Chemical Suppression of Contaminant Metal Ions Using a Metastable

Direct fluorometric detection of paramagnetic and heavy metal ions at sub-amol level using an aromatic polyaminocarboxylate by CZE: Combination of pre...
0 downloads 0 Views 137KB Size
Download PDF
Anal. Chem. 2005, 77, 5332-5338

Chemical Suppression of Contaminant Metal Ions Using a Metastable State in Precolumn Derivatizing HPLC: An Ultratrace Fluorometric Detection of Al(III) Shingo Saito,*,† Tetsuya Anada,‡ Suwaru Hoshi,† and Kunihiko Akatsuka†

Department of Applied and Environmental Chemistry, Kitami Institute of Technology, Kitami, Hokkaido, 090-8507, Japan, and Sagamihara R&D Center, Shino-Test Corporation, 2-29-14, Oonodai Sagamihara, Kanagawa 229-0011, Japan

The contamination of metal ions from reagents used frequently restricts the practical detection limit of the metal ion, which itself is a source of contamination. We have found a novel solution to this problem, a chemicalsuppressing method of contaminant metal ions on a reversed-phase HPLC for Al3+ with a detection limit of 7.6 × 10-11 mol dm-3 (2.1 ng dm-3) by only adding a certain agent into all stock solutions without any preconcentration or purification steps. This technique decreases the concentration of the contaminant Al3+ originating from the reagents by more than 1 order of magnitude using selective derivatization of sample Al3+ ions to a powerful fluorescent complex at a metastable state in the precolumn chelation processes. Meanwhile, the contaminant Al3+ remains as a nonfluorescent complex with a blocking reagent in order to suppress the contamination. This selective derivatization is achieved by the accumulation of several complexation processes based on the difference of formation, dissociation, and ligand-exchange kinetics and the thermodynamics between the derivatizing reagent, the 4′,5′-geometorical isomer of calcein, and the blocking reagent, o,o′-dihydroxyazobenzene. This simple and smart HPLC system was validated through recovery tests of environmental and biological samples. Solving contamination problems is ever more urgent, yet it has proven to be difficult in analytical chemistry so far. There are few researchers who can effectively handle the necessary but timeconsuming and delicate process of purifying all of the reagents and using aqueous solutions. However, neglecting contamination problems will lead to instability and low sensitivity in detection systems. There are many processes that contribute to the contamination problem in chemical analysis: (1) reagents used, such as pH buffer and complexing reagents, (2) sampling and analytical apparatus, and (3) the atmosphere. For metal ion contamination from reagents, the major countermeasures are purification and the employment of a clean room. These are, however, complicated, troublesome, and expensive. In addition, * To whom correspondence should be addressed: (e-mail) [email protected]; (tel) +81-157-26-9411; (fax) +81-157-24-7719. † Kitami Institute of Technology. ‡ Shino-Test Corp.

5332 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

purification is a limited concept, with its degree of success being largely dependent upon the concentration of metal ion in the solutions, but there are a number of other variables that need to be taken into account as well. Therefore, a method enabling the contaminant to be much more easily suppressed from reagents is in great demand, particularly with respect to measurements of elements with large contaminant content, such as Ca2+, Mg2+, Al3+, and Zn2+. Especially, the detection of ultratrace Al3+ in drugs and dialysis solution is important in the field of pathology in terms of Alzheimer disease and amynotrophic lateral and dialysis dementia. However, ppt level detection of Al3+ in those samples is very difficult due to its contamination. High performance liquid chromatography (HPLC) is one of the most accurate and simple techniques for the detection of tracelevel metal ion.1,2 Excellent systems called “kinetically differentiation modes”3-7 have been developed for Al3+ by Yotsuyanagi et al., who have suggested that only kinetically stable complexes are detectable through the on-column dissociation process. Those systems have potential sub ppb level detection limits,8-10 occasionally restricted by contamination from the derivatizing reagents and the pH buffers used. One of these detection systems is commercially available for serum samples.11,12 In these methods the key to lower detection limit is the control of the contamination metal ion. In this paper, we introduce a precolumn derivatization HPLC method with chemical suppression of the contaminant Al3+ by assembling various kinetics and thermodynamics and by using selectively fluorescent detection of Al3+ ions in sample solutions. (1) Wang, P.; Lee, H. K. J. Chromatogr., A 1997, 789, 437-451. (2) Timerbaev, A. R.; Buchberger, W. J. Chromatogr. Libr. 1998, 60, 963. (3) Hoshino, H. Bunseki Kagaku 2002, 51, 594-604. (4) Hoshino, H.; Yotsuyanagi, T. Chem. Lett. 1984, 1445-1446. (5) Hoshino, H.; Nakano K.; Yotsuyanagi, T. Analyst 1990, 115, 133-137. (6) Kaneko, E.; Hoshino, H.; Yotsuyanagi, T.; Watabe R.; Seki, T. Bull. Chem. Soc. Jpn. 1992, 65, 3192-3194. (7) Kaneko, E.; Hoshino, H.; Yotsuyanagi, T.; Gunji, N.; Sato, M.; Kikuta, T.; Yuasa, M. Anal. Chem. 1991, 63, 2219-2222. (8) Uehara, N.; Kanbayashi, M.; Hoshino, H.; Yotsuyanagi, T. Talanta 1989, 36, 1031-1035. (9) Hoshino, H.; Saitoh, Y.; Nakano, K.; Takahashi, T.; Yotsuyanagi, T. Bull. Chem. Soc. Jpn. 2001, 74, 1279-1284. (10) Matsumiya, H.; Iki, N.; Miyano, S. Talanta 2004, 62, 337-342. (11) Sato, M.; Yoshimura, H.; Shimmura, T.; Obi, H.; Hatakeyama, S.; Kaneko, E.; Hoshino, H.; Yotsuyanagi, T. J. Chromatogr., A 1997, 789, 361-367. (12) 2 Sato, M.; Yoshimura, H.; Obi, H.; Hatakeyama, S.; Kaneko, E.; Hoshino, H.; Yotsuyanagi, T. Chem. Lett. 1996, 203-204. 10.1021/ac050380c CCC: $30.25

© 2005 American Chemical Society Published on Web 07/02/2005

Scheme 1. Energy Diagram of Precolumn Derivatization Processes

Figure 1. Chemical structure of 4′,5′-calcein.

Using this method, we have achieved low-nanogram per cubic decimeter level detection of Al3+ without any purification or preconcentration steps. In the developed derivatizing method, two reagents are employed, both of which play an important role; the 4′,5′-geometorical isomer of calcein (4′,5′-calcein or Cal, shown in Figure 1) is used as a florescent labeling compound, which forms an Al3+-calcein complex with fast kinetics, and o,o′dihydroxyazobenzene (DHAB) is used as a blocking compound of contaminant Al3+, which forms the Al3+-DHAB complex and dissociates slowly. To our knowledge, there have been no reports of a method that enables us to discriminate between the same chemical species in the sample and in the contaminant by kinetic masking using two different ligands. Reports on some methods using calcein for the fluorometric determination of metal ions13,14 have tended not to include detailed discussions of its composition or the structure of calcein complexes. Perhaps this can be explained in that commercially available calcein usually includes different geometric isomers of calcein.15-18 The large number of impurity peaks in this calcein would have made the ligand unsuitable for separation systems, such as HPLC. Indeed, our own investigation has revealed that the unpurified calcein produces an unavailable chromatogram due to its many unknown peaks. In recent years, however, the 4′,5′geometorical isomer of calcein has been purified and has been made commercially available. The purity of this ligand makes it sufficiently amenable for use as an analytical reagent for HPLC (>97%).19 Using this 4′,5′-calcein as a precolumn derivatizing HPLC, we were able to determine, for the first time, the stoichiometry of metal to ligand. An investigation about the formation reaction of Al3+-DHAB complex was reported by Yotsuyanagi et al.20 According to their work, while the formation equilibrium constants are very large (βAl-L ) 1016.4, βAl-L2 ) 1029.1), the formation rate of the DHAB complex is very slow even at pH 7.8.3,5 The reagent, DHAB, was considered best for our purposes because of its slow kinetics and large thermodynamic stability on formation of the Al3+ complex. (13) Berregi, I.; Durand, J. S.; Casado, J. A. Talanta 1999, 48, 719-728. (14) Lista, A. G.; Palomeque, M. E.; Band, B. S. F. Talanta 1999, 50, 881-885. (15) Wallach, D. F. H.; Surgenor, D. M.; Soderberg, J.; Delano, E. Anal. Chem. 1959, 31, 456-460. (16) Epsztejn, S.; Kakhlon, O.; Glickstein, H.; Breuer, W.; Cabantchik, Z. I. Anal. Biochem. 1997, 248, 31-40. (17) Galba´n, J.; Marcos, S. D.; Vidal, J. C.; Dı´az, C.; Azna´rez, J. Anal. Sci. 1990, 6, 187-190. (18) Azna´rez, J.; Galba´n, J.; Dı´az, C.; Rabada´n, J. M. Anal. Chim. Acta 1987, 198, 281-286. (19) According to Fluka, the structure of 4′,5′-calcein was confirmed by NMR spectroscopy. (20) Mizuguchi, H.; Kaneko, E.; Yotsuyanagi, T. Analyst 2000, 125, 16671671.

EXPERIMENTAL SECTION Principle. To suppress the contaminant metal ion and to form a derivatized complex with metal ion in the sample selectively, two types of ligand were employed. The labeling ligand (L) had to be able to form a fluorescent complex with the target metal ion with fast kinetics (M + L f ML; kL, fast) and the blocking ligand (B), which was added to all stock solutions used, had to be able to form a thermodynamically more stable complex than the ML complex (KMB . KML), but exhibit very slow kinetics on the formation reaction (M + B f MB; kB, slow). The energy diagram of this system is shown in Scheme 1. The addition of B in solutions of L, pH buffer, mobile phase, and dilution water was necessary to completely block the contaminant metal ions. Since the thermodynamic stability constant KB had to be the largest in those of all M species including M-L, all contaminant ions form M-B complexes in the equilibrium state after standing a long time, regardless of the slow formation rate. When sample metal ions are mixed with the labeling reagents (L) containing B, the target ions only form M-L complex due to the large rate of the complex formation. It is noted that the system temporarily settled down in a metastable state where all contaminants and sample metal ions had separately bound to the B and the L, respectively. The ligand-exchange reaction (M-L + B f M-B + L; kex, slow) had to have been very slow in order to retain the metastable state for a while. In this state, a sample of the mixture was injected to the HPLC. In addition, the kinetic stability of M-L and M-B complexes is required for the detection of the M-L complex (k-L and k-B are very small) because a spatially isolated complex from the ligand in the separation column was subjected to a strong driving force to dissociate due to the absence of the ligand in the mobile phase.3 This kind of reaction kinetics has often been shown to provide detection selectivity for certain metal ions.3-7 The kinetic and thermodynamic requirements of this system should be as follows: (1) KB . KL, (2) kL . kB, and (3) k-L, k-B and kex must all be very small. Although the principle of this technique is simple, finding a system that satisfies all the requirements is not such an easy task. This study was designed to ascertain whether the system of calcein-DHAB as a L-B combination satisfies the above requirements. Chemicals. The reagents, 4′,5′-bis[N,N′-bis(carboxymethyl)aminomethyl]fluorecein (calcein, >97% purity), and DHAB, obtained from Fluka (Buchs, Switzerland) and Dojindo Labs (Kumamoto, Japan), respectively, were dissolved in deionized water Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5333

by the Milli-Q SP. TOC. system (Millipore Co., Billerica, MA) to give a concentration of 10-2 mol dm-3. The standard solutions of metal ions were prepared by dissolving the chloride salts (99.9% purity, Wako Pure Chemical Industries, Osaka, Japan) in deionized water with a few drops of concentrated hydrochloric acid. The 0.1 mol dm-3 pH buffer solutions of 2-(N-morpholino)ethanesulfonic acid (MES)-NaOH (Dojindo Labs) and 2-amino-2(hydroxymethyl)-1,3-propanediol (Tris-HCl; Wako Pure Chemical Industries) were used for the pH range 6.0-7.5. Apparatus. Fluorescent spectra were measured using a Shimadzu model RF-1500 spectrofluorometer. The used HPLC setup consisted of a model LC10-AD pump unit, a model RF-10AXL fluorescence detector (Shimadzu Corp., Kyoto, Japan), and a Rheodyne model 7725 sample injection valve with a 20-µL loop. To avoid any serious contaminant Al3+, polyethylene and Teflon wares were exclusively employed. The analytical reversed-phase column used was a STR ODS-II from Shinwa Chemical Industries, Ltd. (Kyoto, Japan) (150 mm L × 4.6 mm i.d. packed with 5-mm particle of porous and spherical silica with 120-Å pore diameter), with the silica packing fully end capped. Procedure. The stock solutions of calcein (1 × 10-5 mol dm-3 including 0.1 mol dm-3 CHES-NaOH at pH 10), pH buffers (8 × 10-2 mol dm-3 at pH 7.5), and deionized water for dilution were prepared by adding DHAB in order to mask the contaminant Al3+. The concentration of DHAB in each stock solution was 5 × 10-6 mol dm-3. These solutions were left in the dark for 2-7 days at room temperature in order to reach an equilibrium state; i.e., all contaminant Al3+ formed complexes with only DHAB. A 10-mL aliquot of calcein and 5 mL of pH buffer solution containing DHAB were added to the sample solution and then made up to 50 mL with diluting water including DHAB. After allowing the mixture to stand for 20 min in a thermostat at 20 °C, an aliquot (∼0.5 mL) of the mixture was loaded into an injection valve (20-µL loop) for HPLC analysis. Typical HPLC conditions were as follows: the mobile-phase solution was a methanol mixture (18 wt %) containing 1 × 10-3 mol dm-3 Tris-HCl pH buffer at pH 7.5 and 1.25 × 10-7 mol dm-3 DHAB. Fluorescent detection was carried out with an excitation of 493 nm and an emission of 520 nm. None of the experiments were carried out in a clean room. RESULTS AND DISCUSSION HPLC Behavior of 4′,5′-Calcein without Blocking Agent. First, the basic behavior of Al3+-calcein complex in reversedphase HPLC was investigated without DHAB inclusion. Although a ligand-centered emission was observed at 510 nm with an excitation at 495 nm for fluorometric measurements of the Al3+calcein complex, the detection wavelength for chromatography was set at 493 and 520 nm for excitation and emission, respectively, where the signal-to-noise ratio was optimal due to interference of the scattering light. When metal-calcein complexes, which are simply formed in precolumn derivatization, were injected, a baseline separation of the Al3+ complex band from those of Zn2+, Pb2+, Ga3+, and In3+ was achieved. The bands of Zn2+, Pb2+, and Ga3+ were broad and could not be separated from the reagent peak. Meanwhile, no peak of Fe3+, Fe2+, Cu2+, Ni2+, and Co2+ was observed due to the paramagnetic quenching of the ligand-centered fluorescence. With respect to Ca2+ and Mg2+ complexes, no peak was found despite their emissive characteristics. Alkaline earth metal ions are classified in a group that generally forms kinetically active complexes in ligand-exchange 5334 Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

Figure 2. Typical chromatogram of Al3+-calcein complex in reversed-phase HPLC. Sample solution, [Al3+] ) 1 × 10-7 mol dm-3; [calcein] ) 1 × 10-6 mol dm-3; [Tris-HCl] ) 8 × 10-3 mol dm-3. Mobile phase: methanol-water 15 wt %; [Tris-HCl] ) 1 × 10-3 mol dm-3 (pH 7.5); [DHAB] ) 1.25 × 10-7 mol dm-3. Detector: λex ) 493 nm; λem ) 520 nm; sensitivity, low. Flow rate: 1.0 mL min-1.

reactions.21 Therefore, it is likely that these metal complexes decompose in the column because of their lability on dissociation reaction. Even so, judging from the detection of a distinct peak, the Al3+-calcein complex appears sufficiently kinetically stable throughout the separation process. A typical chromatogram of Al3+-calcein complex is shown in Figure 2. Although a baseline separation between Al3+ and calcein was achieved, there was some Al3+ contaminant observed corresponding to 1.6 × 10-8 mol dm-3 in the calcein and pH buffer solutions. Without the blocking agent, the detection limit of this system was 1.8 × 10-9 mol dm-3 (3σ basis on the contamination peak, n ) 9) and the linear range of 1 × 10-6 -1 × 10-8 mol dm-3 was obtained. The control of contamination is the key for more sensitive detection, though this system itself provided high sensitivity. To determine the stoichiometry of the Al3+-calcein complex, the mole ratio method was carried out using HPLC since it was very difficult to analyze the mole ratio in the batch solution, in which the fluorescence and absorbance spectra did not indicate any spectral shift and intensity alteration between the Al3+ complex and the free ligand. The plots of the peak height obtained from HPLC versus the mole ratio are shown in Figure 3. The results strongly suggest that the Al3+-calcein complex is a one-to-one complex (the value of [Al3+]/[calcein] is 1.03 at the intersection point of the two straight lines in Figure 3). This is also supported by the fact that the elution time of the Al3+ peak was the same as those of all [Al3+]/[calcein] ratios. If one-to-two complexes, Al3+(calcein)2, or any other stoichiometry complexes could form, these complex peaks would be observed at different elution times, and no band would be detected at the elution time of the one-to-one complex with the increase in ligand concentration due to the difference of hydrophobicity between each of those complexes. With respect to the chemical structure of the Al3+-calcein complex, there are two possibilities. One is that the two iminodiacetate groups bind to an Al3+ ion as a fully coordinated hexadentate complex like EDTA. The other is that the iminodiacetate and the phenolic hydroxide in the fluorecein framework bind to the metal ion as unsaturated tetradentate coordination. Whether or not structure forms is unclear in this state. (21) Lincoln, S. F.; Merbach, A. E. Advances in Inorganic Chemistry; Academic Press: New York, 1995; Vol. 42.

kL′ ) kL[calcein]0

Figure 3. Mole ratio method for Al3+-calcein complex using HPLC. [Al3+] ) 2.0 × 10-7 mol dm-3, [Cal] ) (0-8.0) × 10-7 mol dm-3, [Tris-HCl] ) 1 × 10-3 mol dm-3. Eluent, methanol-water 35 wt %, [Tris-HCl] ) 1 × 10-3 mol dm-3. Other experimental conditions are the same as for those in Figure 2.

(2)

The formation rate constant of Al3+-calcein, kL, was obtained from the slope as 2.7 × 103 s-1 M-1. Meanwhile, the formation reaction of Al3+-DHAB complex involves succeeding reactions: 1:1 and 1:2 complex formation reaction. Regarding the suppression of aluminum contaminant, the 1:1 complex formation kinetics is rather important since the 1:1 complex formation is usually the rate-determining step for succeeding reactions. A fluorometric measurement (ex 490 nm, em 575 nm) was carried out in the batch system in order to obtain the reaction curve of the 1:1 complex, because only the [Al-dhab]+ complex is a fluorescent species among all DHAB species including [Al-(dhab)2]- and DHAB.20,23 The reaction curve of [Al-dhab]+ was obtained by the time trace of a mixture of Al3+ and DHAB with a large excess of DHAB (Figure 5). Here, the rate law for the succeeding reaction is represented as

d[Al-dhab]/dt ) kB[Al][DHAB] - kB2[Al-dhab][DHAB] (3)

Figure 4. Formation reaction curve of Al3+-calcein complex using HPLC. [Al3+] ) 5 × 10-8 M, [Cal] ) 1.5 × 10-6 M, [Tris-HCl] ) 1 × 10-3 M, pH 7.5. Eluent, methanol-water 18 wt %, [Tris-HCl] ) 1 × 10-3 M (pH 7.5). Flow rate, 1.0 mL min-1. λex ) 493 nm, λem ) 510 nm. Temperature, 293 K. The solid line is the results of nonlinear least squares curve calculation.

kB and kB2 are the formation rate constants for 1:1 and 1:2 from the 1:1 complex, respectively. Considering the pseudo-first-order conditions with a large excess of DHAB and material balances,

[Al-dhab] ) aI )

kB′ [Al]0{exp[(-kB′t)] kB′ - kB2′ exp[(-kB2′t)]}

Kinetic and Thermodynamic Requirements for Precolumn Derivatization. To achieve the chemical suppression of contaminant Al3+, the kinetics of the complex formation and the ligand-exchange reaction play essential roles described in the Principle section. It is, however, difficult to determine the individual formation rate constants of the complexation processes since the reaction solution is a mixture of various Al3+ species, such as Al3+, [Al(OH)]2+, [Al(OH)2]+, [Al(OH)4]-, [Al(OH)3], [Al2(OH)2]4+, [Al13(OH)34]5+,22 and the Al3+ complex with buffer reagent, and the concentrations of each species cannot be estimated. Accordingly, the apparent rates including all aluminum species at pH 7.5 are measured for Al3+-calcein and Al3+-DHAB formation reactions and a ligand-exchange reaction of Al3+-calcein with DHAB. For Al3+-calcein formation reaction, pseudo-first-order reaction curves were obtained by the variation of peak height of Al3+calcein in HPLC against precolumn reaction time with a large excess of calcein (Figure 4). The rate law is written as

- d[Al]/dt ) kL[Al][calcein] ) kL′[Al]

(1)

The observed rate constant, kL′, was determined through fitting a first-order reaction curve, and values of kL′ linearly depended on the calcein concentration with no interception between 5 × 10-7 and 2 × 10-6 M of calcein concentration.

)

kB[DHAB]0 kB[DHAB]0 - kB[DHAB]0

[Al]0 ×

{exp[(-kB[DHAB]0)t] - exp[(-kB2[DHAB]0)t]} (4) where I, a, kB′, and kB2′ are fluorescent intensity, the correction factor, kB[DHAB]0, and kB2[DHAB]0, respectively. The nonlinearleast-squares curve calculation of eq 4 is presented as a solid line in Figure 5, which thoroughly agrees with the experimental plots. The obtained kB′ values from the calculation are practically proportional to DHAB concentrations without any significant interception as kB′ ) kB[DHAB]0. The value of the slope, kB, is 69 s-1 M-1, which is significantly smaller than that of the calcein complex (2.3% of kL). The half-lives of each of the formation reactions in our typical procedure are 4 (kLobs ) kL[calcein]0 ) 2.7 × 10-3 s-1) and 82 min (kB′ ) kB[DHAB]0 ) 1.4 × 10-4 s-1) for calcein and DHAB, respectively. Metal complex formation kinetics is generally described by the Eigen-Tamm mechanism.24 According to the mechanism, the formation rate depends on the outer-sphere reaction equilibrium constant, KOS, and the water molecule-exchange rate constant, kwater, (kf ) KOSkwater). In our case, as the metal ion is the same for each ligand, it seems that the (22) Baes, C. F., Jr.; Mesmer, R. E. The Hydrolysis of Cations; Krieger Publishing Co.: Malabar, FL, 1986; pp 112-123. (23) Kaneko, E.; Ishida, A.; Deguchi, Y.; Yotsuyanagi, T. Chem. Lett. 1994, 1615-1618. (24) Martell, A. E. Coordination Chemistry; American Chemical Society: Washington, DC, 1978; Vol. 2, p 11.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5335

Figure 5. Formation reaction curve of Al3+-dhab complex in a batch solution. [Al3+] ) 1 × 10-7 M, [DHAB] ) 1 × 10-6 M, [Tris-HCl] ) 1 × 10-3 M, pH 7.5. λex ) 490 nm, λem ) 575 nm. Temperature, 293 K. The solid line is the result of the nonlinear least-squares curve fit of eq 4.

Figure 7. Relationship between Al peak heights of blank and DHAB concentrations in pH Buffer stock solutions. Mobile phase: methanolwater 18 wt %; [Tris-HCl] ) 1 × 10-3 mol dm-3 (pH 7.5). Sample: [Cal] ) 2 × 10-6 mol dm-3, [Tris-HCl] ) 8 × 10-3 mol dm-3 (pH 7.5).

Scheme 2. Overall Reaction in Precolumn Chelation

Figure 6. Overall reaction profile of precolumn derivatization using HPLC. Sample, [Al3+] ) 1 × 10-7 M, [calcein] ) 1 × 10-6 M, [DHAB] ) 2 × 10-6 M, [Tris-HCl] ) 8 × 10-3 M, pH 7.5. Chromatographic conditions are the same as Figure 4. The solid line is the result of nonlinear least-squares curve calculation of eq 7.

charge of the ligand controls the kinetics, which has a radical effect on the KOS value.24 The reagents DHAB and calcein mainly exist as the no-charged (pKa1 ) 8.20, pKa2 ) 11.6)23 and -4 form (pKa4 ) 5.5, pKa5 ) 10.8),15,25 respectively, at pH 7.5. Although there are many reports about Al complex formation kinetics with various ligands,26-31 the kinetics are very complicated, and all of the factors involving the formation processes, such as the numbers of atoms participating in the chelate ring and the rigidity of the ligand, have yet completely elucidated. Finally, the overall reaction of a precolumn derivatization was measured to confirm that the Al-dhab complex is thermodynamically most stable at equilibrium under experimental conditions and that the ligand-exchange reaction is sufficiently slow. The reaction curve was obtained by investigating the relationship between the peak heights of the Al-calcein complexes and the reaction times of precolumn derivatization on HPLC detection (Figure 6). The complexation of Al-calcein is completed in ∼20 min for the concentrations designed in our protocol. After that, the ligand-exchange reaction into the DHAB complex takes place, which is ascertained by the decrease in the Al-calcein peak. After 50 h, the peak height of the Al-calcein complex decreases to almost null. This certifies that the Al-dhab complex is a most thermodynamically stable species. The ligand-exchange reaction (25) (26) (27) (28) (29) (30) (31)

Iritani, N.; Miyahara, T. Bunseki[ ]Kagaku 1973, 22, 174-178. O Ä ’Coinceanainn, M.; Hynes, M. J. J. Inorg. Biochem. 2001, 84, 1-12. Tomany, C. T.; Hynes, M. J. Inorg. React. Mech. 1999, 1, 137-144. Murakami, S. J. Inorg. Nucl. Chem. 1979, 41, 209-214. Perlmutter-Hayman, B.; Tapuhi, E. Inorg. Chem. 1979, 18, 875-877. Perlmutter-Hayman, B.; Tapuhi, E. Inorg. Chem. 1977, 16, 2742-2745. Secco, F.; Venturini, M. Inorg. Chem. 1975, 14, 1978-1981.

5336

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

is slow enough that the sample can be injected at the most suitable derivatizing time (20 min). To estimate the rate constant of the ligand-exchange reaction, a reaction profile was analyzed. The overall reaction is summarized in Scheme 2. Here, the back reaction, i.e., the dissociation reaction of the DHAB complex, was not taken into account because the DHAB complex was thermodynamically stable, and a large excess of DHAB over Al3+ was contained. The ligand-exchange process is actually a first-order decay curve, as shown in Figure 6. From Scheme 2, the rate law can be set up, as

d[Al-Cal]/dt ) kL[Al][Cal] - kex[Al-Cal][DHAB] ) kL′[Al] - kex′[Al-cal]

(5)

[Al] ) [Al]0 exp{-(kL′ + kB′)t}

(6)

(

[Al-cal] ) [Al]0

)

kL′ (exp{(-kL - kB′)t} kex′ - kB′ - kL′ exp(-kex′t)) (7)

The integrated eq 7 was fitted to the overall reaction curve using a nonlinear least-squares calculation (the solid curve in Figure 6). The previously obtained rate constants for kL and kB (2.9 × 103 and 69 s-1 M-1, respectively) were substituted, and kex′ was treated as a variable. The calculated line was in complete agreement with the experimental plots, with a correlation coefficient of R2 ) 0.993 and the quantitative recovery of the added sample Al3+ (94%) was comparable to that obtained from the calculation curve (92%), which strongly suggests that the measured rate constants (kL and kB) and the rate law (eq 7) are valid. The estimated kex′ value was obtained as 3.9 × 10-6 s-1; the halflife of the Al-calcein complex in the ligand-exchange process is 4.9 h. As such, it is verified that this reaction system satisfies all

Figure 8. Suppression of contaminant Al with HPLC. Left: Typical chromatograms with/without blocking agent, DHAB. Mobile phase: methanolwater 18 wt %; [Tris-HCl] ) 1 × 10-3 mol dm-3 (pH 7.5). Sample: (a) blank, [Cal] ) 2 × 10-6 mol dm-3, [Tris-HCl] ) 8 × 10-3 mol dm-3 (pH 7.5); (b) blank + DHAB, [DHAB] ) 2.0 × 10-6 and 1.25 × 10-7 mol dm-3 in sample and eluent, respectively; (c) (b) + Al, [Al3+] ) 1.0 × 10-7 mol dm-3. Other conditions are the same as (b). Detector: λex ) 493 nm; λem ) 520 nm; sensitivity, low. Flow rate: 1.0 mL min-1. Right: extent of contamination from each reagent. pH buffer + DHAB: [DHAB] ) 5.0 × 10-6 mol dm-3 in Tris-HCl buffer stock solution; calcein + DHAB: [DHAB] ) 1.0 × 10-6 mol dm-3 in calcein stock solution; eluent + DHAB: [DHAB] ) 1.25 × 10-7 mol dm-3 in eluent.

Table 1. Influence of Interfering Ions on the Determination of Al3+ a diverse ion

molar ratio to Al3+

recovery (%)

diverse ion

molar ratio to Al3+

recovery (%)

Cu2+ Ni2+ Co2+ Zn2+ Pb2+ Fe3+ Ga3+

200 250 500 2000 2000 2000 200

96.3 94.0 93.8 96.0 102 98.2 98.3

In3+ Na+ K+ Mg+ Ca+ F-

1000 100000 100000 100000 100000 10000

111 99.3 102 107 109 103

a

[Al3+] ) 5 × 10-10 mol dm-3.

Table 2. Analytical Results of Various Samples (n ) 5)a Al added (ppb)

Al found (ppb)

2.1

13.8 19.2 3.8 0.47 2.03 35.7 26.3 2.4 26.7

river water (JAC 0031) tap water K city S city normal saline solution KCl transfusion 27 human urine 27 a

recovery (%)

other method 13.4b 14.9c 5.13c

96.7 97.3 98.8

Dilution factor for all samples was 100. b Certified value. c FL-AAS.

the requirements of the kinetics and thermodynamics mentioned in the Principle section. HPLC Behavior with Blocking Reagent and Determination of Al3+. The blocking agent, DHAB, was added to each solution of pH buffers, calcein, diluted water, and eluent solutions. The DHAB concentrations in stock solutions were optimized. The concentration level of contamination in stock solutions lies within several 10-8 mol dm-3. DHAB concentrations of ∼500 times the contaminant (5 × 10-6 mol dm-3) were added to each stock

solution. It is confirmed that this amount of DHAB sufficiently suppressed the contaminant (Figure 7). By adding DHAB to each solution, the Al3+ contaminant was drastically suppressed by a factor of 18.5 in the HPLC detection. The contaminant originating from stock solutions was completely suppressed since more than 99% of the contaminant was pressed down, even when Al3+ concentrations of 5 × 10-7 mol dm-3 were added to each stock solution as a contaminant. This was also ascertained from equilibrium calculations based on the conditional equilibrium constant of Al-dhab complex.20 The extent of the contamination removal is shown in Figure 8. Judging from the peak heights, the contribution to the overall contamination from the pH buffer, calcein, and eluent solutions were 91.4, 2.5, and 0.7%, respectively. The deionized water contains hardly any Al3+ ion content (in ng dm-3 level). The remaining Al3+ contamination peak of 5.4% corresponds to 8.6 × 10-10 mol dm-3. It is likely that the remains of the contaminant peak originate from the apparatus used, which is equipped with a sample loop and pipes made from stainless steel. No fluctuation of the remaining peak was observed. This gives a view that the random contamination from dust in the atmosphere does not substantially affect peak height. When Al3+ was added to the sample, it was recovered (Figure 8) and the slope of the calibration curve in this method was almost the same, at 96%, as the case without kinetic suppression. A linear calibration curve was obtained in a concentration range between 1 × 10-10 and 1 × 10-7 mol dm-3 (a correlation coefficient of linearity, R2 ) 0.998). A detection limit (DL) of 2.1 ng dm-3 (7.6 × 10-11 mol dm-3, 1.5 fmol in amount basis) was achieved. The achieved sensitivity is a 17-fold improvement over the previous DHAB-free method. The DL of our proposed method is the most sensitive of all HPLC methods cited (SHA, 1 × 10-8 mol dm-3;8 8-quinolinol, 1.5 × 10-8 mol dm-3;11,12 sulfonylcalix[4]arenetetrasulfonate, 8.8 × 10-9 mol dm-3;10 Morin, 2 × 10-9 mol dm-3 32) and is comparable to inductively coupled plasma mass (32) Lian, H.-Z.; Kang, Y.-F.; Yasin, A.; Bi, S.-P.; Shao, D.-L.; Chen, Y. -J.; Dai, L.-M.; Tian, L.-C. J. Chromatogr., A 2003, 993, 179-185.

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

5337

spectroscopy (ICPMS). Moreover, no remarkable interference from other ions was observed, as is summarized in Table 1. The suitability of this system for practical use was also demonstrated. Several real samples were tested: human urine, normal saline, KCl transfusion, river water, and tap water solutions. These sample solutions were once acidified to pH 2 with 6 mol dm-3 hydrochloride solution and then diluted by a factor of 100. The method was verified by recovery tests or cross-checks against other methods (Table 2). This system showed good performance for those samples and was amenable to various practical samples. CONCLUSION We designed a novel ultratrace HPLC for Al3+ using a chemical reaction assembly without any preconcentration. These reactions are able to kinetically and thermodynamically control the contaminant Al3+. This method for suppression of the contaminant is very simple, requiring only the addition of a blocking agent to stock solutions. It needs to be emphasized that this technique (33) Saito, S.; Sasamura, S.; Hoshi, S. Analyst 2005, 130, 659-663. (34) Saitoh, T.; Hoshino, H.; Yotsuyanagi, T. Anal. Sci. 1991, 7, 495-497. (35) Carr, J. D.; Swartzfager, D. G. J. Ame. Chem. Soc. 1975, 97, 315-321.

5338

Analytical Chemistry, Vol. 77, No. 16, August 15, 2005

recognizes the same species but differentiates between the sample and the contaminant. Furthermore, this method could prevent the increase in peak height caused by contamination entering stock solutions after preparation. This system has a remarkably high sensitivity, comparable to instrumental analytical methods such as ICPMS, and is tough enough to be applied to samples with complex matrixes. The principle of this method could be applied to other detection systems for various metal ions, such as Zn2+and Ca2+, for which ultratrace-level detection is difficult even using instrumental analysis due to large amounts of contamination. Although metal ions such as Ca2+ and Zn2+ have a tendency to form labile complexes due to the intrinsic kinetic activity, there are several complexing systems to form inert complexes with Zn2+ 33,34 and Ca2+.35 We are continuing the research in the direction of the chemical suppression techniques for better trace detection of Ca2+ and Zn2+.

Received for review March 3, 2005. Accepted June 9, 2005. AC050380C