ARTICLE pubs.acs.org/ac
Iron Oxide Nanomatrix Facilitating Metal Ionization in Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry Rofeamor P. Obena,†,‡ Po-Chiao Lin,†,§ Ying-Wei Lu,†,§ I-Che Li,^ Florian del Mundo,‡ Susan dR. Arco,‡ Guillermo M. Nuesca,‡ Chung-Chen Lin,§ and Yu-Ju Chen*,†,^ †
Institute of Chemistry, Academia Sinica, Taipei, Taiwan Institute of Chemistry, University of the Philippines Diliman, Quezon City, Philippines § Department of Chemistry, National Tsing Hua University, Hsinchu, Taiwan ^ Department of Chemistry, National Taiwan University, Taipei, Taiwan ‡
bS Supporting Information ABSTRACT: The significance and epidemiological effects of metals to life necessitate the development of direct, efficient, and rapid method of analysis. Taking advantage of its simple, fast, and high-throughput features, we present a novel approach to metal ion detection by matrix-functionalized magnetic nanoparticle (matrix@MNP)-assisted MALDI-MS. Utilizing 21 biologically and environmentally relevant metal ion solutions, the performance of core and matrix@MNP against conventional matrixes in MALDI-MS and laser desorption ionization (LDI) MS were systemically tested to evaluate the versatility of matrix@MNP as ionization element. The matrix@MNPs provided 20- to >100-fold enhancement on detection sensitivity of metal ions and unambiguous identification through characteristic isotope patterns and accurate mass (100-fold (Supporting Information, Table 2s, columns 12 15). In terms of the number of ionizable metals, 17 21 metals were detected by SiO2@MNP and matrix@MNP compared to only 8 11 metals by the commercial matrixes and LDI. DHB@MNP performed best by detecting all of the 21 metals (100%), followed by the SiO2@MNP (20 metals), SA@MNP (19 metals), and CHCA@ MNP (17 metals). By the estimation on the ion counts of each metal ion, moreover, the signal of each metal was enhanced 20 to >100 times in DHB@MNP compared to the rest of the matrix@MNP combinations. Taken together, a signal enhancement was clearly achieved for these metals with SiO2@MNP or matrix@MNP. Influence of Laser Intensity and Ionization Potential. To gain insight into the ionization mechanism of the MNP-assisted MALDI MS in metal detection, the influence of laser intensity on metal ionization was studied using copper as a model (Figure 3). In Figure 3a, the Cu ion intensity using core Fe3O4 MNP (control) and matrix@MNP was shown as a function of laser intensity. As expected, the increase in laser intensity substantially enhanced the signal intensity of the Cu ions. Under identical laser power (Figure 3a), the nanoparticles facilitated the ionization of Cu in the following order: DHB@MNP > core MNP (control) > SA@MNP > CHCA@MNP. Among the four MNPs, the use of DHB@MNP showed the steepest curve with minimum threshold laser intensity, indicating that the conjugation of DHB matrix on MNPs provided enhanced ionization for Cu compared to the naked core MNP. Conversely, the surface conjugation of either CHCA or SA did not improve the ability of the core MNP to facilitate Cu ionization even at higher laser power. Such ionization ability of DHB@MNP might have been caused by a more efficient absorption and transfer of energy to the metal analyte by surface-bound DHB than CHCA and SA molecules. Furthermore, the role of ionization potential (IP) in metal ionization was investigated via a comparison of metal ions having
ARTICLE
different IPs: Cs (3.89 eV), Ba (5.21 eV), Ga (6.00 eV), Co (7.88 eV), Cd (8.99 eV), and Zn (9.39 eV) using DHB@MNP (Figure 3b). The ionization tendency curves clearly revealed that the IP is the critical determinant to influence the metal ionization process. Interestingly, three different patterns were observed for metals with low IP (3 4 eV, e.g., Cs), moderately high IP (5 6 eV, Ba and Ga), and high IP (>7 eV, Co, Cd, and Zn). Among the metals, Cs had the lowest IP (3.89 eV), which was almost equivalent to one photon of laser energy (3.68 eV, N2 laser; 3.49 eV Nd:YAG laser) and required the least laser threshold intensity. Metals with moderately high IPs (Ba, 5.21 eV; Ga, 6.00 eV) required higher thresholds (5000 5500), and those with high (Co, 7.88 eV; Cd, 8.99) and very high (Zn, 9.39 eV) IPs needed laser intensities beyond 5500. Thus, both the laser intensity and the IP of metal analyte were critical factors affecting metal ionization by the iron oxide type of matrix-conjugated nanoparticles. Metal Ion Detection in Tap Water Sample. To evaluate the practicability of matrix@MNP to detect metal ion in real samples, 100 μL tap water samples spiked with various amounts of Cu (5, 2.5, and 1.25 μmol) (Figure 4a c) were prepared. One microliter each of these solutions was taken and mixed with DHB@MNP (1000 ppm in ddH2O), spotted on plate, dried, and analyzed by MALDI MS. As shown in Figure 4, strong peak intensities were observed in all spectra in a concentrationdependent manner (2.5- to 10-fold decrease in signal). The current absolute detection limit was estimated to be 0.5 μmol (≈318 mg L 1) Cu in tap water (inset in Figure 4c). The United States Environmental Protection Agency (USEPA) recommends 1.3 mg L 1 Cu as maximum contaminant level (MCL) in drinking water.26 Thus, for practical application of the proposed approach to quantitative trace metal analysis, a preconcentration step was necessary. The present results simply demonstrated the applicability of the nanoprobe-assisted MALDI MS approach to metal ion analysis in real samples.
’ DISCUSSION Ionization Mechanism of Metal Detection by NanoparticleAssisted MALDI MS. Nanomaterials, such as carbon nanotubes,15,16
gold nanoparticles (AuNPs),19 silicon nanoparticles,27 titanium dioxide nanoparticles (TiO2NPs),21 and silver nanoparticles (AgNPs),18 have been employed as laser desorption matrixes to improve the MALDI MS analysis of small molecules. The use of matrix@MNP has been previously reported as a background-free matrix for the soft ionization of a variety of small molecules.23,24 To our best knowledge, this study is the first report for metal ion detection by MALDI MS, as most studies on ionization by laser desorption/ablation are usually performed via tandem LAICPMS.8 11 Despite the advantages of these nanomaterials to efficiently induce desorption and ionization in MALDI MS, there has been a lack of literature describing their mechanism of absorption and transfer of energy to the analyte. Although various ionization models have been proposed to nicely explain the MALDI phenomena, previously reported results are based on the use of a free matrix for the analysis of small molecules or biomolecules. In nanomaterials, the intrinsic size,28,29 and quantum30,31 and electronic properties32 may account for their unique ability to absorb and transfer energy. For instance, in desorption ionization on silicon (DIOS), efficient energy absorption and transfer is provided by the porosity and pore size of the silicon surface.12 Thus, 9340
dx.doi.org/10.1021/ac2017184 |Anal. Chem. 2011, 83, 9337–9343
Analytical Chemistry
ARTICLE
Figure 4. MALDI MS spectra of Cu ion spiked in tap water (100 μL). (a) 5, (b) 2.5, and (c) 1.25 μmol Cu. Inset in (c) shows the current detection limit at 0.5 μmol Cu (S/N = 10.8). All spectra were obtained at 5200 laser intensity. / denotes background-derived peak.
Figure 5. Proposed mechanism of nanoparticle-mediated MALDI process for metal ionization. (a) Formation of metal ligand ion pair and coordination or electrostatic interaction of metal ion with salicylate on DHB or silane on MNP surface in solution. (b) Condensation of metal MNP complex after solvent evaporation. (c) Absorption, conversion (to heat) and pooling of energy by the DHB@MNP and core MNP from the laser. (d) Desorption of metal ion from the surface-bound DHB molecule into the plume as gaseous metal ion. M+ L represents metal ligand ion pair in solution.
the ionization/desorption of analyte on the surface-functionalized MNP may proceed through a nanoparticle-dependent energy absorption. In contrast to protonation-induced ionization by a free matrix,33 the lack of protonated or salt-adducted metal ions also suggests a mechanism different from that found in conventional MALDI MS. As shown in this study (Supporting Information, Table 2s), the efficiency of a free matrix to induce metal ionization was only comparable to the LDI method, which suggested that the energy absorbed and pooled by the unconjugated matrix was insufficient for desorption and ionization from the DHB-metal crystallization. In addition to the presence of matrix molecules, the most salient features of the nanoparticle-assisted MALDI MS lie in the core particle, coating composition, and size effect,23,24 all of which may contribute to the absorption, transfer, and pooling of energy during the ionization/desorption process. Significantly, the UV-absorption of the matrix was greater compared to silane and iron oxide in the MALDI region (355 nm) (Supporting
Information, Figure 3s; CHCA and SA not shown). The data indicated that the core MNP and silane might contribute in the pooling and dissipation of energy as heat, which in turn may provide tremendous energy to overcome the ionization threshold of the metal. Therefore, a model to rationalize the metal ionization mechanisms in nanoparticle-assisted MALDI MS was proposed and discussed below. In MALDI, a direct multiphoton ionization process is the most straightforward explanation for ion formation.34 The detection of metal ions by LDI in this study may be explained by desorption and ionization via low-energy electrons (DIET) from the surface of metal plates35 in a mechanism similar to the photoelectric effect.36 Direct photoionization of the metal is also probable because most of the metals (the alkali, Ba, Al, Ga, Ti, Mn, and Ni), except Pt (Supporting Information, Table 2s), have IPs almost equivalent to one-photon (3.72 eV for 337 nm; 3.53 eV for 355 nm) and two-photon (7.44 eV for 337 nm; 7.06 eV for 355 nm) energy employed in MALDI. Compared to the IPs of 9341
dx.doi.org/10.1021/ac2017184 |Anal. Chem. 2011, 83, 9337–9343
Analytical Chemistry the commonly used matrixes, DHB (8.05 eV),37 CHCA (8.5 eV), and SA (8.7 eV),38 the ionization thresholds of most metals used in this study were lower for the generation of cationic metal ions. Though Pt had a high IP value (8.96 eV, >2 photon energy) and a three-photon ionization yield was expected to be inefficient, the platinum ion was observed in LDI with relatively high intensity and good resolution (Table 2s, also see Supporting Information, Figure 4s). Pt itself is a thermally conductive element, which might synergistically combine with the absorption of the twophoton energy to overcome its high IP. However, the weak signals obtained in LDI and the free matrix suggest the inefficient desorption/ionization of the metal by these two conventional methods (Figure 1a d). In contrast, the observed strong intensity and characteristic isotope distribution for most metal ions in the presence of silane-coated Fe3O4 (MNP) or its matrix-conjugated derivative (matrix@MNP) suggest that the functionalized iron oxide nanoparticle could play an important role in the desorption and ionization of the metal (Figure 1e h), which might stem from the superior energy absorption contributed by exciton pooling in the metal oxides (Fe3O4 and SiO2) and immobilized organic matrix molecules. On the basis of the above inferences, we propose a matrix@MNP-mediated ionization mechanism of the metal ions, exemplified by DHB@MNP, shown in Figure 5. When the metal salt was mixed with the nanomatrix solution, it separated into ion pairs (metal cation ligand anion pairs) upon solvation. Subsequently, the salicylate or carboxylate group on DHB may have chelated the metal cation through coordination39 and electrostatic interactions40 (step 1, Figure 5a). Similarly, the metal coordination may also occur between the metal ion and the partially negative charged silane on the MNP surface.41 Condensation of the matrix@MNP metal complex allowed the preconcentration of the metal analytes around the matrix@MNP (step 2, Figure 5b) by carboxylate- or salicylate-type coordination. Upon irradiation (step 3, Figure 5c) at the UV-MALDI region (355 or 337 nm), the DHB molecule absorbed the laser energy, and the excited DHB molecules may have pooled the energy needed to yield a highly excited matrix molecule within the nanoparticle, thereby causing the whole matrix@MNP to heat up. The MNP, which was now the center of the clustered analytes, could serve as a reservoir and mediator for the efficient absorption and transfer of energy; this unequivocally caused the excitation and simultaneous dissociation ionization of the gaseous metal ion into the plume (step 4, Figure 5d). The presence of excitons or delocalized excited states, a type of energy mobility in the solid state,42 on these metal oxides (Fe3O4 and SiO2),43,44 could enable the matrix@MNP to absorb and generate tremendous amounts of energy,45,46 which when coupled with quantum size and large surface area,47 might have provided the ability to store enormous amounts of energy.48 For the silane-coated MNP, instead of providing a carboxylate or salicylate for metal binding in step 2, the metal analyte might have been “sandwiched” between two hydroxyl oxygen atoms of the silane on the MNP surface. Although both core iron oxide and silane MNPs have lower UV-absorption abilities than the surfacebound matrix molecule, their combined absorption and energy pooling effect (step 3) may efficiently provide energy to overcome the IP and release the gaseous metal ion into the plume. It has been reported that the energy absorption capacity of silica-coated organic ligand-modified iron oxide was greater than that of uncoated iron oxide nanoparticles.49 Thus, the presence of silane or the organic matrix may allow changes in the band structure of
ARTICLE
the iron oxide, leading to an enhancement in optical properties46 and in the absorption and energy transfer ability of the MNP. When the matrix molecules were conjugated on the silanecoated MNP, they converged and concentrated around the MNP surface forming a network around the MNP. Such a network may lead to more efficient pooling of energy than that in the free matrix, which has been reported to enhance energy transfer and ionization.48 Among the three matrixes used in this study, DHB showed a significant improvement in ionization capability, while the CHCA@MNP and SA@MNP showed comparable efficiencies to the core MNP. The observed differences (Supporting Information, Table 2s, columns 13 15) were caused by the inherent chemical structure of the conjugated molecules. DHB has a hydroxyl group ortho to the carboxylate group, which may serve as binding site via either the carboxylate-type or the salicylate-type coordination.39,40 Because of the distant positions of the hydroxyl group, CHCA and SA could only coordinate with the metal analyte via the carboxylate group. However, the cyanogroup on CHCA was a good electron-withdrawing group, which made the carboxylate oxygen a poor electron-donor compared to SA. Thus, of the three matrix@MNPs, CHCA@MNP had the least improvement on the metal ion concentration effect, which in turn may have reduced the energy pooling effect. Role of IP and Lattice Energy on Metal Ionization. The metals’ IP critically influence the metal ionization process, on the basis of the observations in Figure 3 and Table 2s (Supporting Information). In DHB@MNP, the IP correlated well with the experimental result, especially for fifth period elements: Y > Zr > Mo; group IIB: Cd > Zn; and lanthanides: Ce > Gd. The detection of some metals (e.g., Cd, Ce, Cu, Gd, Mo, Zn, Zr) with moderate (5 6 eV) to high (7 8 eV) IPs using the matrix@MNP (with no detection in LDI or with commercial matrixes) may have been attributed to the energy pooling from the nanoparticle and its surface silane and matrix. In the solid state, the lattice energy (LE) could play a role in the desorption and ionization tendency of metals. To investigate their correlation, either the tabulated25 or estimated lattice energy of the metal salts based on the Kapustinskii equation50 are presented in Supporting Information, Table 2s. For alkali and alkaline earth elements, the metal ion intensity decreased with an increase in LE for all the MALDI methods described here (Supporting Information, Table 2s). However, there was no general conclusion for the relationship of LE with ionization compared to the rest of the metal ions. The lack of correlation may be due to the discrepancy between the theoretical and empirical values of LE.
’ CONCLUSIONS In this study, matrix functionalized iron oxide nanoparticles were employed to facilitate the unambiguous detection of metal ions through their isotope pattern and accurate mass by MALDITOF MS technique. Compared to conventional MALDI or LDI methods, iron oxide nanoparticles demonstrated a superior ionization capability to generate discrete peaks, reduced signal suppression, and increased abundance of desorbed/ionized metals over a wide range. Moreover, on the basis of the influence of laser power and matrix selection on the nanoparticle surface with respect to the ionization potential (IP) of the metal analytes, we proposed a model wherein matrix@MNPs could serve as a preconcentrator, absorber, and reservoir of energy in matrix@ MNP-assisted MALDI MS. Further efforts will be made to 9342
dx.doi.org/10.1021/ac2017184 |Anal. Chem. 2011, 83, 9337–9343
Analytical Chemistry understand the mechanistic implications of the direct involvement of MNPs in the transfer of laser energy to a metal analyte. With further functionalization of the MNPs for enrichment of metal ions, the present matrix@MNP-assisted MALDI MS method offers an efficient, simple, and direct alternative tool to analyze trace amount of metals having significant roles in plant and animal life.
’ ASSOCIATED CONTENT
bS
Supporting Information. Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Address: Institute of Chemistry, Academia Sinica 128 Academia Road, Section 2, Nankang, Taipei, Taiwan. E-mail: yjchen@ chem.sinica.edu.tw.
’ ACKNOWLEDGMENT This work was supported by the Academia Sinica (Research Project on Nanoscience and Technology), National Tsing Hua University, and National Science Council of Taiwan. The authors would like to thank Dr. Mitch Chiang for measuring the magnetic susceptibility of the nanoparticles, and Miss Chiu-Yun Chen for helping prepare the manuscript. ’ REFERENCES (1) Rojas, E.; Herrera, L. A.; Poirier, L. A.; Ostrosky-Wegman, P. Mutat. Res. 1999, 443, 157–181. (2) Sanz-Medel, A. Analyst 2000, 125, 35–43. (3) Olesik, J. W. In Inorganic Mass Spectrometry: Fundamentals and Applications; Barshick, C. M.; Duckworth, D. C.; Smith, D. H., Ed.; Marcel Dekker, Inc.: New York, 2000; pp 67 158. (4) Bacon, J. R.; Greenwood, J. C.; Van Vaeck, L.; Williams, J. G. J. Anal. At. Spectrom. 2003, 18, 955–997. (5) Andrews, L.; Rohrbacher, A.; Laperle, C. M.; Continetti, R. E. J. Phys. Chem. 2000, 104, 8173–8177. (6) Koumenis, I. L.; Vestal, M. L.; Yergey, A. L.; Abrams, S.; Deming, S. N.; Hutchens, T. W. Anal. Chem. 1995, 67, 4557–4567. (7) Song, K.; Cha, H.; Kim, D.; Min, K. Bull. Korean Chem. Soc. 2004, 25, 101–105. (8) Arrowsmith, P. Anal. Chem. 1987, 59, 1437–1444. (9) Moenke-Bankenburg, L.; Schumann, T.; G€unther, D. J. Anal. At. Spectrom. 1992, 7, 251–254. (10) G€unther, D.; Horn, I.; Hattendorf, B. Fresenius J. Anal. Chem. 2000, 368, 4–14. (11) G€unther, D.; v. Quadt, A.; Wirz, R.; Cousin, H.; Dietrich, V. J. Microchim. Acta. 2001, 136, 101–107. (12) Peterson, D. Mass Spectrom. Rev. 2007, 26, 19–34. (13) Wei, J.; Buriak, J. M.; Siuzdak, G. Nature 1999, 399, 243–246. (14) Trauger, S. A.; Go, E. P.; Shen, Z. J.; Apon, V.; Compton, B. J.; Bouvier, E. S.; Finn, M. G.; Siuzdak, G. Anal. Chem. 2004, 76, 4484–4489. (15) Xu, S.; Li, Y.; Zou, H.; Qiu, J.; Guo, Z.; Guo, B. Anal. Chem. 2003, 75, 6191–6195. (16) Ren, S. F.; Zhang, L.; Cheng, Z. H.; Guo, Y. L. J. Am. Soc. Mass Spectrom. 2005, 16, 333–339. (17) Zhang, H.; Cha, S.; Yeung, E. S. Anal. Chem. 2007, 79, 6575–6584. (18) Hua, L.; Chen, J. R.; Ge, L. Y.; Tan, S. N. J. Nanopart. Res. 2007, 9, 1133–1138.
ARTICLE
(19) Su, C. L.; Tseng, W. L. Anal. Chem. 2007, 79, 1629–1633. (20) Saino, T.; Sato, H.; Yamamoto, A.; Nemoto, A.; Tomimura, M.; Tao, H. Anal. Chem. 2007, 79, 4827–4832. (21) Yuan, M.; Shan, Z.; Tian, B.; Tu, B.; Yang, P.; Zhao, D. Microporous Mesoporous Mater. 2005, 78, 37–41. (22) Shan, Z.; Han, L.; Yuan, M. J.; Deng, C. H.; Zhao, D. Y.; Tu, B.; Yang, P. Y. Anal. Chim. Acta 2007, 593, 13–19. (23) Lin, P. C.; Tseng, M. C.; Su, A. K.; Chen, Y. J.; Lin, C. C. Anal. Chem. 2007, 79, 3401–3408. (24) Tseng, M. C.; Obena, R. P.; Lu, Y. W.; Lin, P. C.; Lin, P. Y.; Yen, Y. S.; Lin, J. T.; Huang, L. D.; Lu, K. L.; Lai, L. L.; Lin, C. C.; Chen, Y. J. J. Am. Soc. Mass Spectrom. 2010, 21, 1930–1939. (25) Lide, D. R. E. CRC Handbook of Chemistry and Physics, 85th ed.; CRC Press; Boca Raton, FL, 2005. (26) United States Environmental Protection Agency (USEPA). Drinking Water Contaminants. National Primary Drinking Water Regulations (http://water.epa.gov/drink/contaminants/index.cfm#1); (accessed May 12, 2011). (27) Wen, X.; Dagan, S.; Wysocki, V. H. Anal. Chem. 2007, 79, 434–444. (28) He, Y. P.; Miao, Y. M.; Li, C. R.; Wang, S. R.; Cao, L.; Xie, S. S.; Yang, G. Z.; Zou, B. S.; Burda, C. Phys. Chem. Rev. B 2005, 71, 125411. (29) Hill, N. A. Phys. Rev. Let. 1995, 75, 1130–1134. (30) Barnakov, Y. A.; Scott., B. L.; Golub, V.; Kelly, L.; Reddy, V.; Stokes, K. L. J. Phys. Chem. Solids 2004, 65, 1005–1010. (31) Shalchian, M.; Grisolia, J.; Assayag, G.; Ben Coffin, H.; Atarodi, S. M.; Claverie, A. Applied Phys. Lett. 2005, 86, 163111-1–163111-3. (32) Cherepy, N. J.; Liston, D. B.; Lovejoy, J. A.; Deng, H.; Zhang, J. Z. J. Phys. Chem. B 1998, 102, 770–776. (33) Zenobi, R.; Knochenmuss, R. Mass Spectrom. Rev. 1998, 17, 337–366. (34) Knochemuss, R. Analyst 2006, 131, 966–986. (35) Menzel, D.; Gomer, R. J. Chem. Phys. 1964, 41, 3311–3328. (36) Frankevich, V. E.; Zhang, J.; Friess, S. D.; Dashtiev, M.; Zenobi, R. Anal. Chem. 2003, 75, 6063–6067. (37) Karbach, V.; Knochenmuss, R. Rapid Commun. Mass Spectrom. 1998, 12, 968–974. (38) Hoteling, A. J.; Nichols, W. F.; Giesen, D. J.; Lenhard, J. R.; Knochenmuss, R. Eur. J. Mass Spectrom. 2006, 12, 345–358. (39) Kiss, T.; Kozlowski, H.; Micera, G.; Erre, L. S. Polyhedron 1989, 8, 647–651. (40) Carrell, C. J.; Carell, H. L.; Erlebacher, J.; Glusker, J. P. J. Am. Chem. Soc. 1988, 110, 8651–8656. (41) Smith, R. L.; Pietrix, D. J. Anal. Chem. 1984, 56, 610–614. (42) Dexter, D. L.; Knox, R. S. Excitons; John Wiley & Sons: New York, 1965. (43) Tsai, T. E.; Griscom, D. L. Phys. Rev. Lett. 1991, 67, 2517–2520. (44) Hashimoto, T.; Yamada, T.; Yoko, T. J. Appl. Phys. 1996, 80, 3184–3190. (45) Glinka, Y. D.; Lin, S. H.; Hwang, L. P.; Chen, Y. T.; Tolk, N. H. Phys. Rev. B 2001, 64, 085421. (46) Aliev, F. G.; Correa-Duarte, M. A.; Mamedov, A.; Ostrander, J. W.; Giersig, M.; Liz-Marzan, L. M.; Kotov, N. A. Adv. Mater. 1999, 11, 1006–1010. (47) Scholes, G. D.; Rumbles, G. Nat. Mater. 2006, 5, 683–696. (48) Setz, P. D.; Knochenmuss, R. J. Phys. Chem. A 2005, 109, 4030–4037. (49) Chen, L. X.; Liu, T.; Thurnauer, M. C.; Csencsits, R.; Rajh, T. J. Phys. Chem. B 2002, 106, 8539–8546. (50) Kapustinskii, A. F. Q. Rev. Chem. Soc. 1956, 10, 283–294.
9343
dx.doi.org/10.1021/ac2017184 |Anal. Chem. 2011, 83, 9337–9343
