Multinuclear NMR and DFT studies of the structure and fluxionality for MIII–ethylenediamine-tetraacetate complexes (M(EDTA)−, M = Al, Ga and In) in solution

Journal of Molecular Liquids 131–132 (2007) 72 – 80 www.elsevier.com/locate/molliq

Multinuclear NMR and DFT studies of the structure and fluxionality for M III –ethylenediamine-tetraacetate complexes (M(EDTA) − , M = Al, Ga and In) in solution Róbert Jószai a , Mihály Purgel a , Imre Pápai b , Hisanobu Wakita c , Imre Tóth a,⁎ a

Department of Inorganic and Analytical Chemistry, University of Debrecen, Debrecen, Hungary b Chemical Research Center of the Hungarian Academy of Sciences, Budapest, Hungary c Fukuoka University, Fukuoka, Japan Available online 8 December 2006

Abstract Multinuclear NMR studies of M(EDTA)− complexes (M = Al, Ga and In) by 1H, 13C and 27Al NMR in solution while varying the temperature, pH or using different solvents have been done. The features of the high-field NMR spectra show that the complexes have similar structure in solution and solid state, i.e., the Al(III) and Ga(III) are octahedrally coordinated by the hexadentate EDTA ligand, whilst In(III) is sevencoordinated in a trigonal prismatic fashion by a hexadentate EDTA and one water molecule. M(EDTA)(OH)2− complexes (M = Ga, In) have also been detected, pK values in D2O are 6.08 ± 0.06 and 9.17 ± 0.07, respectively. In the case of the octahedral Al(III) and Ga(III) complexes, the fluxional rearrangement of the chelate rings can be followed by 1H NMR, but In(H2O)(EDTA)− does not show fluxionality for structural reasons. The exchange between the axial and equatorial acetate arms of Al(EDTA)−, located at different N-atoms of the ligand, could be a protoncatalyzed reaction assisted with water molecule(s). Density functional theory (DFT) calculations have been carried out to reveal the effect of protonation and the role of solvent. © 2006 Elsevier B.V. All rights reserved. Keywords: EDTA; NMR; DFT; structure; fluxionality; AI; Ga; In

1. Introduction Interest in metal–aminopolycarboxylate complexes dates back several decades [1]. Fueled by the need for analytical chemistry and different applications in industry and medicine, the equilibria of the complexes including group 13 elements (M = Al, Ga, In and Tl) have been studied [2]. The stability constants of the parent complexes, M(EDTA)− are in the range of log K = 16–37, and protonated M(HEDTA) and mixed M (EDTA)(OH)2− complexes are also known. Migration of metal ions, especially in the case of EDTA-containing, radioactive waste waters, has renewed the research interest in the chemistry of this classical ligand [2–4]. The transport of metal complexes in the environment, i.e., the interaction of these species with the components of soil (for example by surface complexation) ⁎ Corresponding author. E-mail address: [email protected] (I. Tóth). 0167-7322/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.molliq.2006.08.030

might be related among others to the hydration and the fluxionality of dissolved complexes. The dilemma of the actual denticity of the potentially hexadentate EDTA ligand and the coordination number of the M (III) ions (M = Al and Ga) in solution is discussed in several papers [5–9], although the molecular structure shows distorted, octahedrally coordinated Al, Ga by hexadentate ligand in the isostructural MIMIII(EDTA)·2H2O solid crystals (MI = K, NH4, MIII = Al, Ga) [8,10,11], with no water molecules in the inner sphere of the central ions. In the Na[In(EDTA)(H2O)]·2H2O crystal, [12] the structure features seven-coordinated In3+ and hexadentate EDTA4− ligand. High-resolution NMR spectroscopy is excellent for studying the symmetry of complexes in solution, but there is an obvious need to take into account the intra-molecular isomerization/ fluxionality of the complexes, because fast rearrangement of the donor atoms (including the water) in the inner sphere might apparently increase the symmetry. The aim of this paper is to

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

73

reinvestigate the M(EDTA)− systems (M = Al, Ga and In) in solution by 1H, 13C and 27Al NMR while varying the temperature, pH or using different solvents to change the rate of the fluxionality. Structures and relative stabilities of various forms of Al(EDTA)− involved in carboxylate exchange processes have also been examined using density functional theory calculations. 2. Experimental procedure 2.1. Materials Al(NO3)3·6H2O, Na2H2EDTA, NaOH, In(NO3)3·6H2O (Reanal, reagent grade) came from commercial sources. GaCl3 and solutions were prepared from 99.99% Ga (Aluminium Work, Ajka, Hungary). KAl(EDTA)·2H2O was supplied by Dojin (Japan).

Fig. 2. 400 MHz 1H NMR spectrum of 0.1 M NaAl(EDTA) in D2O (20%)– DMSO-d6 (80%) at 281 K. (The signal at 2.65 ppm belongs to the DMSO solvent.)

2.2. Preparation of the complexes Solid complexes NaAl(EDTA)·2H2O, NaGa(EDTA)·2H2O and Na[In(EDTA)(H2O)]·2H2O were crystallized from about 0.1 mol dm−3 solutions (containing EDTA and M(III) in a 1:1 molar ratio at pH = 4) by adding an excess (at least four volumes) of ethanol. The solid was filtered, washed with 96% ethanol and dried in air. 1H NMR spectra showed the signals of only the complexes. 2.3. pH measurements The electrode was calibrated [13a] as pH = −log[H+]; therefore, the derived constants were stoichiometric constants. In D2O, pD was calculated as pHreading + 0.4 [13b]. The speciation diagram was calculated using the Medusa program [14]. 2.4. NMR measurements 1

H NMR spectra were recorded at 200, 360, 400 and 500 MHz. 13C NMR spectra were recorded at 50, 90, 100 and

125 MHz using Bruker AM200, Avance 360, AM400, DMX500 spectrometers using 5 mm inverse probes (except the Avance 360 with QNP probe) in locked mode. Typical acquisition parameters for 1H (and 13C) NMR spectra were flip angle ∼ 5 (13) μs (30°), pulse repetition time 0.2–0.5 (15) s, spectral window 210–1100 (25,250) Hz, number of scans 8–32 (256– 8000). Some 13C NMR spectra were collected with power gated or, for quantitative purposes, with the inverse gated decoupling pulse sequence. The chemical shifts are reported in ppm relative to TMS* (δ = 0.00 ppm) as an external standard. Spectral analyses were done using the Bruker WIN-NMR software. The temperature of the probe heads was checked by the methanolthermometer method [15]. 104.23 MHz 27Al NMR spectra were recorded on a JEOL 400 instrument. The chemical shift refers to 0.01 M AlCl3 solution containing 0.01 M HCl, δ = 0 ppm. Some 400 MHz 1H NMR spectra were also recorded using this equipment. 2.5. DFT calculations The structures of the investigated complexes have been fully optimized at the B3LYP/6-31G* level of DFT, where B3LYP refers to the applied exchange-correlation functional [16–18] and 6-31G* is the standard split-valence basis set. The relative energies reported in the paper have been obtained at the same level of theory. The calculations have been carried out using the Gaussian 03 software package [19]. 3. Results and discussion 3.1. NMR study of M(EDTA)− complexes

Fig. 1. The temperature dependence of 400 MHz 1H NMR spectra in a sample of 0.1 M NaAl(EDTA), (self pD ≈ 4, D2O solvent). The temperature values from the top to the bottom are 323, 301, 283 and 273 K.

3.1.1. Al(EDTA)− complex The free EDTA ligand shows two singlet 1H NMR signals, one of the acetate arms (8H) and one of the ethylenic group (4H), and three 13C NMR signals, assigned to the already mentioned groups and another to the carboxylate carbons. However, 1 H and 13C NMR spectra of Al(EDTA)− show different pattern at different fields indicating the presence of exchange processes, i.e.,

74

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

Fig. 3. Distribution diagram of Al(III)-containing species at different pH in aqueous solution [2], cAl = cEDTA = 0.05 M.

the line shape depends on the actual NMR time scale. Changing the temperature affects the rate of the exchange, resulting in a different line shape, as one can see in Fig. 1 showing typical 400 MHz 1H NMR spectra of Al(EDTA)−. At 0 °C, the 1H NMR signal of the ethylenic protons is split into two broad peaks and they are almost lost in the base line. The separation indicates the non-equivalence of these H-atoms. At 10 °C these signals are in coalescence, and at 28 and 50 °C this resulting peak becomes narrower as it should be for time averaged signals in the fast exchange regime. The acetate methylene protons show an AB quartet at 50 °C for the first look, although one can count 6 peaks of 8, i.e., two partially overlapping sets of AB quartets might be assigned. According to Day and Really, the magnetic nonequivalence of the –CH2– COO−, i.e., –CHAHB–COO− H-atoms can be explained by the long-lived metal–N bond compared to the actual NMR time scale [20]. The two sets for both the ethylenic H-atoms (detectable at lower temperature) and acetate methylene H-atoms (detectable at higher temperature) are in agreement with the solid structure of the MIAl(EDTA)·2H2O consisting of two acetate

Fig. 4. 400 MHz 1H NMR spectra of 0.1 M KAl(EDTA) (Dojin) at I = (0.6− cNaOH) M NaCl in D2O with 0 M NaOH (pD= 3.7, bottom trace), 0.04 M NaOH (pD = 5.9, medium trace) and 0.1 M NaOH (pD = 7.1, top trace, not comparable in intensity), at 28 °C. (The narrow peak at 3.38 ppm belongs to an impurity.)

Fig. 5. 104.23 MHz 27Al NMR spectra of 0.1 M KAl(EDTA) at I = (0.6 − cNaOH) M NaCl in D2O with 0 M NaOH (pD = 3.7, bottom trace), 0.04 M NaOH (pD = 5.9, bottom medium trace), 0.06 M NaOH (pD = 6.4, medium trace), 0.08 M NaOH (pD = 6.6, upper medium trace) and 0.1 M NaOH (pD = 7.1, top trace, not comparable in intensity) at 28 °C.

groups in the equatorial plane together with two N-donors and two axial acetate groups [8,10]. The strange shape, the broadening at 10 and 0 °C, might be attributed to the decreased exchange rate for the spin–spin coupled signals in the fast exchange regime, i.e., the two pairs of acetate arms are transformed to four non-equivalent ones, but the freezing of the sample prevents us from doing experiments at lower temperatures in aqueous solutions, vide infra. 13 C NMR spectra recorded at 100 MHz in the same aqueous sample at similar temperature values indicate broadening of the signals for both –CH2–(COO−) and –(CH2)–COO− carbon atoms, but no separation of the peaks can be achieved prior to freezing. A dimethyl-sulphoxide (DMSO), 80%–water, 20% mixture is often used in NMR to extend the range of variable temperature measurements below 0 °C. Moreover, the DMSO is known as a less polar and stronger donor solvent compared to water, and the ligand exchange reactions usually have a lower rate in DMSO compared to water [21,22]. The separation of the 13 C NMR signals of both –CH2–(COO−) and –(CH2)–COO− carbon atoms has already been observed in DMSO-d6 in the

Fig. 6. 104.23 MHz 27Al NMR spectra of NaAl(EDTA)·2H2O in D2O at 298 K (0.1 M, bottom), in DMSO (0.1 M, bottom medium), in methanol (saturated, c b 0.1 M, upper medium) and in CH3CN (saturated, c b 0.1 M, top).

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

75

Fig. 9. 400 MHz 1H NMR spectrum of 0.04 M NaIn(EDTA) in D2O (20%)– DMSO-d6 (80%) at 298 K.

Fig. 7. 400 MHz 1H NMR spectra of 0.1 M NaGa(EDTA) in D2O (20%)– DMSO-d6 (80%) at different temperatures.

case of Al(EDTA)− and Ga(EDTA)− at 125 MHz [8]. A typical 1 H NMR spectrum is shown at Fig. 2. (A detailed temperature dependence study of 1H and 13C NMR spectra is under progress in our laboratory.) In fact, the Al(EDTA)− complex does not exist exclusively in aqueous solution: protonation at lower pH and hydrolysis at higher pH may occur. Typical distribution curves can be seen in Fig. 3, showing that the HAl(EDTA) protonated complex is present at pH b 5 and the formation of Al(EDTA)(OH)2− starts at pH N 4. The coexistence of these species makes the quantitative separation of their kinetic contributions to the fluxional processes rather difficult. Nevertheless, to get an insight into the

Fig. 8. pD dependence of the 1H NMR chemical shift of H atoms in acetate arms (+, centre of AB) and ethylenic group (×) in 0.1 M aqueous Ga(EDTA)− solution recorded at 200 MHz, 231 K. The lines represent curves fitted by the flowing parameters: pK = 6.05 ± 0.06, δet = 3.26 ± 0.003 ppm and 3.13 ± 0.003 ppm, and pK = 6.13 ± 0.06, δac = 3.75 ± 0.003 ppm and 3.63 ± 0.003 ppm for Ga(EDTA)− and Ga(EDTA)(OH)2−, respectively.

pH dependence of the kinetics, aqueous samples at different pD values have been measured. Typical 1H NMR spectra are shown in Fig. 4. It is clear that the splitting of the ethylenic signal is complete at pD = 7.1 even at 28 C; thus, the kinetics are much slower at pD = 5.9 and 7.1 compared to pD = 4, where cooling to 0 °C is needed to get a similar line shape, see Fig. 1. This experimental finding might be attributed to the decreased contribution of the proton-catalyzed exchange pass. The catalytic role of protons in dissociation or central ion exchange reactions of metal–aminopolycarboxylates is well documented in the literature, e.g., Brücher et al. [23]. Therefore we have carried out DFT calculations to rationalize the structure and energetics of possible

Fig. 10. pD dependence of the 1H NMR chemical shift of H atoms in acetate arms (+, centre of AB) and ethylenic group (×) in 0.1 M aqueous In(EDTA)− solution recorded at 200 MHz, 231 K. The lines represent curves fitted by the flowing parameters: pK = 9.17 ± 0.06, δet = 3.03 ± 0.003 ppm and 2.92 ± 0.003 ppm, and pK = 9.18± 0.07, δac = 3.44 ± 0.003 ppm and 3.36± 0.002 ppm for In(EDTA)− and In(EDTA)(OH)2−, respectively.

76

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

protonated intermediates involved in the fluxional rearrangement of Al(EDTA)−, vide supra. The formation of Al(EDTA)(OH)2− takes place in the pH range of pKAl(EDTA) = ±2, i.e., according to the reaction, Al(EDTA)− + H2O ⇌ Al(EDTA)(OH) 2− + H+ , K = [Al(EDTA)(OH) 2− ][H+ ]/ [Al(EDTA) − ], pK = 5.8, see Fig. 3 [2]. One can ask if Al(EDTA) (OH)2− has or has not an Al(III) entity with similar symmetry as in Al(EDTA)−. 27Al NMR shift and line width are good indicators for the symmetry [24]. Similar solutions as used for 1H NMR (at pD = 3.6, 5.9, 6.4, 6.6 and 7.1) have been measured by 27Al NMR. The signals shown in Fig. 5 are only shifted slightly to high field by increasing pD, but the substantial broadening means lowered symmetry, in accordance with the 1H NMR spectra. There is no signal for Al(OH)4− at 80 ppm; thus, complete dissociation of Al(EDTA)− is negligible at our experimental condition. (In a sample at 0.05 M NaAl(EDTA) and about pD = 8.5, the Al(OH)4− signal is already visible together with a very broad, probably composite signal.) The effect of solvent on the line shape was also studied by 27 Al NMR, see Fig. 6. Spectra of NaAl(EDTA) in D2O, DMSO, methanol and CH3CN show quite different line width values (in the case of CH3CN, the signal is almost lost in the base line), but the position is almost unchanged except for the somewhat uncertain value with CH3CN as solvent. 27 Al is a quadrupolar nucleus (I = 5/2); the rate of its relaxation, i.e., the half width of the signal, depends very much on the symmetry of the species and the viscosity of the solvent.

Separation of these effects would need a systematic study. However, the broadening might be attributed to the lowered symmetry, dominantly related to the smaller rate of the intramolecular exchange reactions in organic solvents with different polarity values [22]. This assumption is supported by the different 1 H and 13C NMR spectra of Al(EDTA)− in water at different temperature values (Fig. 1) and DMSO (Fig. 2) explained by different fluxionality, vide supra. 3.1.2. Ga(EDTA)− complex The features of the 500 MHz 1H NMR spectra of Ga(EDTA)− (see Fig. 7) in D2O (20%)–DMSO-d6 (80%) are similar to those of Al(EDTA)− (seen in Fig. 2); however, it is not obvious from a first glance because of the overlapping of the acetate and ethylenic signals in the range of 3.2–3.4 ppm. The appearance of the double AB signals for the acetate arms (8 peaks) and the (nearly) AB quartet signal of the ethylenic H-atoms (4 peaks) can clearly be seen at 253 or 263 K. This is in full agreement with the solid structure of the complex, having octahedrally coordinated Ga(III) and a hexadentate EDTA ligand [8,11]. The change in line-shape of the Ga(EDTA)− spectra can be rationalized in a similar way as for AlEDTA− (vide supra), although the fluxional rearrangement of the Ga-complex is certainly faster compared to the Al-complex providing the temperature is the same for both systems. The formation of Ga(EDTA)(OH)2− can also be followed by 1 H NMR. Fig. 8 shows an up-field shift of both signals (δac = 3.63

Fig. 11. Optimized geometries and selected bond distances (in Å) of Al(EDTA)− (1) and unsaturated complexes derived by the de-coordination of axial or equatorial acetate arms (1a and 1e, respectively). Crystallographic data available for 1 are shown in brackets, and relative stabilities (in kcal/mol) of the unsaturated complexes (with respect to 1) are given in parentheses.

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

and δet = 3.13) of the mixed Ga(EDTA)(OH)2− complex compared to the shift values of the parent Ga(EDTA)− complex (δac = 3.75 and δet = 3.26). The limiting values belong to these two species, while the values in the intermediate pH range are the time-averaged signals being in fast exchange in the actual NMR time scale. The pK1 value for the reaction, Ga(EDTA)− + H2O ⇌ Ga(EDTA)(OH) 2− + H + , K = [GaEDTA)(OH) 2− ][H + ]/[Ga (EDTA)−], can be estimated from the curve. pK = 6.08 ± 0.06 measured in D2O is in reasonable agreement with the literature value of 5.6 in H2O [2]. The high field shift can be attributed to the elongated Ga(III)-donor atom distances in the mixed complex resulting in a smaller de-shielding effect at the covalently bonded H-atoms of the ligand. At the same time, the line shape of the AB signals for both complexes is very similar, indicating similar symmetry for the species. (An interesting finding is for Ga(EDTA)−, that the acetate groups take part in a selective deuteration reaction, which is much faster than the one for Al(EDTA)−, In(EDTA)− and Tl(EDTA)−.) 3.1.3. In(EDTA)− complex 1 H NMR spectra of In(EDTA)− show a signal with AB pattern of the acetate methylene protons (8H) and a singlet one of the ethylene group (4H). The shapes of the NMR spectra are almost identical in water and in water–DMSO solvent mixtures, see Fig. 9, although the chemical shifts are different. Moreover,

77

the spectral changes are quite small in the studied temperature range of 345–248 K. Having only one AB signal indicates four equivalent carboxylate groups, as having only one singlet suggests equivalent H-atoms at the ethylene sites. This finding can be explained if we assume a similar structure in solution as measured in the solid, Na[In(H2O)(EDTA)]·2H2O [12], where the seven-coordinated In3+ is surrounded “hemispherically” by the hexadentate ligand. In fact, the donor atoms build up a mono-capped trigonal prism around the central metal ion. The oxygen donor atoms of the four carboxylate groups form a quasi-plane with dIn–O = 2.2 Å averaged bond length. The In3+ion is located nearly in this plane and the N-donor atoms of the ligand are located under the plane with dIn–N = 2.35 Å, while the water is coordinated from the other direction. Considering a fast exchange of the carboxylate groups, there are two symmetry planes of the molecule. Strictly speaking, the four carboxylate groups should be arranged into two groups, but the chemical shift values must be very close to each other because of the very similar chemical surroundings. Moreover, the exchange between these two sites does not likely require the dissociation of the In–O bond(s); therefore, it can be quite fast, both in mixed solvent and at low temperature (although some broadening appears at low temperature). As expected, the 13C NMR spectrum is in full agreement with this explanation, only three signals of the ethylene, the

Fig. 12. Optimized geometries and selected bond distances (in Å) of Al(HEDTA) complexes derived by protonation of the acetate oxygens. Relative stabilities (in kcal/ mol) with respect to the most stable structure (2e) are given in parentheses.

78

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

methylene and the carboxylate carbon atoms can be recorded in the same sample as used for 1H NMR. pH dependence of 1H NMR spectra for the In(EDTA)− is shown in Fig. 10. The experimental finding and its explanation are in full agreement with the Ga(EDTA)− system, vide supra. The pK1 = 9.17 ± 0.06 in D2O for the reaction, In(EDTA)− + H2O ⇋ In(EDTA) (OH)2− + H+, K = [In(EDTA)(OH)2−][H+]/[In(EDTA)−], is in reasonable agreement with the literature value of 8.2 in H2O [2,7]. 4. DFT calculations on Al(EDTA)−, AlH(EDTA) and Al (EDTA)(H2O)− systems The axial–equatorial exchange of the acetate arms in EDTA complexes could occur via two distinct routes. The acetate arms belonging to the same imino-diacetate entity may change the ax/eq position through the complete de-coordination of an entire imda group followed by the N-inversion [20]. This route can be considered as direct exchange. Alternatively, an indirect exchange between the axial and equatorial groups belonging to different imda groups is also feasible, and it results in the change of the helicity of the acetate arms. The indirect rearrangement involves several elementary steps and should operate ‘back and forth’. In order to reveal the mechanism of exchange processes between the axial and equatorial carboxylate sites of the Al (EDTA)− complex, we initiated a theoretical study in terms of DFT calculations. Here we present our preliminary results con-

cerning the effect of protonation and the presence of water molecules on the de-coordination of a single acetate arm. The optimized structure of the Al(EDTA)− complex is depicted in Fig. 11. The calculated bond distances associated with the metal–ligand bonds in 1 are fairly close to those obtained from X-ray measurements indicating the level of accuracy of the applied methodology for structural investigations. The displacement of the acetate arms out of the coordination sphere of 1 is predicted to be highly endothermic in this model as indicated by the calculated relative energies of 1a and 1e. The structures of the unsaturated complexes show that the coordination sphere may significantly rearrange after the arm displacement. This rearrangement is related to the distortion of AlNCCO pentagons associated with each coordinated acetate arms, which become rather flexible after the de-coordination of an arm. Four different protonated forms can be derived from the hexacoordinated complex depending on whether the protonation takes place on the axial/equatorial or the inner/outer O atom of acetate arms. The optimized structures of these complexes are displayed in Fig. 12 along with their relative stabilities. As expected, the protonation of the carboxylate oxygens weakens the Al–O bonds, particularly if the protonation takes place on the atom bound directly to the metal center. However, from the energetic point of view, the protonation of the outer oxygens is found to be preferred according to the relative

Fig. 13. Optimized geometries and selected bond distances (in Å) of Al(HEDTA)(H2O). Relative stabilities (in kcal/mol) with respect to structure 3a are given in parentheses.

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80

stabilities of the four species. Structures 2a and 2e are calculated to be close in energy implying that both types of arms can undergo the protonation process. Our calculations indicate that the energy required to dissociate the protonated acetate arms is much smaller than in the Al(EDTA)− complex. For instance, the de-coordinated form of 2a lies only 3.1 kcal/mol above the saturated complex. Our preliminary results regarding the activation barrier of arm displacement indicate that the de-coordination of a protonated acetate might not be the ratedetermining step in the ligand exchange process, because the energy gap estimated for the arm displacement is about 9 kcal/ mol. Further work to explore additional elementary steps involved in the axial–equatorial exchange of the acetate arms is in progress in our group. The results presented above refer to gas-phase calculations, because the effect of the water solvent has been fully neglected. Although the experimental findings show that the water exchange rate between the coordinated and bulk water is much faster compared to the rate of the overall carboxylate exchange [25–27] it is quite likely that solvent molecules play an important role not only as proton donors, but also as the solvent medium. As a first step in understanding the latter effects, we have considered the interaction of a single water molecule with possible intermediates characteristic for the carboxylate exchange process. The structures and their relative stabilities are shown in Fig. 13. We mention first that a water molecule is able to coordinate to the metal centre of a saturated Al(HEDTA) complex to form a hepta-coordinated species (3a). The binding energy of H2O in 3a is 7.6 kcal/mol, which is only slightly larger than the bond strength between water molecules in bulk water. The weakening of the Al–N bond related to the protonated acetate in 3a suggests that water coordination may facilitate the arm displacement. Once the protonated arm is de-coordinated, an incoming water molecule may occupy the vacant coordination site giving rise to hexa-coordinated species (3b). The Al–OH2 bond distance in 3b is notably longer than those in other identified Al (HEDTA)(H2O) structures indicating that the occupation of the axial site might not be the energetically most preferred one in the open isomer of the Al(HEDTA) complex. Indeed, complexes 3c and 3d formed after the rearrangement of the coordinated arms and subsequent H2O coordination are clearly more stable than 3b. In these complexes, one of the acetate arms is shifted to the axial position and the water molecule is bound in trans position to the N atom of an imda unit. One can conclude from these results that water exchange processes are likely involved in the rearrangement of the arms that remain attached to the Al ion in Al(HEDTA). Further calculations are required, however, to explore the transition states related to the above exchange processes and gain more detailed insight into the role of water medium in carboxylate exchange processes. It is worth mentioning that during the axial–equatorial exchange, non-octahedral species might be formed, e.g., pentacoordinated (1a) or hepta-coordinated (3a) species. We cannot estimate the population and/or the lifetime of these species, but we emphasize that Al K XANES spectroscopy results [9] can be explained by co-existence of different species in solution.

79

Acknowledgement We acknowledge OTKA T 038296 and K63388 for the financial support. M.P. thanks the support from DETEK. The authors also thank Dr. David Lawrence (Linköping University) for linguistic correction of this manuscript. References [1] I.G. Farbenindustrie A–G Germ. Pat., 638071; Chem. Abstr., 31 (1937) 1044. [2] G. Anderegg, Critical Stability Constants of EDTA Complexes, IUPAC Chemical Data Series, vol. 14, Pergamon, 1977; A.E. Martell, R.M. Smith, Critical Stability Constants, Plenum, 1974; W. Hummel, G. Anderegg, I. Puigdomenech, L. Rao, O. Tochiyama, Chemical Thermodynamics of Compounds and Complexes of U, Np, Pu, Am, Tc, Se, Ni and Zr with Selected Organic Ligand, OECD/NEA Chemical Thermodynamics Series, vol. 9, Elsevier Science, 2005. [3] N.L. Hakem, P.G. Allen, E.R. Sylwester, J. Radioanal. Nucl. Chem. 250 (1) (2001) 47. [4] P.M. Jardie, D.L. Taylor, Geoderma 67 (1–2) (1995) 125. [5] O.W. Howarth, P. Moore, N. Winterton, J. Chem. Soc. Dalton Trans. (1974) 2271. [6] J.K. Hovey, P.R. Tremaine, J. Phys. Chem. 89 (1985) 5541. [7] I. Tóth, E. Brücher, L. Zékány, V. Veksin, Polyhedron 8 (1989) 2057; J. Nemes, I. Tóth, L. Zékány, J. Chem. Soc. Dalton Trans. (1998) 2707. [8] W. Jung, Y.K. Chung, D.M. Shin, S. Kim, Bull. Chem. Soc. Jpn. 75 (2002) 1263. [9] S. Matsuo, H. Wakita, Struct. Chem. 14 (2003) 69; S. Matsuo, K. Shirozu, Y. Tateishi, H. Wakita, T. Yokoyama, Adv. Quantum Chem. 42 (2002) 407. [10] T.N. Polynova, N.P. Bel'skaya, D. Tyurk de Garciya Banus, M.A. PoraiKoshits, L.I. Martynenko, Zh. Strukt. Khim. 11 (1970) 164. [11] K. Namakura, T. Kurasaki, H. Wakita, T. Yamaguchi, Acta Crystallogr., C Cryst. Struct. Commun. 52 (1996) 573. [12] A.B. Ilyukhin, M.A. Maliarik, M.A. Porai-Koshits, R.L. Davidovich, V.B. Loginova, Kristallografia 40 (1995) 656. [13] a. H.M.N.H. Irving, M.G. Miles, L.D. Pettit, Anal. Chim. Acta 38 (1967) 475; b. A.K. Covington, M. Paabo, R.A. Robinson, R.G. Bates, Anal. Chem. 40 (1968) 700. [14] I. Puigdomenech, MEDUSA (Windows Interface to the MS-DOS Version of INPUT, SED and PREDOM Fortran Program Drawing Chemical Equilibrium Diagrams), 1997, Program is available at http://www.kemi. kth.se/medusa/ or http://web.telia.com/~u15651596/. [15] L.A. van Geet, Anal. Chem. 42 (1970) 679. [16] A.D. Becke, J. Chem. Phys. 98 (1993) 5648. [17] C. Lee, W. Yang, R.G. Parr, Phys. Rev., B 37 (1988) 785. [18] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M.J. Frisch, J. Phys. Chem. 98 (1994) 11623. [19] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, J.A. Montgomery Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komáromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Revision C.02, Gaussian, Inc., Wallingford CT, 2004. [20] R.J. Day, C.N. Reilley, Anal. Chem. 36 (1964) 1073; R.J. Day, C.N. Reilley, Anal. Chem. 37 (1965) 1326.

80 [21] [22] [23] [24] [25]

R. Jószai et al. / Journal of Molecular Liquids 131–132 (2007) 72–80 A. Bodor, I. Bányai, L. Zékány, I. Tóth, Coord. Chem. Rev. 228 (2002) 163. C. Reichardt, Pure Appl. Chem. 76 (2004) 1903. E. Brücher, I. Bányai, J. Inorg. Nucl. Chem. 42 (1980) 749. J.W. Akitt, Prog. Nucl. Magn. Reson. Spectrosc. 21 (1989) 1. B.L. Phillips, W.H. Casey, N.S. Crawford, Geochim. Cosmochim. Acta 61 (1997) 3041.

[26] B.L. Phillips, J.A. Tossell, W.H. Casey, Environ. Sci. Technol. 32 (1998) 2865. [27] A. Bodor, I. Tóth, I. Bányai, Z. Szabó, G.T. Hefter, Inorg. Chem. 39 (2000) 2530.