Inorg. Chem. 2002, 41, 5024−5033
A Ligand-Field Analysis of the trensal (H3trensal ) 2,2′,2′′-Tris(salicylideneimino)triethylamine) Ligand. An Application of the Angular Overlap Model to Lanthanides Bernadine M. Flanagan,† Paul V. Bernhardt,† Elmars R. Krausz,‡ Stefan R. Lu1 thi,† and Mark J. Riley*,† Department of Chemistry, UniVersity of Queensland, St. Lucia 4072, Australia, and Research School of Chemistry, Australian National UniVersity, Canberra 0200, Australia Received December 12, 2001
The spectral and geometric trends of Ln(trensal) complexes (H3trensal ) 2,2′,2′′-tris(salicylideneimino)triethylamine) along the lanthanide series are analyzed. Low-temperature polarized absorption and luminescence spectra are reported for nine of the complexes with transitions suitable for analysis. Both the angular and radial geometry variations as a function of the lanthanide ion are quantified. The structural parameters are related to the trend in the ligand field found from an analysis of the f−f transitions. The ligand-field analysis is fitted globally across the whole series accounting for the contraction of the f orbitals as a function of atomic number. This study establishes the utility of describing the ligand field of f electrons within the one-electron operator approach of the angular overlap model.
Introduction There have been few comparative ligand-field analyses along the lanthanide series. Most commonly these have focused on Ln(III)-doped ionic lattices such as LiYF4,1 YAlO3,2 YAG,3 Cs2NaYCl6,4 LaCl3,5 and LaF36 where the guest ion substitutes for either yttrium(III) or lanthanum(III). These studies have the disadvantage that the trends in the ligand-field parameters are related to the inferred geometry changes based on the site symmetry of the host and the expected change in ionic radii of the guest ions. Alternatively, in neat crystal systems such as Na3[Ln(oda)3]‚ 2NaClO4‚6H2O (Ln ) Ce-Yb; oda ) oxydiacetate),7 although the isomorphous nature of the series has been * Author to whom correspondence should be addressed. E-mail: riley@ chemistry.uq.edu.au. † University of Queensland. ‡ Australian National University. (1) Jayasankar, C. K.; Reid, M. F.; Richardson, F. S. Phys. Status Solidi B 1989, 155, 559. (2) Deb, K. K. J. Phys. Chem. Solids 1982, 43, 189. (3) Morrison, C. A.; Wortman, D. E.; Karayianis, N. J. Phys. C 1976, 9, L191. (4) Tanner, P. A.; Kumar, V. V. R. K.; Jayasankar, C. K.; Reid, M. F. J. Alloys Compd. 1994, 215, 349. (5) Jayasankar, C. K.; Richardson, F. S.; Reid, M. F. J. Less-Common Met. 1989, 148, 289. (6) Carnell, W. T.; Goodman, G. L.; Rajnak, K.; Rana, R. S. J. Chem. Phys. 1989, 90, 3443. (7) Hopkins, T. A.; Metcalf, D. H.; Richardson, F. S. Inorg. Chem. 1998, 37, 1401.
5024 Inorganic Chemistry, Vol. 41, No. 20, 2002
established, a fully refined crystal structure of each member is lacking; thus, the local coordination environment (bond lengths and angles) is not known. Recently the relationship between two different descriptions of the ligand field for the Er(trensal) (H3trensal ) 2,2′,2′′-tris(salicylideneimino)triethylamine) complex was investigated.8 The angular overlap model (AOM) σ- and π-bonding parameters were related to the symmetrydetermined “crystal-field” coefficients of the spherical harmonic expansion of the ligand field. Reasonably good agreement was found between these two approaches, and it was proposed that the AOM may prove useful as a “first guess” of the ligand fields of lanthanide complexes of lower symmetry. In low-symmetry complexes, there are many crystal-field parameters and the parameter space can often have multiple solutions when the difference between observed and calculated energy levels is minimized. It is often unclear which parameter set best describes the ligand field. In the present work, we treat a complete series of Ln(III) complexes (except for radioactive Pm(III)) where there are small and systematic changes in the ligand geometry about the metal ion. We also expect small and systematic changes in the ligand field so that the best fit for a particular complex must also fit the trend across the series. (8) Flanagan, B. M.; Bernhardt, P. V.; Lu¨thi, S. R.; Krausz, E. R.; Riley, M. J. Inorg. Chem. 2001, 40, 5401.
10.1021/ic011276q CCC: $22.00
© 2002 American Chemical Society Published on Web 09/06/2002
A Ligand-Field Analysis of the Trensal Ligand
The heptadentate tripodal ligand “trensal” forms stable complexes with lanthanide(3+) ions. This ligand has the important advantage that it forms a complete isomorphous and isostructural series with all lanthanide complexes. A three-dimensional structure determination leaves no doubt of the geometry of the lanthanide complex unlike a doped crystal system. An analysis of the Er(trensal) complex has shown that this ligand has one of the largest ligand fields observed for a lanthanide complex, 3-fold greater than that of the simple halide LaCl3 host, for example.8 The large ligand field and complete structural data over the whole isomorphous series allow us to examine a global fit of crystal-field parameters and to relate the AOM parameters to the observed small structural changes. We analyze the trend in terms of the one-electron-operator, additive ligandfield approach of the AOM. Experimental Section Synthesis and Structure. Details of the synthesis9 and structural characterization10 of these compounds have been described previously. It is possible to crystallize all compounds in the trigonal space group P3hc1. Care must be taken as it has been found11,12 that polymorphic (Ln ) Pr, Nd, Eu, Gd, Tb) trensal compounds are known which include solvent molecules in the structure, which crystallizes in the P3h space group under different synthetic conditions. Our compounds are all exclusive of solvent and have the Ln(III) ions lying on a C3 site of the P3hc1 space group. Typically the crystals have a needlelike habit with the crystallographic c axis coinciding with the long crystal axis. Figure 1 shows the Er(trensal) complex. The heptadentate trensal ligand binds in a N4O3 fashion, resulting in a trigonally capped distorted octahedron, with the C3 axis passing through the Ln(III) ion and the apical tertiary amine (N1). The 3-fold symmetry relates the N2 imine and O1 atoms. These three independent metal-ligand bond lengths for all Ln(trensal) complexes (barring Pm) are given in Table 1. Also included are the structural parameters for the trensal ligand complexed to Cr(III) for comparison. With the smaller transition-metal ions, the apical N1 of the trensal ligand is not bound and the coordination is that of a distorted octahedron. Spectroscopic Investigations. Polarized absorption spectra were recorded at 10 K on a single-beam absorption instrument as described previously.8,13 The near-IR absorption spectra were collected using a liquid-N2-cooled InSb detector with a 600 lines/ mm grating blazed at 1600 nm, while measurements of the visible spectra were done using an S-20 photomultiplier tube with a 1200 lines/mm grating blazed at 500 nm. Luminescence spectra were obtained using the 356.4 and 350.7 nm lines of a Kr+ ion laser as the excitation source. For absorption and luminescence experiments at 10 K, the samples were cooled with a helium flow-tube cryostat. Raman measurements were made on powdered samples at room temperature on a Perkin-Elmer FT-Raman 2000 instrument. Computational Procedures. We used the Reid suite of programs14 to calculate the fn multiplet energies. The large ligand field (9) Bernhardt, P. V.; Flanagan, B. M.; Riley, M. J. Aust. J. Chem. 2000, 53, 229. (10) Bernhardt, P. V.; Flanagan, B. M.; Riley, M. J. Aust. J. Chem. 2001, 54, 229. (11) Kanesato, M.; Yokoyama, T.; Itabashi, O.; Suzuki, T. M.; Shiro, M. Bull. Chem. Soc. Jpn. 1996, 69, 1297. (12) Kanesato, M.; Yokoyama, T. Chem. Lett. 1999, 137. (13) Krausz, E. R. Aust. J. Chem. 1993, 46, 1041. (14) Reid, M. F. J. Chem. Phys. 1987, 87, 2875.
Figure 1. Molecular structure of Er(trensal). (a) Viewed down the c axis, showing the atom labels. Coordinating atoms are shown in black. (b) Edgeon view to the N2-Er-O1 bidentate ring showing the angle that it makes with the trigonal axis.
encountered for the trensal ligand8 implies extensive mixing of the atomic levels, and care must be taken to use an adequate basis. We used the full basis for f1-3(11-13) and the truncated bases of 754 for f5, 477 for f6, 789 for f8, 594 for f9, and 559 for f10. In all cases the truncated atomic levels were >40000 cm-1 above the ground state.
Results Experimental Spectra. A representative overview of the absorption spectrum of Sm(trensal) is shown in Figure 2 for a crystal of ∼200 × 200 µm. The spectrum is typical of those measured across the lanthanide series. The noise at 5300 and 7200 cm-1 is due to the absorption of water overtone bands. The onset of ligand absorption occurs at ∼25000 cm-1. Particularly intense transitions have the potential to be saturated at an absorbance >1.0, due to the small crystal size. This is dependent on the optical quality of the crystal; with some samples it was possible to measure an absorbance of up to 2.5. A representative luminescence spectrum of Sm(trensal) is shown in Figure 3. As observed previously for Er(trensal),8 efficient energy transfer from the ligand-based absorption to the metal-based luminescence occurs. Figure 3 also shows that luminescence occurs at room temperature. This is unusual for a lanthanide complexed to an organic ligand as the high-energy vibrations of the ligand are usually efficient in nonradiative relaxation processes. The luminescent hot bands due to the thermal population of higher levels of the excited-state multiplet are indicated (by an H). The observed Inorganic Chemistry, Vol. 41, No. 20, 2002
5025
Flanagan et al. Table 1. Structural Parameters of the Ln(trensal) Complexes Ce
Pr
Nd
Sm
Eu
Gd
Tb
Dy
Ho
Er
Tm
Yb
Lu
Cr
M-N1/Å 2.836 2.816 2.796 2.780 2.756 2.705 2.739 2.746 2.713 2.713 2.698 2.693 2.697 3.244 M-N2/Å 2.607 2.583 2.576 2.539 2.533 2.508 2.491 2.476 2.450 2.460 2.444 2.428 2.424 2.102 M-O/Å 2.287 2.280 2.273 2.246 2.237 2.240 2.222 2.202 2.194 2.186 2.172 2.161 2.156 1.940 θN2/deg 64.54 65.07 65.43 65.80 65.60 66.30 66.14 66.34 66.37 66.51 66.63 67.20 66.77 59.2 θO/deg 118.5 119.2 119.6 120.0 120.2 120.9 121.2 121.4 122.0 122.1 121.8 122.2 122.9 125.2 φO/deg 46.94 47.30 47.59 49.49 49.84 48.92 50.33 51.07 51.64 51.64 52.36 52.29 52.59 63.8 ψN2/deg 31.55 31.07 33.88 35.76 34.44 33.77 36.07 37.67 37.07 36.57 37.05 37.62 37.05 43.34 ψO/deg 62.10 60.90 61.47 63.27 62.30 61.87 61.87 62.88 62.58 63.00 64.07 64.21 62.99 47.64
Figure 2. Overview absorption spectrum of Sm(trensal) at 10 K in σ polarization.
Figure 3. Overview luminescence spectrum of Sm(trensal) at 10 K and room temperature. H denotes hot bands present in the room-temperature spectrum only.
f-f transition energies of Ln(trensal) and their assignments are given in Table 2. Spectra were not obtained for Ln(trensal) where Ln ) Ce, Gd, or Yb. The f7 Gd complex does not have absorption bands in the 0-25000 cm-1 optical window of the ligand, while the small number of transitions in the f1 Ce and f13 Yb complexes, compared to the number of fitting parameters required for the C3 ligand field, precludes an analysis. Assignment of the Spectra. In C3 symmetry the energy levels of the Ln(trensal) complexes are split into two sets of irreducible representations. For even-electron systems these are the singly and doubly degenerate irreducible representations Γ1 and (Γ2, Γ3), respectively, using Bethe Γ notation.15 For odd-electron systems they are the Kramers doublets (Γ4, Γ5) and (Γ6, Γ6). In what follows we will refer to these as
5026 Inorganic Chemistry, Vol. 41, No. 20, 2002
Γ1, Γ2,3 and Γ4,5, 2Γ6 for even- and odd-electron cases, respectively, with the understanding that Γ2,3 is doubly degenerate and that Γ4,5 and 2Γ6 are each Kramers doublets. The absorption spectra will show polarization anisotropy parallel and perpendicular to the crystallographic c axis, and the selection rules for transitions between these two sets of states for electric-dipole-allowed transitions are given in Table 3 for both even- and odd-electron cases. For all evenelectron Ln(trensal) complexes the ground state is Γ1, while for all odd-electron complexes the ground state is Γ4,5. The spectroscopy of the even-electron systems allows one to determine whether the C3 symmetry of the complex in the room-temperature crystal structure is maintained at 10 K, as the Γ2,3 state will split in lower symmetry while states of other irreducible representations will not. In no cases was a splitting observed that could be attributed to a lowering of the C3 symmetry of the room-temperature structure. Figures 4 and 5 show a more detailed representative spectrum in the IR and visible regions, this time for Nd(trensal). Here the intensities of the calculated peaks have been obtained using the intensity parameters found for the Er(trensal) complex without change.8 While some polarization information is evident, it is not clear-cut as implied by the selection rules in Table 3, which may indicate that a phase change occurs at low temperature. Some complexes (Tb and Ho) have lowenergy excited states (
