Article pubs.acs.org/IC
Solvent Exchange and Electron-Spin Relaxation on Homoleptic Acetonitrile Complexes of Trivalent Lanthanides Gabriella Bodizs and Lothar Helm* Institut des sciences et ingénierie chimiques, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland S Supporting Information *
ABSTRACT: Homoleptic acetonitrile complexes [Nd(CH3CN)9][Al(OC(CF 3 ) 3 ) 4 ] 3 , [Dy(CH 3 CN) 9 ][Al(OC(CF 3 ) 3 ) 4 ] 3 , and [Tm(CH3CN)8][Al(OC(CF3)3)4]3 have been studied in anhydrous acetonitrile by 14N and 1H NMR relaxation. Solvent-exchange rate constants increase from (22 ± 6) × 106 s−1 (Nd3+) and (160 ± 40) × 106 s−1 (Dy3+) for the nonasolvated ions to (360 ± 40) × 106 s−1 (Tm3+) for the octasolvated ions. Electron-spin relaxation of the lanthanide ions studied is similar to that found in aqua ions. This dependence on the binding properties of the coordinating molecules is consistent with the model proposed by Fries et al. for fast electron-spin relaxation of lanthanide ions other than Gd3+.
■
INTRODUCTION On the basis of luminescence studies on TbIII ions in solution, Bünzli and co-workers1,2 established a solvent affinity sequence showing that acetonitrile is the most weakly interacting solvent, at least among those studied. As a consequence of the weak interaction between acetonitrile molecules and lanthanide ions, it is very difficult to synthesize its homoleptic complexes, [Ln(CH3CN)q]3+. Even well-known weakly coordinating counterions such as ClO4− and CF3SO4− compete with CH3CN ligands for coordination sites.3 Using the bulky anion [Al(OC(CF3)3)4]−, we achieved the synthesis of homoleptic nine- and eight-coordinated acetonitrile complexes of Ln3+ ions in the solid state and in solution.4 NMR (14N and 1H) and electron paramagnetic resonance (EPR) experiments performed on anhydrous solutions of [Gd(CH3CN)9]3+ and [Eu(CH3CN)9]2+ allowed the first determination of exchange rate constants of CH3CN on lanthanide ions.5 For both solvate complexes, acetonitrile exchange is slower for the less strongly nitrogen-binding solvent than for the more strongly oxygen-binding water. This has been explained by the exceptional behavior of water exchange on eight-coordinated lanthanide ions:6 the nine-coordinated transition state has an energy close to the eight-coordinated ground state, leading to a very fast water exchange in an associatively activated exchange reaction. Lanthanide ions, which constitute the longest series of chemically similar metal ions, offer the unique opportunity to study solvent-exchange reactions as a function of the ionic radius. The ionic radius shrinks about 0.18 Å upon going from La3+ to Lu3+ for both eight- and nine-coordinated complexes.7 It has been shown on isostructural lanthanide complexes8,9 that variation of the Ln−X distances is best described by a secondorder polynomial, with X being a Lewis-basic donor like oxygen or nitrogen. The few solvent-exchange studies performed on © XXXX American Chemical Society
aqueous and nonaqueous solutions showed different variations of the exchange rate constants, kex, as a function of the ionic radius. In the case of water10−12 and tetramethylurea (TMU),13,14 kex decreases monotonically with decreasing Ln− X distance; in the case of dimethylformamide (DMF),15 kex shows a minimum at Ho3+. It has to be mentioned that these studies have been performed on lanthanide ions of the second half of the series (from Gd3+ to Yb3+), with coordination numbers varying from q = 6 to 8. For water exchange, a maximum of kex has been predicted at Gd3+ on the basis of complex formation rates, and there was a change of the activation mode from associative for the second half of the series to dissociative for the first half.6 The reason for the lack of data from the first part of the lanthanide series is the less favorable magnetic properties of these ions. Lower limits of kex determined for Pr3+ and Nd3+ are in line with the prediction of a maximum of lability close to the change in the coordination number.16 The major problem in the experimental study of solventexchange reactions on Ln3+ ions other than the isoelectronic Gd3+ and Eu2+ is the extremely fast electron-spin relaxation of these paramagnetic ions (the diamagnetic ions La3+ and Lu3+ do not allow the measurement of very fast reaction rate constants using NMR). As a consequence, residence times of solvent molecules in the first coordination sphere of Ln3+, τm = 1/kex, cannot be determined using the usual Swift and Connick approach.17,18 Cossy and co-workers developed an approach10 based on the measurement of nuclear relaxation rates 1/T1 and 1/T2 of the directly bound solvent atom, which is nitrogen in the case of CH3CN. Received: December 3, 2014
A
DOI: 10.1021/ic5028493 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
acetonitrile were measured at 4.7 T: 0.104 s−1 (238.15 K) and 0.068 s−1 (298.15 K). Data Analysis. The individual least-squares fits of 14N NMR as well as the simultaneous fit of 1H NMRD and 14N 1/T1r data were performed by the Visualizeur/Optimiseur programs30 running on a Matlab 6.5 platform. The errors of the fitted parameters correspond to 1 standard deviation.
Besides the chemical dynamics on solvate complexes of Ln3+, the very fast electron-spin relaxation is an interesting phenomenon itself.19,20 It has been studied by 1H NMR relaxation in aqueous solution in the past,21−23 but so far no data are available on nonaqueous systems. The electron-spin relaxation has been characterized by time constant τS, which is 4 T, where Curie relaxation gives important contributions to r1.
Concerning 1H NMRD profiles, the significance of innerand outer-sphere relaxation as well as the contribution of Curie relaxation to it is exemplified in Figure 4 for [Dy(CH3CN)9]3+ (for [Tm(CH3CN)8]3+, see the SI). Both r1is and r1os arise from the corresponding dipolar and Curie relaxation contributions. At low magnetic fields, B0 < 3 T (127 MHz 1H), the dipolar relaxation dominates r1is and r1os. For the Dy3+ (Tm3+) ion, r1is contributes 72% (75%) to the total relaxivity at 20 MHz (0.47 T) and 298 K. At 800 MHz (18.8 T), the contribution to r1 due to Curie relaxation is 63% (Dy3+) and 43% (Tm3+). The relative contribution due to outer-sphere relaxation changes only slightly over the wide range of magnetic fields studied: from 28% (Dy3+, low field) to 32% (Dy3+, high field). The corresponding data for Tm3+ are 25% (low field) and 26% (high-field) for r1os. The relative contribution of outer-sphere relaxation to relaxivity is markedly higher for CH3CN than for H2O. From a reanalysis of Bertini’s data22 on [Dy(H2O)8]3+ including r1os, we obtained a contribution of 18% of r1os to r1 (at low field). The orientation of second-sphere acetonitrile molecules has already been discussed in the former work on [Gd(CH3CN)9]3+.5 A tangential or an antiparallel arrangement of second-sphere or bulk solvent molecules with respect to firstcoordination-sphere CH3CN can explain the relatively small difference between the first-sphere (rLnH) and outer-sphere (aLnH) distance. Both are markedly longer for the rodlike Nbound CH3CN molecule than for the more spherical H2O. This has a marked influence on the inner-sphere relaxation terms, which both vary with r−6 LnH, but only a relatively small influence on outer-sphere relaxation, varying roughly as a−1 LnH. G
DOI: 10.1021/ic5028493 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry
acetonitrile (0.34 cP at 25 °C)47 with respect to water (0.89 cP at 25 °C).48 The rotational correlation times increase slightly to the end of the lanthanide series, which can be attributed to stronger interaction with molecules in the second coordination sphere. The very short electron-spin relaxation times for all three lanthanide ions studied are close to those found for aqueous solvate complexes and other complexes in solution. This is in contrast to gadolinium, for which τS of the acetonitrile complex is at least 10 times longer than τS of the aqua ion. This finding is in perfect agreement with the model proposed by Fries and Belorizky,24 which predicts that electron-spin relaxation on lanthanide ions other than Gd3+ varies little with the coordinating ligand and is practically independent of the magnetic field strength.
The shorter rotational correlation times τR in acetonitrile compared to those measured in aqueous solution reflect the lower viscosity in the organic solvent (Table 4). The value recalculated including r1os for [Dy(H2O)8]3+ is in perfect agreement with 17O NMR data on [Gd(H2O)8]3+.44 The small increase in τR with decreasing ionic radius can probably be attributed to a small increase in interaction with second-sphere solvent molecules. Electron-Spin Relaxation. As was already mentioned, the electron-spin relaxation, τS, is the dominating correlation time for dipolar interaction between protons and the lanthanide electron spins for Ln3+ ions, except Gd3+. The low-field relaxivities, which are dominated by 1/TH1dd, give therefore direct access to τS. Electron-spin relaxation times obtained on [Ln(CH3CN)8/9]3+ are compared with literature data in Table 4. Electron-spin relaxation for Gd3+ ions in solution is slow and varies by at least 1 order of magnitude between solutions in CH3CN and H2O. This difference is a consequence of the zerofield-splitting energy, which is larger for the aqua complex compared to the acetonitrile complex. Electron-spin relaxation of all other lanthanide ions is very fast with τS < 1 ps, as expected. A direct comparison between the results on aqueous and acetonitrile solutions shows that τ298 S values are similar. For solutions of Dy3+, the only ion where a direct comparison is possible between Bertini’s results in water with our results in acetonitrile, electron-spin relaxation is identical if r1os is taken into account in data analysis. The activation energies, ES, in both systems are also very close: 3.8 ± 1.5 kJ mol−1 ([Dy(CH3CN)9]3+) and 2.7 ± 1.6 mol−1 (([Dy(H2O)8]3+).5 Electron-spin relaxation times measured on lanthanide complexes [LnL(D2O)] in D2O and CD3OD are also very similar (Table 4).46 In a recent paper, Fries and Beloritzky24 interpreted the fast relaxation by rapid modulation of the ligand field due to distortions of the complexes induced by collisions with solvent molecules. They found that the effective electronic relaxation time Teff 1e = τS is related to the norm of the transient ligand field and its correlation time τv. Following their model, the symmetry of the transient ligand field and therefore also the coordination number of bound solvent molecules is of minor importance.24 Within the experimental errors on the determination of τS and the assumptions made in the theoretical model, our data are fully consistent with the proposed model for fast electron-spin relaxation of lanthanide ions other than Gd3+.
■
ASSOCIATED CONTENT
* Supporting Information S
Equations used for the combined analysis of NMR relaxation and chemical shifts, calculations justifying the approximations used in data analysis (Table S-1), reanalysis including outersphere relaxation of Bertini’s data on aqua ions for [Tm(H2O)8]3+ (Figure S-1 and Table S-2), contribution of outersphere relaxation to the NMRD profile of [Tm(CH3CN)8]3+ (Figure S-2), and data for μeff obtained from bulk magnetic susceptibility (Table S-3). This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: lothar.helm@epfl.ch. Phone: +41 21 693 9876. Fax: +41 21 693 9855. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS André E. Merbach is acknowledged for numerous helpful and stimulating discussions. This work was supported by the Swiss National Science Foundation.
■
REFERENCES
(1) Bünzli, J.-C. G.; Vuckovic, M. M. Inorg. Chim. Acta 1983, 73, 53− 61. (2) Bünzli, J.-C.; Milicic-Tang, A. In Handbook on the Physics and Chemistry of Rare Earths; Gscheidner, K. A., Eyring, H., Eds.; Elsevier: Amsterdam, The Netherlands, 1995; Vol. 21, pp 305−366. (3) Di Bernardo, P.; Choppin, G. R.; Portanova, R.; Zanonato, P. L. Inorg. Chim. Acta 1993, 207, 85−91. (4) Bodizs, G.; Raabe, I.; Scopelliti, R.; Krossing, I.; Helm, L. Dalton Trans. 2009, 5137−5147. (5) Bodizs, G.; Helm, L. Inorg. Chem. 2012, 51, 5881−5888. (6) Helm, L.; Merbach, A. E. Chem. Rev. 2005, 105, 1923−1960. (7) Shannon, R. D. Acta Crystallogr., Sect. A 1976, 32, 751−767. (8) Quadrelli, E. A. Inorg. Chem. 2002, 41, 167−169. (9) Seitz, M.; Oliver, A. G.; Raymond, K. N. J. Am. Chem. Soc. 2007, 129, 11153−11160. (10) Cossy, C.; Helm, L.; Merbach, A. E. Inorg. Chem. 1988, 27, 1973−1979. (11) Cossy, C.; Helm, L.; Merbach, A. E. Inorg. Chem. 1989, 28, 2699−2703. (12) Micskei, K.; Powell, D. H.; Helm, L.; Brücher, E.; Merbach, A. E. Magn. Reson. Chem. 1993, 31, 1011−1020. (13) Pisaniello, D. L.; Lincoln, S. F.; Williams, A. F.; Jones, A. Aust. J. Chem. 1981, 34, 495−500.
■
CONCLUSION Variable-temperature and magnetic field 14N and 1H NMR experiments have been performed in anhydrous CH3CN solutions of [Nd(CH3CH)9]3+, [Dy(CH3CH)9]3+, and [Tm(CH3CH)8]3+. Acetonitrile exchange rate constants increase continuously from the bigger Nd3+ ion to the smaller Tm3+ ion. Despite the negative activation entropies, a dissociative mechanism for acetonitrile exchange on nonacoordinated lanthanide solvates is proposed. The main reason for that is the increasing repulsion between the first-sphere solvent molecules, leading to an easier dissociation. The jump to a more negative activation entropy for the eight-coordinated [Tm(CH3CH)8]3+ is an indication for a change of the mechanism from dissociative to associative. This change happens between Dy and Tm, which is shifted to smaller lanthanide ions compared to aqua ions. Rotational correlation times are shorter compared to those of aqueous solution, which is due to the lower viscosity of H
DOI: 10.1021/ic5028493 Inorg. Chem. XXXX, XXX, XXX−XXX
Article
Inorganic Chemistry (14) Lincoln, S. F.; White, A. H. Inorg. Chim. Acta 1990, 168, 265− 270. (15) Pisaniello, D. L.; Helm, L.; Meier, P. F.; Merbach, A. E. J. Am. Chem. Soc. 1983, 105, 4528−4536. (16) Powell, D. H.; Merbach, A. E. Magn. Reson. Chem. 1994, 32, 739−745. (17) Swift, T. J.; Connick, R. E. J. Chem. Phys. 1962, 37, 307−320. (18) Swift, T. J.; Connick, R. E. J. Chem. Phys. 1964, 41, 2553−2554. (19) Banci, L.; Bertini, I.; Luchinat, C. Nuclear and Electron Relaxation; VCH: Weinheim, Germany, 1991. (20) Bertini, I.; Luchinat, C.; Parigi, G. Solution NMR of Paramagnetic Molecules; Current Methods in Inorganic Chemistry; Elsevier: Amsterdam, The Netherlands, 2001; Vol. 2. (21) Alsaadi, B. M.; Rossotti, F. J. C.; Williams, R. J. P. J. Chem. Soc., Dalton Trans. 1980, 2147−2150. (22) Bertini, I.; Capozzi, F.; Luchinat, C.; Nicastro, G.; Xia, Z. J. Phys. Chem. 1993, 97, 6351−6354. (23) Vigouroux, C.; Belorizky, E.; Fries, P. H. Eur. Phys. J. D 1999, 5, 243−255. (24) Fries, P. H.; Belorizky, E. J. Chem. Phys. 2012, 136, 074513/1− 10. (25) Ammann, C.; Meier, P.; Merbach, A. E. J. Magn. Reson. 1982, 46, 319−321. (26) Hugi, A. D.; Helm, L.; Merbach, A. E. Helv. Chim. Acta 1985, 68, 508−521. (27) Micskei, K.; Helm, L.; Brücher, E.; Merbach, A. E. Inorg. Chem. 1993, 32, 3844−3850. (28) Vold, R. L.; Waugh, J. S.; Klein, M. P.; Phelps, D. E. J. Chem. Phys. 1968, 48, 3831−3832. (29) Meiboom, S.; Gill, D. Rev. Sci. Instrum. 1958, 29, 688−691. (30) Yerly, F. Visualiseur/Optimiseur; EPFL: Lausanne, Switzerland, 2003. (31) Zimmerman, J. R.; Brittin, W. E. J. Chem. Soc., Faraday Trans. 1 1957, 61, 1328−1333. (32) Helm, L.; Nicolle, G. M.; Merbach, A. E. Adv. Inorg. Chem. 2005, 57, 327−379. (33) Harley, S. J.; Ohlin, C. A.; Casey, W. H. Geochim. Cosmochim. Acta 2011, 75, 3711−3725. (34) Lewis, W. B.; Jackson, J. A.; Lemons, J. F.; Taube, H. J. Chem. Phys. 1962, 36, 694−701. (35) Bleaney, B. J. Magn. Reson. 1972, 8, 91−100. (36) Granot, J.; Fiat, D. J. Magn. Reson. 1975, 19, 372−381. (37) Halle, B.; Wennerström, H. J. Magn. Reson. 1981, 44, 89−100. (38) Koenig, S. H.; Brown, R. D. In Handbook of metal−ligand interactions in biological fluids.Bioiniorganic chemistry; Berthon, G., Ed.; Dekker: New York, 1995; pp 1093−1108. (39) Ayant, Y.; Belorizky, E.; Alizon, J.; Gallice, J. J. Phys. (Paris) 1975, 36, 991−1004. (40) Hwang, L. P.; Freed, J. H. J. Chem. Phys. 1975, 63, 4017−4025. (41) Freed, J. H. J. Chem. Phys. 1978, 69, 4034−4037. (42) Holz, M.; Weingärtner, H. J. Magn. Reson. 1991, 92, 115−125. (43) Dunand, F. A.; Helm, L.; Merbach, A. E. Adv. Inorg. Chem. 2003, 54, 1−69. (44) Powell, D. H.; Ni Dhubhghaill, O. M.; Pubanz, D.; Helm, L.; Lebedev, Y. S.; Schlaepfer, W.; Merbach, A. E. J. Am. Chem. Soc. 1996, 118, 9333−9346. (45) Rast, S.; Borel, A.; Helm, L.; Belorizky, E.; Fries, P. H.; Merbach, A. E. J. Am. Chem. Soc. 2001, 123, 2637−2644. (46) Funk, A. M.; Fries, P. H.; Harvey, P.; Kenwright, A. M.; Parker, D. J. Phys. Chem. A 2013, 117, 905−917. (47) Ibuki, K.; Nakahara, M. J. Phys. Chem. 1990, 94, 8370−8373. (48) Dhondge, S. S.; Dahasahasra, P. N.; Paliwal, L. J.; Deshmukh, D. W. J. Chem. Thermodyn. 2014, 76, 16−23.
I
DOI: 10.1021/ic5028493 Inorg. Chem. XXXX, XXX, XXX−XXX
