Anal. Chem. 1996, 68, 2277-2280
Tributyl Phosphate as a Sensitivity-Enhancing Solvent for Organotin in Carbon Furnace Atomic Absorption Spectrometry Hui Li, Benling Gong, and Kazuko Matsumoto*
Department of Chemistry, Waseda University, Tokyo 169, Japan
Tributyl phosphate (TBP) has been found to be a sensitivity-enhancing solvent for organotin compounds in graphite furnace atomic absorption spectrometry; (C4H9)2Sn(O2CCH3)2, (C4H9)2Sn(O2CC11H23)2, (C4H9)3SnCl, and (C4H9)4Sn all give 1 order of magnitude higher sensitivities in TBP than in toluene or ethyl acetate. The sensitivities are enhanced further 1-2 orders of magnitude in TBP, when PdCl2(CH3CN)2 is added as a matrix modifier in the organic solvent. Among the four organotin compounds, (C4H9)2Sn(O2CCH3)2 and (C4H9)2Sn(O2CC11H23)2 give better sensitivities than (C4H9)3SnCl and (C4H9)4Sn in the absence of palladium in any organic solvent, which suggests that the oxygen atom in the tin compound might form tin oxides that are resistant to volatilization loss during ashing. Scanning electron microscopic, electrothermal vaporization ICPMS, and powder X-ray diffraction studies show that the final products before atomization include phosphorus-containing compounds Sn2P2O7, SnP2O7, and Pd9P2, besides tin-palladium alloys, PdSn, Pd3Sn, Pd2Sn, Pd3Sn2, and PdSn3. These phosphoruscontaining compounds would more efficiently stabilize tin and suppress tin vaporization loss during ashing, to give higher sensitivity. It is known that the sensitivity of inorganic tin in graphite furnace atomic absorption spectrometry (GFAAS) is, among many other elements, strongly dependent on the matrix elements, such as chloride or oxide. This is caused by the different intermediate compounds formed during the ashing stage, depending on the matrix elements.1-4 The sensitivity of organotin is much lower than inorganic tin, and so heavily dependent on matrix elements that it is almost useless to try to measure organotin by GFAAS. Regarding inorganic tin measurement, the oxide or carbide of tin is more resistant to volatilization loss and is reported to give a higher sensitivity than chloride.5 As is widely known, organotin is determined only with very poor sensitivity.6 We have reported that the sensitivity of organotin can be increased by the addition of PdCl2(CH3CN)2 as a matrix modifier in organic solvents and by employing a longer ashing (1) Campbell, W. C.; Ottaway, J. M. Talanta 1974, 21, 837-844. (2) Barnett, W. B.; Mclaughlin, E. A., Jr. Anal. Chim. Acta 1975, 80, 285296. (3) Sturgeon, R. E.; Chakrabarti, C. L.; Langford, C. H. Anal. Chem. 1976, 48, 1792-1807. (4) Rayson, G. D.; Holcombe, J. A. Anal. Chim. Acta 1982, 136, 249-260. (5) Wendl. W.; Mu ¨ ller-Vogt, G. Spectrochim. Acta 1984, 39B, 237-242. (6) Parks, E. J.; Blair, W. R.; Brinckman, F. E. Talanta 1985, 32, 633-639. S0003-2700(95)01082-1 CCC: $12.00
© 1996 American Chemical Society
time.7,8 Considering the fact that the sensitivity of inorganic tin is improved by using oxidizing agents, such as Cr2O72-, MoO42-, or WO42-,9-11 it is expected that the sensitivity of organotin might also be improved by using other matrix modifiers that contain oxygen in the molecules and promote oxide formation. With this point in mind, the effect of several organic solvents having various numbers of oxygen atoms in a molecule was examined. PdCl2(CH3CN)2 was used as a modifier,12 and the sensitivities of tin for several organotin compounds were compared. The effect of oxygen atoms in the solvent was considered for toluene, ethyl acetate, and tributyl phosphate (TBP), and TBP was found to give remarkably high sensitivity.13 In the present study, the enhancing mechanism is studied by electrothermal vaporization ICPMS (ETV-ICPMS), scanning electron microscopy, and powder X-ray diffraction. EXPERIMENTAL SECTION Apparatus. A Hitachi Z-9000 atomic absorption spectrophotometer was used with pyrolyzed graphite tubes. The atomic absorption measurements were carried out with the Sn atomic line of 224.6 nm and Zeeman background correction. The Ar purge gas flow was stopped during the atomization stage. The heating conditions were as follows: dry, 50-120 °C for 40 s, 120200 °C for 20 s; ash, 200-600 °C for 20 s, 600-1000 °C for 10 s; atomization, 2700 °C for 7 s; and cleaning, 2800 °C for 3 s. The ETV-ICPMS was a Yokokawa Analytical Systems Hp-4500 ICPMS spectrometer equipped with ETV 2000 from the same company. In the ETV furnace a carbon tip with dimensions of 60 × 3.0 × 1.7 mm was placed to make the atomization environment more similar to the carbon furnace of AAS. The heating conditions for ETV were the same as those for AAS. The operating conditions of the ICP were as follows: rf power 1.2 kW, rf matching 1.8 V, sampling depth 5.5 mm, carrier gas (Ar) flow rate 1.2 L/min. The solution amount used in the AAS measurement was 10 µL, whereas 1 µL was used in the ETV-ICPMS measurement. The organopalladium modifier, PdCl2(CH3CN)2, was added as a toluene solution before measurement, to the final Pd concentration of 10 µg/mL. The detection limit was determined as the absolute amount corresponding to twice the signal of the background. (7) (8) (9) (10)
Katsuta, T.; Kato, F.; Matsumoto, K. Anal. Sci. 1990, 6, 909-910. Matsumoto, K.; Katsuta, T.; Kato, F. Anal. Sci. 1991, 7 (suppl.), 451-454. Taga, M.; Yoshida, H.; Sakurada, O. Bunseki Kagaku 1987, 36, 597-600. Pinel, R.; Benabdallah, M. Z.; Astruc, A.; Astruc, M. Anal. Chim. Acta 1986, 181, 187-193. (11) Wendl, W.; Mu ¨ ller-Vogt, G. T. J. Anal. At. Spectrom. 1988, 3, 63-66. (12) Katsura, T.; Kato, F.; Matsumoto, K. Anal. Chim. Acta 1991, 252, 77-81. (13) Li, H.; Gong, B.; Ochiai, T.; Matsumoto, K. Anal. Sci. 1993, 9, 707-709.
Analytical Chemistry, Vol. 68, No. 13, July 1, 1996 2277
Table 1. Relative Sensitivities of Organotin Compounds and the Detection Limits in Organic Solvents detection limitb (ng)
enhancement ratio Sn compound (C4H9)2Sn(O2CCH3)2 (DBTDA) (C4H9)2Sn(O2CC11H23)2 (DBTDL) (C4H9)3SnCl (TBTC) (C4H9)4Sn (TeBT) Sn (1 M HCl)
Pda
Oa
+ + + + + -
+ + + + + +
toluene
ethyl acetate
TBP
1200 55 1100 61 1000 3 1200 1
1200 66 1400 88 1200 4 1200 1
2500 578 2500 406 3400 27 2600 3
water
8400 1500
toluene
ethyl acetate
TBP
0.019
0.013
0.003
0.041
0.009
0.003
0.047
0.030
0.003
0.015
0.027
0.001
water
0.011
a +, Pd added or oxygen present; -, Pd not added or oxygen not present. b The detection limit is defined as the concentration corresponding to 2 S/N.
The scanning electron microscope (SEM) was the Horiba EMAX-2200 equipped with a Hitachi S570 X-ray analyzer. All the X-ray measurements were carried out under vacuum. To each TBP solution of (C4H9)4Sn, (C4H9)3SnCl, and (C4H9)2Sn(O2CCH3)2 with a tin concentration of 500 µg/mL, an organopalladium solution (100 µg of Pd/mL) was added; 10 µL of the mixed solution was injected 20 times to the carbon furnace, which was heated to 600 or 1000 °C. The furnace surface of the solution injection site was measured. The X-ray diffraction of the scratched powder of the carbon furnace used for SEM was measured on a Rigaku RADΙΙC diffractometer with monochromated Cu KR (1.540 50 Å) radiation. Reagents. The organotin compounds used in the experiment were tetrabutyltin [(C4H9)4Sn (TeBT)], tributyltin chloride [(C4H9)3SnCl (TBTC)], dibutyltin diacetate [(C4H9)2Sn(O2CCH3)2 (DBTDA)], and dibutyltin dilaurate [(C4H9)2Sn(O2CC11H23)2 (DBTDL)]. All these reagents were used as purchased and were dissolved in toluene to 1000 µg of Sn/mL. The organopalladium standard solution was prepared by dissolving PdCl2(CH3CN)2 into a small amount of acetone. Each standard solution was diluted to appropriate concentrations with the solvent used in the experiment.
RESULTS AND DISCUSSION Sensitivity of Organotin Compounds in Various Solvents. The sensitivities and the detection limits were determined for the organotin compounds in three different organic solvents, and the result is summarized in Table 1. The sensitivity of inorganic tin with and without Pd(NO3)2 addition in aqueous solution is also shown in Table 1 for comparison. The relative sensitivities are expressed in reference to TeBT without a modifier (whose relative sensitivity is set to 1) and are determined as the relative slopes of the calibration lines. The results indicate that the sensitivities for organotin compounds containing oxygen atoms are, as had been expected, higher than those of others that do not contain oxygen atoms, when organopalladium is not added. The low sensitivity for TeBT and TBTC would be due to their volatile nature; the existence of oxygen in DBTDA and DBTDL gives better sensitivity, probably because tin oxide is formed during ashing. When organopalladium is added, the sensitivities are enhanced by 100-1000 times to almost an equal sensitivity. 2278 Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
Both in the presence and in the absence of Pd, the sensitivities in toluene and ethyl acetate are almost comparable, whereas in TBP, the sensitivities are remarkably enhanced for all of the compounds. The reason for this large enhancing effect of TBP would be, for one thing, the presence of oxygen in TBP, which probably leads to tin oxide formation during the ashing stage and suppresses tin volatilization loss. However, other factors may also be operative in such a distinct sensitivity enhancement. The details about the enhancement mechanisms will be discussed later together with the results of other physical measurements. It should be noted that TBTC gives a remarkably high sensitivity in TBP, compared to the other three compounds when Pd is added. This would probably be due to the chloride in TBTC, which is easily substituted by TBP. From these results, TBP seems to be a favorable solvent for organotin. Result of ETV-ICPMS Measurement. ETV-ICPMS would be a suitable tool to examine when the vaporization loss occurs during the heating process and what elements remain on the carbon furnace until the final step of the atomization. In order to confirm that the high sensitivity in TBP and the enhancing effect of the organopalladium are also observed in ETV-ICPMS to the extent comparable to GFAAS, the relative sensitivities of the organotin compounds were compared with ETV-ICPMS, which showed a result almost identical to that of GFAAS (Table S1, supporting information). The MS spectra for the Sn, O2, P, and Pd channels were measured for all the combinations of tin compounds and solvents. Figure 1 shows the MS spectra observed at a Sn 118 channel for two organotin compounds in toluene and TBP without Pd addition. The other two organotin compounds, TBTC and DBTDA, show spectra similar to TeBT and DBTDL, respectively. The MS spectra for organopalladium addition are shown in Figure 2. Note that, in the absence of Pd, tin loss is observed for both of the tin compounds during ashing in both solvents. These tin losses almost disappear in the presence of Pd (Figure 2). If these ashing and atomization processes are observed at O2, P, and Pd channels, the results are as those shown in Figures 3 and 4. Oxygen and phosphorus evolves gradually all throughout the heating process with two distinct peaks. The first peaks in the O2 and P spectra during the ashing process correspond to the temperature region where Sn volatilization is also observed in the Sn MS spectra (Figure 1). The second peaks of O2 and P correspond to the atomization
Figure 1. ETV-ICPMS spectra of TeBT and DBTDL observed at Sn 118 for toluene and TBP solutions without organopalladium addition. The tin concentration was 10 µg/mL. (a) Toluene, TeBT; (a′) toluene, DBTDL; (b) TBP, TeBT; (b′) TBP, DBTDL. Figure 3. ETV-ICPMS spectra of TeBT and DBTDL observed at O2 and P channels for toluene and TBP solutions without organopalladium addition. The tin concentration was 10 µg/mL. (a) Toluene, TeBT; (a′) toluene, DBTDL; (b) TBP, TeBT; (b′) TBP, DBTDL.
Figure 2. ETV-ICPMS spectra of TeBT and DBTDL observed at Sn 118 for toluene and TBP solutions with organopalladium addition. The tin concentration was 1.0 µg/mL, whereas that of palladium was 10 µg/mL. (a) Toluene, TeBT; (a′) toluene, DBTDL; (b) TBP, TeBT; (b′) TBP, DBTDL.
peak of Sn. These facts suggest that the final Sn product(s) formed just before atomization contain phosphorus as well as oxygen. With organopalladium addition, the Pd peaks in the MS spectra appear only at the atomization stage, suggesting that the final Sn compound(s) also contain palladium. Result of SEM Observation. The surface morphology of the carbon furnaces applied with the TBP solutions of TeBT, TBTC, and DBTDA was examined after heating to either 600 or 1000 °C. By heating to 600 °C, the surface of the furnace shows sticky residues whose constituent elements were found to be Sn, Pd, P, and O, as measured with an X-ray microanalyzer. It was also confirmed that the four elements exist on an identical spot of the surface. The surface morphology and the elements detected in the residue were almost same for the three organotin compounds. The relative intensities of the four elements were always in the order Pd > P > Sn > O for the three tin compounds. This order changed when the carbon furnace was heated to 1000 °C; the order was P > Pd > Sn > O for the three tin compounds, and the SEM pictures of the furnace surface for the three compounds
show existence of small particles, on which P, Pd, Sn, and O are distributed, associated with each other. The increase of the P concentration on increasing the temperature from 600 to 1000 °C implies that tin forms compound(s) with the phosphorus of TBP on heating to 1000 °C. The SEM picture of a furnace heated to 1000 °C is shown in Figure 5 for TBTC. The existence of the four elements at an identical spot suggests that tin might be present before atomization as phosphate salts and alloys with palladium. Result of the X-ray Diffraction (XRD) Study. The surface of the carbon furnace was scratched and was subjected to powder X-ray diffraction analysis. Three alloys, PdSn, Pd3Sn, and Pd3Sn2 were detected for TeBT and DBTDA with Pd addition after heating to 600 °C (Table S2, supporting information). On the other hand, the alloys detected were different for TBTC, for which only Pd3SnC0.5 was observed. In the case of DBTDA, pyrophosphate salt SnP2O7 and palladium phosphide Pd9P2 were detected as well as the three Sn-Pd alloys. The higher sensitivity of DBTDA in TBP must be due to pyrophosphate and phosphide formation. When the ashing temperature was raised to 1000 °C, Pd2Sn was observed for TeBT, and PdSn3 was observed for TBTC, in addition to PdSn, Pd3Sn, and Pd3Sn2 observed for ashing at 600 °C (Table S3, supporting information). At 1000 °C, pyrophosphates SnP2O7 and Sn2P2O7 were observed not only for DBTDA but also for TeBT and TBTC. For comparison, XRD measurement was carried out for the same three tin compounds in toluene with organopalladium addition. The result showed that only PdSn, Pd3Sn, and Pd3Sn2 were formed for all the three compounds. From the above experiments, it is concluded that the high sensitivity of organotin in TBP with organopalladium addition would result from the formation of tin pyrophosphate as well as of tin-palladium alloys. The present study has identified various products formed during the ashing stage when palladium is used as a matrix modifier. Although Pd is widely used as a modifier for other volatile elements, only a few ashing products have been confirmed Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
2279
Figure 5. SEM picture of a carbon furnace applied with a TBP solution of TBTC and heated to 1000 °C.
Figure 4. ETV-ICPMS spectra of TeBT and DBTDL observed at O2, P, and Pd channels for toluene and TBP solutions with organopalladium addition. The tin concentration was 1.0 µg/mL, whereas that of palladium was 10 µg/mL. (a) Toluene, TeBT; (a′) toluene, DBTDL; (b) TBP, TeBT; (b′) TBP, DBTDL.
with sufficiently direct evidence; for germanium measurement with Pd(NO3)2 and Mg(NO3)2 addition, Ge9Pd23, GePd2, and MgGeO3 have been found by X-ray diffraction.14 In lead measurement with (NH4)2H2PO4 addition, Pb3(PO4)2, Pb8P2O13, PbO, and Pb were detected, while with Pd(NO3)2 addition, Pd, Pd3Pb2, and Pd3Pb were detected by X-ray diffraction.11 For tin measurement with Pd(NO3)2 addition in aqueous solution, the formation of Pd3Sn2, Pd2Sn, and Pd3Sn has been claimed from the discussion of the experimentally obtained Sn atomic vapor temperature and the PdSn phase diagram.15 Van Loon reported the existence of Pd8Sn3 apparently without any supporting evidence.16 In our previous (14) Xuan, W. K. Spectrochim. Acta 1992, 47B, 545-551. (15) Oishi, K.; Yasuda, K.; Hirokawa, K. Anal. Sci. 1991, 7, 883-887. (16) Brzezinska-Paudyn, A.; Van Loon, J. C. Frezenius’ Z. Anal. Chem. 1988, 331, 707-712. (17) Gong, B.; Li, H.; Ochiai, T.; Zheng, L. T.; Matsumoto, K. Anal. Sci. 1993, 9, 723-726.
2280 Analytical Chemistry, Vol. 68, No. 13, July 1, 1996
experiment with X-ray diffraction for inorganic tin, Pd3Sn2 was identified when both Pd(NO3)2 and serum were added to tin in aqueous solutions, whereas no Pd-Sn alloy was detected when only Pd(N03)2 was added.17 This fact, together with the result of the present experiment for organic solution, suggests that organic materials seem to help the formation of the alloys. Considering these literature discussions, the present study is the first to directly verify the existence of various Pd-Sn alloys and tin pyrophosphates by X-ray diffraction. The organic solvent TBP is highly useful for organotin measurement, and the mechanism of such an unexpected sensitivity enhancement might involve the formation of tin pyrophosphates. Finally, the authors add the comment that the present results were obtained by using several hundred times larger amounts of tin than that used for usual AAS measurement. The results obtained here must be considered with the sample size effect when the results are applied to real sample analysis. SUPPORTING INFORMATION AVAILABLE Relative sensitivities of organotin compounds as observed with ETV-ICPMS and the result of the X-ray diffraction (6 pages). Ordering information is given on any current masthead page. Received for review October 30, 1995. Accepted April 8, 1996.X AC951082Q X
Abstract published in Advance ACS Abstracts, May 15, 1996.
