 1198 s 1226 s 1234 s 1278 vw 1324 vw 1365 vs 1390w 1403 w 1437 w 1460 vw 1483 w 1643 m 1665 vs B 559 w 614 m 627 w 659 m 686 m 748 m 760 w 773 m 786 m 838 vw 889 m 894 m 957 w 996 w 1024 w 1058 vw 1076 vw 1102 vw 1133 w 1185 m 1228 s 1260 vw 1296 vw 1317 vw 1339 vw 1360 vw 1379 s 1390 s 1422 vw 1459 vw 1661 vs 1708 w 1742 w C 557 w 596 vw 614 vw 632 w 673 s 736 w 746 m 760 m 786 m 854 m 868 w 889 m 895 sh 922 vw 955 w 970 vw 984 vw 1018 w 1060 vw 1101 vw 1116 w 1136 w 1188 m 1209 m 1233 s 1284 vw 1302 vw 1332 vw 1376 vs 1386 sh 1418 m 1447 w 1474 w 1664 vs 1711 w 1735 w 1751 w D 564 w 586 vw 613 vw 640 vw 671m 751 m 790 m 822 m 860 vw 910 w 942 w 966 w 1021 m 1057 w 1102 w 1143 w 1186 m 1206 m 1241 s 1267 sh 1375 s 1412 m 1442 vw 1463 vw 1484 w 1655 vs E 554 m 588 w 643 m 633 vw 707 vw 744 w 760 w 780 m 805 m 864 w 895 w 927 vw 973 vw 1008 m 1050 w 1065 vw 1107 w 1128 vw 1148 vw 1180m 1194 sh 1236 s 1275 m 1285 sh 1374 vs 1388 w 1419 m 1447 vw 1466 w 1653 vs 1691 w F 555 w 585 w 636 m 714 vw 732 w 763 vw 780 m 800 sh 845 m 878 m 907 m 975 vw 1004 sh 1016s 1040 w 1073 vw 1093 w 1150 vw 1175 w 1215 w 1226 m 1244 m 1287 w 1319 w 1364 s 1376 s 1387 s 1422 sh 1432 m 1475 w 1653 vs 1671 vs 1715 w G 556 w 590 w 637 m 707 w 728 w 749 vw 779 m 800 sh 842 m 874 nm 908 m 943 vw 972 vw 1009 s 1040 vw 1050 vw 1062 w 1104 w 1143 w 1166w 1177 w 1225 m 1239 m 1283 w 1318 w 1361 s 1373 s 1384 s 1420 m 1436 m 1472 w 1654 sh 1665 vs Assignment ribose ring vo-P ribose v,c-0-c ribose vCC, vC0, ribose vCC, vC0, ribose vCC, vC0, ribose vp0;- ribose ribose vC0, vCC, ribose vC0, ribose v-ring v-ring v-ring v-ring, 6CH3 CH, ribose CH, ribose CH, ribose YC = c, vC(4) = 0 vC(2) = 0 acetyl vC = 0108 Analyst, February 1996, Vol.121 The above spectral correlations are reflected in Table 2. Some of these correlations appear to be similar to those established for the A and B forms of polynucleotides.2 As can be seen in Table L I I I I I 600 1000 t 400 I800 cm-I Fig. 3 (lower). Raman spectra of crystalline 3’5‘DOATM (upper) and H2(TMP) L I I I I I 1 I 600 1000 I400 I800 cm -1 Fig. 4 Raman spectra of crystalline Ca2+, (upper) Sr2+ (middle) and Ba2+ (lower) salts of H2(TMP). 2, the spectral features depend appreciably on the pseudorota- tion angle, P, of the ribose moiety.The following points are noteworthy. (a) A medium intensity band is found in the 1205-12 10 cm- interval of C(2’)-endo-anti nucleoside con- formers and is absent in the other sugar puckers. (b) The strong Raman band in the 750-800 cm-* range appears at higher frequency (785-795 cm- l) for higher P [C(2’)-endo-anti family, which also includes the C(2’)-endo/C(3’)-ex~-anti and C(3)’-exo-anti puckers]. The corresponding strong band for the C(3’)-endo-anti conformers (lower P) is located at lower frequencies, between 775 and 780 cm-1 (Figs. 5-6). (c) The strong band generated by the C(2’)-endo-anti and C(3’)-exo-anti nucleoside conformations falling in the 665-675 cm-1 range splits into two (659 and 686 cm-1) for the 3’0ATM crystal, where thymidine residue adopts the C(2’)-endo/C(3’)-exo-anti conformation. The corresponding band for the C(3’)-endo-anti nucleoside structure appears at lower frequencies, between 635 and 645 cm-l.(4 The two medium intensity bands located in the 735-775 cm- 1 range of the C(2’)-endo-anti pucker family are replaced by a medium-strong band in the 735-755 cm-I interval upon phosphorylation. Recently,g-’ it was proposed that furanose sugar puckering in the C(2’)-endo family may account for the enhanced activity of thymidine nucleosides against human immunodeficiency virus- type I . The titration of C(2’)-endo-anti thymidine residues is, thus, particularly interesting from this point of view. With this aim, the intensity of the band near 670 cm-’ generated by the 600 700 800 cm-1 Fig.5 crystalline TM, 3’0ATM, 3’5’DOATM and H2(TMP). Raman spectra in the 600-850 cm- region of (from top to bottom) Table 2 Raman marker bands of compounds involving thymidine residues Nucleoside conformation Compound Frequencies ((d-A - dT), B form* 668 k 2 748 k 2 C(2’)-endo/C(3’)-exo-anti 3’0ATM 67 1 75 1 746 760 673 614 748 659 773 686 737 737 772 C(3’)-exo-anti Thymidine 675 C( 3’)-endo-anti * Frequencies from ref. 2. (dA - dT)n A form* 642 f 2 [Ca(TMP)6H20 Sr(TMP).3H20 643 636 Ba(TMP)-6H20 637 791 + _ 2 1142 f 2 1208 f 2 790 I206 786 1209 786 793 777 k 2 780 780 779 1239 k 2 1236 1244 I239Analyst, February 1996, Vol. 121 109 0.00 above structure was measured. The titration curve fits the least- squares equation I = 4.5% - 4.355, where c is the concentration expressed as mass percentage and I is the intensity of the 670 cm-1 band relative to the intensity of the band at 606 cm-1 due to dimethylarsinic acid buffer used as an internal standard (see Fig.7). The detection limit, defined as the concentration that produces a peak height two times that of the background noise, was 1.8% (m/m). The errors for intensity measurements were determined as the s values for four independent determinations of each standard solution at concentration levels of 2.0,5.0,6.9, 7.8, 11.7 and 13.5% (m/m) and are represented as error bars on the intensities in Fig. 7. The least-squares correlation coefficient 0.98, and the non-zero intercept, results from the detection limit or sensitivity of this technique. Taking into account the equation of the calibration graph, one cannot obtain any Raman signal at 670 cm-1 when c d 0.95%.The samples of the calibration set were chosen to B 12 . I include the lowest concentration range detectable by the Raman instrument. Although the concentrations in the corresponding biochemical assays can be lower than those reported here, preconcentration of the biological samples can lead to concen- trations falling into the range of this Raman method. In conclusion, we have reported for the first time the Raman spectra of 3’0ATM, 3’5’DOATM, H,(TMP), Ca(TMP).H20, Sr(TMP)-3H20 and Ba(TMP).6H20. The Raman spectra of these compounds have added to those reported in previous works relative to other thymidine derivatives.1.9 The C(2’)- endo-anti, C(3’)-endo-anti and C(2’)-endo/C(3’)-exo-anti thy- midine structures can be characterized by Raman spectroscopy in the 600-800 cm-1 and 1200-1300 cm-1 ranges.A Raman spectroscopic analytical method is described here for the estimation of the percentage rate of the C(2’)-endo-anti thymidine nucleoside conformation, which is believed to enhance the antiviral activity of thymidine residue in nucleo- sides and nucleotides. W 51 0. c I I I I I 600 700 800 cm-’ Fig. 6 Raman spectra in the 600-850 cm-’ region of crystalline Ca2+ (upper), Sr2+ (middle) and Ba2+ (lower) salts of H2(TMP). 80.00 r I 40.00 0 -a ‘- E -2- 0 i? 20.00 Concentration (m/m %) of C(Z’)-ende-anti form Fig. 7 Calibration graph of C(2’)-endo-anti thymidine nucleoside struc- ture. The intensity of the band near 670 cm-1 was measured relative to the 606 cm-’ band of 0.1 mol 1-l buffer. Data also represent s values at each point. Financial support from the Direcci6n General de Investigaci6n Cientifica y TCcnica is gratefully acknowledged (Project No. PB93-0131). References 1 2 3 4 5 6 7 8 9 10 11 12 Katahira, M., Nishimura, Y., Tsuboi, M., Sato, T., Mitsui, Y., and Iitaka, Y., Biochim. Biophys. Acta, 1986, 867, 256. Thomas, G. J., Jr., in Spectroscopy of Biological Systems, ed. Clark, R. J. H., and Hester, R. E., Wiley, Chichester, 1986, vol. 13, ch. 5, Trueblood, K. N., Horn, P., and Luzzati, V., Acta Crystallogr., 1961, 14, 965. Wilson, C. C., Low, J. N., Tollin, P., and Wilson, H. R., Acta Crystallogr., 1984, C40, 1712. Eccleston, R. S., Wilson, C. C., and Howie, R. A., Acta Crystallogr., 1988, C44, 1424. Benevides, J. M., Stow, P. L., Ilag, L. L., Incardona, N. L., Thomas, G. J., Jr., Biochemistry, 1991, 30, 4855. Saenger, W., Principles of Nucleic Acid Structure, Springer Verlag, New York, 1984, pp. 51-104. Young, D. W., Tollin, P., and Willson, H. R., Acta Crystallogr., 1969, B25, 1423. Dijkstra, S., Benevides, J. M., and Thomas, G. J., Jr., J. Mol. Struct., 1991,242, 283. Ghomi, M., Letellier, R., and Taillandier. E., Biopolymers, 1988,27, 605. Ghomi, M., Letellier, R., Liquier, J., and Taillandier, E., Znt. J . Biochem., 1990,22, 691. Tsuboi, M., Ueda, T., Ushizawa, K., Sasatake, Y., Ono, A., Kainosho, M., and Ishido, Y., Bull. Chem. Soc. Jpn., 1994, 67, 1483. pp. 233-309. Paper 5104651 C Received July 17, 1995 Accepted September 20,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100105

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (111-117) 111 Comparison of Chemical Modifiers for the Determination of Vanadium in Water and Oil Samples by Electrothermal Atomization Atomic Absorption Spectrometry* Nikolaos S. Thomaidis and Efrosini A. Piperakit Laboratory of Analytical Chemistry, Chemistry Department, University of Athens, Panepistimiopolis, 157 71, Athens, Greece Chemical modifiers (isomorphous metals and other compounds such as NHdSCN and ascorbic acid) for the determination of vanadium are described. Ascorbic acid (100 pg) eliminated the interferences from NaCl, CaC12 and FeC13 salts. The mass of the metallic modifiers and the maximum permissible pyrolysis temperature (tpyr) have been carefully optimized. Magnesium nitrate (20 pg), rhodium (2 pg) and platinum (2.5 pg) increased tpyr from 1100 "C, without modifiers, to 1500,1700 and 1600 "C, respectively.The characteristic mass without modifiers was 21.2 pg, whereas in the presence of 20 pg of Mg(NO&, 1 pg of Rh and 1 pg of Pt, the characteristic mass was 20.5 pg, 14.7 pg and 14.4 pg, respectively, measuring the peak absorbance with a moderately used graphite tube (around 100 atomization cycles). The LOD was generally 0.5 pg 1-1 without chemical modifier and the same with 1 pg of Pt. The efficiency of the above metals as chemical modifiers was tested in the analysis of oil standards diluted with isobutyl methyl ketone (IBMK). It was found that 1 pg of Pt increased the tpyr from 1000 "C to 1400 "C and decreased the characteristic mass from 27.5 to 15.3 pg, measuring the peak absorbance with a new graphite tube.Magnesium nitrate was inadequate for this type of sample. The ageing of the graphite tube surface affected the vanadium determination. A slow drift in sensitivity appeared with increasing atomization cycles. This was controlled with periodic recalibration using integrated absorbance measurements. Keywords: Vanadium; electrothermal atomization atomic absorption spectrometry; chloride interference; chemical modifier; oil analysis Introduction ETAAS is widely used for the determination of vanadium at trace levels. However, two main problems affect the analytical results by using this technique: ( i ) the formation of thermally stable vanadium and (ii) the interferences in the presence of common chloride matrices.1+7 The first problem is enhanced by the ageing of the pyrolytic coating of the graphite tubes"l0 where a decrease in the sensitivity is observed with increasing number of firings. The formation of vanadium carbides is also promoted in the presence of other carbide-forming element~6.97~ 1712 and, there- fore, a strong decrease in the signal is observed.In order to counter the second problem, several approaches have been developed including e~tractionl3.1~ and chemical modifi~ation.5,~>9J~J5 Chemical modifiers have been used * Presented at The SAC '95 Meeting, Hull, UK, July 11-15, 1995. f To whom correspondence should be addressed. either to increase the volatility of the interfering compound or to stabilize the analyte. Ascorbic acid was used to suppress the iron chloride interference on vanadium.5 The most widely used metallic chemical modifier is magnesium nitrate. This ther- mally stabilizes vanadium up to 1700 OC12 and 1500 OC15 in water12715 and in geological samplesl2, respectively.A mixed modifier consisting of ascorbic acid, 8-hydroxyquinoline and magnesium nitrate improved the precision and sensitivity from different atomization surfaces.9 Palladium was used in the determination of vanadium in acid digests of soils in order to eliminate the interferences caused by the matrix.7 The aim of the present study was to test some new chemical modifiers for the determination of vanadium in water and oil samples. Vanadium is often determined in oil matri~es6.16.1~ by AAS or ICP spectrometry. Platinum and rhodium gave comparable and sometimes better results than the widely used modifier, Mg(N03)2, especially in the determination of vana- dium in an oil matrix.Interferences from common chloride salts were investigated and NH4SCN and ascorbic acid were tested as to whether they could eliminate these interferences and the results are reported. The ageing of the surface of the graphite tube was found to decrease the vanadium sensitivity so that recalibration, using integrated absorbance, was necessary. Experimental Apparatus A Perkin-Elmer Model 5000 atomic absorption spectrometer equipped with an HGA-400 graphite furnace was used for the atomic absorption measurements. Pyrolytically-coated graphite tubes were used throughout the study. A 20 yl volume of the sample solutions was dispensed in the graphite tubes with an AS-1 autosampler and a 5 p1 volume of the modifier solution was injected with an Eppendorf (Westbury, NY, USA) micropipette with disposable poly(propy1ene) tips. The in- strumental settings and the graphite furnace programme are summarized in Table 1.Reagents All chemicals used in this study were of analytical-reagent grade. All glass- and poly(propy1ene)-ware were stored in 10% v/v HN03 (1.6 moll-l) for at least 12 h and then rinsed with 1 % v/v HN03 (0.16 moll-]) and distilled water before use. The acids were of Suprapur grade (Merck, Darmstadt, Germany). Vanadium standards were prepared by diluting a 1000 pg 1-1 stock solution of vanadium (as VOSO4) (Titrisol, Merck) with de-ionized distilled water and acidified to a final HN03 concentration of 0.1% v/v.Stock solutions of modifiers were prepared by dissolving appropriate amounts of their salts in acid media and diluting to a final volume with water. The modifiers studied were: Mg, Ca, Sr (as nitrates); and Rh, Pd, Pt (as112 Analyst, February 1996, Vol. 121 chlorides). All these metals are isomorphous with vanadium.'g A 10 g 1-1 NhSCN solution was prepared from solid NH4SCN (Merck). Ascorbic acid solution (10 g 1-1) was prepared from L- (+)-ascorbic acid powder (Merck). Solutions (10 g 1-I) of each chloride salt (NaC1, CaC12 and FeC13) were prepared from the solid salts (Merck). A Conostan oil multi-element standard (Conostan, Ponca City, OK, USA) was used for experiments on the determination of vanadium in oil. Procedures Interferences by common chloride salts The effect of NaCl, CaC12 and FeC13 on the atomic absorption signal of 1 ng of vanadium was investigated, by introducing increasing amounts of these salts in the graphite tube alone, and in the presence of 100 pg of NH4SCN or ascorbic acid.The temperature programme in Table 1 was followed at a tpyr of 1100 "C. The tpyr of 1 ng of vanadium with 200 pg of NaC1,lOO pg of CaC12 or 150 pg of FeC13 was investigated for an explanation of the interference mechanism. Tube Lifetime To study the effect of ageing of the graphite tube surface on the atomic absorption signal of 1 ng of vanadium, both the integrated and the peak absorbance was measured with increasing atomization cycles in the absence and in the presence of 20 pg Mg(N03)~ or 1 pg of Rh.Comparison of Chemical Modifiers To choose the maximum permissible pyrolysis temperature (tpyr) and the optimum mass of the modifiers, 5 p1 of the modifier solution was dispensed into the graphite tube, followed by 20 pl of the vanadium solution (1 ng per injection). Calibration curves were constructed by injecting 20 p1 of standard solutions containing 5.0, 10.0,20.0,30.0,50.0 and 100 pg 1-1 of vanadium in the graphite tube. The characteristic mass, m, (pg), was calculated from the slope (b) of the standard curve, using the equation m, = 0.0044 X 20/b for a sample volume of 20 pl. The LOD (pg 1-I) was calculated from the equation LOD = 3 X sBL/b, where SBL was the standard deviation of ten blank firings. Table 1 Instrumental operating conditions and temperature programmes for the determination of vanadium Spectrometer- Wavelength 318.5 nm Bandwidth 0.7 nm Lamp current 25 mA BG corrector D2-ON Sample volume 20 pl Modifier volume 5 p1 Temperature programme- Ar flow Tempera- Ramp Hold rate/ ture/oC time/s time/s ml min-' Reading Drying 120 10 20 300 Pyrolysis Variable* 10 20 300 Cool-down 20 1 14 300 Atomization 2650 0 7 0 ON Cleaning 2650 1 2 300 * Variable tPYr was 1100 "C without modifiers, 1500 "C with 20 pg Mg(N03)2, 1600 "C with 2.5 pg Pt and 1700 "C with 2 pg Rh.Determination of Vanadium in Oil A Conostan standard (S-21), containing 100 mg kg-1 of vanadium, was used throughout this study. A portion of 0.05 g of this standard was weighed accurately and diluted with IBMK (isobutyl methyl ketone) to a final volume of 5 ml, giving a stock solution of 1 mg 1-l.The working standards were prepared from the stock solution with appropriate dilution with IBMK. A calibration curve was constructed as described above. The modifiers tested were magnesium nitrate and platinum. The instrumental parameters and temperature programme in Table 1 were followed. The tpyr without modifiers was 1000 OC, whereas with 1 pg of Pt it was 1400 "C and with 20 pg Mg(N03)~ it was 1500 "C. Results and Discussion Interferences by Common Chloride Salts During preliminary experiments comparing chemical modi- fiers, the chloride salts of Pt and Pd decreased the signal of V. The same effect was observed in a previous study for Cr and it was found that 20 yg of NHdSCN had to be added in order to overcome this problem.19 Therefore, NH4SCN and ascorbic acid were tested as interference suppressors for common chloride salts, such as NaCl, CaC12 and FeC13 in the determina- tion of vanadium. The influence of increasing masses of ascorbic acid and NH4SCN was investigated in order to assess whether there was a positive or negative effect on the V signal. It was found that masses of these compounds up to 200 pg influenced neither the signal nor the pyrolysis temperature of V. The only effect was a small shift of the signal to earlier times with increasing amounts of both compounds. Hence, 100 pg of each compound was used in this study. The effect of increasing amounts of these compounds on the V signal is shown in Fig. 1. It is apparent that 100 yg of ascorbic acid could eliminate these interferences.Ammonium thiocya- nate showed less tolerance to interferences. Fig. 1 (a) shows that masses of NaCl higher than 50 pg depressed the V signal, even at a tpyr of 1100 "C. Fig. l(b) shows that CaC12 interfered with V in a different way. The signal of V increases with increasing amounts of CaC12. Fig. l(c) shows that FeC13 decreased the signal of V, even at 5 pg. In order to elucidate any possible mechanism of their action, the tpyr of 1 ng of vanadium was measured in the absence and presence of these salts. The results are presented in Fig. 2. Irregular peaks, with overcorrection, were obtained at tPYr <900 "C in the presence of NaC1. Therefore, NaCl is not removed from the graphite tube at this temperature which is consistent with the literature.2&Z2 How- ever, the signal was not recovered at a higher tpyr, indicating that a portion of V was lost during the ramp time of the pyrolysis step. This is consistent with the calculations of Frech et al.23 who showed that in the presence of chloride salts the volatile vanadium chloride compounds (VOC13 and VC14) were more stable than any condensed V compound and were removed from the furnace during pre-atomization steps.Byrne et aZ.21 showed that even at tpyr of 1050 OC, a temperature at which the excess of NaCl was removed, the interference on the Mn signal was still observed, because a small amount of NaCl was released simultaneously with the analyte at the beginning of the atomization step. Only 30% of the V signal was recovered in the presence of 200 pg NaC1.22 As no pyrolysis step was used in this study, it was concluded that vapour-phase interferences were responsible, since V showed less tolerance to interferent : ana- lyte mole ratio and has the higher M-C1 dissociation energy (477 kJ mol- 1).22 However, different behaviour was observed with the other two salts.Both salts were expected to hydrolyse during theAnalyst, February 1996, Vol. 121 113 pyrolysis step. The losses at low tpyr were expected to originate from the co-volatilization of the analyte and its expulsion from the furnace together with the HCl(,, generated by the hydrolysis reacti0n.2~ At a tpyr of 1100 "C the atomic absorption (AA) signal of V increased in the presence of CaC12 and decreased in the presence of FeC13 (Fig.1). The background absorption in the presence of CaC12 was elevated and increased with increasing amounts of CaC12. Irregular peaks of V were observed with a dip in the peak absorbance. This effect was present up to a tpyr of 1600 "C where V losses began to occur (Fig. 2). In order to be certain whether this effect was due to chloride or calcium compounds, 150 pg Ca(N03)2 (containing the same amount of Ca, 36 pg, as 100 pg of CaC12) was injected into the furnace with 1 ng of vanadium at tpyr of 1100-1500 "C. There was no interference or background absorption. We therefore assumed that chloride was present, even at an elevated tPYr, raising the background absorption at the atomization stage. The presence of chloride (from MgC12) in the atomization step at a rpyr of 900 "C was observed by Byrne et al.24 with ETV-ICP-MS, even though at this tPYr chloride has been removed from the furnace.Littlejohn25 also observed an interference from MgC12 at high tPYr, although chloride was removed from the surface of the furnace and determined by ion chromatography. Iron chloride showed different behaviour; no background absorption was observed, although the shape of V signal was altered. A decrease in the AA signal of V was noticed at a tpyr of 700-1200 "C, followed by a small increase up to 1600 "C. Frech et al.*3 reported that high amounts of Cl,,, and FeC12(,, were present in the gas phase during the pyrolysis and vaporization of 100 yg FeC13. Therefore, in the presence of these compounds, formation of volatile vanadium+hloride compounds was expected.However, both compounds, CaC12 and FeC13, stabilized V up to 1600 "C (Fig. 2), probably acting as chemical modifiers in the form of their oxides.20 0.00 j 1 , , , , , 1 0 88 175 263 350 Sodium chloride mass I p g 0.30 7 0 s B 2 C 2 0.20 0 30 M) 90 120 I50 180 Calcium chloride mass /pg 0.30 I I 8 0.24 P C p 0.18 B 0 A E 0.06 3 0 50 100 IS0 2W 250 Iron (111) chloride /pg Fig. 1 Influence of increasing amounts of (a) NaCl, (h) CaC12 and (c) FeC13 on the integrated absorbance of 1 ng of V without (A) and with 100 pg of NaSCN (B) or 100 pg of ascorbic acid (C). Ammonium thiocyanate (100 pg) may partly remove these interferences. It is likely that NH4Cl was formed during the pyrolysis step, and higher masses were necessary to remove the interferences completely.However, this was not investigated further. Ascorbic acid suppressed the interferences completely. Early in the pyrolysis step ascorbic acid suppressed the interference from NaCl by the formation of HCl(,).21 At a higher temperature, the carbon residue of its pyrolysis removed the remaining chloride.2l A different mechanism was proposed for its action with MgC12. Ascorbic acid retarded the hydrolysis of this ~ a l t , 2 ~ eliminating the expulsion of the analyte with the forming HCl. This mechanism is also applicable to other hydrated salts which undergo hydrolysis during the pyrolysis step (such as CaC12 and FeC13). In a recent study, 5 pg of CaC12 at a tpyr of 1000 "C and tatom of 2200 "C produced high molecular absorbance by CaC1(,).26 This may be the reason why background absorbance appeared in the presence of CaC12 in this study.It could also have been due to the extensive light scattering measured in the 300-350 nm range during the vaporization of 5 pg of CaC12.27 Ascorbic acid eliminated this effect and a small background absorption appeared only at 150 pg CaC12, which was corrected. The presence of ascorbic acid clearly reduced this light scattering.27 In the presence of NH,SCN, background absorption appeared at 100 pg CaC12. Nevertheless, the mechanisms of the action of the interfering compound and modifiers require more investigation and current work is in progress. Effect of Ageing on Graphite Tubes It was observed that the condition of the surface of pyrolytic- ally-coated tubes greatly affected the AA signals of V.After successive firings at 2650 "C, a gradual deterioration of analytical results appeared. The effect of increasing atomization cycles on the peak and integrated absorbance of 1 ng of V is shown in Fig. 3. Both peak- and integrated absorbance were constant for 80 atomization cycles. More firings resulted in a decrease of AA signal. The peak height was more constant in the absence of chemical modifier, but integrated absorbance was more constant in the presence of 1 ,ug of Rh. The same trends were observed in the presence of 1 pg of Pt. Similar results have been reported.7-10 This effect is believed to appear due to degradation of the pyrolytic coating with increasing atomization cycles, especially at such a high temperature. The deterioration of this coating promoted the formation of vanadium carbides. This was reflected on the peak profiles of V, which appeared later and the rate of atom formation was lower for the old tubes.The tpyr curve was constructed with an old tube (about 200 firings) and compared with the curve of a new one. The results are shown in Fig. 4. The sensitivity was lower but the tpyr could increase up to 1600 "C without significant loss of the signal. In addition, the rate of losses at higher temperatures 0.3 I I . c1 0.0 500 700 900 1100 1300 1500 1700 1900 2100 Pyrolysis temperature /"C Fig. 2 (B), 100 pg of CaCI2 (C) or 150 pg of FeC13 (D). Pyrolysis curves of 1 ng of V without (A) and with 200 pg of NaCl114 Analyst, February 1996, Vol. 121 0.30 8 0.20 - B e .sf n .t 0.10 - 0.00 is lower. These results gave more evidence to the hypothesis of carbide formation. Nevertheless, each tube could be used for up to 250 measurements, but after 100 measurements periodic recalibration (using integrated absorbance measurements) was necessary in order to obtain accurate and precise results. (a) 9 , , , , , , , , , , , , , , , , , , 1 , , 1 Comparison of Metallic Modifiers Optimization of modifier mass and temperature programme In the first part of this study we compared the maximum pyrolysis temperature (tpy,) and the sensitivity obtained when all the metals mentioned in the Experimental section were used as chemical modifiers. Calcium and Sr increased the pyrolysis temperature up to 1400 "C but enhanced the peak tailing and shifted the appearance temperature to lower values.As no advantages were observed in comparison to Mg, these modifiers were not tested further. Palladium chloride decreased the AA signal of V. The signal was recovered by the addition of 20 pg of ascorbic acid. An amount of 10 pg of Pd with 20 pg ascorbic acid increased tpyr to 1400 OC, but did not increase the sensitivity. Thus, this modifier was not tested further. Mg(N03)2, Pt and Rh appeared to thermally stabilize V and increased the sensitivity. These modifiers were chosen for further study. 0.1 & 0 50 100 150 Atomization cycles 0.15 1 150 50 100 Atomization cycles Fig. 3 Influence of increasing atomization cycles ( a ) on the peak absorbance and (b) on the integrated absorbance of 1 ng of V in the absence (A) and in the presence of 20 yg of Mg(NO& (B) or 1 pg of Rh (C).Each point represents the average of ten measurements. 0.3 0.0 b l 900 1200 1500 1800 2100 Pyrolysis Temperature /"C Pyrolysis curves of 1 ng of V from a new graphite tube (A) and a Fig. 4 used tube (B). The influence of the mass of the modifiers on the V sensitivity and tpyf was investigated. The modifier mass influences the sensitivity significantly, high modifier masses decreasing the signal due to secondary adsorption at the cooler ends of graphite tube,28329 but stabilizing the analyte to higher temperatures.29.30 Therefore, a careful optimization of the modifier mass was deemed necessary in order to achieve the maximum of sensitivity and thermal stabilization. The influence of increasing masses of Mg(NO&, Pt and Rh on the V signal at a tpyr of 1500 "C is shown in Fig.5 . These experiments were carried out in the presence of 20 pg of ascorbic acid in order to ensure that chlorine was driven off the tube and reduction to the pure metal was achieved. The integrated absorbance was recovered with 10 pg Mg(N0& and with 0.25 pg of Pt or Rh. However, the modifiers showed a different effect on the peak height. Magnesium nitrate recovered the V signal in an optimum range of 10-50 yg. An amount of 20 pg was used for the remaining study. Experiments with Pt and Rh showed that the peak height signal of V was higher at 2 yg of these metals. The signal of V was decreased gradually when masses greater than 3 pg were used, but this decrease was more obvious with peak height measurements. These phenomena were ascribed to the trapping of analyte atoms at the cooler ends of the tube in the presence of modifier~.~833* The effect of different modifier masses on the pre-atomization losses of V at a pyrolysis temperature of 1500 "C was examined with a new graphite tube !? 0.25 4 i I 0.05 1 1 0.0 1.0 2.0 3.0 4.0 5.0 Platinum mass lpg 0.4 A (4 0.1 I ' I I ' 1 ' I .I ' 0.0 1.0 2.0 3.0 4.0 5.0 Rhodium mass Ipg Fig. 5 Influence of increasing amounts of (a) Mg(NO&, (b) Pt and (c) Rh on the peak height (A) and integrated absorbance (B) of 1 ng of V at a pyrolysis temperature of 1500 "C. Vanadium solution included 20 yg of ascorbic acid in the experiments with Pt and Rh.Analyst, February 1996, Vol. 121 115 and it was found that these losses were decreased even when 0.25 yg of Rh was present and V could be pyrolysed for 10 s at 1500 "C without significant losses.The pyrolysis (tpyr) and atomization temperatures (tat,,) of V determination were optimized in the absence and in the presence of the above chemical modifiers with a new graphite tube. The pyrolysis and atomization curves from peak ab- sorbance measurements are presented in Fig. 6. The same trends were observed with integrated absorbance measurements. When no chemical modifier was present, the maximum t p y was 1100 "C. In the presence of 20 pg of Mg(N03)2 the maximum tpyr was 1500 "C. An amount of 2 pg of Rh could stabilize V up to 1700 "C. Platinum (2.5 pg) increased tpyr up to 1600 "C. The tatom was 2650 "C. The appearance temperature of V increased substantially in the presence of Rh (Fig.6). The same was observed in the presence of Pt. The first V atoms were produced at about 1800 O C , without modifiers or in the presence of Mg(N03)2. In the presence of 2 yg Rh this temperature was about 2200 "C. This means that V atoms require much more energy to be released from the Rh or Pt matrix. The low tPYr (1 100 "C) which could be applied with new tubes in the absence of modifiers could be ascribed to V losses as volatile oxides (VO and V02). These oxides were observed in the gas phase with MS323 and were the precursors of V atoms. Manning and Slavin12 reported that oxygen reduced the temperature at which V was lost and recommended that tpyr was decreased to 1100 "C in order to avoid such losses.They recommended the use of Mg(N03)2 as chemical modifier. The increased thermal stability of V in the presence of the tested modifiers could be ascribed to the prevention of formation of these volatile V oxides. If the losses occurred at the active sites on the graphite surface, blockage of these sites by the modifier species and atomization through the bulk of the modifier mass is a possible explanation. The atomization signals for I ng of V in the presence of different chemical modifiers are shown in Fig. 7. It is apparent that Mg(N03)2 did not influence the peak characteristics of V significantly. It is not expected that the mechanism of atomization of V will change in the presence of this modifier. In the presence of Pt, the appearance of V atoms was delayed and the atomization rate increased substantially, shown by the increase of the sharpness of the peak.The same was observed for Rh. This could be attributed to the formation of intermetallic bonding and/or to a formation of a compact solid solution during the pre-atomization step. When the atomization temperature was high enough this solid solution was damaged and V atoms were produced rapidly. It is interesting to observe that the difference between appearance time and peak time is only 0.2 s in the presence of 1 pg of Pt, 0.52 I I t \ / B f 3 0.09 a -0.02 800 1100 1400 1700 2000 2300 2600 2900 Temperature 1°C 'o.20L Fig. 6 Pyrolysis curve (hollow symbols) and atomization curve (filled symbols) of 1 ng of V in the absence (A) and in the presence of 20 pg of Mg(NO& (B) or 2 pg of Rh + 20 pg of ascorbic acid (C) from a new graphite tube.Similar trends were observed with integrated absorbance measurements. A tat,, of 2650 "C was used for the pyrolysis curve and a tpyr of 1100 "C was used for the atomization curve. whereas without modifiers or with Mg(N03)2 this time is approximately 0.6 s. However, integrated absorbance did not change significantly. This meant that the atomization efficiency remained constant. From the peak profiles, extensive tailing (even in the presence of modifiers) was observed. This is consistent with the formation of vanadium carbides. It was shown that vanadium carbides are produced at temperatures of 1500 "C and precede the formation of V(g).*,3,23 This mecha- nism could be applied both in the absence and in the presence of Mg(N03)~. In the absence of Pt or Rh, vanadium carbides could not be formed before the atomization of V, because the carbides are thermally stable compounds and dissociate slowly, whereas in the presence of Pt or Rh sharp peaks were observed.However, V exhibits high affinity with carbon and after the production of V(g) atoms, re-adsorption or condensation could lead to formation of carbides. Platinum and Rh atomized at the tatom of 2650 "C permitting V to interact with carbon at this high temperature. No memory effects were observed due to carbide formation, because the high internal argon flow at the cleaning step removed any residual of the refractory V compounds. Nevertheless, the mechanism of V in the presence of modifiers requires more investigation and should be the purpose of future work.Analytical figures of merit Calibration curves were constructed following the procedure described in the experimental session, and applying the temperature program given in Table 1. A moderately used graphite tube (around 100 atomizations) was used for this experiment. This was repeated several times and consistent results were achieved. These results are summarized in Table 2. The characteristic masses and limits of detection found were in 0.550 0.450 - 8 0.350 e 0.250 0.150 0.050 0 2 P -0.050 I I I 0 2 4 5 7 Atomization time i s Fig. 7 Atomization signals for 1 ng of V: A, without modifier, tpyr = 1100 "C, peak height 0.200, integrated absorbance 0.234 s; B, with 20 pg of Mg(NO&, tpyr = 1500 "C, peak height 0.215, integrated absorbance 0.253 s; and C, with 1 pg of Pt + 20 pg of ascorbic acid, tPYr = 1500 OC, peak height 0.442, integrated absorbance 0.248 s.Table 2 Sensitivity, LOD and precision of vanadium determination in aqueous solutions. PH = peak height absorbance and IA = integrated absorbance mo*lpg LOD+/pg 1 - 1 &* (%) Modifiers PH IA PH IA PH IA None 21.2 18.7 0.47 0.52 3.5 3.2 1 vg Pt§ 14.4 18.4 0.44 0.66 1.4 1.7 1 Rh 14.7 19.2 0.79 1.05 1.4 1.5 * Characteristic mass (mass of the analyte giving an absorbance of 0.0044). t Limit of detection (3aBL). * n = 9 at 1 ng V. § With 20 pg ascorbic acid. 20 pg Mg(N03)2 20.5 18.0 0.91 1.50 2.8 1.8116 Analyst, February 1996, Vol. 121 agreement with those previously rep0rted.~*J5,34 Better charac- teristic masses were realised with Rh and Pt, using peak height absorbance measurements.The s, of nine replicate injections of 1 ng of V was slightly improved in the presence of chemical modifiers. Determination of V in Oil Matrix Vanadium is determined in fuel or crude oil samples6J6.17 because it acts as 'catalyst poison' and causes corrosion to engines. It is also released into the atmosphere from oil-fired power station~.~O>~~ It is determined in lubricating oils as a marker of the wear of oil-lubricating machines. Therefore, there is a need for the development of a direct, fast, precise and accurate method for the determination of V in oil matrices that is free of interferences. The direct determination of V in an oil multi-element standard (diluted only with IBMK) was investigated with and without chemical modifiers. Only Pt and Mg(NO& were tested.The pyrolysis curves of V in an oil matrix are shown in Fig. 8. It was observed that without modifiers, V losses started at temperatures higher than lo00 "C. In the presence of 1 yg Pt, the tpyr could be raised up to 1400 "C and peak height absorbance increased significantly. Magnesium nitrate de- 0.3 I 1 \ a :::*\ 0.0 500 900 1300 1700 2100 2500 Pyrolysis Temperature/"C 0.2 , I 0 . 0 1 , I ' , , I ' , Pyrolysis Temperature / "C Fig. 8 Pyrolysis curve for 1 ng of V in oil standard diluted with IBMK using (a) peak height absorbance or (b) integrated absorbance measure- ments, without chemical modifiers (A) and with 20 yg of Mg(NO& (B), or with 1 yg of Pt (C).500 900 1300 1700 2100 2500 Table 3 Sensitivity, LOD and precision of vanadium determination in oil. PH = peak height absorbance and IA = integrated absorbance mo'lpg LOD+/pg 1-1 SP Modifiers PH IA PH IA PH IA None 27.5 25.0 2.3 2.1 7.1 7.7 20 pg Mg(N03)Z 41.1 27.8 3.1 2.4 10.6 8.7 1 pg PtS 15.3 24.0 2.8 2.4 7.4 3.4 * Characteristic mass (mass of the analyte giving an absorbance of 0.0044). + Limit of detection (3aBL). * n = 9 at 1 ng V. creased the AA signal of V in an oil matrix until a tpyr of 1300 "C was used. This stabilized V up to 1500 "C. Characteristic masses, limits of detection and reproducibility results are presented in Table 3. It is obvious that Pt improved the over-all performance of the method. However, the better s, obtained from integrated absorbance measurements in the presence of Pt could not be explained.In the determination of V, higher characteristic mass was found in oil- than in water-matrices. The thermal stability was also decreased, probably owing to formation of volatile V compounds during the pyrolysis step and/or the early formation of V carbides, which was enhanced in the presence of this organic matrix. Addition of ascorbic acid was not necessary for Pt, as the organic matrix probably promoted the early formation of metallic Pt. Conclusions Ascorbic acid and NH4SCN were tested as interference suppressors for common chloride salts. Ascorbic acid effec- tively eliminates these interferences. From the comparison of the metallic modifiers, it is concluded that Mg(N03)2, Pt and Rh thermally stabilized V, preventing pre-atomization losses.The peak absorbance sensitivity increased significantly in the presence of Pt and Rh. Platinum improved the determination of V in a standard oil matrix. Current work is in progress to elucidate the mechanism of the action of these modifiers. N.T. would like to thank the Greek State Scholarships Foundation for financial support for this work. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Slavin, W., Graphite Furnace AAS. A Source Book, Perkin-Elmer, Norwalk, CT, USA, 1984, pp. 148-149. Wendl, W., and Muller-Vogt, G., Spectruchim. Acta, Part B, 1985, 40, 527. Matousek, J. P., and Powell, H. K. J., Spectrochim. Acta, Part B , 1988, 43, 167. Studnicki, M., Anal. Chem., 1980, 52, 1762.Tominaga, M., and Umezaki, Y., Anal. Chim. Acta, 1982, 139, 279. Barbooti, M. M., and Jasim, F., Tafanta, 1982, 29, 107. Lechotycki, A., J . Anal. At. Spectrom., 1990, 5, 25. Dymott, T. C., Wassall, M. P., and Whiteside, P. J., Analyst, 1985, 110, 467. Pantano, P., and Sneddon, J., Appl. Spectrosc., 1989, 43, 504. Carneiro, M. C., Campos, R. C., and Curtius, A. J., Talanta, 1993,40, 1815. Hulanicki, A., Karwowska, R., and Stanczak, J., Talanta, 1980, 27, 214. Manning, D. C., and Slavin, W., Spectrochim. Acta, Part B , 1985,40, 461. Buchet, J. P, Knepper, E., and Lauwerys, R., Anal. Chim. Acta, 1982, 136, 243. Bermejo-Barrera, P., Beceiro-Gonzalez, E., Bermejo-Barrera, A., and Bermejo-Martinez, F., Analyst, 1990, 115, 545. Bermejo-Banera, P., Beceiro-Gonzalez, E., and Bermejo-Banera, A., Anal. Chim. Acta, 1990, 236, 475. Murillo, M., and Chirinos, J., J . Anal. At. Spectrom., 1994, 9, 237. Bettinelli, M., and Tittarelli, P., J . Anal. At. Spectrom., 1994, 9, 805. Tsalev, D. L., Slaveykova, V. I., and Mandjukov, P. B., Fifth Colloquium Atomspektrometrische Spurenanalytik, ed. Welz, B., Perkin-Elmer, Uberlingen, Germany, 1989, pp. 177-205. Thomaidis, N. S., Piperaki, E. A., Polydorou, C. K., and Efstathiou, C. E., J. Anal. At. Spectrom., in the press. Chaundhry, M. M., Mouillere, D., Ottaway, B. J., Littlejohn, D., and Whitley, J., J . Anal. At. Spectrom., 1992, 7, 701. Byrne, J. P., Lamoureux, M. M., Chakrabarti, C. L., Ly, T., and Gregoire, D. C., J . Anal. At. Spectrum., 1993, 8, 599. Carroll, J., Miller-Ihli, N. J., Hamly, J. M., O'Haver, T.C., and Littlejohn, D., J. Anal. At. Spectrum., 1992, 7, 533. Frech, W., Lundberg, E., and Cedegren, A., Prog. Anal. At. Spectrosc., 1985, 8, 257.Analyst, February 1996, Vol. 121 117 24 25 Littlejohn, D., personal communication. 26 27 28 29 Byme, J. P., Chakrabarti, C. L., Gregoire, D. C., Lamoureux, M. M., and Ly, T., J. Anal. At. Spectrom., 1992, 7, 371. Thomaidis, N. S . , Piperaki, E. A., and Efstathiou, C. E., unpublished work. Shekiro, J. M. Jr., Skogerboe, R. K., and Taylor, H. E., Anal. Chem., 1988,60,2578. Frech, W., Li, K., Berlung, M., and Baxter, D. C . , J. Anel. At. Spectrom., 1992, 7, 141. Thomaidis, N. S., Piperaki, E. A., and Efstathiou, C. E., J . Anal. At. Spectrom., 1995, 10, 221. 30 31 32 33 34 Mandjukov, P. B., Vassileva, E. T., and Simeonov, V. D., Anal. Chem., 1992,64,2596. L’vov, B. V., and Frech, W., Spectrochim. Acta, 1993,48B, 425. Styris, D. L., and Kaye, J. H., Anal. Chem., 1982, 54, 864. Styris, D. L., and Redfield, D. A., Spectrochim. Acta Rev., 1993, 15, 71. L’vov B. V., Spectrochim. Acta, 1990, 45B, 633. Paper 51051 65G Received August 2,1995 Accepted August 21, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100111

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (119-125) 119 Gas Phase Detection of Cocaine by Means of Immunoanalysis* Torsten Ziegler", Oliver Eikenberg", Ursula Bilitewski"?? and Michael Grolb a Gesellschaft fur Biotechnologische Forschung (GBF) mbH, Department Enzymology, Mascheroder Weg 1,38124 Braunschweig, Germany 82377 Penzberg, Germany Boehringer Mannheim GmbH, R & D Diagnostic Reagents, Nonnenwald 2 , Immunoanalytical techniques based on an indirect competitive ELISA to determine cocaine in the gas phase are described. A test-gas generator was developed and evaluated by determining saturation vapour pressures and the sublimation enthalpy of cocaine base. To achieve quantitative recovery of the drug, the saturated air stream was sucked through a retention fluid. For validation of the test gas generator, samples were analysed with a microtitre plate cocaine-ELISA based on monoclonal antibodies.To simplify the sampling and analysis procedures by analysing the retention fluid directly in the sampling vessel, particle-based immunoassay formats were developed. According to the ELISA format, avidin was immobilized on the particles. The application of various solid particles consisting of different materials with a wide range of diameters (0.5-550 pm) to the analysis of cocaine standards and gas samples showed good correlation with the results obtained with the microtitre plate ELISA with an average IC50 value (end-point of the test; concentration at 50% binding) of 9.7 ng ml-1. The particle-based assays showed IC50 values in the range of 5-54 ng ml-l and signal background ratios ranging from 2 to 11.The application of particle-based assays for direct analysis of cocaine in the sample fluid was successfully performed with glass beads, precoated with avidin and a biotin-cocaine conjugate. Recovery of cocaine from the gas phase depended on the volume of sample fluid and on the geometry of the sample tube. Keywords: Cocaine; vapour pressure; sublimation enthalpy; enzyme-linked immunosorbent assay; particle-based immunoassay; avidin Introduction Owing to the stimulating effect of cocaine on the central nervous system, it is one of most frequently consumed drugs of abuse.' In view of the low volatility of cocaine, few approaches have been made for its detection from the gas phase. The vapour pressure of solid cocaine base was first determined by Lawrence et a1.2 and further investigation had shown that the sublimation vapour pressure of cocaine base was 1 X 10-5 Pa at 20 "C.3 The vapour pressure of the corresponding cocaine hydrochloride at the same temperature was 3 X 10-7 Pa.3 Numerous methods for cocaine analysis in the liquid phase have been described, usually investigating samples of medical or toxicological interest.Sensitive cocaine determinations have been performed with chromatographical methods such as GC,4 GC-MS5 or HPLC6 and with immunoanalytical techniques * Presented at The SAC '95 Meeting, Hull, UK, July 11-15, 1995. + To whom correspondence should be addressed. such as immunosensors7 and immunoassays.8~9 Compared with chromatographical determination methods which require pre- treatment of the liquid samples, such as extraction and derivatization of cocaine, immunoassays offer the possibility to determine cocaine in liquid samples without sample pre- treatment due to selective recognition of the analyte by means of highly specific antibodies.Application of GC to the analysis of vapours of cocaine and other drugs requires solid-phase extraction from the gas phase.2 Interesting approaches to direct and sensitive analysis of airborne vapours have been made by the application of ion mobility spectroscopy. Owing to the low resolution ob- tained with this ion separation technique, the application to the analysis of simple mixtures with well defined components has been proposed.12 The determination of drug vapours without enrichment procedures by means of biosensor devices was reviewed by Guilbault and Schmid.13 Considerable interest has been focused on the application of lipid-modified surface acoustic wave sensors to the direct detection of target molecules from the gas phase via non-specific hydrophobic inter- actions.14>15 The modification of piezoelectric sensors with specific antibodies has been applied to pesticide analysis and to the detection of cocaine in the gas phase.l6 The results reported by Rajakovic et al.17 on vapour analysis with antibody-coated piezoelectric devices could not prove a specific immunoreac- tion on the solid/gas interface.This led to the conclusion that determination of cocaine in the gas phase with immunoanalyt- ical techniques requires the presence of an aqueous environment during the immunoreaction.A new attempt at the application of immunoanalytical techniques to cocaine determination in the gas phase is being made by collecting cocaine vapours in the liquid phase and subsequent analysis with cocaine-specific immunoassays in the liquid phase. A test-gas generator was developed to produce vapours of cocaine. The performance of the vapour preparation was evaluated with reference measurements. Quantitative analysis of cocaine was carried out with a cocaine-ELISA. To simply vapour sampling and determination, the immunoanalyt- ical procedure was transferred from the microtitre plate into particle-based assays. Experimental Reagents and Materials All chemicals were of analytical-reagent grade and purchased from Merck (Darmstadt, Germany) unless stated otherwise.Water was de-ionized and doubly distilled. Cocaine base and hydrochloride, bovine serum albumin (BSA; fraction V), 3-aminopropyltriethoxysilane (APTS) and 1 -ethyl-3-(3-di- methylaminopropy1)carbodiimide (EDC) were purchased from Sigma (Deisenhofen, Germany). Avidin (> 78%; ex hen egg- white), N-(2-hydroxyethyl)piperazine-N-ethanesulfonic acid (HEPES), 2-aminoethanol and glutardialdehyde (GDA; 25%120 Analyst, Februaiy 1996, Vol. 121 solution in water) were obtained from Fluka (Neu-Ulm, Germany). N-hydroxysulfosuccinimide (Sulfo-NHS) was ob- tained from Pierce (Miinchen, Germany). Ethanol (chro- matographic grade) was supplied by Baker (GroB-Gerau, Germany). Glass beads (diameter = 220-550 pm) were from Clauss (Frankfurt, Germany), acid-washed glass beads (diam- eter = 150-212 pm) and avidin on acrylic beads were purchased from Sigma (Miinchen, Germany).Activated poly- (methylmethacrylate)-microcarriers (PMMA) were from Loewe Biochemika (Otterfing, Germany). Activated melamin- microcarriers were kindly supplied by U. Rothe, Martin-Luther- Universitat (Halle, Germany). Amino-coated magnetic particles Biomag 4100 were purchased as a particle suspension from Paesel and Lorei (Frankfurt, Germany). Anti-cocaine antibod- ies, streptavidin-coated microtitre plates and the other reagents for the cocaine-ELISA (see below) were supplied by Boehringer Mannheim (Penzberg, Germany). Test Gas Generator Apparatus The test-gas generator which produces vapours of low volatile compounds such as cocaine was developed in our laboratory and is schematically shown in Fig.1. This allowed the control of the drug concentration, humidity, temperature and flow rate of the gas phase. The entering flow of synthetic air was dried, filtered and divided into the three sub-flows: (i) humid air, which could be saturated with humidity by passing through a water-vessel and subsequently thermostated to process tem- perature with a Liebig-cooler, (ii) drug-saturated air, which was saturated with cocaine by flowing through a thermostated vessel containing cocaine base coated on Raschig rings and (iii) unmodified dry air to dilute the resulting gas mixture. The apparatus was thermostated in an incubation cupboard (1 520 X 1350 X 760 mm) obtained from Elsbrock and Partner (Scharmbeck, Germany). Glass fittings were sealed with PTFE fittings.Temperature control was performed with several nickel thin film thermocouples (GMS, St. Georgen, Germany). Humidity was controlled with a combined humidity-tem- perature sensor Testo 601 from Testotherm (Lenzkirch, Ger- many). The sub-flows were regulated by precision reduction Fig. 1 Schematic diagram of the test-gas generator for low volatile substances. Air was saturated with cocaine base by flowing through drug- coated Raschig rings. T = temperature sensor, H-T = humidity- temperature sensor, M = manometer, RV = reduction valve, F = flow meter, Th = thermostat, and LC = Liebig-cooler. valves from Fairchild (Winston-Salem, NC, USA) and con- trolled with inductive mass flow meters (Kobold, Hofheim, Germany).The gas left the generator through the funnel outlet and was transported through the sampling vessels with a GSA 50-1 suction pump (GSA, Neuss, Germany). Sampling devices were made from glass and are described below. Glass sample tubes (12 X 100 mm) for particle-based assays were purchased from Schott (Mainz, Germany). Smaller polystyrene tubes (12 x 70 mm) for microcarrier assays were from Nunc (Roskilde, Germany). The cocaine concentration in the retention fluid was determined by immunoassays (see below). Flow control of the suction gas flow was performed for flow rates from 0.25-1 1 min-1 with a rotameter (Kobold, Hofheim, Germany) and for flow rates > 1 1 min-1 with a rotameter manufactured by Rota- Jogawa (Wehr, Germany).Waste air from the outlet and the suction pump was transported to a laboratory hood to avoid contamination of room air. Sampling procedures For evaluation measurements, drug-saturated dry gas of differ- ent temperatures was sucked through a sample vessel (20 mm diameter X 300 mm), which was filled with 30 ml of ethanol. One end of the vessel was placed deep in the funnel outlet of the test-gas generator, the other end was connected to the suction pump. The suction gas flow entered the retention fluid through a glass capillary equipped with a frit tip submerged about 15 cm. The suction flow rates were varied from 0.25 to 5 1 min-1 and sampling times ranged from 80 to 400 min. To ensure complete retention, a control vessel of the same type was placed in the gas flow behind the sample vessel.After sampling, the cocaine solutions were transferred into 50 ml glass flasks, the ethanol removed in vacuum and the cocaine dissolved in 2-5 ml of the ELISA buffer [0.05 moll-' HEPES-buffered saline (HBS), pH 7.41 containing 5% ethanol. The cocaine solutions were diluted in glass tubes and 50 pl of the diluted solutions were transferred into the microtitre plates and analysed with the cocaine-ELISA (see below). The concentration of the undiluted cocaine solution was calculated from the average of the concentration values obtained with the corresponding dilutions producing optical densities in the linear detection range of the ELISA. Therefore, the determined concentrations of the diluted solutions were multiplied by the dilution factor.The mass of cocaine base mcoc collected in the vessels was obtained by multiplication of the concentrations of the undiluted cocaine solutions by the volume of HBS used for dissolving. Sampling with a solution containing carrier particles was carried out at a temperature of 30 "C in small glass tubes or polystyrene tubes with a reduced sample volume. The tubes containing particles precoated with cocaine-biotin conjugate and 0.5-2 ml of 0.05 mol I-' HBS (pH 7.4) were placed in the sample vessels used for the evaluation measurements. The cocaine-saturated gas was sucked through the fluid at a flow rate of 0.33 1 min-1 for 10-20 min. The cocaine concentrations obtained in the sample solutions were in the linear detection range of the particle-based immunoassays. For complete retention of cocaine from the gas phase, the capillary had to submerge at least 7 cm into the retention fluid.After sampling, particles were separated and the whole supernatant removed. One aliquot of 100 pl sample fluid was transferred back into the original sampling tube and into a control tube containing particles precoated with avidin and cocaine-biotin conjugate. Then the particle-based immunoassay was carried out as described below, starting with the addition of 100 pl of anti- cocaine antibody solution to the sample and control tubes. For reference measurements the cocaine solutions were stored at 4 "C and analysed with the microtitre plate cocaine-ELISA.Analyst, February 1996, Vol. I21 121 Immunoassays The immunoassay for the detection of cocaine was an indirect competitive ELISA.Avidin was coated on solid supports, i.e. beads of different materials (microtitre plates were purchased with a precoating of streptavidin) to allow binding of a cocaine- biotin conjugate. Cocaine from the sample competed with the immobilized-cocaine conjugate for the limited binding sites of the cocaine-specific monoclonal antibodies. Binding of solid- phase bound antibodies was determined with an enzyme- labelled secondary antibody. ELISA protocol for microtitre plates Flat bottomed polystyrene microtitre plates with a streptavidin precoating were coated with 100 microlitres per well of a cocaine-biotin conjugate solution (BZE-DADOO-X-Biotin BMO 15.360039 Ch.02, 5 ng ml-1 in 0.05 mol 1-1 HBS, pH 7.4).After incubation for 1 h at room temperature, the plate was washed three times with water containing 0.9% NaCl and 0.05% Tween 20 (washing buffer). In the competition step 50 1-11 of monoclonal antibody solution (MAK < BZE > M-094-10676 IgG(DE) Ch.08/1, 100 ng ml-1 in 0.05 mol 1 - 1 HBS, pH 7.4) and 50 pl of cocaine hydrochloride standard solution or the sample in the same buffer was added to each well and incubated at room temperature for 1 h. After a washing step, bound antibody was labelled with a secondary antibody labelled with horseradish peroxidase (PAK < M-Fcy > S-Fab(1S)-POD Ch. 12944700, PAB-HRP-conjugate) by incubation with 100 pl of a 40 mU ml-I (1 U = 16.67 nkat) PAB-HRP conjugate solution in 0.05 mol 1-1 HBS, pH 7.4, for 1 h at room temperature.After a final washing step, 100 1-11 of TMB- substrate solution BM-Blue (Boehringer Mannheim, Germany) was added to each well. The enzyme reaction was stopped after 30 min by adding 50 p1 of 2 mol 1-1 HzS04 per well. The absorbances in the microtitre plate were measured with an ELISA-reader v-max (Molecular Devices, Grafelfing, Ger- many) at 450 nm and 690 nm as the reference wavelength. The software packet Softmax (Molecular Devices) was used for data analysis. Immobilization of avidin to particles Avidin was immobilized on the carrier particles with different covalent procedures. Different amounts of particles were incubated in an avidin solution of 0.1 mg ml-1 in 0.05 mol 1-1 HBS, pH 8.5 containing 0.01% sodium azide (except for magnetic particles and PMMA-microcarriers).Glass beads. For covalent immobilization of avidin to glass particles, hydroxy groups on the glass surface were silanized with APTS, creating amino groups on the surface. The protein was linked to the silanized particles by using GDA, a bifunctional aldehyde. The procedure was modified on the basis of the procedure of Weetall.18 Glass beads (1 g) were pre- treated with 4 ml of HNO3 (10% in water) for 1 h at 90 "C. After repeated washing with water to pH-neutrality, the beads were dried overnight at 200 "C. Silanization was performed by addition of 6 ml of APTS (10% in dry acetone) and thorough shaking overnight. After repeated washing with acetone, the silanized beads were dried at 45 O C for 2 h and stored at room temperature until activation.For GDA-activation, 1 g of beads was incubated with 4 ml of a GDA solution (2.5% in 0.1 moll-' potassium phosphate buffer (PPB) pH 6, for 90 min under thorough shaking, and subsequently washed three times with the incubation buffer. For the avidin coupling, 1 g of activated beads was shaken overnight at room temperature with 4 ml of avidin solution. After repeated washing with 0.05 moll-' HBS, pH 7.4, the remaining activated groups were blocked by incubating the beads with 4 ml of a BSA-solution (1% in 0.05 mol 1-1 HBS, pH 7.4) for 1 h at room temperature. After repeated washing with HBS, the avidin-coated beads were stored at 4 "C in HBS containing 0.01% sodium azide. Magnetic beads. The magnetic particles (Biomag 4100) had amino groups on the surface and were coated with avidin according to the instructions of the manufacturer applying a GDA-activation.A 0.25 ml volume of Biomag 4100 suspension (containing 12.5 mg of amino-coated particles) was repeatedly washed with 0.5 ml of water containing 0.15 mol 1-1 NaCl (washing buffer). Separation of particles was carried out in a magnetic separation rack (Baker), GDA-solution (0.5 ml, 5% in 0.0 1 moll- 1 PPB, pH 7) was added to the beads and the mixture was thoroughly shaken for 3 h at room temperature. After repeated washing with 0.5 ml of washing buffer, a solution of 2.5 mg avidin in 1 mlO.01 moll-' PPB, pH 7 was added to the particles and shaken overnight at room temperature. The particles were separated and incubated with 1.25 ml of a 1 moll-' glycine solution (pH 8) for 15 rnin and subsequently washed with washing buffer.The avidin-coated particles were stored in 4 "C in 10 ml of PPB with a concentration of 1.25 milligrams of particles per millilitre. PMMA-microcarrier. The immobilization of avidin to acti- vated PMMA-microcarriers exploited the high density of carboxylic groups on the carrier surface (970 pmol m-2).19 Because of the solubility of PMMA in most organic solvents, an activation method suitable for aqueous solutions was applied. The rapid one-step activation with EDC and Sulfo-NHS was performed according to Johnsson et a1.20 Dry carriers ( 100 mg) were incubated with 5 ml of an aqueous solution of 0.1 rnol 1-1 EDC and 0.1 moll-' Sulfo-NHS. After 10 min shaking at room temperature, the activated microcarriers were separated by centrifugation (4000 rpm) and rapidly washed two times with 10 ml of water.For protein-coupling, 10 ml of a solution of 0. I mg ml-1 avidin in 0.01 moll-' acetate buffer (pH 5) was added and the mixture thoroughly shaken for 1 h at room temperature. After repeated washing with 10 ml of 0.05 mol of 1-1 HBS (pH 7.4) excessive activated esters were inactivated with 5 ml of 1 mol 1-1 of ethanolamine in 0.1 mol 1-1 acetate-buffer (pH 6). After thorough shaking for 30 rnin at room temperature and washing, the microcarriers were incubated with BSA-solution (1% in 0.05 moll-' of HBS, pH 7.4) for 1 h and subsequently washed with HBS. The carriers were stored at 4 "C in HBS containing 0.01 % sodium azide with a concentration of particles of 10 mg ml-1.Melarnin-rnicrocarrier. Melamin-microcarriers were pur- chased already activated with cyanuric acid. For covalent immobilization of avidin via the amino groups 50 mg of particles were suspensed in 10 ml avidin solution and shaken overnight at room temperature. I 9 After centrifugation (4000 rpm) and repeated washing with 0.05 mol 1-1 HBS, pH 7.4, the microcarriers were blocked with BSA and stored according to the procedure applied to PMMA-microcarriers. ELISA-protocol using avidin-coated particles The immunological procedure of the ELISA was transferred into particle-based assays and the concentrations of the solutions of anti-cocaine antibody and cocaine-biotin conjugate were optimized for each assay format, but proved to be same as those applied in the microtitre plate ELISA.Cocaine hydro- chloride standards for gas sampling experiments were measured in glass tubes or in polystyrene tubes in the case of PMMA and melamin-microcarriers. The avidin-coated particles were incu- bated with the assay solutions according to the microtitre plate ELISA-procedure. Amounts of particles used in the assays are shown in Table 1. Considering the dimensions of the test tubes, the volumes of the test solutions from the microtitre plate assay were doubled except for the magnetic particles. Washing steps were carried out with 400 pl of washing buffer. PMMA and melamin-microcarriers were separated by centrifugation, mag-122 Analyst, February 1996, Vol. 121 netic particles by means of a magnetic separation rack and other particles by gravitation. The supernatant fluid was removed with a capillary attached to a water jet vacuum pump.Generally the substrate reaction time was 30 min. In the case of the glass beads with a diameter of 220-550 pm and the melamin- microcarriers, the enzyme reaction was stopped after 15 min. For measurement, 100 pl of the solutions were transferred into microtitre plates and the absorption determined with the ELIS A-reader. Data analysis Data analysis was performed with the software packet Softmax (Molecular Devices). The absorbances were normalized accord- ing to: x 100 A - Aexcess AO - Aexcess (%) B/Bo = where A is the absorbance, Ao is the absorbance at zero concentration cocaine and AexceSS is the absorbance at excess concentration of analyte. Further analysis of the sigmoidal calibration curves was done by application of the 4-parameter model21 to the raw data according to: a - d 1 + (X/C)b + d Y = with y being the absorbance A and x the analyte concentra- tion.The parameters a and dare the upper and lower asymptotes of the sigmoidal curve and equivalent to AexceSS and Ao. Parameter b is the slope at inflection point and c is the concentration at inflection point. The parameter c is identical with the IC50 value (concentration of 50% binding) and used as a representative parameter for the sensitivity of the assay. Parameter b is often defined as the sensitivity of the assay.22 As parameters and concentration values x were calculated from the ELISA curves with cocaine hydrochloride standards, the values of the gas samples of cocaine base were corrected with the factor F [F = 303.4/339.8 (quotient of molecular masses)].Results and Discussion Evaluation of the Test Gas Generator Both cocaine base and cocaine hydrochloride possess very low sublimation pressures.23 Direct and indirect methods for the determination of vapour pressures have been de~cribed.~3 A direct determination is only applicable to substances of significantly higher vapour pressures. Therefore, an indirect determination method was chosen for our experiments, where Table 1 Results of cocaine ELISAs using microtitre plates and avidin-coated particles. Where d = diameter, n = number of standard curves, and IC50 = concentration at 50% binding Assay results Assay format Cocaine-ELIS A Glass beads (polystyrene microtitre plate) d = 220-550 pm Glass beads d = 150-210ym Acrylic beads (Sigma) d = 100-190ym PMMA-microcarrier d = 5-10ym Melamin-microcarrier d = 2 y m Biomag 4100 d = 0.5-1.5 pm Avidin layer adsorptive covalent (streptavidin) (1) APTS-silanization (2) GDA-activation (1) APTS-silanization (2) GDA-activation covalent covalent covalent (1) EDC-Sulfo-NHS activation (1) cyanuric acid- activation (1) GDA-activation covalent covalent Particles Per testlmg - 50 40 0.5 0.5 0.5 0.025 ICs0/ng ml-' 9.7 k 3.6 54 f 12 14.7 k 2.1 5.9 5.0 f 0.8 13.4 fi 6.2 12.0 fi 1.3 Signal-to- background ratio 9.2 k 1.7 10.7 k 3.2 7.6 f 2.4 1.8 5.7 fi 1.8 5.8 f 0.8 2.5 k 0.0 Table 2 Determination of sublimation vapour pressure of cocaine base at 20 "C at different gas flow rates and sample gas volumes.The mcoc was calculated from the cocaine concentrations in the vessels determined by ELISA Suction flow rate/l min-1 0.25 1 1.5 3 3 5 5 Total gas volume/l 100 400 400 400 400 400 400 Sample vessel 256 552 1028 928 620 558 823 Control vessel 10 5 26 24 17 25 10 Average value cgas*/ng 1-' 2.66 1.39 2.69 2.38 1.59 1.46 2.08 2.04 -I 0.52 Pc,J10-5 Pa 2.14 1.12 2.16 1.91 1.28 1.17 1.67 1.64 k 0.42 * Calculated from mcoc in both vessels.Analyst, February 1996, Vol. 121 123 an inert carrier gas (i.e. air) passed along the solid cocaine base dissipated on Raschig rings. Dry air was saturated with cocaine base at different temperatures, collected in sample vessels and subsequently analysed with the cocaine-ELISA. The vapour pressure was determined indirectly from the collected amount of cocaine (see below).In Table 2 data obtained for various gas flow rates and total gas volumes of 100 1 and 400 1 at 20 "C are summarized. The mass of cocaine base mcoc found in the sample and control vessels was determined with the microtitre plate ELISA (see above). It could be shown that more than 95% of cocaine base was already collected in the first sample vessel even at the highest flow rate of 5 1 min- l. The gas phase concentrations of cocaine base cgas were calculated from the mass of cocaine base mcoc in both sample vessels and the total volume of gas sucked through the fluid. Assuming ideal behaviour, the partial pressure of cocaine base P,,, was obtained from: (3) where T = 293.16 K is the absolute temperature, R = 8.314 J mol-1 K-l is the universal gas constant and M = 303.4 g mol-l is the molar mass of cocaine-base.At 20 "C the mean value of the sublimation vapour pressure of cocaine base was 1.6 k 0.4 X 10-5 Pa equivalent to a gas phase saturation concentration of 2.0 k 0.5 ng 1- l. The obtained data correlated well with data known from literature determined by other methods to be 1.2 X 10-5 Pa and 1 X 10-5 Pa at the same temperature.2.3 Investigations at different temperatures allowed the deter- mination of the sublimation enthalpy of cocaine base AHsub according to the equation of Clausius-Clapeyron: A In P,,, = - - R (4) When In P versus T-1 is plotted, a straight line results with a slope of -15293 K (r2 = 0.995) leading to a sublimation enthalpy of 127.2 kJ mol-1.The obtained value again correlated well with literature data of 112.3 kJ mol-l.2 These results demonstrated that the test gas generator was suitable for the generation of cocaine saturated gas phases, that cocaine vapours were retained in the sample vessels and that cocaine concentra- tions could be determined via ELISA. Immunoassays Microtitre plate ELISA The ELISA for the determination of cocaine was based on an indirect competitive assay principle, i.e. a cocaine-conjugate was immobilized. To achieve a homogenous immobilization on the surface a cocaine-biotin conjugate was bound to streptavi- din-coated microtitre plates. A typical calibration curve of the microtitre plate cocaine-ELISA is shown in Fig. 2(a), charac- terized by the IC50 value which was determined to be 9.7 ng ml- with an average signal-to-background ratio of 9 (defined as the ratio of maximum absorption A.and the substrate blank absorption). Particle-based immunoassays The indirect ELISA principle was adapted to assay formats with avidin-coated particles, to allow a cocaine determination with precoated particles directly in the retention fluid. Avidin is a tetramer containing four identical sub-units of relative mol- ecular mass 15 000 Da each and four high-affinity binding sites for biotin with dissociation constants of approximately 10- l5 mol- 1.24 Streptavidin which was coated on the microtitre plates has an almost identical structure and binding affinity for biotin, but a lower relative molecular mass of 40000 Da.24 The strength of the avidin-biotin interaction allowed a stable precoating of microtitre plates and particles with the cocaine- biotin conjugate.Avidin was chosen for the immobilization procedures due to its lower price. Regarding the different chemical surface properties of the particles, different covalent immobilization procedures was applied. The acrylic beads were commercially available with a covalent coating of avidin. The particle-based immunoassays were optimized with respect to the amounts of particles per test and the concentra- tions of cocaine-biotin conjugate and cocaine-antibody solu- tions to achieve low IC50-values in combination with sufficient absorptions. Typical sigmoidal standard curves of the particle- based immunoassays are shown in Fig. 2(a) and (b). The amounts of particles used in each test tube are summarized in Table 1.Except for the porous acrylic beads, increasing particle size correlated with an increasing amount of particles per test. The optimized concentrations of cocaine-biotin conjugate and anti-cocaine antibody solutions were identical with the concen- trations used in the cocaine-ELISA. It can be seen from the results summarized in Table 1 and the standard curves in Figs. 2(a) and (b) that the IC50 values of the particle-based assays were generally comparable to those obtained with the microtitre plate assay. The PMMA-micro- carrier assay was the most sensitive assay with an average IC50 value of 5 ng ml-1, while the assay with the big glass beads revealed the highest IC50 value of 54 ng ml-1. Porous avidin- coated acrylic beads showed a low IC50 value.However, they 100 80 60 40 20 0 % 0.08 1 10 100 1000 L l T 100 80 60 40 20 0 I- I . I , . . . , . I I I I . , , . , I I I , .,- 1 10 100 1c Cocaine hydrochloridehg ml-' 10 Fig. 2 Sigmoidal standard curves of cocaine hydrochloride with ELISA and particle-based assays. (a) A = acrylic beads, B = ELISA, C = glass beads (150-210 pm), D = glass beads (220-550 pm) and (b) E = PMMA- microcarrier, F = melamin-microcarrier, G = magnetic particles.1 24 Analyst, February 1996, Vol. 121 were not considered for further experiments, due to the high background observed. Besides the assay results, handling characteristics had to be considered. The handling of the magnetic particles with fast magnetic separation and washing steps was remarkably easy, compared to the centrifugation required for the separation of the microcarriers. Glass beads sedimented in the vessels in a few seconds.Gas Phase Determination of Cocaine During the evaluation of the test generator, retention of cocaine base from the gas phase was performed in sample vessels containing 30 ml of ethanol. Owing to the high flow rates of up to 5 1 min-l, the ethanol was evaporated and had to be refilled in long term experiments. After sampling, the cocaine solution were transferred into flasks and the ethanol was removed in vacuum. For analysis with the microtitre plate ELISA, the cocaine in the flasks was dissolved in ELISA-buffer, diluted and the diluted solutions transferred into the microtitre plate. To simplify the sampling procedure and to reduce the sampling time cocaine determination was to be carried out directly in the sampling tube.For that reason the retention of cocaine was performed in small test tubes containing a suspension of the assay particles precoated with cocaine-biotin conjugate in ELISA buffer. The small dimensions of the sample tubes allowed us to reduce the sampling volume to 1 ml and the sampling time to 10 min. Gas was saturated with cocaine base at 30 "C with a gas phase concentration determined to be 7.0 k 2.1 ng 1-1. With a sampling time of 10 min and a suction flow rate of 0.33 1 min-1,23 ng of cocaine base were sucked through the retention fluid, the volume of which was varied from 0.5 to 2 ml. As can be seen from Table 3, using a sample volume of 0.5 ml led to a retention of only 25% and a minimum volume of 1 ml was required to achieve quantitative retention of cocaine base from the gas phase.Data from the sample tubes were compared to those obtained with control tubes and the ELISA, i.e. cocaine concentrations in the sample solutions were also determined by a particle-based assay in control tubes and by the microtitre plate ELISA. The agreement between all three determinations was good. Table 3 shows the cocaine recoveries observed with different particles using a sample volume of 1 ml. After sampling, each sample solution was analysed as described above in the sample tube, in a control tube and with the microtitre plate ELISA. The best results were obtained with the large glass beads. Not only quantitative retention of cocaine was observed, but also good agreement for all three cocaine determinations of the sample solution.For all other particles significant deviations between the three determinations or only partial retention of cocaine were observed. This can be explained partially with the standard deviations of the calculated saturation gas phase concentrations observed in the evaluation experiments (see Table 2), due to e.g., inhomogeneities in gas flow or substance losses during sample transfer. Other reasons can be difficulties with the handling of these small particles. It was observed that even with the lowest controllable flow rate of 0.33 1 min-l, the removal of particles from the sample tube could not be prevented. This led to changes in the amount of particle-bound cocaine-biotin conjugate in the sample tube and also to losses of sample fluid.Moreover, decantation of the supernatant was more difficult with small particles, i.e. particles may also be removed and transferred to control tubes and the microtitre plates and supernatant may not be completely removed from the sample tubes. Another parameter can be unspecific binding of cocaine base to porous and small particles. Therefore the use of non- porous glass beads with a diameter of 220-250 pm was preferable, which were easy to handle and gave reliable results. Conclusion The good agreement of sublimation vapour pressures and sublimation enthalpy of cocaine base with reference data demonstrated the applicability of the proposed indirect method for vapour pressure determination and the reliability of the cocaine determination with the microtitre plate ELISA.The ELISA format was successfully transferred to particle-based assays by immobilization of avidin on various types of solid supports. The detection ranges and signal-to-background ratios of the developed assays were comparable with those of the original microtitre plate assay. A simple procedure for the determination of cocaine base in the gas phase was developed, collecting the drug in a sample tube with particle-bound cocaine-biotin-conjugate and assay buffer as retention fluid. This allowed the performance of the immunoassay directly in the sample tube. From all particles best results were obtained with non-porous glass beads with a diameter of 220-550 pm.The authors wish to thank the Deutsche Aerospace AG Munchen, Germany, for technical advice, U. Rothe of the Martin-Luther-Universitat Halle, Germany, for kindly supply- ing activated melamin-microcarriers and D. Hanisch for his excellent technical assistance. Financial support of this work was obtained from the German Ministry of Education and Research. ~~~~~ Table 3 Effect of sample volume on the recovery of cocaine base analysed with a glass bead (220-550 pm) immunoassay and recoveries of cocaine base from the gas phase with different particle-based immunoassays with a sampling time of 10 min, comparing direct sampling, measurement of a control with the particle assay and reference measurements with the cocaine-ELISA. With Biomag 4100 sampling times were 5 , 10 and 20 min.n.d. = not detected Particles Concentration of cocaine*/ng ml-' Recovery (%) Retention Sample Control Reference Sample Control Reference volume/ml Expected tube tube ELISA tube tube ELISA Glass beads (220-550 pm) 0.5 46 10.1 10.7 10.1 21.9 23.2 21.9 I 23 20.5 23.2 17.9 88.0 99.5 79.2 1 .s 15 15.2 14.3 13.2 101.3 95.3 88.0 2 12 12.5 13.4 12.4 104.2 111.7 103.3 Glass beads (150-210 pm) 1 23 26.2 10.6 10.9 112.1 45.6 66.4 PMMA*-microcarrier 1 23 5.2 n.d. 23.8 22.4 n.d. 101.8 Melamin-microcarrier 1 23 16.5 8.2 18.8 70.8 35.2 80.4 Biomag 1 12 11.8 7.4 9.2 101.3 63.6 79.1 1 23 24.0 10.0 6.8 103.0 42.7 29.3 1 46 14.3 5.2 11.8 30.6 11.1 25.2 * Average of at least two determinations.Analyst, February 1996, Vol. 121 125 References 1 2 3 4 5 6 7 8 9 10 11 12 Hoffman, K.D., Pharm. Ztg., 1995, 140, 36. Lawrence, A. H., Elias, L., and Authier-Martin, M., Can. J . Chem., 1984,62, 1886. Hilpert, R., Binder, F., Grol, M., Hallermayer, K., Josel, H.-P., Klein, C., Maier, J., Oberpriller, H., Ritter, J., and Scheller, F., Proc. SPIE- Int. SOC. Opt. Eng., 1994, 2276, in the press. Javaid, J. I., Dekirmenjian, H. Davis, J. M., and Schuster, C. R., J . Chromatogr., 1978, 152, 105. Goenechea, S., Rucker, G., Neugebauer, M., and Zerell, U., Fresenius’ Z. Anal. Chem., 1986, 323, 326. Jatlow, P., and Nadim, H., Clin. Chem. (Winston-Salem, N.C.), 1990, 36, 1436. Ogert, R. A., Kusterbeck, A. W., Wemhoff, G. A., Burke, R., and Ligler, F. S., Anal. Lett., 1992, 25, 1999. Robinson, K., and Smith, R. N., J . Pharm. Pharmacol., 1984, 36, 157. Armbruster, D. A., Schwarzhoff, R. H., Hubster, E. C., and Liseno, M. K., Clin. Chem. (Winston-Salem, N.C.), 1993, 39, 2137. Eiceman, G. A., Snyder, A. P., and Blyth, D. A., Znt. J . Environ. Anal. Chem., 1989,38,415. Eiceman, G. A., Shoff, D. B., Harden, C . S., Snyder, A. P., Martinez, P. M., Fleischer, M. E., and Watkins, M. L., Anal. Chem., 1989,61, 1093. St. Louis, R. H., and Hill, Jr., H. H., Crit. Rev. Anal Chem.. 1900,21, 321. 13 14 15 16 17 18 19 20 21 22 23 24 Guilbault, G. G., and Schmid, R. D., Biotechnol. Appf. Biochem., 1991, 14, 133. Chang, S.-M., Tamiya, E., and Karube, I., Biosens. Bioelectron., 1991, 6, 9. Chang, S.-M., Ebert, B., Tamiya, E., and Karube, I., J . Biotechnof., 1990, 11, 1. Guilbault, G. G., and Luong, J. H., J . Biotechnol., 1988, 9, 1. Rajakovic, L., Ghaemmaghami, V., and Thompson, M., Anal. Chim. Acta, 1989, 217, 11 1. Weetall, H. H., in Methods of Enzymology, ed. Mosbach, K., Academic Press, New York, 1976, vol. 44, pp. 134-148. Rothe, U., Habilitation, Martin Luther Universitat Halle, Germany, 1990. Johnson, B., Lofas, S., and Lindquist, G., Anal. Biochem., 1991,198, 268. Chan, D. W., in Immunoassay: A Practical Guide, ed. Chan, D. W., and Perlstein, M. T., Academic Press, Orlando, 1987, pp. 1-23. Hock, B., Z. Wasser-Abwasser. Forsch., 1989, 22, 78. Kohlrausch, K., Praktische Physik, Teubner, Stuttgart, 1960, vol. 1, Hermanson, G. T., Mallia, A. K., and Smith, P. K., Immobilized Affinity Ligand Techniques, Academic Press, San Diego, 1992, Paper 5105888K Received September 6 , 1995 Accepted November 17, I995 pp. 363-366. pp. 41-50.

ISSN:0003-2654    

DOI:10.1039/AN9962100119

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (127-131) 127 Caesium Ion-selective Electrodes Based on Crowned Benzoquinones* Michael G. Fallon, David Mulcahy, William S. Murphy and Jeremy D. Glennont Department of Chemistry, University College Cork, Ireland Three new derivatives of benzo-15-crown-5 were examined as potential caesium ion-selective ionophores in PVC membrane electrodes. The ionophores were 2,3-benzoquino[ 15lcrown-5, 2-bromo-1,4-dihydroxybenzo[15]crown-5 and 5-bromo- 2,3-benzoquino[ 151crown-5. PVC membranes plasticized with 2-nitrophenyl octyl ether, and incorporating the secondary ion exchanger potassium tetrakis(4-chloropheny1)borate were used throughout this study. Interferent studies for each electrode were carried out to assess electrode selectivity in the presence of ammonium ion, alkali and alkaline earth ions and selected transition metals. A graphical method of representing selectivity coefficients is provided together with the separate solution technique for selectivity determination. Dynamic response, precision, lifetime and temperature studies was also reported. The benzoquinone crown ethers proved to be good caesium ionophores, displaying near-Nernstian responses to this ion, in the range 10-1-10-4.5 mol 1-1.The electrode incorporating the 2,3-benzoquino[ 15]crown-5 ionophore had the highest selectivity for caesium ions over alkali and alkaline earths studied. Keywords: 2,3-Benzoquino[l5]crown-5; 2 -bromo-l,4-dihydroxybenzo[l5]crown-5; 5- bromo-2,3-benzoquino[l5]crown-5; potentiometry ; caesium ion-selective electrode Introduction Crown ethers and related macrocyclic compounds have attracted great attention as artificial ionophores.' Their struc- tural properties have also led them to be successfully applied in areas as diverse as ion chromatography,* near-IR redox-active fluoroionophores3 and as supported liquid membrane (SLM)4 modifiers.A number of model systems have been developed as enzyme mimic^,^ in which several functional groups, appro- priately cited, cooperate in their specific reactions. The effect of structural modification to model crown compounds has been extensively investigated, particularly in the areas of optimizing solvent extraction efficiencies6 and electrode ionophore selectivities toward target analyte~.~ Crown ethers have been used for some time as neutral ionophores in Cs+ ion-selective electrodes (ISEs).Those based on 15-crown-5 functionality have produced excellent Nernstian response slopes, limits of detection (LODs), and short response times of less than 1 min. However, their selectivities for Rb+, K+, and Na+ were quite low.8 The bis(benzo-18-crown-6) derivatives seem to produce the best Cs+ electrodes,g since the cis-form is capable of forming sandwich-type Cs+ complexes. The electrode developed by Attiyat et al. 10 which incorporates * Presented at The SAC '95 Meeting, Hull, UK, July 11-15, 1995. + To whom correspondence should be addressed. a TMC-crown formazane produced (16,17-dihydro-5H, 15H- dibenzo[b,i] [ 1,11,4,5,7,8]dioxatetra-azacylotetradecin-7-car- bonitrile) a five- and twenty-fold selectivity over Rb+ and K+, respectively.More recently a dibenzo-24-crown-8 has been proposed as a promising Cs+ PVC-based ISE.11 The plasticizing agent was dibutyl phthalate (DBP) as opposed to the more frequently employed 2-nitrophenyl octyl ether (2-NPOE). Moreover, the influence of membrane components such as secondary ion exchangers,12 plasticizers l 3 and polymer base l 4 on electrode performance has been extensively examined in the literature. Crowned natural products15 are of interest in biological and pharmacological studies. Coenzyme Q (also called ubiquinone- 10, n = 10, Fig. 1) is an important redox carrier in the mitochrondrial respiratory chain. Crown ether derivatives of ubiquinone-0, n = 0 can be synthesized, by replacement of the two methoxy groups by oligo-ethylene glycol bridges.167'7 Such crowned benzoquinones are capable of multisite interactions via ion-binding and charge-transfer complexation, in addition to redox activity. Cyclic voltammetry studies of 2,3-benzoquino[ 15lcrown-5 interaction with alkali metals in N,"-dimethylformamide (DMF) has been reported.18 Anodic shifts of the Q/semi-Q peak (El) for K+ and Na+ were approximately +lo0 mV, while the shift for Li+ was +60 mV.Corresponding shifts for 1,4-benzo- quinone resulted in El potential shifts of +5, +20 and +40 mV, respectively.19~20 Therefore, cation-dependent E l potential shifts observed for it are consistent with complexation of the metal cations to the crown ether moiety, which in turn makes the quinone more labile towards reduction to the corresponding semi-quinone. It was also shown that ion-binding by the crown group will tend to increase the electron affinity of the quinone moiety and conversely, charge transfer complex formation (i.e., partial reduction of quinone) will tend to enhance the ion- binding ability of the crown group in a co-operative manner reminiscent of the positive effectors of allosteric enzymes.18 Macrocycles, such as those capable of such coorperative interactions, could lead to the development of novel potentio- metric sensors for ion and organic analysis.In this work, novel crowned ubiquinone derivatives are examined as ionophores in caesium-selective PVC membrane electrodes. Experimental Materials The ionophores synthesized and employed in this work are shown in Fig.2. The membrane components, PVC (ISE grade), 0 /I = 6-10 Chemical structure of ubiquinone. Fig. 1128 Analyst, February 1996, Vol. 121 potassium tetrakis(4-chloropheny1)borate (KTpClPB), 2-NPOE and analytical-reagent grade tetrahydrofuran (THF) were ob- tained from Fluka AG (9470 Buchs, Switzerland). The nitrate salts of silver(I), cobalt(r~), nickel(rr), copper(II), lead(II), cadmium(II), mercury(I1) were of analytical-reagent grade as were the chloride salts of the alkali and alkaline earth metals. All solutions and standards were made up in doubly distilled water. Synthesis of Benzo-15-crown5 Derivatives The crown-ether I was prepared according to slight modifica- tion of the procedures of Merz and Rauschel17 and Hayakawa et al.18,21 mp 122-124 "C (from propan-2-01) (literature value, 121-123 "C). Treatment of the quinone (I) with hydrobromic acid (47% aq, 1 .5 equiv.) gave the 2-bromo-l,4-dihydroxy- benzo[lS]crown-5 (11) as a white solid in 80% yield mp 123-1 25 "C (from hexane).This was readily oxidized to 5-bromo-2,3-benzoquino[ 151crown-5 (111) by stirring for 5 min with a 20% dispersion of potassium dichromate on silica with methylene chloride as solvent.22 The crowned bromobenzo- quinone was obtained as a red oil in 77% yield. For ionophore 11, found: C, 44.49; H, 5.16; Br, 21.22. C14H19Br07 requires: C, 44.33; H, 5.01; Br, 21.1 1; vmaX (KBr)/ cm-1 3385.8, 2865.5, 1485.4, 1447.8; 6H (270 MHz; CDC13) 3.71-3.76 (8H, crown H), 3.88-3.97 (4H, M, crown H), 4.254.30 (4H, M, crown H), 5.87 (2H, br s, D20 exchangeable 2 X ArOH), 6.87 (IH, S, ArH); bC (67.5 MHz; CDC13) 70.27 71.04 (CHz), 73.08 (CHz), 73.37 (CHT), 103.29 (C-Br), 113.32 (CHz), 70.42 (CHZ), 70.51 (CHz), 70.60 (CH2), 70.82 (CH2), (C-H), 139.40 (C-0), 139.86 (C-0), 140.42 (C-0), 143.17 (C-0).For ionphore 111, found: M+, 376.0079. C14H1779Br07 requires M, 376.0157; m/z 379, 378, 377, 376, 246, 231, 218, 216, 192, 165, 134, 82; Y,,, (film)/cm-' 2868.5, 1660.5, 1585.5, 1451.5; aH (270 MHz CDC13) 3.62-3.72 (8H, M, crown H), 3.80-3.84 (4H, M, crown H), 4.46-4.49 (2H, M, crown H), 4.54-4.58 (2H, M, crown H), 7.08 (IH, S, quinone-H); 6c (67.5 MHz; CDC13) 70.20 (CH2), 70.31 (CHZ), 70.60 (CHZ), 70.66 (CHZ), 70.71 (CH2), 70.91 (CHZ), 73.17 (CH;?), 73.30 (CHZ), B P O ? I 134.98 (=CBr), 135.58 (HC=C), 144.16 (=C-0), 145.15 (0-C=), 176.62 (C--O), 181.59 (C=O).Electrode Preparation The membrane components outlined in Table 1 were mixed and dissolved in THF overnight. The resulting homogeneous syrup was poured into a 25 mm diameter ground glass casting ring,23 and the solvent was allowed to evaporate off at room temperature, over a period of 48 h. An approximately 0.5 mm thick semi-transparent flexible membrane was obtained from which a working membranes of 7 mm diameter were cut using a cork borer. These disks were then pasted, using THF, to an interchangeable PVC tip which was clipped on to the end of the electrode body. This was in turn connected to a silver wire (1.63 mm diameter, Merck, Poole, Dorset, UK) contact which was previously chloridized by immersion in 40% m/m sodium hypochlorite for 30 min.Each electrode was stored in 0.1 moll-' CsCl solution when not in use. Measurement of Electrode Potentials All measurements were carried out in a thermostated poten- tiometric cell. The potential readings were measured using a Metrohm (Herisau, Switzerland) 654 millivolt/pH-meter rela- tive to a Metrohm 6.0702.100 SCE reference electrode. The electrochemical systems for the study were as follows: Ag/AgCl I 10-1 moll-' CsCl I PVC membrane 1 sample I KCl,,. I Hg2C12/Hg for alkali and alkaline earth metals, and Ag/AgCl I 10-1 mol 1-I CsCl I PVC membrane I sample I salt bridge I 1 mol 1-1 KN03 I KCl,,,, I Hg2Clflg for heavy metals. The salt bridge was prepared by dissolving 10 g of KN03 and 1.5 g of agar in 100 ml of doubly distilled water by heating.This was poured into a glass salt bridge to which two rubber tubes were attached, which on retraction allowed surface renewal on a daily basis. A dynamic method was used to construct calibration graphs of each analyte by performing regular injections at 1 min intervals. Absolute as opposed to relative potentials were used to construct each calibration graph. Dynamic responses were measured by injecting 90.9 pl of 1 moll-' CsCl into 10 ml of 10-3 moll-' CsCl thus producing a ten-fold increase in concentration while monitoring the potential change at a chart speed of 12 cm min-I. This plot was used to calculate the slope for a particular electrode. The LOD was taken at the point of intersection of the extrapolated linear segments of the caesium calibration curve.Repeatability was estimated by immersing the electrode alternatively into 1 0-2 and 10-3 mol 1-1 CsCl solutions at 25 "C. Potential readings were noted after 2 min, and the relative standard deviation (s,) was calculated. Selectivity coefficients were determined using the separate solution method24 at a concentration of lo-' moll-1 unless otherwise stated. The calibration plots were used in order to determine the change of KpotCsm as a function of concentration using the rearranged form of the Nicolsky equati0n:~5 I L 0 - J OH I1 Table 1 Membrane components used in ISE membrane construction 111 Fig. 2 Ionophores employed in this work: I, 2,3-benzoquino[ 151crown-5; 11, 2-bromo- 1 ,bdihydroxy[ 151crown-5; and 111, 5-bromo-2,3-benzo- quino[ 151crown-5.Component Amount (% m/m) Masslg Ionophore 0.66 0.0034 KTpClPB 0.17 0.0009 2-NPOE 65.84 0.3357 PVC 33.33 0. I700 THF - 3 mlAnalyst, February 1996, Vol. 121 129 r 1 2.303 RT zcs+ F where S = Electrode lifetimes were examined by monitoring the slope of Cs+ calibrations periodically. The temperature dependence of each electrode was investigated by noting the potential change in solutions of lO-3,10-2, and 10-1 moll-' CsCl taken over a temperature range of 5-55 "C. These values were rearranged to construct isotherms for each electrode at selected temperatures of 25 and 45 "C. The intercept of these isotherms represented the isopotential point of each respective electrode. Discussion Three crown ether derivatives containing quinone or dihy- droxybenzo groups were incorporated into PVC membrane electrodes and were shown to give acceptable linear response for E versus log (Cs+ activity) (Fig.3). Previous electrodes based on 15-crown-5-phosphotungstic acid precipitates give Nernstian responses also in a wide Cs+ activity range.' The alkali metal ion calibration plot of the electrode based on ionophore I, i.e., 2,3-benzoquino[ 15lcrown-5, is given in Fig. 4. The corresponding plots for this electrode for alkaline earths and selected transition metal ions indicate high Cs+ selectivity over divalent ions. There is however a slight curvature for Cs+ response at approximately - 1.5 of log concentration. All cations with the exception of Rb+ produced responses with slopes and LOD values worse than the Cs+ response. The cation responseorder for alkalimetalsisCs+ > Rb+ > K+ > NH4+ > H+ > Na+ > Li+.This series is similar to that for the 2-NPOE plasticized PVC electrodes incorporating the KTpClPB secon- dary ion exchanger, where selectivity is inversely proportional to the cation hydration enthalpy. The effect of functional group changes adjacent to the crown ether ring on the ionophores used in this work on cation response order would thus be important to study. The cation response order for alkaline earth ions was Ba2+ > Ca2+ > Sr2+ > Be2+ > Mg2+. The response for Ba2+ was 15.5 mV decade-', whereas responses for other divalent cations were considerably less. The transition metal ion responses were very poor with the response order being Pb2+ > Ni2+ =: Co2+ > Cu2+ but the slope for Ag+ (45.1 mV decade-') approaches that for Cs+ (51.9 mV decade-').The Ag+ calibration plot has a higher LOD (10-3.7 mol 1-1) than the Cs+ calibration (Table 2). The cation selectivity of a second electrode incorporating ionophore 11, i.e., 2-bromo-1,4-dihydroxybenzo[ 15lcrown-5, is significantly different. The cation response order is Cs+ > Rb+ = K+ > Na+ > H+ > N&+ > Li+. Of the diverse cations, Rb+ gave the best slope and LOD, closely followed by K+. Constant slope values (in the range 3.6-13.2 mV decade-') in the concentration range investigated were obtained for alkaline earth ions, with the response order Be2+ > Ba2+ > Ca2+ > Mg2+. Unlike the electrode based on I there was a linear response for Ca2+.Transition metal ions gave better responses for this electrode. A Nernstian response was obtained for Ag+ (56.4 mV decade-') in the concentration range 10-4.5-10-2 moll-1, followed by a sharp drop in potential at 10-1 moll-'. The order of responses for the other metals is Pb2+ > Cu2+ > The cation response order of the third electrode incorporating ionophore 111, i.e., 5-bromo-2,3-benzoquino[ 151crown-5, for alkali metals is Cs+ 2 K+ > Rb+ > N&+ > H+ > Na+ > Li+. The corresponding series for alkaline earth metals is Ba2+ > Be2+ > Ca2+ = Mg2+. All the alkaline earth metal ions with the exception of Ba2+ gave sub-Nernstian linear responses across the concentration range investigated. Transition metals fol- lowed the series Pb2+ > Cu2+ > Co2+ =.Ni2+, similar to electrode 11. Both Co2+ and Ni2+ produced no response (zero slope) across the calibration concentration range. Co2+ = Ni2+. I-CCS +H +Li +Na -K -Rb -NH4I 43 5 4 3 2 1 0 Log [CI Fig. 4 Calibration of the ionophore I electrode for alkali metal and ammonium ions. I I I I I I I -6 -5 -4 -3 -2 -1 0 Log [CS'] Fig. 3 ionpohores 1-111. Plot of responses for caesium ion-selective electrodes incorporating Table 2 Response slope, LOD, and s, (%) (of potential readings) properties of each ISE s, (%) at 25 "C (n = 10) Slopel LOD 10-2 10-3 Ionophore mV decade-' (-Log [Cs+]) moll-' mol I-' I 51.9 4.3 1.7 2.4 I1 48.6 4.4 11.4 4.7 I11 52.2 4.6 12.5 3.1130 Analyst, February 1996, Vol. 121 The slope of caesium response for each electrode is presented in Table 2 and the corresponding dynamic response trace is shown in Fig.5. Each electrode exhibits a near-Nernstian response for caesium. The response time for electrodes 1-111 is less than 3 s. Electrode response stability is in the order I = I11 > 11, with the slope of electrode I1 gradually decreasing over time to produce a stable sub-Nernstian slope of approximately 26 mV decade-' after 30 d. The LOD and repeatability for each electrode is also presented in Table 2. The LOD of the electrodes increase in the order I = I1 > 111. The high s, value for electrode I1 and I11 at high concentration was due to a positive potential response drift throughout the experiment. The selectivity coefficients (Log K P O t c s ~ ) determined by the separate solution method at 10-l moll-' are presented in Table 3 for a wide range of cations including alkali, alkaline earth and transition metal ions. From this tabulation, the major inter- ferents can be readly identified.For electrode I, Rb+, Ag+ if"d Hg2+ are the major interferents for the unsubstituted benzoquino crown ether. The incorporation of a bromo substituent at position 5 on the benzoquinone for electrode I11 increases the level of interference from Cu2+ significantly and Rb+ to a lesser extent. The Hg2+ and Ag+ interferences are lowered somewhat. Electrode I1 incorporating the bromo-substituted dihydroxy- benzo functional crown ether is significantly different from the i f Fig. 5 Dynamic responses of the three electrodes for a ten-fold increase in caesium concentration.Table 3 Interferent selectivity values (logKPotCs~) using the separate solution method Interferent I I1 I11 H' Li+ Na+ K+ Rb+ NH4+ Be2+ Mg2+ Ca2+ Sr2+ B a2+ Ag+ co2+ Ni2+ cu2+ Pb2+ Cd2+ Hg2+ -2.06 -3.00 -2.38 -0.99 -0.47 - 1.40 - 3.62 -4.03 -3.44 -3.10 -2.88 -2.59 -2.47 -2.42 -2.00 -2.1 1 +2.12 +0.94* -0.45 -1.44 -0.65 +0.04 -0.10 -1.79 -1.73 -2.37 -2.21 -1.64 +1.13t -1.83 -2.11 -1.68 -0.75 -0.40 - +2.93 * Determined at 10-2 mol 1-1. -1.10 -2.27 - 1.94 -0.89 -0.39 -0.99 -3.17 -2.77 -2.70 - -2.51 +0.47* -2.43 -2.56 -0.15 -3.05 -3.37 +1.83 other electrodes with increased responses to the interferent ions Na+, K+, Rb+, Ag+, Pb2+, Cd2+ and Hg2+. This is accompanied by an increase in the response to H+ ions and is attributed to the presence of phenolic groups capable of proton release and coordination to heavy metal ions.Further insight into the level of interference by diverse ions can be obtained graphically from plots of selectivity coefficients as a function of interferent concentration. This plot for electrode I11 is illustrated in Fig. 6 for monovalent interferents. Interferents that give calibration plots of similar slopes and LOD inflection points as the primary analyte ion produce corresponding selectivity plots which are parallel to the abscissa. This is the case for both monovalent and divalent interferences. The degree of selectivity as a function of concentration is reflected in the vertical position of the plotted lines across the calibration range to 10-l mol 1-l which encompases the Nernstian range of each electrode.This graphical representation also highlights the selectivity behavi- our obtained for Ag+ at 10- moll- which is not representative of its behaviour across the calibration range 10-4-10-2 moll-'. The result of temperature studies are outlined in Table 4. The isopotential point for electrodes I to I1 are reasonably constant and slightly above the working concentrations of the internal solutions employed throughout these studies. The remaining electrode had a significantly higher isopotential point which suggested the requirement of an almost two-fold concentration of the internal solution in order to optimize the temperature properties of the electrode. . + Conclusions Caesium ion-selective electrodes incorporating novel benzo crown ether derivatives based on the important biological redox -2.5 -1 5 4 3 2 1 Log concentration Fig. 6 Plot of selectivity coefficients for monovalent ions as a function of concentration for the ionophore I11 electrode.Table 4 Temperature studies on each ISE Temperature coefficient (6E/611 Isopotential 10-1 mol 10-2 rnol 10-3 rnol point*/ Ionophore 1 - 1 Cs+ 1-1 CS' I-' CS' moll-' I -0.166 -0.353 -0.428 0.129 I1 -0.167 -0.629 -1.215 0.117 -0.681 - 1.282 0.252 I11 -0.28 1 * Determined from isotherms at 25 and 45 "C.Analyst, February 1996, Vol. 121 131 carrier ubiquinone-0, have been investigated. Structural varia- tions within the benzo crown are reflected in the selectivity of the electrodes which was assessed for a wide range of cations, including heavy metals. The responses obtained for Ag+ ions are worthy of further investigation, particularly in conjunction with synthesis of newer benzocrown derivatives.The authors acknowledge D. Compagnone for his helpful comments. References 1 2 3 4 5 6 7 8 9 10 Kimura, K., and Shono, T. in Cation Binding by Macrocycles. Complexation of Cationic Species by Crown Ethers, ed. Inoue, Y. and Gokel, G. W., Marcel Dekker, New York, 1990, pp. 429-463. Blasius, E., Jansen, K. P., Klein, W., Klotz, H., Nguyen-Tien, T., Pfeiffer, R., Scholten, G., Simon, H., Stockemer, H., and Toussaint, A., J. Chromatogr., 1980, 201, 147. Das, S., Thomas, K. G., Thomas, K. J., George, M. V., Bedja, I., and Kamat, P. V., Anal. Proc., 1995, 32, 213. Partasarathy, N., and Byffle, J., Anal. Chim. Acta., 1991, 254, 1.Lehn, J. M., Science, 1985, 227, 849. Walkowiak, W., Charewicz, W. A., Kang, S. I., Yang, I., Pugia, M. J.. and Bartsch, R. A., Anal. Chem., 1990,62,2018. Ohki, A., Lu, J., Hallman, J. L., Huang, X., and Bartsch, R. A., Anal. Chem., 1995,67, 2405. Wang, D., and Shih, S. J., Analyst, 1985, 110, 635. Fung, K. W., and Wong, K. H., J. Electroanal. Chem., 1980, 111, 359. Attiyat, A. S., Ibrahim, Y. A., and Christian, G. D., Microchem. J., 1988, 37, 122. 1 1 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Srivastava, S. K., Gupta, V. K., Dwivedi, M. K., and Jain, S., Anal. Proc., 1995, 32, 21. Eugster, R., Gehrig, P. M., Morf, W. E., Spichiger, U. E., and Simon, W., Anal. Chem., 1991, 63, 2285. Eugster, R., Rosatzin, T., Rusterholz, B., Aebersold, B., Pedrazza U., Riiegg, D., Schmid, A., Spichiger, U. E., and Simon, W., Anal. Chim. Acta., 1994, 289, 1. Tsujimura, Y., Yokoyama, M., and Kimura, K., Anal. Chem., 1995, 67, 2401. Vogtle, F., and Jansen, B., Tetrahedron Lett., 1976, 4895. Dietl, F., Gierer, G., and Merz, A., Synthesis, 1985, 626. Merz, A., and Rauschel, M., Synthesis, 1993, 797. Hayakawa, K., Kido, K., and Kanematsu, K., J . Chem. SOC. Perkin Trans. I., 1988, 511. Peover, M. E., and Davis, J. D., J . Electroanal. Chem., 1963, 6, 46. Nagaoka, T., Okazaki, S., and Fujinaga, T., J. Electroanal. Chem., 1982, 133, 89. Hayakawa K., Kid0 K., and Kanematsu, J . Chem. Soc., Chem. Commun., 1986,268. Fischer A., and Henderson G. N., Chem. Commun., 1985,641. Moody, G. J., and Thomas, J. D. R., in Chemical Sensors, ed. Edmonds, T. E., Blackie, London, 1988. Guilbault, G. G., Durst, R. A., Frant, M. S., Freiser, H., Hansen, E. H., Light, T. S., Pungor, E., Rechnitz, G., Rice, N. M., Rohm, T. J., Simon, W., and Thomas, J. D. R., Pure Appl. Chem., 1976, 48, 127. Nicolsky, B. P., Z. Fiz. Khim., 1937, 10,495. Paper 5107253K Received November 3,1995 Accepted December 8, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100127

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (133-138) 133 Observations on the Behaviour of Some Trifluoroacetophenone Derivatives as Neutral Carriers for Carbonate Ion-selective Electrodes* Tomasz Sokalski, Dariusz Paradowski, Joanna Ostaszewska, Magdalena Maj -Zurawska,t Jozef Mieczkowski, Andrzej Lewenstam and Adam Hulanicki Department of Chemistry, University of Warsaw, 02 093 Warsaw, Poland Properties of three derivatives of trifluoacetophenone [4-(n-butyl)-l-trifluoroacetylbenzene (BTFAB), 4-(n-hexadecyl)-l-trifluoroacetylbenzene (HDTFAB), p-dodecyloxytrifluoroacetylbenzene (DDTFAB)] as neutral carriers for carbonate were examined. The best electrode characteristic (slope, S = 29 mV decade-', lowest detection limit) was for HDTFAB. The sensitivity towards carbonates was for DDTFAB pH dependent.This eliminates DDTFAB from practical applications. A theoretical treatment concerning carbonate and hydrogencarbonate selectivity under investigated conditions has been performed. Preliminary determination of carbonate in bovine serum was made using electrodes with HDTFAB. Keywords: Carbonate, trifluoroacetophenone derivative, ion-selective electrode, serum analysis Introduction Accurate and rapid determination of total carbon dioxide species (as C02, HC03- and C032-) is required in physiolog- ical, industrial and environmental samples. This can be attained by potentiometry with the use of carbonate-selective electrodes. Trifluoroacetophenone derivatives have been found to act as neutral carriers for carbonate ions.1-8 These compounds are Lewis acids with an electron deficit on the carbonyl carbon.The mechanisms of interaction with carbonate have been discussed: simple donor-acceptor interaction,g 1 : 1 and 2 : 1 adduct formation7-10 and the hydrogen-bonding that occurs after the ketone is hydrated to a geminal diol in the membrane.9.11-13 Selectivity of the compound is related to electrophilicity of substituents which is correlated with the Hammett constant ( o ) . ~ , ~ O Another parameter influencing selectivity is a presence of a liphophilic cation in a membrane, which hinders cationic ~electivity.~?~ The characterization of three trifluoroacetophe- none derivatives in a solvent-polymeric membrane containing bis(2-ethylhexy1)sebacate (DOS) or o-nitrophenyl octyl ether (0-NPOE) as plasticizer is reported.Experimental Reagents Doubly distilled water was used for preparing all solutions, additionally water for measurements of C032- and pH response was freshly boiled and bubbled with Ar purified from CO2 by bubbling through concentrated NaOH solution. Salts used were * Presented at The SAC '95 Meeting, Hull, UK, July 11-15, 1995. ' To whom correspondence should be addressed pure for analysis or pure (POCh, Gliwice, Poland). Tetra- hydrofuran (THF) was obtained from Merck (Darmstadt, Germany), PVC, DOS, o-NPOE were from Fluka (Buchs, Switzerland). Methyltridodecylammonium chloride (MTDDACl) was from Polysciences (Niles, IL, USA) and N - tris[hydroxymethyl]methyl-2-amino-ethanesulfonic acid (TES) was from Sigma (St. Louis, MO, USA). 4-(n-Butyl)-l-tri- fluoroacetylbenzene (BTFAB) was from Orion (Espoo, Fin- land), and Quality Serum Sample (LOT3800 I B) was obtained from KONE Diagnostics (Espoo, Finland).Each membrane contained 0.076 g of PVC as matrix, 0. I0 g of DOS or o-NPOE as plasticizer, neutral ionophore (5.0 X 10-5 mol BTFAB, 1.6 X rnol DDTFAB) and MTDDACl (0-150 mol% versus the ionophore). After dissolution in 3 ml of THF, the mixture was poured into a 28 mm diameter glass ring placed on a glass plate and the solvent was allowed to evaporate. After 48 h round pieces of the membrane were cut out for the preparation of electrodes as described earlie1-1~ or put in the Phillips IS 561 electrode body. mol HDTFAB or 1.6 X Procedure and Equipment All measurements were made at 25 L- 1 "C in cells set as follows: Ag/AgCl,KCl,,,,, I 1 mol dm-3 KC1 I sample solution I membrane I internal solution; AgCl/Ag Two kinds of internal filling solutions were used: ( i ) 0.138 mol dm-3 NaCl, 0.01 mol dm-3 TES buffer (pH 7.4), 0.1 mol dm-3 NaHCO3; and (ii) 0.1 rnol dm-3 NaH2P04, 0.1 mol dm-3 Na2HP04, 0.01 mol dm-3 NaCl.In both cases electrodes were stored in the conditioning solution: 0 1 rnol dm-3 NaHC03. The behaviour of the electrodes was independent of the internal solution used. The external reference electrode was a double-junction Ingold 98 1 1 elec- trode. For the monitoring of pH, glass electrodes Ingold I047 1 1 or Hydromet ERH 11 (Warsaw, Poland) were used. Single-ion activities were calculated according to Debye- Huckel theory. l5 For carbonates and phosphates the concen- tration of each species was calculated from protonation equilibria at measured pH.The emf measurements were performed on a Solartron (Farnborough, Hampshire, UK) LM 1604 DC digital voltmeter with multi-electrode sampler. Each reading was carried out 20 min after change of sample solution, when no further significant emf drift (0.2 mV min-1) was observed. The measurements were made in the unstirred solutions. For determination of the selectivity coefficients by the separate-solution method, 0.1 mol dm-3 solutions of appro-134 Analyst, February 1996, Vol. 121 priate sodium salts were used except when carbonate was the main ion, in which instance 0.1 mol dm-3 NaHC03 was also used.8 The Henderson equationls provided corrections for the liquid-junction potential.Where data was not available for salicylate, corresponding parameters of benzoate were used. Potential response characteristics of electrodes was deter- mined by stepwise addition of weighed amounts of NaHC03 (to obtain the lowest concentrations, dilution was used instead) to TES (0.01 mol dm-3). The pH was kept constant by the addition of dilute sulfuric acid (0.1 and 0.01 moll-1). The effect of pH change was investigated by the addition of dilute sulfuric acid to 0.025 mol dm-3 NaHC03. Analytical recovery investigation of carbonate concentration in aqueous solutions and in spiked bovine serum (KONE Diagnostics Control Serum LOT38001 B) was performed by a Microlyte 6 (KONE, Finland) clinical analyser in which the carbonate electrode was inserted. Synthesis All melting and boiling points are uncorrected.The IR spectra were obtained on a Nicolet Magna (Madison, USA) 550 ST spectrophotometer (using KBr pellets or film). The 1H NMR spectra were recorded on a JEOL-4H- 100 spectrometer (Tokyo, Japan) and 13C NMR on a Bruker 90 spectrometer (Karlsruhe, Germany) in deuteriated solvent CDC13 using TMS as internal standard. Phenylpalmitoy lketone To 17 g (0.121 mol) of anhydrous aluminium trichloride suspended in 100 ml of benzene, 30 g (0.109 mol) of palmitoyl chloride was added. The reaction mixture was stirred and heated for 2 h at 60"C, then poured on ice and acidified with 20% aqueous hydrochloric acid. The benzene layer was dried with anhydrous magnesium sulfate and then the benzene was evaporated. The crude product was distilled (bp, 200"C, 10 Torr).Yield 25.1 g (72.7%) of pale yellow oil which solidified after standing. The analytical sample following recrystallization from pentane had mp, 60-61 "C. Elemental analysis found: C, 83.31; H, 11.51%. Calc. for C22H360: C, 83.54; H, 11.39%. IR(KBr): 2930, 2860, 1690, 1600, 1470, 1380, 740 cm-l. IH NMR (CDC13, 6): 8.15-7.87 (m, 2 H) 7.65-7.37 (m, 2 H) 2.97 (t, 3 H) 1.91-1.61 (m, 10 H), 0.91 (t, 3 H). \ Hexadecy lbenzene To 100 g of activated zinc powder (Zn/Hg couple) suspended in 600 ml of 15% aqueous hydrochloric acid, 20 g (0.0632 mol) of phenylpalmitoylketone was added and stirred for 2 h at 30 "C, then 500 ml of concentrated hydrochloric acid was added and heated under reflux for 10 h. The organic layer, extracted with benzene, was dried with anhydrous calcium chloride, then evaporated, and the crude product was distilled (bp, 230 "C, 15 Torr) Yield 15 g (78.5%) mp, 20 "C.Elemental analysis found: C, 87.21; H, 12.72. Calc. for C22H38: C, 87.41; H = 12.58%. IR(KBr): 2930, 2870, 1380, 1130 cm-1. 1H NMR (CDC13, 6): 7.53-7.08 (m, 5 H), 2.60 (t, 2 H), 1.78-1.01 (m, 14 H), 1.91-1.61 (m, 10 H) 0.91 (t, 3 H). 4-(n-Hexadecyl)- 1 -trifluor-oacetylbenzene To 8 g (0.0599 mol) of anhydrous aluminium trichloride, suspended in 100 ml of methylene dichloride 6.3 g (0.03 mol) of trifluoroacetic anhydride at -20 "C was added and stirred for 1 h at the same temperature. Hexadecylbenzene (14 g in 20 ml of methylene dichloride) was added and stirred for 1 h at room temperature. The reaction mixture was poured on ice, acidified with 20% aqueous hydrochloric acid.The organic layer was dried with anhydrous sodium sulfate, evaporated and the solid remainder chromatographed on silica gel (Merck, 100-200 mesh) eluted with hexane. After chromatography 6 g of unreacted hexadecylbenzene was recovered and 3.3 g (3 1.3%) of ionophore was obtained, which solidified on standing in a refrigerator. The analytical sample was recrystallized from pentane (mp, 34-36 "C). Elemental analysis found: C, 72.10; H, 9.48%. Calc. for C24H370F3: % C, 72.36, H = 9.29%. IR(KBr): 2940, 2870, 1725,1625,1480,1205,1150,950 cm-l. lH NMR (CDC13,S): 8.05 d, 2 H), 7.38 (d, 2 H), 2.68 (t, 2 H), 1.91-0.76 (m, 14 H). 13C NMR (CDC13, 6): 180.09 ppm (q,Z34.2 Hz) CF3-C=O. p-Chloro- 1 -dodecyloxybenzene To 35 g (0.272 mol) of p-chlorophenol dissolved in 300 ml of dry dimethylformamide 61 g (0.298 mol) of dodecyl chloride and 70 g (0.5 mol) of anhydrous potassium carbonate was added.The reaction mixture was heated for 4 h under reflux, then poured into water and extracted with benzene. On evaporation of the solvents, the crude product was distilled (bp, 220 "C, 10 Torr) and crystallized from methanol. Yield 61.5 g (76.4%), mp, 32-35 "C. Elemental analysis found: C, 72.51; H 10.02%. Calc. for Ci8H2&10: C, 72.84; H, 9.78%. IR(KBr): 2940, 2870, 1600, 1500, 1200 cm-1. p-Dodecy loxytrifluoroacety lbenzene To 15 g (0.0505 mol) of p-chloro- 1 -dodecyloxybenzene in absolute THF (100 ml) cooled to -78 "C, 3 1.6 ml(O.0505 mol) of butyllithium in hexane was added and stirred at this temperature for 0.5 h.The reaction mixture was siphoned to 5 ml of methyl trifluoroacetate in SO ml of THF. After 0.5 h it was poured into aqueous ammonium chloride and extracted with benzene. After evaporation of solvents, the crude reaction mixture was chromatographed on silica gel (Merck, 200-300 mesh) eluted with hexane-benzene mixture 9 : 1. Yield 2.6 g (30%) of ionophore and recovered 8 g of unreacted p-chloro- 1 -dodecyloxybenzene. The product was crystallized from pentane (mp, 30-33 "C). Elemental analysis found: C, 66.05; H, 8.21%. Calc. for C20H2902F3: C, 67.03; H, 8.10%. IR(KBr): 2950, 2870, 1730, 1600, 1500, 1290, 1200, 980 cm-1. 1H NMR (CDC13, 6): 7.77-7.41 (m, 2 H), 7.05-7.88 (m, 2 H), 4.05 (t, 2 H), 2.0-1.05 (m, 7 H), 0.9 (t, 3 H). 13C NMR (CDC13, 6): 182.61 ppm (4, I , 37.8 Hz) CF,-C=O.Results and Discussion Selectivity and Characteristics of Investigated Trifuoracetophenone Derivatives The selectivity of investigated membranes is shown in Figs. 1 and 2. It is dependent on molar ratio of lipophilic ammonium salt to the ionophore; it is expected that for the HDTFAB the behaviour is similar to that for BTFAB (Fig. 1). The addition of this salt is necessary in order to avoid cationic selectivity of the membrane. The higher the concentration of the salt, the more similar is the selectivity to the selectivity of the membrane with MTDDACl only (last column in Fig. 1). The only exception is the membrane containing DDTFAB in DOS, where non- monotonic changes are observed. The expected selectivity improvement with the change of the plasticizer from DOS (permittivity EO = 4.2) to o-NPOE (EO = 20) was not observed.Small changes in membrane composition with DDTFAB have caused bad reproducibility of membrane behaviour. The difference in selectivity coefficients at 20 and 40 mol% of MTDDACl for all ions (except Cl) is small. Because of possibleAnalyst, February 1996, Vol. 121 135 loss of lipophylic ammonium salt due to elution from the membrane, a higher concentration of MTDDACl is more favourable to hinder the influence of cations. For further investigations the membranes with DOS and 40 mol% of MDDACl were selected. In Fig. 2 the selectivity of three investigated ionophores with DOS and 40 mol% of MDDACl is shown. The Hammett substituent constant (0) is the essential parameters to predict the selectivity of used trifluoroacetophene derivatives.This constant is -0.17 for BTFAB and HDTFAB and -0.27 for DDTFAB.16 The selectivities of BTFAB and HDTFAB are similar but differ from that of DDTFAB. Another important parameter to characterize the ionophore is its lipophilicity. This can be estimated as log P according Hansch and Le0.17 Calculated values of log P were 5.6 for BTFAB, 1 1.6 for HDTFAB and 8.4 for DDTFAB. The detection limit of electrodes should be lower with higher lipophilicity . Our observations are in a good agreement with this (Fig. 3(a)]. Electrode characteristics at different pH (7.00 & 0.05, 7.30 k 0.05, 7.80 k 0.05) are shown in Fig. 3. The first potential value was measured in the basic TES solution after removal of CO:! 8 6 4 2 0 -2 -4 -6 I I , I , , , I I I I - , I I I I I \ \ !......... \ ... 1 -2 I I I , I I I I I I I I I , , t I , , I I I I I I I I I I I I I I I I I I I I I I I I I I I I I .... .... .... ........... / ................ I I I I I I I I 0 20 40 60 80 100 150 no DDTFAB mol% MTDDACI Fig. 1 o-NPOE on MTDDACl mol%. Dependence of selectivity of membranes containing (a) BTFAB and DOS, (h) BTFAB and o-NPOE; (c) DDTFAB and DOS; and (4 DDTFAB and136 Analyst, February 1996, Vol. 121 0.14 - 0.12 - h $ 0.10 - v ; 0.08 - 0.06 - 0.04 - 0.02 traces by Ar having appropriate pH and taken as a blank. The potentials of all three electrodes were measured at the same time (i.e., in the same solution). The DDTFAB characteristic is pH dependent: the higher pH the worse response to carbonate. At constant pH, the concentration of hydrogen carbonate ions is directly proportional to carbonate ions concentration, i.e.relative changes in concentrations of both ions are the same. Fig. 3 shows clearly the influence of OH- ions because very small emf changes were observed at pH 7.8. This eliminates DDTFAB from ionophores suitable to carbonate determina- tion. 0 3 6 9 1 2 I I I I Theoretical Treatment of Carbonate-selective Electrode Behaviour So far it was either assumed that carbonate electrodes are selective to carbonate ions2 or said in general terms that they are selective to ‘hydrogencarbonate specie^'.^ It seems that this question should be examined more thoroughly. If we assume that carbonate is the main ion we obtain the following equation: I RT 2F E = Eo- 2.303 - log { [C032-] + kyt [HC03-I2} (1) I Alternatively if we assume hydrogencarbonate to be the main ion we get: RT F E = E o - 2.303 - log { [HC03-] + k r t [C032- The unusual and specific quality of this system is that we cannot talk about ‘main’ and ‘interfering’ ion in the usual sense.Normally concentrations of those ions are fixed and do not change in the solution. In the case of the carbonate-hydrogen- carbonate system the ratio of those two ions changes with the change of the pH. This means the ‘main ion’ can be converted into the ‘interfering ion’ and vice versa. That makes the situation even more complicated as the ion charge changes. Before further discussion let us consider how to calculate concentrations of hydrogencarbonate species.The exact method involves iterative calculations and is rather time-consuming. Fortunately enough, the relative difference between results I I - Salk obtained by ‘exact’ and ‘simplified’ way, especially in the physiological range (Fig. 4), is so small that it can be neglected and we can use simplified calculation method. (i) Assuming that the electrode answers only to the carbonate ions, and Ap(C03*-) = -Alog[C032-] = 1. Then AE = -29.6 mV. Because of the relation between carbonate and 70 t 60 1 100 (4 I 70 I- :: 70 I I 60 -6 -5 -4 -3 Log aco32- Fig. 3 Electrode characteristics at different pH (a) 7.00 k 0.05; (b), 7.30 & 0.05; and (c) 7.80 & 0.05. Membranes contained DOS, 40 mol% MTDDACl and HDTFAB (m), BTFAB (a), and DDTFAB ( A ).l- - Saiic. 1 0.16 I 2 i - AcO- - - so:- -6 c H,PO,- \ H,PO,- I H,PO,-- BTFAB HDTFAB DDTFAB Fig. 2 Selectivity of membranes containing BTFAB, HDTFAB, DDTFAB and 40 mol% MTDDACl in DOS.Analyst, February 1996, Vol. 121 137 hydrogencarbonate concentrations [HC03-] = [H+] [C03*-]/ K2, at constant pH carbonate and hydrogencarbonate concentra- tion are directly proportional. If we calculate the slope of the calibration curve, we should obtain following values, depending whether we use eqn. (1) or (2): Log{ [C032-] + k r t [HC03-I2} (1) AE ( a ) krt << 1 then S = = -29.6 mV AP(C03) AE (b) kyt << then S = = -14.8 mV 2AP(CO3) AE ( a ) k r t << then S = = -29.6 mV APtCO3) 2AE (b) kfl"' >> 1 then S = = -59.2 mV AP(C03) (ii) If the electrode is selective only to the hydrogencarbonate ions, and Ap(HC03-) = -Alog[HC03-] = 1, then of course AE = -59.2 mV.Again we can expect following values: Log{ [C032-] + k;.Ot [HC03-12} (1) AE (a) k y t << then S = = -59.2 mV APtHC03) 2APtHC03) AE (b) kFt << then S = = -29.6 mV Log{ [HC03-] + k;.Ot [C032-]l/2} AE (a) k r t then S = = -59.2 mV AP(HC03) 2AE (b) kyt >> 1 then S = = -118.4mV AP(HC03) Table 1 presents slopes calculated according to eqns. (1) and (2) from the experimental data for the typical calibration curve of HDTFAB. Comparison of obtained results to the limiting values Table 1 Experimental dependence of carbonate electrode (HDTFAB) slope on the assumed selectivity coefficients Slope (SlmV decade-') Assumed kij Main ion: C03*- Main ion: HC03- 0.001 -29.9 0.0 1 -27.8 0.1 -22.0 1 - 17.3 10 -15.0 100 -14.3 1000 - 14.1 -28.3 -28.4 -29.5 -35.8 -50.5 -59.0 -60.4 given above shows clearly that in those conditions the electrode is selective almost exclusively towards carbonates.Further confirmation of this statement can be obtained from the dependence of the electrode potentials on pH changes at constant total carbonate concentration of 25 mmol dm-3 (Fig. 5(a)]. If one recalculates hydrogencarbonate concentration one obtains apparently strangely shaped curves [Fig. 5(c)]. How- ever, these curves prove, that electrodes are certainly not sensitive only to hydrogencarbonate concentration. If electrodes had been sensitive only to hydrogencarbonate, potentials for the same hydrogencarbonate concentration (before and after max- imum) would have been the same.Additionally, Fig. 5 illustrates the sensitivity of the DDTFAB electrode towards OH- ions (increase of electrode slope with increasing pH). 160 I 140 - 120 - 100 - 80 - 60- 40 - 20- 4 5 6 7 8 9 1 0 1 1 0 0 -10 -9 -8 -7 6 -5 4 -3 -2 -1 160 140 - 120 - loo - 8 0 - 60 - 40 - 20 - 'I- (') -3.5 -3.0 -2.5 -2.0 -1.5 aHC03- Fig. 5 Dependence of electrode potential on pH changes (a) and corresponding changes of carbonate (h) and hydrogen carbonate (c) concentration (total carbonate concentration was 25 mmol dm-3). Mem- branes contained DOS, 40 mol% MTDDACl and HDTFAB (M), BTFAB (a), and DDTFAB ( A ).138 Analyst, February 1996, Vol. 121 Preliminary Application of HDTFAB Electrode as Carbonate Sensor in Blood Serum To estimate the possibility of application of the carbonate electrodes for biological samples, pH, C1- and C032- concen- tration were determined using KONE Microlyte 6 clinical analyser.Carbonate concentrations were determined for the series of aqueous solutions, having (i) ionic composition close to human serum and (ii) bovine serum (Quality Serum Sample, KONE) spiked with NaHC03. Calibration curves were linear having regression equations Y = 1.OOX - 0.39, r = 0.97 ( n = 9) and Y = 1.05X - 0.65, r = 0.96 (n = 6) for (i) and (ii), respectively. The correlation of added and determined HC03- both in aqueous solutions and in serum is satisfactory. However, one must stress the importance of exact knowledge of k r t , Eo, S , and pH to obtain proper results.Conclusions Properties of three derivatives of trifluoacetophenone BTFAB, HDTFAB and DDTFAB) as neutral carriers for carbonate were examined. Under the investigated conditions the studied electrodes show exclusive sensitivity towards carbonate but not hydrogencarbonate. The best analytical parameters are dis- played by electrodes with HDTFAB carrier in the membrane. In the literature the influence of plasticizer polarity on selectivity properties is sometimes stressed.l*J9 However, in our case the polarity of used the plasticizers (DOS, o-NPOE) was of no practical consequence. In final experiments DOS was used due to its better behaviour in contact with biological samples. Preliminary determination of carbonates in bovine serum was made using electrode with HDTFAB in potentiometric KONE Microlyte 6 clinical analyser.These results indicates that this sensor can be used for rapid and simple potentiometric determination of HC03- in blood. This study was partially supported by the project BW- 12 19/2/94 and KONE Instruments. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Wise, W. M., US Pat., 3723281, 1973. Herman, H. B., and Rechnitz, G. A., Anal. Chim. Acta, 1975, 76, 155. Herman, H. B., and Rechnitz, G. A., Anal. Lett., 1975, 8, 147. Greenberg, J. A., and Meyerhoff, M. E., Anal. Chim. Acta, 1982,141, 57. Smirnova, A. L., Grekovich, A. L., and Materova, E. A., Electro- khimiya, 1985, 21, 1221. Smirnova, A. L., Grekovich, A. L., and Materova, E. A., Electro- khimiya, 1985, 21, 1335. Meyerhoff, M. E., Pretch, E., Welti, D. H., and Simon, W., Anal. Chem., 1987,59, 144. Behringer, Ch., Lehmann, B., Hang, J. P., Seiler, K., Morf, W. E., Hartman, K., and Simon, W., Anal. Chim. Acta, 1990, 233, 41. Ampilogova, N. A., Karavan, V. S., and Beloshapko, M. I., Zh. Anal. Khim., 1985,40, 895. Bart, T. Ya., Karavan, V. S., Grekovich, A. L., Ampilogova, N. A., Yurinskaya, V. E. and Nikiforov, V. A., Zh. Anal. Khim., 1990, 45, 1364. Scott, W. J., Chapoteau, E., and Kumar, A., Clin. Chem. (Winston- Salem, N.C.), 1986, 32, 137. Wang, K., Seiler, K., Hang, J. P., Lehmann, B., West, S., Hartman, K., and Simon, W., Anal. Chem., 1991, 63, 970. Seiler, K., Wang, K., Kuratli, M., and Simon, W., Anal. Chim. Acta, 1991,244, 151. Grass, A., Moody, G. J., and Thomas, J. D. R., J. Chem. Educ., 1974, 51 (8), 541. Meier, P. C., Amman, D., Morf, W. E., and Simon, W., in Medical and Biological Applications of Electrochemical Devices, ed. Koryta, J., Wiley, New York, 1980, ch. 2. McDaniel, D. H., and Brown, H. C., J. Org. Chem., 1958, 23, 420. Leo, A., Hansch, C., and Elkins, D., Chem. Rev., 1971, 71, 525. Pretsch, E., Badertscher, M., Welti, M., Maruizumi, T., Morf, W. E., and Simon, W., Pure Appl. Chem., 1988,60,567. Sokalski, T., Maj-Zurawska, M., and Hulanicki, A., Mikrochim. Acta, 1991, 19911, 285. Paper 5103526K Received June 2,1995 Accepted August 16, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100133

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (139-161) 139 Recent Developments in the Determination of Precious Metals A Review Yi Bin Qu Institute of Precious Metals, China National Non-ferrous Metals Industry Cooperation, Kunming, China Summary of Contents Introduction Sampling, Sample Preparation and Standards Preconcentration and Separation Fire Assay Solvent Extraction Adsorption and Ion Exchange Chromatography, Flotation and Other Techniques Gravimetric Methods Volumetric Methods Gravimetric and Volumetric Methods Atomic Absorption Spectrometry Atomic Emission Spectrometry UV-VIS Absorption Spectrometry Luminescence Analysis Chemiluminescence Fluorescence Catalytic Kinetic Methods Detection by Spectrophotometry Detection by Electrochemical Methods Flow Injection Chromatography Electrochemical Methods Mass Spectrometry Radiochemical Analysis X-ray Spectrometry Other Methods References Keywords: Precious metal; determination; preconcentration; separation; sampling Introduction The determination of precious metals attracted the interest of analysts and developed rapidly because these metals are valuable and rare, yet also very important for many industrial processes and products.Their concentration levels are very low in many natural resources, metallurgical intermediates and environmental samples. This review deals with the develop- ments in the determination of precious metals in recent years with emphasis on applications. Precious metals here refers to the six platinum group metals (PGE), gold and silver. Most of the cited references were obtained from a survey of Chemical Abstracts, starting with Vol.116, No. 1, and extending to Vol. 120, No. 26. Some important topics are traced back to earlier literature. Some of the books and reviews that have been published are the following: a book on selected instrumental methods for the determination of precious metals' and several reviews that deal with the development of precious metal assay techniques in Europe and Hungary,2 the development of analytical methods for precious metals in China (the third biennial re vie^),^ prospects for the determination of platinum group metals and gold in geological materials: analytical methods for determin- ing gold in geological samples,5 a comparison of analytical methods for gold,6 the determination of gold in gold alloys or bullion by Standards Au~tralia,~ and analyses for microelements in silver and silver-based alloys.8 Others will be cited in the relevant sections.Sampling, Sample Preparation and Standards Because the rocks and ores that contain precious metals are usually heterogeneous, sampling and preparing homogeneous samples, especially those used as reference materials, are important and constant subjects for the analyst. The methods of determining sample mass for the chemical analysis of ores were reviewed and compared using a gold ore as an example.9 A simple process was reported for preparing representative samples from complex gold ores by crushing, screening and finely grinding.lO However, a knowledge of the state and particle size of the analyte will help to select an adequate method for sampling or sample preparation.It was found that if the size of metallic Au particles in gold ores is < 0.074 mm (200 mesh), a sample ground to a particle size of 80 mesh will be adequate." For ores containing natural silver, accurate results for Ag were only obtained when the samples were ground to 200 mesh.12 New techniques for sample preparation in fire assay were presented based on the experi- mental results that Au and Ag can be extracted from materials crushed to a grain size of 1 mm, when sufficiently large ratios of the flux mass to the mass of the analytical specimen were used.13 Four Canadian reference ores of gold were analysed by instrumental neutron activation analysis (NAA) to determine their sampling constant.l 4 When gold-bearing till samples were stored in plastic containers, it was found that the superficial layer of all samples was reduced in Au whereas the basal layers and walls of the containers were enriched, the sorting of Au grains and Au-bearing minerals being enhanced by both gravity separation and electrostatic adhesive forces between them and the walls of the plastic containers.15 A review was published on decomposition techniques for geochemical samples, including those containing precious metals. l6 The analytical results for 26 geochemical reference samples decomposed by aqua regia and HF were higher than those decomposed by aqua regia only, preconcentration by extraction with isobutyl methyl ketone (IBMK) and determi- nation by electrothermal furnace atomic absorption spec- trometry (ETAAS) being used in both cases.17 By comparison of analytical results for Au in vegetation with and without high- temperature (875 "C) ashing, no evidence was found of the loss of Au in the former instance.18 A study was carried out to develop an alternative method for analysing 'classically unassayable' Au in gold ores and this method has been patented.'9140 Analyst, February 1996, Vol. 121 Two chromitite samples from the Shetland Islands (Scotland) were processed as reference samples for the determination of PGE and Au, a detailed assessment of the sample homogeneity being presented together with the results of a cooperative study by 35 international geochemical laboratories.20 After studying the precision of the analytical results for Au, Ir, Pd and Pt in rocks, it was concluded that reference materials for the analysis of sulfur-poor mafic and ultramafic rocks with background levels of noble metals will be more satisfactory if volcanic rocks are selected.21 The analytical data for Ag, Au, Pd and Pt in 26 Geological Survey of Japan (GSJ) geochemical reference samples (1991)Z2 and the results for Au concentrations in five reference samples23 were presented and discussed.The manufacturing process and the methodology used to assess the homogeneity of a set of reference fine gold materials and to determine the concentrations of various trace metals in these materials were described.24 The preparation and certifi- cation of a rhodium standard reference material solution were rep0rted.~5 Preconcentration and Separation Fire Assay Fire assay is a unique technique for preconcentrating precious metals. It is especially useful where large samples are required owing to their extremely low concentrations in some materials and the heterogeneity of ores containing these metals.Lead, copper, antimony, bismuth, tin, copper-iron-nickel alloy and nickel sulfide have been used as collectors. An excellent review on fire assay for the preconcentration of PGE and Au has been published.26 A report has been presented discussing the fire assay technique with respect to its potential and limitations.27 Lead fire assay is an effective and widely used technique for the preconcentration of Au, Ag, Pt and Pd when the concentra- tions of the last two elements are low.This technique has long been known as an analytical method for Au and Ag, yet there were reports discussing the conditions and skills in this technique,28 especially at the cupellation stage.29-32 An equa- tion for calculating the loss of Ag in lead fire assay was also proposed and it can be used for the correction of the results obtained.33 For improvement of the operation, a domestic microwave oven was used instead of the ancient furnace for extracting Au and Ag by fire assay.34 Nickel sulfide fire assay is very effective for the collection of PGE and Au and is now widely used. The ratio of flux, nickel, sulfur and sample in this method generally follows that in the procedure established by Williamson and Savage in 1965 (see ref. 26). Some studies were made to use a minimized NiS button in recent years.35.36 By reducing the button size to below 1 g, the precious metals were successfully determined directly in the button with NAA37 and ICP-MS using laser ablation for sample introduction.38 A recent report also focused on the pre- concentration of precious metals in geological materials by NiS fire assay for their determination by NAA.39 For understanding the effects of fusion charge composition on the collection of PGE into a minimized NiS button, the concept of the extraction ratio E was proposed, which is expressed as E = (msample + mflux)/(mNi + ms), where m is the mass of the component indicated by the subscript.40 According to this expression, E is about 5 for general procedures.After a detailed study of the relationship between E and the recovery of PGE, it was concluded that E 3 30 is feasible only when the effect of the sample composition, particular its sulfur content, is fully considered and carefully controlled.Because a minimized NiS button offers obvious advantages for the subsequent stages of analytical procedures, more detailed study may lead to its wider application. Antimony fire assay was devised early to preconcentrate PGE and Au from samples with simpler compositions. The separa- tion of precious metals, including Os, from antimony can be achieved by cupellation. An antimony fire assay method for silver has been reported.4' A patent claims a new fire assay procedure for precious metals in ores and concentrates containing iron oxide, using iron as a collector.42 The iron is embrittled by adding Te or a Te compound during the pyrometallurgical process and is easy to separate by electrolysis or selective dissolution.Solvent Extraction Solvent extraction is one of the important techniques for the separation and preconcentration of precious metals. Many common extractants such as IBMK, tributyl phosphate (TBP) and trialkylphosphine oxide (TAPO) are still widely used for the separation and preconcentration of precious metals. Two new extractants, dibutyl selenide and dibutyl telluride, were proposed and compared with dibutyl sulfide in the extraction separation of Au and Ag.43 It was found that dibutyl sulfide and selenide are preferred for Au"', and dibutyl selenide and telluride for Ag'. Silver was quantitatively separated from large amounts of common metals and chloride by extraction with 1,2-bis[alkyl(aryl)thio]ethane.44 Bis(2-ethylhexyl) sulf- oxide (BESO) exhibits a strong extraction ability towards Pd, and trace and macro amounts of Pd can be quantitatively extracted from about 8 moll-' HN03 solutions of pH down to 2 by BESO in toluene.45 A mixture containing an organic sulfide and a diluent with the addition of tert-alkylamine or alkylaniline was selective for the extraction of Pd.46 The extraction of PtTV with bis(dodecylsulfiny1)ethane was studied and applied to the separation of Pt from noble metal and common metal ions.47 The separation of Pd from Au, Cu, Ag and Hg was achieved by the extraction of its dibutyldithiophosphate into nonane.48 The extraction of Ptkv with rubeanic acid in TBP, with TBP and with thenoyltrifluoroacetone in butanol-acetophenone was investigated and optimized conditions were found.49 The extraction of Pd by triphenylphosphine sulfide in benzene has been investigated.50 Au was quantitatively extracted by tris(2- ethylhexyl) phosphate and thus separated from Pd, Pt, Pb and cu.51 Silver was extracted from iodide solution by diantipyryl- propylmethane in CHC13 and thus separated from Sb1n.52 The quantitative extraction of Ag' as the thiosulfato complex was achieved by the use of a high molecular mass secondary amine, Amberlite LA-2, as extractant.53 Ru in acidic solution was extracted with a high yield and rate by using acetone as a solvent and a&-dipyridyl or 1,lO-phenanthroline as a reagent after neutralizing the initial solution with NaOH and cooling to Effective solid extractants have been developed recently.The extraction of Pt and Au from chloride solution and leaching pulp was achieved by using a solid extractant based on quaternary ammonium halide, trioctylphosphine oxide (TOPO) and TBP.55 Au was recovered from HCl solution by extraction with a solid extractant containing paraffin as matrix and derivatives of 18-crown-6 as active c0mponents.5~ 55-63 OC.54 Adsorption and Ion Exchange Studies on the separation and preconcentration of precious metals by adsorption have been very extensive. Several traditional sorbents, such as activated carbon, sulfhydryl cotton and polyurethane foam, are still widely used and efforts are being made to establish improved procedures. Many new sorbents have been synthesized, some of which are specific for certain precious metals.T1"' adsorbed simultaneously with Au"' on activated carbon was extracted with dilute nitric acid and thus separated fromAnalyst, February 1996, Vol. 121 141 ~ ~ Au.57 The preconcentration of Au with an activated carbon column was greatly improved after removing the colloidal matter in ore sample solutions by the addition of polyethylene oxide and filtration.58 To speed up the process of adsorbing AuC14- by polyurethane foam, a high-speed dynamic proce- dure59 and a static procedure under reduced pressure60 were developed. Rapid desorption of Au from the foam was achieved by elution with a thiourea solution containing Fe11160 and the thiourea in the eluate was decomposed rapidly by heating on a water bath in the presence of H202 and KC1.6' Gold was also eluted from the foam by acetone and determined directly in the eluate by a spectrophotometric method.59~~~ The preconcentra- tion of Au and Ag with polyurethane diphenylthiourea- loaded foam64 and IBMK-loaded foam65 was proposed. Since the adsorption ability of polyurethane foam varied greatly with its origin, caution must be exercised when it is used.Silver was selectively adsorbed by nitrocotton as a chelate of 1,lO-phenan- throline, eluted with dilute HN03 and determined directly in the eluate by AAS.66 The main characteristics of the type of sorbent with the trade name of Polyorgs were reviewed, in which those containing pyrazole, imidazole, 2-mercaptobenzothiazole and thioglycol- anilide groups exhibit high selectivity for precious metals and have been applied to the preconcentration of these metals in rocks, ores, industrial products and natural waters.67 A macroporous poly(viny1thiopropionamide) chelating resin was synthesized and applied to the preconcentration of AulI1, PtIV, Pd" and Ir'v.68 Investigations by Fourier transform IR spectrometry and electron spectroscopy revealed that Au, Pt and Pd ions were chelated mainly with the thioketo form of the thiopropionamide group in the resin, forming a quadridentate chelate.The other resins recommended are the following: a newly synthesized macroporous poly(viny1aminoacetone) che- lating resin for the preconcentration of Au"', Pd", Rh"' and Ru1";69 a chelating resin containing a quinaldinic acid amide group for the separation of PtIV and Pd" from each other and from noble-base mixtures, and also from native Ag s01utions;~O new chelating sorbents of the pyrazolone type for the separation of Au, Pd and Ag from base metals;7' and a rhodanine chelating resin72 and a rhodanine fibre73 for the preconcentration of Au, Pd and Ag.When used for preconcentration of Au, the rhodanine fibre is better than polyurethane foam owing to the constancy of its adsorption ability and its stability. It can be used repeatedly many times. The separation of Au"', PtIV and Pd" from base metals and from each other was achieved by using a phosphorus-containing chelating resin with tetraethylenepenta- mine? Rhuthenium was separated from Ir by thiopropionyl- amide resin.75 Gold was preconcentrated by sorption of its 2-aminobenzolthiazole complex with Amberlite XAD-4 resin.76 The enrichment of Pd (as PdC14*-) from aqueous solution was carried out by means of a liquid membrane containing trioctylamine in the form of an emulsion.77 An 853 1 fibre was synthesized by linking a neutral phosphine oxide to an inert fibre.78 It is an excellent sorbent for Au and has been used in a FI procedure.It will be discussed in the later part of this review. Studies of more specific sorbents such as this will be beneficial. The following adsorbents were also applied to the precon- centration and separation of precious metals: mercaptoace- tamide resin for Au and Pt;79 chelating resin containing NjV- dialkyl-N'-benzoylthiourea for Pd, Pt, Ru, Rh and I.r;80 macroporous TSC-chelating resin for Au, Pd and Pt;81 piper- idine resin 4 10 for Au and Ag;s2 2-hydroxybenzalrhodanine- loaded resin for Pd;s3 poly(2-methyl-5-vinylpyridine)-grafted fibre for Pd, Pt and Rh;84 and amidoxime chelating fibre for Ag.85 The resins which have been investigated for the preconcentration of Au are propanediamide resin,86J7 AP resin and other amide resins,s7 polyamide,88 4 10 guanidine resin,89 macroporous resin containing o-hydroxyazo group,'O Unithiol modified copolymer91 and a copolymer prepared from Amber- lite XE 305 and 1,3,4-thiadiazole-2-amin0-5-thiol.92 An anion-exchange resin with a quaternary phosphonium chloride group in the structure exhibiting high sorption selectivity for precious metal ions such as Au"' and Pt'" was prepared and applied to the separation of Au"' from C U ' ' .~ ~ The macroporous styrene anion-exchange resin D37 1 was used for the preconcentration of Au.94,95 Trace Pt and Pd in ores were concentrated by using a 7 17 strongly basic anion exchanger and activated carbon.96 The positively charged mixed complexes of Au with ethylenediamine, chloride and bromide are readily transported through the cation-exchange membrane and can be separated from PtIV, Ir"' and Rh"I.97 Silver in waste fixing solution was adsorbed by trialkylmethylammonium chloride Levextrel resin as an ion associate, Dithiocarbamatexhitin99 and N-(2-pyridylmethyl)chito- sanloo were synthesized. The former was used for the preconcentration of Au"', Pd" and Ru"', while the later exhibits absorptive activity for Pd" and PtrV.Tetradecyl- dimethylbenzylammonium iodide on naphthalene was used as an adsorbent for Pd after complexing it with protocatechuic acid. 101 Resins impregnated with tribenzylamine,102 N,N-di(sec- octyl)acetamide,1°3 tri(2-ethylhexy1)phosphine oxide104 and TOP0105 were used for the preconcentration of some precious metals. The adsorption mechanism of Pd with primary amine N1923- impregnated silica106 and with trialkylmethylammonium chlo- ride-silica and trialkylamine-silica107 was studied. Silica chemically modified by amine, NH4+, guanidinium, morphol- inium and triphenylphosphonium groups was used for the preconcentration of PGE and Au.108 Silica gel modified by thiourea derivatives was applied to the separation of Pd from PGE and nonferrous metals.lo9 Tin(I1) hexacyanoferrate(I1) exhibits high affinity to Pd" and was used for the separation of Pd" from PtIV and Ir'v.IIO 3R3CH3N+*Ag( S203)23-.98 Chromatography, Flotation and Other Techniques The use of N,N-dialkyl-N'-benzoylthioureas as complexing agents and their recent applications to the enrichment of trace PGE by TLC and HPLC were reviewed.lll The efficient separation of PGE using centrifugal partition chromatography was discussed.' l 2 Platinum and U were quantitatively separated by electrochromatography on papers impregnated with thorium antimonate cation exchanger.] l 3 TLC on chitin and chitosan was used for the separation of six cations including Ag'.' l4 The effective chromatographic separation of Pd" and Pt'" was based on the differing affinities of their chromotrope 2B complexes to Aliquat 336.*15 Mixtures of Pd and Pt were separated by extraction chromatography, after complexation with calmagite, on silca gel columns modified with trioctylamine hydroch- loride.116 The effect of electronegativity of donor atoms on the RF values of Ru1''-tris((3-diketonato) complexes by TLC on silica gel was studied.117 Au and Ag in industrial waste water were concentrated by solvent flotation from an HCl-KI medium, using cetylpyridi- nium bromide as a collector and IBMK as an extraction solvent.118 Gold in ores was preconcentrated by flotation of the ion association complex of Janus Green and chloroaurate, using toluene as a solvent.119 Gravimetric and Volumetric Methods Gravimetric Methods Gold was separated and determined gravimetrically by reducing it to the metal with hydrazinium hydrate in ammonia solution at142 Analyst, February 1996, Vol.I21 pH 10.5-11.120 The gravimetric determination of Au in jewellery was accomplished by reducing it with sodium sulfite.'21 6-Nitro- 1 -hydroxy- 1,2,3-benzotriazoIe122 and sub- stituted triazinone thiones123 were used as complexing agents for the gravimetric determination of Ag. Palladium was precipitated quantitatively and determined gravimetrically with 2-hydroxy- 1 -acetonaphthone oxime and its isomer. 124 The determination of Ag, Au, Pt, Pd, Ru and Rh in pharmaceuticals was accomplished by reducing these metals with noramidopyr- ine and weighing after drying or calcining.125 on the chemical form of Ir present in the solution.The conditions for the amperometric titration of 0svt1' by EDTA were optimized. l53, ls4 Ag in spent photoplates was determined by amperometric titration of its halides with 8-mercaptoquino- line.155 An amperometric titration method for Pt and Ir was established after studying the catalytic effect of FerI1, Cur' and Celv on the reactions of Ir'" and PtIV with 8-mercaptoquinoline and some other sulfur-containing reagents.156 A rapid conductimetric titration method for Agl Au"' and Pd" using diphenylthiocarbazone as a titrant was reported.157 Volumetric Methods Atomic Absorption Spectrometry The volumetric methods here include those the end-points of which are detected by physico-chemical techniques. New volumetric methods developed are scarce. Some efforts have been made to perfect the methods already established.1,3,4,6-Tetrahydropyrimidine-2-thione, 126 DL-cysteine 127 and sodium sulfite128 are recommended as releasing agents for the selective complexometric titration of Pd. The Volhard method was applied to the determination of Ag in a waste fixing bath after centrifugation separation of the Ag2S precipitated with Na2S,129 and of high contents of Ag in rocks after its preconcentration with sulfhydryl cotton.130 A new vanadate- auric indicator I1 was applied to the titrimetric determination of Auo and Aul*"' in natural waters using Mohr's salt as a titrant after preconcentrating them ~eparately.1~~ The iodimetric determination of Au and Ag in electrolytes used in the gold industry was reported.132 A patent was issued for the photo- metric titration of 0 s using a Ba salt solution as titrant and a bisazo-substituted chromotropic acid as indicator.133 The conditions for improving the precision of the spectrophoto- metric titration of Ag with EDTA were optimized.134 A patent was issued for the potentiometric titration of a dimeric RulV chloro complex with ferrocene.135 Pd" was determined in palladium solution for glazing ceramics by zero- current oscillopotentiometric titration at pH 2.0-3.0, using N - cetylpyridinium chloride as a titrant.136 An EDTA titration method for Pd was developed potentiometrically using a Pt electrode modified by a complex of Pd with 4-(3,5-dichloro- 2-pyridy1azo)- 1,3-diaminobenzene as end-point indicator.137 2-Mercaptoquinoline was recommended as a titrant for the differential potentiometric titration of RuVin combination with Osvl, Pd'' and AgI.138 This method has been applied to the analysis of Cu-Ni alloy and enriched slime. A new potentio- metric titration method for the direct determination of Au was developed based on the redox reaction between Au"' and CoL' in the presence of 1,lO-phenanthroline or a,a'-dipyridyl.139 The iodimetric titration of Au"' was improved by using a direct potentiometric procedure in HC104-H2S04 solution and can be used for the routine assay of Au in different samples.*40 Ag' in solutions as a stable complex with ammonium, sulfite, thiocya- nate or thiosulfate was directly determined by potentiometric titration with sodium diethyldithiocarbamate, using a sulfide ion-selective electrode to ensure a well defined potential break.141 A catalytic potentiometric titration method for Ag and Pd was developed based on the catalytic effect of excess titrant, KI or NaC1, on the CeIV-AslI1 redox r e a ~ t i 0 n .I ~ ~ A similar procedure was used for the determination of Ag, Pd and Au in alloys.143 Several other potentiometric titration methods were also presented for the determination of Ag. 144-149 Micro-amounts of Pd" were determined by inverse amper- ometric titration with 2-(2-phenyl-2-hydroxyiminoethyl)-2-iso- quinolinium chloride.150 The amperometric titration of Ir using FeS04 as a titrant was carried out after its oxidation to IrCkj2- with NaC103,'51,*52 the interference of Au"' and Pd" being eliminated by displacing them with active copper powder. 152 The accuracy and precision of this method are highly dependent AAS has been widely used for the determination of precious metals in various materials, especially rocks, ores and metallur- gical intermediates.Separation, preconcentration and some- times dissolution of samples are the vital steps in many procedures, owing to the very low concentration of these metals in many materials, the complexity of the matrix and the refractory nature of some metals. Among the AAS methods developed, the determinations for Ag, Au, Pd and Rh are relatively sensitive. Efforts have been made to make methods reliable and practical. Hence, the precision, accuracy and reproductibility of these methods should be carefully checked.These methods will be discussed in more detail from the point of view of applications. A review was published of enrichment and separation for the AAS determination of Ag.158 A scheme was developed for the rapid determination of precious metals in ores, their processing products and catalysts by AAS and ICP atomic emission spectrometry (ICP-AES). 159 The following methods have been used for the determination of precious metals in geological materials. The determination of pg g-1 and ng g-1 amounts of Ru, Rh, Pt, Ir, Pd, Ag and Au was accomplished by attacking the samples with HF and aqua regia, removing the interfering elements with a cation-exchange resin and measuring in a simultaneous multi-element graphite furnace atomic absorption spectrometer equipped with a Zeeman-effect background correction device.160 An FAAS method for Au and Pd was achieved by extraction of their complexes with 2-mercapto- benzothiazole into IBMK and measuring in the organic phase.161 Au and Pd at ppb levels were determined by ETAAS after their coprecipitation with Hg.162 A ETAAS method was applied to the determination of Pt, Pd, Rh, Ir and Ru in chromite ores using a modified NiS fire assay preconcentration procedure in which CaF2 was added to increase the flux temperat~re.16~ Platinum and Rh were determined by ETAAS after their preconcentration using a VS-I1 anion-exchange fibre.AAS methods for determining Pt, Pd and Au were applied after their extraction with dibutyl sulfide-diisobutyl ket0ne.1~59~66 AAS methods for the determination of 0 s are rare.One report described optimized conditions for the ETAAS determination of Os, which includes a vacuum fire assay process and may be applied to the analysis of ores.167 ETAAS or FAAS methods for the determination of precious metals after their preconcentra- tion by sorption were developed.6345*72,81 The US Bureau of Mines developed an FAAS method for the determination of Au at levels between 0.001 and several hundred troy ounces per ton (1 troy ounce = 31.1035 g) in geological materials and metallurgical test products having complex variations in mineralogy by decomposing the samples with an acid attack procedure, extracting bromoaurate with amyl acetate and measuring in the organic phase.168 A procedure carried out at the UNZA School of Mines during 1990-91 for the determination of Au in soils and rocks was accomplished by bottle rolling of the samples in a cyanide solution, preconcentration by solvent extraction and measuring in the organic phase.169Analyst, February 1996, Vol.121 143 ~~ Numerical studies on the AAS determination of Au in ores and geological samples have been focused on the decomposi- tion stage, or preconcentration stage, or both. The samples were decomposed with HCl and Br2 at room temperature,170 with aqua regia and KBr03,171 with HC1 and bleaching powder,172 sequentially with HBr-Br2 and aqua regia173 and with various combination of NH4HF2, H2S04, HN03, HC104 and KC103 according to the nature of ore samples.*74 The following preconcentration methods have been used: extraction with IBMK7171,l7S--'77 TBP,178 and di-tert-butyl sulfide; 179 sorption with polyurethane foam,60,174,180-182 chelating resin,89y94-95 chelating fibre,73 VS-I1 type anion-exchange fibre183 and resins impregnated with various extractants;l03-105 coprecipitation with Te,170,173 Hg,172 and Cu;177 flotation;119 and fire assay.lg4 Some of these methods were accomplished by ETAAS,180,181 the process time being reduced from 79 to 18 and 25 s through rapid drying and ashing or eliminating the ashing step.181 The sensitivity and precision in the AAS determination of Au in geological samples have been discussed.185 Attention has also been paid on the determination of Ag in rocks, minerals and ores by AAS. Silver in 73 geochemical reference samples was dissolved in HN03-HC104-HF with the addition of iron perchlorate, extracted as iodide with IBMK and determined in the organic phase.186 Microwave digestion was used for decomposing copper concentrates for AAS determina- tion of Ag.187 Several AAS methods for Ag have been reported after its preconcentration by extraction with thiobenzoylamino- benzene in ethyl acetate; ' 8 8 by adsorption with xanthated cotton , 1 89 die th ylammonium die thy ldi thiocarbamate-loaded polyurethane foam190 and polyurethane sponge as an Ag- 1,lO-phenanthroline complex;191 by coprecipitation with CuS as Ag2S;192 and by other means.193 An FAAS method using a quartz boat was proposed for Ag after its concentration by extraction with IBMK.1g4 ETAAS methods for trace Ag were developed, using various sampling techniques and different means of eliminating interferences.195-198 Among the preconcentration methods used for the determi- nation of precious metals in metallurgical intermediate products are the following: synergistic extraction of Pd (in copper anodic slime) with quinolin-8-01 and TBP in CHC13, and back- extraction with dilute HCl containing thiourea;199 extraction of Pt and Pd in sulfate solutions with alkylaniline hydrochloride and petroleum sulfide in toluene and measuring in the organic phase;200 extraction of Au (in retrieved slime material) with Aliquat 336 in diisobutyl ketone;201 and adsorption of Au and Ag with triphenylphosphonate-loaded foam.202 The interferences of co-existing elements, such as Cu and Fe, in the FAAS determination of Au in mill cyanide solutions were studied and it was found that the interference of Fe is nullified owing to the releasing action of Cu.203 Gold and Ag in industrial waste water were determined by AAS after their preconcentra- tion by solvent flotation.118 Gold was extracted from cyanide- containing wastewater into tri-Cg-l -alkylmethylammonium chloride for direct FAAS determination.204 Gold in waste water was preconcentrated on a TOP0 chemically modified tungsten wire and measured by ETAAS.205 Silver in waste water was determined by ETAAS after its preconcentration by sorption on a Nafion-modified tungsten wire electrode206 and on dithizone particles in an ultrasonic field,207 and by coprecipitation with HgS.208 A new design of graphite atomizer crucible with independent electric heating was proposed for the determination of Ag and base metals in aqueous suspensions collected on ultrafiltration membranes, the detection limits of the elements being lowered by factors of 3-10.209 Palladium in environ- mental water samples was separated by a dithizonesulfonate- immobilized anion-exchange resin and measured by ETAAS using a resin-water suspension sampling technique.210 Nickel@) was used as a matrix improver for the direct determination of Au in effluent by AAS using an angular pyrolytic graphite platform.211 The preconcentration and AAS determination of Ag in natural waters were r e p ~ r t e d .~ 1 2 v 2 Gold, Pd and Pt in high-purity silver were determined by ETAAS with two different solid sampling techniques, one being the in situ dissolution of a solid sample by addition of dilute HN03 in a cup in a tube atomizer, and the other a slurry method by which Ag samples were dissolved in HN03 and then agitated to suspend the undissolved elements.214 Closed-vessel micro- wave dissolution followed by ETAAS was employed for the determination of Au, Pd and Pb in platinum to reduce contamination.215 Silver in high-purity cathode Cu was deter- mined after its preconcentration by sulfhydryl cotton.2l6 Au and Ir in non-ferrous metals were determined by AAS after their preconcentration with chelating resin.68 A rapid FAAS method was employed to determine Pt in Pt-Pd alloy for quality control.217 Platinum in catalysts was determined by FAAS directly in the solutions obtained by decomposing samples with HF2I8 or by ETAAS after its separation by extraction.219 Pd in a palladium catalyst was determined by FAAS.102 Iridium in environmental samples was determined by ETAAS after liquid anion-exchange separation.z2O Platinum in biotic and environmental samples was enriched by electro- deposition into a graphite tube packed with reticulated vitreous carbon and measured by ETAAS.221 There have also been reports describing AAS methods for the determination of Pt in human fluids222 and animal tissues,223 Ru in biological tissues,224 Au in Chrysanol oily injections,225 Ag in and silver227 or silver stearates228 in workplace air.Palladium was used as a modifier to increase the char temperature for the determination of Ag in biological ~amples.~2~ AAS methods were developed for the determination of Au and Ag in activated carbon,230 of Ag in electroplating baths231 and in high-temperature superconductive ceramics232 and of Pd in photographic film,233 in phenols234 and in manganese and nickel c0mpounds.~~5 To remove co-existing ions and salts for AAS analysis, Pt and Au were extracted with trioctylmethylammonium chloride,236 Pd was extracted with ammonium tetramethylene- dithiocarbamate236 and with 3-ethyl-2-methylbenzothiazolium iodide,237 Pd, Ag, Pt and Au were sorbed with xanthate- or dithiocarbamate-immobilized silica gels,238 Pd was coprecipi- tated with zirconium hydroxide239 and Au was preconcentrated by Donnan dialy~is.2~0 Some efforts have been made to improve the atomization stage, the sample introduction stage and others.These will be described briefly.Cu(N03)2 was used as a chemical modifier for the determi- nation of Au by ETAAS to eliminate the interferences caused by co-existing elements, such as Na, K, Fe and Ni.241 The interferences of complexing agents in the FAAS determination of Au were eliminated by the addition of an excess of sodium diethyldithi~carbamate.~~~ The harmful effect of heteropoiyan- ions on the AAS determination of Au and Ag caused by the accommodation of the analyte on the surface and cavities of crystals formed by the heteropolyanions can be eliminated by adding excess of cyanide, resorcinol, ethylenediamine or EDTA.243 The interferences of certain cations and anions on the FAAS signals of analytes such as Pd and Au were eliminated by the addition of excess cyanide, which scavenges the oxidizing species in the flame to produce a reducing atmosphere favouring carbide formation and the production of measurable analyte atoms.244 In a ETAAS method for the determination of Pd, a 20000-fold excess of lead can be separated by an optimized temperature programme before Pd atomization starts.245 In the FAAS determination of Rh, the interference of cations and anions can be completely eliminated by the use of excess of Brilliant Green as a releasing agent, which has two roles: first, it forms a stable complex with the analyte, so that the144 Analyst, February 1996, Vol.121 Rh species will reach the flame independently; second, the excess of Brilliant Green scavenges the oxidizing species in the flame, providing favourable conditions for Rh atomization.246 IBMK and ethanol were recommended as solvents and heptan- 1-a1 as a buffering agent for the FAAS determination of Rh in organic solutions containing toluene, hex- 1 -ene, heptan- 1 -01 and heptan- 1 -al.247 In the study of Ag analyte loss as a function of temperature and time during the thermal pretreatment stage for ETAAS, the presence of Mg, both alone and with Pd, displayed a temporal stabilizing effect on the Ag anal~te.~~X By optimizing the conditions of slotted quarts tube atom- trapping AAS, the sensitivities of Au and Ag were increased 62- and 106-fold, respectively, in comparison with conventional FAAS.249 By combination of a high-efficiency nebulizer, high- performance hollow cathode lamp, slotted quartz tube and flow injection analysis, a technique named microamount FAAS was developed and applied to the direct determination of Au in ores with detection limits down to 0.2 ng m1--1.250 The effects of various factors on the determination of Ag by microwave-induced plasma AAS with an ultrasonic nebuliza- tion sample introduction system251 and with an electrothermal sample-introduction device252 were studied. A slurry ETAAS method was developed for the determination of Au after its accumulation by bacterial ~ e l l s .~ ~ 3 Laser-enhanced ionization combined with rod-flame and graphite furnace atomizers- ionizers was used for the direct determination of Au in solid silver nitrate samples.254 After studying the effects of atomization temperature on the characteristic masses and the atomic absorption coefficient of Ag, a standardless ETAAS method was established and appplied to the determination of Ag in real samples.255 The absolute analysis by ETAAS established earlier has been extended to the determination of Ag.256 The roll-over effect caused by stray light originating from a pulsed hollow-cathode lamp and the other factors affecting the sensitivity of the ETAAS determination of Au have been in~estigated.~57 Using a shadow spectral filming technique, the Ag layers in ETAAS were found to be relatively uniform but may be distorted by the internal gas flow and non-isothermality of the atomizer.258 A study of the atomization of Ag in ETAAS was presented.259 Atomic Emission Spectrometry Many studies in AES were concentrated on the determination of impurities in fine precious metals and their alloys.Attention was also paid on the determination of these metals in ores and materials of industrial importance. A method of decomposing placer platinum samples was proposed and this made it possible to determine Pt, Pd, Rh, Ir, Os, Ru, Au and some base metals simultaneously.260 The decomposition procedure consists in treating the samples with potassium tetrafluorobromate and converting the complex fluorides into soluble complex chlorides. Ultra-trace levels of Pt, Pd and Au in geological samples were preconcentrated from 10 g of sample into a I mg silver bead by minimized lead fire assay and determined with AES by arcing the bead in a graphite electrode along with 0s-Sb alloy, osmium acting as a spectrochemical carrier.261 In the ICP-AES determination of PGE and Au in geological samples following the NiS fire assay and Te coprecipitation, the recovery of these elements was improved by employing screw-capped Teflon bombs for dissolving the NiS beads or the Te precipitates.262 The precipitation of Crystal Violet-precious metals-tin" chloride has been applied to the enrichment of Pt, Pd, Rh, Ir and Au for ICP-AES determination.263 The importance of standard preparations and measurement control in the ICP-AES determination of precious metals was explained with examples.264 Comparisons of the ICP-AES results with gravimetric results for precious metals in different routine samples showed excellent agreement.265 It was sugges- ted that precious metals at low levels in geological samples should be determined by a combination of ICP-AES with fire assay.266 A spark emission spectrometer with improved sensi- tivity was recommended for the determination of precious metals.267 For the AES determination of Au in ores and geological samples, the following techniques were found useful: a mixture of HN03, bromine water and table salt for decomposing gold- bearing ores,268 a mixture of 6% H3P04 and 4% NH4HF2 as chelating agent to prevent the co-adsorption of Fe with Au in a polyurethane foam separation procedure,269 and an arc AES method with gas chamber profile electrodes using NH41 as a spectrochemical carrier for the determination of ultra-trace amounts of Au.~~O A sequential ICP-AES method was presented for the determination of Au in sludge and soil samples after its preconcentration with sulfhydryl cotton fibre.27' Gold in copper concentrates was determined by ICP-AES after its pre- concentration by extraction with IBMK or by coprecipitation with copper.177 Gold in black shales was determined by chemical spectral analysis.272 ICP-AES methods proposed for the determination of impuri- ties in fine metals are the following: the direct determination of I3 impurities in 99.98% palladi~rn;~73 the direct determination of 13 impurities in 99.99% platin~m27~ and palladium275 with the equivalent concentration subtraction method in which the matrix element did not add to the calibration standards and the interference caused by the background of the matrix was corrected by subtraction of a value equivalent to the concentra- tion of the analyte; the determination of 18 elements276 or 28 elements277 in gold after removing the matrix by extraction with diethyl ether, the equivalent concentration subtraction method being used in the later procedure to eliminate the inter-element interferences: the determination of impurities in silver;278,279 and the determination of Ag in copper.280 The advantage of equivalent concentration subtraction method is that the proce- dure is simple and a large amount of precious metals can be saved.A rapid AES method was established for the determina- tion of five impurities in gold ingots by direct arcing the solid samples.281 The AES analysis of 0 s powder of 99.5-99.94% was performed after evaporation of 0 s in a quartz boat using graphite powder as an impurities catcher.282 A review was presented describing ICP-AES methods for the determination of Au in gold jewellery alloys.283 ICP-AES methods were applied to the determination of the minor alloy elements and main impurities in gold jewellery alloys,284 of 22 impurities in Pd-Ag alloys after their separation from Pd and Ag by coprecipitation with Y(OH)3 and precipitating Ag as AgC1,285 of Rh and 12 impurities in Pt-Rh and Pd-Rh alloys,286 of Ir and 12 impurities in Pt-Ir and Pd-Ir alloy^,^*^ of Ag, Pd and Au as the main components in some alloys,143 of trace Ag in superalloy after separation with sulfhydryl cotton288 and of impurities in silver all0ys.~78 An ICP-AES method with dynamic background correction was proposed for the direct determination of Pt and Pd in catalysts after decomposing the sample with a mixture of HNO?, HF and HC104.289 Pt290 and Ru29I in catalysts were separated by extraction with xanthates and determined by ICP- AES.Ag in uranium oxide292 and Pd in stannic oxide293 were also determined by ICP-AES. For control purpose, the major components and 12 potentially detrimental impurity elements in gold electroplating solutions were determined by ICP-AES.294 The conditions for precipita- tion of Pd and Ag with CuS have been studied and applied to the preconcentration of these elements from copper salt solutions for AES determination.295 ICP-AES methods were developed for the determination of Au and Ir in non-ferrous metals68 and of Ag in waters.296 AES methods were also developed for theAnalyst, February 1996, Vol.121 145 determination of Au, Pd, Pt and Rh in technical sol~tions,29~ and of Pd and Pt in high-purity nitric a~id.~98 Styloscopy was employed for the analysis of alloys based on silver and platinum.299 Flame AES was applied to the detection of 0 s after its extraction as Os04 into IBMK.300 ICP atomization combined with laser excitation was employed for the single- and double-resonance atomic fluorescence spec- trometry of Agl, Au', Pd' and Pt', and this technique has been applied to the determination of Ag in c0pper.3~~ UV/VIS Absorption Spectrometry The application of UV/VIS absorption spectrometry to the determination of precious metals is still popular in many laboratories, especially in developing countries. Many new organic reagents have been synthesized and various highly sensitive methods developed with molar absorptivities of 105-106 or even higher.In studies for establishing new methods, much attention was paid to sensitivity. However, specificity and selectivity are also very important. Developing more methods with good selectivity will be beneficial for applications. Since so many spectrophotometric methods have been developed and most of them can be applied to the analysis of simpler samples, the application aspects of these methods will not be described except for some complex materials. The methods not described in the text will be found in Table 1. A critical review302 focused on the ultra-sensitive spectro- photometric methods for precious metals with molar absorpti- vities of 2 2 X lo5.Other reviews dealt with spectrophotome- tric methods for Pd since 1976303 and for Ag in the period 1970-1990.304 Gold in ores was extracted into phenyl sulfoxide-IBMK and developed colour in the organic phase with 4,4'-bis(dimethy- 1amino)thiobenzophenone (thio-Michler's ketone, TMK).30s Several methods for the determination of Au in minerals, ores, geological samples and metallurgical tailings with TMK were developed after its preconcentration by various tech- niques.59,62~306-3~~ A patent for the rapid determination of Au in geological samples was i~sued.3~0 Simultaneous determinations of Au and Ag with TMK-Tween-80 by Kalman filter spectrophotometry,31 of Ag and Pd with TMK-Tween-80312 and of Pt and Pd with TMK3'3 were reported.Silver in crude lead was directly determined with TMK.314 A liquid bead extraction-spectrophotometric method was developed for Au in geological samples by preconcentrating it with polyurethane foam and subsequently extracting with two drops of TBP containing 4,4'-bis(diethy1amino)thiobenzophenone (gold rea- gent).3Is A similar method was also developed for the determination of both Au and Ag.316 A fully differential spectrophotometric method was proposed for Ag in geological samples by developing colour with gold reagent in the presence Dual-wavelength spectrophotometry for Rh and Ir was recommended with extraction of their complexes with SnBr2 and 2-mercaptobenzothiazole into ethyl acetate and has been used for the determination of these elements in metallurgical intermediate pr0ducts.3~8 Trace Pt in glass was extracted from HC1 solution containing SnC12 and dithizone on to polyurethane foam, eluted with acetone and measured in the eluate.319 Platinum and Pd were determined simultaneously by dithizone extraction and dual-wavelength spectrophotometry.320,321 A patent claims a test kit for the determination of a number of elements, including Pt and Pd, using dithizone as a major reagent.322 The derivatives of thiourea used in spectrophotometry are as follows: 1 -hydroxy- 1 -(p-ethoxyphenyl)-3-alkyl/arylthiourea~23 and 2-hydroxy- 1 -acetonaphthone thio~emicarbazone3~4 for Pt'"; 4-(dimethy1amino)phenyl- 1 -phenylthiosemicarbazide,325 o-hy- droxyacetophenone thiosemicarbazone,326 5-bromosalicylalde- of op.317 hyde-4-phenyl-3-thiosemicarbazone (extraction with CHC13)327 and 2-hydroxy- l-naphthaldehyde 4-phenyl-3-thio- semicarbazone (extraction with CHC13)328 for Pd"; and bis(thio- phene-2-a1dehydo)thiocarbohydrazone for Ru"' and Ir'II.329 The simultaneous determination of Pd and Pt was proposed by first- derivative spectrophotometry of their complexes with salicy- laldehyde thiosemicarbazone.33" N-Allyl-"-(sodium p-benze- nesu1fonate)thiourea was suggested for the identification of Ag', 0s""' and Ru1".331 The determination of Pd in carbenicillin sodium and catalysts was accomplished by developing the colour with glyoxal bis(4-phenyl-3-thiosemicarbazone in aque- ous DMF solution.332 Table 1 Reagents used for the photometric determination of precious metals Element Reagent Ref.Ag Pd Ru, Rh Pd Pt Pd Au Ag Pd 0 s Au Pd Pd, Rh 0 s Pd Pd, Ru Pt Pd Ag Ag Pd Ag Ag Pd Pd Rh Ru Pd Pt Au Ag Pd 0 s Cyclopentanone-thiodiglycolic acid 450 Hexy lthiogly colate 45 I Cyclohexylthioglycolate 452 Amyl thioglycolate 453 Amyl thioglycolate 454 N-Thiobenzoyl-N-p-chlorophenylhydrox y lamine 457 Sodium pentamethylene dithiocarbamate 45 9 Thiosalicylic acid and hexylamine, extraction with CHC13 455 Di thiooxamide 456 N-Methylaniline carbodithioate 45 8 Sodium pentamethylene dithiocarbamate 460 1 -Hydroxy-2-pyridinethione (Na salt) 46 I 2,2'-Diaminodiphenyl disulfide 462 Cyclopentanespiro-2'-( 1 -methyl-2',4'-dithio)-s-triazine 463 6-(2'-Hydroxyphenyl)-2,4-dithio- 1 -phenyl- 1,3,5- triazine 464 Bis(thiophene-2-a1dehyde)thiocarbohyrazone 465 Thiopyrine, extraction 466 Phenylthiohydantoic acid 467 Cyclohexyl xanthate 468 4,7-Dimethyl- 1,3-dioxaphosphepane-2-thiol-2-thione 469 Schiff base derived from 3-methyl-4-amino-5- mercapto- 1,2,4-triazole and piperonal 470 5-Amino-4-arylazo-3-phenylpyrazole 47 1 2-Chloro-4-nitrobenzeneazoaminoazobenzene, Triton X-100 472 5 -(4'-Amino-2'-hydroxybenzeneazo)tetrazole 47 3 Triton X- 100 474 3-Phenylazoh ydrox ylamine 475 2-(4-Antipyrinylazo)-5-dimethylaminoaniline 476 Semi Xylenol Orange 477 Bis(benzoy1acetone)-meso-stil benediimine 478 &,a'-Dipyridyl, I-, extraction with CHC13 479 a-Benzoin oxime, extraction with IBMK 480 Sodium 2-(8-hydroxyquinolin-5-ylazo)benzoate 48 1 Ferron, extraction with tribenzylamine-CHC13 482 Cadion 2B, nonylphenol polyethoxylate, SDS 483 2,4-Dibromo-6-carboxybenzenediazoaminoazobenze, Triton X- 100 484 cx-Isonitroso-2-acetonaphthone, extraction with 485 Isonitroso-p-methylacetophenone, extraction with 486 Isonitrosomalondiani1ide, extraction with CHC13 487 Pyridine-2-acetaldehyde salicyloylhydrazone, extraction with CHC13 488 N-m-Tolyl-m-nitrobenzohydroxamic acid 489 Sulfochlorophenolazorhodanine 490 4-Methylbenzeneazorhodanine 49 1 Dimethylaminobenzylidenerhodanine, SDS 492 3-Ethoxycarbonylmethyl-5-@- 2,6-Dibromo-4-sulfamoylphenyldiazoaminobenzene, CHC13 CHC13 diethylaminophenylamino) rhodanine 493 p-Hydroxyphenylfjuorone, CTAB 494 2,3,7-Trihydroxy-9-@-nitrophenyl)fluorone, CTAB 495146 Analyst, February 1996, Vol.121 Several new pyridylazo reagents were synthesized, among which 4-(5-nitro-2-pyridylazo)resorcinol reacted with Pd", PtIV and Oslv to form coloured complexes.333 Propanol was suggested for enhancing the colour reaction between Rhl" and 4-(2-pyridylazo)resorcinol (PAR).334 The spectrophotometric determination of Pd was performed after extraction of the ion associate formed by the Pd-PAR complex and tetraphenylar- sonium ion3335 or tetraphenylstibonium ion336 into CHC13.Iterated target transformation factor analysis was applied to resolve the overlapping spectral data and led to the simul- taneous determination of Pd and Rh after developing their colour with 2-(5-bromo-2-pyridylazo)-5-dimethylaminophenol and Triton X- A number of substituted pyridylazo compounds were used for the determination of Pd,338-343 olylazo reagents for Pd were synthesized by introducing an arsonic group ortho to the azo group,351-353 the best of which is 2-(2-benzothiazolylazo)-5-dimethylamino-4-tolylarsonic acid. A method was described for Pd by converting it into a coloured compound with 4-(4-methyl-2-thiazolylazo)resorcinol in the presence of cetylpyridinium chloride (CPC) and extracting with CHC13.354 Three thiazolylazo chromogenic reagents, namely, 2-(2-benzothiazolylazo)-5 -sulfomethylamino-4-tolylphos- phonic acid, 2-(2-benzothiazolylazo)-5-sulfomethylamino- 4-chlorobenzenephosphonic acid and 2-(2-benzothiazolylazo)- 5-sulfomethylamino-4-tolylarsonic acid, were synthesized for Pd.355,356 Several other thiaolylazo derivatives were applied to the determination of Pd357-360 and Rh.360-365 The derivatives of 5-arylazo-8-aminoquinoline synthesized and applied to the determination of precious metals are the following: 5-(4-acetylaminophenylazo)-8-aminoquinoline,~~~ 5-(4-methoxyphenylazo)-8-aminoquinoline3~~~~~~ 5-(4-methyl- phenylazo)-8-aminoquinol ine,368 5 -(2-benzothiazolylazo)- 8-aminoquinoline,369 5-[(phenylazophenyl)azo]-8-amino- quinoline370 and some derivatives of 5-thiazolylazo- 8-aminoquinoline37* for Pd; 5-(4-sodium sulfonatopheny1azo)- 8-aminoquinoline,372 5-(4-arsonophenylazo)-8-aminoquin- oline373 and 8-aminoquinolyl-5-azo-p-benzoic a ~ i d 3 ~ ~ for Au; and 5-(4-nitrophenylazo)-8-~4-toluenesulfonamido)quinoline for Pt.375 Cationic surfactants were used to improve the sensitivity in many cases.In one report,372 the effect of surfactants was studied by changing the SO3H2 group in the reagent to CH30, CH3C0, NO2, COOH and As03H2, and it was found that cationic surfactants only affect the complexation reaction between Au"' and reagents which have an anionic substituent group with an increase in sensitivity. One diazo derivative of chromotropic acid, chlorophosphonazo-mA, was used for the direct determination of Pd in Ag-Pd alloys in the presence of cetyltrimethylammonium nitrate.376 Some other derivatives of chromotropic acid were used for the determina- tion Ru377 and Rh.378 The effects of surfactants and solvents on the analytical characteristics of complexes of Rh with azo dyes were examined.379 The possible use of some naphthoic acid azo dyes for Pd380 and of some symmetric monomethinecyanines based on pyrrolo- and imidazo[ 1,2-a]benzimidazole for A u ~ ~ ~ was reported.Spectrophotometric methods for precious metals using porphyrins and porphines as reagents are very sensitive. meso- Tetrakis(4-acetoxypheny1)porphyrin was used for the determi- nation of Au382 and Rh.383 The complex of Au"' with this reagent in the presence of sodium dodecyl sulfate (SDS), Triton X-100 and Cd" exhibits a molar absorptivity of 1.62X 106 and has been applied to the determination of gold in geological samples. The reaction of Ag with 5,10,15,20-tetrakis(3-chloro- 4-sulfopheny1)porphyrin in the presence of cetyltrimethy- lammonium bromide (CTAB) and Pb" has been applied to the direct determination of trace silver in anode mud and photo- graphic film.384 meso-Tetrakis(p-sulfopheny1)porphyrin was proposed for the determination of Rh"' in the presence of CTAB Rh,344,345 Pt,34637 Os,348 Au343 and Ag.349350 Eleven thiaz- ~ ~ ~~~ with a molar absorptivity of 1.1 X 107 by measuring the fourth derivative of the absorbance and this method has been applied to the determination of Rh in ores after its preconcentration by fire assay and solvent extraction.385 This reagent was also used for the determination of R u " ' .~ ~ ~ Other reagents include meso- tetra(3-methoxy-4-hydyoxyphenyl)porphine for the simultane- ous determination of Pt and Rh,387 meso-tetra(4-chloro- pheny1)porphyrin for the determination of Ag388 and meso- tetra(4-trimethylammoniumphenyl)porphyrin for the determi- nation of A u . ~ ~ ~ Some water-soluble porphines were synthe- sized and their complexes with Ag, Pd and Au were studied.390 The reactions between 0 s and some N'-benziloyl-N-(sul- fony1)hydrazines were highly selective, and procedures for the extraction-photometric determination of 0 s in industrial com- plex solutions have been e~tablished.39~-393 The following reagents were also recommended: ferroin-yielding as-triazines for R ~ , 3 9 ~ a-substituted alicyclic oximes for Rh395 and azastyrene Schiff bases for Au and Ag.396 The ion associates usually formed between the cations of basic dyes and the anion complexes of precious metals with halide, thiocyanate and tin(i1) chloride are always highly sensitive.A review discussing two ultrasensitive groups of reaction systems, namely the PGE-SnCl2-basic dye-surfactant system and PGE-KI-basic dye-surfactant system, has been published.397 Gold in ores was preconcentrated with sulfhydryl cotton and determined spectrophotometrically after extracting its ion associate AuC14--phenosaframine into isoamyl ace- tate398 or developing the colour with the use of the Au"'-SCN-- Rhodamine B (RB)-arabic gum ~ystem.39~ A photometric method was proposed for Au in recycled soldering materials by extracting it with CHC13 in the presence of DMF and developing the colour with Methylene Blue in the organic phase.400 By measuring with a total reflection long capillary cell, the sensitivity of the bromoaurate-ethyl-Rhodamine B ion associate in benzene was greatly increased, covering a linear range of 20-60 ng ml-l of Au.~OI To increase the sensitivity, a dual-wavelength spectrophotometric method was recom- mended for the determination of Ag by measuring the difference in absorbance of the butyl-Rhodamine B-Ag132-- Tween 80 system at 605 nm (positive peak) and 555 nm (negative peak).40* The ion association complexes formed in Pt1"-SCN--RB,403 Ag'-I--Ethyl Violet404 and Agl-adenine- eosine405 systems were proposed for determining the respective elements.The chromogenic systems of Pd"-SCN--Methylene Blue extracted with CHC13,406 of Pd"-SCN--2-[ 1-(5-dimethy- lamino-2-thienyl)vinyl]- 1,3,3-trimethy1-3H-indole extracted with benzene or toluene,407 of PdI1-SCN--2[2-(5-dimethyla- mino-2-thienyl)vinyl]- 1,3,3-trimethy1-3H-indolium extracted with several solvents,408 of Pd"-I--tetrabutylammonium ex- tracted with CHC13,409 of Agl-I--l, 1'-tetramethy- lenebisindocarbocyanine extracted with toluene410 and of AuC14--Crystal Violet extracted with toluene41 1 were pro- posed for determining the corresponding precious metals.A patent was issued for determining Au by extracting it from HCl solution into amyl acetate and subsequently treating the extract with azide ion as a negative ligand and N,N'-dimethy- lindodicarbocyanine as a counter part to form an associate.412 Pt-SnCl2-buty1-Rhodamine B,413 Rh-SnC12-Chrompyrazol 1,414 0s-SnCl2-Brilliant Green415 and 0s-SnCl2-Crystal Vio- let416 systems were proposed for determining Pt, Rh and Os, respectively. The ion association complex of Ru"' with 2,4,6-tris(2'-pyridyl)- 1,3,5-triazine as a primary ligand and picrate as a counter ion was extracted into 1,2-dichloroethane for the determination of Ru.417 Centrifugation spectrophotometry and flotation spectro- photometry for precious metals are usually connected with the formation of ion association complexes.A review described the solvent flotation separation and the subsequent spectro-147 Analyst, February 1996, VoE. 121 photometric determination of 0 s and Ru by their ion associates with basic dyes.418 One method for Au was developed involving the centrifuga- tion separation of the bromoaurate-RB and chloroaurate- Rhodamine 6G (R6G) ion associates and dissolving the precipitates in ethanol with molar absorptivities of 3.39 X 10" and 2.25 X 108, respectively.419 The mechanism of these surprising ultrasensitive colour reactions was attributed to the adsorption of a large excess of dyes by the species [AuBr4--RB+] and [AuC14-.R6G+], but the reason is unknown.This method has been applied to the direct determination of Au in geological samples. Another centrifugation-spectrophoto- metric method for Au in ores with a molar absorptivity of 3.3 X 106 was reported using the Au-SnCl2-Malachite Green (MG) system and dissolving the precipitate in ethanol.420 Pt-SnC12- MG was recommended for the determination of Pt with a similar procedure.421 The flotation spectrophotometry of Au was reported by separating the Au-I--Methylene Blue ion associate on shaking with cyclohexane and dissolving in methanol and has been applied to the determination of Au in ores, anodic slime and cyclone dust.422 The other systems used in flotation spec- trophotometry are Ag-I--SnC12-MG in the presence of poly(viny1 alcoho1),423 Ru (and/or Os)-SnC12-basic and Pt-didodecyldithiooxamide in the presence of SnC12 and CTAB.425 Various spectrophotometric methods with the use of different multicomponent complexes have been reported.Osmium reacted with pyrocatechol and N-hydroxy-NJV-diphenylben- zamidine to form a coloured species extracted into CHCl3 with a molar absorptivity of 3.95 X 106 and very high selectivity.426 Silver in copper amalgams was directly determined by using its ternary complex with 1,lO-phenanthroline and Bromopyro- gallol Red.427 Silver was also determined with the use of 1,lO-phenanthroline and tetrachlorotetrabromofluorescein in the presence of OP,428 and with o-hydroxybenzenediazoami- noazobenzene and triethanolamine in the presence of Triton X- 100.429 The mechanism of the synergistic sensitizing effect of mixed ionic and non-ionic surfactants on some multicomponent complexes of Ag was discussed.430,431 The determination of Pd was reported by the use of 1 ,lo-phenanthroline, cadion A and Peregal 0,432,433 and of 1,lO-phenanthroline and 1 -(2-carbox- yphenyl)-3-(4-phenylazophenyl)triazene in the presence of T ~ e e n - 8 0 .~ 3 ~ Osmium was determined by adsorptive extraction of its ternary complex with 3-(4-phenyl-2-pyridyl)-5,6-diphenyl- 1,2,4-triazine and tetraphenylborate anion on to microcrystal- line naphthalene and subsequently dissolving the coloured naphthalene mixture with DMF for measurement.435 Similar procedures were described for Pd by extraction of its complexes with N-p-tolylthiobenzohydroxamic acid,436 4-(dimethylami- no)benzaldehyde thiosemicarbazone437 and 1 -allyl-3-(2-pyri- midyl)thi~urea,~~* and for Ag by extraction of its complex with l-allyl-3-(5-chloro-2-pyridyl)thiourea.4~~ Microcrystalline p - dichlorobenzene was used for extracting the complex of Ru with 3-hydroxy-2-methyl- 1,4-naphthoquinone 4-oxime.440 Several traditional methods were recommended for the determination of precious metals in some specific objects: Rh in electrolyte by directly oxidizing it to produce a blue-violet colour with sodium bi~muthate;~~l Ru in resistive paste and powdery oxide samples for integrated electronic circuits by oxidation to R u O ~ ~ - with K2S208;442 Pd in ammonium chloride electroplating baths by measuring the photoabsorption of PdC142-,443 and by other means;444 Ag" oxide by measuring Mn3+ produced from its oxidation of Mn2+;445 Rh4463447 and Pt448 in various materials by developing the colour with SnC12; and Ru and 0 s in the same solution by second-order derivative spectrophotometry after converting them into complexes with SnC12.449 Luminescence Analysis The luminescence determination of precious metals has devel- oped rapidly in recent years.These methods are very sensitive, with detection limits usually in the range 0.1-1 .O ng ml-I. Chemiluminescence A review has been published on the application of chemilu- minescence (CL) for the determination of precious metals.496 A method for Au based on the direct combination of solvent extraction with luminol CL in reversed micelles was developed and has been applied to the determination of trace Au in silver- based alloy.4973498 When a solution containing dodecylbenzene sulfonate, B-cyclodextrin and acetone is used as eluent, Au eluted from foamed plastic can be stabilized and determined directly in the eluate by the luminol-H202-AuC14- CL method.499 Solid-surface CL analysis methods have been developed.In one, Au"' was adsorbed on the surface of foamed plastic and then reacted with lumino1.500.501 In another, Ag' was selectively retained on the centre of a filter-paper by the ring- oven technique and determined by the CL reaction of Ag' with luminol and K2S208.502 The CL of 4-diethylaminophthalhydrazide oxidized with water-soluble oxygen catalysed by Pt'" provided the basis for the determination of Pt.503 The CL of the 6,7-dihydroxy- 2,4-dimethylbenzopyrylium chloride-H202 system catalysed by Au"' was utilized for the determination of Au in ores after its preconcentration using sulfhydryl cotton.504 A method for the determination of trace Ag in waste water was developed based on the CL reaction of the 9,lO-dimethylacridinium fluorosul- fOnate-K&08 system catalysed by AgI.505 The electrochemi- luminescence of a new reagent, 6-[2-hydroxy-4-(diethylami- no)phenylazo] -2,3-dihydro- 1,4-phthalazine- I ,4-dione, was employed for the determination of Ag as it exhibits a significant effect on the efficiency of light emission of the reagent.506 The CL of Ru complexes containing a,a'-dipyridyl used as a possible detection technique for HPLC and FI has been reviewed.507 Several papers dealt with CL methods for the determination of Ir by the use of its complex with a,&'- dipyridy1,508-510 in which one was accomplished after selective sorption of Ir on silica-based anion exchangers,508 and the other was applied to the determination of Ir in standard reference samples of platinum concentrates.509 A CL method for the determination of Rh using its a,a'-dipyridyl complex has been reported.511 Fluorescence A method for Ag was proposed by measuring the intensity of the fluorescence of the reaction system fluorescein-mercury- sulfide-silver(I)-ethanol and has been applied to the determi- nation of Ag in silicate samples.512 A patent claims a method and an apparatus for the rapid determination of Au in soil or rocks by extracting it with a crown ether, labelling with a chromophore and measuring using fluorimetry.513 Lumines- cence methods were developed for the determination of Ru and 0 s after the formation of complexes with 1,lO-phenanthroline (phen) and have been applied to the determination of these elements in ores.514-515 The limits of determination were lowered ten-fold by sorption of Ru-phen or 0s-phen complexes on Silochrom silica.514 Agr exhibits enhanced effect on the fluorescence of 9-[(methylamino)thiocarbonyl]anthracence.516 A fluorescence quenching method for the determination of Ag was developed by the use of a newly synthesized reagent, 3,5-dibromosalicylaldehyde thiosemicarbazone, and has been applied to the determination of Ag in anode slime.517 The fluorescence quenching of Rhodamine 6G (R6G) by the formation of [RhC12(SnC13)2]-R6G was employed for the determination of Rh in the presence of other PGE.518 Fluores-148 Analyst, February 1996, Vol.121 cence quenching methods for the determination of Pd, Au and Ag were developed based on the reactions of Pd with Eriorubine B5 l9 and with rneso-tetrakis(acetylpheny1)porphyrin in the presence of OP,520 of Au with a,fi,y,&tetrakis(4-trirnethy- lammoniumphenyl)porphyrin,521 and of Ag with cyclic tri- and tetraamines containing fluorescent groups,522 respectively. Platinum and Au were determined by extracting their ion associates with acriflavin into IBMK, and measuring the fluorescence of cationic dye in the organic pha~e.5~3 Platinum was also determined by extracting its ion associate with Nile Blue B into a mixture of dichloroethene and trichloroethylene and measuring the fluorescence in the extractant.524 The extraction-fluorimetric method for Au by using its associate with Rhodamine B (RB) has been applied to the determination of Au in ores and technological ~ a t h o l y t e s .~ ~ ~ Laser-induced fluorescence and a microcomputer-controlled system for the determination of trace gold have been developed by the use of the ion associate [AuCL-1-RB in diisopropyl ether with a detection limit of 0.01 ng rnl-l.s26,527 Catalytic Kinetic Methods Since the precious metals are excellent catalysts, catalytic kinetic methods (henceforth, catalytic methods) of these metals have long been investigated by many workers. The catalytic methods considered here are referred to those based on the chemical changes of indicator reactions catalysed by ions to be determined.The indicator reactions in these methods were usually monitored by photometry. Electrochemical detection techniques were also used recently. These methods are very sensitive, many of them giving detection limits of 0.01-0.1 ng ml-1. However, they usually suffer severely from interfer- ences from many other ions and often require strict control of the experimental conditions and complete remove of interfer- ences, The determination of 0 s and Ru in complex samples is easier to carry out, because they can be completely separated from all interferences by distilling them in the form of tetraoxide. A review relating to the application of kinetic methods for the determination of platinum group metals for rapid monitoring of technological processes has been published.528 Detection by Spectrophotometry A catalytic spectrophotometric method for the determination of Ag was developed based on the catalytic effect of Ag on the oxidation of C1- with Ce'" in the presence of a,&'-dipyridyl and monitoring the Ce"' produced with alizarin-3-methylamine- NJV-diacetic acid and fluoride as chromogenic reagents.529 This method has been applied to the determination of trace Ag in ores after preconcentration with sulfhydryl cotton. Trace amounts of Rh, Ru and Ir were separated from a 107-fold excess of Fe by a cation-exchange technique and determined by catalytic meth- ods.530 A procedure was presented for the catalytic determina- tion of Os, Ru, Ir and Rh in a single sample of copper Highly sensitive and selective methods for Ir with detection limits of 2 X 10-5 and 8 X 10-5 pg ml-1 were developed by using the oxidation of N-methyldiphenylamine-4-sulfonic acid with KIO4 and KI03 catalysed by Irrv.532 Other organic reagents decolorized by catalytic oxidation with KI04 and used for the determination of precious metals are the following: 3-(4-car- boxylazobenzene)-6-(4-acetylazobenzene)-4,5-dihydroxy- 2,7-naphthyldisulfonic acid for Ru533 and Ir;534 3-(2-car- boxylazobenzene)-6-(4-acetylazobenzene-4,5 -dihy droxy -2,7 - naphthyldisulfonic acid for Ru535 and Ir;536 Rhodamine B (RB) for Ru,537 Ir538 and Rh;539,540 and Malachite Green (MG) for R u .~ ~ ~ Among these methods, the RB-KI04-precious metal ion systems are very sensitive. aiioy.53 1 The indicator reactions used for the catalytic determination of precious metals by decoloration of organic reagents oxidized with other inorganic oxidants are the following: RB-H202 for Pd,542 Pyrogallol Red-H202543 and pyrogallolphthale-in-H202 in the presence of Brij 35544 for Os, phenosafranine-K&Og for Ag,545 Amido Black 1 OB-K2S208 for Ru,s46 carmine-K2S208 in the presence of a,&'-dipyridyl for Ag,547 Bromocresol Green-KBrO3 in the presence of &,a'-dipyridyl for Ru548 and N-methyldiphenylamine-4-sulfonic acid-NH4V03 for Au.549 The following reactions have been used to develop catalytic spectrophotometric methods for the catalysts involved.Silver(1) catalyses the oxidation of K4Fe(CN)6 by dissolved 0 2 forming a blue c01our,55~ the reaction between ammonium tetramethy- lenedithiocarboxylic acid and phenoF and the ligand dis- placement reaction between potassium ferrocyanate and o- ~henanthroline.5~~ Platinum(1v) in ammonia solution catalytically inhibits the decoloration reaction of oxidizing Xylene Blue FF by H202.553 Pd" catalyses the oxidation of C1- by Mne1.554 Silver(1) and AulI1 catalyse the ligand displacement reaction between pentacyanoaminoferrate(I1) and a,&'-di- p ~ r i d y l .~ S ~ A method was proposed for the simultaneous determination of two catalysts through a single catalytic run based on the difference in apparent orders of the indicator reaction catalysed by each catalyst, using the CelV-Aslll reaction catalysed by 0 s and Ru as an example.556 Data processing methods were proposed for the above determinations.556.557 The determina- tion of Rh and Rh-Pt mixtures was based on the acceleratory effect of Rh on the CerV-H202 reaction and the inhibitory effect of Pt on the same reaction.558 Detection by Electrochemical Methods The determination of Ir based on its catalytic effect on the redox reaction between MG and periodate was proposed with oscillopolarographic detection of MG, with a detection limit of 8 ng 1-* using a fixed reaction time of 10 min.559 The other catalytic reactions monitored by oscillopolarography and used for the determination of the respective catalysts are the following: RB-104-560,561 and Methyl Orange-Br03-562 cat- alysed by Ru; Brilliant Green-PO23- by Pd;563 Fast Green- 1 0 4 - by Rh;564 Ce'V-Aslll by Ru565 and Ir;566 As111-C103- by 0 ~ 3 5 6 7 and Safranine T-S208- by Ag in the presence of &,a'- dipyridyl.568 A method was described for the determination of Au"' based on the redox reaction between CeIV and Hgl, and the Cerv unreacted in the catalytic reaction reacts with benzilic acid to form benzophenone, which can be detected by linear scanning voltammetry at a dropping mercury This method has been applied to the determination of Au in ores.The oxidation of C1- with CeIV catalysed by Ag+ was employed for the determination of Ag, using a chloride ion-selective electrode to measure the potential change.570 Flow Injection Flow injection has been applied to the determination of precious metals only relatively recently.There are few reports in this area, but some promising developments have been made by coupling an efficient preconcentration procedure with an adequate on-line detection method. Palladium in ores was determined by a method in which an optical fibre detector was connected to an FI system and with the use of one of three chromogenic reagents, i.e., 1,8-dihy- droxy-2-(4-chloro-2-phosphonophenylazo)-7-(6,8-disulfo- naphthylazo)naphthalene, 1,8-dihydroxy-2-(4-chlor0-2- phosphonophenylazo)-7-(4-sulfonamidophenylazo)-3,6-Analyst, February 1996, Vol. 121 149 disulfonaphthalene and 1,8-dihydroxy-2-(4-chloro-2- phosphonophenylazo)-7-(p-hippuric acid azo)-3,6-disulfonaph- thalene.571 A laboratory-made microwave flow injection ana- lyser and a chromogenic reagent, chlorophosphonazo-mN, were used to establish a flow injection spectrophotometric method for determining Pd in ore and anode mud ~amples.57~ Several other flow injection spectrophotometric methods were developed for the determination of Pd based on the reaction of Pd with DCS- arsenazo,573 DBC-arsenaz~,S~~ Orange G in the presence of hexadecylpyridinium bromide575 and sulfonitrophenol.576 Some of these methods have been applied to the determination of Pd in ores, anode mud and other metallurgical samples.The colour reactions of Pt and Pd with SnC12 were used for the FI of the two elements in mixtures.577 The reaction of Au and Ag with sulfochlorophenolazorho- damine provided the basis for the flow injection spectropho- tometric determination of the two elements.576 The flow injection determination of Ag in waste water determined by dual-beam spectrophotometry using p-dimethylaminobenzal- rhodanine as chromogenic reagent was twice as sensitive as that by conventional FI spectrophotometry.578 Flow injection spectrophotometric methods were developed for the determination of Ru based on its catalysis on the 104- oxidation of the 1, 1 0-phenanthroline-Fell complex579 and on the 104- oxidation of Amido Black 10B.580 A continuous-flow analysis method was described for the determination of 0 s based on its catalytic effect on the reaction between p - phenylenediamine and H202.5*1 A flow injection CL method was proposed for Ag based on its catalysis on the oxidation of lucigenin by H202 in the presence of acetone and NaOH.5g2 A flow injection on-line 8531 fibre column separation and preconcentration system was proposed for the efficient FAAS determination of trace Au in ores and metallurgical samples, with 40-60 injections performed per hour.78 Trace gold in cyanide process solutions was preconcentrated by extraction of the dicyanoaurate anion into a supported liquid membrane in a flow-injection manifold and then washed off with an organic solvent for FAAS determination.583 An F1 method with on-line preconcentration using a minicolumn loaded with dialkyldithio- carbamate immobilized on controlled-pore glass and detected by AAS was described for the determination of Rh"' and some base metak584 A flow injection on-line extraction-FAAS method was developed for the determination Ag in geological samples using diphenylthiourea in IBMK as an extractant.s8s Flow injection systems incorporating microcolumns of Amber- lyst A-26 and sulfhydryl cotton were combined with ICP-AES for the on-line preconcentration and determination of Au in water.s86 A two-channel FIA system with an Ag ion-selective electrode using ammonium nitrilotriacetate solution as a carrier stream was used for the indirect potentiometric determination of Ag.587 Chromatography Palladium, Rh and Ru were separated and determined in ore concentrates as chelates with 2-(2-thiazolylazo)-5-diethyl-m- aminophenol on a Separon Clx column by reversed-phase HPLC, using CTAB as a mobile phase additive to improve the separation.588 2-(6-Methyl-2-benzothiazolylazo)-S-diethylam- inophenol was synthesized and used as a precolumn derivatiz- ing reagent for the determination of Ru, Rh, Pd, Os, Ir and Pt by reversed phase HPLC.s89,590 This method has been applied to the analysis of mineral and anode slime samples.A reversed- phase HPLC method was applied to the simultaneous determi- nation of Ru, 0 s and Pd using 4-(2'-thiazoly- lazo)resacetophenone oxime as a precolumn derivatizing reagent, octadecyl-bonded silica as the stationary phase and methanol-H20 as the mobile pha~e.~91 This method has been applied to the determination of these elements in an anode slime. Among the other azo compounds used as precolumn derivatizing reagents for the separation and determination of precious metals by HPLC are the following: 2-(S-nitro- 2-pyridylazo)-5-dimethylaminobenzoic acid"2 and 1-(2-thiaz- olylazo)-2-naphthol-3,6-disulfonic acids93 for Pd; 7-phenylazo- 8-hydroxyquinoline-5-sulfonic acid for binary mixtures of Ir-Pd, Ru-Pd, Rh-Pd and Pt-Pd;594 2-(6-methyl-2-benzo- thiazolylazo)-5-diethylaminophenol for Rh and IqSY5 and 7-(2-phenylazo)-8-hydroxyquinoline-5-sulfonic acid for Pd, Pt, Ir, Rh and Ru.596 A method for the determination of Pd in platinum concen- trates was developed involving reversed-phase HPLC on a Separon CI8 column with the use of its EDTA complex.597 A reversed-phase HPLC method was described for the separation and determination of Pd", Pt" and some other ions by the use of their complexes with bis-(isova1erylacetone)ethylene- diimine.598 The retention behaviour and recovery of Pd", as chelates with N-alkylisonitrosoacetylacetone imine, was inves- tigated.599-601 Other methods for the determination of Pd include GC and HPLC with the use of their complexes with tetradentate Schiff bases602-6m and ion-pair chromatography utilizing on-line complexation and ion-pair formation in a post- column reactor.605 Gold([) as Au(CN)2- in a sample matrix containing other metal cyano complexes was separated and determined by anion- exchange chromatography using a perchlorate-cyanide-ammo- nia eluent and detected spectrophotometrically at 2 IS nm.6(I6 An ion-pair chromatographic method for the determination of Au and Ag in electroplating cyanide electrolytes was carried out in an ion chromatograph with a conductimetric detector, using a Silasorb CI 8 sorbent and tetrabutylammonium butyrate as the mobile phase.607 HPLC methods were also developed for the determination of Au in electrolytes.608,60Y Capillary zone electrophoresis with on-column UV detection at 214 nm was used to detect and determine Au' and Ag' cyanide complexes in alkaline cyanide solution and has been applied to the determjna- tion of Au and Ag cyanide in ore samples.6'" The chromatographic behaviour of some anionic, uncharged and cationic Pt" complexes was investigated and a method for their separation in one run was developed.61' This method has been applied to the purity determination of cisplatin and to studies the reactions of the Pt" complexes in solutions.An HPLC method was developed for Pt by the use of the polynuclear complexes of Pt" and Pd" with 4,4'-bis(diethy1ami- n0)thiobenzophenone and has been applied to the determination of Pt in the anticancer drug cisplatin and carboplatinum.612 A method for the micro-determination of Ru was developed by using the Weise ring-oven technique, 2-nitroso- 1 -naphthol being used as a chromogenic reagent.613 Electrochemical Methods Electrochemical analysis is one of the important techniques utilized for the determination of precious metals.A variety of electrochemical methods have been established based on the versatile electrochemical properties of these metals and their compounds. The preconcentration of platinum group metal materials in vol tammetric analysis has been reviewed. 6 14 Differential-pulse polarographic methods were developed for the determination of Pt,6ls Pd61"61* and R U .~ ' ~ The measure- ment of Pt was carried out in a supporting electrolyte containing formaldehyde, hydroxylamine and borate with a detection limit of 2 X 10-" mol lk1.615 Pd and Ni in an electroplating bath were determined simultaneously in ammonia-ammonium chlo- ride buffer with the addition of a small amount of EDTA or dimethylglyoxime.618 Differential-pulse or square-wave vol- tammetric methods were used to determine Ag ion at micro-150 Analyst, February 1996, Vol. 121 molar levels in water using a redox-switchable ligand, 1,l'- (1,4,10,13-tetraoxa-7,16-diazacyclooctadecane-7,16-diyldi- methy1)ferrocene .619 The product from the oxidation of alizarin violet by AulI1 exhibits a reduction wave, which was employed for the determination of Au by oscillopolarography.620 A voltammetric method for determining Pd was accomplished by preliminary electrochemical oxidation of Pd" to Pd'" in the presence of azide ion on a Pt electrode and subsequent measurement of the limiting current for reduction of the resulting azido complex of PdIV to Pd" on a rotating disc electrode.621 Polarographic and voltammetric methods were developed for the determination of the following precious metals: IT1" by reduction at a rotating disc electrode;622 Pt and Pd after extraction with molten l-naph- t h ~ l a m i n e ; ~ ~ ~ Pt and Pd after extraction of their complexes with aryloxy-substituted 8-mercaptoquinolines into p-dichloroben- zene melt;624 Pd by the use of its complex with histidine;625 Pd after extraction of its diacetyl oxime thiosemicarbazone com- plex into molten naphthalene and dissolution in DMF;626 Pt in cisplatin and related Pt-derived anticancer drugs;627 Ag in silver plating electrolytes;628 and Rh"' in mixtures with associated metals.629 An adsorptive differential-pulse voltammetric method was developed for the determination of trace Ru based on the adsorptive accumulation of Ru"-salicylaldehyde thiosemicar- bazone on the surface of a hanging mercury drop electrode, followed by reduction of the adsorbed complex during the cathodic ~can.6~0 The adsorptive waves of Au in a medium containing N-8-quinolyl-N'-benzoylthiourea,63~ of Au in a solution containing bismuthiol 11,632 of Pd-di(2-ethylhexyl- )sulfoxide complex633 and of 0s-N-( 8-quinolyl)-N'-benzoylthi- ourea c0mplex63~ were used for the determination of the respective elements.The complexes of Rh with 2-(3,5-dichloro-2-pyridylazo)- 5-dimethylaminoaniline635 and 5-(5-chloro-2-pyridylazo)- 2,4-diaminotoluene,636 of Ru with biologically active ligands637 and of Ir with cysteine638 exhibit a catalytic hydrogen wave. Some of these reactions have been applied to the determination of trace precious metals in al10ys.635~6~6~~3~ The catalytic hydrogen wave of complexes of PGE with nitrogen- and sulfur- containing ligands were examined and methods for the determination of Ru, Rh, 0 s and Pt were developed.639 The catalytic hydrogen wave of the Pt-formazone complex was ultilized for the determination of Pt in human body fluids.640 The adsorptive catalytic wave of an Ag-ammine complex in basic solution was exploited for the determination of Ag.641 By application of various solid electrodes and chemically modified solid electrodes, many stripping voltammetric meth- ods have been established.Because these methods include an accumulation step which is also a good preconcentration and separation step, some of them can be applied to the direct analysis of complex materials. A glassy carbon electrode chemically modified with 8,9,17,18-dibenzo-l,7-dioxa- 10,13,16-triazacyclooctadecane has been used for the preconcentration of Au for its subsequent determination by anodic stripping voltammetry. This method is very sensitive and selective and has been applied to the determination of trace Au in a geological sample.642 A carbon paste electrode modified with flower cells, Datura innoxia, was prepared for the accumulation of trace Au"' under open-circuit conditions, followed by measurement of Au with cathodic stripping voltammetry.643 Numerous other cathodic stripping v o l t a m m e t r i ~ ~ ~ ~ ~ ~ ~ and anodic stripping voltammetric653~654 methods have been deve- loped for the determination of trace gold.In these methods, the electrodes employed for the preconcentration of Au are the following: a carbon paste electr0de;64~ carbon paste electrodes modified with triisooctylamine,645 thioben~anilide,~~7 Rhoda- mine B,648 hexadecyl ester of Rhodamine B,649 tricresyl phosphate650 and polyaniline;652 a hanging mercury drop electrode;651 a carbon fibre electr0de;~53 a glassy carbon electrodeF4 and a glassy carbon electrode modified with trioctylphosphine oxide.646 ' Anodic stripping methods for the determination of Au in rocks, soils and sedirnents655,656 and Pd in retrieval solutions657 with portable voltammeter were developed.A paper has been published on the determination of the content of PGE in natural and industrial materials by stripping ~oltammetry.6~8 Other stripping voltammetric methods for determining Pt,659 0 ~ 6 6 0 Pd6617662 and Ag663-669 were deve- loped. Among the electrodes used for electrodeposition or preconcentration are the following: Au rotating disc;659 a graphite electrode;660 a graphite electrode modified with 8,9,17,18-dibenzo- 1,7-diaza- 10,16-dioxacyclooctadecane crown ether,661 662 and with various macrocyclic thia ligands and Nafion fi1m;667,668 a carbon paste electrode;663 a carbon paste electrode modified with 2,9-dichloro- 1,lO-phenanthrol- a carbon fibre electrode activated by potentiostatic treatment;665 a glassy carbon electrode modified with Nafion 130;666 and polymer modified with hexacyanoferrate(11).669 A method was developed for determining Ag by cyclic voltammetry with a previous step of accumulation by elec- trodeposition at constant potential and then measuring the anodic current peak.670 A patent was issued on the coulometric determination of Pt and Ag by electrochemically reducing PtIV and Ag' sequentially to Pt" and Ago on a Pt electrode and measuring the respective amounts of electricity c0nsumed.~71 For determining Ru in resistive pastes, coulometry with controlled potential was developed based on the single-electron electroreduction of dimeric chloride complexes of Ru'" to an intermediate com- pound, i.e., a dimer containing RuIV and Ru"I.6723673 The coulometric determination of Au is the best choice for the precise analysis of materials with high content of Au.A review entitled 'Progress in the Coulometric Determination of Gold' has been published.674 Au in 98-99.9% gold was determined by coulometric titration with electrogenerated CuI.675 For eliminating the interference of Fell1 in the coulo- metric determination of Au in a semi-finished gold ingot, Au"' and Fell1 were titrated with excess of electrogenerated Cu' and then the excess of Cul and Fe" was back-titrated with electrogenerated AuC14-, the content of Au" being calculated by difference.676 A gold-plated glassy carbon crucible and TMP-4 carbon fibre were used as electrode materials for determining Au"' by controlled-potential coulometry in the presence of Cu", PtIV, Pd" and Ag' at mass ratios of 7 : 1, 1 : 2, 1 : 1 and 1 : 100, re~pectively.~7~ Gold, Pd, Pt and Rh were chemically concentrated on an Hg- graphite electrode and determined by currentless stripping chronopotentiometry using KMn04 as an oxidizing agent.678 Ag was determined by stripping chronopotentiometry after deposition on a glassy carbon disc electrode surface.679 A review has been published on electrochemical sensors for water analysis, including the preparation of an Ag+-sensitive mem- brane.680 Trace Ag was determined by potentiometry with the use of a carbon paste electrode as Ag+ sensor, which exhibits a linear response to the logarithm of Ag concentration from 5 X 10-7 to 1 x 10-2 mol 1-1.681,682 A chemically modified field- effect transistor with plasticized PVC membranes containing macrocyclic thioethers exhibits potentiometric Ag+ selectiv- ity.683 A potentiometric method with an ion-selective electrode based on a liquid anion exchanger was developed for the determination of Ag in thiosulfate solutions.684 Other ion- selective electrodes developed for Ag are the following: poly(viny1 chloride) membrane electrode with sulfur- and/or nitrogen-containing calixarene derivatives as ionoph0res;~85 a carbon paste electrode or wire film electrode with an N,N'- diphenyldithiooxamide-based membraneF HgI2-Ag2S-Analyst, February 1996, Vol.I21 151 AsZS3 chalcogenide glass electrodes;687 chemically deposited and modified chalcogenide glass electrodes;688 and a ceramic electrode of magnesium titanate.689 A gold-selective electrode was developed using a poly(viny1 chloride) liquid membrane modified with a Langmuir-Blodgett film of trialkylammoniumchloroauric acid.690 A patent was issued for the determination of Au"' by using an ion-selective electrode with a liquid membrane containing an electrode- active substance.691 A selective method for the determination of Au was based on its electrodeposition on a gold-plated quartz piezoelectric detector, the concentration of Aul" being proportional to the slope of the frequency versus time plot under selected conditions.692 A patent claims an apparatus and a method for directly indicating the Au content of a metal object by electrochemical reaction.693 Other patents claim an electro- chemical monitor method for Ag and other metal and an electrolytic method for determining Ag in photographic pro- cessing solution.695 The electrochemical behaviour of Pd that could possibly be used in its determination was investigated.696-698 Mass Spectrometry ICP-MS is a powerful tool for the determination of trace amounts of precious metals.The following methods were developed for the determination of precious metals in geo- logical samples by ICP-MS. PGE and Au were determined after the samples had been treated by dry chlorination in the presence of NaCl at 580°C and dissolved in 10% HCl, and the results obtained were comparable to those obtained by fire a ~ s a y .~ ~ 9 , 7 ~ 0 Because these elements are not distributed uniformly in rocks, large samples were chlorinated in a specially designed chlorina- tion tube rotated at a definite rate.70° Low concentrations of PGE and Au in metallic and non-metallic fractions of rocks were determined in solutions obtained by dry chlorination and HF-HN03-HC1 microwave dige~tion.70~ A simple fusion procedure was developed for preconcentrating parts per 1012 levels of PGE and Au, allowing direct analysis of the digested fusion beads by ICP-MS .702 Slurry nebulization introduction of the solid sample, after grinding to < 5 pm and suspending in Na4P207 and aqua regia, was proposed for the determination of PGE and Au with limits of quantification in the range 0.04-0.2 pg g- 1.703 Laser ablation-ICP-MS38 and spark ablation-ICP- MS704 were used for the direct determination of some precious metals in fire assay beads.Nanogram amounts of Pt in environmental dust samples were determined by stable isotope dilution (ID) ICP-MS after its separation by anion e~change.~Os Platinum in natural waters and sediments and Ir in sediments were determined by flow injection ID-ICP-MS after dissolving the sediments by micro- wave digestion and separating Pt and Ir by anion exchange.706 ID-ICP-MS methods were also used to determine ultra-trace amounts of Ag in pure copper707 and pure iron.708 An ID-MS method was developed for the accurate determination of 0 s abundance in molybdenite after acid decomposition of samples by microwave digestion with the addition of K2Cr207 and purification of 0 s by distillation.709 A method was developed to analyse fresh waters for Pd and Pt by ICP-MS, after adsorption of the analytes on activated charcoal, using either nebulization or electrothermal vapori- zation for sample introduction.710 From the experimental result that solutions of hydrochloric acid containing 20-32% of hydrogen chloride allow one to obtain a stable ion current under definite conditions, a spark-source MS method was developed to analyse inorganic solutions for 47 elements including Au, Ag, Pd, Pt, Rh and Ru.711 Laser ablation-ICP-MS was applied to the characterization of the trace element composition fingerprint of selected Au samples from Australia and South Africa for the purpose of identifying the provenance of stolen Au.712 Trace impurities in gold bonding wire used in the semiconductor industry were determined by ID-ICP-MS after removing the matrix Au by electrolysis at controlled potential.713 Resonance ionization MS with laser ablation for sample vaporization was used to determine trace amounts of Au in a copper rnatrix.7l4 Silver and other inorganic ions were determined by capillary electro- phoresis-mass spectrometry using an ionspray-sheath flow interface coupling.715 ICP-MS was applied to the direct determination of Pt in blood plasma716 and Ag in waste oil717 after dilution with suitable solvents.The weakly bound Au in humus samples was determined by ICP-MS after its extraction with 1.2 mol I-' Two strong and narrow resonance excitation lines of autoionizing states of the Ru atom, lying approximately 160 cm-1 above the first ionization limit, were observed by the technique of laser resonance ionization time-of-flight MS.This method was applied to the measurement of Ru concentration in solution samples with a detection limit of 0.05 ng m1-l.719,720 ~ ~ 1 . 7 1 8 Radiochemical Analysis Some radiochemical methods have been developed. However, their applications are restricted because special and expensive equipment is required, especially when NAA methods are used. An NAA method was developed for the determination of Pd, Pt, Ir, Au and Ag at ppb and sub-ppb levels in rocks by thermal neutron irradiation, alkali fusion with carriers, acid dissolution, Te coprecipitation and measuring the y-ray spectra produced.721 PGE and Au were determined in reference materials by NAA through direct irradiation of the NiS button obtained by fire assay with 0.5 g of Ni, the problem of standardizing the button mass being overcome by using a spiking te~hnique.3~ Platinum in environmental and geological samples was determined by NAA via the indicator radionuclidel99 Au, based on the separation of Au on polyurethane f0am.7~~ The use of reactor neutron activation for accurate determination of Ru, Os, Ir, Pt and Au in komatiite rock723 and the NAA determination of Ir in rocks724 were reported with enrichment by NiS fire assay.N extraction combined with NAA was used for the determination of the abundance of Ir and its chemical species in geological samples.725 The NAA determination of Au and Ag in rocks was developed after their extraction with 0,O-diisopropyldithio- phosphate in chloroform.726 Other NAA methods developed to analyse geological samples are the following: 0 s and Ru after their separation by distillation and coprecipitation with CuS;727 Pt and Au after their preconcentration by solid extraction;55 isotopic abundance of 0 s and Ru in meteorites;728 and Pd, Ir, Pt and Au in meteorites after their separation with the use of polyaniline as an anion exchanger.729 The use of epithermal INAA for the determination of Au in geological and ore samples was rep0rted730.73~ and the underestimation of gold (198Au) in ore samples due to the self- shielding of epithermal neutrons was measured and dis- cussed.731 The assay of gold ores by NAA was investigated, and methods for reducing systematic errors caused by resonance self-shielding were de~cribed.7~~ The use of a multi-elemental automated INAA system733 and an americium-beryllium neutron source734 for the determination of Au in ores was reported.A radiometric procedure was described for the determination of trace Ag in galena by precipitating it as insoluble AgI with an excess of 131I-labelled KI, coprecipitating with Zr(HP04)2 and measuring the loss of 1311 activity in solution.735 The use of NAA and chemical proton activation analysis for Ir and Pt in metallic rhodium was described.736,737 INAA and152 Analyst, February 1996, Vol. 121 ICP-AES were applied to the comprehensive analysis of S i c powders for 66 elements including Ir, Ag, Pd, Pt and Ru.738 0 s and Ru in ancient coins were determined simultaneously by NAA with the use of target transformation factor analysis to resolve the mixed y-ray spectra.739 NAA methods were applied to the determination of Au and Ag in natural waters,740 Au in human blood serum741 and human tissues742 and I06Ru in human excreta.743 X-ray Spectrometry PGE in ores were enriched by NiS fire assay, collected on a Millipore filter after removing Ni and reducing them to metal with the use of sodium tetrahydroborate and measured by (XRF) spectroscopy using Au as an internal standard.744 XRF methods were developed for the determination of Au and Ag in non-ferrous metallurgical ores and products after sorbing on a complexing sorbent, VVT,745 and of Pt and Au in geological samples after their preconcentration by solid-phase extrac- t i ~ n .~ ~ Operational experience in the determination of Au746 and progress in the determination of Au and Pt747 in ores using the automated AZTEC XRF analyser were reported. The composi- tion of native Au was determined using an automated electron probe mi~roanalyser.~~~ An X-ray absorption method was developed for the determination of Os, Ir and some other elements in non-ferrous metallurgical materials with a deter- minable concentration range from 0.1 to 100%.T49 An EDXRF method was developed for determining low contents of Au in ores after preconcentration with activated carbon.750 EDXRF analysis using the fundamental parameters approach was used for the multi-elemental analysis of Prestea (Ghana) gold XRF methods were also developed for the determination of Au in ores, rocks and environmental samples after its preconcentration by ion exchange,7S2 by extraction with a solid extractant based on TBP753 and by extraction with melts of mixtures of aliphatic monocarboxylic acids and additives.754 Ag in ores was determined by XRF following its precipitation on a microporous film by reduction with hydrazine hydrate, using Au as a carrier and internal standard.7s5 An X-xay radiometric method was developed for the determination of Ag in complex ores by the use of gamma irradiation stimulating characteristic X-ray radiati0n.75~ For expanding the range of determinable concentrations of Ag in non-metallic powders and increasing the accuracy of X-ray spectrometry, silver in samples was enriched with tin metal by fusing the test piece with flux and Sn.757 XRF analysis was used to analyse gold reference materials for Au, Ag, Cu and Zn using the chemometrics X-ray spectrometry software package V2.1 .75* An XRF method was developed for the measurement of the K value in gold jewellery using scattered source lines as an internal standard and exciting the samples by an annular 241Am source with an activity of 18.5 GBq.759 Analytical parameters were given for the determination of Ag and some other elements in Au-Cu alloys by XRF using an ARL 8420 X-ray spe~trometer.7~~ Palladium in Pd-C catalysts was determined by XRF.761 Patents were issued for the non-destructive determination of the amount and distribution of Pt dispersed on a catalyst support by XRF,762 and for producing standards with a sillenite structure for the determination of Pt by XRF.763 A method for the determination of Ru and Pd was developed by a relatively new technique, delayed X-ray spectrometry after fast neutron activation.764 It was declared that interferences from the sample matrix can be suppressed to an extent that makes the method almost independent of the matrix.PIXE analysis was applied to the analysis of an Au-A1 alloy film for A1,765 of pyrite and arsenopyrite for Au,766 of gold-bearing mineral samples767 and of platinum group minerals .768 0 t her Met hods A review was given of the existence, form and occurrences of gold in 0res.~~9 Phase analysis methods for Au770 and Ag771 in geological materials were developed.Patents were issued for the determination of Au content in ingots772 and the K value in gold j e ~ e l l e r y ~ ~ ~ by density measurement. A method was developed for the determination of Au in cathode copper.774 X-ray photoelectron spectroscopy was used to study the oxidation state of Pd in fired palladium electrode ink materi- als.775 Other instrumental analysis methods that have been developed include the following: electron-spin resonance for the determination of Ag in sol~tions;7~6 low-energy ion scattering spectrometry for the analysis of the surface composi- tion of a Pt loNi90 single crysta1;7'7 Auger electron spectrometry for the analysis of Au-Cu alloys,77* PtSi and Pd2Si;779 two-jet plasmatron for scintillation determination of Au in ores;780 photoacoustic method for the determination of O S ; ~ ~ ~ and resonance ionization spectrometry for the determination of Au and Ru.782 A spot test reagent for Osvlli and Aui1',7*3 a detection agent for organic gold compounds in waste gas784 and a complexing agent for Ag785 were reported.References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Van Loon, J. C., and Barefoot, R. R., Determination of the Precious Metals. Selected Instrumental Methods, Wiley, Chichester, 199 1. Vitezi, A., and Fulep, I., Kohaszat, 1993, 126, 383. Dong, S., Fenxi Shiyanshi, 1992, 11(4), 78. Paukert, T., Geol. Pruzkum, 1992, 34(2), 43. Kontas, E., Tutkimusrap.-Geol. Tutkimuskeskus, 1993, 114, 5. Griffin, H..Kill, B., McNulty, R., and Wilcox, D., Precious Met., 1993, 17th, 557. Lea, D. E., AusIMM Bull., 1993, (3, 74. Yu, W., Guijinshu, 1993, 14(1), 70. Mestek, O., Vlacil, F., and Suchanek, M., Chem. Listy, 1992, 86, 242. Zhao, D., Yankuang Ceshi, 1993, 12, 68. Zhu, M., Fenxi Shiyanshi, 1990, 9(1), 66. Hu, H., Lihua Jianyan, Huaxue Fence, 1992,28,54. Shvetsov, V. A., and Pakhomova, V. V., Zavod. Lab., 1991, 57(7), 16. Pany, S. J., Analyst, 1992, 117, 1347. Hamalainen, L., and Kontas, E., Appl. Geochem., 1992, 7, 287. Chao, T. T., and Sanzolone, R. F. J., Geochem. Explor., 1992, 44, 65. Terashima, S . , Itoh, S., and Ando, A., Geostand. Newsl., 1992,16( l), 9. Hall, G. E. M., Rencz, A. N., and Maclaurin, A. I., J. Geochem. Explor., 1991, 41, 291. Butler, D., Randol Gold Forum, 1991 (Pub.1992), 225. Potts, P. J., Gowing, C. J. B., and Govindaraju, K., Geostand. Newsl., 1992, 16(1), 81. Evans, N. J., and Crocket, J. H., Geostand. Newsl., 1992, 16(1), 71. Terashima, S., Itoh, S., and Ando, A., Chishitsu Chosasho Geppo, 1992, 43(3), 141. Kontas, E., Tutkimusrap.-Geol. Tutkimuskeskus, 1993, 114, 39. Kogan,V., Hinds, M. W., Ocampo, G., and Valente, G., Precious Met., 1993, 17th, 101. Beck, C. M., II., Salit, M. L., Watters, R. L., Jr., Butler, T. A., and Wood, L. J., Anal. Chem., 1993, 65, 2899. Parry, S. J., in Preconcentration Techniques for Trace Elements, ed. Alfassi, Z . B., and Wai, C. M., CRC Press, Boca Raton. FL, 1992, pp. 401-415. Gasparrini, C., Precious Met., 1993, 17th, 581. Shvetsov, V. A., and Pakhomova, V. V., Zavod.Lab., 1991, 57(9), 13. Shvetsov,V. A., Bashkirova, T. V., Chicheva, V. P., and Pak- homova,V. V., Zavod. Lab., 1992, 58(1 l), 9. Shvetsov,V. A., and Bashkirova, T. V., Zavod. Lab., 1992, 58(12), 16. Shvetsov,V. A., Bashkirova, T. V., Chicheva, V. P., and Pak- homova,V. V., Zavod. Lab.,1993, 59( l), 20.Analyst, February 1996, Vol. 121 153 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 Shvetsov,V. A., and Bashkirova, T. V., Zavod. Lab., 1992, 58(8), 5. Sheng, J., Fenxi Shiyanshi, 1992, 11(4), 66. Ching, J., Schmidt, B., and Mojica, A., Precious Met., 1993, 17th, 131. Asif, M., and Parry, S. J., Analyst, 1989, 114, 1057. Asif, M., and Parry, S. J., Min. Pet., 1990, 42, 321.Asif, M., Parry, S. J., and Malik, H., Analyst, 1992, 117, 1351. Javis, K. E., Williams, J. G., Parry, S. J., and Malik, H., in Abstract oj Papers, Third International Symposium Analytical Chemistry in the Exploration, Mining and Processing of Materials, South Africa, August, 1992, Paper No. 206. Parry, S. J., Zh. Anal. Khim., 1994, 49, 65. Paukert, T., and Rubeska, I., Anal. Chim. Acta, 1993, 278, 125. Fu Y. X., Disanjie Quanguo Guijinshu Fenxi Xueshu Huiyi Ji Diwujie Huangjin Fenxi Nianhui Lunwen Zhaiyaoji, November, 1991, p. 26. Shubert, R. H., US Pat., 5238485, 1993. Burmaa, G., Potapov, V. A., Amosova, S. V., and Khlebnikova, A. A., Zh. Anal. Khirn., 1992, 47, 930. Shao, P., Yao, X., Wang, H., Wang, W., Liu, B., Li, M., Luo, L., and Xu, D., Gaodeng Xuexiao Huaxue Xuebao, 1991, 12, 143.Shukla, J. P., Singh, R. K., Sawant, S. R., and Varadarajan, N., Anal. Chim. Acta, 1993, 276, 181. Vasil’eva, A. A., Kalish, N. K., Torgov. V. G., and Tatarchuk, V. V., Russ. Pat., 1341 798, 1991. Li, H., Xu, H., Rong, Q., Lin, S., and Gu, X., Fenxi Shiyanshi, 1993, 12(2), 13. Zhurdvlev, A. A., Ribnov, V. V., and Ganiev, A. G., Uzh. Khim. Zh., 1992, (3-4), 5. Parent, M., Cornelis, R., Dams, R., and Alt, F., Anal. Chim. Acta, 1993, 281, 153. Daoud, J. A., El-Reefy, S. A., and Aly, H. F., J . Radioanal. Nucl. Chem., 1992, 166, 441. Singh, H., and Shinde, V. M., Fresenius’ J . Anal. Chem., 1993,345, 471. Kordyukevich, V. O., Giruts, V. L., and Kuznetsov, V. I., Zh. Anal. Khim., 1993, 48(1 l ) , 50. Chatterjee, A., and Basu, S.J., Radioanal. Nucl. Chern., 1992, 162, 259. Leshchinskij, B. L., Polyaks, A. G., and Vejdemanis, K. Yu., Russ. Pat., 1791 752, 1993. Dmitriev, S., Krasnov, A., Shishkina, T., Chimeg, J., and Drobina T., Anal. Sci., 1991,7 (Suppl., Proc. Int. Congr. Anal. Sci., 1991, Pt. l), 65. Yakshin, V. V., Vilkova, 0. M., and Tsarenko, N. A., Zh. Anal. Khim., 1993, 48, 184. Koshima, H., and Onishi, H., Anal. Sci., 1993, 9, 143. Cai, S., Lihua Jianyan, Huaxue Fence, 1992, 28, 61. Ren,Y., Lei, Z., and Zhang, J., Fenxi Huaxue, 1993, 21, 99. Liu, J.. Chen, Y., and Zhu, B., Fenxi Huaxue, 1992, 20, 982. Ping, X., Yankuang Ceshi, 1991, 10, 325. Zhang, J., Ren, Y., He, H., and Lei, Z., Yingyong Huaxue, 1991.8(5), 57. Wang, J., and Zhao, J., Fenxi Ceshi Tongbao, 1991, 10(4), 56. Tian, S., and Dai, Y., Yejin Fenxi, 1991, 11(2), 25.Chen, J., Zeng, F., and Shen, H., Yankuang Ceshi, 1993, 12, 85. Zhou, Z., Li, H., and Xu, J., Fenxi Shiyanshi, 1993, 12(6), 65. Myasoedova, G. V., Fresenius’ J. Anal. Chem., 1991, 341, 586. Su, Z., Chang, X., Xu, K., Luo, X., and Zhan, G., Anal. Chirn. Acta, 1992, 268, 323. Chang, X., Luo, X., Zhan, G., and Su, Z., Talanta, 1992, 39, 937. Konar, B., Basu, S., and Das, J., Indian J . Chem., Sect. A , 1992,31A, 735. Ivanova, E., Todorova, O., and Stoimenova, M., Fresenius’ J . Anal. Chem., 1992,344, 316. Guo, X., Su, Z., Zhang, J., and Cao, D., Guangpuxue Yu Guungpu Fenxi, 1990, 10(3), 48. Guo, X., Zhang, J., and Su, Z., Fenxi Shiyanshi, 1992, 11(5), 10. Mori, H., Fujimura, Y., Yamaguti, T., Onoda, Y., and Shinba, H., Nippon Kagaku Kaishi, 1993, 1244.Zhang, Y., Lu, F., Xu, Z., and Zhu, Y., Guijinshu, 1991, 12(2), 66. Li, H., Rong, Q., Liu, Y., and Feng, J., Huaxue Shiji, 1991, 13, 368. Wang, Z. Li, Y., Li, Y., and Guo, Y., Anal, Lett., 1994, 27, 957. 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 Qi, W., Wu, X., Zhou, C., Wu, H., and Gao, Y., Anal, Chim. Acta, 1992,270, 205. Su, Z., Cao, D., Xu, K., Zhang, Q., and Qi, L., Yejin Fenxi, 1992, 12(3), 4. Szczepaniak, W., Chern. Anal., 1992,37,257. Yu, Z., Wang, H., Li, C., and Wei, X., Yankuang Ceshi, 1991, 10, 9. Li, Y., Rao, Z., and Zhu, Y., Yankuang Ceshi, 1993, 12, 1. Chen, J., Zhou, Z., and Tang, F., Hangzhou Daxue Xuebao, Ziran Kexueban, 1992, 19(2), 179.Simanova, S. A., Pormnov, G. N., Konovalov, L. V., Molodkina, G. N., and Lysenko, A. A., Zh. Prikl. Khim., 1991, 64, 2358. Zhang, D., Wei, Q., Wang, L., and Xu, T., Lizi Jiaohuan Yu Xifu, 1991, 7, 412. Li, L., An, Y., Jing, L., and Qian, T., Fenxi Shiyanshi, 1991, 10(6), 6. Li, L., Fu, J., An, Y., and Lin. X., Fenxi Shiyanshi, 1992, 11(5), 7. Chen, H., and Zhang, R., Fenxi Ceshi Tongbao, 1991, 10(6), 34. Yu, D., Jing, Y., and Wang, W., Lihua Jiunyan, Huaxue Fence, 1991, 27, 340. Zhang, B., and Zhao, G., Yejin Fenxi, 1991, 11(2), 6. Morozova, L. V., Mikhaleva, A. I., Golentovskaya, I. P., Shaulina, L. P., and Trofimov, B. A., Dokl. Akad. Nauk., 1992, 326, 835. Hudson, M. J., and Leung, B. K. O., Ion Exch. Ad\>., Proc.IEX ’92, 1992, 398. Fujiwara, M., Matsushita, T., Kobayashi, T., Yamashoji, Y., and Tanaka, M., Anal. Chim. Acta, 1993, 274, 293. Jing, Y., and Yu, D., Anal. Sci., 1991, 7 (Suppl., Proc. Int. Congr. Anal. Sci.,1991, Pt. 2), 1297. Yu, D., Yejin Fenxi, 1991, 11(3), 25. Fan, F., Yankuang Ceshi, 1992, 11, 287. Pyrzynska, K., Anal. Chim. Acta, 1991, 255, 169. Cheng, D., Xia, S., and He, Y., Fenxi Huaxue, 1992, 20, 259. Terada, K., and Kawamura, H., Anal. Sci., 1991,7 (Suppl., Proc. Int. Congr. Anal. Sci., 1991, Pt. l), 71. Bara, Y., and Hirakawa, H., Chem. Lett., 1992, 1905. Satake, M., Lee, J. R., Puri, B. K., and Katyal, M., Indian J . Chem., Sect. A , 1992,31A, 67. Tang, F., Mao, X., Wu, G., and Zhang. Y ., Gaodeng Xuexiao Huaxue Xuebao, 1993, 14, 27 1. Chen, J., Zhou, Z., and Tang, F., Yingyong Huaxue, 1992, 9(3), 15.Li, L., Li, S., and Di, P., Fenxi Huaxue, 1992, 20, 1287. Chen, J., Jin, H., Zhou, Z., and Tang, F., Yejin Fenxi, 1992, 12(1), 53. Wu, Y., and Zhou, Z., Fenxi Huaxue, 1992,20, 871. Wu, Y., and Zhou, Z., Fenxi Huaxue, 1993, 21, 56. Runov, V. K., Strepetova, T. V., Pukhovskaya, V. M., Trofimchuk, A. K., and Kuz’min, N. M., Zh. Anal. Khirn., 1993, 48( 1 l), 43. Skopenko, V. V., Trofimchuk, A. K., and Yanovskaya, E. S., Ukr. Khim. Zh. (Russ. Ed.), 1993, 59, 549. Rozowski, J., and Mojaki, M., Chem. Anal. (Warsaw), 1993, 38, 33. Schuster, M., Fresenius’ J. Anal. Chern., 1992, 342, 791. Surakitbanham, Y., 1992, PhD Thesis, available from University Microfilms, Order No. DA9309034; Chern. Abstr., 1994, 120, 2349 19e.Misra, A. K., Asian J. Chem., 1992. 4, 39. Rozylo, J. K., Gwis-Chomicz, D., and Malinowska, I., Chern. Anal. (Warsaw), 1991, 36, 279. Flieger, A., Chem. Anal. (Warsaw), 1991, 36, 31 1. Flieger, A., Chem. Anal. (Warsaw), 1993, 38, 483. Tesic, Z. L., Janjic, T. J., Tosic, R. M., and Celap, M. B., Chromatographia, 1993, 37, 599. Ren, F., and Wu, L., Zhongnan Kuangye Xueyuan Xuebao, 1991,22, 582. Fang, W., Wu, S., and Zhang, J., Lihua Jianyan, Huaxue Fence, 1993, 29,18. Donova, I., and Siftar, J., Acta Chim. Slov., 1993, 40(1), 47. Liu, Y., Lu, X., and Lin, X., Lihua Jianyan, Huaxue Fence, 1993,29, 308. Verma, B. B., Singh, D., and Parmar, M. S., Asian J . Chern., 1992,4, 655. Khadar, A. M. A., and Gajendragad, M. R., Asian J . Chem., 1993,5, 578. Balaknshna, V.V., and Raju, N. A., Acta Cienc. Indica, Chem., 1990, 16C(2), 161.154 Analyst, February 1996, Vol. 121 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 I48 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 Wisniewski, M., Fresenius’ J. Anal. Chern., 1992, 343, 360. Narayana, B., Rao, B. M., Kuchinad, G. T., and Karunakaran, V., Asian J . Chem., 1993, 5, 29. Narayana, B., Joseph, A., Rao, B. M., and Chandra, P., Turk. J. Chem., 1993, 17(1), 13. Rao, B. M., and Narayana, B., Ann. Chirn., 1993, 83, 431. Cai, D., and Xiong, D., Huaxue Shijie, 1991, 32, 410. Hua, D., Dong, L., and Zhao,Y., Yejin Fenxi, 1992, 12(3), 51. Li, Y., Fenxi Shiyanshi, 1993, 12(4), 69. Ermolaev, M. N., Goiodaeva, E.B., and Vinogradova, V. P., Sovrem. Metody Anal. Prom. Mater. i Prirod. Oh’ektov: Mater. Nauch.- Tekhn. Kratkosroch. Sernin., 0-vo ‘Znanie’ RSFSR, Leningr. Dorn Nauch.-Tekhn. Prop., S.-Peterburg, 1992, (Ch. l), 75. Kuznetsov, V. V., and Reshetnikova, V., Russ. Pat., 1704066, 1992. Kiciak, S., Gontarz, H., Goszczynska, H., Koschel, P., and Mehdi, M. A., Chem. Anal. (Warsaw), 1989,34,413. Bakhvalova, I. P., Volkova, G. V., and Dmitrieva, G. A., Russ. Pat., 1 647 399, 199 1. Yun, J., Lihua Jianyan, Huaxue Fence, 1991, 27, 355. Raba, J., Cortinez, V. A., and Sustersic, M. G., An. Asoc. Quirn. Argent., 1993, 81(1), 49. Fedorova, N. G., Avdienko, T. N., Naumenko, E. S., and Supruno- vich, V. I., Vopr. Khirn. Khim. Tekhnol., 1991, 96, 30. Rao, €3. V., Anal.Chirn. Acta. 1992, 258, 167. Robles, L. C., Garcia-Olalla, C., and Aller, A. J., Fresenius’ J. Anal. Chem., 1993,345,441. Mojski, M., Jarosz, M., and Gluch, I., Analusis, 1992, 20, 567. Abramovic, B., Gaal, F., and Horvath, K., Glas. Hem. Tehnol. Maked., 1989, 7, 293. Abramovic, B. F., Horvath, K. S., Gaal, F. F., and Gaal, 0. P., Microchem. J., 1993, 48, 137. Kozlov, V. K., and Simakova, V. K., Russ. Pat. 1 769 127, 1992. Berestetskii, V. I., and Fedoruk, L. I., Tekh. Xino Telev., 1991, (7), 43. Berestetskii, V. I., Zavod. Lab., 1993, 59( 1 I), 6. Parczewski, A., and Madej, K., Chem. Anal. (Warsaw), 1991, 36, 55. Garcia, R, M., Quim. Ind., 1991, 37 565. Kudrev, A. G., and Belousov, V. V., Zh. Anal. Khim., 1994, 49, 195. Koricanac, Z., and Stankovic, B., Arh.Farm., 1991, 41(1-2), 15. Guo, Q., Guijinshu, 1991, 12(3), 54. Guo, Q., Guijinshu, 1993, 14(3), 42. Gevorgyan, A. M., Tsagaraev, E. T., Davidyants, V. G., Novgor- odova, N. J., and Agzamova, D. T., Dokl. Akad. Nauk Resp. Uzb., 1992, (2), 32. Gevorgyan, A, M., Tsagaraev, E. T., Sirlibaev, T. S., Davidyants, V. G., and Novgoradova, N. Yu., Dokl. Akad. Nauk Resp. Uzh., 1992, (3), 31. Sukhoruchkina, A. S., Suprunovich, V. I., Shevchenko, Yu. I., and Sharonina, R. M., Ukr. Khim. Zh. (Russ. Ed.), 1991, 57, 948. Suprunovich, V. I., Velichko, V. V., Shevchenko, Yu. I., Peschan- skaya, I. V., and Shamseeva, N. A., Zh. Anal. Khirn., 1993, 48, 1060. Zaky M., and El-Sheikh, R., Aswan Sci. Technol. Bull., 1991, 12, 91. Zhu, M., and Zhu, X., Yejin Fenxi, 1991, 11(4), 30. Kuz’min, N.M., Kubrakova, I. V., Pukhovskaya, V. M., and Kudinova, T. F., Zh. Anal. Khim., 1994, 49, 199. Sen Gupta, J. G., Talanta, 1993, 40, 791. Xiong, S., Hou, X., Wang, W., and Xu, P., Fenxi Shiyanshi, 1991, 10(3), 32. Kontas, E., Tutkirnusrap.-Geol. Tutkirnuskeskus, 1993, 114, 29. Zereini, F., Urban, H., and Lueschow, H. M., Erzrnetall, 1994,47( 1), 45. Bao C., Zhang, K., and Sun, Q., Fenxi Huaxue, 1994,22, 179. Ojaniemi, E., Tutkimusrap.-Geol. Tutkimuskeskus, 1993, 114, 25. Penttinen, U., Tutkirnusrap.-Geol. Tutkimuskeskus, 1993, 114, 2 1. Kolosova, L. P., Ushinskaya, L. A., and Kopylova, T. N., Zh. Anal. Khirn., 1993,48,644. Davidson, R. A., At. Spectrosc., 1992, 13, 199. Nugteren, H. W., and Morgan, R. K., Chem. Abstr., 1993, 119, 2405 54c. Reddi, G. S., Ganesh, S., Rao, C.R. M., and Ramanan, V., Anal. Chim. Acta, 1992, 260, 13 1. 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 I97 I98 199 200 20 1 202 203 204 205 206 207 208 209 210 21 1 212 213 214 215 216 217 218 219 220 Noras, P., Tutkirnusrap.-Geol. Tutkirnuskeskus, 1993, 114, 17. Tewari R. K., and Gupta, J. R., At. Spectrosc., 1993, 14, 152. Chattopadhyay, P., and Sahoo, B. N., Analyst, 1992, 117, 1481. Wu, R., Lihua Jianyan, Huaxue Fence, 1991,27, 303. Jara Facundo, M. A., Bol. SOC. Quirn. Peru, 1991, 57(4), 226. Yu, Y., Zhuang, X., and Li. S., Guangdong Gongxueyuan Xuebao, Ziran Kexueban, 1991, 8(2), 63. Arpadjan, S., Jordanova, L., and Karadjova, I., Fresenius’ J. Anal. Chern., 1993,347,480. Xu, C., Youkuangye, 1991, 10,51.Burmaa, G., Amar, G., Amosova, S. V., and Potapov, V. A., Zh. Anal. Khim., 1992, 47, 934. Cai, S., Yejin Fenxi, 1992, 12(2), 48. Yuan Y., Yankuang Ceshi, 1993,12, 194. Wang, X., Lihua Jianyan, Huaxue Fence, 1991,27,220. Bao, C., Sun, Q., Zhang, K., and Fan, Q., Fenxi Huaxue, 1993, 21, 1363. Juvonen, R., and Vaananen, P. J., Tutkimusrap.-Geol. Tutki- muskeskus, 1993, 114, 13. Jara Facundo, M. A., Bol. SOC. Quirn. Peru, 1992, 58(4), 240. Terashima, S., Geostand. Newsl., 1991, 15(2), 195. Ma, H., Wu, S., Chen, Y., Chen, C., and Hu, H., Yankuang Ceshi, 1993, 12, 28. Lu, Z., Lu, Y., Tong, J., and Hong, L., Lihua Jianyan, Huaxue Fence, 1993, 29, 291. Cai, S., Yejin Fenxi, 1992, 12(6), 55. Rigas, J., Palagyi, S., Holeczyova, G., and Matherny, M., Chem.Pap., 1991, 45, 509. Zhou, Z., Li, H., and Xu, J., Yejin Fenxi, 1993, 13( l), 20. Dolaberidze, L. D., Todradze, G. A., Shanidze, M. K., and Chikvaidze, G. G., Zavod. Lab., 1993, 59(8), 27. Tian, S., Lihua Jianyan, Huaxue Fence, 1991,27, 249. Yuan, J., and Yuan, S., Lihua Jianyan, Huaxue Fence, 1992, 28, 57. Zhang, X., Li, B., Zhou, D., Yu, G., and Zhang, X., Guangpuxue Yu Guangpu Fenxi, 1992,12(5), 83. Yang, Y., Yankuang Ceshi, 1992,11, 310. Chen, Y., Yankuang Ceshi, 1991, 10,291. Feng, Z., Li, Y., and Xie, F., Fenxi Ceshi Xuebao, 1993, 12(2), 57. Lee, G. H., and Chung, K. S., Anal. Sci., 1991, 7(Suppl., Proc. Int. Congr. Anal. Sci., 1991, Pt. 2), 1339. Belova, V. V., and Vasil’eva, A. A., Zavod. Lab., 1991, 57(9), 2. Curra-Lava, R. A., and Jimenez-Prieto, J.A., At. Spectrosc., 1991,12, 210. Shao, W., and Zhang, M., Yankuang Ceshi, 1993,12,57. Brooks, D. S., and Flatt, J. R., Anal. Chim. Acta, 1992, 264, 107. Zhou, J., Ding, Y . , Yu, L., and Lu, J., Youkuangye, 1991, 10(1), 56. Lu, G., Xu, J., Xu, T., Jin, L., and Fang, Y., Talanta, 1992, 39, 51. Jin, L., Xu, J., Qian, J., and Tong, W., Fenxi Ceshi Tongbao, 1992, 11(2), 56. Hiraide, M., Zhou, S. H., and Kawaguchi, H., Analyst, 1993, 118, 1441. Quigley, M. N., and Vernon, F., Anal. Proc., 1992, 29(1), 1. Oreshkin, V. N., Okeanologiya, 1993, 32, 954. Aoki, H., Murakami, H., and Chikuma, M., Bunseki Kagaku, 1993, 42( 1 l), T147. Sung, S., Song, X., and Kang, W., Youkuangye, 1993,12,204. El’nazarov, S., Kompleksoobraz. v Rastvorakh. Tadzh. Gos. Un-t Dushanbe, 1991, 189.Oreshkin, V. N., Malofeeva, G, I., and Vnukovskaya, G. L., Okeanologiya, 1993,33, 784. Hinds, M. W., Spectrochim. Acta, Part B, 1993, 48, 435. Hinds, M. W., Littau, S., and Moulinie, P., Analyst, 1992, 117, 1473. Zhang, H., and Liu, F., Lihua Jianyan, Huaxue Fence, 1992, 28, 280. Ariizumi, N., and Kamijo, M., Yarnanashi-ken Kogyo Gijutsu Senta Kenkyu Hokoku, 1993,7,94. Zelentsova, L. V., Shchipachev, V. A., and Levchenko, L. M., Zh. Anal. Khim., 1992, 47, 1889. Kalpana, G., and Koshy, V. J., Indian J. Chem., Sect. A, 1992, 31A, 39. Mazzucotelli, A., Vivaldi, B., De Paz, F., and Magi, E., Ann. Chim., 1992, 82, 107.Analyst, February 1996, Vol. 121 155 ~~ 22 1 222 223 224 225 226 227 228 229 230 23 1 232 233 234 235 236 237 238 239 240 24 1 242 243 244 245 246 247 248 249 250 25 1 252 253 254 255 256 257 25 8 259 260 26 1 262 263 264 265 Beinrohr, E., Lee, M.L., Tschoepel, P., and Toelg, G., Fresenius’ J. Anal. Chem., 1993,346, 689. Guo, J., and Chen, Y., Guangpuxue Yu Guangpu Fenxi, 1991,11(2), 48. Bader, R., Reichelmayr-Lais, A. M., and Kirchgessner, M., Agribiol. Res., 1991, 44, 212. Tamura, H., Arai, T., Nagase, M., and Ichinose, N., Bunseki Kagaku, 1992, 41(2), T13. Arzamastsev, A. P., Aliakbarova, S. A., and Listov, S. A., Farmatsiya, 1992, 41(4), 31. Listov, S. A., Farmatsiya, 1991, 40(6), 38. Gaweda, E., Pr. Cent. Inst. Ochr. Pr., 1991, 41(148-149), 39. Nakonechnyi, I. I., Kushnir, S. V., and Deineka, S. E., Gig. Sanit., 1992, (3), 79. Littlejohn, D., Egila, J. N., Gosland, R. M., Kunwar, U. K., Smith, C., and Shan, X., Anal.Chim. Acta, 1991, 250, 71. Das, B. C., AusIMM Proc., 1992, 297(1), 25. Indyk, L. I., Mel’nik, V. V., and Knutova, 0. V., Zavod. Lab., 1992, 58(9), 26. Shaburova, V. P., Yudelevich, I. G., and Kolotii, 0. Yu., Sib. Khim. Zh., 1991, (3), 58. Wu, X., Cui, G., and Peng, Y., Fenxi Huaxue, 1992, 20, 743. Frame, E. M. S., King, J. A., Jr., Anderson, D. A., and Balz, W. E., Appl. Spectrosc., 1993, 47, 1276. Elci, L., Anal. Lett., 1993, 26, 1025. Matunaga, H., Tohoku Kogyo Gijutsu Shikensho Hokoku, 1993, 26, 21. Mateeva, N., Arpadzhan, S., Deligeorgiev, T., and Miteva, M., Analyst, 1992, 117, 1599. Aihara, M., Watanabe, H., and Kiboku, M., Anal. Sci., 1991, 7 (Suppl., Proc. Int. Congr. Anal. Sci., 1991, Pt. l), 87. Mihara, Y., and Yokota, K., Annu.Rep. Tohoku Coll. Pharm., 1990, 37, 247. Pyrzynska, K., Talanta, 1994, 41, 381. Zeng, X., You, Q., and Zhong, Y., Guangpuxue Yu Guangpu Fenxi, 1991, 11(4), 39. Mostafa, M. A., Analusis, 1992, 20, 283. Abdallah, A. M., El-Shahawi, M., El-Defrawy, M. M., and Askar, E. E., Chem. Anal. (Warsaw), 1992, 37, 589. El-Defrawy, M. M., Analusis, 1991, 19, 292. Frigge, C., and Jackwerth, E., Spectrochim. Acta, Part B , 1992, 47, 787. El-Defrawy, M. M., Abdallah, A., and El-Shamy, M., Asian J . Chem., 1993, 5, 1047. Kauppinen, M., and Smolander, K., Anal. Chim. Acta, 1994, 285, 45. Rayson, G. D., and Fresquez, M. R., Can. J. Appl. Spectrosc., 1993, 38(4), 114. Huang, G., Qian, S., Liu, C., and Mou, J., Fenxi Huaxue, 1993, 21, 935. Gao, Y., Wu, X., Chi, K., and Wu, H., Fenxi Shiyanshi, 1993,12( l), 77.Duan, Y., Zhang, H., Lu, H., and Jin, Q., Fenxi Huaxue, 1992, 20, 383. Duan, Y., Li, X., Hou, M., and Jin, Q.. Gaodeng Xuexiao Huaxue Xuebao, 1993,14, 1204. Carlos Robles, L., Garcia-Olalla, C . , and Javier Aller, A., J. Anal. At. Spectrom., 1993, 8, 1013. Chekalin, N. V., and Vlasov, I. I., J. Anal. At. Spectrom., 1992, 7, 225. Zheng, Y., and Su, X., Can. J. Appl. Spectrosc., 1993, 38(4), 109. Fagioli, F., Locatelli, C., and Torsi, G., Ann. Chim., 1992, 82, 283. Aller, A. J., Anal. Chim. Acta, 1993, 284, 361. Gilmutdinov, A. Kh., Zakharov, Yu. A., Ivanov, V. P., and Voloshin, A. V., Zh. Anal. Khim., 1993, 48, 28. Rohrer, C., and Wegscheider, W., Spectrochim. Acta, Part B , 1993, 48, 315. Mit’kin, V. N., Zemskov, S. V., Zaksas, B. I., Petrova, E.A., and Kurskii, A. N., Zh. Anal. Khim., 1991, 46, 2416. Lin, Y., Shen, Z., and Hu, J., Yankuang Ceshi, 1991, 10, 247. Sun, M., Jain, J., Zhou, M., and Kemch, R., Can. J. Appl. Spectrosc., 1993, 38(4), 103. Xie, S., and Zhao, M., Guijinshu, 1992, 13(3), 53. Hofmans, D., and Adrisenssens, E., Precious Met., 1993, 17th, 235. Wissmann, F., and Nordheim, U., Precious Met., 1993, 17th, 117. 266 267 268 269 270 27 1 272 273 274 275 276 277 278 279 280 28 1 282 283 284 285 286 287 288 289 290 29 1 292 293 294 295 296 297 298 299 300 30 1 302 303 304 305 306 307 308 309 310 31 1 3 12 313 Basi, M. S., Blumberg, P., and Kyriacou, M., Precious Met., 1993, 17th, 549. Roberts, G. J., Precious Met., 1992, 16th, 9. Teng, S., Lihua Jianyan, Huaxue Fence, 1992, 28, 370.Zheng, M., Fenxi Shiyanshi, 1993, 12(2), 79. Xu, P., J. Anal. At. Spectrom., 1992, 7, 775. Xu, L., and Schramel, P., Fresenius’ J. Anal. Chem., 1992, 342, 179. Kulikova, A. B., and Kulikov, A. A., Soversh. Metodov Anan. Veshchestva i Obrab. Dannykh pri Geokhim. Issled. v Zabaikal’e AN SSSR. SO. Buryat. Geol. In-t, Ulan-Ude, 1991, 79. Zhang, X., Li, H., and Yang, Y., Guangpuxue Yu Guangpu Fenxi, 1991, 11(5), 38. Zhang, X., Li, H., and Yang, Y., Fenxi Huaxue, 1991,19, 1199. Zhang, X., Li, H., and Yang, Y., Fenxi Shiyanshi, 1991, 10(3), 70. Provodenko, L. B., Kulikov, Yu. V., and Andreev, V. M., Zavod. Lab., 1991, 57(8), 36. Liu, J., Qiu, H., and Li, L., Fenxi Huaxue, 1993, 21, 1188. Feng, Z., and Xu, G., Guijinshu, 1991, 12(4), 53. Venkatasubramanian, R., Biswas, S.S., and Murty, P. S., Indian J . Technol., 1991, 29, 605. Hara, H., Jpn. Pat., 05 157 696, 1993. Chen, J., Yejin Fenxi, 1992, 12(6), 54. Gao, T., Yu, C., and Xian, C., Guijinshu, 1993, 14(3), 57. Brill, M., and Wiedemann, K. H., Gold Bull. Gold Pat. Dig., 1992, 25(1), 13. Liu, Y., Lu, X., and Lin, X., Fenxi Ceshi Tongbao, 1991, 10(4), 53. Qiu, H., Liu, J., and Li, L., Guijinshu, 1993, 14(3), 53. Zhang, X., Li, H., and Yang, Y., Lihua Jianyan, Huaxue Fence, 1993, 29. 281. Zhang, X., Li, H., and Yang, Y., Guangpuxue Yu Guangpu Fenxi, 1993, 13(1), 107. Zhu, L., Liu, H., Li, Z., Hu, Z., and Zhang, J., Yejin Fenxi, 1991, 11(4), 26. Liu, S., and Zhu, T., Guangpuxue Yu Guangpu Fenxi, 1992, 12(5), 71. Watanabe, H., Tadani, I., and Aihara, M., Hil-oshima-kenritsu Seibu Kogyo Gijutsu Senta Kenkyu Hokoku, 1991,34, 18.Watanabe, H., Tadani, I., Aihara, M., and Kiboku, M., Chem. Express, 1993, 8(2), 77. Guan, J., Gao, B., Wang, W., and Liu, Y., Yuanzineng Kexue Jishu, 1991, 25(6), 1. Kim, S. T., Min, K. J., Lee, Y. H., Park, J. A., and Choi, B. S., .I. Korean Chem. SOC., 1992, 36, 243. Thiel, R. D., Chem. Abstr., 1993, 118, 13884211. Alexandrov, S., Arpadjan, S., and Matova, M., God. Sofi. Univ. ‘Sv. Kliment Okhridski’, Khim. Fak., 1992, 80, 134. Spectro Hispania, S. L., Tec. Lab., 1991, 13(164), 319. El’nazarov, S., Rustamov, S., Khamidov, B. O., and Egorova, L. A., Kompleksoobraz. v Rastvorakh. Tadzh. Gos. Un-t Dushanbe, 1991, 186. Talalaev, B. M., Saradzhev, L. V., and Lomonosova, 0. S., Pr-vo Azot. Udobrenii, M., 1991, 114. Subacheva, N.D., and Nikitenko, N. A., Spektr. Anal. Tr. Mosk. Kollok., 1990, AN USSR. Otd-nie Obshch. Fiz. i Astron. Nauch. Sov. PO Spektroskopii, M., 1991, 2, 116. Akhoond, M., Ensafi, A. A., Massoumi, A., and Safavi, A., Microchem. J., 1992,45, 365. Simeonsson, J. B., Ng, K. C., and Winefordner, J. D., Appl. Spectrosc., 1991, 45, 1456. Wu, R., Lihua Jianyan, Huaxue Fence, 1992,28, 121. Zabeen, R., Rezaei, B., and Purohit, D. N., Rev. Anal. Chem., 1991, 10, 115. Xiong, C., Zhu, T., and Zhang, J., Yejin Fenxi, 1993, 13(3), 30. Gu, X., Lu, Q., and Rong, Q., Lihua Jianyan, Huaxue Fence, 1992, 28, 229. Te, A., and Deng, Y., Yejin Fenxi, 1991, 11(2), 47. Ding, Y., Yejin Fenxi, 1991, 11(2), 60. Ren, Y., Lei, Z., and Zhang, J., Yejin Fenxi, 1992, 12(5), 25. Gong, Z., Song, J., and Luo, J., Yankuang Ceshi, 1993, 12, 189.Song, L., Chin. Pat., 1 074 536, 1993. Liu, X., Li, L., and Gao, P., Fenxi Shiyanshi, 1992, 11(4), 38. Lei, Z., Ren, Y., and Wang, Z., Fenxi Shiyanshi, 1993, 12(6), 26. Li, J., and Zuo, B., Hebei Daxue Xuebao, Ziran Kexueban, 1993, 13(2), 28.156 Analyst, February 1996, Vol. 121 3 14 3 15 316 317 318 319 3 20 32 1 322 323 3 24 325 326 327 328 329 330 33 1 332 333 334 335 336 337 338 339 340 34 1 342 343 344 345 346 347 348 349 350 35 1 352 353 354 355 356 357 358 359 360 361 362 363 Zhang, D., and Li, Q., Lihua Jianyan, Huaxue Fence, 1993, 29, 288. Xue, G., Yejin Fenxi, 1991, 11(2), 19. Xue, G., Fenxi Shiyanshi, 1993, 12(5), 41. Xue, G., Yankuang Ceshi, 1991, 10,310. Li, Z., Fenxi Huaxue, 1993, 21, 224. Kundu, D., and Roy, K., Talanta, 1992, 39,415.Li, H., and Yi, L., Fenxi Huaxue, 1992, 20, 774. Li, H., and Yi, L., Yankuang Ceshi, 1993,12,43. Yeaw, D. C., and Hoffmann, A. F., Eur. Pat. Appl., 536059, 1993. Bhargava, C., and Tandon, V. K., Asian J. Chem., 1992, 4, 664. Murthy, G. V. R., and Reddy, T. S., Asian J. Chem., 1993, 5, 1133. Gyepes, E., Galusova, M., Beno, A., Tothova, E., and Grolmusova, A., Acta Fuc. Rerum Nat. Univ. Comenianae, Chim., 1990, 37, 53. Murthy, G. V. R., and Reddy, T. S., Talanta, 1992, 39, 697. Uesugi, K., Nagahiro, T., Kumagai, T., and Nishioka, H., Chim. Acra Turc., 1991, 19, 245. Uesugi, K., Nagahiro, T., Kumagai, T., and Nishioka, H., Orient. J. Chem., 1991, 7(4), 176. Chaudhury, S. P., and Shome, S. C., J. Indian Chem. SOL.., 1991,68, 430.Palaniappan, R., Chim. Acta Turc., 1990, 18, 483. Ma, D., Wang, Y., and Wang, J., Anal. Lett., 1991, 24, 2307. Jimenez, A. M., Herrador, A., and Aspero, A. G., Afinidad, 1993, 50(445), 160. Zhang, J., Mou, W., and Cheng, J., Huaxue Shiji, 1992, 14, 279. Pilipenko, A. T., D’yachenko, N. A., and Falendysh, N. F., Ukr. Khim. Zh. (Russ. Ed.), 1991, 57, 639. Xang, H., and Yin, A., Fenxi Shiyanshi, 1991, 10(6), 26. Yin, A., and Xiang, H., Lihua Jianyan, Hua,uue Fence, 1992, 28, 83. Wang, Y., Zhou, G., Wang, H., Wang, D., and Chen, S., Fenxi Huaxue, 1993,21, 1317. Mao, X., Tang, F., and Jin, Q., Fenxi Huaxue, 1992, 20, 951. Zhu, Y., and Yu, C., Fenxi Huaxue, 1994,22, 210. Han, Q., Yang, X., and Wu, Q., Fenxi Shiyanshi, 1992, 11(5), 34. Zhang, H., Liu, Q., and Yu, J., Huaxue Shiji, 1993, 15, 135.Hu, S., Yi, R., and Zhang, R., Tianranqi Huagong, 1992, 17(6), 43. Abu Zuhri, A. Z., and Rady, A. H., J. Chern. Technol. Biotechnol., 1992, 54( l), 39. Zhang, G., and Liu, X., Lihua Jianyan, Huaxue Fence, 1991, 27, 201. Liu, X., Yang, H., and Zhang, G., Fenxi Shiyanshi, 1992, 11(1), 37. Rong, Q., and Li, D., Guijinshu, 1993, 14(2), 55. D’yachenko, N. A., Falendysh, N. F., and Sukhan, V. V., IzI>. Vyssh. Uchehn. Zaved., Khim. Khim. Tekhnol., 1992,35(7), 30. Mori, I., Fujita, Y., Toyoda, M., and Hasegawa, Y., Talanta, 1992, 39, 535. Chai, H., and Zhang, X., Fuda Xuehao, Ziran Kexuehan, 1992,31( I), 115. Jarosz, M., Oszwaldowski, S., and Marczenko, Z., Chem. Anal. (Warsaw), 1992,37, 335. Liu, H., Liu, B., Wu, B., and Wu, C., Lihua Jianyan, Huaxue Fence, 1992, 28, 13 1.Liu, B., Liu, H., and Wu, C., Huaxue Shiji, 1992, 14, 261. Liu, B., Liu, H., and Wu, C., Talanta, 1991,38, 1 143. Pilipenko, A. T., D’yachenko, N. A., Falendysh, N. F., and Karetnikova, E. A., Russ. Pat., 1 663 539, 1991. Zhang, F., Wu, B., Liu, H., and Wu, C., Microchem. J., 1993, 48, 104. Zhang, F., Wu, B., Liu, H., and Wu, C., Fenxi Shiyanshi, 1993,12( l), 70. Zhu, Y., Liu, J., Liu, B., and Zhou, X., Huaxue Shtji, 1993, 15, 274. Zhang, G., Wang, Y., and Au, B., Guijinshu, 1991, 12(3), 40. Zhang, F., Wu, B., Liu, H., Wu, C., and Li, Z., LihuaJianyan, Huaxue Fence, 1993, 29, 79. Zhang, G., Wang, Y., Zhang, Z., and Hu, Q., Lihua Jianyan, Huaxue Fence, 1992, 28, 264. Sukhan, V. V., D’yachenko, N. A., and Falendysh, N. F., R i m .Pat., 1797023, 1993. Zhang, X., and Zhang, G., Fenxi Huaxue, 1994, 22, 61. Yang, H., Zhang, G., Mu, Q., and Huang, H., Femi Huaxue, 1992,20, 725. 364 365 366 367 368 369 370 37 1 372 373 3 74 375 376 377 378 379 380 38 1 382 383 384 385 386 3 87 388 389 390 39 1 392 393 394 395 3 96 397 398 399 400 40 1 402 403 404 405 406 407 408 409 410 41 1 412 Yang, H., Zhang, G., Mu, Q., and Zhou, C., Huaxue Shiji, 1992,14, 328. Zhang, G., Zhang, X., Hu, Y., Wang, X., and Quan, C., Yejin Fenxi, 1991, 11(5), 21. Zhao, J., and Xu, Q., Fenxi Huaxue, 1992,20,619. Zhao, J., and Xu, Q., Guijinshu, 1991, 12(3), 45. Shi, G., and Yang, G., Huaxue Shiji, 1992, 14, 85. Yang, G., Yang, X., Ren, C., and Yang, J., Fenxi Shiyanshi, 1991, 10(6), 3. Yang, G., Yu, Z., and Yu, R., Fenxi Shiyanshi, 1992, 11(5), 1.Lu, W., Xu, Q., and Zeng, Z., Fenxi Huaxue, 1993,21, 1065. Zeng, Z., and Xu, Q., Talunta, 1992,39, 409. Zeng, Z., Huang, L., and Xu, Q., Lihua Jianyan, Huaxue Fence, 1992, 28, 70. Zeng, Z., Zhao, Y., and Xu, Q., Yankuang Ceshi, 1992, 11, 207, Zhao, J., Zeng, Z., and Xu, Q., Fenxi Shiyanshi, 1991, 10(5), 17. Zhu, L., and Wu, R., Guijinshu, 1991, 12(2), 57. Badawy, S. S., Omar, M. M., and Soliman, A. A., Egypt. 1. Chem., 1989 (Pub. 1991), 32,419. Zhu, L., Guijinshu, 1993, 14(3), 47. D’yachenko, N. A., Falendysh, N. F., and Sukhan, V. V., Izv. Vyssh. Uchehn. Zaved., Khim. Khim. Tekhnol., 1992, 35(2), 44. El-Ansary, A. L., Abdel Fattah, H. M., and Ghamrawy, S. S., Asian J . Chem., 1993, 5, 1057. Chernovyants, M. S., Askalepova, 0. I., Anisimova, V.A., and Bagdasarov, K. N., Ukr. Khim. Zh. (Russ. Ed.), 1992, 58, 257. Wang, D., Hu, W., Zhu, Z., Huang, Z., and He, S., Fenxi Huusue, 1993, 21, 375. Wang, D., Wu, Q., Hu, W., Zhuang, L., Huang, Z., and He, S., Yejin Fenni, 1991, 11(5), 18. Tang, F., Wang, L., Mao, X., and Lin, M., Yejin Fenxi, 1992, 12(4), 12. Li, H., and Yao, Y., Lihuu Jianyan, Huaxue Fence, 1991, 27, 301. Chen, L., and Li, Y., Xiangfan Daxue Ziran Kexue Xuehao, 1991, 13(1), 101. Li, Y., and Zheng, Y., Chem. Res. Chin. Univ., 1991, 7(2), 91. Cheng, S., and Zhang, Y., Fenxi Huaxue, 1992,20,7. Huang, G., Wang, H., and Xia, H., Yankuang Ceshi, 1993, 12, 198. Tong, S., and Li, N., Huaxue Shiji, 1993, 15, 139. Vorob’eva, N. E., Pavlov, P. T., and Zhivopistsev, V. P., Zh. Anal. Khim., 1991, 46, 1088.Vorob’eva, N. E., Pavlov, P. T., and Zhivopistsev, V. P., Izv. Vyssh. Uchehn Zaved., Khim. Khim. Tekhnol., 1991,34(3), 17. Vorob’eva, N. E., Pavlov, P. T., Zhivopistsev, V. P., Vergizova, T. V., and Ermolina, G. I., Zavod. Lab., 1991, 57( 1 l), 13. Islam, M. A., and Stephen, W. I., Talanta, 1992, 39, 1429. Varhelyi, C., Makkay, F., Benko, A., and Kazinczy, E., Chem. Anal., 1992, 37, 219. Moustafa, M. E., Mabrouk, E. M., Dessouki, H. A., and Amine, A. S., Microchem. J., 1991, 44, 311. Xe, X., Guijinshu, 1991, 12(2), 68. Chen, W., Lihua Jianyan, Huaxue Fence, 1991, 27, 161. Xu, S., and Bai, F., Fenxi Shiyanshi, 1991, 10(2), 18. Abalakina, V. M., Kleimenova, 0. K., and Dedkov, Yu. M., Zavod. Lab., 1992, 58(10), 1. Liao, Y., Feng, M., He, Q., Wang, W., and Wang, T., Fenxi Huaxue, 1993, 21.70. Li, Y., Rao, Z., and Zhu, Y., Fenli Huaxue, 1992, 20, 1333. Li, S., Li, S., and Zhang, Y., Fenxi Shiyanshi, 1992, 11(2), 35. Bai, J., Cui, X., and Yao, H., Taiyuari Gongye Dame Xuebao, 1993, 24( l), 57. Fujita, Y., Mori, I., Toyoda, M., and Matsuo, T., Anal. Sci., 1993, 9, 829. Manko, R. W., C‘hem. Anal. (Warsaw), 1991, 36, 763. Bazel, Ya. R., Kish, P. P., Studenyak, Ya. I., and Zimomrya, I. I., RUSS. Pat., 1668940, 1991. Bazel, Ya. R., Studenyak, Ya. I., and Kish, P. P., Izv. Vyssh. Uchehn. Zaved.. Khim. Khim. Tekhnol., 1992, 35(7), 25. Dadfamia, S., and Shamsipour, M., Iran. J. Chem. Chem. Eng., 1992, 11(1), 14. Balog, J. S., Mushkalo, I. L., and Zimomrya, I. I., Russ. Pat., 1728741, 1992. Mestek, O., Rudy, 1991, 39(9-lo), 160.Balog, 1. S., Kish, P. P., Rigan, M. Yu., and Zimomrya, 1. I., Russ. Pat., 1647400, 1991.Analyst, February 1996, Vol. 121 157 413 414 415 416 417 418 419 420 42 1 422 423 424 425 426 427 428 429 430 43 1 432 433 434 435 436 437 438 439 440 44 1 442 443 444 445 446 447 448 449 450 45 1 452 453 454 455 456 457 458 459 460 Lu, Y., Wu, Q.. Tang, L., and Lu, L., Shanxi Daxue Xuehao, Ziran Kexuehan, 1992,15, 152. Lazarev, A. I., and Gerko, V. V., Zavod. Lab., 1993, 59(7), 7. Lu, Y., Chen, Z., Zhou, H., Sun, W., and Ning, B., Fenxi Shiyanshi, 1993, 12(3), 48. Zhao, M., and Hu, S., Fenxi Huaxue, 1992, 20, 1147. Morales, A., Toral, M. I., Richter, P., and Silva, M.,Anal. Left., 1992, 25, 1765. Balcerzak, M., Analusis, 1992, 20, 295. Wu, S., and Fang, W., Fenxi Huaxue, 1993, 21, 148.Li, Z., Guijinshu, 1991, 12(4), 45. Li, Z., Fenxi Shiyanshi, 1992, 11(2), 29. Balcerzak, M., Chem. Anal. (Warsaw), 1992, 37, 683. Zhang, S., Chen, X., and Wang, Z., Huaxue Shiji, 1992, 14, 239. Balcerzak, M., Pr. Nauk.-Politech. Warsz., Chem., 1991, 55, 3. Li, B., Shiyou Huagong, 1991, 20, 558. Deb, M. K., and Mishra R. K., Int. J . Environ. Anal. Chem., 1992, 48(3-4), 151. Yaroshenko, 0. P., Gavrilova, V. N., and Sumskaya, N. R., Zavod. Lab., 1991, 57(9), 4. Zhu, Y., Ye, X., and Chen, H., Hangzhu Daxue Xuehao, Ziran Kexueban, 1992, 19( l), 71. Chen. W., Jiang, X., Pan, J., and Xu, Z., Lihua Jianyan, Huaxue Fence, 1991,27, 267. Qi, W., Zhu, L.. and Luo, H., Microchim. Acta, 1992, 106(1-2), 27. Qi, W., and Yan, X., Fenxi Ceshi Tonghao, 1992, 11(1), 18.Zhu. Y., and Yang, L., Fenxi Huaxue, 1991, 19, 778. Zhu. Y., and Yang, L., Anal. Lett., 1993, 26, 309. Zhu, Y., Cao, H., and Yu, C., Fenxi Huaxue, 1992, 20, 15. Kumar, A., Shukla, R. K., Agarwal, R. S., Satake, M., and Nagahiro, T., Chim. Acta Turc., 1990, 18, 477. Swami, V. K., and Mathur, S. P., Indian J. Chem. Sci., 1991, 5 , 69. Shetty, A. N., and Gadag, R. V., Bull. Chem. Soc. Jpn., 1993, 66, 2536. Panwar, 0. S., Sharma, A,, and Mathur, S. P., Chem. Environ. Res., 1992, 1(2), 99. Mathur, P., Mathur, S, P., and Gupta, B. D., Asian J . Chem., 1993.5, 777. Sharma, R. K., Bull. Chem. Soc. Jpn., 1993, 66, 1084. Parkhomenko, E. P., Falendysh, N. F., and Pilipenko, A. T., Zavod. Lab., 1991, 57(9), 9. Bakurdzhieva, V., and Ivanov, N., God. Vissh. Khim-Tekhnol.Inst., Sofiya, 1990. 30(2), 184. Parkhomenko, E. P., Falendysh, N. F., and Pilipenko, A. T., Za\)od. Lah., 1991, 57(12), 7. Kannan, R., Basu, B. B., Sreepathy, S., and Rajagopalan, S. R., Precious Met., 1993, 17th, 95. Yi, D., and Huang, B., Kuangye Gongcheng, 1991, 11(2), 66. Grzedzicki, K., Bujewski, A., and Dobrowolski, J., Zesz. Nauk. Politech. Slask., Chem., 1989, 1035(121), 55. Topalov, A. S., Kosanic, M. M., and Balac, A. S., Zb. Rad. Prir.-Mat. Fak., Unit’. Nov. Sadu, Ser. Hem., 1988 (Pub. 1990), 18, 9. Gudzenko, L. V., and Gaiduk, 0. V., Zavod. Lab., 1993, 59(1 I ) , 4. Balcerzak, M., and Kus, S., Anal. Sci., 1994, 10, 65. Dressier, V. L., Kielling, J. A., and Martins, A. F., Ecletica Quim., 1992, 17, 17. Gupta, U., Kuinar, A., Aradbana, and Rao, A.L. J., Chim. Acta Turc., 1991 (Pub. 1992), 19,63. Gupta, U., Rao, A. L. J., and Jeet, V., Asian J . Chem., 1993, 5, 296. Gupta, U., and Rao, A. L. J., Asian J . Chem., 1992, 4, 257. Malik, A. K., Gupta, U., and Rao, A. L. J., J. Indian Chem. SOC., 1991, 68, 587. Chhakkar, A. K., and Kakkar, L. R., Fresenius’ J . Anal. Chem., 1993, 347, 483. Sarwar, M., and Maqsood, A., Sci. Int., 1992, 4, 271. Swami, V. K., Arora, S. K., and Mathur, S. P., J . Inst. Chem., 1992, 64(3), 128. Chopra, S., Malik, A. K., and Rao, A. L. J., J. Indian Chem. Soc., 199 1, 68, 637. Ramana, V. V., and Saraswathi, K., J. Indian Chem. Soc., 1993, 70, 274. Saraswathi. K., Lakshmi, A. V. S., and Ramana, V. V., J . Indian Chem. So(.., 1992, 69, 795. 46 1 462 463 464 465 466 467 468 469 470 47 I 472 473 474 475 476 477 478 479 480 48 1 482 483 484 485 486 487 488 489 490 49 1 492 493 494 495 496 497 498 499 500 501 502 503 504 505 Ahmadzadeh, H., Safavi, A., and Massoumi, A., J .Sci. Islamic Repub. Iran, 1990, 1, 266. Chakrabarti, S. K., Chakrabarti, A. K., and Bag, S. P., Proc. Indian Natl. Sci. Acad.. Part A , 1991, 57, 459. Chavan, J. D., and Joshi, V. P., Asian J. Chem., 1992, 4, 361. Chavan, J. D., and Joshi, V. P., J. Indian Chem. Soc., 1991, 68, 309. Chowdhury, A., Das, H. R., and Shome, S. C., J. Indian Chem. Soc., 1992, 69,796. Petrov, B. I., Chukin, V. M., and Yakovleva, T. P., Russ. Pat., 1734008, 1992. Paria, P. K., Sarkar, A., and Majumdar, S. K., Indian J . Chem., Sect. A , 1993,32A, 647. Malik, A. K., Rashmi, and Rao, A. L. J., Proc. Natl.Acad. Sci., India, Sect. A , 1992, 62, 125. Atanasova, D., and Shishkov, A., Nauchni Tr.-Plovdivski Univ., Khim., 1989, 27(5), 171. Moily, S., and Khader, A. M. A., Asian J. Chem., 1993, 5, 1 140. Abdel Razik, A., Int. J . Chem., 1992, 3(2), 35. Sun, D., and Wang, L., Guijinshu, 1993, 14(1), 51. Israilov, M. A., and Golushko, R. N., Izv. Akad. Nauk Tadzh. SSR, Otd. Fiz.-Mat., Khim. Geol. Nauk, 1989, (3), 83. Wang, L.. and Sun, P., Guijinshu, 1991, 12(3), 50. Swami, V. K., and Mathur, S. P., Asian J. Chem., 1992, 4, 393. Ai, P., and Xu, Q., Guijinshu, 1991, 12(2), 60. Abd El Hamied Hafez, M.. and Kenawy, I. M. M., Bull. Soc. Chim. Fr., 1991, 837. Tanwari. Z., and Khuhawar, M. Y., J. Chem. Soc. Pak., 1993, 15, 237. Chiugureanu D. Gh., Khetrush, P. M., and Savva, P.V., Bul. Acad. Stiinte Repub. Mold., Stiinte Biol. Chim., 1993, (4), 62. Dakshinamoorthy, A., Singh, R. K., and Iyer, R. H.., J . Radioanal. Nucl. Chem., 1994, 177, 327. Saran, N. R., Umashanker, V., and Ramanaiah, G. V., Bull. Chem. Soc Jpn., 1992, 65, 2291. Chhakkar, A. K., and Kakker, L. R., Ann. Chinz., 1992, 82(3-4), 203. Zhu, Y., and Yang, L., Lihua Jianyan, Huaxue Fence, 1991, 27, 223. Wang, L.. and Sun, P., Yankuang Ceshi, 1992, 11, 326. Chhakkar, A. K., and Kakker, L. R., Anal. Sci., 1992, 8, 243. Mabajan, H. A., and Patil, M. R., Chem. Anal. (Warsaw), 1992, 37, 239. Chaudhari, R. A., and Sawant, A. D., J. Indian Chem. Soc., 1991,68, 635. Sinha, S. H., and Sawant, A. D., Bull. Chem. Sou. Jpn., 1992, 65, 1622. Mahajan, H. A., and Dhadke, P. M., Proc.Natl. Acad. Sci., India, Sect. A , 1991, 61, 565. Bochkova, L. P., Borisova, L. V., and Korchemnaya, E. K., Zh. Anal. Khim., 1992, 47, 809. Telia, N. M., Mandzhgaladze, 0. V., and Chichua, D. G., Izv. Akad. Nauk Gruz., Ser. Khim., 1991, 17(2), 98. Ajt, A. O., Alfimov, M. V., Galashin, A. E., Zajtsev, N. K., and Kopneva, Y. B., Russ. Pat., 1 774 25 1, 1992. Zimenkovskij, B. S., Muzychenko, V. A., Mynka, A. F., Muraveva, A. A., and Muratova, L. A., Russ. Pat., I755 135, 1992. Zhao, L., and Hu, S., Shiyou Huagong, 1992, 21, 695. Zhao, L., and Hu, S., Fenxi Huaxue, 1992, 20, 584. Geng, Z., Guijinshu, 1992, 13(1), 66. Imdadullah, Fujiwara, T., and Kumamaru, T., Anal. Sci., 1991, 7 (Suppl., Proc. Int. Congr. Anal. Sci., 1991, Pt. 2), 1399. Imdadullah, Fujiwara, T., and Kumamaru, T., Anal.Chenz., 1993,65, 421. Lu, J., Zhang, X., Zhu, L., Zhou, S., and Zhang, Z., Felzvi Huaxue, 1993,21, 1334. Lu, J., Zhang, X., Fan, W., and Zhang, Z., Fenxi Huaxue, 1992, 20, 104. Lu, J., Zhang, X., Fan, W., and Zhang, Z . , Anal. Chim. Acta, 1992, 262, 225. Zhang, X., Lu, J., Feng, M., and Zhang, Z., Fenxi Huaxue, 1993,21, 575. Zaporozhets, 0. A., Pogasii, L. B., and Sukhan, V. V., Zh. Anal. Khim., 1993, 48( 1 l), 75. Zhang, F., Huang, S., and Lin, J., Fenxi Huaxue, 1991, 19, 809. Lin, J., and Zhang, F., Chin. J. Chem., 1991, 9, 499.158 Analyst, February 1996, Vol. 121 506 507 508 5 09 5 10 51 1 512 513 5 14 5 15 516 517 518 519 520 521 522 523 524 525 526 5 27 528 529 530 53 1 532 533 534 535 536 537 538 539 540 54 1 542 543 544 545 546 547 548 549 550 551 An, J., and Lin, J., Chem.Res. Chin. Univ., 1991, 7(1), 32. Uchikura, K., Kuromatogurafi, 1993, 14( I), 7. Losev, V. N., Runov, V. K., Strepetova, T. V., and Trofimchuk, A. K., Zh. Anal. Khim., 1993, 48, 1915. Losev, V. N., and Runov, V. K., Zavod. Lab., 1991, 57(9), 6. Losev, V. N., and Runov, V. K., Sovrem. Metody Anal. Kontrolya na Prom. Predpriyatiyakh: Muter. Semin. 0-vo ‘Znanie’ RSFSR, Mosk. Dom Nauch.-tekhn. Prop., M., 1991, 67. Losev, V. N., and Runov, V. K., Izv. Vyssh. Uchebn. Zaved., Khim. Khim. Tekhnol., 1992,35(8), 31. Jiang, W., Zhu, G., Yang, J., and Gao, F., Lihua Jianyan, Huaxue Fence, 1992,28,90. Berman, H. L., Derbalian, G. P., Jaw, H. R. C., Lee, J. J., and Beigler, M. A., PCT Int. Appl., WO 93/07 472, 1993. Golovanova, N.V., Runov, V. K., Sadvakasova, S. K., and Khvostova, V. P., Zh. Anal. Khim., 1991, 46, 2336. Tikhomirova, T. I., Fadeeva, V. I., and Kudryavtsev, G. V., Anal. Chim. Acta, 1992, 257, 109. Chae. M. Y., and Czamik, A. W., J. Am. Chem. SOC., 1992, 114, 9704. Yang, C., Yang, Z., and Zhang, C., Fenxi Huaxue, 1993,21, 1272. Song, Y., Gong, X., Liu, X., and Ma, Z., Xibei Shifan Duxue Xuebao, Ziran Kexuehan, 1991, 27(2), 39. Lu, J., Wei, S., and Wang, H., Yejin Fenxi, 1992, 12(5), 36. Tian, L., and Zhong, H., Nanjing Daxue Xuehao, Ziran Kexue, 1991, 27, 396. Huang, G., and Wang, Z., Yankuang Ceshi, 1993, 12, 21. Fujiwara, M., Matsushita, T., and Wakita, H., Chem. Express, 1992, 7(1), 17. Yamamoto, D., Tsukada, M., Nagasawa, T., and Takeuchi, H., Meiji Daigaku Nogakubu Kenkyu Hokoku, 1992,95, 1.Geokchyan, N. O., Ovsepyan, E. N., and Khachatryan, A. G., Ukr. Khim. Zh. (Russ. Ed.), 1991, 57, 405. Kachin, S. V., Poddubnykh, L. P.. and Runov, V. K., Zavod. Lab., 1991, 57(9), 1. Wang, A., Liu, M., Zhang, Y., and Cao, L., Guangpuxue Yu Guangpu Fenxi, 1992, 12(6), 109. Zhang, Y., Liu, M., Wang, A., and Cao, L., Guangpuxue Yu Guangpu Fenxi, 1993, 13(1), 17. Fedotova, I. A., Tikhonova, L. P., Ustinova, N. V., and Solov’eva, G. N., Zh. Anal. Khim., 1993, 48, 19. Zhu, H., Peng, A., and Wu, Z., Fenxi Shiyanshi, 1992, 11(4), 31. Khomutova, E. G., Rysev, A. P., Fedorina, L. I., Solomonov, V. A., and Ershova, I. V., Zh. Anal. Khirn., 1993, 48, 1653. Khomutova, E. G., Romanovskaya, L. E., Zhitenko, L. P., and Rysev, A. P., Zavod. Lab., 1992, %(lo), 3.Mushtakova, S. P., Gumenyuk, A. P., Khmelev, S. S., and Batasheva, N. A,, Zh. Anal. Khim., 1993, 48, 1218. Liu, X., Kou, Z., Chen, X., and Hu, Z., Fenxi Huaxue, 1992, 20, 345. Kou, Z., and Liu, X., Yejin Fenxi, 1992, 12(6), 48. Kou, Z., Liu, X., Chen, X., and Hu, Z., Fenxi Shiyanshi, 1992, ll(4) 48. Liu, X., Kou, Z., Chen, X., and Hu, Z., Guijinshu, 1992, 13(1) 51. Jiang, Z., Huaxue Tongbao, 1992, (l), 38. Jiang, Z., and Dai, G., Guijinshu, 1992, 13(1), 54. Zhang, Z., and Zhan, H., Chin. Sci. Bull., 1992,37, 615. Zhang, Z., Zhan, H., and Guo, H., Fenxi Huaxue, 1994,22, 193. Jiang, Z., Tang, Q., and Wu, D., Guijinshu, 1992, 13(2), 51. Li, F., and Chen, S., Lihua Jianyan, Huaxue Fence, 1991, 27, 286. Ensafi, A. A., and Rezaei, B., Anal. Lett., 1993, 26, 1771.Mori, I., Fujita, Y., Toyoda, M., and Takeuchi, Y., Fresenius‘ J. Anal. Chem., 1992,344,292. Prasad, K. M. M. K., and Subhashchandran, K. P., Asian J. Chem., 1992, 4, 715. Chen, X., Gong, H., Jiang, H., and Hu, Z., Fenxi Huaxue, 1993, 21, 701. Wu, H., Zheng, Z., Wang, X., and Zhang, X., Guijinshu, 1991,12(4), 58. Yang, W., Zhou, X., Zhang, G., Yu, G., He, X., and Shi, H., Yankuang Ceshi, 1993, 12, 89. Mushtakova, S. P., Khmelev, S. S., Kozhina, L. F., and Bylinkina, N. N., Zh. Anal. Khim., 1991, 46, 2411. Zhou, J., Lihua Jianyan, Huaxue Fence, 1992,28, 227. Gao, H., Huagong Yejin, 1993, 14(2), 168. 552 553 554 555 556 557 558 559 560 56 I 562 563 564 565 566 567 568 5 69 570 57 I 572 573 574 575 576 577 578 579 580 58 1 582 583 584 585 586 587 588 589 590 59 1 592 593 594 595 596 597 598 ~~ ~~ Guo, G., Zhou, X., and Mei, G., Lihua Jianyan, Huaxue Fence, 1993, 29, 35.Chen, G., Luo, G., and Zhang, F., Fenxi Huaxue, 1993,21, 1081. Tikhonova, L. P., Goncharik, V. P., Kozhara, L. I., and Svarkovskaya, 1. P., Zh. Anal. Khim., 1991, 46, 2009. Mehra, M. C., Francoeur, B., and Katyal, M., Orient. J. Chem., 1992, 8(2), 81. Gu, Z., Zhu, Z., Chen, R., Lu, B., and Han, C., Analyst, 1993, 118, 1055. Zhu, Z., Gu, Z., and Wang, X., Talanta, 1993, 40, 1013. Milovanovic, G. A., and Bilandzic, M.,J. Serb. Chem. Soc., 1991,56, 671. Jiang, Z., Talanta, 1992, 39, 1317. Jiang, Z., and Wang, L., Anal. Lett., 1993, 26, 1967. Jiang, Z., Liang, L., and Tang, Q., Guijinshu, 1992, 13(4), 48. Jiang, Z., Li, T., Wu, Q., Li, C., and Dai, G., Huaxue Tongbao, 1993, (71, 3 1.Jiang, Z., Wang, L., Li, X., and Tang, Q., Guijinshu, 1993, 14(2), 45. Wang, L., and Jiang, Z., Chem. Res. Chin. Univ., 1993, 9(1), 85. Jiang, Z., Electroanalysis, 1991, 3, 823. Jiang, Z., Wang, L., Liang, L., and Deng, Y., Guijinshu, 1992,13(4), 42. Qin, C., Li, X., and Jiang, Z., Chin. Chem. Lett., 1991, 2, 641. Jiang, Z., Guijinshu, 1991, 12(2), 34. Jiang, Z., Talanta, 1993, 40, 1823. Zhu, H., Peng, A., and Yang, X., Fenxi Huaxue, 1992,20, 742. Xu, Y., Chen, X., Liu, M., and Hu, Z., Anal. Chim. Acta, 1992,262, 97. Xu, Y., and Hu, Z., Anal. Lett., 1994, 27, 793. He, X., Cai, Y., and Hu, Z., Anal. Chim. Acta, 1992, 256, 113. He, X., Cai, Y., Gong, H., and Hu, Z., Fenxi Shiyanshi, 1990, 9(1), 18. Ma, L., Chen, X., and Hu, Z., Fenxi Huaxue, 1992,20, 339.Shpigun, L. K., and Gur’eva, R. F., Zh. Anal. Khim., 1991, 46, 2187. Koch, K. R., and Auer, D., Talanta, 1993, 40, 1975. Zhao, Z., Ma, C., Chen, H., and Sun, H., Fenxi Shiyanshi, 1992, 11(3), 15. Almuaibed, A. M., and Townshend, A., Microchem. J., 1993, 48, 210. Chen, X., Hu, Z., and Zheng, H., Bull. SOC. Chim. Belg., 1992, 101, 989. Rodionova, T. V., Beklemishev, M. K., Chigvintseva, E. V., and Zolotov, Yu. A., Vestn. Mosk. Univ., Ser. 2: Khim., 1992, 33, 472. Liu, Y . , and Yu, Z., Fenxi Huaxue, 1992, 20, 474. Taylor, M. J. C., Bames, D. E., and Marshall, G. D., Anal. Chim. Acta, 1992, 265, 71. Townshend, A., and Habib, K. A. J., Microchem. J., 1992, 45, 210. Tang, Z., Jin, Z., Lu, Q., and Tian, F., Yankuang Ceshi, 1993, 12, 14. Gomez, M. M., and Mcleod, C.W., J . Anal. At. Spectrom., 1993, 8, 461. Shpigun, L. K., and Bazanova, 0. V., Zh. Anal. Khim., 1992, 47, 1588. Shapovalova, E. N., Mishenina, 1. V., Basova, E. M., Bol’shova, T. A., and Shpigun, 0. A., Zh. Anal. Khim., 1991, 46, 1503. Liu, Q., Wang, Y., Liu, J., and Cheng, J., Fenxi Huaxue, 1992, 20, 1088. Liu, Q., Wang, Y., Liu, J., and Cheng, J., Anal. Sci., 1993, 9, 523. Ramesh, A., Talanta, 1994, 41, 355. Liu, Q., Zhang, H., and Cheng, J., Fresenius’ J. Anal. Chem., 1992, 344, 356. Basova, E. M., Bol’shova, T. A., and Shpigun, 0. A., Zh. Anal. Khim., 1991,46, 2362. Basova, E. M., Kazakova, 0. V., Bondareva, L. G., Polenova, T. V., and Ivanov, V. M., Zh. Anal. Khim., 1993, 48, 1531. Liu, Q., Zhang, H., and Cheng, J., Chem. Res. Chin. Univ., 1993,9( I), 18.Bondareva, L. G., Basova, E. M., and Ivanov, V. M., Vestn. Mosk. Univ., Ser. 2: Khim., 1992, 33(2), 176. Basova, E. M., Bondareva, L. G., and Ivanov, V. M., Zh. Anal. Khim., 1992,47, 1712. Khuhawar, M. Y., and Soomro, A. I., J. Liq. Chromatogr., 1992,15, 647.Analyst, February 1996, Vol. 121 159 599 600 60 1 602 603 604 605 606 607 608 609 610 61 1 612 613 614 615 616 617 618 619 620 62 1 622 623 624 625 626 627 62 8 629 630 63 1 632 633 634 635 636 637 638 639 640 Kim, Y. J., Kewon, J. H., and Lee, W., Anal. Sci. Technol., 1992, 5, 63. Kim, I. W., Shin, H. C., Lee, M. H., Yoon, T. K., Kang, C. H., and Lee, W., Anal. Sci. Technol., 1992, 5 , 389. Kim, I. W., Kang, S. H., Lim, J. H., Park, K. S., and Lee, W., Punsok Kwahak, 1989, 2(1), 61. Khuhawar, M.Y., and Soomro, A. I., Anal. Chim. Actu. 1992, 268, 49. Khuhawar, M. Y., and Bozdar, R. B., J . Chem. SOC. Pak., 1991,13, 197. Khuhawar, M. Y., and Soomro, A. I., J . Chem. Soc. Pak., 1992, 14, 206. Siren, H., and Riekkola, M. L., J . Chromatogr., 1992, 590, 263. Otu, E. O., Robinson, C. W., and Byerley, J. J., Analyst, 1992, 117, 1145. Padarauskas, A., Kazlauskas, R., and Petrukhin, 0. M., Zavod. Lab., 1991, 57(11), 16. Powers, M., Hoppe, F., and Gannotti, J., Plat. Suif Finish., 1991, 78( 1 l), 76. Stoelting, J., Maurer, K., and Bacher, W., Metalloberflache, 1994, 48(3), 172. Aguilar, M., Farran, A., and Martinez, M., J . Chromatogr., 1993,635, 127. Macka, M., and Borak, J., J . Chromatogr., 1993, 641, 101. Zhang, X., Lin, C., Zhen, L., and Qu, J., Fenxi Huaxue, 1993, 21, 1365.Hanif, M., Ijaz, I., Hussain, S., and Bhatti, M. S., Sci. Znt., 1993, 5, 41. Ulakhovich, N. A., Budnikov, G. K., and Medyantseva, E. P., Zh. Anal. Khim., 1992, 47, 1546. Czea, M. Z., Lee, S. W., Chung, K. S., and Lee, D. W., J . Korean Chem. Soc., 1992,36,616. Sawangikar, N., and Joshi, A. P.. Indian J . Chem., Sect. A, 1994,33A, 77. Tarapcik, P., and Vanickova, M., Chem. Pap., 1992, 46, 293. Basu, B. J., and Rajagopalan, S. R., Analyst, 1992, 117, 1623. Rojas, M. T., Medina, J. C., Gokel, G. W., and Kaifer, A. E., Supramol. Chem., 1993, 2( 1), 5. Wang, Y., and Zhou, S., Yejin Fenxi, 1991, 11(6), 19. Shirokova, V. I., and Kabanova, 0. L., Russ. Pat., 1 616 333, 1992. Bessarabova, I. M., and Ross, E. V., Izv. Akad. Nauk Kaz. SSR, Ser. Khim., 199 1, (Z), 36.Mashkina, S. V., Gareev, R. F., Ulakhovich, N. A., and Budnikov, G. K., Zh. Anal. Khim., 1993, 48, 1486. Mashkina, S. V., Ulakhovich, N. A., Budnlkov, G. K., Sturis, A. P., Asaks, J., and Janson, G., Zh. Anal. Khim., 1991, 46, 770. Wang, C., Fu, X., Wang, Z., and Li, H., Fenxi Hua,uue, 1994, 22, 77. Agarwal, R. S., Asolkar, A., Kumar, A., Sharma, A., and Saxene, G. C., Bull. Electrochem., 1991, 7, 436. Wenclawiak, B., and Briese M. A., Fresenius’ J . Anal. Chem., 1992, 343, 117. Kiryushov, V. N., Shustova, L. I., and Kletenik, Yu. B., Gal’vano- tekhn. i Obrab. Poverkhnosti, 1992, l(1-2), 75. Obr, L., and Popl, M., Sh. Vys. Sk. Chem.-Tecknol. Praze H: Anal. Chem., 1991, H24,29. Palaniappan, R., and Kumar, T. A., Analyst, 1993, 118, 293. Zheng, Q., and Wang, Z., Fenxi Shiyanshi, 1992, 11(2), 6.Zhang, Q., Xu, X., and Fan, Q., Fenxi Huaxue, 1993,21,744. Liu, X., Lu, W., Zheng, P., and Zhu, H., Fenxi Huaxue, 1993, 21, 488. Zheng, Q., Wang, Z., Lu, W., and Huang, K., Yejin Fenxi, 1992, 12(5), 12. Zhang, Z., Zhang, G., and Wu, J., Fenxi Huaxue, 1993,21, 557. Zhang, Z., Zhang, G., and Feng, Y., Huaxue Xuebao, 1992, 50, 11 17. Medyantseva, E. P., Budnikov, G. K., Balakaeva, T. A,, and Zakharova, I. A., Zh. Anal. Khim., 1991, 46, 1573. Wang, C., Xin, W., Fu, X., Zhang, Y., and Yang, Y., Fenxi Huaxue, 1992, 20, 368. Medyantseva, E. P., Budnikov, G. K., and Romanova, 0. N., Zh. Anal. Khim., 1991, 46, 2004. Messerschmidt, J., Alt, F., Tolg, G., Angerer, J., and Shaller, K. H., Fresenius’ J . Anal. Chem., 1992, 343, 391.64 1 642 643 644 645 646 647 648 649 650 65 1 652 653 654 655 656 657 658 659 660 66 1 662 663 664 665 666 667 668 669 670 67 1 672 673 674 675 676 677 678 679 680 68 1 682 683 684 685 Zhang, Y., Lihua Jianyan, Huaxue Fence, 1991,27, 135. Turyan, I., and Mandler, D., Anal. Chem., 1993, 65, 2089. Wang, J., Tian, B., and Rayson, G. D., Talanta, 1992, 39, 1637. Peng, T., Wang, G., and Zhu, P., Yingyong Huaxue, 1992, 9(2), 76. Peng, T., Li, H., and Wang, S., Analyst, 1993, 118, 1321. Lu, G., Zhu, M., Jin, L., and Fang, Y., Gaodeng Xuexiao Huaxue Xuebao, 1991, 12,735. Cai, X., Kalcher, K., Neuhold, C., Diewald, W., and Magee, R. J., Analyst, 1993, 118, 53. Koelbl, G., Kalcher, K., and Voulgaropoulos, A., Fresenius’ J . Anal. Chem., 1992,342, 83. Talaber, C . , Kalcher, K., Kobl, G., Raber, G., and Cai, X., Sci.Int., 1992, 4, 347. Vytras, K., Svancara, I., Renger, F., Srey, M., Vankova, R., and Hvizdalova, M., Collect. Czech. Chem. Commun., 1993, 58, 2039. Abu Zuhri, A. Z., El-Shahawi, M. S., and Kamal, M. M., Anal. Chim. Acta, 1993, 282, 133. Li. G., Fang, H., and Chen, H., Gaodeng Xuexiao Huaxue Xuebao, 1993,14, 1522. Zhang, Y., and Chen, G., Fenxi Huaxue, 1993,21,959. Zou, J., and Ma, X., Nanjing Huagong XueyuanXuebao, 1991,13(4), 64. Hall, G. E. M., and Vaive, J. E., Chem. Geol., 1992, 102(14), 41. Yang, Y., Zhu, G., and Gao, H., Fenxi Huaxue, 1991,19, 805. Yang, Y., Zhu, G., and Gao, H., Fenxi Huaxue, 1992,20, 1174. Kolpakova, N. A., Kropotkina, S. V., and Shabalkina, T. G., Sovrem. Metody Anal. Prom. Mater. i Prirod.Ob’ektov: Mater. Nauch.- Tekhn. Kratkosroch. Semin., 0-vo ‘Znanie’ RSFSR, Leningr. Dom Nauch.-Tekhn. Prop., S.-Peterburg, 1992, (Ch. l), 86. Sharipova, N. S., Karimov, A. N., and Maulyasheva, E. M., Rum Pat., 1765761, 1992. Kolpakova, N. A., Kropotkina, S. V., and Sukhomlinova, L. V., Russ. Pat., 1746285, 1992. Brainina, Kh. Z., Tsymbal, M. V., and Temerdashev, Z. A., Zavod. Lab., 1992, 58(3), 11. Brainina, Kh. Z., Turyan, I..Ya., Tsymbal, M. V., Temerdashev, Z. A., and Trizin, Y. G., Russ. Pat., 1 746 286, 1992. Nikitin, A. A,, Shcherbakova, L. P., and Zhulanova, A. G., Izv. Akad. Nauk Kaz. SSR, Ser. Khim., 1991, (l), 38. Huang, S., Li, B., Lin, H., Liao, S., Zhang, Z., and Yu, R., Fenxi Huaxue, 1993,21, 1423. Peng, T., Lu, R., Li, H., and Zhu, P., Fenxi Huaxue, 1991, 19, 1336.Wu, Q., Yu, X., Liu, Y., Zhang, S., Ying, S., and Liu, K., Lihua Jianyan, Huaxue Fence, 1992,28,233. Labuda, J., Vanickova, M., Pavlishchuk, V. V., and Holdt, H. J., Zh. Anal. Khim., 1992, 47, 1297. Labuda, J., and Vanickova, M., Elecrroanalysis (N.Y.), 1993, 5, 141. Iorenzo, E., and Abruna, H. D. J., Electroanal. Chem., 1992, Rivas, C., and Ortiz, P. I., Anal. Lett., 1994, 27, 751. Demkin, A. M., and Kabanova, 0. L., Russ. Pat., 1658065, 1991. Afanas’eva, M. V., and Ezerskaya, N. A., Zavod. Lab., 1991,57(9), 8. Ezerskaya, N. A., and Afanas’eva, M. V., Russ. Pat., 1673947, 1991. Yang, B., Yejin Fenxi, 1991, 11(5), 33. Dong, S., Fenxi Shiyanshi, 1991, 10(5), 19. Dong, S., Guijinshu, 1992, 13(1), 45. Markova, 1. V., Zh. Anal. Khim., 1993, 48, 5 13.Toropova, V. F., Polyakov, Yu. N., and Krut, S. Kh., Izv. Vyssh. Uchebn. Zaved., Khim. Khim. Tekhnol., 1993, 36(2), 111. Suturovic, Z. J., Marjanovic, N. J., and Jankovits, I. F., J . Serb. Chem. SOC., 1992, 57, 187. Jaffrezic-Renault, N., RGE, Rev. Gen. Electr., 1992, (3), 56. Pei, J., Yin, Q., and Zhong, J., Talanta, 1991, 38, 1185. Pei, J., Zhong, J., and Yin, Q., Gaodeng Xuexiao Huaxue Xuebao, 1991,12, 1173. Brzozka, Z., Cobben, P. L. H. M., Reinhoudt, D. N., Edema, J. J. H., Buter, J., and Kellogg, R. M., Anal. Chim. Acta, 1993, 273, 139. Kuchkarev, E. A., Urusov, Yu. I., Bashir, E., and Petrukhin, 0. M., Russ. Pat., 171 1065, 1992. O’Connor, K. M., Svehla, G., Harris, S. J., and McKervey, M. A., Talanta, 1992, 39, 1549. 328( 1-2), 11 1.160 Analyst, February 1996, Vol.121 ~ 686 687 688 689 690 69 1 692 693 694 695 696 697 698 699 700 70 1 702 703 704 705 706 707 708 709 710 71 1 712 713 714 715 716 717 718 719 720 72 1 722 723 724 725 726 Shvedene, N. V., Sheina, N. M., Barabanskaya, E. A., and Kochulin, N. V., Vestn. Mosk. Univ., Ser. 2: Khim., 1993, 34, 582. Liang, Z., and Cheng, J., Huadong Huagong Xueyuan Xuebao, 1991, 17, 516. Legin, A. V., Bychkov, E. A., and Vlasov, Yu. G., Sens. Actuators B, Park, M. Y., Cho, D. H., Lee, J. H.. and Cho, Y. J., Anal. Sci. Technol., 1991, 4, 41. Wu, J., Guo, D., Zhang, Y., and Wang, X., Fenxi Huaxue, 1993,21, 1131. Byr’ko, V. M., Agasyan, P. K., and Nemilova, M. Yu., Russ. Pat., 1672339, 1991. Sanchez-Pedreno, C., Ortuno, J. A., and Martinez, D., Anal. Chim. Acta, 1992, 263, 143.Moment, N. J., and Nelson, 0. L., US Pat., 5080766, 1992. Vasu, K. I., Chandra, N., Subramanian, G. G., Birlasekaran, S., and Vincent, S., Indian Pat., 170767, 1992. Ninomya, Y., and Oda, S., Jpn. Pat., 04276541, 1992. Adekola, F. A., and Bauer, D., Electrochim. Acta, 1992, 37, 2009. Adekola, F. A., Diaw, M., Colin, C., and Bauer, D., Electrochim. Acta, 1992, 37, 2491. Tsymbal, M. V., Brainina, Kh. Z., and Temerdashev, Z. A., Zh. Anal. Khim., 1993, 48, 1032. Perry, B. J., Van loon, J. C., and Speller, D. V., J . Anal. At. Spectrom., 1992, 7, 883. Perry, B. J., Speller, D. V., Barefoot, R. R., and Van Loon, J. C., Can. J . Appl. Spectrosc., 1993, 38(5), 13 1. Perry, B. J., Barefoot, R. R., Van Loon, J. C., Naldrett, A. J., and Speller, D. V., Spec. Pub1.-R.Soc. Chem., 1993, (124), 91. Godfrey, J., and McCurdy, E., Spec. Puh1.-R. Soc. Chem., 1993, (124), 64. Totland, M., Jarvis, I., and Jarvis, K. E., Chem. Geol., 1993, 104( I+), 175. Van Hoven, R. L., Chem. Abstr., 1994,120, 288946e. Akatsuka, K., Hoshi, S., McLaren, J. W., and Berman, S. S., Bunseki Kagaku, 1994,43,61. Colodner, D. C., Boyle, E. A., and Edmond, J. M., Anal. Chem., 1993, 65, 1419. Chiba, K., Inamoto, I., and Saeki, M. J., Anal. At. Spectrom., 1992,7, 115. Inamoto, I., and Chiba, K., Tetsu to Hagane, 1993, 79(2), 175. Suzuki, K., Lu, Q., Shimizu, H., and Masuda, A., Analyst, 1992, 117, 1151. Hall, G. E. M., and Pelchat, J. C., J . Anal. At. Spectrom., 1993, 8, 1059. Saprykin, A. I., and Kovalev, Yu.V., Fresenius’ J . Anal. Chem., 1992, 342, 15.Watling, R. J., Herbert, H. K., Delev, D., and Abell, I. D., Spectrochim. Acta, Part B, 1994, 49, 205. Lee, G. H., Yang, S. R., Park, C. J., and Lee, K. W., Bull. Korean Chem. Soc., 1993, 14, 696. McCombes, P. T., Borthwick, I. S., Jennings, R., Ledingham, K. W. D., and Singhal, R. P., Inst. Phys. Con$ Ser., 1991, (1 13), 163. Huggins, T. G., and Henion, J. D., Electrophoresis (Weinheim, Fed. Repuh. Ger.), 1993, 14, 531. Allain, P., Berre, S., Mauras, Y., and Le Bouil, A., Biol. Mass Spectrom., 1992, 21(3), 141. Williams, M. C., ASTM Spec. Tech. Publ., 1991, (1 109), 96. Perry, B. J., and Van Loon, J. C., Appl. Plasma Source Mass Spectrom. (Sel. Pap. Int. Conf.), 2nd, 1990 (Pub. 1991), 38. Ma, W. Y., Li, L. Q., Zhao, W. Z., and Chen, D. Y., Opt. Commun., 1991, 85, 408. Ma, W., Li, L., Zhao, W., and Chen, D., Yuanzi Yu Fenzi Wuli Xuebao, 1992,9, 2346. Stone, W. E., and Crocket, .I. H., Chem. Geof., 1993, 106(34), 219. Wildhagen, D., and Krivan, V., Anal. Chim. Acta, 1993, 274, 257. Perlwitz, A., Erasmus, C. S., and Pillay, A. E., Nucl. Geophys., 1991, 5 , 303. Bichler, M., Radioact. Radiochem., 1991, 2(3), 50. Chai, Z., Kong, P., Mao, X., Ma, S., Ma, J., Zhou, Y., Wang, J., and Liu, Y., Fenxi Shiyanshi, 1993, 12( l), 43. Turaev, Kh. Kh., Khudaibergenov, U., Ganiev, A. G., and Turdiku- lov, A. Kh., Dokl. Akad. Nuuk Resp. Uzb., 1992, (lo-ll), 64. 1993, 15 (1-3), 184. ~ 727 728 729 730 73 1 732 733 734 735 736 737 738 739 740 74 1 742 743 744 745 746 747 748 749 750 75 1 752 753 754 755 756 757 758 759 760 76 1 762 763 764 765 766 767 Wang, J., and Chai, Z., He Huaxue Yii Fangshe Huaxue, 1992, 14, 239. Liu, Y., and Chai, Z., He Huaxue Yu Fangshe Huaxue, 1993, 15, 26. Kumar, S., Verma, R., and Gangadharan, S., Analyst, 1993, 118, 1085. Rosenberg, R. J., Zilliacus, R., and Lipponen, M., Tutkimusrap.-Geol. Tutkimuskeskus, 1993, 114, 33. Malik, H., and Parry, S. J., Analyst, 1992, 117, 1945. Ward, C. B., Nucl. Geophys., 1991, 5 , 291. Miransky, I. A., Kist, A. A., Bakhrieva, F. B., andTeplyakov, P. V., .I. Radioanal. Nucl. Chem., 1993, 168, 329. Ahmad, K., Osae, E. K., and Amuh, G., Nucl. Tech. Enplor. Exploit. Energy Miner. Resour., Proc. Int. Symp., 1990 (Pub. 1991), 591. Das, N. R., and Maity, H. P., J. Radioanal. Nucl. Chem., 1993, 176, 45. Albert, P., Bajard, J. C., Delmas, R., and Fedoroff, M., J . Radioanal. Nucl. Chern., 1993, 168, 337. Albert, P., Bajard, J. C., Delmas, R., and Fedoroff, M., J. Radioanal. Nucl. Chem., 1993, 171, 51 1. Franek, M., and Krivan, V., Fresenius’ J . Anal. Chem., 1992, 342. 118. Choi, J. U., Chun, K. S., Lee, J. D., Lee, C., and Chung, K. S., Bull. Korean Chem. Soc., 1993, 14, 367. Turaev, Kh. Kh., Khudaibergenov, U., and Ganiev, A. G., Dokl. Akad. Nauk Resp. Uzh., 1992, (12), 36. Van Renterghem, D., Cornelis, R., and Vanholder, R., Anal. Chim. Acta, 1992, 257, 1. Rajadhyaksha, M. M., and Turel, Z. R., J. Radioanal. Nucl. Chem., 1992, 156, 341. Zhu, Z., and Wang, Y., Fushe Fanghu, 1991, 11, 224. Eddy, B. T., Stuckenberg, B. L., and Pansi, G., Adv. X-Ray Anal., 1991, 34, 277. Frolova, M. M., Golentovskaya, I. P., Smagunova, A. N., Zagu- mennova, V. D., Morozova, L. V., and Trofimov, B. A., Zavod. Lab., 1991, 57(8), 33. Robertson, M. E. A., Murray-Smith, R. de L., and Prosser, M. J., Randol Gold Forum, 1991 (Pub. 1992), 219. Barfoot, K. M., Hinde, J. P., Laws, M. J., Murray-Smith, R. de L., and Robertson, M. E. A,, Int. Congr. Appl. Mineral., 1991, Vol. 1, Paper 4. Kuznetsova, A. I., Lavrent’ev, Yu. G., and Maiorova, 0. N., Zavod. Lab., 1992, 58(6), 20. Obrazovakii, E. G.. Gil’bert, E. N., and Bulaeva, I. P., Zh. Anal. Khim., 1992, 47, 2057. Peraniemi, S., Parkkinen, J., Smolander, K., Mustalahti, H., and Ahlgren, M., Fresenius’ .I. Anal. Chem., 1992, 343, 292. Aboh, I. J . K., Bamford, S. A., and Tetteh, G. K., X-Ray Spectrom., 1992, 21(3), 119. Cojocaru, V., and Spiridon, S., J . Radioanal. Nucl. Chem., 1993,170, 259. Dmitriev, S. N., Shishkina, T. V., Zhuravleva, E. L., Shishkin, S. V., and Gerbish, Sh., Process Metall., 1992, 7A, 509. Lobanov, F. I., Logunova, S. A., Popov, A. M., and Krasnopevtseva, E. V., Zh. Anal. Khim., 1994, 49, 295. Zheng, H., and Qin, X., Yankuang Ceshi, 1992, 11, 217. Startsev, Y. S., and Shvartsman, M. M., Russ. Pat., 1735209, 1992. Alekseeva, T. Yu., and Orlov, V. V., Russ. Pat., 1732245, 1992. Luo, L., An, Q., and Song, E., Yankuang Ceshi, 1993, 12, 234. Wei, C., Li, D., Hu, J., Fan, Q., and Liu, Y., Hejishu, 1992, 15, 531. Stroba, A., and Zimna, K., Rudy Met. Niezelaz., 1992, 37(5), 136. Hu, G., Huaxue Shijie, 1992, 33, 464. Blossfeld, D. H., and Schnieder, E. W., US Pat., 5 185773, 1993. Kutvitskii, V. A., Osokin, E. N., Khomutova, E. G., Chernyshova, L. M., Ustinov, I. V., Kozik, A. V., and Kamal’dinov, Sh. A., Russ. Pat., 1636 747, 1991. Pillay, A. E., and Mashilo, N., S. Afr. J . Chem., 1993, 46(1-2), 20. Demortier, G., Steukers, C., Vandesteene, N., and Mathot, S., Nucl. Instrum. Methods Phys. Res., Sect. B, 1994, 85(1-4), 311. Huston, D. L., Sie, S. H., Large, R. R., Bottrill, R. S., andRyan, C. G., Proc. Aust. Con$ Nucl. Tech. Anal., 1991, 7th, 129. Ren, C., Zhou, S., Hu, W., Huang, F., Tang, J., Yang, F., Wang, K., and Zhou, Y., Chin. Sci. Bull., 1992, 37, 375.Analyst, February 1996, Vol. 121 161 768 769 770 77 1 772 773 774 775 776 777 778 779 Criddle, A. J., Tamana, H., Spratt, J., Reeson, K. J., Vaughan, D., and Grime, G., Nucl. Instrum. Methods Phys. Res., Sect. B, 1993,77( l a ) , 444. Tang, X., and Yao, J., Yankuang Ceshi, 1992, 11, 162. 781 Zhang, M., and Gong, M., Yejin Fenxi, 1991, 11(3), 1. Gong, M., and Zheng, W., Yejin Fenxi, 1991, 11(5), 15. 782 De Oliveira e Filho Silva, J., Bruz. Put. 93/01 973, 1993. Wu, Z., Gin, Z., and Wu, X., Chin. Pat., 1071 754, 1993. 783 Thompson, E. C., PreciousMet., 1993, 17th, 539. Brown, N. M. D., Hewitt, J. A., and Meenan, B. J., Surf. Interface 784 Anal., 1993, 20, 215. 785 Solozhenkin, P. M., Usova, S. V., Pupkov, V. S., Yakovenko, I. V., Bykova, N. O., and Maron, R. S., Russ. Pat., 1727054, 1992. Weigand, P., Jelinek, B., Hofer, W., and Varga, P., Nucl. Instrum. Methods Phys. Res., Sect. B, 1994, 85( 1 4 ) , 424. Harada, T., and Tanuma, S., Bunseki Kaguku, 1991, 40, 71 1. Wirth, T., Fresenius' J . Anal. Chem., 1993, 346, 308. 780 Saichenko, A. N., Saichenko, L. A., and Engel'sht, V. S., Nov. Meyody Anal. Vysokochist. i Tekhn. Muter. Akud. Nauk. SSSR, In-t Neorgan. Khimii, Novosibirsk, 1990, 2 16. Golovanova, N. V., Runov, V. K., Sadvakasova, S . K., and Khvostova, V. P., Russ. Pat., 1647392, 1991. Wen, K., Chen, D., Ma, W., Hui, Q., Zhao, W., Li, L., and Cheng, Y., C-MRS Int. Symp. Proc., 1990 (Pub. 1991), 1, 189. Radhakrishnamurty, C., and Rao, G. S., Acta Cienc. Indica, Chem., 1990, 16C(1), 113. Shimada, T., and Hatakeyama, T., Jpn. Pat., 05 322 876, 1993. Jia, X., Wang, Y. Wang, C., and Ma, D., Huaxue Shiji, 1993, 15, 369. Paper 5103826J Received June I3 , 1995 Accepted October I I , 1995

ISSN:0003-2654    

DOI:10.1039/AN9962100139

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. I21 (163-168) 163 Com bined Benedett i-Pich ler/Stange-Poole Sampling Equation for Two-component Particulate Mixtures Lu Zheng and Byron Kratochvil" Department of Chemistry, University of Alberta, Edmonton, Alberta, T6G 2G2, Canada General equations for the estimation of statistical uncertainty in the sampling of mixtures of two types of particles were developed that take into account variability in particle size and composition. Different expressions were obtained for situations where the component of interest is present in bulk or trace amounts. The general equation for a bulk component was verified experimentally by sampling KCI-NaCI and KC1-sucrose mixtures with various particle sizes and compositions with a spinning rimer and determining the relative standard deviation of the percentage of C1 or KCI under conditions where the analytical uncertainty was small.Keywords: Sampling; statistics; chemical analysis; spinning rifler Introduction The sampling of particulate mixtures is important in many areas of chemical and drug production. The statistical estimation of sampling error has been related to particle size, number and density and also the fraction of each particle type.l About 40 years ago, Benedetti-Pi~hler~.~ and Stange4 independently showed that appreciable random error may occur when sampling two-component mixtures whose particles differ significantly in composition or size. The Benedetti-Pichler equation was based on mixtures of two types of particles of uniform size, with each type containing a differing quantity of a component of interest.An example is a ground ore. The Stange equation, on the other hand, is designed for a range of particle sizes but with the analyte of interest present in only one particle type, as in many drug formulations. No expressions are available that deal with mixtures of particles that vary in both size and composition. In this work, we developed and confirmed experimentally a general equation for the estimation of statistical sampling uncertainty for mixtures of such particles. We also developed a similar equation based on a Poisson rather than a Gaussian distribution. Background Benedetti-Pichler derived a quantitative expression for the sampling relative standard deviation R, in parts per hundred, of the percentage P of a component of interest as ~ ~ ~~~~ * To whom correspondence should be addressed. From eqn.(l), it follows that the number of particles n required per sample to hold R to any specified value can be calculated by where P = overall average percentage of sought component in a well mixed sample of n units; P1 = percentage of type A particles; P2 = percentage of type B particles; p = fraction of particles in the mixture that are type A; dl = de_nsity of type A particles; d2 = density of type B particles; d = weighted density of the mixture. In eqn. (l), Benedetti-Pichler assumed all of the particles to be the same size and each type of particle to contain a different percentage of the analyte of interest. By also assuming a uniform particle shape, such as spherical, the relationship between test portion mass and a specified sampling uncertainty can be calculated.5 In practice, however, particle sizes in a material often cover a range. For a well mixed pair of powders that differ in density and particle size range, Stange4 used the binomial distribution to obtain an expression for the standard deviation of the mass fraction of type A particles per sample, a(W): where W = g l / g , fraction of type A particles in the sample by mass; x = fraction by mass of type A particles in mixture; y = fraction by mass of type B particles -- in - mixture; g = mass of sample taken from the mixture; ww,, w2) = mean mass of the components of the mixture; C,2 = Sx2/wx2, relative variance in mass of type A particles; Cy2 = Sy2/Gy2, relative variance in mass of type B particles.Here S, and Sy are the standard deviations in masses of the two components of the mixture. Poole et aL6 have shown that eqn. (3) may be expressed as wheref, = fraction by mass of component of interest in particle size class i (note that Xfi = unity); mi = mean mass of particles consisting of component of interest in a given particle size class; Zf,rni = mean effective particle mass of the component over all i particle size classes. Use of eqn. (4), which we shall call the Stange-Poole equation, to estimate the content uniformity of drugs has been reviewed by Orr.7 If the analyte of interest is present at trace levels compared with the diluent, and the random selection of a trace particle as164 Analyst, February 1996, Vol.121 opposed to a diluent particle is a rare event, the Poisson distribution applies. Johnson8 derived an equation for this case in which the variance due to particle size distribution was considered to be a linear combination of independent randomly distributed variables: V;(2r1)3 +...+ fi(2ri)3] xdlgl Qd2 =7 so ( 5 ) where gl = mass of trace component per sample; dl = density of particles of trace component; r = radius of particles of trace component. Egermann et al.9-12 studied several applications of eqn. (5). Wilson13 modified eqn. (2) to express the standard deviation in terms of sample mass g rather than number of particles: a(P) = (PI - P2)+ (y) (;) Here a uniform particle volume v is assumed. For sample particles having a wide size distribution, Wilson calculated a weighted average volume v, assuming the distribution of particle size is the same for each type of particle: h = l where = weighted volume of each particle in the sample; f b = fraction of the hth group by mass; vh = volume of a particle in the hth group. Grant and Peltonl4.15 compared experimental relative stan- dard deviations with three different ways of estimating weighted particle volumes and found that average particle volumes agreed better with the experimental results than volumes of the largest particles.Because eqn. (7) requires the size distribution of each type of particle to be similar, its application is limited. In this regard, the Stange-Poole equation [eqn. (4)] has an advantage. Theoretical Expanded Benedetti-Pichler Equation A more useful expression would accommodate cases where two types of particles, each type uniform in size, are present in a sample but the particle size of each type differs from that of the other.This can be done by expansion of the Benedetti-Pichler equation (see Appendix for derivationl6) to give an expression for the standard deviation of P, the target component in the mixtures: Defining Vl/V2 as c, the relative standard deviation R , in parts per hundred, is then given by We shall call eqn. (9) the extended Benedetti-Pichler equation. The original equation can be shown to be a special case of the extended version by noting that if the particle size in a mixture of the two types A and B is uniform, c = 1, whereupon (cnl + n2)4 = (nl + n2)4. Therefore, since n1 + 122 = n, eqn.(9) reduces to eqn. (1). Combined Benedetti-PichlerlStange-Poole Equation Eqn. (9), the extended Benedetti-Pichler equation, requires all particles of each type to be of the same size, whereas in practice a range of sizes is often encountered. A more generally applicable expression can be obtained by combining the Stange-Poole and extended Benedetti-Pichler equations (see Appendix). The resulting combined equation for the standard deviation o(P) is and for the relative standard deviation R (again in parts per hundred), Eqn. (1 1) relates the sampling error to the relative difference in composition between the two types of material, the fraction of mass contributed by each type, the mass of sample and the mean effective particle masses. Combined equation for trace amounts of one type of particle in a mixture Eqn.(4) is derived from the binomial distribution, whereas eqn. ( 5 ) applies the Poisson distribution to situations where only a small fraction of the particles in a mixture contain the component of interest. In eqn. (5) the background particles are the same size or larger than the particles of interest, and the particles of interest have a range of sizes. The standard deviation in the concentration of the trace component in a sample according to eqn. (5) is given by When the substance of interest is present in trace amounts in both types of particles, and the fraction of one type is small, a combination of the Benedetti-Pichler and Johnson equations is applicable. Substituting o(g1) for Ag1 [see eqn.(Al) in the Appendix], the standard deviation o(P) is given by and the relative standard deviation R of the trace component in parts per hundred by Ege~manng.~o reported the Johnson equation [eqn. ( 5 ) ] to be preferable to the Stange-Poole equation [eqn. (4)] when the difference in particle sizes is large. Accordingly, eqn. (12) is more suitable than eqn. (1 1) when one type of particle is present in trace amounts (defined here as less than I%), when the size of the major particle type is significantly larger than that of the minor and when both types contain the analyte of interest. A summary of conditions under which the various equations discussed here apply is given in Table 1. These equations are valid only for homogeneous mixtures. Because differences in composition, size and density can cause segregation of particles during and after mixing, carefulAnalyst, February 1996, Vol.121 165 sampling is necessary to minimize bias during collection. This problem was minimized in the experimental work described here to confirm the validity of eqn. (1 1) by use of a spinning riffler. It was not possible to confirm eqn. (12) with the system used here because the analytical uncertainty became too large to allow useful comparisons. Experimental After consideration of several criteria, including uniformity in shape, purity, similar densities, ease of handling and measure- ment and availability over a range of particle sizes, potassium chloride, sodium chloride and sucrose, all of analytical-reagent grade, were chosen for the experimental work.The materials were sieved into several size fractions and stored in a desiccator over saturated potassium carbonate solution at a constant relative humidity of 43% at room temperature17 to reduce problems with static charge during weighing. The size ranges used here were 250-300, 300-425, 425-500 and 500-600 Pm* Evaluation of the statistical sampling expressions required laboratory sampling procedures that introduced minimum bias. Three techniques were tested: scoop sampling, either from a well mixed material poured into a cone on a clean flat surface or directly from a bottle that was rotated slowly end-over-end about ten times between withdrawal of sample portions; use of a laboratory-built sample splitter in which the original material was split a number of times to yield roughly equal test portions suitable for titration; and use of a commercial spinning riffler (Model SP-225, Gilson, Worthington, OH, USA), which divided the sample into 20 equal test portions on a rotating tray.To accommodate the small amounts studied here, the sample feed chute of the riffler was narrowed and the sample hopper replaced with a small glass funnel to provide feed times of 3-5 min for 2-4 g of material. Typically, two sets of eight alternate tray segments of 0.2-0.4 g each were taken from each run as test portions for analysis. For each set, the percentage of chloride or potassium chloride in the mixture was determined, along with the relative standard deviation in per cent., by titration with silver nitrate using dichlorofluorescein as adsorption indicator.'* The 0.05 moll- silver nitrate titrant was standardized against pure potassium chloride in alternation with the samples. Typically eight samples and four KC1 standards were analysed per experiment. Relative standard deviations for the standardizations ranged from 0.3 to 1.6 parts per thousand.The relative standard deviation of sampling, R,, was obtained for each set from the equation R, = d R o 2 - R,2, where R, is the overall relative standard deviation and R, the relative standard deviation of the ana1y~is.l~ From the masses of KCl, NaCl and sucrose, the theoretical percentage of chloride (or percentage of potassium chloride in the KC1-sucrose mixtures) was calculated, along with the mass fraction of each type of material, in the mixtures.Average test portion masses were calculated for each of eight test portions, and average particle volumes and masses were determined from one or more of measurement of sieve size, measurement of particle size by microscope or counting the number of particles and calculation of size from the density and total mass. The three methods gave similar values. From these parameters the relative standard deviations in the distribution of the analyte of interest were calculated using eqns. (9) and (1 1) for mixtures of two types and sizes of particles. Prediction of Relation Between Relative Standard Deviation of Sampling (R) and Particle Parameters Using the equations developed here, R values were calculated for combinations of particle size, composition and fraction using the spreadsheet program EXCEL (Microsoft) on a Macintosh SE computer.On the basis of these calculations, mass ratios of the two components were selected to provide values large enough to permit useful comparison of the experiments with theory. Fig. 1 shows the relationship predicted between sampling relative standard deviation for chloride and mass fraction of NaCl in mixtures of NaCl and KC1. Data for the figure were calculated by eqn. (1 1) with particle size and test portion mass held constant. The maximum value of R for chloride occurs at a mass fraction for NaCl of approximately 0.4. Fig. 2 illustrates the change in R with test portion mass, again calculated by eqn. (1 1) for constant particle size and a mass fraction of 0.4 NaCl.On the basis of these plots, experimental conditions were chosen to ensure a sufficiently large sampling variance that it could be determined with adequate precision and would not be swamped by the variance of the measurements. Results and Discussion Evaluation of Experimental Sampling Procedures To compare the effectiveness of the scoop, split riffler and spinning riffler methods, 1 + 1 mixtures by mass of potassium Table 1 Summary of conditions under which the equations described for estimation of sampling error in mixtures of two types of particles are applicable Equation Equation No. Particle size Conditions* Substance of interest Benedetti-Pichler (1) Uniform Non-trace Present in both types of Wilson (6) Same distribution for each of Non-trace Present in both types of S tange-Poole (4) Differing distributions between Non-trace Present in only one of the two types of particles Johnson ( 5 ) Differing distributions between Trace Present in only one of the two types of particles Extended (9) Different but uniform sizes for Non-trace Present in both types of Combined (1 1) Differing distributions between Non-trace Present in both types of throughout particles two types of particles particles each of two types of particles each of two types of particles Benedetti-Pichler each of two types of particles particles Benedetti-Pichler/ each of two types of particles particles Stange-Poole Benedetti-Pichler/ each of two types of particles particles Johnson Combined (12) Differing distributions between Trace Present in both types of * Trace conditions defined as the component of interest being present at 1% or less under conditions where the Poisson distribution applies.166 Analyst, February 1996, Vol.121 chloride (particle size range 500-600 pm) and sucrose (425-500 pm) were sampled. Test portions of about 0.2 g were collected and analysed for KCl content. Table 2 gives the results; runs with the split riffler are listed individually because the portion masses varied. The spinning riffler provided the most representative test portions, and was therefore used, after a series of experiments to optimize the operating parameters, for all subsequent evaluations of the equations. Variables studied in the optimization included the rate of sample feed to the riffler and the location of riffler segments from which test portions 2 0.8 r I G I I ~ a 01 I I I I J 0 0.2 0.4 0.6 0.8 1 Mass fraction of NaCl Fig.1 Relation between relative standard deviation for chloride and mass fraction of NaC1, calculated by eqn. (1 l), for mixtures of 600 pm (28 mesh) NaCl and 500 pm (32 mesh) KCI at a constant test portion mass of 0.8 g. 1.5 r 1 I I I I Test portion masslg Fig. 2 Relation between relative standard deviation for chloride and mass of test portion, calculated by eqn. (1 l), for a mixture of NaCl and KCI is in Fig. 1 but at a constant mass fraction of 0.4 NaC1. 0 0.4 0.8 1.2 Table 2 Comparison of scoop, split riffler and spinning riffler sampling methods for a mixture of 500-600 pm (28-32 mesh) KC1 and 425-500 ym (32-35 mesh) sucrose. Mass fraction of KC1 = 0.5; test portions = 0.2 g; n = 8.Theoretical values based on eqn. ( 5 ) (Stange-Poole equation) Relative standard deviation in % KC1 Calculated Calculated using using eqn. (4) and top sieve average sieve eqn. (4) and Sampling method size size Experimental Scoop, from cone* 2.8 2.4 4.1, 7.4 Scoop, from vial? 2.8 2.4 4.1,*, 6.2, 4.8, Split riffler, trial 1 3.0 2.7 4.6 Split riffler, trial 3 2.5 2.2 2.8 Spinning riffler 2.9 2.6 3.1, 2.6, 2.7, 2.4, 4.5s Split riffler, trial 2 2.4 2.1 2.2 2.6, 2.4, 3.2, 2.6; av. 2.7 * Dip sampling from coned material. t Dip sampling from vial, with mixing after removal of each sample. * n = 14. 5 n = 10. were collected. Table 3 compares experimental results for KCl content and R of sampling for five sampling rates and two collection patterns for populations containing 49549.7% KCl (calculated from the masses of KCl and sucrose taken) and R = 2.9 parts per hundred [calculated by eqn.(1 1) using a particle count to determine n]. Calculations assuming spherical particles with diameters equal to the top sieve size gave R = 3.0 parts per hundred; an assumption of cubic particles with cube edges equal to top sieve size gave R = 4.2. Examination by microscope showed the sucrose particles were generally spherical and the KC1 was mostly cubic with rounded corners, plus a few rectangular shapes corresponding to two cubes fused together. Since elongated shapes yield larger particle volumes than those calculated on the basis of cubic or spherical shapes, counting of particles was chosen over use of sieve size or estimation of particle size and shape distribution by microscope.To test the ability of the spinning riffler to collect a valid sample of a highly segregated material, individual weighed portions of NaCl and KCl were passed sequentially through it and collected in 20 portions. The results, shown in the first row of Table 4, indicate that when correctly used, a spinning riffler allows accurate sampling of even poorly mixed material. Table 3 Comparison of calculated and measured relative standard deviations values for chloride in test portions from spinning riffler as a function of sampling time. Samples were a mixture of 500-600 pm (28-32 mesh) KC1 and 425-500 pm (32-35 mesh) sucrose Relative standard deviation (pph) Chloride (%) Calc.* Ordered? Random+ Calc.Ordered? Random* Time/min 2.9 3.1 2.6 2.9 2.7 2.4 2.9 2.6 2.4 2.9 2.1 2.7 2.9 3.2 2.6 3.3 2.7 3.0 2.0 2.4 2.6 2.6 2.9 3.5 2.2 49.6 49.8 48.8 49.6 49.5 49.5 49.5 49.1 49.9 49.6 48.7 49.9 49.7 49.8 49.5 49.0 20.5 49.5 49.6 12 49.4 49.9 9.5 49.1 49.1 9.8 49.5 49.7 6 49.6 * Calculated by extended Benedetti-Pichler equation [eqn. (9)] from masses of KC1 and sucrose taken. t Eight test portions for each data set taken in sequential order from the 20 portions collected in the spinning riffler. * Eight test portions for each data set taken at random from the 20 portions collected in the spinning riffler. Table 4 Comparison of calculated and experimentally determined relative standard deviations [in parts per hundred (pph)] and percentage of chloride for mixtures of 500-600 pm (28-32 mesh) NaCl and 250-300 pm (4840 mesh) KCI.Nominal test portion mass 0.08 g throughout 90% confidence Calculated Measured intervals R (pph) C1 (%) R (pph) C1 (%) ForR* For % Cll 0.53 0.5 1 0.5 1 0.5 1 0.5 1 0.5 1 0.5 1 0.5 1 0.52 0.5 1 52.73 52.73 52.76 52.76 52.80 52.80 52.76 52.76 52.77 52.77 0.41 0.45 0.61 0.30 0.52 0.63 0.35 0.70 0.4 1 0.60 52.61 52.77 52.63 52.73 52.83 52.7 1 52.64 52.72 52.67 52.81 * Degrees of freedom = 8 - 1. + n = 8. 0.29-0.74 0.32-0.81 0.21-0.54 0.37-0.93 0.44-1.13 0.25-0.63 0.29-0.74 0.43-1.10 0.49-1.26 0.42-1.08 5 2.47-5 2.7 5 52.6 1-52.93 52.5 3-52.93 52.48-52.78 52.64-53.02 52.48-52.94 52.5 1-52.77 52.47-52.82 52.52-52.85 52.60-53.02Analyst, February 1996, Vol. 121 167 Comparison of Com bined Ben edetti-Pic h lerlstange-Poole Equation [Eqn (1 I ) ] With Experimental Results For these experiments, particle sizes for NaCl of 500-600 pm (28-32 mesh) and for KC1 of 250-300 pm (48-60 mesh) were used.Test portion masses were about 0.08 g and the mass fraction of NaCl was held at 0.4. Table 4 compares the results for ten repetitions of the experiment with values calculated by eqn. (11). Comparison by Hartley's F,,, test20 of the ten experimental relative standard deviations indicates that they are homogeneous, that is, they can be considered to come from the same population. The average of the experimental relative standard deviations, 0.50 parts per hundred, agrees well with the value calculated using eqn. (1 l), 0.51 parts per hundred. A standard F-test indicates that these variances are not sig- nificantly different at better than a 95% confidence level. With the reasonable assumption that all experimental means and relative standard deviations came from the same popula- tion, an over-all relative standard deviation of 0.49 was obtained for the ten experiments (80 measurements) in Table 4, with a 90% confidence interval of 0.56-0.43.This experimental R compares well with a value of 0.51 calculated using eqn. (1 1). For comparison, the Benedetti-Pichler equation [eqn. ( l)] gives a value of 0.62, using a particle size of 600 pm, the larger end of the NaCl sieve range. Results of experiments in which four other particle size ranges of NaCl and KCl were studied are shown in Table 5. Here, too, values calculated by eqn.(1 1) for R fall within the 90% confidence intervals for the experimental data. In summary, the experiments described here confirm that eqn. (11), a general expression obtained by combining the Stange-Poole and extended Benedetti-Pichler equations, can be successfully applied to real systems containing two types of particles. The expression takes into account not only relative differences in composition of the two types of particles but also a range of particle sizes. The parallel eqn. (12) for trace components, developed by combining the Johnson and ex- tended Benedetti-Pichler equations, should also be more broadly applicable than past expressions. This equation could not be tested experimentally by the method used for eqn. (1 1) because the measurement method was too imprecise for the low concentrations involved.The standard deviations obtained by these expressions represent the best that can be achieved i.e., using perfect experimental sampling techniques. Additional uncertainty will be introduced if the population being sampled is not well mixed or if the physical operations of sample collection are biased. We also confirm that a spinning riffler, when correctly used, can subsample particulate laboratory samples with negligible error and is much superior to scoop or split riffler sampling. Appendix Derivation of Extended Benedetti-Pichler Equations [Eqns. (8) and (9)]16 The percentage of X in a mixture of two types of particles can be given as nldlvlP1 + nZd2~2P2 nldlvl + n2dZv2 P = where nl = number of particles of type A in sample; 112 = number of particles of type B in sample; dl = density of type A; d2 = density of type B; v l = volume of a particle of type A; v2 = volume of a particle of type B; P1 = percentage of X in type A; Pz = percentage of X in type B.Defining c as vI/v2, cnldlPl + n2d2P2 cn ldl + n2d2 P = For a constant sample mass, the sampling error of a mixture is related to the number n1 + Anl of particle type A and n2-Anl of particle type B instead of to the theoretical numbers n1 and n2. Therefore, the sampling error in the percentage of X, AP, is a function of An Differentiating, Let Since v1 = cv2, ~~~~~ ~~ Table 5 Comparison of calculated [eqn (1 l)] and experimentally determined values for relative standard deviation of sampling R in parts per hundred (pph) and percentage of chloride in mixtures of KCI and NaC1.Duplicate results shown for each set of conditions. Nominal test portion masses 0.08 g 90% confidence Calculated Determined intervals Particle size ranges used R (pph) Cl(%) R (pph) CI (%) ForR' For % Clt NaCl(500-600 pm) KCl(425-500 pm) NaCl(500-600 pm) KCl (425-500 pm) NaCl (500-600 pm) KCI (250-300 Fm) NaCl(500-600 pm) KCI (100-150 pm) NaCl (300-425 pm) KCl (425-500 pm)* and and and and and 0.58 52.88 0.52, 0.50 0.58 52.79 0.54, 0.37 0.5 1 52.77 0.41, 0.60 0.49 52.71 0.42, 0.49 0.60 54.09 0.65, 0.54 52.89, 52.76 52.77, 52.69 52.67, 52.81 52.64, 52.75 53.96, 54.00 5.63, 0.10, 5.20 1.28 6.07, 0.20, 2.85 1.42 4.52, 1.31, 9.69 0.36 5.14, 0.86, 7.00 0.42 8.22, 1.13, 5.67 0.98 * By chi-squared test.Degrees of freedom for each set of measurements = 8 - 1; at 90% confidence level, Xi,o5 = 2.17, 90% confidence level, t = 1.90. * Nominal test portion mass 0.05 g. = 14.07. + For n = 8, at168 Analyst, February 1996, Vol. 121 and therefore Since the mass fraction W = gl/g, then According to the binomial distribution, the standard deviation is given by hence Anl = = V n p ( 1 - p ) and the relative standard deviation R in parts per hundred is given by Derivation of Corn bined Benedetti-Pic hlerf Stange-Poole Equation [Eqn. (ll)] For a mixture that contains two types of particles, A and B, each of which can have a different size distribution, let P = the total percentage of sought component in the mixture; P I = percentage in particle type A; P2 = percentage in particle type B; g = total mass of sample; gl = mass of particle type A in sample; g2 = mass of particle type B in sample.The percentage of sought component in the mixture can then be calculated by glpl +g2p2 P = <P Since g = gl + g2 and g2 = R - g l , this equation can be written as R As before, sampling error arises when the masses of the two types of particles in the sample become gl + A g , and g2 - Agl instead of the theoretical masses gl and g2. Since the content P + AP of the actual sample is then a function of g l + A g l , P = f ( g l ) , and differentiation gives p1 - p2 AP =- R X Agl = o(P) (Al) A g l in eqn. (Al) can be substituted as follows. Since the standard deviation o(w> for different particle sizes in a two particle type random mixture of any composition is given by Stange4 as (g constant and W = g l / g ) rearranging according to Poole6 gives where (CfimJX = mean effective particle mass of type A particles; (Z:fimJy = mean effective particle mass of type B particles.Therefore, Substituting eqn. (A2) into eqn. (Al) for Ag1, the standard deviation o(P) = AP is given by Converting standard deviation to relative standard deviation R in parts per hundred gives the combined expression Eqn. (1 1) can be shown algebraically16 to incorporate eqns. (l), ( 5 ) , (8) and (12). References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Kratochvil, B., Wallace, D., and Taylor, J., Anal. Chem., 1984, 56, I13R. Benedetti-Pichler, A. A., in Physical Methods in Chemical Analysis, ed. Berl, W. G., Academic Press, New York, 1956, vol. 111, pp. Benedetti-Pichler, A. A., Essentials of Quantitative Analysis, Ronald Press, New York, 1956, ch. 19, pp. 309-316. Stange, K., Chem. Eng. Tech., 1954, 26, 331. Harris, W. E., and Kratochvil, B., Anal. Chem., 1974, 46, 313. Poole, K. R., Taylor, R. F., and Wall, G. P., Trans. insf. Chem. Eng., 1964,42, T305. Orr, N. A., in Progress in Quality Control in Medicine, ed. Deasy. P. B., and Timoney, R. F., Elsevier Biomedical Press, Amsterdam, Johnson, M. C. R., Pharm. Acta Helv., 1972, 47, 546. Egermann, H., Pharm. Acta Hehi., 1985, 60, 322. Egermann, H., Pharm. Acta Helv., 1986, 61, 10. Egermann, H., Kemptner, I., and Pichler, E., Drug Dev. Ind. Pharm., 1985, 11, 663. Egermann, H., Int. J. Pharm. Tech. Prod. Manuf., 1982, 3, No. 2, 59-66. Wilson, A. D., Analyst, 1964, 89, 18. Grant, C. L., and Pelton, P. A., ASTM STP 540, 1973, 16. Grant, C . L., and Pelton, P. A,, Adr). X-Ray Anal., 1974, 17,44. Zheng, L., MSc Thesis University of Alberta, 1991. Harris, W., and Kratochvil, B., An introduction to Chemical Analysis, Saunders, Philadelphia, 198 1, ch. 21. Harris, W., and Kratochvil, B., An introduction to Cheniical Analysis, Saunders, Philadelphia, 1981, ch. 4. Youden, W. J., Statistical Methods for- Chemists, R. E. Krieger, New York, 1977. Sokal, R. R., and Rohlf, F. J., Biometry, Freeman, San Francisco, 2nd edn., 1981, p. 403. Paper 510031 6 0 Received January 18, I995 Accepted September 25, 1995 183-2 17. 1981, pp. 193-256.

ISSN:0003-2654    

DOI:10.1039/AN9962100163

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (169-1 72) 169 Multicomponent Analysis by Flow Injection Using a Partial Least-squares Model. Determination of Copper and Zinc in Serum and Metal Alloys Oscar Hernandez, Francisco Jimenez, Ana Isabel Jimenez and Juan JosC Arias Departamento de Quimica Analitica, Bromatologia y Toxicologia, Universidad de La Laguna, E-38204 La Laguna, Tenerife, Spain A partial least-squares multivariate calibration method for the resolution of mixtures of metal ions from the spectra of the complexes produced in a flow injection (FI) system has been developed. The method has been applied to the simultaneous determination of Cull and Znii with 4-(4'-methyl-2'-thiazolylazo)-2-methylresorcinol. The reagent exhibits a high absorbance and its spectrum strongly overlaps with the spectra of the complexes that it forms with copper and zinc ions.Chemical and FI variables influencing the system have been optimized and the proposed method validated on the determination of Cu and Zn in blood serum and metal alloys. Keywords: Partial least squares; copper; zinc; flow injection; serum; alloy Introduction Flow injection (FI) is a rapid, precise technique with many applications in studies involving agricultural, environmental, geochemical, clinical, pharmaceutical and industrial control samples.' Flow injection has so far been primarily used for quantitative determinations of a single analyte. A number of traditional chemical processes, such as dilution, extraction and derivatization, have been adapted for use in this type of continuous-flow system; in this way, the effect of potential interferences is suppressed and a specific signal for each analyte studied is obtained.Current hyphenated approaches using multidimensional instrumentation such as diode array detectors that allow entire spectra to be recorded for samples in a fraction, and powerful computational methods [multiple linear regression (MLR), principal components regression (PCR), partial least squares (PLS), Kalman filtering, factor analysis (FA)] capable of processing the vast amount of information produced by the former, have expanded the scope of application of the flow injection technique to multicomponent determinations. The significance achieved by FI combinations for simul- taneous determinations is clearly reflected in the growing number of reported applications of this type,2 most of which resolve analytes by using concentration or pH gradients.In this work we developed a new method for the simultaneous determination of copper and zinc by their reaction with the chromogenic reagent 4-(4'-methyl-2'-thiazolylazo)-2-methyl- resorcinol (MTAMR) using FI and a diode-array detector, the multiwavelength data thus obtained being processed by PLS methodology. The ensuing method was applied to the determi- nation of copper and zinc in blood serum and metal alloys. Experimental Apparatus The FI manifold used consisted of a Gilson (Worthington, OH, USA) Minipuls-2 four-channel peristaltic pump furnished with speed control and a Rheodyne (Cotati, CA, USA) 5041 manual injection valve. Absorbance measurements were made on a Hewlett-Packard (Avondale, PA, USA) 8452 A spectro- photometer furnished with a Hellma 174.717 QS flow cell of 18 1.11 inner volume and 10 mm pathlength.The instrument was interfaced to a Vectra ES computer, also from Hewlett- Packard. All tubes, injection loops and connectors used were made from PTFE tubing of 0.5 mm id. Whenever required, pH measurements were made with a Radiometer PHM84 digital pH-meter equipped with a com- bined glass-saturated calomel electrode. Software Absorption spectra were recorded and processed by using the spectrophotometer software. Data were transferred onto a PC/ AT 486 DX2 personal computer for processing by the program Unscrambler.3 Reagents A 1 X 10-3 moll-' standard solution of MTAMR in methanol was prepared by weighing the required amount of reagent.4 Copper(I1) and zinc(I1) nitrate standard solutions containing 106.7 and 134.8 mmol 1-1, respectively, were prepared.Boric acid-sodium borate buffer solutions (0.5 mol 1-1) of pH 7.42-9.56 were used. Working-strength solutions were pre- pared by appropriate dilution of the stock solutions. All reagents and solvents used were of analytical-reagent grade. Procedure Batch experiments In preliminary experiments, the absorption spectra for the complexes of the metal ions studied were recorded. For this purpose, a 25 ml calibrated flask was filled, in this order, with 0.5 ml of CU" or Zn" solution, 2.5 ml of buffer, 2.5 ml of methanol, 2.5 ml of I X 10-3 mol 1-I MTAMR and distilled water to the mark. The spectra of the solutions thus made were recorded between 500 and 700 nm against a blank prepared in the same way but containing neither analyte.170 Analyst, February 1996, Vol.121 Flow injection experiments A volume of 192 yl of sample was aspirated into the loop of the peristaltic pump and injected into a carrier, with which a buffer stream (CT = 0.1 mol 1-l, pH 9) and a reagent solution containing a reagent concentration of 1 x 10-4 moll-' in 10% v/v methanol-water had been previously merged. Absorption spectra were recorded from 520 to 620 nm at 1 s intervals for 40 s, using an integration time of 0.4 s. Determination of Cu and Zn in blood serum To 2 ml of serum 0.5 ml of 20% trichloroacetic acid was added. The mixture was stirred for 2 min and then centrifuged at room temperature for 5 min at 3000 rpm.The supernatant was neutralized using NaOH to a pH of approximately 7 and analysed as described above. Determination of Cu and Zn in metal alloys Accurately weighed amounts of approximately 1 g of metal were treated with 10 ml of concentrated (13.14 moll-1) HN03 and evaporated to near-dryness three times. The residue was diluted with 20 ml of distilled water and filtered through Whatman No. 41 filter paper. The solution was transferred into a 100 ml calibrated flask and made up to the mark with distilled water. The volume required to obtain a Cu and Zn concentration within the linear range of the calibration graph on dilution to 25 ml was taken and analysed as described above. Results and Discussion MTAMR is an azo dye that was synthesized and characterized as described elsewhere.4 Its spectral and acid-base features and those of its complexes with various metal ions have also been described.Thus, MTAMR has previously been used to carry out the determination of copper and vanadium in metal alloys,5-7 gallium in biological samples8 and zinc in waters.4.9 Fig. 1 shows the absorption spectra for the Cu-MTAMR and Zn-MTAMR complexes. As can be seen, the simultaneous determination of both metal ions with MTAMR using conven- tional spectrophotometric techniques is hindered by the ex- tensive spectral overlap; however, a flow injection spec- trophotometric method in combination with 0.5 1 partial 500 520 540 560 580 600 620 640 Wave leng tWnm Fig. 1 Absorption spectra of MTAMR and the Cu2+ and Zn2+ MTAMR complexes versus reagent blank at various pH levels.CMTAMR, moll-1; ccu, 8.44 X 10-6 mol 1-1; and cZn, 4.88 X 10-6 mol I-'. pH: 1, 7.8; 2, 8.5; and 3, 9.0. least-squares calibration allows the resolution of both ions in mixtures. Influence of Variables The performance of the flow manifold was affected by two broad types of variables: chemical and flow variables. Their individual effects were investigated for each ion in order to determine the optimal conditions for application of the proposed method. The final selected values were those which resulted in the maximum possible sensitivity and reproducibility in the FI peak recorded at 556 nm (the intermediate wavelength between the absorption maxima for the two complexes). Chemical Variables The height of the FI peak, which was studied over the pH range 7.4-9.5, remained constant above pH 8.5; therefore a working pH of 9 was selected.The effect of the buffer concentration was investigated between 0.05 and 0.5 mol 1-l. The peak height increased with increasing concentration of buffer; however, subtraction of the signal for a blank containing the same buffer concentration showed the absorbance due to the complex formed to be independent of such a concentration, so the signal increment was ascribed to changes in the refractive index arising from the salinity gradient of the medium. A buffer concentration of 0.1 mol 1-1 was chosen in order to minimize this effect. The influence of MTAMR concentration was studied over the range 5 X 10-5-2.5 X mol l-l? throughout which the peak height was found to be independent of MTAMR concentration.A concentration of 10-4 mol 1-1 was chosen for subsequent experiments as it gave rise to a moderately low absorbance at the absorption wavelengths for the complexes, thereby avoiding potential saturation of the spectrophotometer. The methanol content was found not to affect the sensitivity over the range from 6 to 20%. A methanol concentration of 10% v/v was chosen for this work as it allowed the reagent to be dissolved in pure methanol and had no adverse effect on the FI manifold. Flow Injection Variables Fig. 2 shows the FI peaks for Cu obtained at a variable flow rate. As can be seen, increasing the flow rate also increased the height of the peaks and decreased their width (and hence the dispersion); a similar behaviour is obtained for Zn.A reproduci- bility experiment involving 10 injections of samples at each flow rate studied revealed peak heights to be quite reproducible [standard deviation (s) less than 10-31 at flow rates below 1.75 ml rnin-l, which was thus chosen for subsequent experiments as it resulted in maximum sensitivity and acceptable reproduci- bility for both ions, in addition to a high sample throughput (over 90 samples per hour). The effect of the reactor length was studied between 10 and 240 cm using tubing of 0.5 mm id. The peak height for the Zn- MTAMR complex remained constant below 80 cm, above which it decreased through increased dilution. On the other hand, the peak height for the Cu-MTAMR complex increased slightly up to 80 cm (the complex was not formed instanta- neously) and then decreased with increase in the reactor length.A length of 80 cm was thus taken to be optimal. The effect of the injected volume, which determined the sensitivity, was studied from 117 to 249 pl. Increasing volumes also resulted in increasing peak height and width. A volume of 192 pl was chosen as it ensured good sensitivity and reasonably wide peaks. After the best flow conditions have been established, the precision of measurements made in a continuous-flow system depends on the frequency at which recordings are acquired and the time during which they are obtained. Therefore, weAnalyst, February 1996, Vol. 121 171 evaluated the reproducibility in 10 consecutive injections of the same sample, the spectrum of which was recorded at 0.5, 0.6, 0.7,0.8,0.9 and 1 .O s intervals using integration times between 0.1 and 0.5 s.The best results (s c 3 X 10-4) were obtained by recording spectra every second and using an integration time of 0.4 s. Table 1 gives the best values of FI variables for the determination of Cu and Zn with MTAMR. Calibration Calibration was carried out with the aid of a PLS multivariate model included in the software package, Unscrambler I1 v. 5.0,3 using the cross-validation method. The calibration matrix was constructed in a 5 X 5 design such as that shown in Table 2. The different absorbance data matrices tested for calibration were only extracted from the FI peak. Initially, the absorption spectra obtained at times from 8 to 16 s were treated as a series10 and used to validate the leverage correction method from centred data about the mean.Of the different times for the FI peak and wavelength ranges tested to construct the calibration matrix, it was found that the spectrum obtained 11 s after the 0.50, L 0.40 - u $0.35 0.30 - I 0-25.1... 0.20 0 10 20 30 40 50 60 70 80 90 Timeis Fig. 2 Variation of flow signals as a function of the flow rate for the Cu- MTAMR complex. A, 556 nm; CR, moll-1; ccur 2.1 1 X 10-5 moll-1; CBuffer, 0.1 mol I-]; pH 9.0; L, 160 cm; At = 1 s; ti = 0.4 s. Flow rates: 1, 0.52; 2,0.70; 3,0.87; 4, 1.04; 5, 1.22; 6, 1.40; 7, 1.57; 8, 1.75; and 9, 1.92 ml min-1. Table 1 Studied ranges and optimum values of the variables Variable Studied range Selected value Physical- Chemical - TemperaturePC PH [Borate]/mol 1-1 [ MTAMR]/mol 1 - 1 [Methanol] (96) Flow rate/ml min-1 Injected sample volume/pl Reactor length/cm FI- 20-40 25 7.0-10.0 9.0 0.05-0.5 0.1 (0.5-2.5) x 10-4 0.10 6-35 10 0.52-1.92 1.75 1 17-249 192 10-240 80 sample was injected (the maximum of the FI peak) between 530 and 600 nm provided the best results.The calibration procedure was repeated under the above- described conditions. The model was then centred about the mean values and the cross-validation method applied to five randomly chosen segments. As few as two factors accounted for 99.5% of the variance in the validation of the concentration and absorbance matrices. This suggests good correlation between the absorbance at the maximum of the FI peak and the Cu and Zn concentrations.One of the most important qualitative inferences a PLS calibration method affords, when a highly ordered experimental design, such as that employed in this work is used, is whether certain samples should be excluded from the calibration set because they contain some error. In the present instance, there was a clear relationship between scores and the concentration of each analyte, and there was not any evidence of outliers; thus, the score for the first factor was related to the concentrations of Cu and Zn, i.e., it increased with increasing concentrations of both metal ions. On the other hand, the score for the second factor bore a direct relationship to the Cu concentration and an inverse one to the Zn concentration. The proposed method was validated by applying it to a series of synthetic samples containing variable concentrations of Cu and Zn.Table 3 gives the results obtained for each analyte using the calibration matrix of Table 2. As can be seen, the amounts added and found were quite consistent, with small relative standard deviations (s,) and errors of less than 5% in every case. The mean square root of the relative error (RRMSE)lO was 2.21% for Cu and 2.95% for Zn. The absence of systematic errors in the determination was checked by subjecting the added and found amounts to the joint confidence region test for the slope and intercept, developed by Mandel and Lining. 1 1 Interferences We investigated the potential interferences of the ions that most frequently accompany Cu and Zn in real samples.A given ion was considered to interfere with the proposed method if its Table 2 Composition of the calibration matrix Added amount/pg ml-1 Sample No. 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 c u Zn 0.268 0.241 0.268 0.562 0.268 0.887 0.268 1.204 0.268 1.525 0.805 0.241 0.805 0.562 0.805 0.887 0.805 1.204 0.805 1.525 1.342 0.241 1.342 0.562 1.342 0.887 1.342 1.204 1.342 1.525 1.879 0.241 1.879 0.562 1.879 0.887 1.879 1.204 1.879 1.525 2.416 0.241 2.4 16 0.562 2.416 0.887 2.416 1.204 2.4 16 1.525172 Analyst, February 1996, Vol. 121 presence resulted in an error above 7% for a mixture containing 1.342 pg ml-1 Cu and 0.887 pg ml-1 Zn. Of the ions tested, K+, Li+, Ba2+, Ca2+, Mg2+, F-, S042-, NO3- and HP042- were found not to interfere at concentrations over 1000 times higher than those of the analytes.On the other hand, V5+, Sn2+, Mn2+, Co2+, Cd2+ and Ti4+ were tolerated at only five-fold higher concentrations than those of the analytes (Pb2+ and Ni*+ in this last ratio had no effect on the Zn determination but resulted in errors of approximately 20 and lo%, respectively, in the Cu determination, the effect disappearing at a 1 : 1 ratio). The most severe interferences were posed by Cr3+, Fe3+ and A13+, and largely arose from precipitation of their respective hydroxides at the working pH. While Cr3+ did not interfere in a 1 : 1 ratio with respect to the analyte, Al3+ and Fe3+ interfered at even lower levels. The interferences of the last two ions can be lessened by using sodium fluoride; thus, in the presence of 0.1 moll-' F-, AP+ and Fe3+ had no adverse effect on the determinations at ratios above 1000 : 1 and 1 : 1, respectively.Applications The proposed method was applied to the determination of copper and zinc in blood serum and in metal alloys. Copper plays an essential role as the activator for a number of enzymes and zinc is a constituent of some enzyme systems and an essential growth factor. Several blood serum samples, prepared as described under Experimental, were analysed. As can be seen from Table 4, the results they provided were quite consistent with those obtained by AAS, which is frequently employed for the determination of copper and zinc in this type of sample, as well as with normal levels in blood serum. The proposed method was also applied to the determination of copper and zinc in various metal alloys.As can be seen from Table 5, the results were consistent with those obtained by AAS and the certified composition of the alloys. Conclusions It can be concluded that the proposed method proves that the coupling of an FI system and a multivariate calibration model allows simultaneous determinations of several analytes to be performed, using a minimum amount of sample, and provides comparable results to those obtained when a single analyte is determined in similar samples. 12-14 However, the low selectivity of azoic dyes makes their application to the determination of metal ions in complex samples very difficult, despite their high sensitivity, because of the interference of other ions present in their matrices.Multivariate calibration allows the simultaneous quantification of several ions, the single spectrophotometric determinations of which are not possible when the analyte to be determined has strongly overlapping spectra. The proposed method highlights the potential of PLS multivariate calibration for the resolution of mixtures of metal ions by complex formation with chromogenic reagents, even when the absorption maxima for the complexes lie at very close wavelengths, the stoichiometries and formation constants of the complexes are unknown and excess of reagent absorbs in the same region as the complexes. This work was financially supported by The Autonomic Government of Canary Islands in the framework of Project 931042. ~~ Table 3 Results of multicomponent FI analysis of synthetic mixtures of Cu and Zn Zn/pg ml-1 Cu/pg ml-I Added Found rfr s Added Found f s 0.562 0.887 1.525 0.887 0.562 1.204 0.241 1.204 0.562 1.204 0.528 f 0.006 0.889 f 0.008 1.539 f 0.007 0.894 k 0.01 1 0.579 k 0.006 1.225 f 0.01 1 0.282 f 0.004 1.218 rfr 0.010 0.552 f 0.008 1.152 f 0.009 0.268 0.268 0.368 0.805 1.342 1.342 1.879 1.879 2.416 2.416 0.242 f 0.005 0.211 k 0.003 0.386 f 0.004 0.808 k 0.006 1.362 f 0.005 1.343 f.0.007 1.849 f 0.008 1.921 f 0.010 2.414 & 0.009 2.386 f 0.012 References Table 4 Simultaneous determination of copper and zinc in blood serum Metal ion found k s/pg ml-' Proposed method AAS method Sample No. c u Zn c u Zn 1 0.608 f0.020 0.680 f0.012 0.592 f.0.013 0.667 k0.013 2 0.589 f 0.019 0.488 f 0.02 1 0.601 k 0.024 0.473 k 0.024 3 0.770f0.021 0.421 f0.017 0.755 k0.018 0.408+0.018 4 0.548k0.016 0.515f0.016 0.564k0.020 0.538f0.029 5 0.662f.0.019 0.570f0.018 0.655 k0.025 0.583 f0.032 6 0.558It0.021 0.538f0.019 0.555k0.027 0.518f0.030 Table 5 Simultaneous determination of copper and zinc in metal alloys Metal ion found f s (%) Proposed method AAS method Sample Cu Zn c u Zn Alloy 1* 50.78 f.I .93 34.01 f. 1.76 5 1.14 k 1.45 34.91 k 1.07 Alloy 2+ 0.91 k0.03 4.74k0.13 0.91 k0.02 4.83f0.08 Alloy 3* 63.24 f. 2.37 14.18 k 0.56 62.45 rt 2.95 14.88 k0.63 * Alloy of Cu, Zn and Ni. Cr, 0.168; Ni, 0.024; Sn, 0.035; V, 0.012; Si, 0.047; Fe, 0.336; Ti, 0.015; Mn, 0.53; Cu, 0.95; Zn, 4.82; Mg, 2.43; Pb, 0.19; Al, 90.00%. * Brass. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Wada, H., Yamada, H., and Yuchi, A., J . Flow Inject. Anal., 1993,10, 114. Luque de Castro, M.D., and Tena, M.T., Talanta, 1995, 42, 151. Unscrambler I1 v.5.0., Computer-Aided Modelling A/S, Trondheim, Norway, 1993. Arias, J. J., JimCnez, F., and Garcia Montelongo, F., An. Quim., 1980, 76B, 452. Arias, J. J., JimCnez, F., Suarez, S., and Garcia Montelongo, F., Quim. Anal., 1984, 3, 193. Garcia Montelongo, F., Arias Lebn, J. J., and JimCnez Moreno, F., Analusis, 198 1, 9, 400. Arias, J. J., JimCnez, F., and Garcia Montelongo, F., Microchem. .I., 198 1, 26, 354. JimCnez, A. I., JimCnez, F., PCrez, J. P., and Arias, J. J., Collect. Czech. Chem. Commug, 1990, 55, 1500. Garcia Montelongo, F., Arias, J. J., and Jimenez, F., Microchem. J., 1980, 25, 410. Whitman, D. A., Seasholtz, M. B., Christian, G. D., Ruzicka, J., and Kowalski, B. R., Anal. Chem., 1991, 63, 775. Mandel, J., and Lining F. J., Anal. Chem., 1957, 29, 743. Hemindez, O., JimCnez, A. I., JimCnez, F., and Arias, J. J., Anal. Chim. Acta, 1995, 310, 53. Liu, R., Liu, D., and Sun, A., Talunta, 1993, 40, 5 11. Garcia Rodriguez, A. M., Garcia de Torres, A., Can0 Pavbn, J. M., and Bosch Ojeda, C., Talunta, 1993, 40, 1866. Paper 5/04426J Received July 6 , 1995 Accepted October 5, 1995

ISSN:0003-2654    

DOI:10.1039/AN9962100169

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 (1 73-1 76) 173 Use of Poly(ethy1ene terephthalate) Plastic Bottles for the Sampling, Transportation and Storage of Potable Water Prior to Mercury Determination* D. D. Copeland, M. Facer, R. Newton and P. J. Walker Thames Water Utilities Ltd., Spencer House Laboratory, Manor Farm Road, Reading, UK RG2 OJN Comparisons were carried out between the use of poly(ethy1ene terephthalate) (PET) plastic bottles and clear glass bottles for the storage of potable water samples for up to 10 days, prior to mercury determination. The PET bottles were found to be as suitable as glass bottles for the above purpose, while also offering significant cost saving and safety advantages. Two chemical preservatives were also compared. The recovery of mercury was significantly enhanced for both bottle types by the pre-addition of either hydrochloric acid or an acid dichromate preservative solution.The latter was preferred on safety grounds owing to its obvious colour. Long-term storage was not tested as this laboratory has a maximum five working day turnaround time for this determinand. The concentration range of interest was 0.05-1.5 pg 1-1 of mercury. Keywords: Mercury determination; potable water; poly(ethy1ene terephthalate) plastic sample bottle Introduction Conventional wisdom'4 holds that water samples for mercury determination should be taken and stored in glass bottles. Polyethylene bottles have been previously shown to be unsuitable for the p ~ r p o s e . ~ , ~ The practice at this laboratory has been to send out sealed 250 ml clear glass bottles (clear flint or borosilicate glass reagent bottles with glass stoppers obtained from Merck, Darmstadt, Germany), pre-acidified with 5 ml of concentrated hydrochloric acid ( = 11.6 mol 1 - l ) for the collection of water samples for mercury determination.Clear glass is preferred as it allows inspection of the colour formed during the bromination reaction used to convert organomercury compounds to the inorganic form required and the subsequent discharge of the excess of bromine with hydroxylammonium chloride solution, which decolourizes the solution, prior to the mercury determination: light may also help to catalyse the bromination r e a c t i ~ n . ~ . ~ ~ There are a number of disadvantages to the use of glass bottles. They are heavy and easily broken, which could expose the samplers, or customers, to concentrated hydrochloric acid, which is colourless and could be mistaken for water.They are an additional bottle type that have to be stored at operational sites. They are also relatively expensive, so have to be re-used, which involves de-labelling, washing in a laboratory washing machine and acid rinsing with a dilute solution of hydrochloric acid and potassium bromidebromate; all labour intensive and therefore costly activities. * The views expressed in this paper represent the views of the authors and do not necessarily represent those of Thames Water Utilities Ltd. For the analysis of all other metallic elements, samples are taken in 300 ml disposable plastic bottles made from poly- (ethylene terephthalate) (PET), which is a hard, clear and relatively robust material.This plastic can be recycled after use. Tests were carried out to assess these bottles for the sampling and storage of water prior to mercury determination. If suitable, the use of this bottle type would avoid the disadvantages of glass bottles described above. The procedure given in the Standard Methods for the Examination of Water and Wastewater.' recommends the use of an acid dichromate preservative for mercury. This was assessed against our existing preservative of 2% hydrochloric acid. Experimental Apparatus The determination of mercury was carried out using a P.S. Analytical Merlin Mercury Fluorescence Analyser fitted with a 48 place random access autosampler (P.S.Analytical, Sevenoaks, Kent, UK). Other apparatus consisted of Grade A calibrated flasks, measuring cylinders and calibrated autopipettes. Reagents ionization. De-ionized water. Purified by reverse osmosis and de- Tap water. Reading tap water. Hydrochloric acid. Merck (Darmstadt, Germany) AristaR Nitric acid. Merck AristaR grade, = 16.0 moll-'. Tin(m) chloride reducing agent. Prepared by dissolving 25 g of SnC12.2H20 (Merck; SpectrosoL) in 100 ml of concentrated hydrochloric acid and then made up to 1 1 with de-ionized water. Brominating solution. Prepared by dissolving 2.78 g of potassium bromate and 10.0 g of potassium bromide (both Merck; AnalaR) and making up to I 1 with de-ionized water. Hydroxylammonium chloride solution. Prepared by dissolving 12.0 g of hydroxylammonium chloride (Merck; SpectrosoL) in 100 ml of de-ionized water.Acid dichromate preservative. Potassium dichromate, 20% m/v (Merck; AnalaR) dissolved in 1 : 1 nitric acid. Mercury stock standard, 1000 mg 1-1. Merck; SpectrosoL. Mercury working standards. By serial dilution of the mercury standard, aqueous working standards were prepared in 250 ml calibrated flasks at the following concentrations: 0,0.2,0.4,0.8 and 1.2 pg 1-1. To each flask 5.0 ml of concentrated hydrochloric acid and 2.5 ml of brominating solution were added and, after mixing, the resulting solutions were allowed to stand for at least 2 h. After this period the excess of bromine was removed by the addition of a few drops of the grade, specific gravity = 11.6 moll-1.174 Analyst, February 1996, Vol.121 hydroxylammonium chloride solution. For the experiments involving the use of the acid dichromate (AD) preservative, this reagent was added to the standards at a concentration equivalent to that in the samples. For comparison 3, which measured mercury concentrations in samples of up to 1.5 pg 1-1, an additional top standard of 1.6 pg 1-1 was prepared, with the 0.2 pg 1-1 standard being omitted. Glassware cleaning solution. Prepared by carefully adding (with stirring) 100 k 2 ml of the brominating solution and 20 k 2 ml of concentrated hydrochloric acid to approximately 500 ml of purified water contained in a beaker, made up to 1 1 with purified water and mixed. Procedure Sample preparation The general sample preparation procedure was to acidify the sample to 2% by volume with concentrated hydrochloric acid (= 1 1.6 mol 1-I), add 1% v/v of the brominating solution, mix and then allow to stand for at least 4 h, which will convert any organomercury compounds to the inorganic form required for the subseqeunt analysis.3.6.7 After this period the excess of bromine was removed by the addition of a few drops of the hydroxylammonium chloride solution.Mercury determination The principle of this determination is to reduce the mercury compounds contained in the standards and the samples to the elemental form by mixing them within the instrument with the tin(xx) chloride reducing solution. The mercury, as a vapour, is then stripped from the aqueous line by a constant flow of argon, or nitrogen,* through a gas-liquid separator and then passed through a mercury fluorescence detector.Within the analytical range, the electronic output of this detector is linear to the concentrations of mercury. By calibration with the known standards, the concentration of mercury in a sample can be determined.3,6-10 Bottle comparison Comparison 1: initial feasibility test. A series of tap water samples and tap water spiked to 1.0 pg 1-1 with mercury were stored in 250 ml glass mercury bottles and 300 ml PET plastic bottles, using three different means of preservation and for different time intervals up to a maximum of 16 d. On the day of analysis samples were acidified to 2% HCl, where necessary, brominated and then analysed by the above mercury determina- tion procedure. Comparison 2: efSect of mercury concentration.Bulk solu- tions of mercury in tap water were prepared from a 1000 yg 1-l calibration standard at the following concentrations: 0, 0.5, 1 .O and 1.5 pg 1-1. Aliquots of these solutions (250 ml) were then poured into a series of plastic and glass bottles containing one of the following preservatives: (a) 5 ml of concentrated hydrochlo- ric acid. Immediately after the addition of the above mercury solutions, 2.5 ml of potassium bromidebromate was added in accordance with the ‘Blue Book’ mercury pro~cedure;~ (b) 0.5 ml of acid potassium dichromate as per the APHA procedure; ( c ) no preservative. Aliquots from these bottles were analysed for mercury on the day of preparation (day 0) and after two and 10 days. For samples prepared as per (b) and ( c ) above, 5 ml of concentrated hydrochloric acid (= 11.6 mol 1-l) and 2.5 ml potassium * Argon gives greater sensitivity.Nitrogen was used in these tests as an argon line was not available at the location of the analyser. However a NS 30 eleven-day validation exercise on this procedure, using nitrogen, demonstrated a limit of detection of 0.042 pg 1-1, which is well below that required for potable water analysis. bromidebromate solution were added approximately 4 h before analysis. Comparison 3: a comparison between the recovery of mercury from glass and PET bottles, with and without the addition of AD preservative, against time. A large volume of tap water was spiked to 1 .O pg 1- 1 with mercury, mixed thoroughly and then poured immediately into a series of 300 ml glass and PET plastic bottles.half of each bottle type being pre-acidified with 0.6 ml of AD solution. At the specified time intervals, one bottle of each type, with and without AD, was acidified with hydrochloric acid and brominated in accordance with standard procedure. After standing for approximately 2 h, a mercury determination was carried out by the normal procedure, modified by adding AD to the standards at a concentration equivalent to that contained on the samples. To check if the addition of AD had any effect on the sensitivity of the determination, one additional 0.8 pg 1-1 standard was prepared without the addition of AD. This gave a result of 0.79 pg 1-l Hg, which is not significantly different from the target value. Comparison 4: effect of the addition of AD on mercury recovery after standing.A series of PET plastic bottles were filled with tap water spiked to 1 yg 1-1 with mercury and allowed to stand for various periods of time. At the end of the stated period, 0.6 ml of AD preservative solution was added. The treated samples were allowed to stand for 2 h and then mercury was determined by the standard procedure, including acidification with hydrochloric acid and bromination. (NB the final day 10 sample was allowed to stand overnight after adding acid dichromate to see if recovery improved with time.) As controls, a further series of bottles were prepared as above, but with the pre-addition of 0.6 ml of AD preservative solution. Comparison 5: assessment of within- batch standard devia- tion for glass and PET plastic bottles.A large volume of tap water was spiked to 1.0 pg 1-1 with mercury, mixed thoroughly and then poured into a series of 300 ml glass and PET plastic bottles that had been pre-acidified with 0.6 ml of AD solution. At specified time intervals five bottles of each type were acidified with hydrochloric acid and brominated in accordance with out standard mercury analysis procedure. After standing for 2 h determination of mercury was carried out by the standard procedure, modified by the addition of acid dichromate to the standards at a concentration equal that contained in the above samples. Please note that day 3 was run twice owing to a later problem with the analyser while running routine samples; as both sets of data were valid, these have been included as separate batches in the results Table 1.Results and Discussion Results from the above comparisons are shown graphically in Figs. 1 4 and in Table 1. Comparison 1 was an initial feasibility study which suggested that there was no significant differences in loss or gain of aqueous mercury at concentrations of zero and 1 pg 1-1 in PET plastic bottles as compared with glass bottles using similar preservatives. (NB. 1 yg 1-1 is the prescribed concentration or value for mercury set by the Drinking Water Inspectorate in the United Kingdom.) Data is shown graphically in Fig. 1 . Comparison 2 was an expanded comparison at mercury concentrations of 0, 0.5, 1.0 and 1.5 pg 1-1 in tap water, with duplicate samples. The missing data on day 0 was owing to a fault in the autosampler which caused it to miss every other sample towards the end of the analytical run.Unfortunately this was not noticed until the data was inspected the following day, so could not be repeated. For given preservatives and concentra- tions there is a good correlation between PET and glass bottles. The non-acidified samples show a considerable decrease in theAnalyst, February 1996, Vol. 121 175 measured mercury concentration with time. The two preserva- tives, hydrochloric acid plus bromate and AD, give similar results within expected experimental error. Data is shown graphically in Fig. 2. Comparison 3 compared PET and glass bottles for tap water spiked to 1 pg 1-1 mercury, with and without preservation with AD, against time.A shorter time scale with more frequent measurements than in the previous comparisons was used to assess how quickly mercury was lost. Again good correlation between PET and glass bottles was obtained, and very similar mean values and standard deviations being obtained. The preserved samples show no significant loss of mercury with time, whereas the non-preserved samples do show some loss of mercury. Data is shown graphically in Fig. 3. Comparison 4 checked the efficiency of acid dichromate as a preservative against time and also tested if lost mercury could be recovered from non-acidified samples by adding acid dichromate after set periods of time. Only PET bottles were used for this comparison and the control sample bottles were pre-treated with the addition of 0.6 ml of AD.The preserved control samples showed no significant loss of mercury, but Table 1 Data from comparison 5 between replicate analysis of 1 pg 1-1 mercury samples in glass and plastic bottles using AD preservative. Results as mercury (pg 1-1) Replicate Plastic Glass Time no. bottles bottles Day 0 1 2 3 4 5 Mean 2 3 4 5 Mean 2 3 4 5 Mean 2 3 4 5 Mean Day 7 I 2 3 4 5 Mean Day 1 1 Day 3A 1 Day 3B 1 0.96 0.96 0.95 0.96 0.95 0.954 0.94 0.95 0.94 0.93 0.95 0.942 1 .oo 1.02 1 .oo 1.02 1 .oo 1.008 0.97 0.95 0.96 0.95 0.95 0.957 1.05 1.05 1.06 1.06 1.05 1.052 0.94 0.95 0.93 0.95 0.93 0.941 0.93 0.92 0.93 0.93 0.93 0.926 0.99 0.97 0.99 0.99 1 .oo 0.987 0.94 0.9 1 0.95 0.944 0.949 0.940 1.07 1.07 1.05 1.08 1.07 1.066 These results were fed into the WRC AQCSOFT validation program to obtain statistical data, which calculated out as follows: Bottle type Plastic Glass Mean 0.9826 0.97 18 s (within batch) 0.0070 0.0106 s (between batch) 0.0465 0.057 1 s (total) 0.047 1 0.0580 Calculated F 0.8857 1.3474 Estimated degrees of freedom 4.1 4.2 These results are not significantly different for 95% confidence limits.significant loss of mercury occurred from the non preserved samples, which was not recovered by the subsequent addition of AD preservative, even after overnight standing. See Fig. 4. Comparison 5 was a direct comparison between glass and PET plastic bottles, both with pre-addition of AD preservative. The data from the two bottle types was analysed statistically and 0.02 0.015 0.8 - 0.7 I ' I I 1 8 16 Day number Fig. 1 Comparison between bottle types and preservatives for samples of (a) tap water and (b) tap water spiked with mercury to 1 pg 1-1, against time.A, Glass, 2% HCl; B, PET, 2% HC1; C, glass, AD; D, PET, AD; and E, PET, no preservative. 0.06 1 I 0.05 0.04 0.03 0.02 0.01 0 0.55 0.5 0.45 0.4 0.35 =i u) 1.1 2 1 0.9 0.8 0.7 0.6 0.5 ' I 1.6 I 1 1.3 1.2 F 1.1 ::: 0 10 Day number Fig. 2 Comparison between bottle types and preservative solutions over a range of mercury concentrations for samples of (a) tap water; and spiked tap water (b) 0.5 pg 1-1; (c) 1.0 pg 1-1; and (d) 1.5 pg 1-1. A, PET + HCl/Br; B, PET + AD; C, PET + no preservative; D, glass + HCl/Br; E, glass + AD; and F, glass + no preservative.176 Analyst, February 1996, Vol. 121 found to be not significantly different.Data is shown in Table 1. Conclusions There seems to be little significant difference between PET plastic bottles and glass bottles in terms of mercury preserv- ation. Some form of acid preservation, either at the time of sampling, or as soon as possible afterwards would appear to be desirable. Although the effect was somewhat variable, possibly relating to the exact pH of the water, there appears to be a fairly rapid and significant loss of mercury from an unpreserved sample. Over a 10 d period, up to 40% of the mercury was lost 1.03 1.02 1.01 1 - 1.99 u) 1.98 1.96 1.95 1.94 1.93 1.92 1.91 - 2 1.97 Fig. 3 L 0 2 4 6 24 48 120 168 Time/h Comparison between PET and glass bottles, with and without AD preservative, for a 1 .O pg 1 - 1 mercury-spiked tap solution, against time.A, PET; B, PET + AD; C, glass; and D, glass + AD. 1.2 -4 1.1 I- 0.5 I ’ I I I t 0 3 5 10 Time/h Fig. 4 Comparison between PET plastic bottles, without (A, test) and with (B, control) AD preservative, against time for tap water spiked with mercury to 1 pg 1-I. Acid dichromate added to test bottle after standing. from unpreserved samples, both in plastic and in glass bottles. The preferred acid preservative is 20% m/v potassium dichromate dissolved in 1 : 1 nitric acid. This is the preservative recommend in the Standard Methods for the Examination of Water and Wastewater. 1 From a safety point of view the use of this preservative should be carefully controlled, but it has an obvious colour and requires a smaller volume than the alternative hydrochloric acid, or hydrochloric acidbromate preservatives.As nitric acid and potassium dichromate are hazardous materials, the AD solution will be prepared only by trained laboratory staff with appro- priate safety precautions. The required volume of this preserva- tive will be added by laboratory staff to the PET sample bottles, to which corrosive hazard warning labels will be affixed and the lids sealed, prior to dispatch to the samplers. A test was also carried out to assess if AD attacked the plastic bottles over a six week period. The preservative turned slightly cloudy over this period, probably owing to dissolution of plasticizer from the plastic, but the bottles did not appear to be weakened and analytical results were not affected. References 1 2 3 4 5 6 7 8 9 10 Standard Methods for the Examination of Water and Wastewater, APHA-AWWA-WEF, 18th edn., Washington D.C., 1992. Methods f o r the E.vaminution of Waters and Associated Materials, Department of the Environment, Standing Committee of Analysts, HM Stationery Office, London, 1978, ISBN 0-1 17-5 1326- 1. Methods for. the E.ramination of U’aters and Associated Materials, Department of the Environment. Standing Committee of Analysts, HM Stationery Office, London, 1985, ISBN 0-1 17-5190’7-3. Leermakers, M., Lansens, P., and Baeyens. W., Fresenius’ 2. Anal. Chem.. 1990, 336, 655. Bothner, M. H., and Robertson, D. E., Anal. Chenr., 1975, 47, 592. Farey, B. J., Nelson, L. A.. and Rolph, M. G., Analyst, 19’78, 103, 656. Farey, B. J., and Nelson, L. A.. Anal. Chem., 1978, 50, 2147. Griepink, B., Pure Appl. Cheni., 1984, 56, 1477. Thompson, K. C., and Reynolds, R. J., Atomic Absorption, Fluores- cence and Flame Emission SpectroscoppA Practical Approach. 2nd end., Charles Griffin, London, 1978, pp. 102-109. Thompson, K. C., and Godden, R. G., Analyst, 1975, 100, 544. Paper 5/04 765J Received July 20, 1995 Accepted Septeniher 28, I995

ISSN:0003-2654    

DOI:10.1039/AN9962100173

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Analyst, February 1996, Vol. 121 ( I 77-182) 177 Development of a Stand-alone Affinity Clean-up for Lysergic Acid Diethylamide in Urine John M. Francis and Derek H. Craston* Laboratory of the Government Chemist, Queens Road, Teddington, Middlesex, UK TWII OLY A total analysis scheme for lysergic acid diethylamide (LSD) from human urine is described. A simple ELISA technique led to the development and optimization of an affinity clean-up cartridge, resulting in high purification factors with a single combined extractionklean-up step. LSD can be measured with a straightforward HPLC-fluorescence technique, which minimizes operating complexity and process implementation time. The method has been applied to urine containing 0.5 ng ml-1 LSD, and the ability of high-affinity materials to preconcentrate a sample into a small volume should allow the working range of the procedure to be adjusted as required.Keywords: Extraction; affinity chromatography; lysergic acid diethylamide; urine; enzyme linked immunosorbent assay; high-performance liquid chromatography Introduction The analysis of lysergic acid diethylamide (LSD) from body fluids is now undergoing rapid development, corresponding to a renewed awareness of the drug's dangers, and in particular the long-term risk of recurrent effects.' However, the effective dose of LSD is so low2 that the rigorous evidence of its ingestion necessary in court cases requires large investments in state-of- the-art equipment such as HPLC-MS, as well as the extensive training and use of highly skilled personnel. There is a case to be made for an inexpensive, simple yet sensitive LSD determination method, which nevertheless affords a high degree of confidence, for applications in the medical, welfare, and occupational health sectors.The high selectivity of immunoaffinity chromatography can be utilized to simplify and combine extraction and full clean-up, allowing LSD to be measured by straightforward HPLC-fluorescence spectrometry. An established capability of good affinity clean- up (ACU) methods is the recovery of a concentrated solution of the analyte from a large sample volume,3-5 which may be termed sample focusing. Antibodies raised against small molecules such as LSD rely, for their selectivity, on the synthesis of an immunogen in which the distinguishing features of the target molecule remain unmodified and accessible.LSD has some structural similarity with other drugs producing related physiological effects, and antibodies with broader drug specificities than are generally desirable for analytical work may be produced if the common structural moiety is recognizable as part of the immunogen.6 Antibody generation is not a priori limited to parts of the molecule bearing functional groups; for instance, the multiple ring systems of steroids, which share a similar size and hydrophobic character with LSD, can produce antibodies of high a f f i n i t ~ . ~ A variety of immunogens (protein-bound *To whom correspondence should be addressed. haptens) were produced during the mid- 1970s for the develop- ment of radioimmunoassays,8-'1 which have been reviewed12.13 and which have continued in use.14-16 These are likely to be superseded for screening purposes by simpler and safer ELISA procedures in the future, and a prototype ELISA for LSD, which was used as a detection technique during the development of ACU, is described here.From the outset the ACU was designed to be simple and fully disposable, as the possibility of carry-over contamination from a positive sample was unacceptable for forensic purposes. LSD ACU directly coupled to HPLC-MS was reported in a significant paper which, however, dealt more extensively with application of similar methodology to the analysis of the medicinal drug propranolol. 17 Direct coupling necessitated column switching by the interposition of a trap column, and the ACU system was not fully disposable.Both antibody and antigen were eluted together from a support containing immobilized protein G, which adversely affected the sub- sequent analysis of the antigen. It was reported that 0.5 ng ml-' LSD in urine could be detected as the molecular ion from 4 ml of sample. Experimental Materials LSD was purchased from Alltech, Camforth, Lancashire, UK. To prepare the LSD stock solution 1 mg of LSD was dissolved in 10 ml of methanol and stored for up to 8 months in an amber glass bottle at -20 "C. Urine specimens were obtained from donors from the authors' laboratory, pooled in equal volumes where required, and stored for up to 2 d at 4 "C. Sephadex and Sepharose gels were from Phannacia, St. Albans, Hertfordshire, UK.Neat antiserum raised against LSD in rabbits (anti-LSD) was obtained via Cozart Bioscience, Abingdon, Oxfordshire, UK. Dimethyl pimelimidate dihydrochloride (DMP) was from Sigma, Poole, Dorset, UK. Other reagents were generally of analytical-reagent grade, or, where appropriate, HPLC grade. Phosphate-buffered saline (PBS) pH 7.4 was prepared by mixing: NaCl 8 g I-l, KC10.2 g 1-1, Na2HP04 1.15 g 1-l, and KHZP04 0.2 g 1-' . Preparation of LSD Affinity Gel A 9 g amount of drained Sepharose CL-4B was mixed well with 0.15 g of dry Protein A Sepharose CL-4B in 10 ml of PBS, allowed to settle in a disposable poly(propy1ene) column tube, then washed with 20 ml of PBS. A 5 ml volume of a 100-fold dilution of anti-LSD in PBS was incubated with the support for 30 min.The column was then washed with 4 ml of PBS, followed by 4 ml of DMP buffer (NaHC03, 100 mmol 1-1, adjusted to pH 9.5 with NaOH, 1 mol 1-1, then 9 g 1-1 NaCl added). A 5 ml volume of DMP solution (DMP freshly dissolved to 20 mmol 1-'in DMP buffer, then adjusted to pH 9.9178 Analyst, February 1996, Vol. 121 with 10 moll-1 sodium hydroxide) was applied to the column which was then allowed to undergo cross-linking for 30 min. A 5 ml volume of DMP blocker (ethanolamine, 200 mmol 1-l, adjusted to pH 8.0 with concentrated HCl) was passed through the column, followed by 15 ml of PBS, the gel being resuspended and mixed in the final 5 ml portion of the PBS. A 5 ml volume of an aqueous solution of sodium azide, 1 g 1-l, was applied, then the material was drained and refrigerated until needed.Affinity Clean-up of LSD A 0.36 g amount of the affinity material was mixed with about 2 nil of PBS, transferred into a disposable polystyrene column tube of a capacity of 2 ml (Pierce & Warriner, Chester, UK), packed according to proprietary instructions, and washed with 6 ml of PBS. Shortly before use, the LSD stock solution was serially diluted, as appropriate, in PBS or urine, and the desired volume was applied to the column, which was allowed to drain for 30 min. The unbound fraction was removed with 4 ml of PBS, followed by 4 ml water to remove buffer salts if elution was to be performed with an organic solvent. Elution conditions are discussed below. ELISA for LSD Polystyrene microtitre plates were coated with an LSD derivative under acidic conditions.These were stored refriger- ated until needed, then washed. For competitive ELISA, appropriate dilutions in PBS of LSD solutions and controls were added (but omitted for assay of anti-LSD recovery), followed by equal volumes of anti-LSD dilution in PBS. Standards contain- ing elution and neutralization buffers corresponding to varia- tions in the clean-up were included on the same plate when necessary. The plate was subsequently developed with perox- idase substrate solution, and the absorbance was measured at 450 nm with a Multiskan MK I1 reader (Labsystems, Helsinki, Finland). HPLC-Fluorescence A Hypersil ODS column (Jones, Hengoed, Mid-Glamorgan, UK), particle size 5 pm, length 250 mm, id 4.6 mm was used at 18 "C, 160 bar, and 1.2 ml min-* flow rate.The solvent composition of the mobile phase was 65% by volume methanol, and 35% by volume aqueous, the aqueous component being ammonium acetate, 100 mmoll-l. The remainder of the system consisted of a Waters (Stockport, UK) 712 WISP autoinjector, Waters 510 Solvent Delivery System, Waters pBondapak C18 Guard-Pak HPLC Precolumn Insert, Shimadzu (Kyoto, Japan) C-R3A Chromatopac integrator and Perkin-Elmer (Beacons- field, UK) LS-4 fluorescence spectrometer. Optical excitation and emission conditions were: wavelengths, 320 and 400 nm; and slit widths, 15 and 20 nm, respectively. For HPLC, ACU fractions were eluted with ethanol and boiled dry in glass vials. Each was reconstituted with methanol followed by sufficient ammonium acetate solution to achieve the over-all composition of the HPLC mobile phase and a total sample volume of 100 1.11.Results and Discussion Prototype LSD ELISA Although an efficient ACU for LSD was the original and primary objective of this work, previous immunoassay experi- ence'* led first to the development of an ELISA method as a sensitive assay tool and a means of characterizing the antiserum. The mean colour developed in the ELISA was proportional to anti-LSD concentration [Fig. 1 (a)], and the standard deviation 0 1 2 3 4 Relative concentration of antiserum I I I 1 100 LSD concentrationlng ml-' I I 0 50 100 Urine in sample (% v/v) Fig. 1 Characteristics of the LSD ELISA. (a) Colour intensity versus concentration of antiserum (no LSD). A relative concentration of 1 corresponds to the dilution used routinely.Error bars represent f l s , n = 8. (b) Colour intensity versus concentration of LSD. Final concentrations of LSD, after mixing with anti-LSD in the microtitre wells, are plotted. Error bars represent range, n = 2. (c) Cobour intensity versus concentration of a urine specimen containing no LSD. Squares, urine, pH 6.1, unadjusted; triangles, adjusted to pH 7.4 with NaOH. Absorbances were corrected for non-specific binding by means of controls containing saturating concentrations of LSD. Error bars represent f l standard error, n = 4; where omitted for clarity, these were of similar magnitude.Analyst, February 1996, Vol. 121 179 (s) approximately so, leading to the choice of a high dilution of antiserum for subsequent assays.The ELISA was sensitive for LSD concentrations of 0.01-1 ng ml-1 [Fig. l(b)]. The sensitivity of the ELISA allows scope for optimization to produce a screening kit (in progress), The colour developed in the ELISA was substantially less in the presence of urine; this was not a pH-related effect [Fig. l(c)]. Dilution of a urine specimen into PBS before ELISA was not in itself sufficient to alleviate the matrix effect [(Fig. l(c)]. These observations illustrate the fundamental problem to be addressed in develop- ing a marketable kit. Development of ACU A cross-linked agarose was chosen as the affinity support material19 because a simple gravity-feed to medium-pressure system was desired. A small proportion of the agarose had been coupled to protein A, the regiospecific affinity of which for many antibodies was expected to optimize the orientation of the antigen binding sites.20 This coupled agarose was blended with a similar uncoupled variety as a bulking agent, to ease the preparation of columns.LSD-specific antiserum was covalently attached to the orienting affinity sites before use. Antibody immobilization procedures can be unreliable, but the present procedure resulted in comparable LSD recoveries for each of 11 batches of affinity gel so far prepared (Table 1). The affinity gel could be stored in the refrigerator for one month in sodium azide solution without appreciable loss in capacity. The possibility that powerful eluents might harm the affinity gel was not of concern because the columns were to be used once only.However, the stability of LSD had to be considered: it was known that ergot alkaloids are light-sensitive,2l but in fact no precautions were necessary other than the avoidance of direct sunlight; also that LSD can be hydrolysed by strong alkali at 110 OC,22 but is stable at 100 "C in water.23 LSD was stable (greater than 90% recovery by HPLC) for 60 min at 18 "C in KCl (50 mmol l-I)-HCl, pH 1.9, allowing ample time for elution and neutralization of the purified fraction [carried out when necessary with Na2HP04 (0.5 mol 1- l)-NaHZPO4 (0.5 mol 1-I), pH 7.41. Good recoveries were obtainable at pH 1.9, as also with KCl (50 mmol l-')-NaOH, pH 12.2. Eventually, elution with 100% ethanol was favoured, because rapid concentration of the purified fraction at its boiling point and reconstitution in the HPLC mobile phase were then possible.Gradient elution with ethanol-water mixtures resulted in a gradual recovery at compositions ranging from 50 to 100% by volume of ethanol, without increased selectivity. The pH elution and ethanol elution both resulted in the transfer of interferences from the affinity gel to the ELISA (Table 2). In the case of alklaine elution, the interference was derived from the urine sample, and persisted even if the sample was pre-adsorbed with Sepharose CL-4B. Interferences were associated with ethanol elution even if no urine was applied to Table 1 Reproducibility of preparation of the LSD affinity gel* Standard Preparation Mean deviation of number recovery (%) recovery (%) 1 55.0 0.9 2 48.0 2.0 3 46.5 1.6 * Batches of the affinity gel prepared on three different days were used for ACU followed by ELISA (n = 4).For each of the tabulated experiments, 1.6 ng of LSD dissolved in 0.18 ml PBS was applied per 0.36 g of affinity gel, the eluent being KC1 (50 mmol l-I)-HCl, pH 2.2 Eight further batches resulted in comparable recoveries; however, one or more parameters of the clean-up were varied in those cases. the affinity gel. It is unllkely that affinity clean-up would be combined with ELISA for routine use, the latter being perceived as a rapid screening technique. The reliability of ACU-HPLC- fluorescence data was not compromised by impurities arising from elution with ethanol (see below). Nevertheless, it is preferable to eliminate non-specific binding to the affinity gel, and this could be achieved most logically by the elimination of the bulking agent Sepharose CL-4B.However, in the authors' experience, the preparation of low-volume disposable columns containing, for instance, 10 mg of affinity gel is labour- intensive. LSD Recovery Polyclonal antibodies were used for the ACU in order to achieve rapid results and to limit development costs. The mixed-affinity nature of such antibodies leads to a gradual approach to saturation with increasing loads of antigen [Fig. 2(a)]. In other words, a balance must be struck between a high recovery of the analyte and the economical use of immunochemical reagents. Additionally, matrix effects are likely to vary between samples, so that while it can be shown with confidence that an action limit for LSD in urine is exceeded, full quantification following ACU is only possible at very high materials costs. Approximately 80 k 10% recovery was obtained from ACU of 1 ng of LSD as a spike in urine and with 0.36 g of the affinity gel.The recoveries obtained were much lower than those which would be expected based on the theoretical two moles of antigen bound per mole of antibody (for immunoglobulin G). An efficient coupling ratio between the protein A-bearing support and the antiserum had been chosen after studying a range of conditions [Fig. 2(b)], so wastage of antibody was not an important limitation on antigen-binding capacity. A reduction in the recovery of small antigens from ACU owing to the urine matrix has been noted previou~ly.~7~7 The antigen-binding capacity of the affinity gel for LSD spiked in urine was 47% of that for a control experiment in which PBS was substituted for urine (standard error, 5% based on eight replicates); this matrix effect is distinct from that observed with the ELISA [Fig.l(c)], but might have similar aetiology. Fractionation of urine on a Sephadex G-25M column demon- strated that most of the ELISA matrix effect is associated with components of low M, (Fig. 3). Nevertheless, components of high M , also had such an effect, and the total effect was less than the sum of the effects derived from each M, fraction. The matrix effect therefore appears to be of a complex and non-specific Table 2 ELISA interferences arising from the ACU procedures* Neat Eluent Acid Alkali ethanol ELISA (%) tested PBS Inhibition in Not 1.1 22 P >0.1 < 0.001 P < 0.001 < 0.001 ( < 0.001) < 0.001 IJrine Inhibition in 23 7.1 (6.2) 14 ELISA (%) * ACU was performed with either PBS alone, or 0.4 ml of a urine specimen which contained no LSD.The acid eluent was KC1 (50 mmol l-l)-HCI, pH 1.9; the alkaline eluent was KCI (50 mmol l-I)-NaOH, pH 12.2. Acid and alkaline eluates were neutralized with double the eluate volume of Na2HP04 (0.5 mol I-1)-NaH2P04 (0.5 mol 1-l) pH 7.4 before ELISA. Evaporated ethanol-containing eluates were reconstituted with ethanol then PBS, 1 + 4 v/v respectively, before ELISA. Differences in colour development, between eluates and controls of matching solvent composition, were evaluated by the two-tailed Student t-test.Figures in parentheses are for ACU before which urine was passed through an equal volume of Sepharose CL-4B.180 Analyst, February 1996, Vol. 121 nature, as observed elsewhere.24 A number of ions and organic constituents of urine were tested individually at normal physiological concentrations25 by ELISA in an attempt to isolate the major contributor to the matrix effect. No one substance produced a strong effect on the colour developed (i.e., the binding of anti-LSD to the solid phase): urea (corroborating ref. 17), uric acid, KCl, NH4C1, Na2S04, and creatinine caused slight (about 10% per substance) or borderline inhibition; creatine, NH4SCN, and NaF had no appreciable effect. It should be borne in mind that some components of the matrix effect in the ELISA could be attributable to disruption of the solid phase, and therefore be irrelevant to ACU; the converse, i.e., disruption of the affinity gel support but not the ELISA solid phase, is also possible.Sample Focusing The rate of uptake of LSD by the affinity material was sufficiently great to allow a low concentration of the drug to be I I I I I 1 2 3 4 5 LSD offereding 0.1 1 10 Antiserum/pl mg-' Protein A Sepharose CL4B Fig. 2 (a) Saturation curve for the retention of LSD by the affinity gel. Several successive equal doses of LSD dissolved in 0.18 ml PBS were applied to a column based on 8 1 mg of Protein A Sepharose CL-4B. Effluent fractions (1 ml) were collected by the application of more PBS between doses, and ELISA for unbound LSD was performed. Error bars represent k 1 s, n = 8.(h) Saturation curve for the retention of anti-LSD by the affinity support material. Sepharose CL-4B containing a variable proportion of Protein A sepharose CL4B was mixed with a constant volume of antiserum, and unbound anti-LSD was subsequently determined by ELISA. Error bars represent kls, n = 7. collected from a relatively large specimen volume. For instance, ACU was performed for 0.8 ng of LSD spiked into either 0.18 or 5 ml of PBS; the two experiments resulted in similar ELISA recoveries, which were not significantly different by Student's t-test (six degrees of freedom). The focusing of very diluted LSD solutions in this way allows the detection limit to be lowered as desired, although at the expense of saturation should a more concentrated specimen be encountered.ACU-HPLC-Fluorescence HPLC-fluorescence experiments were based on a pre-existing system for the measurement of LSD from urine after solid phase extraction (SPE). A concentration of 0.5 ng ml-l LSD was measurable in standard solutions, recovery at low concentra- tions being more quantitative when the HPLC sample injection volume was increased (Fig. 4). Although the total sample 1.5 I T high low hn, M, Assay Gel filtration Neat buffer fractions urine Fig. 3 ELISA colour developed ~.~crsus M, distribution for components of urine. A 2.5 ml volume of a single urine specimen was applied to a pre- packed 9 ml Sephadex G-2SM column, and eluted with 2.5 ml portions of PBS, followed by ELISA for fractions 2-5 and the appropriate controls.Error bars represent +Is, n = 8. I 1 L 0 1 2 log (LSD concentrationlng rnl-') Fig. 4 Typical HPLC standard curves for LSD. Injection volume: 10 pl (solid line); 30 ~1 (dotted line). The LSD stock solution was diluted into the HPLC mobile phase to provide the most concentrated standard, and thence further diluted in the same to provide the rest of the standards.Analyst, February 1996, Vol. I21 181 (4 LS D I 1 Fig. 5 Fluorescence intensity versus HPLC retention time for: ( a ) neat urine spiked with LSD; ( h ) the corresponding ACU fraction; ( c ) the corresponding fraction from SPE. The LSD stock solution was diluted to a final concentration of 0.5 ng tnl-1 in a single urine specimen. ACU was performed on 2 ml of spiked urine, and the evaporated eluate was reconstituted in 100 pl of the mobile phase for HPLC (i.e., a concentration factor of 20).For SPE, 5 ml of spiked urine was processed by an in-house method using a Bond-Elut Certify LRC column (Varian, Walton-on- Thames, Surrey), again with an over-all concentration factor of 20. volume available for HPLC injections is limited by the amount of antibody used for ACU, it proved difficult accurately to reconstitute and handle ACU fractions in volumes of less than 100 pl of the HPLC mobile phase. Consequently, injection volumes of, for instance, 30 pl are feasible if the number of repeat injections to be made is limited. The retention time for LSD was 5.2 min (range 0.1 min within-day, 0.3 min over 4 4 . Fig. S(a) shows the complex and intense pattern of fluores- cent interferences characteristic of raw urine.This level of contamination can carry over to subsequent runs, and is likely to shorten the lifetime of the HPLC column. Furthermore, the identity of any peak having a retention time close to that of LSD must be open to question in an unknown sample. ACU-HPLC- fluorescence resulted in the resolution of the LSD peak from the few detectable contaminants, and the corresponding peak area was enhanced by sample focusing [Fig. 5(b)]. ACU-HPLC- fluorescence of LSD from a pool of six urine specimens gave a similar purification. SPE was carried out in parallel with ACU, using the same over-all sample-focusing factor for both methods. The fluores- cence intensity contributed by contaminants as LSD eluted from the HPLC column was actually greater following SPE than for raw urine [Fig.5(c*) I’ersus Fig. 5(a)], although the correspond- ing increase in the LSD signal intensity provided some analytical advantage. The materials costs for SPE were lower than for ACU; the choice of extraction/clean-up technique therefore depends on the particular application and on cost constraints. The composition of urine depends markedly on environ- mental and circumstantial factors;25 consequently, the possi- bility that a newly observed contaminant might be mistaken for LSD was given further consideration. In a control experiment, HPLC was performed in parallel for LSD after ACU from urine and for the unused eluent (absolute ethanol). Nearly all interferences were derived from the eluent (Fig. 6), and so could probably have been eliminated by purchasing ethanol of higher quality.Since a single urine specimen was used to obtain the data of Figs. 5 and 6, the late-eluting contaminant in Fig. 5(b) is Fig. 6 Fluorescence intensity ~’ersus HPLC retention time for: (a) the residue from evaporated ethanol; (6) the ACU fraction from LSD spiked in urine. A single urine specimen was spiked with LSD to a final concentration of 0.5 ng ml-1. ACU was performed with 2 ml of the spiked urine. The eluate was evaporated alongside an equal volume of unused eluent. and each residue was reconstituted for HPLC in 100 p1 of the mobile phase. spurious. In short, no detectable contaminant has been shown to be derived from urine in the foregoing ACU-HPLC-fluores- cence experiments; therefore current experience suggests that the risk of a false peak assignment is negligible provided contamination during the course of the analysis can be avoided.Conclusion This affinity clean-up allows LSD to be well resolved from contaminants detected by HPLC-fluorescence, so that quan- tification of 0.5 ng ml-1 LSD in urine is well within the capability of the method. The quality of the clean-up is superior to that obtained with SPE (which, however, is cheaper than affinity clean-up, given the same detection technique). Consequently, reliable results can be obtained in combination with a relatively inexpensive detection technology. Optimized for a disposable cartridge format, the affinity clean-up eliminates the risk of carry-over and is easy to introduce into a pre-existing scheme of analysis.The prototype LSD ELISA developed alongside this clean-up has sufficient sensitivity for optimization as a commercial kit (paper in preparation), although alterations will be needed to minimize the matrix effect if it is to be used with urine samples. The authors thank the LSD analysis group, and P. Norris, at the Laboratory of the Government Chemist for their contributions to this work. References 1 2 3 4 5 6 7 8 9 10 Leikin, J. B., Krantz, A. J., Zell-Kanter, M., Barkin, R. L., and Hryhorczuk, D. O., Med. Toxicol. Adverse Drug Exper., 1989, 4, 324. United Nations Division of Narcotic Drugs (Vienna), Recommended Methods for Testing Lysergide (LSDwanuaE for Use hy National Narcotics Laboratories, United Nations, New York, 1989, p.7. Hooijerink, H., Schilt, R., van Bennekom, E. O., and Huf, F. A., J . Chromatogr. B , 1994,660, 303. Farjam, A., Vreuls, J. J., Cuppen, W. J. G. M., Brinkman, U. A. Th., and de Jong, G. J., Anal. Chem., 199 I , 63, 248 1. de Frutos, M., and Regnier, F. E., Anal. Chem., 1993, 65, 17A. van Vunakis, H., Farrow, J. T., Gjika, H. B.. and Levine, L., Proc. Nutl. Acad. Sci. U.S.A., 1971, 68, 1483. Arevalo, J. H., Taussig, M. J., and Wilson, I. A., Nature (London), 1993,365, 859. Castro, A., Grettie, D. P., Bartos, F., and Bartos, D., Res. Commun. Chem. Pathol. Pharmacol., 1973, 6, 879. Taunton-Rigby, A., Sher, S. E., and Kelley, P. R., Science (Washington, D.C., 1883), 1973, 181, 165. Loeffler, L. J., and Pierce, J. V., J . Phui-m. Sci., 1973,62, 1817.182 Analyst, February 1996, Vol. 121 11 12 13 14 15 16 17 18 19 Ratcliffe, W. A., Fletcher, S. M., Moffat, A. C., Ratcliffe, J. G., 20 Harland, W. A., and Levitt, T. E., Clin. Chem. (Winston-Salem, N.C.), 1977,23, 169. 21 Landon, J., and Moffat, A. C., Analyst, 1976, 101, 225. 22 Castro, A., and Malkus, H., Res. Commun. Chem. Pathol. Pharma- col., 1977, 16, 291. 23 Twitchett, P. J., Fletcher, S . M., Sullivan, A. T., and Moffat, A. C., J. Chromatogr., 1978, 150, 73. 24 Peel, H. W., and Boynton, A. L., Can. SOC. Forensic Sci. J., 1980,13, 23. 25 Stead, A. H., Watton, J., Goddard, C. P., Patel, A. C., and Moffat, A. C., Forensic Sci. Int., 1986, 32, 49. Rule, G. S., and Henion, J. D., J. Chromatogr., 1992, 582, 103. Francis, J. M., and Craston, D. H., Analysr, 1994, 119, 1801. Narayanan, S. R., and Crane, L. J., Trends Biotechnol., 1990, 8, 12. Antibodies-A Laboratory Manual, ed. Harlow, E., and Lane, D., Cold Spring Harbor Laboratory, New York, 1988, pp. 51 1-552. Hellberg, H., Acta Chem. Scand., 1957, 11, 219. Lopatin, D. E., Winkelhake, J. L., and Voss, E. W., Mol. Pharmacol., 1974, 10, 767. Gettner, H. H., Rolo, A., and Abramson, H. A., J. Psychol., 1970,75, 35. Maxey, K. M., Maddipati, K. R., and Birkmeier, J., J. Clin. Immunoassay, 1992,15, 116. Hawk’s Physiological Chemistry, ed. Oser, B. L., McGraw-Hill, New York, 1965, pp. 1153-1209. Paper SlO6076A Received September 14, I995 Accepted October 25,1995

ISSN:0003-2654    

DOI:10.1039/AN9962100177

出版商:RSC    

年代:1996

数据来源: RSC